From 16d990a15491cf76cd6eef0846e1b4100e63261a Mon Sep 17 00:00:00 2001 From: Junrui Luo Date: Wed, 15 Apr 2026 17:26:55 +0800 Subject: [PATCH 001/333] KVM: s390: pci: fix GAIT table indexing due to double-scaling pointer arithmetic kvm_s390_pci_aif_enable(), kvm_s390_pci_aif_disable(), and aen_host_forward() index the GAIT by manually multiplying the index with sizeof(struct zpci_gaite). Since aift->gait is already a struct zpci_gaite pointer, this double-scales the offset, accessing element aisb*16 instead of aisb. This causes out-of-bounds accesses when aisb >= 32 (with ZPCI_NR_DEVICES=512) Fix by removing the erroneous sizeof multiplication. Fixes: 3c5a1b6f0a18 ("KVM: s390: pci: provide routines for enabling/disabling interrupt forwarding") Fixes: 73f91b004321 ("KVM: s390: pci: enable host forwarding of Adapter Event Notifications") Reported-by: Yuhao Jiang Cc: stable@vger.kernel.org Signed-off-by: Junrui Luo Reviewed-by: Christian Borntraeger Reviewed-by: Matthew Rosato Tested-by: Matthew Rosato Signed-off-by: Christian Borntraeger --- arch/s390/kvm/interrupt.c | 3 +-- arch/s390/kvm/pci.c | 6 ++---- 2 files changed, 3 insertions(+), 6 deletions(-) diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c index 7cb8ce833b62..f48f25c7dc8f 100644 --- a/arch/s390/kvm/interrupt.c +++ b/arch/s390/kvm/interrupt.c @@ -3307,8 +3307,7 @@ static void aen_host_forward(unsigned long si) struct zpci_gaite *gaite; struct kvm *kvm; - gaite = (struct zpci_gaite *)aift->gait + - (si * sizeof(struct zpci_gaite)); + gaite = aift->gait + si; if (gaite->count == 0) return; if (gaite->aisb != 0) diff --git a/arch/s390/kvm/pci.c b/arch/s390/kvm/pci.c index 86d93e8dddae..eed45af1a92d 100644 --- a/arch/s390/kvm/pci.c +++ b/arch/s390/kvm/pci.c @@ -290,8 +290,7 @@ static int kvm_s390_pci_aif_enable(struct zpci_dev *zdev, struct zpci_fib *fib, phys_to_virt(fib->fmt0.aibv)); spin_lock_irq(&aift->gait_lock); - gaite = (struct zpci_gaite *)aift->gait + (zdev->aisb * - sizeof(struct zpci_gaite)); + gaite = aift->gait + zdev->aisb; /* If assist not requested, host will get all alerts */ if (assist) @@ -357,8 +356,7 @@ static int kvm_s390_pci_aif_disable(struct zpci_dev *zdev, bool force) if (zdev->kzdev->fib.fmt0.aibv == 0) goto out; spin_lock_irq(&aift->gait_lock); - gaite = (struct zpci_gaite *)aift->gait + (zdev->aisb * - sizeof(struct zpci_gaite)); + gaite = aift->gait + zdev->aisb; isc = gaite->gisc; gaite->count--; if (gaite->count == 0) { From 846c76ecc02973b05ae909dd4248c11bfa277fc1 Mon Sep 17 00:00:00 2001 From: KaFai Wan Date: Tue, 21 Apr 2026 23:58:01 +0800 Subject: [PATCH 002/333] bpf: Reject TCP_NODELAY in TCP header option callbacks A BPF_SOCK_OPS program can enable BPF_SOCK_OPS_WRITE_HDR_OPT_CB_FLAG and then call bpf_setsockopt(TCP_NODELAY) from BPF_SOCK_OPS_HDR_OPT_LEN_CB or BPF_SOCK_OPS_WRITE_HDR_OPT_CB. In these callbacks, bpf_setsockopt(TCP_NODELAY) can reach __tcp_sock_set_nodelay(), which can call tcp_push_pending_frames(). >From BPF_SOCK_OPS_HDR_OPT_LEN_CB, tcp_push_pending_frames() can call tcp_current_mss(), which calls tcp_established_options() and re-enters bpf_skops_hdr_opt_len(). BPF_SOCK_OPS_HDR_OPT_LEN_CB -> bpf_setsockopt(TCP_NODELAY) -> tcp_push_pending_frames() -> tcp_current_mss() -> tcp_established_options() -> bpf_skops_hdr_opt_len() -> BPF_SOCK_OPS_HDR_OPT_LEN_CB >From BPF_SOCK_OPS_WRITE_HDR_OPT_CB, tcp_push_pending_frames() can call tcp_write_xmit(), which calls tcp_transmit_skb(). That path recomputes header option length through tcp_established_options() and bpf_skops_hdr_opt_len() before re-entering bpf_skops_write_hdr_opt(). BPF_SOCK_OPS_WRITE_HDR_OPT_CB -> bpf_setsockopt(TCP_NODELAY) -> tcp_push_pending_frames() -> tcp_write_xmit() -> tcp_transmit_skb() -> tcp_established_options() -> bpf_skops_hdr_opt_len() -> bpf_skops_write_hdr_opt() -> BPF_SOCK_OPS_WRITE_HDR_OPT_CB This leads to unbounded recursion and can overflow the kernel stack. Reject TCP_NODELAY with -EOPNOTSUPP in bpf_sock_ops_setsockopt() when bpf_setsockopt() is called from BPF_SOCK_OPS_HDR_OPT_LEN_CB or BPF_SOCK_OPS_WRITE_HDR_OPT_CB. Fixes: 7e41df5dbba2 ("bpf: Add a few optnames to bpf_setsockopt") Closes: https://lore.kernel.org/bpf/d1d523c9-6901-4454-a183-94462b8f3e4e@std.uestc.edu.cn/ Reported-by: Quan Sun <2022090917019@std.uestc.edu.cn> Reported-by: Yinhao Hu Reported-by: Kaiyan Mei Signed-off-by: KaFai Wan Signed-off-by: Martin KaFai Lau Reviewed-by: Jiayuan Chen Link: https://patch.msgid.link/20260421155804.135786-2-kafai.wan@linux.dev --- net/core/filter.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/core/filter.c b/net/core/filter.c index 5fa9189eb772..96849f4c1fbc 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -5833,6 +5833,12 @@ BPF_CALL_5(bpf_sock_ops_setsockopt, struct bpf_sock_ops_kern *, bpf_sock, if (!is_locked_tcp_sock_ops(bpf_sock)) return -EOPNOTSUPP; + /* TCP_NODELAY triggers tcp_push_pending_frames() and re-enters these callbacks. */ + if ((bpf_sock->op == BPF_SOCK_OPS_HDR_OPT_LEN_CB || + bpf_sock->op == BPF_SOCK_OPS_WRITE_HDR_OPT_CB) && + level == SOL_TCP && optname == TCP_NODELAY) + return -EOPNOTSUPP; + return _bpf_setsockopt(bpf_sock->sk, level, optname, optval, optlen); } From 54377fcab51f6f1f8807827d3751be42279e1a6a Mon Sep 17 00:00:00 2001 From: KaFai Wan Date: Tue, 21 Apr 2026 23:58:02 +0800 Subject: [PATCH 003/333] bpf: Reject TCP_NODELAY in bpf-tcp-cc A BPF TCP congestion control program can call bpf_setsockopt() from its callbacks. In current kernels, if it calls bpf_setsockopt(TCP_NODELAY) from cwnd_event_tx_start(), the call can re-enter the TCP transmit path before the outer tcp_transmit_skb() has completed and advanced the send head. This can re-trigger CA_EVENT_TX_START and lead to unbounded recursion: tcp_transmit_skb() -> tcp_event_data_sent() -> tcp_ca_event(sk, CA_EVENT_TX_START) -> cwnd_event_tx_start() -> bpf_setsockopt(TCP_NODELAY) -> tcp_push_pending_frames() -> tcp_write_xmit() -> tcp_transmit_skb() This leads to unbounded recursion and can overflow the kernel stack. Reject TCP_NODELAY with -EOPNOTSUPP for bpf-tcp-cc by introducing a dedicated setsockopt proto for BPF_PROG_TYPE_STRUCT_OPS TCP congestion control programs. To keep it simple, all tcp-cc ops is rejected for TCP_NODELAY. Fixes: 7e41df5dbba2 ("bpf: Add a few optnames to bpf_setsockopt") Suggested-by: Martin KaFai Lau Signed-off-by: KaFai Wan Signed-off-by: Martin KaFai Lau Reviewed-by: Jiayuan Chen Link: https://patch.msgid.link/20260421155804.135786-3-kafai.wan@linux.dev --- include/linux/bpf.h | 1 + net/core/filter.c | 24 ++++++++++++++++++++++++ net/ipv4/bpf_tcp_ca.c | 2 +- 3 files changed, 26 insertions(+), 1 deletion(-) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index b4b703c90ca9..01e203964892 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -3725,6 +3725,7 @@ extern const struct bpf_func_proto bpf_for_each_map_elem_proto; extern const struct bpf_func_proto bpf_btf_find_by_name_kind_proto; extern const struct bpf_func_proto bpf_sk_setsockopt_proto; extern const struct bpf_func_proto bpf_sk_getsockopt_proto; +extern const struct bpf_func_proto bpf_sk_setsockopt_nodelay_proto; extern const struct bpf_func_proto bpf_unlocked_sk_setsockopt_proto; extern const struct bpf_func_proto bpf_unlocked_sk_getsockopt_proto; extern const struct bpf_func_proto bpf_find_vma_proto; diff --git a/net/core/filter.c b/net/core/filter.c index 96849f4c1fbc..2914f5330310 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -5688,6 +5688,30 @@ const struct bpf_func_proto bpf_sk_getsockopt_proto = { .arg5_type = ARG_CONST_SIZE, }; +BPF_CALL_5(bpf_sk_setsockopt_nodelay, struct sock *, sk, int, level, + int, optname, char *, optval, int, optlen) +{ + /* + * TCP_NODELAY triggers tcp_push_pending_frames() and re-enters + * CA_EVENT_TX_START in bpf_tcp_cc. + */ + if (level == SOL_TCP && optname == TCP_NODELAY) + return -EOPNOTSUPP; + + return _bpf_setsockopt(sk, level, optname, optval, optlen); +} + +const struct bpf_func_proto bpf_sk_setsockopt_nodelay_proto = { + .func = bpf_sk_setsockopt_nodelay, + .gpl_only = false, + .ret_type = RET_INTEGER, + .arg1_type = ARG_PTR_TO_BTF_ID_SOCK_COMMON, + .arg2_type = ARG_ANYTHING, + .arg3_type = ARG_ANYTHING, + .arg4_type = ARG_PTR_TO_MEM | MEM_RDONLY, + .arg5_type = ARG_CONST_SIZE, +}; + BPF_CALL_5(bpf_unlocked_sk_setsockopt, struct sock *, sk, int, level, int, optname, char *, optval, int, optlen) { diff --git a/net/ipv4/bpf_tcp_ca.c b/net/ipv4/bpf_tcp_ca.c index 008edc7f6688..791e15063237 100644 --- a/net/ipv4/bpf_tcp_ca.c +++ b/net/ipv4/bpf_tcp_ca.c @@ -168,7 +168,7 @@ bpf_tcp_ca_get_func_proto(enum bpf_func_id func_id, */ if (prog_ops_moff(prog) != offsetof(struct tcp_congestion_ops, release)) - return &bpf_sk_setsockopt_proto; + return &bpf_sk_setsockopt_nodelay_proto; return NULL; case BPF_FUNC_getsockopt: /* Since get/setsockopt is usually expected to From 52b6b5334924d8f083a2abe8edeface9206e13ee Mon Sep 17 00:00:00 2001 From: KaFai Wan Date: Tue, 21 Apr 2026 23:58:03 +0800 Subject: [PATCH 004/333] selftests/bpf: Test TCP_NODELAY in TCP hdr opt callbacks Add a sockops selftest for the TCP_NODELAY restriction in BPF_SOCK_OPS_HDR_OPT_LEN_CB and BPF_SOCK_OPS_WRITE_HDR_OPT_CB. With BPF_SOCK_OPS_WRITE_HDR_OPT_CB_FLAG enabled, bpf_setsockopt(TCP_NODELAY) returns -EOPNOTSUPP from BPF_SOCK_OPS_HDR_OPT_LEN_CB and BPF_SOCK_OPS_WRITE_HDR_OPT_CB, avoiding unbounded recursion and kernel stack overflow. Other cases continue to work as before, including BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB. Signed-off-by: KaFai Wan Signed-off-by: Martin KaFai Lau Reviewed-by: Jiayuan Chen Link: https://patch.msgid.link/20260421155804.135786-4-kafai.wan@linux.dev --- .../selftests/bpf/prog_tests/tcp_hdr_options.c | 4 ++++ .../bpf/progs/test_misc_tcp_hdr_options.c | 15 ++++++++++++++- 2 files changed, 18 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/bpf/prog_tests/tcp_hdr_options.c b/tools/testing/selftests/bpf/prog_tests/tcp_hdr_options.c index 56685fc03c7e..80e6315da2a5 100644 --- a/tools/testing/selftests/bpf/prog_tests/tcp_hdr_options.c +++ b/tools/testing/selftests/bpf/prog_tests/tcp_hdr_options.c @@ -507,6 +507,10 @@ static void misc(void) ASSERT_EQ(misc_skel->bss->nr_hwtstamp, 0, "nr_hwtstamp"); + ASSERT_TRUE(misc_skel->bss->nodelay_est_ok, "nodelay_est_ok"); + ASSERT_TRUE(misc_skel->bss->nodelay_hdr_len_reject, "nodelay_hdr_len_reject"); + ASSERT_TRUE(misc_skel->bss->nodelay_write_hdr_reject, "nodelay_write_hdr_reject"); + check_linum: ASSERT_FALSE(check_error_linum(&sk_fds), "check_error_linum"); sk_fds_close(&sk_fds); diff --git a/tools/testing/selftests/bpf/progs/test_misc_tcp_hdr_options.c b/tools/testing/selftests/bpf/progs/test_misc_tcp_hdr_options.c index d487153a839d..ed5a0011b863 100644 --- a/tools/testing/selftests/bpf/progs/test_misc_tcp_hdr_options.c +++ b/tools/testing/selftests/bpf/progs/test_misc_tcp_hdr_options.c @@ -29,6 +29,10 @@ unsigned int nr_syn = 0; unsigned int nr_fin = 0; unsigned int nr_hwtstamp = 0; +bool nodelay_est_ok = false; +bool nodelay_hdr_len_reject = false; +bool nodelay_write_hdr_reject = false; + /* Check the header received from the active side */ static int __check_active_hdr_in(struct bpf_sock_ops *skops, bool check_syn) { @@ -300,7 +304,7 @@ static int handle_passive_estab(struct bpf_sock_ops *skops) SEC("sockops") int misc_estab(struct bpf_sock_ops *skops) { - int true_val = 1; + int true_val = 1, false_val = 0, ret; switch (skops->op) { case BPF_SOCK_OPS_TCP_LISTEN_CB: @@ -316,10 +320,19 @@ int misc_estab(struct bpf_sock_ops *skops) case BPF_SOCK_OPS_PARSE_HDR_OPT_CB: return handle_parse_hdr(skops); case BPF_SOCK_OPS_HDR_OPT_LEN_CB: + ret = bpf_setsockopt(skops, SOL_TCP, TCP_NODELAY, &true_val, sizeof(true_val)); + if (ret == -EOPNOTSUPP) + nodelay_hdr_len_reject = true; return handle_hdr_opt_len(skops); case BPF_SOCK_OPS_WRITE_HDR_OPT_CB: + ret = bpf_setsockopt(skops, SOL_TCP, TCP_NODELAY, &true_val, sizeof(true_val)); + if (ret == -EOPNOTSUPP) + nodelay_write_hdr_reject = true; return handle_write_hdr_opt(skops); case BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB: + ret = bpf_setsockopt(skops, SOL_TCP, TCP_NODELAY, &false_val, sizeof(false_val)); + if (!ret) + nodelay_est_ok = true; return handle_passive_estab(skops); } From 2c7e33f1fc2e75fcfb4aa5d840bcd2e8b53c1847 Mon Sep 17 00:00:00 2001 From: KaFai Wan Date: Tue, 21 Apr 2026 23:58:04 +0800 Subject: [PATCH 005/333] selftests/bpf: Verify bpf-tcp-cc rejects TCP_NODELAY Add a bpf_tcp_ca selftest for the TCP_NODELAY restriction in bpf-tcp-cc. Update bpf_cubic to exercise init() and cwnd_event_tx_start(), and check that both callbacks reject bpf_setsockopt(TCP_NODELAY) with -EOPNOTSUPP. Signed-off-by: KaFai Wan Signed-off-by: Martin KaFai Lau Link: https://patch.msgid.link/20260421155804.135786-5-kafai.wan@linux.dev --- .../testing/selftests/bpf/prog_tests/bpf_tcp_ca.c | 4 ++++ tools/testing/selftests/bpf/progs/bpf_cubic.c | 14 ++++++++++++++ 2 files changed, 18 insertions(+) diff --git a/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c b/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c index f829b6f09bc9..fe30181e6336 100644 --- a/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c +++ b/tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c @@ -112,6 +112,10 @@ static void test_cubic(void) ASSERT_EQ(cubic_skel->bss->bpf_cubic_acked_called, 1, "pkts_acked called"); + ASSERT_TRUE(cubic_skel->bss->nodelay_init_reject, "init reject nodelay option"); + ASSERT_TRUE(cubic_skel->bss->nodelay_cwnd_event_tx_start_reject, + "cwnd_event_tx_start reject nodelay option"); + bpf_link__destroy(link); bpf_cubic__destroy(cubic_skel); } diff --git a/tools/testing/selftests/bpf/progs/bpf_cubic.c b/tools/testing/selftests/bpf/progs/bpf_cubic.c index ce18a4db813f..ebd5a1e69f56 100644 --- a/tools/testing/selftests/bpf/progs/bpf_cubic.c +++ b/tools/testing/selftests/bpf/progs/bpf_cubic.c @@ -16,6 +16,7 @@ #include "bpf_tracing_net.h" #include +#include char _license[] SEC("license") = "GPL"; @@ -170,10 +171,18 @@ static void bictcp_hystart_reset(struct sock *sk) ca->sample_cnt = 0; } +bool nodelay_init_reject = false; +bool nodelay_cwnd_event_tx_start_reject = false; + SEC("struct_ops") void BPF_PROG(bpf_cubic_init, struct sock *sk) { struct bpf_bictcp *ca = inet_csk_ca(sk); + int true_val = 1, ret; + + ret = bpf_setsockopt(sk, SOL_TCP, TCP_NODELAY, &true_val, sizeof(true_val)); + if (ret == -EOPNOTSUPP) + nodelay_init_reject = true; bictcp_reset(ca); @@ -189,8 +198,13 @@ void BPF_PROG(bpf_cubic_cwnd_event_tx_start, struct sock *sk) { struct bpf_bictcp *ca = inet_csk_ca(sk); __u32 now = tcp_jiffies32; + int true_val = 1, ret; __s32 delta; + ret = bpf_setsockopt(sk, SOL_TCP, TCP_NODELAY, &true_val, sizeof(true_val)); + if (ret == -EOPNOTSUPP) + nodelay_cwnd_event_tx_start_reject = true; + delta = now - tcp_sk(sk)->lsndtime; /* We were application limited (idle) for a while. From 375e4e33c18dfa05c5dfd5f3dfffeb29343dd4c7 Mon Sep 17 00:00:00 2001 From: Weiming Shi Date: Tue, 21 Apr 2026 23:54:12 -0700 Subject: [PATCH 006/333] bpf: Fix NULL pointer dereference in bpf_sk_storage_clone and diag paths bpf_selem_unlink_nofail() sets SDATA(selem)->smap to NULL before removing the selem from the storage hlist. A concurrent RCU reader in bpf_sk_storage_clone() can observe the selem still on the list with smap already NULL, causing a NULL pointer dereference. general protection fault, probably for non-canonical address 0xdffffc000000000a: KASAN: null-ptr-deref in range [0x0000000000000050-0x0000000000000057] RIP: 0010:bpf_sk_storage_clone+0x1cd/0xaa0 net/core/bpf_sk_storage.c:174 Call Trace: sk_clone+0xfed/0x1980 net/core/sock.c:2591 inet_csk_clone_lock+0x30/0x760 net/ipv4/inet_connection_sock.c:1222 tcp_create_openreq_child+0x35/0x2680 net/ipv4/tcp_minisocks.c:571 tcp_v4_syn_recv_sock+0x123/0xf90 net/ipv4/tcp_ipv4.c:1729 tcp_check_req+0x8e1/0x2580 include/net/tcp.h:855 tcp_v4_rcv+0x1845/0x3b80 net/ipv4/tcp_ipv4.c:2347 Add a NULL check for smap in bpf_sk_storage_clone(). bpf_sk_storage_diag_put_all() has the same issue. Add a NULL check and pass the validated smap directly to diag_get(), which is refactored to take smap as a parameter instead of reading it internally. bpf_sk_storage_diag_put() uses diag->maps[i] which is always valid under its refcount, so diag->maps[i] is passed directly to diag_get(). Fixes: 5d800f87d0a5 ("bpf: Support lockless unlink when freeing map or local storage") Reported-by: Xiang Mei Acked-by: Amery Hung Signed-off-by: Weiming Shi Signed-off-by: Martin KaFai Lau Link: https://patch.msgid.link/20260422065411.1007737-2-bestswngs@gmail.com --- net/core/bpf_sk_storage.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/net/core/bpf_sk_storage.c b/net/core/bpf_sk_storage.c index 14eb7812bda4..dc3e8fce8809 100644 --- a/net/core/bpf_sk_storage.c +++ b/net/core/bpf_sk_storage.c @@ -172,7 +172,7 @@ int bpf_sk_storage_clone(const struct sock *sk, struct sock *newsk) struct bpf_map *map; smap = rcu_dereference(SDATA(selem)->smap); - if (!(smap->map.map_flags & BPF_F_CLONE)) + if (!smap || !(smap->map.map_flags & BPF_F_CLONE)) continue; /* Note that for lockless listeners adding new element @@ -531,10 +531,10 @@ bpf_sk_storage_diag_alloc(const struct nlattr *nla_stgs) } EXPORT_SYMBOL_GPL(bpf_sk_storage_diag_alloc); -static int diag_get(struct bpf_local_storage_data *sdata, struct sk_buff *skb) +static int diag_get(struct bpf_local_storage_map *smap, + struct bpf_local_storage_data *sdata, struct sk_buff *skb) { struct nlattr *nla_stg, *nla_value; - struct bpf_local_storage_map *smap; /* It cannot exceed max nlattr's payload */ BUILD_BUG_ON(U16_MAX - NLA_HDRLEN < BPF_LOCAL_STORAGE_MAX_VALUE_SIZE); @@ -543,7 +543,6 @@ static int diag_get(struct bpf_local_storage_data *sdata, struct sk_buff *skb) if (!nla_stg) return -EMSGSIZE; - smap = rcu_dereference(sdata->smap); if (nla_put_u32(skb, SK_DIAG_BPF_STORAGE_MAP_ID, smap->map.id)) goto errout; @@ -596,9 +595,11 @@ static int bpf_sk_storage_diag_put_all(struct sock *sk, struct sk_buff *skb, saved_len = skb->len; hlist_for_each_entry_rcu(selem, &sk_storage->list, snode) { smap = rcu_dereference(SDATA(selem)->smap); + if (!smap) + continue; diag_size += nla_value_size(smap->map.value_size); - if (nla_stgs && diag_get(SDATA(selem), skb)) + if (nla_stgs && diag_get(smap, SDATA(selem), skb)) /* Continue to learn diag_size */ err = -EMSGSIZE; } @@ -665,7 +666,7 @@ int bpf_sk_storage_diag_put(struct bpf_sk_storage_diag *diag, diag_size += nla_value_size(diag->maps[i]->value_size); - if (nla_stgs && diag_get(sdata, skb)) + if (nla_stgs && diag_get((struct bpf_local_storage_map *)diag->maps[i], sdata, skb)) /* Continue to learn diag_size */ err = -EMSGSIZE; } From 6451d58a355642b612f2bf948ad39108c998ac2a Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Mon, 20 Apr 2026 19:48:41 +0000 Subject: [PATCH 007/333] sockmap: Fix sk_psock_drop() race vs sock_map_{unhash,close,destroy}(). syzbot reported a splat in sock_map_destroy() [0], where psock was NULL even though sk->sk_prot still pointed to tcp_bpf_prots[][]. The stack trace shows how badly the path was excercised, see inet_release() calls tcp_close(), not sock_map_close() yet, but finally reaching sock_map_destroy(). The root cause is a lack of synchronisation. Even if sk_psock_get() fails to bump psock->refcnt, it does not guarantee that sk_psock_drop() has finished, and thus sk->sk_prot might not have been restored to the original one. Commit 4b4647add7d3 ("sock_map: avoid race between sock_map_close and sk_psock_put") attempted to address this, but it was insufficient for two reasons. It did not cover sock_map_unhash() and sock_map_destroy(), and it missed the corner case where sk_psock() is NULL. On non-x86 platforms, sk_psock_restore_proto(sk, psock) and rcu_assign_sk_user_data(sk, NULL) can be reordered because there is no address dependency between sk->sk_prot and sk->sk_user_data. sk_psock_get() returning NULL implies nothing about sk->sk_prot. Let's simply retry sk_psock_get() in the unlikely case. Note that we cannot avoid loop even if we added memory barrier in sk_psock_drop() and sock_map_psock_get_checked(). Also note that sock_map_destroy() cannot be called from softirq while sock_map_close() has also been running. It is because sock_map_destroy() requires SOCK_DEAD, so sock_map_destroy() cannot happen until sock_map_close() has finished the saved_close() (which is tcp_close()). [0]: WARNING: CPU: 1 PID: 8459 at net/core/sock_map.c:1667 sock_map_destroy+0x28b/0x2b0 net/core/sock_map.c:1667 Modules linked in: CPU: 1 UID: 0 PID: 8459 Comm: syz.0.1109 Not tainted syzkaller #0 PREEMPT_{RT,(full)} Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/12/2025 RIP: 0010:sock_map_destroy+0x28b/0x2b0 net/core/sock_map.c:1667 Code: 8b 36 49 83 c6 38 4c 89 f0 48 c1 e8 03 42 80 3c 38 00 74 08 4c 89 f7 e8 93 62 22 f9 4d 8b 3e e9 79 ff ff ff e8 a6 2b c3 f8 90 <0f> 0b 90 eb 9c e8 9b 2b c3 f8 4c 89 e7 be 03 00 00 00 e8 0e 4e bc RSP: 0018:ffffc9000d067be8 EFLAGS: 00010293 RAX: ffffffff88fb30aa RBX: ffff888024832000 RCX: ffff888024283b80 RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000 RBP: 0000000000000001 R08: 0000000000000000 R09: 0000000000000000 R10: dffffc0000000000 R11: ffffed100862e946 R12: dffffc0000000000 R13: ffff888024832000 R14: ffffffff995b2208 R15: ffffffff88fb2e20 FS: 0000555579a7d500(0000) GS:ffff8881269c2000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00002000000048c0 CR3: 000000003713a000 CR4: 00000000003526f0 Call Trace: inet_csk_destroy_sock+0x166/0x3a0 net/ipv4/inet_connection_sock.c:1294 __tcp_close+0xcc1/0xfd0 net/ipv4/tcp.c:3262 tcp_close+0x28/0x110 net/ipv4/tcp.c:3274 inet_release+0x144/0x190 net/ipv4/af_inet.c:435 __sock_release net/socket.c:649 [inline] sock_close+0xc0/0x240 net/socket.c:1439 __fput+0x45b/0xa80 fs/file_table.c:468 task_work_run+0x1d4/0x260 kernel/task_work.c:227 resume_user_mode_work include/linux/resume_user_mode.h:50 [inline] exit_to_user_mode_loop+0xec/0x110 kernel/entry/common.c:43 exit_to_user_mode_prepare include/linux/irq-entry-common.h:225 [inline] syscall_exit_to_user_mode_work include/linux/entry-common.h:175 [inline] syscall_exit_to_user_mode include/linux/entry-common.h:210 [inline] do_syscall_64+0x2bd/0x3b0 arch/x86/entry/syscall_64.c:100 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0033:0x7f265847ebe9 Code: ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 a8 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007ffd158dfbd8 EFLAGS: 00000246 ORIG_RAX: 00000000000001b4 RAX: 0000000000000000 RBX: 000000000002ddb0 RCX: 00007f265847ebe9 RDX: 0000000000000000 RSI: 000000000000001e RDI: 0000000000000003 RBP: 00007f26586a7da0 R08: 0000000000000001 R09: 0000000e158dfecf R10: 0000001b30a20000 R11: 0000000000000246 R12: 00007f26586a5fac R13: 00007f26586a5fa0 R14: ffffffffffffffff R15: 00007ffd158dfcf0 Fixes: 1aa12bdf1bfb ("bpf: sockmap, add sock close() hook to remove socks") Fixes: b05545e15e1f ("bpf: sockmap, fix transition through disconnect without close") Fixes: d8616ee2affc ("bpf, sockmap: Fix sk->sk_forward_alloc warn_on in sk_stream_kill_queues") Reported-by: syzbot+b0842d38af58376d1fdc@syzkaller.appspotmail.com Closes: https://lore.kernel.org/bpf/69cec5ef.050a0220.2dbe29.0009.GAE@google.com/ Signed-off-by: Kuniyuki Iwashima Signed-off-by: Martin KaFai Lau Reviewed-by: Jiayuan Chen Link: https://patch.msgid.link/20260420194846.1089595-1-kuniyu@google.com --- net/core/sock_map.c | 39 +++++++++++++++++++++++++-------------- 1 file changed, 25 insertions(+), 14 deletions(-) diff --git a/net/core/sock_map.c b/net/core/sock_map.c index 02a68be3002a..99e3789492a0 100644 --- a/net/core/sock_map.c +++ b/net/core/sock_map.c @@ -1630,18 +1630,23 @@ void sock_map_unhash(struct sock *sk) void (*saved_unhash)(struct sock *sk); struct sk_psock *psock; +retry: rcu_read_lock(); psock = sk_psock(sk); if (unlikely(!psock)) { rcu_read_unlock(); saved_unhash = READ_ONCE(sk->sk_prot)->unhash; + if (unlikely(saved_unhash == sock_map_unhash)) + goto retry; } else { saved_unhash = psock->saved_unhash; sock_map_remove_links(sk, psock); rcu_read_unlock(); + + if (WARN_ON_ONCE(saved_unhash == sock_map_unhash)) + return; } - if (WARN_ON_ONCE(saved_unhash == sock_map_unhash)) - return; + if (saved_unhash) saved_unhash(sk); } @@ -1652,20 +1657,25 @@ void sock_map_destroy(struct sock *sk) void (*saved_destroy)(struct sock *sk); struct sk_psock *psock; +retry: rcu_read_lock(); psock = sk_psock_get(sk); if (unlikely(!psock)) { rcu_read_unlock(); saved_destroy = READ_ONCE(sk->sk_prot)->destroy; + if (unlikely(saved_destroy == sock_map_destroy)) + goto retry; } else { saved_destroy = psock->saved_destroy; sock_map_remove_links(sk, psock); rcu_read_unlock(); sk_psock_stop(psock); sk_psock_put(sk, psock); + + if (WARN_ON_ONCE(saved_destroy == sock_map_destroy)) + return; } - if (WARN_ON_ONCE(saved_destroy == sock_map_destroy)) - return; + if (saved_destroy) saved_destroy(sk); } @@ -1676,32 +1686,33 @@ void sock_map_close(struct sock *sk, long timeout) void (*saved_close)(struct sock *sk, long timeout); struct sk_psock *psock; +retry: lock_sock(sk); rcu_read_lock(); - psock = sk_psock(sk); + psock = sk_psock_get(sk); if (likely(psock)) { saved_close = psock->saved_close; sock_map_remove_links(sk, psock); - psock = sk_psock_get(sk); - if (unlikely(!psock)) - goto no_psock; rcu_read_unlock(); sk_psock_stop(psock); release_sock(sk); cancel_delayed_work_sync(&psock->work); sk_psock_put(sk, psock); + + /* Make sure we do not recurse. This is a bug. + * Leak the socket instead of crashing on a stack overflow. + */ + if (WARN_ON_ONCE(saved_close == sock_map_close)) + return; } else { saved_close = READ_ONCE(sk->sk_prot)->close; -no_psock: rcu_read_unlock(); release_sock(sk); + + if (unlikely(saved_close == sock_map_close)) + goto retry; } - /* Make sure we do not recurse. This is a bug. - * Leak the socket instead of crashing on a stack overflow. - */ - if (WARN_ON_ONCE(saved_close == sock_map_close)) - return; saved_close(sk, timeout); } EXPORT_SYMBOL_GPL(sock_map_close); From 1081de1accb2b224516cca7071122c59532d0b22 Mon Sep 17 00:00:00 2001 From: Weiming Shi Date: Thu, 23 Apr 2026 11:38:32 -0700 Subject: [PATCH 008/333] bpf: Fix NULL pointer dereference in bpf_skb_fib_lookup() When tot_len is not provided by the user, bpf_skb_fib_lookup() resolves the FIB result's output device via dev_get_by_index_rcu() to check skb forwardability and fill in mtu_result. The returned pointer is dereferenced without a NULL check. If the device is concurrently unregistered, dev_get_by_index_rcu() returns NULL and is_skb_forwardable() crashes at dev->flags: KASAN: null-ptr-deref in range [0x00000000000000b0-0x00000000000000b7] Call Trace: is_skb_forwardable (include/linux/netdevice.h:4365) bpf_skb_fib_lookup (net/core/filter.c:6446) bpf_prog_test_run_skb (net/bpf/test_run.c) __sys_bpf (kernel/bpf/syscall.c) Add the missing NULL check, returning -ENODEV to be consistent with how bpf_ipv4_fib_lookup() and bpf_ipv6_fib_lookup() handle the same condition. Fixes: 4f74fede40df ("bpf: Add mtu checking to FIB forwarding helper") Reported-by: Xiang Mei Signed-off-by: Weiming Shi Signed-off-by: Martin KaFai Lau Acked-by: Paul Chaignon Link: https://patch.msgid.link/20260423183831.1325480-2-bestswngs@gmail.com --- net/core/filter.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/core/filter.c b/net/core/filter.c index 2914f5330310..bc96c18df4e0 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -6473,6 +6473,8 @@ BPF_CALL_4(bpf_skb_fib_lookup, struct sk_buff *, skb, * against MTU of FIB lookup resulting net_device */ dev = dev_get_by_index_rcu(net, params->ifindex); + if (unlikely(!dev)) + return -ENODEV; if (!is_skb_forwardable(dev, skb)) rc = BPF_FIB_LKUP_RET_FRAG_NEEDED; From b5c111f4967ba4fdecdd318923ec7b081e9ef95f Mon Sep 17 00:00:00 2001 From: Amery Hung Date: Thu, 23 Apr 2026 15:23:55 -0700 Subject: [PATCH 009/333] bpf: Fix sk_local_storage diag dumping uninitialized special fields Call check_and_init_map_value() after the copy_map_value() to zero out special field regions. diag_get() copies sk_local_storage map values into a netlink message using copy_map_value{_locked}(), which intentionally skip special fields. However, the destination buffer from nla_reserve_64bit() is not zeroed and the skipped regions contain uninitialized skb data can be sent to userspace. Fixes: 1ed4d92458a9 ("bpf: INET_DIAG support in bpf_sk_storage") Signed-off-by: Amery Hung Signed-off-by: Martin KaFai Lau Link: https://patch.msgid.link/20260423222356.155387-1-ameryhung@gmail.com --- net/core/bpf_sk_storage.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/core/bpf_sk_storage.c b/net/core/bpf_sk_storage.c index dc3e8fce8809..ecd659f79fd4 100644 --- a/net/core/bpf_sk_storage.c +++ b/net/core/bpf_sk_storage.c @@ -557,6 +557,7 @@ static int diag_get(struct bpf_local_storage_map *smap, sdata->data, true); else copy_map_value(&smap->map, nla_data(nla_value), sdata->data); + check_and_init_map_value(&smap->map, nla_data(nla_value)); nla_nest_end(skb, nla_stg); return 0; From 5f69165b7e4215f02247b0c64052c71b2f66d73a Mon Sep 17 00:00:00 2001 From: "Mukesh Kumar Chaurasiya (IBM)" Date: Sun, 26 Apr 2026 15:17:25 +0530 Subject: [PATCH 010/333] rust/drm: import ARef from sync crate ARef is defined in sync and is getting used from types causing the build to fail. Fix this by using ARef from sync module. Fixes: 80df573af9ef ("rust: drm: gem: shmem: Add DRM shmem helper abstraction") Signed-off-by: Mukesh Kumar Chaurasiya (IBM) Link: https://patch.msgid.link/20260426094725.2188668-2-mkchauras@gmail.com [ Add missing Fixes: tag. - Danilo ] Signed-off-by: Danilo Krummrich --- rust/kernel/drm/gem/shmem.rs | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/rust/kernel/drm/gem/shmem.rs b/rust/kernel/drm/gem/shmem.rs index d025fb035195..e1b648920d2f 100644 --- a/rust/kernel/drm/gem/shmem.rs +++ b/rust/kernel/drm/gem/shmem.rs @@ -19,10 +19,8 @@ }, error::to_result, prelude::*, - types::{ - ARef, - Opaque, // - }, // + sync::aref::ARef, + types::Opaque, // }; use core::{ ops::{ From 15e8bae5d930c91b8739a87d75db0a6efca3cb32 Mon Sep 17 00:00:00 2001 From: Danilo Krummrich Date: Thu, 23 Apr 2026 14:46:35 +0200 Subject: [PATCH 011/333] MAINTAINERS: nova: update mailing list The nouveau mailing list has some issues (e.g. with stripping Cc entries from replies when using notmuch + b4 based workflows). Besides that, having a separate mailing list for nova also helps to better distinguish nova from nouveau and makes it easier to track nova-specific discussions. Replace the nouveau mailing list with the new nova-gpu@lists.linux.dev mailing list for both nova-core and nova-drm, and remove the patchwork entries, since those are bound to the nouveau mailing list and not used by nova anyway. Link: https://lore.kernel.org/all/bc2517c2-6772-4cbd-8fd7-6dbdcdd13eab@nvidia.com/ Reviewed-by: Joel Fernandes Reviewed-by: John Hubbard Acked-by: Alexandre Courbot Link: https://patch.msgid.link/20260423124649.38793-1-dakr@kernel.org Signed-off-by: Danilo Krummrich --- MAINTAINERS | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index 2fb1c75afd16..5c9272622033 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -8193,10 +8193,9 @@ F: include/uapi/drm/nouveau_drm.h CORE DRIVER FOR NVIDIA GPUS [RUST] M: Danilo Krummrich M: Alexandre Courbot -L: nouveau@lists.freedesktop.org +L: nova-gpu@lists.linux.dev S: Supported W: https://rust-for-linux.com/nova-gpu-driver -Q: https://patchwork.freedesktop.org/project/nouveau/ B: https://gitlab.freedesktop.org/drm/nova/-/issues C: irc://irc.oftc.net/nouveau T: git https://gitlab.freedesktop.org/drm/rust/kernel.git drm-rust-next @@ -8205,10 +8204,9 @@ F: drivers/gpu/nova-core/ DRM DRIVER FOR NVIDIA GPUS [RUST] M: Danilo Krummrich -L: nouveau@lists.freedesktop.org +L: nova-gpu@lists.linux.dev S: Supported W: https://rust-for-linux.com/nova-gpu-driver -Q: https://patchwork.freedesktop.org/project/nouveau/ B: https://gitlab.freedesktop.org/drm/nova/-/issues C: irc://irc.oftc.net/nouveau T: git https://gitlab.freedesktop.org/drm/rust/kernel.git drm-rust-next From 0cfe660559e857d7c00ab86c73e4510ce069086f Mon Sep 17 00:00:00 2001 From: Matthew Rosato Date: Fri, 24 Apr 2026 15:39:00 -0400 Subject: [PATCH 012/333] KVM: s390: pci: Fix aisb calculation The current implementation of aisb calculation will erroneously index via an unsigned long * as well as multiply by 8B for every 64-bits in the offset; only one or the other is required. This throws off aisb calculations once the number of devices exceeds 64, and can result in out-of-bounds access as well as failure to indicate summary bits associated with those devices in guests. Fix this by converting to a physical address before applying the offset, as is already done in arch/s390/pci/pci_irq.c. Fixes: 3c5a1b6f0a18 ("KVM: s390: pci: provide routines for enabling/disabling interrupt forwarding") Signed-off-by: Matthew Rosato Reviewed-by: Niklas Schnelle Signed-off-by: Christian Borntraeger --- arch/s390/kvm/pci.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/s390/kvm/pci.c b/arch/s390/kvm/pci.c index eed45af1a92d..5b075c38998e 100644 --- a/arch/s390/kvm/pci.c +++ b/arch/s390/kvm/pci.c @@ -166,7 +166,7 @@ static int kvm_zpci_set_airq(struct zpci_dev *zdev) fib.fmt0.noi = airq_iv_end(zdev->aibv); fib.fmt0.aibv = virt_to_phys(zdev->aibv->vector); fib.fmt0.aibvo = 0; - fib.fmt0.aisb = virt_to_phys(aift->sbv->vector + (zdev->aisb / 64) * 8); + fib.fmt0.aisb = virt_to_phys(aift->sbv->vector) + (zdev->aisb / 64) * 8; fib.fmt0.aisbo = zdev->aisb & 63; fib.gd = zdev->gisa; @@ -308,7 +308,7 @@ static int kvm_s390_pci_aif_enable(struct zpci_dev *zdev, struct zpci_fib *fib, /* Update guest FIB for re-issue */ fib->fmt0.aisbo = zdev->aisb & 63; - fib->fmt0.aisb = virt_to_phys(aift->sbv->vector + (zdev->aisb / 64) * 8); + fib->fmt0.aisb = virt_to_phys(aift->sbv->vector) + (zdev->aisb / 64) * 8; fib->fmt0.isc = gisc; /* Save some guest fib values in the host for later use */ From 58c0ac6125d89bf6ec65a521eaeb52a0e8e20a9f Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Mon, 20 Apr 2026 08:42:03 +0000 Subject: [PATCH 013/333] iommu/amd: Use maximum Event log buffer size when SNP is enabled on Family 0x19 Due to CVE-2023-20585, the Event log buffer must use the maximum supported size (512K) on Milan/Genoa (Family 0x19) systems when SNP is enabled, to mitigate a potential security vulnerability. All other systems continue to use the default Event log buffer size (8K). Apply the errata fix by making the following changes: * Introduce new global variable (amd_iommu_evtlog_size) to have event log buffer size. Adjust variable size for family 0x19. * Since 'iommu_snp_enable()' must be called after the core IOMMU subsystem is initialized, it cannot be moved to the early init stage. The SNP errata must also be applied after the 'iommu_snp_enable()' check. Therefore, 'alloc_event_buffer()' and 'iommu_enable_event_buffer()' are now called in the IOMMU_ENABLED state, after the errata is applied. * Adjust alloc_event_buffer() and iommu_enable_event_buffer() to handle all IOMMU instances. * Also rename EVT_* macros to make it more readable. Link: https://www.amd.com/en/resources/product-security/bulletin/amd-sb-3016.html Cc: Borislav Petkov Cc: Suravee Suthikulpanit Cc: Joerg Roedel Signed-off-by: Vasant Hegde Tested-by: Dheeraj Kumar Srivastava Signed-off-by: Joerg Roedel --- drivers/iommu/amd/amd_iommu.h | 2 + drivers/iommu/amd/amd_iommu_types.h | 10 ++- drivers/iommu/amd/init.c | 120 +++++++++++++++++++--------- drivers/iommu/amd/iommu.c | 2 +- 4 files changed, 91 insertions(+), 43 deletions(-) diff --git a/drivers/iommu/amd/amd_iommu.h b/drivers/iommu/amd/amd_iommu.h index 1342e764a548..f1c486dcf0f3 100644 --- a/drivers/iommu/amd/amd_iommu.h +++ b/drivers/iommu/amd/amd_iommu.h @@ -11,6 +11,8 @@ #include "amd_iommu_types.h" +extern int amd_iommu_evtlog_size; + irqreturn_t amd_iommu_int_thread(int irq, void *data); irqreturn_t amd_iommu_int_thread_evtlog(int irq, void *data); irqreturn_t amd_iommu_int_thread_pprlog(int irq, void *data); diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h index c685d3771436..c3430c09bc5c 100644 --- a/drivers/iommu/amd/amd_iommu_types.h +++ b/drivers/iommu/amd/amd_iommu_types.h @@ -15,6 +15,7 @@ #include #include #include +#include #include #include #include @@ -141,7 +142,6 @@ #define MMIO_STATUS_GALOG_INT_MASK BIT(10) /* event logging constants */ -#define EVENT_ENTRY_SIZE 0x10 #define EVENT_TYPE_SHIFT 28 #define EVENT_TYPE_MASK 0xf #define EVENT_TYPE_ILL_DEV 0x1 @@ -259,8 +259,12 @@ #define MMIO_CMD_BUFFER_TAIL(x) FIELD_GET(MMIO_CMD_TAIL_MASK, (x)) /* constants for event buffer handling */ -#define EVT_BUFFER_SIZE 8192 /* 512 entries */ -#define EVT_LEN_MASK (0x9ULL << 56) +#define EVTLOG_ENTRY_SIZE 0x10 +#define EVTLOG_SIZE_SHIFT 56 +#define EVTLOG_SIZE_DEF SZ_8K /* 512 entries */ +#define EVTLOG_LEN_MASK_DEF (0x9ULL << EVTLOG_SIZE_SHIFT) +#define EVTLOG_SIZE_MAX SZ_512K /* 32K entries */ +#define EVTLOG_LEN_MASK_MAX (0xFULL << EVTLOG_SIZE_SHIFT) /* Constants for PPR Log handling */ #define PPR_LOG_ENTRIES 512 diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index 56ad020df494..d8dc5c6db29d 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -132,6 +132,8 @@ struct ivhd_entry { u8 uid; } __attribute__((packed)); +int amd_iommu_evtlog_size = EVTLOG_SIZE_DEF; + /* * An AMD IOMMU memory definition structure. It defines things like exclusion * ranges for devices and regions that should be unity mapped. @@ -865,35 +867,47 @@ void *__init iommu_alloc_4k_pages(struct amd_iommu *iommu, gfp_t gfp, } /* allocates the memory where the IOMMU will log its events to */ -static int __init alloc_event_buffer(struct amd_iommu *iommu) +static int __init alloc_event_buffer(void) { - iommu->evt_buf = iommu_alloc_4k_pages(iommu, GFP_KERNEL, - EVT_BUFFER_SIZE); + struct amd_iommu *iommu; - return iommu->evt_buf ? 0 : -ENOMEM; -} - -static void iommu_enable_event_buffer(struct amd_iommu *iommu) -{ - u64 entry; - - BUG_ON(iommu->evt_buf == NULL); - - if (!is_kdump_kernel()) { - /* - * Event buffer is re-used for kdump kernel and setting - * of MMIO register is not required. - */ - entry = iommu_virt_to_phys(iommu->evt_buf) | EVT_LEN_MASK; - memcpy_toio(iommu->mmio_base + MMIO_EVT_BUF_OFFSET, - &entry, sizeof(entry)); + for_each_iommu(iommu) { + iommu->evt_buf = iommu_alloc_4k_pages(iommu, GFP_KERNEL, + amd_iommu_evtlog_size); + if (!iommu->evt_buf) + return -ENOMEM; } - /* set head and tail to zero manually */ - writel(0x00, iommu->mmio_base + MMIO_EVT_HEAD_OFFSET); - writel(0x00, iommu->mmio_base + MMIO_EVT_TAIL_OFFSET); + return 0; +} - iommu_feature_enable(iommu, CONTROL_EVT_LOG_EN); +static void iommu_enable_event_buffer(void) +{ + struct amd_iommu *iommu; + u64 entry; + + for_each_iommu(iommu) { + BUG_ON(iommu->evt_buf == NULL); + + if (!is_kdump_kernel()) { + /* + * Event buffer is re-used for kdump kernel and setting + * of MMIO register is not required. + */ + entry = iommu_virt_to_phys(iommu->evt_buf); + entry |= (amd_iommu_evtlog_size == EVTLOG_SIZE_DEF) ? + EVTLOG_LEN_MASK_DEF : EVTLOG_LEN_MASK_MAX; + + memcpy_toio(iommu->mmio_base + MMIO_EVT_BUF_OFFSET, + &entry, sizeof(entry)); + } + + /* set head and tail to zero manually */ + writel(0x00, iommu->mmio_base + MMIO_EVT_HEAD_OFFSET); + writel(0x00, iommu->mmio_base + MMIO_EVT_TAIL_OFFSET); + + iommu_feature_enable(iommu, CONTROL_EVT_LOG_EN); + } } /* @@ -984,15 +998,20 @@ static int __init alloc_cwwb_sem(struct amd_iommu *iommu) return 0; } -static int __init remap_event_buffer(struct amd_iommu *iommu) +static int __init remap_event_buffer(void) { + struct amd_iommu *iommu; u64 paddr; pr_info_once("Re-using event buffer from the previous kernel\n"); - paddr = readq(iommu->mmio_base + MMIO_EVT_BUF_OFFSET) & PM_ADDR_MASK; - iommu->evt_buf = iommu_memremap(paddr, EVT_BUFFER_SIZE); + for_each_iommu(iommu) { + paddr = readq(iommu->mmio_base + MMIO_EVT_BUF_OFFSET) & PM_ADDR_MASK; + iommu->evt_buf = iommu_memremap(paddr, amd_iommu_evtlog_size); + if (!iommu->evt_buf) + return -ENOMEM; + } - return iommu->evt_buf ? 0 : -ENOMEM; + return 0; } static int __init remap_command_buffer(struct amd_iommu *iommu) @@ -1044,10 +1063,6 @@ static int __init alloc_iommu_buffers(struct amd_iommu *iommu) ret = remap_command_buffer(iommu); if (ret) return ret; - - ret = remap_event_buffer(iommu); - if (ret) - return ret; } else { ret = alloc_cwwb_sem(iommu); if (ret) @@ -1056,10 +1071,6 @@ static int __init alloc_iommu_buffers(struct amd_iommu *iommu) ret = alloc_command_buffer(iommu); if (ret) return ret; - - ret = alloc_event_buffer(iommu); - if (ret) - return ret; } return 0; @@ -2893,7 +2904,6 @@ static void early_enable_iommu(struct amd_iommu *iommu) iommu_init_flags(iommu); iommu_set_device_table(iommu); iommu_enable_command_buffer(iommu); - iommu_enable_event_buffer(iommu); iommu_set_exclusion_range(iommu); iommu_enable_gt(iommu); iommu_enable_ga(iommu); @@ -2957,7 +2967,6 @@ static void early_enable_iommus(void) iommu_disable_event_buffer(iommu); iommu_disable_irtcachedis(iommu); iommu_enable_command_buffer(iommu); - iommu_enable_event_buffer(iommu); iommu_enable_ga(iommu); iommu_enable_xt(iommu); iommu_enable_irtcachedis(iommu); @@ -3070,6 +3079,7 @@ static void amd_iommu_resume(void *data) for_each_iommu(iommu) early_enable_iommu(iommu); + iommu_enable_event_buffer(); amd_iommu_enable_interrupts(); } @@ -3399,6 +3409,23 @@ static __init void iommu_snp_enable(void) #endif } +static void amd_iommu_apply_erratum_snp(void) +{ +#ifdef CONFIG_KVM_AMD_SEV + if (!amd_iommu_snp_en) + return; + + /* Errata fix for Family 0x19 */ + if (boot_cpu_data.x86 != 0x19) + return; + + /* Set event log buffer size to max */ + amd_iommu_evtlog_size = EVTLOG_SIZE_MAX; + pr_info("Applying erratum: Increase Event log size to 0x%x\n", + amd_iommu_evtlog_size); +#endif +} + /**************************************************************************** * * AMD IOMMU Initialization State Machine @@ -3435,6 +3462,21 @@ static int __init state_next(void) case IOMMU_ENABLED: register_syscore(&amd_iommu_syscore); iommu_snp_enable(); + + amd_iommu_apply_erratum_snp(); + + /* Allocate/enable event log buffer */ + if (is_kdump_kernel()) + ret = remap_event_buffer(); + else + ret = alloc_event_buffer(); + + if (ret) { + init_state = IOMMU_INIT_ERROR; + break; + } + iommu_enable_event_buffer(); + ret = amd_iommu_init_pci(); init_state = ret ? IOMMU_INIT_ERROR : IOMMU_PCI_INIT; break; @@ -4037,7 +4079,7 @@ int amd_iommu_snp_disable(void) return 0; for_each_iommu(iommu) { - ret = iommu_make_shared(iommu->evt_buf, EVT_BUFFER_SIZE); + ret = iommu_make_shared(iommu->evt_buf, amd_iommu_evtlog_size); if (ret) return ret; diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c index 01171361f9bc..157505d96fdd 100644 --- a/drivers/iommu/amd/iommu.c +++ b/drivers/iommu/amd/iommu.c @@ -1010,7 +1010,7 @@ static void iommu_poll_events(struct amd_iommu *iommu) iommu_print_event(iommu, iommu->evt_buf + head); /* Update head pointer of hardware ring-buffer */ - head = (head + EVENT_ENTRY_SIZE) % EVT_BUFFER_SIZE; + head = (head + EVTLOG_ENTRY_SIZE) % amd_iommu_evtlog_size; writel(head, iommu->mmio_base + MMIO_EVT_HEAD_OFFSET); } From 1f44aab79bac31f459422dfb213e907bb386509c Mon Sep 17 00:00:00 2001 From: Vasant Hegde Date: Mon, 20 Apr 2026 08:42:04 +0000 Subject: [PATCH 014/333] iommu/amd: Use maximum PPR log buffer size when SNP is enabled on Family 0x19 Due to CVE-2023-20585, the PPR log buffer must use the maximum supported size (512K) on Genoa (Family 0x19, model >= 0x10) systems when SNP is enabled, to mitigate a potential security vulnerability. Note that Family 0x19 models below 0x10 (Milan) do not support PPR when SNP is enabled. Hence the PPR log size increase is only applied for model >= 0x10. All other systems continue to use the default PPR log buffer size (8K). Apply the errata fix by making the following changes: - Introduce global new variable (amd_iommu_pprlog_size) to have PPR log buffer size. Adjust variable size for Genoa family. - Extend 'amd_iommu_apply_erratum_snp()' to also set the PPR log buffer size to maximum for Family 0x19 model >= 0x10 when SNP is enabled. - Rename PPR_* macros to make it more readable. Link: https://www.amd.com/en/resources/product-security/bulletin/amd-sb-3016.html Cc: Borislav Petkov Cc: Suravee Suthikulpanit Cc: Joerg Roedel Signed-off-by: Vasant Hegde Tested-by: Dheeraj Kumar Srivastava Signed-off-by: Joerg Roedel --- drivers/iommu/amd/amd_iommu.h | 1 + drivers/iommu/amd/amd_iommu_types.h | 11 ++++++----- drivers/iommu/amd/init.c | 13 ++++++++++++- drivers/iommu/amd/ppr.c | 8 +++++--- 4 files changed, 24 insertions(+), 9 deletions(-) diff --git a/drivers/iommu/amd/amd_iommu.h b/drivers/iommu/amd/amd_iommu.h index f1c486dcf0f3..834d8fabfba3 100644 --- a/drivers/iommu/amd/amd_iommu.h +++ b/drivers/iommu/amd/amd_iommu.h @@ -12,6 +12,7 @@ #include "amd_iommu_types.h" extern int amd_iommu_evtlog_size; +extern int amd_iommu_pprlog_size; irqreturn_t amd_iommu_int_thread(int irq, void *data); irqreturn_t amd_iommu_int_thread_evtlog(int irq, void *data); diff --git a/drivers/iommu/amd/amd_iommu_types.h b/drivers/iommu/amd/amd_iommu_types.h index c3430c09bc5c..f9f718087893 100644 --- a/drivers/iommu/amd/amd_iommu_types.h +++ b/drivers/iommu/amd/amd_iommu_types.h @@ -267,11 +267,12 @@ #define EVTLOG_LEN_MASK_MAX (0xFULL << EVTLOG_SIZE_SHIFT) /* Constants for PPR Log handling */ -#define PPR_LOG_ENTRIES 512 -#define PPR_LOG_SIZE_SHIFT 56 -#define PPR_LOG_SIZE_512 (0x9ULL << PPR_LOG_SIZE_SHIFT) -#define PPR_ENTRY_SIZE 16 -#define PPR_LOG_SIZE (PPR_ENTRY_SIZE * PPR_LOG_ENTRIES) +#define PPRLOG_ENTRY_SIZE 0x10 +#define PPRLOG_SIZE_SHIFT 56 +#define PPRLOG_SIZE_DEF SZ_8K /* 512 entries */ +#define PPRLOG_LEN_MASK_DEF (0x9ULL << PPRLOG_SIZE_SHIFT) +#define PPRLOG_SIZE_MAX SZ_512K /* 32K entries */ +#define PPRLOG_LEN_MASK_MAX (0xFULL << PPRLOG_SIZE_SHIFT) /* PAGE_SERVICE_REQUEST PPR Log Buffer Entry flags */ #define PPR_FLAG_EXEC 0x002 /* Execute permission requested */ diff --git a/drivers/iommu/amd/init.c b/drivers/iommu/amd/init.c index d8dc5c6db29d..3bdb380d23e9 100644 --- a/drivers/iommu/amd/init.c +++ b/drivers/iommu/amd/init.c @@ -133,6 +133,7 @@ struct ivhd_entry { } __attribute__((packed)); int amd_iommu_evtlog_size = EVTLOG_SIZE_DEF; +int amd_iommu_pprlog_size = PPRLOG_SIZE_DEF; /* * An AMD IOMMU memory definition structure. It defines things like exclusion @@ -3423,6 +3424,16 @@ static void amd_iommu_apply_erratum_snp(void) amd_iommu_evtlog_size = EVTLOG_SIZE_MAX; pr_info("Applying erratum: Increase Event log size to 0x%x\n", amd_iommu_evtlog_size); + + /* + * Set PPR log buffer size to max. + * (Family 0x19, model < 0x10 doesn't support PPR when SNP is enabled). + */ + if (boot_cpu_data.x86_model >= 0x10) { + amd_iommu_pprlog_size = PPRLOG_SIZE_MAX; + pr_info("Applying erratum: Increase PPR log size to 0x%x\n", + amd_iommu_pprlog_size); + } #endif } @@ -4083,7 +4094,7 @@ int amd_iommu_snp_disable(void) if (ret) return ret; - ret = iommu_make_shared(iommu->ppr_log, PPR_LOG_SIZE); + ret = iommu_make_shared(iommu->ppr_log, amd_iommu_pprlog_size); if (ret) return ret; diff --git a/drivers/iommu/amd/ppr.c b/drivers/iommu/amd/ppr.c index e6767c057d01..1f8d2823bea4 100644 --- a/drivers/iommu/amd/ppr.c +++ b/drivers/iommu/amd/ppr.c @@ -20,7 +20,7 @@ int __init amd_iommu_alloc_ppr_log(struct amd_iommu *iommu) { iommu->ppr_log = iommu_alloc_4k_pages(iommu, GFP_KERNEL | __GFP_ZERO, - PPR_LOG_SIZE); + amd_iommu_pprlog_size); return iommu->ppr_log ? 0 : -ENOMEM; } @@ -33,7 +33,9 @@ void amd_iommu_enable_ppr_log(struct amd_iommu *iommu) iommu_feature_enable(iommu, CONTROL_PPR_EN); - entry = iommu_virt_to_phys(iommu->ppr_log) | PPR_LOG_SIZE_512; + entry = iommu_virt_to_phys(iommu->ppr_log); + entry |= (amd_iommu_pprlog_size == PPRLOG_SIZE_DEF) ? + PPRLOG_LEN_MASK_DEF : PPRLOG_LEN_MASK_MAX; memcpy_toio(iommu->mmio_base + MMIO_PPR_LOG_OFFSET, &entry, sizeof(entry)); @@ -201,7 +203,7 @@ void amd_iommu_poll_ppr_log(struct amd_iommu *iommu) raw[0] = raw[1] = 0UL; /* Update head pointer of hardware ring-buffer */ - head = (head + PPR_ENTRY_SIZE) % PPR_LOG_SIZE; + head = (head + PPRLOG_ENTRY_SIZE) % amd_iommu_pprlog_size; writel(head, iommu->mmio_base + MMIO_PPR_HEAD_OFFSET); /* Handle PPR entry */ From f9471dc1a7177f594acf8106cc4a6ab33bf0bcb6 Mon Sep 17 00:00:00 2001 From: Mostafa Saleh Date: Fri, 24 Apr 2026 11:50:51 +0000 Subject: [PATCH 015/333] iommu/pages: Fix iommu_pages_flush_incoherent() for non-x86 The dma_sync_single_for_device() function expects a dma_addr_t, but iommu_pages_flush_incoherent() was incorrectly passing a virtual address. Since iommu_pages_start_incoherent() enforces a 1:1 mapping between DMA addresses and physical addresses (checked via WARN_ON), we can convert the virtual address to a physical address before passing it to the DMA API. This also matches the behaviour of the other non-x86 in iommu_pages_free_incoherent(), which uses virt_to_phys(virt); Fixes: 36ae67b13976 ("iommu/pages: Add support for incoherent IOMMU page table walkers") Signed-off-by: Mostafa Saleh Reviewed-by: Lu Baolu Reviewed-by: Pranjal Shrivastava Reviewed-by: Jason Gunthorpe Signed-off-by: Joerg Roedel --- drivers/iommu/iommu-pages.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/iommu/iommu-pages.h b/drivers/iommu/iommu-pages.h index ae9da4f571f6..e9e605b5fa3a 100644 --- a/drivers/iommu/iommu-pages.h +++ b/drivers/iommu/iommu-pages.h @@ -137,7 +137,7 @@ static inline void iommu_pages_flush_incoherent(struct device *dma_dev, void *virt, size_t offset, size_t len) { - dma_sync_single_for_device(dma_dev, (uintptr_t)virt + offset, len, + dma_sync_single_for_device(dma_dev, virt_to_phys(virt) + offset, len, DMA_TO_DEVICE); } void iommu_pages_stop_incoherent_list(struct iommu_pages_list *list, From 3f91484f6c13c434bd573ca6b6779c26adb0ddab Mon Sep 17 00:00:00 2001 From: Aaro Koskinen Date: Mon, 13 Apr 2026 21:49:12 +0300 Subject: [PATCH 016/333] USB: omap_udc: DMA: Don't enable burst 4 mode Commit 65111084c63d7 ("USB: more omap_udc updates (dma and omap1710)") added setting for DMA burst 4 mode. But I think this should be undone for two reasons: - It breaks DMA on 15xx boards - transfers just silently stall. - On newer OMAP1 boards, like Nokia 770 (omap1710), there is no measurable performance impact when testing TCP throughput with g_ether with large 15000 byte MTU size. It's also worth noting that when the original change was made, the OMAP_DMA_DATA_BURST_4 handling in arch/arm/plat-omap/dma.c was broken, and actually resulted in the same as the OMAP_DMA_DATA_BURST_DIS i.e. burst disabled. This was fixed not until a couple kernel releases later in an unrelated commit 1a8bfa1eb998a ("[ARM] 3142/1: OMAP 2/5: Update files common to omap1 and omap2"). So based on this it seems there was never really a very good reason to enable this burst mode in omap_udc, so remove it now to allow 15xx DMA to work again (it provides 2x throughput compared to PIO mode). Fixes: 65111084c63d ("[PATCH] USB: more omap_udc updates (dma and omap1710)") Cc: stable Signed-off-by: Aaro Koskinen Link: https://patch.msgid.link/ad06qHLclWHeSGnV@darkstar.musicnaut.iki.fi Signed-off-by: Greg Kroah-Hartman --- drivers/usb/gadget/udc/omap_udc.c | 4 ---- 1 file changed, 4 deletions(-) diff --git a/drivers/usb/gadget/udc/omap_udc.c b/drivers/usb/gadget/udc/omap_udc.c index 91139ae668f4..f3ca79cece1b 100644 --- a/drivers/usb/gadget/udc/omap_udc.c +++ b/drivers/usb/gadget/udc/omap_udc.c @@ -733,8 +733,6 @@ static void dma_channel_claim(struct omap_ep *ep, unsigned channel) if (status == 0) { omap_writew(reg, UDC_TXDMA_CFG); /* EMIFF or SDRC */ - omap_set_dma_src_burst_mode(ep->lch, - OMAP_DMA_DATA_BURST_4); omap_set_dma_src_data_pack(ep->lch, 1); /* TIPB */ omap_set_dma_dest_params(ep->lch, @@ -756,8 +754,6 @@ static void dma_channel_claim(struct omap_ep *ep, unsigned channel) UDC_DATA_DMA, 0, 0); /* EMIFF or SDRC */ - omap_set_dma_dest_burst_mode(ep->lch, - OMAP_DMA_DATA_BURST_4); omap_set_dma_dest_data_pack(ep->lch, 1); } } From 0b9fcab1b8608d429e5f239afb197de928d4de7d Mon Sep 17 00:00:00 2001 From: Felix Gu Date: Tue, 7 Apr 2026 21:21:22 +0800 Subject: [PATCH 017/333] usb: ulpi: fix memory leak on ulpi_register() error paths Commit 01af542392b5 ("usb: ulpi: fix double free in ulpi_register_interface() error path") removed kfree(ulpi) from ulpi_register_interface() to fix a double-free when device_register() fails. But when ulpi_of_register() or ulpi_read_id() fail before device_register() is called, the ulpi allocation is leaked. Add kfree(ulpi) on both error paths to properly clean up the allocation. Fixes: 01af542392b5 ("usb: ulpi: fix double free in ulpi_register_interface() error path") Cc: stable Signed-off-by: Felix Gu Reviewed-by: Heikki Krogerus Link: https://patch.msgid.link/20260407-ulpi-v1-1-f3fafe53f7b2@gmail.com Signed-off-by: Greg Kroah-Hartman --- drivers/usb/common/ulpi.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/usb/common/ulpi.c b/drivers/usb/common/ulpi.c index b34fb65813c4..9b69148128e5 100644 --- a/drivers/usb/common/ulpi.c +++ b/drivers/usb/common/ulpi.c @@ -286,12 +286,15 @@ static int ulpi_register(struct device *dev, struct ulpi *ulpi) ACPI_COMPANION_SET(&ulpi->dev, ACPI_COMPANION(dev)); ret = ulpi_of_register(ulpi); - if (ret) + if (ret) { + kfree(ulpi); return ret; + } ret = ulpi_read_id(ulpi); if (ret) { of_node_put(ulpi->dev.of_node); + kfree(ulpi); return ret; } From 2909f0d4994fb4306bf116df5ccee797791fce2c Mon Sep 17 00:00:00 2001 From: Amit Sunil Dhamne Date: Tue, 14 Apr 2026 00:58:32 +0000 Subject: [PATCH 018/333] usb: typec: tcpm: reset internal port states on soft reset AMS Reset internal port states (such as vdm_sm_running and explicit_contract) on soft reset AMS as the port needs to negotiate a new contract. The consequence of leaving the states in as-is cond are as follows: * port is in SRC power role and an explicit contract is negotiated with the port partner (in sink role) * port partner sends a Soft Reset AMS while VDM State Machine is running * port accepts the Soft Reset request and the port advertises src caps * port partner sends a Request message but since the explicit_contract and vdm_sm_running are true from previous negotiation, the port ends up sending Soft Reset instead of Accept msg. Stub Log: [ 203.653942] AMS DISCOVER_IDENTITY start [ 203.653947] PD TX, header: 0x176f [ 203.655901] PD TX complete, status: 0 [ 203.657470] PD RX, header: 0x124f [1] [ 203.657477] Rx VDM cmd 0xff008081 type 2 cmd 1 len 1 [ 203.657482] AMS DISCOVER_IDENTITY finished [ 203.657484] cc:=4 [ 204.155698] PD RX, header: 0x144f [1] [ 204.155718] Rx VDM cmd 0xeeee8001 type 0 cmd 1 len 1 [ 204.155741] PD TX, header: 0x196f [ 204.157622] PD TX complete, status: 0 [ 204.160060] PD RX, header: 0x4d [1] [ 204.160066] state change SRC_READY -> SOFT_RESET [rev2 SOFT_RESET_AMS] [ 204.160076] PD TX, header: 0x163 [ 204.162486] PD TX complete, status: 0 [ 204.162832] AMS SOFT_RESET_AMS finished [ 204.162840] cc:=4 [ 204.162891] AMS POWER_NEGOTIATION start [ 204.162896] state change SOFT_RESET -> AMS_START [rev2 POWER_NEGOTIATION] [ 204.162908] state change AMS_START -> SRC_SEND_CAPABILITIES [rev2 POWER_NEGOTIATION] [ 204.162913] PD TX, header: 0x1361 [ 204.165529] PD TX complete, status: 0 [ 204.165571] pending state change SRC_SEND_CAPABILITIES -> SRC_SEND_CAPABILITIES_TIMEOUT @ 60 ms [rev2 POWER_NEGOTIATION] [ 204.166996] PD RX, header: 0x1242 [1] [ 204.167009] state change SRC_SEND_CAPABILITIES -> SRC_SOFT_RESET_WAIT_SNK_TX [rev2 POWER_NEGOTIATION] [ 204.167019] AMS POWER_NEGOTIATION finished [ 204.167020] cc:=4 [ 204.167083] AMS SOFT_RESET_AMS start [ 204.167086] state change SRC_SOFT_RESET_WAIT_SNK_TX -> SOFT_RESET_SEND [rev2 SOFT_RESET_AMS] [ 204.167092] PD TX, header: 0x16d [ 204.168824] PD TX complete, status: 0 [ 204.168854] pending state change SOFT_RESET_SEND -> HARD_RESET_SEND @ 60 ms [rev2 SOFT_RESET_AMS] [ 204.171876] PD RX, header: 0x43 [1] [ 204.171879] AMS SOFT_RESET_AMS finished This causes COMMON.PROC.PD.11.2 check failure for TEST.PD.VDM.SRC.2_Rev2Src test on the PD compliance tester. Signed-off-by: Amit Sunil Dhamne Fixes: 8d3a0578ad1a ("usb: typec: tcpm: Respond Wait if VDM state machine is running") Fixes: f0690a25a140 ("staging: typec: USB Type-C Port Manager (tcpm)") Cc: stable Reviewed-by: Badhri Jagan Sridharan Acked-by: Heikki Krogerus Link: https://patch.msgid.link/20260414-fix-soft-reset-v1-1-01d7cb9764e2@google.com Signed-off-by: Greg Kroah-Hartman --- drivers/usb/typec/tcpm/tcpm.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/usb/typec/tcpm/tcpm.c b/drivers/usb/typec/tcpm/tcpm.c index dfbb94ddc98a..cd5560d6f19e 100644 --- a/drivers/usb/typec/tcpm/tcpm.c +++ b/drivers/usb/typec/tcpm/tcpm.c @@ -5935,6 +5935,8 @@ static void run_state_machine(struct tcpm_port *port) /* remove existing capabilities */ tcpm_partner_source_caps_reset(port); tcpm_pd_send_control(port, PD_CTRL_ACCEPT, TCPC_TX_SOP); + port->vdm_sm_running = false; + port->explicit_contract = false; tcpm_ams_finish(port); if (port->pwr_role == TYPEC_SOURCE) { port->upcoming_state = SRC_SEND_CAPABILITIES; From f6ec9bb4acc7182b25a793ad094a764e1cb819a7 Mon Sep 17 00:00:00 2001 From: Xu Yang Date: Fri, 24 Apr 2026 15:40:09 +0800 Subject: [PATCH 019/333] usb: typec: tcpm: fix debug accessory mode detection for sink ports The port in debug accessory mode can be either a source or sink. The previous tcpm_port_is_debug() function only checked for source port. Commit 8db73e6a42b6 ("usb: typec: tcpm: allow sink (ufp) to toggle into accessory mode debug") changed the detection logic to support both roles, but left some logic in _tcpm_cc_change() unchanged, This causes the state machine to transition to an incorrect state when operating as a sink in debug accessory mode. Log as below: [ 978.637541] CC1: 0 -> 5, CC2: 0 -> 5 [state TOGGLING, polarity 0, connected] [ 978.637567] state change TOGGLING -> SRC_ATTACH_WAIT [rev1 NONE_AMS] [ 978.637596] pending state change SRC_ATTACH_WAIT -> DEBUG_ACC_ATTACHED @ 180 ms [rev1 NONE_AMS] [ 978.647098] CC1: 5 -> 0, CC2: 5 -> 5 [state SRC_ATTACH_WAIT, polarity 0, connected] [ 978.647115] state change SRC_ATTACH_WAIT -> SRC_ATTACH_WAIT [rev1 NONE_AMS] It should go to SNK_ATTACH_WAIT instead of SRC_ATTACH_WAIT state. To fix this, add tcpm_port_is_debug_source() and tcpm_port_is_debug_sink() helper to explicitly identify the power mode in debug accessory mode. Update the state transition logic in _tcpm_cc_change() to ensure the state machine transitions comply with Type-C specification. Also update the logic in run_state_machine() to keep consistency. Fixes: 8db73e6a42b6 ("usb: typec: tcpm: allow sink (ufp) to toggle into accessory mode debug") Cc: stable Signed-off-by: Xu Yang Acked-by: Heikki Krogerus Reviewed-by: Amit Sunil Dhamne Link: https://patch.msgid.link/20260424074009.2979266-1-xu.yang_2@nxp.com Signed-off-by: Greg Kroah-Hartman --- drivers/usb/typec/tcpm/tcpm.c | 25 ++++++++++++++++--------- 1 file changed, 16 insertions(+), 9 deletions(-) diff --git a/drivers/usb/typec/tcpm/tcpm.c b/drivers/usb/typec/tcpm/tcpm.c index cd5560d6f19e..55fee96d3342 100644 --- a/drivers/usb/typec/tcpm/tcpm.c +++ b/drivers/usb/typec/tcpm/tcpm.c @@ -732,9 +732,14 @@ static const char * const pd_rev[] = { (tcpm_cc_is_source((port)->cc2) && \ !tcpm_cc_is_source((port)->cc1))) +#define tcpm_port_is_debug_source(port) \ + (tcpm_cc_is_source((port)->cc1) && tcpm_cc_is_source((port)->cc2)) + +#define tcpm_port_is_debug_sink(port) \ + (tcpm_cc_is_sink((port)->cc1) && tcpm_cc_is_sink((port)->cc2)) + #define tcpm_port_is_debug(port) \ - ((tcpm_cc_is_source((port)->cc1) && tcpm_cc_is_source((port)->cc2)) || \ - (tcpm_cc_is_sink((port)->cc1) && tcpm_cc_is_sink((port)->cc2))) + (tcpm_port_is_debug_source(port) || tcpm_port_is_debug_sink(port)) #define tcpm_port_is_audio(port) \ (tcpm_cc_is_audio((port)->cc1) && tcpm_cc_is_audio((port)->cc2)) @@ -5176,7 +5181,7 @@ static void run_state_machine(struct tcpm_port *port) tcpm_set_state(port, SNK_UNATTACHED, PD_T_DRP_SNK); break; case SRC_ATTACH_WAIT: - if (tcpm_port_is_debug(port)) + if (tcpm_port_is_debug_source(port)) tcpm_set_state(port, DEBUG_ACC_ATTACHED, port->timings.cc_debounce_time); else if (tcpm_port_is_audio(port)) @@ -5434,7 +5439,7 @@ static void run_state_machine(struct tcpm_port *port) tcpm_set_state(port, SRC_UNATTACHED, PD_T_DRP_SRC); break; case SNK_ATTACH_WAIT: - if (tcpm_port_is_debug(port)) + if (tcpm_port_is_debug_sink(port)) tcpm_set_state(port, DEBUG_ACC_ATTACHED, PD_T_CC_DEBOUNCE); else if (tcpm_port_is_audio(port)) @@ -5454,7 +5459,7 @@ static void run_state_machine(struct tcpm_port *port) if (tcpm_port_is_disconnected(port)) tcpm_set_state(port, SNK_UNATTACHED, PD_T_PD_DEBOUNCE); - else if (tcpm_port_is_debug(port)) + else if (tcpm_port_is_debug_sink(port)) tcpm_set_state(port, DEBUG_ACC_ATTACHED, PD_T_CC_DEBOUNCE); else if (tcpm_port_is_audio(port)) @@ -6362,10 +6367,10 @@ static void _tcpm_cc_change(struct tcpm_port *port, enum typec_cc_status cc1, switch (port->state) { case TOGGLING: - if (tcpm_port_is_debug(port) || tcpm_port_is_audio(port) || + if (tcpm_port_is_debug_source(port) || tcpm_port_is_audio(port) || tcpm_port_is_source(port)) tcpm_set_state(port, SRC_ATTACH_WAIT, 0); - else if (tcpm_port_is_sink(port)) + else if (tcpm_port_is_debug_sink(port) || tcpm_port_is_sink(port)) tcpm_set_state(port, SNK_ATTACH_WAIT, 0); break; case CHECK_CONTAMINANT: @@ -6373,9 +6378,11 @@ static void _tcpm_cc_change(struct tcpm_port *port, enum typec_cc_status cc1, break; case SRC_UNATTACHED: case ACC_UNATTACHED: - if (tcpm_port_is_debug(port) || tcpm_port_is_audio(port) || + if (tcpm_port_is_debug_source(port) || tcpm_port_is_audio(port) || tcpm_port_is_source(port)) tcpm_set_state(port, SRC_ATTACH_WAIT, 0); + else if (tcpm_port_is_debug_sink(port)) + tcpm_set_state(port, SNK_ATTACH_WAIT, 0); break; case SRC_ATTACH_WAIT: if (tcpm_port_is_disconnected(port) || @@ -6397,7 +6404,7 @@ static void _tcpm_cc_change(struct tcpm_port *port, enum typec_cc_status cc1, } break; case SNK_UNATTACHED: - if (tcpm_port_is_debug(port) || tcpm_port_is_audio(port) || + if (tcpm_port_is_debug_sink(port) || tcpm_port_is_audio(port) || tcpm_port_is_sink(port)) tcpm_set_state(port, SNK_ATTACH_WAIT, 0); break; From aad35f9c926ec220b0742af1ada45666ae667956 Mon Sep 17 00:00:00 2001 From: Selvarasu Ganesan Date: Fri, 17 Apr 2026 12:03:11 +0530 Subject: [PATCH 020/333] usb: dwc3: Move GUID programming after PHY initialization The Linux Version Code is currently written to the GUID register before PHY initialization. Certain PHY implementations (such as Synopsys eUSB PHY performing link_sw_reset) clear the GUID register to its default value during initialization, causing the kernel version information to be lost. Move the GUID register programming to occur after PHY initialization completes to ensure the Linux version information persists. Fixes: fa0ea13e9f1c ("usb: dwc3: core: write LINUX_VERSION_CODE to our GUID register") Cc: stable Reported-by: Pritam Manohar Sutar Signed-off-by: Selvarasu Ganesan Acked-by: Thinh Nguyen Link: https://patch.msgid.link/20260417063314.2359-1-selvarasu.g@samsung.com Signed-off-by: Greg Kroah-Hartman --- drivers/usb/dwc3/core.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c index 58899b1fa96d..65213896de99 100644 --- a/drivers/usb/dwc3/core.c +++ b/drivers/usb/dwc3/core.c @@ -1359,12 +1359,6 @@ int dwc3_core_init(struct dwc3 *dwc) hw_mode = DWC3_GHWPARAMS0_MODE(dwc->hwparams.hwparams0); - /* - * Write Linux Version Code to our GUID register so it's easy to figure - * out which kernel version a bug was found. - */ - dwc3_writel(dwc, DWC3_GUID, LINUX_VERSION_CODE); - ret = dwc3_phy_setup(dwc); if (ret) return ret; @@ -1398,6 +1392,12 @@ int dwc3_core_init(struct dwc3 *dwc) if (ret) goto err_exit_phy; + /* + * Write Linux Version Code to our GUID register so it's easy to figure + * out which kernel version a bug was found. + */ + dwc3_writel(dwc, DWC3_GUID, LINUX_VERSION_CODE); + dwc3_core_setup_global_control(dwc); dwc3_core_num_eps(dwc); From 7a400c6fe3617e31e690e3f7ca37bb335e0498f3 Mon Sep 17 00:00:00 2001 From: Greg Kroah-Hartman Date: Mon, 20 Apr 2026 18:11:03 +0200 Subject: [PATCH 021/333] usb: usblp: fix heap leak in IEEE 1284 device ID via short response usblp_ctrl_msg() collapses the usb_control_msg() return value to 0/-errno, discarding the actual number of bytes transferred. A broken printer can complete the GET_DEVICE_ID control transfer short and the driver has no way to know. usblp_cache_device_id_string() reads the 2-byte big-endian length prefix from the response and trusts it (clamped only to the buffer bounds). The buffer is kmalloc(1024) at probe time. A device that sends exactly two bytes (e.g. 0x03 0xFF, claiming a 1023-byte ID) leaves device_id_string[2..1022] holding stale kmalloc heap. That stale data is then exposed: - via the ieee1284_id sysfs attribute (sprintf("%s", buf+2), truncated at the first NUL in the stale heap), and - via the IOCNR_GET_DEVICE_ID ioctl, which copy_to_user()s the full claimed length regardless of NULs, up to 1021 bytes of uninitialized heap, with the leak size chosen by the device. Fix this up by just zapping the buffer with zeros before each request sent to the device. Cc: Pete Zaitcev Assisted-by: gkh_clanker_t1000 Cc: stable Link: https://patch.msgid.link/2026042002-unicorn-greedily-3c63@gregkh Signed-off-by: Greg Kroah-Hartman --- drivers/usb/class/usblp.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/usb/class/usblp.c b/drivers/usb/class/usblp.c index 669b9e6879bf..e9b848622a3a 100644 --- a/drivers/usb/class/usblp.c +++ b/drivers/usb/class/usblp.c @@ -1377,6 +1377,7 @@ static int usblp_cache_device_id_string(struct usblp *usblp) { int err, length; + memset(usblp->device_id_string, 0, USBLP_DEVICE_ID_SIZE); err = usblp_get_id(usblp, 0, usblp->device_id_string, USBLP_DEVICE_ID_SIZE - 1); if (err < 0) { dev_dbg(&usblp->intf->dev, From b38e53cbfb9d84732e5984fbd73e128d592415c5 Mon Sep 17 00:00:00 2001 From: Greg Kroah-Hartman Date: Mon, 20 Apr 2026 18:11:04 +0200 Subject: [PATCH 022/333] usb: usblp: fix uninitialized heap leak via LPGETSTATUS ioctl Just like in a previous problem in this driver, usblp_ctrl_msg() will collapse the usb_control_msg() return value to 0/-errno, discarding the actual number of bytes transferred. Ideally that short command should be detected and error out, but many printers are known to send "incorrect" responses back so we can't just do that. statusbuf is kmalloc(8) at probe time and never filled before the first LPGETSTATUS ioctl. usblp_read_status() requests 1 byte. If a malicious printer responds with zero bytes, *statusbuf is one byte of stale kmalloc heap, sign-extended into the local int status, which the LPGETSTATUS path then copy_to_user()s directly to the ioctl caller. Fix this all by just zapping out the memory buffer when allocated at probe time. If a later call does a short read, the data will be identical to what the device sent it the last time, so there is no "leak" of information happening. Cc: Pete Zaitcev Assisted-by: gkh_clanker_t1000 Cc: stable Link: https://patch.msgid.link/2026042011-shredder-savage-48c6@gregkh Signed-off-by: Greg Kroah-Hartman --- drivers/usb/class/usblp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/usb/class/usblp.c b/drivers/usb/class/usblp.c index e9b848622a3a..746414763da5 100644 --- a/drivers/usb/class/usblp.c +++ b/drivers/usb/class/usblp.c @@ -1178,7 +1178,7 @@ static int usblp_probe(struct usb_interface *intf, } /* Allocate buffer for printer status */ - usblp->statusbuf = kmalloc(STATUS_BUF_SIZE, GFP_KERNEL); + usblp->statusbuf = kzalloc(STATUS_BUF_SIZE, GFP_KERNEL); if (!usblp->statusbuf) { retval = -ENOMEM; goto abort; From cfcbfe5cb11650d53f7cafd7adfd556690b77114 Mon Sep 17 00:00:00 2001 From: Keith Busch Date: Tue, 21 Apr 2026 08:06:44 -0700 Subject: [PATCH 023/333] PCI: Don't fallback to bus reset after failed slot reset If a bus has hotplug slots that implement the slot's reset_slot callback, it is not safe to do the non-slot specific bus reset, so don't fallback to it. If a slot reset does fail, the subsequent bus reset will attempt a 2nd link reset on top of previous and fail to handle the hotplug events. Fixes: 8238cb69c01fe ("PCI: Make reset_subordinate hotplug safe") Signed-off-by: Keith Busch Signed-off-by: Bjorn Helgaas Link: https://patch.msgid.link/20260421150644.3543733-1-kbusch@meta.com --- drivers/pci/pci.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-) diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c index 8f7cfcc00090..d34266651ad0 100644 --- a/drivers/pci/pci.c +++ b/drivers/pci/pci.c @@ -5607,13 +5607,14 @@ static int pci_try_reset_bus(struct pci_bus *bus) * reset for affected devices * * This function will first try to reset the slots on this bus if the method is - * available. If slot reset fails or is not available, this will fall back to a + * available. If slot reset is not available, this will fall back to a * secondary bus reset. */ static int pci_reset_bridge(struct pci_dev *bridge, bool restore) { struct pci_bus *bus = bridge->subordinate; struct pci_slot *slot; + int ret = 0; if (!bus) return -ENOTTY; @@ -5627,19 +5628,17 @@ static int pci_reset_bridge(struct pci_dev *bridge, bool restore) goto bus_reset; list_for_each_entry(slot, &bus->slots, list) { - int ret; - if (restore) ret = pci_try_reset_slot(slot); else ret = pci_slot_reset(slot, PCI_RESET_DO_RESET); if (ret) - goto bus_reset; + break; } mutex_unlock(&pci_slot_mutex); - return 0; + return ret; bus_reset: mutex_unlock(&pci_slot_mutex); From a6715d7ec472a476db17787697a4abda62962284 Mon Sep 17 00:00:00 2001 From: Evangelos Petrongonas Date: Fri, 10 Apr 2026 01:16:05 +0000 Subject: [PATCH 024/333] kho: skip KHO for crash kernel kho_fill_kimage() unconditionally populates the kimage with KHO metadata for every kexec image type. When the image is a crash kernel, this can be problematic as the crash kernel can run in a small reserved region and the KHO scratch areas can sit outside it. The crash kernel then faults during kho_memory_init() when it tries phys_to_virt() on the KHO FDT address: Unable to handle kernel paging request at virtual address xxxxxxxx ... fdt_offset_ptr+... fdt_check_node_offset_+... fdt_first_property_offset+... fdt_get_property_namelen_+... fdt_getprop+... kho_memory_init+... mm_core_init+... start_kernel+... kho_locate_mem_hole() already skips KHO logic for KEXEC_TYPE_CRASH images, but kho_fill_kimage() was missing the same guard. As kho_fill_kimage() is the single point that populates image->kho.fdt and image->kho.scratch, fixing it here is sufficient for both arm64 and x86 as the FDT and boot_params path are bailing out when these fields are unset. Fixes: d7255959b69a ("kho: allow kexec load before KHO finalization") Signed-off-by: Evangelos Petrongonas Reviewed-by: Mike Rapoport (Microsoft) Link: https://patch.msgid.link/20260410011609.1103-1-epetron@amazon.de Signed-off-by: Mike Rapoport (Microsoft) --- kernel/liveupdate/kexec_handover.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/liveupdate/kexec_handover.c b/kernel/liveupdate/kexec_handover.c index 94762de1fe5f..4fde8325c49f 100644 --- a/kernel/liveupdate/kexec_handover.c +++ b/kernel/liveupdate/kexec_handover.c @@ -1702,7 +1702,7 @@ int kho_fill_kimage(struct kimage *image) int err = 0; struct kexec_buf scratch; - if (!kho_enable) + if (!kho_enable || image->type == KEXEC_TYPE_CRASH) return 0; image->kho.fdt = virt_to_phys(kho_out.fdt); From 0fb1daf0b78d0e23b63b6b65de56d4a3fd83bc14 Mon Sep 17 00:00:00 2001 From: David Carlier Date: Wed, 15 Apr 2026 06:23:00 +0100 Subject: [PATCH 025/333] mm/memfd_luo: report error when restoring a folio fails mid-loop memfd_luo_retrieve_folios() initialises err to -EIO, but the per-iteration calls to mem_cgroup_charge(), shmem_add_to_page_cache() and shmem_inode_acct_blocks() reuse and overwrite err. Once any iteration completes successfully, err becomes zero. If a later iteration's kho_restore_folio() returns NULL, the failure path jumps to put_folios without resetting err, so the function returns 0. The caller memfd_luo_retrieve() then takes the success path, sets args->file and reports the restore as successful, leaving userspace with a partially populated memfd and no indication that anything went wrong. Set err to -EIO in the kho_restore_folio() failure branch so the error is propagated to the caller. Signed-off-by: David Carlier Reviewed-by: Pratyush Yadav Fixes: b3749f174d68 ("mm: memfd_luo: allow preserving memfd") Link: https://patch.msgid.link/20260415052300.362539-1-devnexen@gmail.com Signed-off-by: Mike Rapoport (Microsoft) --- mm/memfd_luo.c | 1 + 1 file changed, 1 insertion(+) diff --git a/mm/memfd_luo.c b/mm/memfd_luo.c index b02b503c750d..35d1247281e0 100644 --- a/mm/memfd_luo.c +++ b/mm/memfd_luo.c @@ -427,6 +427,7 @@ static int memfd_luo_retrieve_folios(struct file *file, if (!folio) { pr_err("Unable to restore folio at physical address: %llx\n", phys); + err = -EIO; goto put_folios; } index = pfolio->index; From 3e75021f615ceee8562e6455c335936b39929ffb Mon Sep 17 00:00:00 2001 From: Troy Mitchell Date: Fri, 24 Apr 2026 16:20:32 +0800 Subject: [PATCH 026/333] clk: spacemit: k3: mark top_dclk as CLK_IS_CRITICAL top_dclk is the DDR bus clock. If it is gated by clk_disable_unused, all memory-mapped bus transactions cease to function, causing DMA engines to hang and general system instability. Mark it CLK_IS_CRITICAL so the CCF never gates it during the unused clock sweep. Fixes: e371a77255b8 ("clk: spacemit: k3: add the clock tree") Reviewed-by: Brian Masney Signed-off-by: Troy Mitchell Signed-off-by: Stephen Boyd --- drivers/clk/spacemit/ccu-k3.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/clk/spacemit/ccu-k3.c b/drivers/clk/spacemit/ccu-k3.c index e98afd59f05c..bb8b75bdbdb3 100644 --- a/drivers/clk/spacemit/ccu-k3.c +++ b/drivers/clk/spacemit/ccu-k3.c @@ -846,7 +846,7 @@ static const struct clk_parent_data top_parents[] = { CCU_PARENT_HW(pll6_d3), }; CCU_MUX_DIV_GATE_FC_DEFINE(top_dclk, top_parents, APMU_TOP_DCLK_CTRL, 5, 3, - BIT(8), 2, 3, BIT(1), 0); + BIT(8), 2, 3, BIT(1), CLK_IS_CRITICAL); static const struct clk_parent_data ucie_parents[] = { CCU_PARENT_HW(pll1_d8_307p2), From 79a1886be1564a009cd2a003bada15ed6153f819 Mon Sep 17 00:00:00 2001 From: Jerome Brunet Date: Wed, 25 Feb 2026 17:55:05 +0100 Subject: [PATCH 027/333] clk: eyeq: use the auxiliary device creation helper MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The auxiliary device creation of this driver is simple enough to use the available auxiliary device creation helper. Use it and remove some boilerplate code. Tested-by: Théo Lebrun # On Mobileye EyeQ5 Signed-off-by: Jerome Brunet Reviewed-by: Luca Ceresoli Signed-off-by: Théo Lebrun Signed-off-by: Stephen Boyd --- drivers/clk/clk-eyeq.c | 57 +++++++++--------------------------------- 1 file changed, 12 insertions(+), 45 deletions(-) diff --git a/drivers/clk/clk-eyeq.c b/drivers/clk/clk-eyeq.c index c1dccedf8d5b..ae1aaf274336 100644 --- a/drivers/clk/clk-eyeq.c +++ b/drivers/clk/clk-eyeq.c @@ -321,38 +321,18 @@ static void eqc_probe_init_fixed_factors(struct device *dev, } } -static void eqc_auxdev_release(struct device *dev) -{ - struct auxiliary_device *adev = to_auxiliary_dev(dev); - - kfree(adev); -} - -static int eqc_auxdev_create(struct device *dev, void __iomem *base, - const char *name, u32 id) +static void eqc_auxdev_create_optional(struct device *dev, void __iomem *base, + const char *name) { struct auxiliary_device *adev; - int ret; - adev = kzalloc_obj(*adev); - if (!adev) - return -ENOMEM; - - adev->name = name; - adev->dev.parent = dev; - adev->dev.platform_data = (void __force *)base; - adev->dev.release = eqc_auxdev_release; - adev->id = id; - - ret = auxiliary_device_init(adev); - if (ret) - return ret; - - ret = auxiliary_device_add(adev); - if (ret) - auxiliary_device_uninit(adev); - - return ret; + if (name) { + adev = devm_auxiliary_device_create(dev, name, + (void __force *)base); + if (!adev) + dev_warn(dev, "failed creating auxiliary device %s.%s\n", + KBUILD_MODNAME, name); + } } static int eqc_probe(struct platform_device *pdev) @@ -364,7 +344,6 @@ static int eqc_probe(struct platform_device *pdev) unsigned int i, clk_count; struct resource *res; void __iomem *base; - int ret; data = device_get_match_data(dev); if (!data) @@ -378,21 +357,9 @@ static int eqc_probe(struct platform_device *pdev) if (!base) return -ENOMEM; - /* Init optional reset auxiliary device. */ - if (data->reset_auxdev_name) { - ret = eqc_auxdev_create(dev, base, data->reset_auxdev_name, 0); - if (ret) - dev_warn(dev, "failed creating auxiliary device %s.%s: %d\n", - KBUILD_MODNAME, data->reset_auxdev_name, ret); - } - - /* Init optional pinctrl auxiliary device. */ - if (data->pinctrl_auxdev_name) { - ret = eqc_auxdev_create(dev, base, data->pinctrl_auxdev_name, 0); - if (ret) - dev_warn(dev, "failed creating auxiliary device %s.%s: %d\n", - KBUILD_MODNAME, data->pinctrl_auxdev_name, ret); - } + /* Init optional auxiliary devices. */ + eqc_auxdev_create_optional(dev, base, data->reset_auxdev_name); + eqc_auxdev_create_optional(dev, base, data->pinctrl_auxdev_name); if (data->pll_count + data->div_count + data->fixed_factor_count == 0) return 0; /* Zero clocks, we are done. */ From a25ab518f355e1f0dcbea24ee26418dfcd6944b5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Th=C3=A9o=20Lebrun?= Date: Wed, 25 Feb 2026 17:55:06 +0100 Subject: [PATCH 028/333] clk: eyeq: add EyeQ5 children auxiliary device for generic PHYs MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Grow our clk-eyeq family; it knows how to spawn reset provider and pin controller children. Expand with a generic PHY driver on EyeQ5. Reviewed-by: Luca Ceresoli Signed-off-by: Théo Lebrun Signed-off-by: Stephen Boyd --- drivers/clk/clk-eyeq.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/clk/clk-eyeq.c b/drivers/clk/clk-eyeq.c index ae1aaf274336..d9303c2c7aa5 100644 --- a/drivers/clk/clk-eyeq.c +++ b/drivers/clk/clk-eyeq.c @@ -110,6 +110,7 @@ struct eqc_match_data { const char *reset_auxdev_name; const char *pinctrl_auxdev_name; + const char *eth_phy_auxdev_name; unsigned int early_clk_count; }; @@ -360,6 +361,7 @@ static int eqc_probe(struct platform_device *pdev) /* Init optional auxiliary devices. */ eqc_auxdev_create_optional(dev, base, data->reset_auxdev_name); eqc_auxdev_create_optional(dev, base, data->pinctrl_auxdev_name); + eqc_auxdev_create_optional(dev, base, data->eth_phy_auxdev_name); if (data->pll_count + data->div_count + data->fixed_factor_count == 0) return 0; /* Zero clocks, we are done. */ @@ -520,6 +522,7 @@ static const struct eqc_match_data eqc_eyeq5_match_data = { .reset_auxdev_name = "reset", .pinctrl_auxdev_name = "pinctrl", + .eth_phy_auxdev_name = "phy", .early_clk_count = ARRAY_SIZE(eqc_eyeq5_early_plls) + ARRAY_SIZE(eqc_eyeq5_early_fixed_factors), From 4ac170432cf74b753cf59bcd0d449dced48585da Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Th=C3=A9o=20Lebrun?= Date: Wed, 25 Feb 2026 17:55:07 +0100 Subject: [PATCH 029/333] reset: eyeq: drop device_set_of_node_from_dev() done by parent MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Our parent driver (clk-eyeq) now does the device_set_of_node_from_dev(dev, dev->parent) call through the newly introduced devm_auxiliary_device_create() helper. Doing it again in the reset-eyeq probe would be redundant. Drop both the WARN_ON() and the device_set_of_node_from_dev() call. Also fix the following comment that talks about "our newfound OF node". Signed-off-by: Jerome Brunet Reviewed-by: Philipp Zabel Acked-by: Philipp Zabel Signed-off-by: Théo Lebrun Signed-off-by: Stephen Boyd --- drivers/reset/reset-eyeq.c | 24 ++---------------------- 1 file changed, 2 insertions(+), 22 deletions(-) diff --git a/drivers/reset/reset-eyeq.c b/drivers/reset/reset-eyeq.c index 791b7283111e..1a3857983897 100644 --- a/drivers/reset/reset-eyeq.c +++ b/drivers/reset/reset-eyeq.c @@ -422,13 +422,6 @@ static int eqr_of_xlate_twocells(struct reset_controller_dev *rcdev, return eqr_of_xlate_internal(rcdev, reset_spec->args[0], reset_spec->args[1]); } -static void eqr_of_node_put(void *_dev) -{ - struct device *dev = _dev; - - of_node_put(dev->of_node); -} - static int eqr_probe(struct auxiliary_device *adev, const struct auxiliary_device_id *id) { @@ -439,21 +432,8 @@ static int eqr_probe(struct auxiliary_device *adev, int ret; /* - * We are an auxiliary device of clk-eyeq. We do not have an OF node by - * default; let's reuse our parent's OF node. - */ - WARN_ON(dev->of_node); - device_set_of_node_from_dev(dev, dev->parent); - if (!dev->of_node) - return -ENODEV; - - ret = devm_add_action_or_reset(dev, eqr_of_node_put, dev); - if (ret) - return ret; - - /* - * Using our newfound OF node, we can get match data. We cannot use - * device_get_match_data() because it does not match reused OF nodes. + * Get match data. We cannot use device_get_match_data() because it does + * not accept reused OF nodes; see device_set_of_node_from_dev(). */ match = of_match_node(dev->driver->of_match_table, dev->of_node); if (!match || !match->data) From 69056753231f483cc1e40db52228aac42fd4c93d Mon Sep 17 00:00:00 2001 From: Brian Masney Date: Wed, 15 Apr 2026 16:30:49 -0400 Subject: [PATCH 030/333] MAINTAINERS: add myself as a reviewer for the clk subsystem I've reviewed a lot clk patches for parts of the subsystem that typically doesn't get much review. Add myself as a reviewer so that I don't miss anything. Link: https://lore.kernel.org/linux-clk/?q=f%3Abmasney%40redhat.com Signed-off-by: Brian Masney Signed-off-by: Stephen Boyd --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 2fb1c75afd16..384ce5ee7323 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -6358,6 +6358,7 @@ F: include/uapi/linux/comedi.h COMMON CLK FRAMEWORK M: Michael Turquette M: Stephen Boyd +R: Brian Masney L: linux-clk@vger.kernel.org S: Maintained Q: http://patchwork.kernel.org/project/linux-clk/list/ From de019f203b0d472c98ead4081ad4f05d92c9b826 Mon Sep 17 00:00:00 2001 From: Johan Hovold Date: Tue, 7 Apr 2026 11:50:27 +0200 Subject: [PATCH 031/333] clk: rk808: fix OF node reference imbalance The driver reuses the OF node of the parent multi-function device but fails to take another reference to balance the one dropped by the platform bus code when unbinding the MFD and deregistering the child devices. Fix this by using the intended helper for reusing OF nodes. Fixes: 2dc51ca822e4 ("clk: RK808: Reduce 'struct rk808' usage") Cc: stable@vger.kernel.org # 6.5 Cc: Sebastian Reichel Signed-off-by: Johan Hovold Reviewed-by: Sebastian Reichel Reviewed-by: Brian Masney Reviewed-by: Heiko Stuebner Signed-off-by: Stephen Boyd --- drivers/clk/clk-rk808.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/clk/clk-rk808.c b/drivers/clk/clk-rk808.c index f7412b137e5e..5a75b5c91555 100644 --- a/drivers/clk/clk-rk808.c +++ b/drivers/clk/clk-rk808.c @@ -153,7 +153,7 @@ static int rk808_clkout_probe(struct platform_device *pdev) struct rk808_clkout *rk808_clkout; int ret; - dev->of_node = pdev->dev.parent->of_node; + device_set_of_node_from_dev(dev, dev->parent); rk808_clkout = devm_kzalloc(dev, sizeof(*rk808_clkout), GFP_KERNEL); From 100201d349edd226ca3470c894c92dccc67ee7a8 Mon Sep 17 00:00:00 2001 From: Fabio Porcedda Date: Mon, 27 Apr 2026 11:17:46 +0200 Subject: [PATCH 032/333] USB: serial: option: add Telit Cinterion LE910Cx compositions Add the following Telit Cinterion LE910Cx compositions: 0x1251: RNDIS + tty (AT/NMEA) + tty (AT) + tty (AT) + tty (SAP) T: Bus=01 Lev=01 Prnt=21 Port=06 Cnt=01 Dev#=108 Spd=480 MxCh= 0 D: Ver= 2.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=1bc7 ProdID=1251 Rev=03.18 S: Manufacturer=Android S: Product=LE910C1-EU S: SerialNumber=0123456789ABCDEF C: #Ifs= 6 Cfg#= 1 Atr=a0 MxPwr=500mA I: If#= 0 Alt= 0 #EPs= 1 Cls=02(commc) Sub=02 Prot=ff Driver=rndis_host E: Ad=82(I) Atr=03(Int.) MxPS= 8 Ivl=32ms I: If#= 1 Alt= 0 #EPs= 2 Cls=0a(data ) Sub=00 Prot=00 Driver=rndis_host E: Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms I: If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=83(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=84(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=85(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=86(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 4 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=87(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=88(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 5 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=05(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=89(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=8a(I) Atr=03(Int.) MxPS= 10 Ivl=32ms 0x1253: ECM + tty (AT/NMEA) + tty (AT) + tty (AT) + tty (SAP) T: Bus=01 Lev=01 Prnt=21 Port=06 Cnt=01 Dev#=121 Spd=480 MxCh= 0 D: Ver= 2.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=1bc7 ProdID=1253 Rev=03.18 S: Manufacturer=Android S: Product=LE910C1-EU S: SerialNumber=0123456789ABCDEF C: #Ifs= 6 Cfg#= 1 Atr=a0 MxPwr=500mA I: If#= 0 Alt= 0 #EPs= 1 Cls=02(commc) Sub=06 Prot=00 Driver=cdc_ether E: Ad=82(I) Atr=03(Int.) MxPS= 16 Ivl=32ms I: If#= 1 Alt= 1 #EPs= 2 Cls=0a(data ) Sub=00 Prot=00 Driver=cdc_ether E: Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms I: If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=83(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=84(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=85(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=86(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 4 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=87(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=88(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 5 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=05(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=89(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=8a(I) Atr=03(Int.) MxPS= 10 Ivl=32ms 0x1254: tty (AT) + tty (AT) T: Bus=01 Lev=01 Prnt=21 Port=06 Cnt=01 Dev#=122 Spd=480 MxCh= 0 D: Ver= 2.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=1bc7 ProdID=1254 Rev=03.18 S: Manufacturer=Android S: Product=LE910C1-EU S: SerialNumber=0123456789ABCDEF C: #Ifs= 2 Cfg#= 1 Atr=a0 MxPwr=500mA I: If#= 0 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=82(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 1 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=83(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=84(I) Atr=03(Int.) MxPS= 10 Ivl=32ms 0x1255: tty (AT/NMEA) + tty (AT) + tty (AT) + tty (SAP) T: Bus=01 Lev=01 Prnt=21 Port=06 Cnt=01 Dev#=123 Spd=480 MxCh= 0 D: Ver= 2.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs= 1 P: Vendor=1bc7 ProdID=1255 Rev=03.18 S: Manufacturer=Android S: Product=LE910C1-EU S: SerialNumber=0123456789ABCDEF C: #Ifs= 4 Cfg#= 1 Atr=a0 MxPwr=500mA I: If#= 0 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=82(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 1 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=83(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=84(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=85(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=86(I) Atr=03(Int.) MxPS= 10 Ivl=32ms I: If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option E: Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=87(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms E: Ad=88(I) Atr=03(Int.) MxPS= 10 Ivl=32ms Cc: stable@vger.kernel.org Signed-off-by: Fabio Porcedda Signed-off-by: Johan Hovold --- drivers/usb/serial/option.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/usb/serial/option.c b/drivers/usb/serial/option.c index c71461893d20..42e4cecd28ac 100644 --- a/drivers/usb/serial/option.c +++ b/drivers/usb/serial/option.c @@ -1513,7 +1513,11 @@ static const struct usb_device_id option_ids[] = { { USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1231, 0xff), /* Telit LE910Cx (RNDIS) */ .driver_info = NCTRL(2) | RSVD(3) }, { USB_DEVICE_AND_INTERFACE_INFO(TELIT_VENDOR_ID, 0x1250, 0xff, 0x00, 0x00) }, /* Telit LE910Cx (rmnet) */ + { USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1251, 0xff) }, /* Telit LE910Cx (RNDIS) */ { USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1252, 0xff) }, /* Telit LE910Cx (MBIM) */ + { USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1253, 0xff) }, /* Telit LE910Cx (ECM) */ + { USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1254, 0xff) }, /* Telit LE910Cx */ + { USB_DEVICE_INTERFACE_CLASS(TELIT_VENDOR_ID, 0x1255, 0xff) }, /* Telit LE910Cx */ { USB_DEVICE(TELIT_VENDOR_ID, 0x1260), .driver_info = NCTRL(0) | RSVD(1) | RSVD(2) }, { USB_DEVICE(TELIT_VENDOR_ID, 0x1261), From 7dd57d7a6350770dfc283287125c409e995200e0 Mon Sep 17 00:00:00 2001 From: Karol Wachowski Date: Thu, 30 Apr 2026 11:56:44 +0200 Subject: [PATCH 033/333] accel/ivpu: Disallow re-exporting imported GEM objects Prevent re-exporting of imported GEM buffers by adding a custom prime_handle_to_fd callback that checks if the object is imported and returns -EOPNOTSUPP if so. Re-exporting imported GEM buffers causes loss of buffer flags settings, leading to incorrect device access and data corruption. Reported-by: Yametsu Fixes: 57557964b582 ("accel/ivpu: Add support for userptr buffer objects") Reviewed-by: Andrzej Kacprowski Signed-off-by: Karol Wachowski Cc: # v6.19+ --- drivers/accel/ivpu/ivpu_drv.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) diff --git a/drivers/accel/ivpu/ivpu_drv.c b/drivers/accel/ivpu/ivpu_drv.c index 2801378e3e19..3b7b008bccfe 100644 --- a/drivers/accel/ivpu/ivpu_drv.c +++ b/drivers/accel/ivpu/ivpu_drv.c @@ -537,6 +537,26 @@ static const struct file_operations ivpu_fops = { #endif }; +static int ivpu_gem_prime_handle_to_fd(struct drm_device *dev, struct drm_file *file_priv, + u32 handle, u32 flags, int *prime_fd) +{ + struct drm_gem_object *obj; + + obj = drm_gem_object_lookup(file_priv, handle); + if (!obj) + return -ENOENT; + + if (drm_gem_is_imported(obj)) { + /* Do not allow re-exporting */ + drm_gem_object_put(obj); + return -EOPNOTSUPP; + } + + drm_gem_object_put(obj); + + return drm_gem_prime_handle_to_fd(dev, file_priv, handle, flags, prime_fd); +} + static const struct drm_driver driver = { .driver_features = DRIVER_GEM | DRIVER_COMPUTE_ACCEL, @@ -545,6 +565,7 @@ static const struct drm_driver driver = { .gem_create_object = ivpu_gem_create_object, .gem_prime_import = ivpu_gem_prime_import, + .prime_handle_to_fd = ivpu_gem_prime_handle_to_fd, .ioctls = ivpu_drm_ioctls, .num_ioctls = ARRAY_SIZE(ivpu_drm_ioctls), From d6cc7c99bf1f73eda7d565d224d791d16239bb41 Mon Sep 17 00:00:00 2001 From: Sanman Pradhan Date: Thu, 16 Apr 2026 21:59:30 +0000 Subject: [PATCH 034/333] hwmon: (ltc2992) Clamp threshold writes to hardware range ltc2992_set_voltage(), ltc2992_set_current(), and ltc2992_set_power() do not validate the user-supplied value before converting it to a register value. This can result in: 1. Negative input values wrapping to large positive register values. For power, the negative long is implicitly cast to u64 in mul_u64_u32_div(), producing an incorrect value. For voltage and current, the negative converted value wraps when passed to ltc2992_write_reg() as a u32. 2. Intermediate arithmetic exceeding the range representable in u64 on 64-bit platforms. In ltc2992_set_voltage(), (u64)val * 1000 can exceed U64_MAX when val is a large positive long. In ltc2992_set_current(), (u64)val * r_sense_uohm can overflow similarly. In ltc2992_set_power(), the computed value may not fit in u64. 3. Register values exceeding the hardware field width. Voltage and current threshold registers are 12-bit (stored left-justified in 16 bits), and power threshold registers are 24-bit. Without clamping, bits above the field width are truncated in ltc2992_write_reg(). Fix by clamping negative values to zero, clamping positive values to the rounded hardware-representable maximum (the value returned by the read path for a full-scale register) to prevent intermediate overflow, and clamping the converted register value to the hardware field width before writing. The existing conversion formula and rounding behavior are preserved. In the power write path, cancel the factor of 1000 from both the numerator (r_sense_uohm * 1000) and the denominator (VADC_UV_LSB * IADC_NANOV_LSB) to also eliminate a u32 overflow of r_sense_uohm * 1000 when r_sense_uohm exceeds about 4.29 ohms. Fixes: b0bd407e94b03 ("hwmon: (ltc2992) Add support") Cc: stable@vger.kernel.org Signed-off-by: Sanman Pradhan Link: https://lore.kernel.org/r/20260416215904.101969-2-sanman.pradhan@hpe.com Signed-off-by: Guenter Roeck --- drivers/hwmon/ltc2992.c | 35 ++++++++++++++++++++++++++++------- 1 file changed, 28 insertions(+), 7 deletions(-) diff --git a/drivers/hwmon/ltc2992.c b/drivers/hwmon/ltc2992.c index 1fcd320d6161..106973619676 100644 --- a/drivers/hwmon/ltc2992.c +++ b/drivers/hwmon/ltc2992.c @@ -431,10 +431,16 @@ static int ltc2992_get_voltage(struct ltc2992_state *st, u32 reg, u32 scale, lon static int ltc2992_set_voltage(struct ltc2992_state *st, u32 reg, u32 scale, long val) { - val = DIV_ROUND_CLOSEST(val * 1000, scale); - val = val << 4; + u32 reg_val; + long vmax; - return ltc2992_write_reg(st, reg, 2, val); + vmax = DIV_ROUND_CLOSEST_ULL(0xFFFULL * scale, 1000); + val = max(val, 0L); + val = min(val, vmax); + reg_val = min(DIV_ROUND_CLOSEST_ULL((u64)val * 1000, scale), + 0xFFFULL) << 4; + + return ltc2992_write_reg(st, reg, 2, reg_val); } static int ltc2992_read_gpio_alarm(struct ltc2992_state *st, int nr_gpio, u32 attr, long *val) @@ -559,9 +565,15 @@ static int ltc2992_get_current(struct ltc2992_state *st, u32 reg, u32 channel, l static int ltc2992_set_current(struct ltc2992_state *st, u32 reg, u32 channel, long val) { u32 reg_val; + long cmax; - reg_val = DIV_ROUND_CLOSEST(val * st->r_sense_uohm[channel], LTC2992_IADC_NANOV_LSB); - reg_val = reg_val << 4; + cmax = DIV_ROUND_CLOSEST_ULL(0xFFFULL * LTC2992_IADC_NANOV_LSB, + st->r_sense_uohm[channel]); + val = max(val, 0L); + val = min(val, cmax); + reg_val = min(DIV_ROUND_CLOSEST_ULL((u64)val * st->r_sense_uohm[channel], + LTC2992_IADC_NANOV_LSB), + 0xFFFULL) << 4; return ltc2992_write_reg(st, reg, 2, reg_val); } @@ -634,9 +646,18 @@ static int ltc2992_get_power(struct ltc2992_state *st, u32 reg, u32 channel, lon static int ltc2992_set_power(struct ltc2992_state *st, u32 reg, u32 channel, long val) { u32 reg_val; + u64 pmax, uval; - reg_val = mul_u64_u32_div(val, st->r_sense_uohm[channel] * 1000, - LTC2992_VADC_UV_LSB * LTC2992_IADC_NANOV_LSB); + uval = max(val, 0L); + pmax = mul_u64_u32_div(0xFFFFFFULL, + LTC2992_VADC_UV_LSB / 1000 * + LTC2992_IADC_NANOV_LSB, + st->r_sense_uohm[channel]); + uval = min(uval, pmax); + reg_val = min(mul_u64_u32_div(uval, st->r_sense_uohm[channel], + LTC2992_VADC_UV_LSB / 1000 * + LTC2992_IADC_NANOV_LSB), + 0xFFFFFFULL); return ltc2992_write_reg(st, reg, 3, reg_val); } From 2da0c1fd01dbd6b22844e8676585153dfc660cbe Mon Sep 17 00:00:00 2001 From: Sanman Pradhan Date: Thu, 16 Apr 2026 21:59:40 +0000 Subject: [PATCH 035/333] hwmon: (ltc2992) Fix u32 overflow in power read path ltc2992_get_power() computes the divisor for mul_u64_u32_div() as r_sense_uohm * 1000. This multiplication overflows u32 when r_sense_uohm exceeds about 4.29 ohms (4294967 micro-ohms), producing a truncated divisor and an incorrect power reading. Cancel the factor of 1000 from both the numerator (VADC_UV_LSB * IADC_NANOV_LSB = 312500000) and the divisor (r_sense_uohm * 1000), giving (VADC_UV_LSB / 1000) * IADC_NANOV_LSB = 312500 as the numerator and plain r_sense_uohm as the divisor. The cancellation is exact because LTC2992_VADC_UV_LSB (25000) is divisible by 1000. This is the read-path counterpart of the write-path fix applied in the preceding patch. Fixes: b0bd407e94b03 ("hwmon: (ltc2992) Add support") Cc: stable@vger.kernel.org Signed-off-by: Sanman Pradhan Link: https://lore.kernel.org/r/20260416215904.101969-3-sanman.pradhan@hpe.com Signed-off-by: Guenter Roeck --- drivers/hwmon/ltc2992.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/hwmon/ltc2992.c b/drivers/hwmon/ltc2992.c index 106973619676..2617c4538af9 100644 --- a/drivers/hwmon/ltc2992.c +++ b/drivers/hwmon/ltc2992.c @@ -637,8 +637,10 @@ static int ltc2992_get_power(struct ltc2992_state *st, u32 reg, u32 channel, lon if (reg_val < 0) return reg_val; - *val = mul_u64_u32_div(reg_val, LTC2992_VADC_UV_LSB * LTC2992_IADC_NANOV_LSB, - st->r_sense_uohm[channel] * 1000); + *val = mul_u64_u32_div(reg_val, + LTC2992_VADC_UV_LSB / 1000 * + LTC2992_IADC_NANOV_LSB, + st->r_sense_uohm[channel]); return 0; } From 5ed26ffe57ffca054b7a141ff0c1a07bf3a80f6c Mon Sep 17 00:00:00 2001 From: Ninad Naik Date: Sat, 18 Apr 2026 00:44:11 +0530 Subject: [PATCH 036/333] Documentation: hwmon: fix link to ideapad-laptop.c file The ideapad-laptop.c file now exists inside drivers/platform/x86/lenovo/ directory. Updating the GitHub link to the correct path. Signed-off-by: Ninad Naik Link: https://lore.kernel.org/r/20260417191411.713958-1-ninadnaik07@gmail.com Signed-off-by: Guenter Roeck --- Documentation/hwmon/yogafan.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/hwmon/yogafan.rst b/Documentation/hwmon/yogafan.rst index c553a381f772..68761947a1a8 100644 --- a/Documentation/hwmon/yogafan.rst +++ b/Documentation/hwmon/yogafan.rst @@ -135,4 +135,4 @@ References 4. **Lenovo IdeaPad Laptop Driver:** Reference for DMI-based hardware feature gating in Lenovo laptops. - https://github.com/torvalds/linux/blob/master/drivers/platform/x86/ideapad-laptop.c + https://github.com/torvalds/linux/blob/master/drivers/platform/x86/lenovo/ideapad-laptop.c From 1a1414c675ee1b637bbe3840241555a49c61b123 Mon Sep 17 00:00:00 2001 From: Sasha Levin Date: Sat, 25 Apr 2026 20:03:19 -0400 Subject: [PATCH 037/333] hwmon: Remove stale CONFIG_SENSORS_SBRMI Makefile reference MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit kconfiglint reports: X001: CONFIG_SENSORS_SBRMI referenced in Makefile but not defined in any Kconfig The SB-RMI hardware monitoring driver was originally introduced in commit 5a0f50d110b3 ("hwmon: Add support for SB-RMI power module") with both a Kconfig entry (CONFIG_SENSORS_SBRMI) and a Makefile line (obj-$(CONFIG_SENSORS_SBRMI) += sbrmi.o) in drivers/hwmon/. Commit e156586764050 ("hwmon/misc: amd-sbi: Move core sbrmi from hwmon to misc") moved the driver to drivers/misc/amd-sbi/ to support additional functionality beyond hardware monitoring. That commit correctly removed the Kconfig entry from drivers/hwmon/Kconfig, moved the source file drivers/hwmon/sbrmi.c to drivers/misc/amd-sbi/sbrmi.c, and created new Kconfig/Makefile entries in drivers/misc/amd-sbi/ with a renamed symbol (CONFIG_AMD_SBRMI_I2C). However, the Makefile line in drivers/hwmon/Makefile was not removed in that commit. The orphaned line references a CONFIG symbol that no longer exists and a source file that is no longer present, so it has no effect on the build — but it is dead code that should be cleaned up. Remove the stale Makefile reference. Assisted-by: Claude:claude-opus-4-6 kconfiglint Signed-off-by: Sasha Levin Link: https://lore.kernel.org/r/20260426000319.55908-1-sashal@kernel.org Signed-off-by: Guenter Roeck --- drivers/hwmon/Makefile | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/hwmon/Makefile b/drivers/hwmon/Makefile index 4788996aa137..982ee2c6f9de 100644 --- a/drivers/hwmon/Makefile +++ b/drivers/hwmon/Makefile @@ -201,7 +201,6 @@ obj-$(CONFIG_SENSORS_PWM_FAN) += pwm-fan.o obj-$(CONFIG_SENSORS_QNAP_MCU_HWMON) += qnap-mcu-hwmon.o obj-$(CONFIG_SENSORS_RASPBERRYPI_HWMON) += raspberrypi-hwmon.o obj-$(CONFIG_SENSORS_SBTSI) += sbtsi_temp.o -obj-$(CONFIG_SENSORS_SBRMI) += sbrmi.o obj-$(CONFIG_SENSORS_SCH56XX_COMMON)+= sch56xx-common.o obj-$(CONFIG_SENSORS_SCH5627) += sch5627.o obj-$(CONFIG_SENSORS_SCH5636) += sch5636.o From 174606451fbb17db506ebaacdd5e203e57773d5f Mon Sep 17 00:00:00 2001 From: Myeonghun Pak Date: Fri, 24 Apr 2026 22:50:51 +0900 Subject: [PATCH 038/333] hwmon: (corsair-psu) Close HID device on probe errors corsairpsu_probe() opens the HID device before sending the device init and firmware-info commands. If either command fails, the error path jumps directly to fail_and_stop and skips hid_hw_close(). Use the existing fail_and_close label for those post-open failures so the open count and low-level close callback are balanced before hid_hw_stop(). Fixes: d115b51e0e56 ("hwmon: add Corsair PSU HID controller driver") Cc: stable@vger.kernel.org Signed-off-by: Myeonghun Pak Reviewed-by: Wilken Gottwalt Link: https://lore.kernel.org/r/20260424135107.13720-1-mhun512@gmail.com Signed-off-by: Guenter Roeck --- drivers/hwmon/corsair-psu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/hwmon/corsair-psu.c b/drivers/hwmon/corsair-psu.c index dddbd2463f8d..76f3e1da68d0 100644 --- a/drivers/hwmon/corsair-psu.c +++ b/drivers/hwmon/corsair-psu.c @@ -796,13 +796,13 @@ static int corsairpsu_probe(struct hid_device *hdev, const struct hid_device_id ret = corsairpsu_init(priv); if (ret < 0) { dev_err(&hdev->dev, "unable to initialize device (%d)\n", ret); - goto fail_and_stop; + goto fail_and_close; } ret = corsairpsu_fwinfo(priv); if (ret < 0) { dev_err(&hdev->dev, "unable to query firmware (%d)\n", ret); - goto fail_and_stop; + goto fail_and_close; } corsairpsu_get_criticals(priv); From ba6b328588f7cb7cf4aca33bae565c2914d74786 Mon Sep 17 00:00:00 2001 From: David Gow Date: Sat, 25 Apr 2026 11:41:23 +0800 Subject: [PATCH 039/333] rust: arch: um: Fix building 32-bit UML with GCC 32-bit UML builds can be configured either by setting CONFIG_64BIT=n or with SUBARCH=i386. Both work with Rust-for-Linux when clang is the compiler, but when SUBARCH=i386, we don't set a bindgen target correctly if gcc is the compiler. Add the appropriate bindgen target configuration for i386, as is done in Makefile.clang. [ For reference, the errors look like: BINDGEN rust/bindings/bindings_generated.rs error: unsupported option '-mno-sse' for target '' ... error: unknown target triple 'unknown' panicked at .../bindgen-0.72.1/ir/context.rs:562:15: libclang error; possible causes include: ... - Miguel ] Fixes: ab0f4cedc355 ("arch: um: rust: Add i386 support for Rust") Signed-off-by: David Gow Link: https://patch.msgid.link/20260425034125.53866-1-david@davidgow.net [ Added space in title. - Miguel ] Signed-off-by: Miguel Ojeda --- rust/Makefile | 2 ++ 1 file changed, 2 insertions(+) diff --git a/rust/Makefile b/rust/Makefile index b361bfedfdf0..b9e9f512cec3 100644 --- a/rust/Makefile +++ b/rust/Makefile @@ -403,6 +403,8 @@ BINDGEN_TARGET_x86 := x86_64-linux-gnu BINDGEN_TARGET_arm64 := aarch64-linux-gnu BINDGEN_TARGET_arm := arm-linux-gnueabi BINDGEN_TARGET_loongarch := loongarch64-linux-gnusf +# This is only for i386 UM builds, which need the 32-bit target not -m32 +BINDGEN_TARGET_i386 := i386-linux-gnu BINDGEN_TARGET_um := $(BINDGEN_TARGET_$(SUBARCH)) BINDGEN_TARGET := $(BINDGEN_TARGET_$(SRCARCH)) From 83ac2870310b694775ab7e8f0244fdd94fc21926 Mon Sep 17 00:00:00 2001 From: Gary Guo Date: Mon, 27 Apr 2026 16:43:00 +0100 Subject: [PATCH 040/333] rust: pin-init: internal: move alignment check to `make_field_check` Instead of having the reference creation serving dual-purpose as both for let bindings and alignment check, detangle them so that the alignment check is done explicitly in `make_field_check`. This is more robust against refactors that may change the way let bindings are created. Cc: stable@vger.kernel.org Reviewed-by: Alice Ryhl Signed-off-by: Gary Guo Link: https://patch.msgid.link/20260427-pin-init-fix-v3-1-496a699674dd@garyguo.net [ Reworded for typo. - Miguel ] Signed-off-by: Miguel Ojeda --- rust/pin-init/internal/src/init.rs | 78 ++++++++++++++---------------- 1 file changed, 37 insertions(+), 41 deletions(-) diff --git a/rust/pin-init/internal/src/init.rs b/rust/pin-init/internal/src/init.rs index daa3f1c6466e..0a6600e8156c 100644 --- a/rust/pin-init/internal/src/init.rs +++ b/rust/pin-init/internal/src/init.rs @@ -249,10 +249,6 @@ fn init_fields( }); // Again span for better diagnostics let write = quote_spanned!(ident.span()=> ::core::ptr::write); - // NOTE: the field accessor ensures that the initialized field is properly aligned. - // Unaligned fields will cause the compiler to emit E0793. We do not support - // unaligned fields since `Init::__init` requires an aligned pointer; the call to - // `ptr::write` below has the same requirement. let accessor = if pinned { let project_ident = format_ident!("__project_{ident}"); quote! { @@ -367,49 +363,49 @@ fn init_fields( } } -/// Generate the check for ensuring that every field has been initialized. +/// Generate the check for ensuring that every field has been initialized and aligned. fn make_field_check( fields: &Punctuated, init_kind: InitKind, path: &Path, ) -> TokenStream { - let field_attrs = fields + let field_attrs: Vec<_> = fields .iter() - .filter_map(|f| f.kind.ident().map(|_| &f.attrs)); - let field_name = fields.iter().filter_map(|f| f.kind.ident()); - match init_kind { - InitKind::Normal => quote! { - // We use unreachable code to ensure that all fields have been mentioned exactly once, - // this struct initializer will still be type-checked and complain with a very natural - // error message if a field is forgotten/mentioned more than once. - #[allow(unreachable_code, clippy::diverging_sub_expression)] - // SAFETY: this code is never executed. - let _ = || unsafe { - ::core::ptr::write(slot, #path { - #( - #(#field_attrs)* - #field_name: ::core::panic!(), - )* - }) - }; - }, - InitKind::Zeroing => quote! { - // We use unreachable code to ensure that all fields have been mentioned at most once. - // Since the user specified `..Zeroable::zeroed()` at the end, all missing fields will - // be zeroed. This struct initializer will still be type-checked and complain with a - // very natural error message if a field is mentioned more than once, or doesn't exist. - #[allow(unreachable_code, clippy::diverging_sub_expression, unused_assignments)] - // SAFETY: this code is never executed. - let _ = || unsafe { - ::core::ptr::write(slot, #path { - #( - #(#field_attrs)* - #field_name: ::core::panic!(), - )* - ..::core::mem::zeroed() - }) - }; - }, + .filter_map(|f| f.kind.ident().map(|_| &f.attrs)) + .collect(); + let field_name: Vec<_> = fields.iter().filter_map(|f| f.kind.ident()).collect(); + let zeroing_trailer = match init_kind { + InitKind::Normal => None, + InitKind::Zeroing => Some(quote! { + ..::core::mem::zeroed() + }), + }; + quote! { + #[allow(unreachable_code, clippy::diverging_sub_expression)] + // We use unreachable code to perform field checks. They're still checked by the compiler. + // SAFETY: this code is never executed. + let _ = || unsafe { + // Create references to ensure that the initialized field is properly aligned. + // Unaligned fields will cause the compiler to emit E0793. We do not support + // unaligned fields since `Init::__init` requires an aligned pointer; the call to + // `ptr::write` for value-initialization case has the same requirement. + #( + #(#field_attrs)* + let _ = &(*slot).#field_name; + )* + + // If the zeroing trailer is not present, this checks that all fields have been + // mentioned exactly once. If the zeroing trailer is present, all missing fields will be + // zeroed, so this checks that all fields have been mentioned at most once. The use of + // struct initializer will still generate very natural error messages for any misuse. + ::core::ptr::write(slot, #path { + #( + #(#field_attrs)* + #field_name: ::core::panic!(), + )* + #zeroing_trailer + }) + }; } } From 68bf102226cf2199dc609b67c1e847cad4de4b57 Mon Sep 17 00:00:00 2001 From: Gary Guo Date: Mon, 27 Apr 2026 16:43:01 +0100 Subject: [PATCH 041/333] rust: pin-init: fix incorrect accessor reference lifetime When a field has been initialized, `init!`/`pin_init!` create a reference or pinned reference to the field so it can be accessed later during the initialization of other fields. However, the reference it created is incorrectly `&'static` rather than just the scope of the initializer. This means that you can do init!(Foo { a: 1, _: { let b: &'static u32 = a; } }) which is unsound. This is caused by `&mut (*#slot).#ident`, which actually allows arbitrary lifetime, so this is effectively `'static`. Somewhat ironically, the safety justification of creating the accessor is.. "SAFETY: TODO". Fix it by adding `let_binding` method on `DropGuard` to shorten lifetime. This results in exactly what we want for these accessors. The safety and invariant comments of `DropGuard` have been reworked; instead of reasoning about what caller can do with the guard, express it in a way that the ownership is transferred to the guard and `forget` takes it back, so the unsafe operations within the `DropGuard` can be more easily justified. Fixes: 42415d163e5d ("rust: pin-init: add references to previously initialized fields") Cc: stable@vger.kernel.org Signed-off-by: Gary Guo Link: https://patch.msgid.link/20260427-pin-init-fix-v3-2-496a699674dd@garyguo.net [ Reworded for missing word. - Miguel ] Signed-off-by: Miguel Ojeda --- rust/pin-init/internal/src/init.rs | 106 +++++++++++++---------------- rust/pin-init/src/__internal.rs | 28 +++++--- 2 files changed, 66 insertions(+), 68 deletions(-) diff --git a/rust/pin-init/internal/src/init.rs b/rust/pin-init/internal/src/init.rs index 0a6600e8156c..487ee0013faf 100644 --- a/rust/pin-init/internal/src/init.rs +++ b/rust/pin-init/internal/src/init.rs @@ -249,18 +249,6 @@ fn init_fields( }); // Again span for better diagnostics let write = quote_spanned!(ident.span()=> ::core::ptr::write); - let accessor = if pinned { - let project_ident = format_ident!("__project_{ident}"); - quote! { - // SAFETY: TODO - unsafe { #data.#project_ident(&mut (*#slot).#ident) } - } - } else { - quote! { - // SAFETY: TODO - unsafe { &mut (*#slot).#ident } - } - }; quote! { #(#attrs)* { @@ -268,51 +256,31 @@ fn init_fields( // SAFETY: TODO unsafe { #write(&raw mut (*#slot).#ident, #value_ident) }; } - #(#cfgs)* - #[allow(unused_variables)] - let #ident = #accessor; } } InitializerKind::Init { ident, value, .. } => { // Again span for better diagnostics let init = format_ident!("init", span = value.span()); - // NOTE: the field accessor ensures that the initialized field is properly aligned. - // Unaligned fields will cause the compiler to emit E0793. We do not support - // unaligned fields since `Init::__init` requires an aligned pointer; the call to - // `ptr::write` below has the same requirement. - let (value_init, accessor) = if pinned { - let project_ident = format_ident!("__project_{ident}"); - ( - quote! { - // SAFETY: - // - `slot` is valid, because we are inside of an initializer closure, we - // return when an error/panic occurs. - // - We also use `#data` to require the correct trait (`Init` or `PinInit`) - // for `#ident`. - unsafe { #data.#ident(&raw mut (*#slot).#ident, #init)? }; - }, - quote! { - // SAFETY: TODO - unsafe { #data.#project_ident(&mut (*#slot).#ident) } - }, - ) + let value_init = if pinned { + quote! { + // SAFETY: + // - `slot` is valid, because we are inside of an initializer closure, we + // return when an error/panic occurs. + // - We also use `#data` to require the correct trait (`Init` or `PinInit`) + // for `#ident`. + unsafe { #data.#ident(&raw mut (*#slot).#ident, #init)? }; + } } else { - ( - quote! { - // SAFETY: `slot` is valid, because we are inside of an initializer - // closure, we return when an error/panic occurs. - unsafe { - ::pin_init::Init::__init( - #init, - &raw mut (*#slot).#ident, - )? - }; - }, - quote! { - // SAFETY: TODO - unsafe { &mut (*#slot).#ident } - }, - ) + quote! { + // SAFETY: `slot` is valid, because we are inside of an initializer + // closure, we return when an error/panic occurs. + unsafe { + ::pin_init::Init::__init( + #init, + &raw mut (*#slot).#ident, + )? + }; + } }; quote! { #(#attrs)* @@ -320,9 +288,6 @@ fn init_fields( let #init = #value; #value_init } - #(#cfgs)* - #[allow(unused_variables)] - let #ident = #accessor; } } InitializerKind::Code { block: value, .. } => quote! { @@ -335,18 +300,41 @@ fn init_fields( if let Some(ident) = kind.ident() { // `mixed_site` ensures that the guard is not accessible to the user-controlled code. let guard = format_ident!("__{ident}_guard", span = Span::mixed_site()); + + // NOTE: The reference is derived from the guard so that it only lives as long as the + // guard does and cannot escape the scope. If it's created via `&mut (*#slot).#ident` + // like the unaligned field guard, it will become effectively `'static`. + let accessor = if pinned { + let project_ident = format_ident!("__project_{ident}"); + quote! { + // SAFETY: the initialization is pinned. + unsafe { #data.#project_ident(#guard.let_binding()) } + } + } else { + quote! { + #guard.let_binding() + } + }; + res.extend(quote! { #(#cfgs)* - // Create the drop guard: + // Create the drop guard. // - // We rely on macro hygiene to make it impossible for users to access this local - // variable. - // SAFETY: We forget the guard later when initialization has succeeded. - let #guard = unsafe { + // SAFETY: + // - `&raw mut (*slot).#ident` is valid. + // - `make_field_check` checks that `&raw mut (*slot).#ident` is properly aligned. + // - `(*slot).#ident` has been initialized above. + // - We only need the ownership to the pointee back when initialization has + // succeeded, where we `forget` the guard. + let mut #guard = unsafe { ::pin_init::__internal::DropGuard::new( &raw mut (*slot).#ident ) }; + + #(#cfgs)* + #[allow(unused_variables)] + let #ident = #accessor; }); guards.push(guard); guard_attrs.push(cfgs); diff --git a/rust/pin-init/src/__internal.rs b/rust/pin-init/src/__internal.rs index 90adbdc1893b..5720a621aed7 100644 --- a/rust/pin-init/src/__internal.rs +++ b/rust/pin-init/src/__internal.rs @@ -238,32 +238,42 @@ struct Foo { /// When a value of this type is dropped, it drops a `T`. /// /// Can be forgotten to prevent the drop. +/// +/// # Invariants +/// +/// - `ptr` is valid and properly aligned. +/// - `*ptr` is initialized and owned by this guard. pub struct DropGuard { ptr: *mut T, } impl DropGuard { - /// Creates a new [`DropGuard`]. It will [`ptr::drop_in_place`] `ptr` when it gets dropped. + /// Creates a drop guard and transfer the ownership of the pointer content. + /// + /// The ownership is only relinguished if the guard is forgotten via [`core::mem::forget`]. /// /// # Safety /// - /// `ptr` must be a valid pointer. - /// - /// It is the callers responsibility that `self` will only get dropped if the pointee of `ptr`: - /// - has not been dropped, - /// - is not accessible by any other means, - /// - will not be dropped by any other means. + /// - `ptr` is valid and properly aligned. + /// - `*ptr` is initialized, and the ownership is transferred to this guard. #[inline] pub unsafe fn new(ptr: *mut T) -> Self { + // INVARIANT: By safety requirement. Self { ptr } } + + /// Create a let binding for accessor use. + #[inline] + pub fn let_binding(&mut self) -> &mut T { + // SAFETY: Per type invariant. + unsafe { &mut *self.ptr } + } } impl Drop for DropGuard { #[inline] fn drop(&mut self) { - // SAFETY: A `DropGuard` can only be constructed using the unsafe `new` function - // ensuring that this operation is safe. + // SAFETY: `self.ptr` is valid, properly aligned and `*self.ptr` is owned by this guard. unsafe { ptr::drop_in_place(self.ptr) } } } From 838d852da8503372f3a1779bfbd1ccb93153ab4e Mon Sep 17 00:00:00 2001 From: Miguel Ojeda Date: Sun, 26 Apr 2026 16:42:00 +0200 Subject: [PATCH 042/333] rust: allow `clippy::collapsible_match` globally The `clippy::collapsible_match` lint [1] can make code harder to read in certain cases [2], e.g. CLIPPY P rust/libmacros.so - due to command line change warning: this `if` can be collapsed into the outer `match` --> rust/pin-init/internal/src/helpers.rs:91:17 | 91 | / if nesting == 1 { 92 | | impl_generics.push(tt.clone()); 93 | | impl_generics.push(tt); 94 | | skip_until_comma = false; 95 | | } | |_________________^ | = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#collapsible_match = note: `-W clippy::collapsible-match` implied by `-W clippy::all` = help: to override `-W clippy::all` add `#[allow(clippy::collapsible_match)]` help: collapse nested if block | 90 ~ TokenTree::Punct(p) if skip_until_comma && p.as_char() == ',' 91 ~ && nesting == 1 => { 92 | impl_generics.push(tt.clone()); 93 | impl_generics.push(tt); 94 | skip_until_comma = false; 95 ~ } | The lint does not have much upside -- when the suggestion may be a good one, it would still read fine when nested anyway. And it is the kind of lint that may easily bias people to just apply the suggestion instead of allowing it. [ In addition, as Gary points out [3], the suggestion is also wrong [4] and in the process of being fixed [5], possibly for Rust 1.97.0: Link: https://lore.kernel.org/rust-for-linux/DI3YV94TH9I3.1SOHW51552497@garyguo.net/ [3] Link: https://github.com/rust-lang/rust-clippy/issues/16875 [4] Link: https://github.com/rust-lang/rust-clippy/pull/16878 [5] - Miguel ] Thus just let developers decide on their own. Cc: stable@vger.kernel.org # Needed in 6.12.y and later (Rust is pinned in older LTSs). Link: https://rust-lang.github.io/rust-clippy/master/index.html#collapsible_match [1] Link: https://lore.kernel.org/rust-for-linux/CANiq72nWYJna_hdFxjQCQZK6yJBrr1Mb86iKavivV0U0BgufeA@mail.gmail.com/ [2] Reviewed-by: Gary Guo Link: https://patch.msgid.link/20260426144201.227108-1-ojeda@kernel.org Signed-off-by: Miguel Ojeda --- Makefile | 1 + 1 file changed, 1 insertion(+) diff --git a/Makefile b/Makefile index e27c91ea56fc..621d84aa4700 100644 --- a/Makefile +++ b/Makefile @@ -486,6 +486,7 @@ export rust_common_flags := --edition=2021 \ -Wclippy::as_ptr_cast_mut \ -Wclippy::as_underscore \ -Wclippy::cast_lossless \ + -Aclippy::collapsible_match \ -Wclippy::ignored_unit_patterns \ -Aclippy::incompatible_msrv \ -Wclippy::mut_mut \ From 2adc8664018c1cc595c7c0c98474a33c7fe32a85 Mon Sep 17 00:00:00 2001 From: Miguel Ojeda Date: Sun, 26 Apr 2026 16:42:01 +0200 Subject: [PATCH 043/333] rust: allow `clippy::collapsible_if` globally Similar to `clippy::collapsible_match` (globally allowed in the previous commit), the `clippy::collapsible_if` lint [1] can make code harder to read in certain cases. Thus just let developers decide on their own. In addition, remove the existing `expect` we had. Cc: stable@vger.kernel.org # Needed in 6.12.y and later (Rust is pinned in older LTSs). Suggested-by: Gary Guo Link: https://lore.kernel.org/rust-for-linux/DGROP5CHU1QZ.1OKJRAUZXE9WC@garyguo.net/ Link: https://rust-lang.github.io/rust-clippy/master/index.html#collapsible_if [1] Reviewed-by: Gary Guo Link: https://patch.msgid.link/20260426144201.227108-2-ojeda@kernel.org Signed-off-by: Miguel Ojeda --- Makefile | 1 + drivers/android/binder/range_alloc/array.rs | 1 - 2 files changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 621d84aa4700..28f4ae452441 100644 --- a/Makefile +++ b/Makefile @@ -486,6 +486,7 @@ export rust_common_flags := --edition=2021 \ -Wclippy::as_ptr_cast_mut \ -Wclippy::as_underscore \ -Wclippy::cast_lossless \ + -Aclippy::collapsible_if \ -Aclippy::collapsible_match \ -Wclippy::ignored_unit_patterns \ -Aclippy::incompatible_msrv \ diff --git a/drivers/android/binder/range_alloc/array.rs b/drivers/android/binder/range_alloc/array.rs index ada1d1b4302e..081d19b09d4b 100644 --- a/drivers/android/binder/range_alloc/array.rs +++ b/drivers/android/binder/range_alloc/array.rs @@ -204,7 +204,6 @@ pub(crate) fn reservation_abort(&mut self, offset: usize) -> Result // caller will mark them as unused, which means that they can be freed if the system comes // under memory pressure. let mut freed_range = FreedRange::interior_pages(offset, size); - #[expect(clippy::collapsible_if)] // reads better like this if offset % PAGE_SIZE != 0 { if i == 0 || self.ranges[i - 1].endpoint() <= (offset & PAGE_MASK) { freed_range.start_page_idx -= 1; From dff205550e714f1cb65f27409586e2612905fa88 Mon Sep 17 00:00:00 2001 From: Gui-Dong Han Date: Thu, 16 Apr 2026 21:57:03 +0800 Subject: [PATCH 044/333] hwmon: (lm63) Add locking to avoid TOCTOU The functions show_fan(), show_pwm1(), show_temp11(), temp2_crit_hyst_show(), and show_lut_temp_hyst() access shared cached data without holding the update lock. This can cause TOCTOU races if the cached values change between the checks and the later calculations. Those cached values are updated in lm63_update_device(). In the general case, the affected functions combine multiple cached values without locking and can therefore observe a mixed old/new snapshot. In addition, show_fan() reads data->fan[nr] locklessly while lm63_update_device() updates data->fan[0] in two steps, which can expose an intermediate torn value and potentially trigger a divide-by-zero error. This means that converting the macro to a function is not sufficient to fix show_fan(). Hold the update lock across the whole read and calculation sequence so that the values remain stable. Check the other functions in the driver as well. Keep them unchanged because they either do not access shared cached values multiple times or already do so under lock. Link: https://lore.kernel.org/linux-hwmon/CALbr=LYJ_ehtp53HXEVkSpYoub+XYSTU8Rg=o1xxMJ8=5z8B-g@mail.gmail.com/ Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Fixes: e872c91e726e ("hwmon: (lm63) Add support for unsigned upper temperature limits") Fixes: d216f6809eb6 ("hwmon: (lm63) Expose automatic fan speed control lookup table") Signed-off-by: Gui-Dong Han Link: https://lore.kernel.org/r/20260416135703.53262-1-hanguidong02@gmail.com [groeck: Use lm63_update_device() to get driver data in temp2_crit_hyst_store] Signed-off-by: Guenter Roeck --- drivers/hwmon/lm63.c | 39 ++++++++++++++++++++++++++++++--------- 1 file changed, 30 insertions(+), 9 deletions(-) diff --git a/drivers/hwmon/lm63.c b/drivers/hwmon/lm63.c index 035176a98ce9..30500b4d2221 100644 --- a/drivers/hwmon/lm63.c +++ b/drivers/hwmon/lm63.c @@ -333,7 +333,13 @@ static ssize_t show_fan(struct device *dev, struct device_attribute *devattr, { struct sensor_device_attribute *attr = to_sensor_dev_attr(devattr); struct lm63_data *data = lm63_update_device(dev); - return sprintf(buf, "%d\n", FAN_FROM_REG(data->fan[attr->index])); + int fan; + + mutex_lock(&data->update_lock); + fan = FAN_FROM_REG(data->fan[attr->index]); + mutex_unlock(&data->update_lock); + + return sprintf(buf, "%d\n", fan); } static ssize_t set_fan(struct device *dev, struct device_attribute *dummy, @@ -366,12 +372,14 @@ static ssize_t show_pwm1(struct device *dev, struct device_attribute *devattr, int nr = attr->index; int pwm; + mutex_lock(&data->update_lock); if (data->pwm_highres) pwm = data->pwm1[nr]; else pwm = data->pwm1[nr] >= 2 * data->pwm1_freq ? 255 : (data->pwm1[nr] * 255 + data->pwm1_freq) / (2 * data->pwm1_freq); + mutex_unlock(&data->update_lock); return sprintf(buf, "%d\n", pwm); } @@ -529,6 +537,7 @@ static ssize_t show_temp11(struct device *dev, struct device_attribute *devattr, int nr = attr->index; int temp; + mutex_lock(&data->update_lock); if (!nr) { /* * Use unsigned temperature unless its value is zero. @@ -544,7 +553,10 @@ static ssize_t show_temp11(struct device *dev, struct device_attribute *devattr, else temp = TEMP11_FROM_REG(data->temp11[nr]); } - return sprintf(buf, "%d\n", temp + data->temp2_offset); + temp += data->temp2_offset; + mutex_unlock(&data->update_lock); + + return sprintf(buf, "%d\n", temp); } static ssize_t set_temp11(struct device *dev, struct device_attribute *devattr, @@ -592,9 +604,14 @@ static ssize_t temp2_crit_hyst_show(struct device *dev, struct device_attribute *dummy, char *buf) { struct lm63_data *data = lm63_update_device(dev); - return sprintf(buf, "%d\n", temp8_from_reg(data, 2) - + data->temp2_offset - - TEMP8_FROM_REG(data->temp2_crit_hyst)); + int temp; + + mutex_lock(&data->update_lock); + temp = temp8_from_reg(data, 2) + data->temp2_offset + - TEMP8_FROM_REG(data->temp2_crit_hyst); + mutex_unlock(&data->update_lock); + + return sprintf(buf, "%d\n", temp); } static ssize_t show_lut_temp_hyst(struct device *dev, @@ -602,10 +619,14 @@ static ssize_t show_lut_temp_hyst(struct device *dev, { struct sensor_device_attribute *attr = to_sensor_dev_attr(devattr); struct lm63_data *data = lm63_update_device(dev); + int temp; - return sprintf(buf, "%d\n", lut_temp_from_reg(data, attr->index) - + data->temp2_offset - - TEMP8_FROM_REG(data->lut_temp_hyst)); + mutex_lock(&data->update_lock); + temp = lut_temp_from_reg(data, attr->index) + data->temp2_offset + - TEMP8_FROM_REG(data->lut_temp_hyst); + mutex_unlock(&data->update_lock); + + return sprintf(buf, "%d\n", temp); } /* @@ -616,7 +637,7 @@ static ssize_t temp2_crit_hyst_store(struct device *dev, struct device_attribute *dummy, const char *buf, size_t count) { - struct lm63_data *data = dev_get_drvdata(dev); + struct lm63_data *data = lm63_update_device(dev); struct i2c_client *client = data->client; long val; int err; From c9e3878ae2f57fd6786279cf5d9dc6e6e1b52f5a Mon Sep 17 00:00:00 2001 From: Timur Tabi Date: Thu, 30 Apr 2026 17:38:29 -0500 Subject: [PATCH 045/333] Revert "drm/nouveau/gsp: add support for GA100" This reverts commit 20e0c197802c545db220157fafd567a10f2b7672. Despite claiming to add GA100 support, that commit actually has quite a few problems. It falsely claims that there is no VBIOS. GA100 does have a VBIOS, but it has no display engine, so it cannot use the PRAMIN method the read VBIOS and must fall back to using PROM. For whatever reason, the VBIOS on GA100 has an "Init-from-ROM" (IFR) header where the PCI Expansion ROM would normally be found. So to find that ROM, Nouveau needs to parse the IFR header. The commit also falsely claimed that there is no graphics (GR) engine. So rather than try to fix that commit, just revert it and start over from scratch. Signed-off-by: Timur Tabi Link: https://patch.msgid.link/20260430223838.2530778-2-ttabi@nvidia.com Signed-off-by: Danilo Krummrich --- .../gpu/drm/nouveau/nvkm/engine/device/base.c | 11 +++++++++-- .../gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c | 4 ++++ .../gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c | 18 +++++------------- 3 files changed, 18 insertions(+), 15 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c index 72848ed80df7..b101e14f841e 100644 --- a/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/engine/device/base.c @@ -2513,6 +2513,7 @@ static const struct nvkm_device_chip nv170_chipset = { .name = "GA100", .bar = { 0x00000001, tu102_bar_new }, + .bios = { 0x00000001, nvkm_bios_new }, .devinit = { 0x00000001, ga100_devinit_new }, .fault = { 0x00000001, tu102_fault_new }, .fb = { 0x00000001, ga100_fb_new }, @@ -2529,7 +2530,6 @@ nv170_chipset = { .vfn = { 0x00000001, ga100_vfn_new }, .ce = { 0x000003ff, ga100_ce_new }, .fifo = { 0x00000001, ga100_fifo_new }, - .sec2 = { 0x00000001, tu102_sec2_new }, }; static const struct nvkm_device_chip @@ -3341,7 +3341,6 @@ nvkm_device_ctor(const struct nvkm_device_func *func, case 0x166: device->chip = &nv166_chipset; break; case 0x167: device->chip = &nv167_chipset; break; case 0x168: device->chip = &nv168_chipset; break; - case 0x170: device->chip = &nv170_chipset; break; case 0x172: device->chip = &nv172_chipset; break; case 0x173: device->chip = &nv173_chipset; break; case 0x174: device->chip = &nv174_chipset; break; @@ -3361,6 +3360,14 @@ nvkm_device_ctor(const struct nvkm_device_func *func, case 0x1b6: device->chip = &nv1b6_chipset; break; case 0x1b7: device->chip = &nv1b7_chipset; break; default: + if (nvkm_boolopt(device->cfgopt, "NvEnableUnsupportedChipsets", false)) { + switch (device->chipset) { + case 0x170: device->chip = &nv170_chipset; break; + default: + break; + } + } + if (!device->chip) { nvdev_error(device, "unknown chipset (%08x)\n", boot0); ret = -ENODEV; diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c index fdd820eeef81..27a13aeccd3c 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/ga100.c @@ -41,11 +41,15 @@ ga100_gsp_flcn = { static const struct nvkm_gsp_func ga100_gsp = { .flcn = &ga100_gsp_flcn, + .fwsec = &tu102_gsp_fwsec, .sig_section = ".fwsignature_ga100", .booter.ctor = tu102_gsp_booter_ctor, + .fwsec_sb.ctor = tu102_gsp_fwsec_sb_ctor, + .fwsec_sb.dtor = tu102_gsp_fwsec_sb_dtor, + .dtor = r535_gsp_dtor, .oneinit = tu102_gsp_oneinit, .init = tu102_gsp_init, diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c index dd82c76b8b9a..19cb269e7a26 100644 --- a/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/gsp/tu102.c @@ -318,13 +318,8 @@ tu102_gsp_oneinit(struct nvkm_gsp *gsp) if (ret) return ret; - /* - * Calculate FB layout. FRTS is a memory region created by the FWSEC-FRTS firmware. - * FWSEC comes from VBIOS. So on systems with no VBIOS (e.g. GA100), the FRTS does - * not exist. Therefore, use the existence of VBIOS to determine whether to reserve - * an FRTS region. - */ - gsp->fb.wpr2.frts.size = device->bios ? 0x100000 : 0; + /* Calculate FB layout. */ + gsp->fb.wpr2.frts.size = 0x100000; gsp->fb.wpr2.frts.addr = ALIGN_DOWN(gsp->fb.bios.addr, 0x20000) - gsp->fb.wpr2.frts.size; gsp->fb.wpr2.boot.size = gsp->boot.fw.size; @@ -348,12 +343,9 @@ tu102_gsp_oneinit(struct nvkm_gsp *gsp) if (ret) return ret; - /* Only boot FWSEC-FRTS if it actually exists */ - if (gsp->fb.wpr2.frts.size) { - ret = nvkm_gsp_fwsec_frts(gsp); - if (WARN_ON(ret)) - return ret; - } + ret = nvkm_gsp_fwsec_frts(gsp); + if (WARN_ON(ret)) + return ret; /* Reset GSP into RISC-V mode. */ ret = gsp->func->reset(gsp); From 845db023a8aeba8b14315a846dcfba31ee727fb1 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Fri, 1 May 2026 19:23:12 +0800 Subject: [PATCH 046/333] ublk: don't issue uring_cmd from fallback task work When ublk_ch_uring_cmd_cb() runs as fallback task work (e.g., because the submitting task is exiting), the command should not be issued as current is a kworker, not the daemon task. This can cause io->task to capture the wrong task in __ublk_fetch(), leading to a task mismatch warning in ublk_uring_cmd_cancel_fn(). Check tw.cancel and return -ECANCELED instead of issuing the command from fallback context. Fixes: 3421c7f68bba ("ublk: make sure io cmd handled in submitter task context") Signed-off-by: Ming Lei Link: https://patch.msgid.link/20260501112312.947327-1-tom.leiming@gmail.com Signed-off-by: Jens Axboe --- drivers/block/ublk_drv.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/block/ublk_drv.c b/drivers/block/ublk_drv.c index 8e5f3738c203..d10460d29e4a 100644 --- a/drivers/block/ublk_drv.c +++ b/drivers/block/ublk_drv.c @@ -3496,8 +3496,10 @@ static void ublk_ch_uring_cmd_cb(struct io_tw_req tw_req, io_tw_token_t tw) { unsigned int issue_flags = IO_URING_CMD_TASK_WORK_ISSUE_FLAGS; struct io_uring_cmd *cmd = io_uring_cmd_from_tw(tw_req); - int ret = ublk_ch_uring_cmd_local(cmd, issue_flags); + int ret = -ECANCELED; + if (!tw.cancel) + ret = ublk_ch_uring_cmd_local(cmd, issue_flags); if (ret != -EIOCBQUEUED) io_uring_cmd_done(cmd, ret, issue_flags); } From 2e42a17b8f6bc3c0cd69d7556b588011d3ec2394 Mon Sep 17 00:00:00 2001 From: Eliot Courtney Date: Thu, 23 Apr 2026 21:36:52 +0900 Subject: [PATCH 047/333] rust: drm: gem: clean up GEM state in init failure case MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Currently, if `drm_gem_object_init` fails, the object is freed without any cleanup. Perform the cleanup in that case. Cc: stable@vger.kernel.org Fixes: c284d3e42338 ("rust: drm: gem: Add GEM object abstraction") Signed-off-by: Eliot Courtney Reviewed-by: Alice Ryhl Reviewed-by: Onur Özkan Link: https://patch.msgid.link/20260423-fix-gem-1-v1-1-e12e35f7bba9@nvidia.com [ Move safety comment closer to unsafe block to avoid a clippy warning. - Danilo ] Signed-off-by: Danilo Krummrich --- rust/kernel/drm/gem/mod.rs | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/rust/kernel/drm/gem/mod.rs b/rust/kernel/drm/gem/mod.rs index 75acda7ba500..01b5bd47a333 100644 --- a/rust/kernel/drm/gem/mod.rs +++ b/rust/kernel/drm/gem/mod.rs @@ -277,8 +277,17 @@ pub fn new(dev: &drm::Device, size: usize, args: T::Args) -> Result` is always treated as pinned. let ptr = KBox::into_raw(unsafe { Pin::into_inner_unchecked(obj) }); From cb48828f06afa232cc330f0f4d6be101067810b3 Mon Sep 17 00:00:00 2001 From: Mark Brown Date: Thu, 23 Apr 2026 20:17:45 +0100 Subject: [PATCH 048/333] selftests/rseq: Don't run tests with runner scripts outside of the scripts The rseq selftests include two runner scripts run_param_test.sh and run_syscall_errors_test.sh which set up the environment for test binaries and run them with various parameters. Currently we list these test binaries in TEST_GEN_PROGS but this results in the kselftest framework running them directly as well as via the runners, resulting in duplication and spurious failures when the environment is not correctly set up (eg, if glibc tries to use rseq). Move the binaries the runners invoke to TEST_GEN_PROGS_EXTENDED, binaries listed there are built but not run by the framework. The param_test benchmarks are not moved since they are not run by run_param_test.sh. Fixes: 830969e7821a ("selftests/rseq: Implement time slice extension test") Signed-off-by: Mark Brown Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Link: https://patch.msgid.link/20260423-selftests-rseq-use-runner-v1-1-e13a133754c1@kernel.org Cc: stable@vger.kernel.org --- tools/testing/selftests/rseq/Makefile | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/tools/testing/selftests/rseq/Makefile b/tools/testing/selftests/rseq/Makefile index 4ef90823b652..0d1947c0d623 100644 --- a/tools/testing/selftests/rseq/Makefile +++ b/tools/testing/selftests/rseq/Makefile @@ -14,12 +14,15 @@ LDLIBS += -lpthread -ldl # still track changes to header files and depend on shared object. OVERRIDE_TARGETS = 1 -TEST_GEN_PROGS = basic_test basic_percpu_ops_test basic_percpu_ops_mm_cid_test param_test \ - param_test_benchmark param_test_compare_twice param_test_mm_cid \ - param_test_mm_cid_benchmark param_test_mm_cid_compare_twice \ - syscall_errors_test slice_test +TEST_GEN_PROGS = basic_test basic_percpu_ops_test basic_percpu_ops_mm_cid_test \ + param_test_benchmark param_test_mm_cid_benchmark slice_test -TEST_GEN_PROGS_EXTENDED = librseq.so +TEST_GEN_PROGS_EXTENDED = librseq.so \ + param_test \ + param_test_compare_twice \ + param_test_mm_cid \ + param_test_mm_cid_compare_twice \ + syscall_errors_test TEST_PROGS = run_param_test.sh run_syscall_errors_test.sh From 2cb68e45120dfc66404c7547d95b8ac6ff0b25ce Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Apr 2026 10:10:19 +0200 Subject: [PATCH 049/333] rseq: Set rseq::cpu_id_start to 0 on unregistration The RSEQ rework changed that to RSEQ_CPU_UNINITILIZED, which is obviously incompatible. Revert back to the original behavior. Fixes: 0f085b41880e ("rseq: Provide and use rseq_set_ids()") Reported-by: Dmitry Vyukov Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224427.271566313%40kernel.org Cc: stable@vger.kernel.org --- kernel/rseq.c | 20 +++++++++----------- 1 file changed, 9 insertions(+), 11 deletions(-) diff --git a/kernel/rseq.c b/kernel/rseq.c index 38d3ef540760..b9f11931ef78 100644 --- a/kernel/rseq.c +++ b/kernel/rseq.c @@ -236,11 +236,6 @@ static int __init rseq_debugfs_init(void) } __initcall(rseq_debugfs_init); -static bool rseq_set_ids(struct task_struct *t, struct rseq_ids *ids, u32 node_id) -{ - return rseq_set_ids_get_csaddr(t, ids, node_id, NULL); -} - static bool rseq_handle_cs(struct task_struct *t, struct pt_regs *regs) { struct rseq __user *urseq = t->rseq.usrptr; @@ -384,19 +379,22 @@ void rseq_syscall(struct pt_regs *regs) static bool rseq_reset_ids(void) { - struct rseq_ids ids = { - .cpu_id = RSEQ_CPU_ID_UNINITIALIZED, - .mm_cid = 0, - }; + struct rseq __user *rseq = current->rseq.usrptr; /* * If this fails, terminate it because this leaves the kernel in * stupid state as exit to user space will try to fixup the ids * again. */ - if (rseq_set_ids(current, &ids, 0)) - return true; + scoped_user_rw_access(rseq, efault) { + unsafe_put_user(0, &rseq->cpu_id_start, efault); + unsafe_put_user(RSEQ_CPU_ID_UNINITIALIZED, &rseq->cpu_id, efault); + unsafe_put_user(0, &rseq->node_id, efault); + unsafe_put_user(0, &rseq->mm_cid, efault); + } + return true; +efault: force_sig(SIGSEGV); return false; } From e9766e6f7d330dce7530918d8c6e3ec96d6c6e24 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Apr 2026 10:14:41 +0200 Subject: [PATCH 050/333] rseq: Protect rseq_reset() against interrupts rseq_reset() uses memset() to clear the tasks rseq data. That's racy against membarrier() and preemption. Guard it with irqsave to cure this. Fixes: faba9d250eae ("rseq: Introduce struct rseq_data") Reported-by: Dmitry Vyukov Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224427.353887714%40kernel.org Cc: stable@vger.kernel.org --- include/linux/rseq.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/include/linux/rseq.h b/include/linux/rseq.h index b9d62fc2140d..f446909551df 100644 --- a/include/linux/rseq.h +++ b/include/linux/rseq.h @@ -119,6 +119,8 @@ static inline void rseq_virt_userspace_exit(void) static inline void rseq_reset(struct task_struct *t) { + /* Protect against preemption and membarrier IPI */ + guard(irqsave)(); memset(&t->rseq, 0, sizeof(t->rseq)); t->rseq.ids.cpu_id = RSEQ_CPU_ID_UNINITIALIZED; } From 010b7723c0a3b9ad58f50b715dbe2e7781d29400 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Tue, 28 Apr 2026 09:34:45 +0200 Subject: [PATCH 051/333] rseq: Don't advertise time slice extensions if disabled If time slice extensions have been disabled on the kernel command line, then advertising them in RSEQ flags is wrong. Adjust the conditionals to reflect reality, fixup the misleading comments about the gap of these flags and the rseq::flags field. Fixes: d6200245c75e ("rseq: Allow registering RSEQ with slice extension") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224427.437059375%40kernel.org Cc: stable@vger.kernel.org --- include/uapi/linux/rseq.h | 5 ++++- kernel/rseq.c | 9 +++++---- 2 files changed, 9 insertions(+), 5 deletions(-) diff --git a/include/uapi/linux/rseq.h b/include/uapi/linux/rseq.h index f69344fe6c08..ca6fe1f9d05e 100644 --- a/include/uapi/linux/rseq.h +++ b/include/uapi/linux/rseq.h @@ -28,7 +28,7 @@ enum rseq_cs_flags_bit { RSEQ_CS_FLAG_NO_RESTART_ON_PREEMPT_BIT = 0, RSEQ_CS_FLAG_NO_RESTART_ON_SIGNAL_BIT = 1, RSEQ_CS_FLAG_NO_RESTART_ON_MIGRATE_BIT = 2, - /* (3) Intentional gap to put new bits into a separate byte */ + /* (3) Intentional gap to keep new bits separate */ /* User read only feature flags */ RSEQ_CS_FLAG_SLICE_EXT_AVAILABLE_BIT = 4, @@ -161,6 +161,9 @@ struct rseq { * - RSEQ_CS_FLAG_NO_RESTART_ON_PREEMPT * - RSEQ_CS_FLAG_NO_RESTART_ON_SIGNAL * - RSEQ_CS_FLAG_NO_RESTART_ON_MIGRATE + * + * It is now used for feature status advertisement by the kernel. + * See: enum rseq_cs_flags_bit for further information. */ __u32 flags; diff --git a/kernel/rseq.c b/kernel/rseq.c index b9f11931ef78..586f58f652c6 100644 --- a/kernel/rseq.c +++ b/kernel/rseq.c @@ -462,10 +462,11 @@ SYSCALL_DEFINE4(rseq, struct rseq __user *, rseq, u32, rseq_len, int, flags, u32 return -EFAULT; if (IS_ENABLED(CONFIG_RSEQ_SLICE_EXTENSION)) { - rseqfl |= RSEQ_CS_FLAG_SLICE_EXT_AVAILABLE; - if (rseq_slice_extension_enabled() && - (flags & RSEQ_FLAG_SLICE_EXT_DEFAULT_ON)) - rseqfl |= RSEQ_CS_FLAG_SLICE_EXT_ENABLED; + if (rseq_slice_extension_enabled()) { + rseqfl |= RSEQ_CS_FLAG_SLICE_EXT_AVAILABLE; + if (flags & RSEQ_FLAG_SLICE_EXT_DEFAULT_ON) + rseqfl |= RSEQ_CS_FLAG_SLICE_EXT_ENABLED; + } } scoped_user_write_access(rseq, efault) { From 01eb80b767430ae868c48ad106c60eb61a508c85 Mon Sep 17 00:00:00 2001 From: Alok Tiwari Date: Fri, 10 Apr 2026 04:20:12 -0700 Subject: [PATCH 052/333] accel/qaic: fix incorrect counter check in RAS message decode The UE and UE_NF cases check ce_count against UINT_MAX before incrementing their respective counters. This is logically incorrect and prevents ue_count and ue_nf_count from incrementing when ce_count reaches UINT_MAX. Fixes: c11a50b170e7 ("accel/qaic: Add Reliability, Accessibility, Serviceability (RAS)") Signed-off-by: Alok Tiwari Reviewed-by: Jeff Hugo Signed-off-by: Jeff Hugo Link: https://patch.msgid.link/20260410112015.592546-1-alok.a.tiwari@oracle.com --- drivers/accel/qaic/qaic_ras.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/accel/qaic/qaic_ras.c b/drivers/accel/qaic/qaic_ras.c index cc0b75461e1a..6791af366cba 100644 --- a/drivers/accel/qaic/qaic_ras.c +++ b/drivers/accel/qaic/qaic_ras.c @@ -497,11 +497,11 @@ static void decode_ras_msg(struct qaic_device *qdev, struct ras_data *msg) qdev->ce_count++; break; case UE: - if (qdev->ce_count != UINT_MAX) + if (qdev->ue_count != UINT_MAX) qdev->ue_count++; break; case UE_NF: - if (qdev->ce_count != UINT_MAX) + if (qdev->ue_nf_count != UINT_MAX) qdev->ue_nf_count++; break; default: From 0c7ae130698e70107430254e79fbe996b4d37ab5 Mon Sep 17 00:00:00 2001 From: Paul Chaignon Date: Sat, 2 May 2026 12:12:40 +0200 Subject: [PATCH 053/333] tools/headers: Regenerate stddef.h to fix BPF selftests MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit With commit dacbfc167808 ("crypto: af_alg - Annotate struct af_alg_iv with __counted_by"), two selftests, test_tag and crypto_sanity, now indirectly rely on the __counted_by macro. On systems with commit dacbfc167808 in the installed UAPI headers, the selftests build fails with: In file included from tools/testing/selftests/bpf/prog_tests/crypto_sanity.c:7: /usr/include/linux/if_alg.h:45:22: error: expected ‘:’, ‘,’, ‘;’, ‘}’ or ‘__attribute__’ before ‘__counted_by’ 45 | __u8 iv[] __counted_by(ivlen); | ^~~~~~~~~~~~ This patch fixes it by regenerating stddef.h in tools/include using the instructions from commit a778f5d46b62 ("tools/headers: Pull in stddef.h to uapi to fix BPF selftests build in CI"). Fixes: dacbfc167808 ("crypto: af_alg - Annotate struct af_alg_iv with __counted_by") Signed-off-by: Paul Chaignon Reviewed-by: Alan Maguire Tested-by: Ihor Solodrai Link: https://lore.kernel.org/r/8da8ef16055aa452d940668ed5359ce54adc6b0b.1777715500.git.paul.chaignon@gmail.com Signed-off-by: Alexei Starovoitov --- tools/include/uapi/linux/stddef.h | 26 +++++++++++++++++++++++++- 1 file changed, 25 insertions(+), 1 deletion(-) diff --git a/tools/include/uapi/linux/stddef.h b/tools/include/uapi/linux/stddef.h index c53cde425406..457498259494 100644 --- a/tools/include/uapi/linux/stddef.h +++ b/tools/include/uapi/linux/stddef.h @@ -3,7 +3,6 @@ #define _LINUX_STDDEF_H - #ifndef __always_inline #define __always_inline __inline__ #endif @@ -36,6 +35,11 @@ struct __struct_group_tag(TAG) { MEMBERS } ATTRS NAME; \ } ATTRS +#ifdef __cplusplus +/* sizeof(struct{}) is 1 in C++, not 0, can't use C version of the macro. */ +#define __DECLARE_FLEX_ARRAY(T, member) \ + T member[0] +#else /** * __DECLARE_FLEX_ARRAY() - Declare a flexible array usable in a union * @@ -52,3 +56,23 @@ TYPE NAME[]; \ } #endif + +#ifndef __counted_by +#define __counted_by(m) +#endif + +#ifndef __counted_by_le +#define __counted_by_le(m) +#endif + +#ifndef __counted_by_be +#define __counted_by_be(m) +#endif + +#ifndef __counted_by_ptr +#define __counted_by_ptr(m) +#endif + +#define __kernel_nonstring + +#endif /* _LINUX_STDDEF_H */ From 0799e5943611006b346b8813c7daf7dd5aa26bfd Mon Sep 17 00:00:00 2001 From: Lyes Bourennani Date: Wed, 22 Apr 2026 00:20:22 +0200 Subject: [PATCH 054/333] batman-adv: fix integer overflow on buff_pos Fixing an integer overflow present in batadv_iv_ogm_send_to_if. The size check is done using the int type in batadv_iv_ogm_aggr_packet whereas the buff_pos variable uses the s16 type. This could lead to an out-of-bound read. Cc: stable@vger.kernel.org Fixes: c6c8fea29769 ("net: Add batman-adv meshing protocol") Signed-off-by: Lyes Bourennani Signed-off-by: Alexis Pinson Signed-off-by: Sven Eckelmann --- net/batman-adv/bat_iv_ogm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/batman-adv/bat_iv_ogm.c b/net/batman-adv/bat_iv_ogm.c index f28e9cbf8ad5..618d1889c04e 100644 --- a/net/batman-adv/bat_iv_ogm.c +++ b/net/batman-adv/bat_iv_ogm.c @@ -335,7 +335,7 @@ static void batadv_iv_ogm_send_to_if(struct batadv_forw_packet *forw_packet, struct batadv_priv *bat_priv = netdev_priv(hard_iface->mesh_iface); const char *fwd_str; u8 packet_num; - s16 buff_pos; + int buff_pos; struct batadv_ogm_packet *batadv_ogm_packet; struct sk_buff *skb; u8 *packet_pos; From 3243543592425beec83d453793e9d27caa0d8e66 Mon Sep 17 00:00:00 2001 From: Jiexun Wang Date: Mon, 27 Apr 2026 14:43:33 +0800 Subject: [PATCH 055/333] batman-adv: reject new tp_meter sessions during teardown Prevent tp_meter from starting new sender or receiver sessions after mesh_state has left BATADV_MESH_ACTIVE. Fixes: 33a3bb4a3345 ("batman-adv: throughput meter implementation") Cc: stable@kernel.org Reported-by: Yuan Tan Reported-by: Yifan Wu Reported-by: Juefei Pu Reported-by: Xin Liu Co-developed-by: Luxing Yin Signed-off-by: Luxing Yin Signed-off-by: Jiexun Wang Signed-off-by: Ren Wei Signed-off-by: Sven Eckelmann --- net/batman-adv/tp_meter.c | 17 ++++++++++++++++- 1 file changed, 16 insertions(+), 1 deletion(-) diff --git a/net/batman-adv/tp_meter.c b/net/batman-adv/tp_meter.c index 2e42f6b348c8..d9a80e459c2e 100644 --- a/net/batman-adv/tp_meter.c +++ b/net/batman-adv/tp_meter.c @@ -947,6 +947,13 @@ void batadv_tp_start(struct batadv_priv *bat_priv, const u8 *dst, /* look for an already existing test towards this node */ spin_lock_bh(&bat_priv->tp_list_lock); + if (atomic_read(&bat_priv->mesh_state) != BATADV_MESH_ACTIVE) { + spin_unlock_bh(&bat_priv->tp_list_lock); + batadv_tp_batctl_error_notify(BATADV_TP_REASON_DST_UNREACHABLE, + dst, bat_priv, session_cookie); + return; + } + tp_vars = batadv_tp_list_find(bat_priv, dst); if (tp_vars) { spin_unlock_bh(&bat_priv->tp_list_lock); @@ -1329,9 +1336,12 @@ static struct batadv_tp_vars * batadv_tp_init_recv(struct batadv_priv *bat_priv, const struct batadv_icmp_tp_packet *icmp) { - struct batadv_tp_vars *tp_vars; + struct batadv_tp_vars *tp_vars = NULL; spin_lock_bh(&bat_priv->tp_list_lock); + if (atomic_read(&bat_priv->mesh_state) != BATADV_MESH_ACTIVE) + goto out_unlock; + tp_vars = batadv_tp_list_find_session(bat_priv, icmp->orig, icmp->session); if (tp_vars) @@ -1464,6 +1474,9 @@ void batadv_tp_meter_recv(struct batadv_priv *bat_priv, struct sk_buff *skb) { struct batadv_icmp_tp_packet *icmp; + if (atomic_read(&bat_priv->mesh_state) != BATADV_MESH_ACTIVE) + goto out; + icmp = (struct batadv_icmp_tp_packet *)skb->data; switch (icmp->subtype) { @@ -1478,6 +1491,8 @@ void batadv_tp_meter_recv(struct batadv_priv *bat_priv, struct sk_buff *skb) "Received unknown TP Metric packet type %u\n", icmp->subtype); } + +out: consume_skb(skb); } From 3d3cf6a7314aca4df0a6dde28ce784a2a30d0166 Mon Sep 17 00:00:00 2001 From: Jiexun Wang Date: Mon, 27 Apr 2026 14:43:34 +0800 Subject: [PATCH 056/333] batman-adv: stop tp_meter sessions during mesh teardown TP meter sessions remain linked on bat_priv->tp_list after the netlink request has already finished. When the mesh interface is removed, batadv_mesh_free() currently tears down the mesh without first draining these sessions. A running sender thread or a late incoming tp_meter packet can then keep processing against a mesh instance which is already shutting down. Synchronize tp_meter with the mesh lifetime by stopping all active sessions from batadv_mesh_free() and waiting for sender threads to exit before teardown continues. Fixes: 33a3bb4a3345 ("batman-adv: throughput meter implementation") Cc: stable@kernel.org Reported-by: Yuan Tan Reported-by: Yifan Wu Reported-by: Juefei Pu Reported-by: Xin Liu Co-developed-by: Luxing Yin Signed-off-by: Luxing Yin Signed-off-by: Jiexun Wang Signed-off-by: Ren Wei Signed-off-by: Sven Eckelmann --- net/batman-adv/main.c | 1 + net/batman-adv/tp_meter.c | 94 +++++++++++++++++++++++++++++++-------- net/batman-adv/tp_meter.h | 1 + net/batman-adv/types.h | 4 ++ 4 files changed, 82 insertions(+), 18 deletions(-) diff --git a/net/batman-adv/main.c b/net/batman-adv/main.c index 3a35aadd8b41..a4d33ee0fda5 100644 --- a/net/batman-adv/main.c +++ b/net/batman-adv/main.c @@ -249,6 +249,7 @@ void batadv_mesh_free(struct net_device *mesh_iface) atomic_set(&bat_priv->mesh_state, BATADV_MESH_DEACTIVATING); batadv_purge_outstanding_packets(bat_priv, NULL); + batadv_tp_stop_all(bat_priv); batadv_gw_node_free(bat_priv); diff --git a/net/batman-adv/tp_meter.c b/net/batman-adv/tp_meter.c index d9a80e459c2e..58ca59a2799e 100644 --- a/net/batman-adv/tp_meter.c +++ b/net/batman-adv/tp_meter.c @@ -12,6 +12,7 @@ #include #include #include +#include #include #include #include @@ -365,23 +366,38 @@ static void batadv_tp_vars_put(struct batadv_tp_vars *tp_vars) } /** - * batadv_tp_sender_cleanup() - cleanup sender data and drop and timer - * @bat_priv: the bat priv with all the mesh interface information - * @tp_vars: the private data of the current TP meter session to cleanup + * batadv_tp_list_detach() - remove tp session from mesh session list once + * @tp_vars: the private data of the current TP meter session */ -static void batadv_tp_sender_cleanup(struct batadv_priv *bat_priv, - struct batadv_tp_vars *tp_vars) +static void batadv_tp_list_detach(struct batadv_tp_vars *tp_vars) { - cancel_delayed_work(&tp_vars->finish_work); + bool detached = false; spin_lock_bh(&tp_vars->bat_priv->tp_list_lock); - hlist_del_rcu(&tp_vars->list); + if (!hlist_unhashed(&tp_vars->list)) { + hlist_del_init_rcu(&tp_vars->list); + detached = true; + } spin_unlock_bh(&tp_vars->bat_priv->tp_list_lock); + if (!detached) + return; + + atomic_dec(&tp_vars->bat_priv->tp_num); + /* drop list reference */ batadv_tp_vars_put(tp_vars); +} - atomic_dec(&tp_vars->bat_priv->tp_num); +/** + * batadv_tp_sender_cleanup() - cleanup sender data and drop and timer + * @tp_vars: the private data of the current TP meter session to cleanup + */ +static void batadv_tp_sender_cleanup(struct batadv_tp_vars *tp_vars) +{ + cancel_delayed_work_sync(&tp_vars->finish_work); + + batadv_tp_list_detach(tp_vars); /* kill the timer and remove its reference */ timer_delete_sync(&tp_vars->timer); @@ -886,7 +902,8 @@ static int batadv_tp_send(void *arg) batadv_orig_node_put(orig_node); batadv_tp_sender_end(bat_priv, tp_vars); - batadv_tp_sender_cleanup(bat_priv, tp_vars); + batadv_tp_sender_cleanup(tp_vars); + complete(&tp_vars->finished); batadv_tp_vars_put(tp_vars); @@ -918,7 +935,8 @@ static void batadv_tp_start_kthread(struct batadv_tp_vars *tp_vars) batadv_tp_vars_put(tp_vars); /* cleanup of failed tp meter variables */ - batadv_tp_sender_cleanup(bat_priv, tp_vars); + batadv_tp_sender_cleanup(tp_vars); + complete(&tp_vars->finished); return; } @@ -1024,6 +1042,7 @@ void batadv_tp_start(struct batadv_priv *bat_priv, const u8 *dst, tp_vars->start_time = jiffies; init_waitqueue_head(&tp_vars->more_bytes); + init_completion(&tp_vars->finished); spin_lock_init(&tp_vars->unacked_lock); INIT_LIST_HEAD(&tp_vars->unacked_list); @@ -1126,14 +1145,7 @@ static void batadv_tp_receiver_shutdown(struct timer_list *t) "Shutting down for inactivity (more than %dms) from %pM\n", BATADV_TP_RECV_TIMEOUT, tp_vars->other_end); - spin_lock_bh(&tp_vars->bat_priv->tp_list_lock); - hlist_del_rcu(&tp_vars->list); - spin_unlock_bh(&tp_vars->bat_priv->tp_list_lock); - - /* drop list reference */ - batadv_tp_vars_put(tp_vars); - - atomic_dec(&bat_priv->tp_num); + batadv_tp_list_detach(tp_vars); spin_lock_bh(&tp_vars->unacked_lock); list_for_each_entry_safe(un, safe, &tp_vars->unacked_list, list) { @@ -1496,6 +1508,52 @@ void batadv_tp_meter_recv(struct batadv_priv *bat_priv, struct sk_buff *skb) consume_skb(skb); } +/** + * batadv_tp_stop_all() - stop all currently running tp meter sessions + * @bat_priv: the bat priv with all the mesh interface information + */ +void batadv_tp_stop_all(struct batadv_priv *bat_priv) +{ + struct batadv_tp_vars *tp_vars[BATADV_TP_MAX_NUM]; + struct batadv_tp_vars *tp_var; + size_t count = 0; + size_t i; + + spin_lock_bh(&bat_priv->tp_list_lock); + hlist_for_each_entry(tp_var, &bat_priv->tp_list, list) { + if (WARN_ON_ONCE(count >= BATADV_TP_MAX_NUM)) + break; + + if (!kref_get_unless_zero(&tp_var->refcount)) + continue; + + tp_vars[count++] = tp_var; + } + spin_unlock_bh(&bat_priv->tp_list_lock); + + for (i = 0; i < count; i++) { + tp_var = tp_vars[i]; + + switch (tp_var->role) { + case BATADV_TP_SENDER: + batadv_tp_sender_shutdown(tp_var, + BATADV_TP_REASON_CANCEL); + wake_up(&tp_var->more_bytes); + wait_for_completion(&tp_var->finished); + break; + case BATADV_TP_RECEIVER: + batadv_tp_list_detach(tp_var); + if (timer_shutdown_sync(&tp_var->timer)) + batadv_tp_vars_put(tp_var); + break; + } + + batadv_tp_vars_put(tp_var); + } + + synchronize_net(); +} + /** * batadv_tp_meter_init() - initialize global tp_meter structures */ diff --git a/net/batman-adv/tp_meter.h b/net/batman-adv/tp_meter.h index f0046d366eac..4e97cd10cd02 100644 --- a/net/batman-adv/tp_meter.h +++ b/net/batman-adv/tp_meter.h @@ -17,6 +17,7 @@ void batadv_tp_start(struct batadv_priv *bat_priv, const u8 *dst, u32 test_length, u32 *cookie); void batadv_tp_stop(struct batadv_priv *bat_priv, const u8 *dst, u8 return_value); +void batadv_tp_stop_all(struct batadv_priv *bat_priv); void batadv_tp_meter_recv(struct batadv_priv *bat_priv, struct sk_buff *skb); #endif /* _NET_BATMAN_ADV_TP_METER_H_ */ diff --git a/net/batman-adv/types.h b/net/batman-adv/types.h index 8fc5fe0e9b05..daa06f421154 100644 --- a/net/batman-adv/types.h +++ b/net/batman-adv/types.h @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include @@ -1328,6 +1329,9 @@ struct batadv_tp_vars { /** @finish_work: work item for the finishing procedure */ struct delayed_work finish_work; + /** @finished: completion signaled when a sender thread exits */ + struct completion finished; + /** @test_length: test length in milliseconds */ u32 test_length; From 0a69ac25bd596d50823d530d0a2004336668c0df Mon Sep 17 00:00:00 2001 From: Eliot Courtney Date: Fri, 1 May 2026 19:49:37 +0900 Subject: [PATCH 057/333] rust: drm: fix unsound initialization in drm::Device::new If pinned initialization of drm::Device::Data fails, it calls drm::Device::release via drm_dev_put. This materializes a reference to &drm::Device, but it's not fully constructed yet, because initializing `data` failed. It should not be dropped either. Instead, if pinned initialization fails, make sure drm::Device::release isn't called. Fixes: 2e9fdbe5ec7a ("rust: drm: device: drop_in_place() the drm::Device in release()") Signed-off-by: Eliot Courtney Reviewed-by: Gary Guo Link: https://patch.msgid.link/20260501-fix-drm-1-v2-1-5c4f681837bc@nvidia.com Signed-off-by: Danilo Krummrich --- rust/kernel/drm/device.rs | 22 ++++++++++++++++------ 1 file changed, 16 insertions(+), 6 deletions(-) diff --git a/rust/kernel/drm/device.rs b/rust/kernel/drm/device.rs index adbafe8db54d..403fc35353c7 100644 --- a/rust/kernel/drm/device.rs +++ b/rust/kernel/drm/device.rs @@ -119,13 +119,20 @@ pub fn new(dev: &device::Device, data: impl PinInit) -> Result()); + // Use a temporary vtable without a `release` callback until `data` is initialized, so + // init failure can release the DRM device without dropping uninitialized fields. + let alloc_vtable = bindings::drm_driver { + release: None, + ..Self::VTABLE + }; + // SAFETY: - // - `VTABLE`, as a `const` is pinned to the read-only section of the compilation, + // - `alloc_vtable` reference remains valid until no longer used, // - `dev` is valid by its type invarants, let raw_drm: *mut Self = unsafe { bindings::__drm_dev_alloc( dev.as_raw(), - &Self::VTABLE, + &alloc_vtable, layout.size(), mem::offset_of!(Self, dev), ) @@ -133,6 +140,10 @@ pub fn new(dev: &device::Device, data: impl PinInit) -> Result) -> Result Date: Tue, 28 Apr 2026 21:37:47 +0530 Subject: [PATCH 058/333] cifs: abort open_cached_dir if we don't request leases It is possible that SMB2_open_init may not set lease context based on the requested oplock level. This can happen when leases have been temporarily or permanently disabled. When this happens, we will have open_cached_dir making an open without lease context and the response will anyway be rejected by open_cached_dir (thereby forcing a close to discard this open). That's unnecessary two round-trips to the server. This change adds a check before making the open request to the server to make sure that SMB2_open_init did add the expected lease context to the open in open_cached_dir. Cc: Reviewed-by: Bharath SM Signed-off-by: Shyam Prasad N Signed-off-by: Steve French --- fs/smb/client/cached_dir.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/fs/smb/client/cached_dir.c b/fs/smb/client/cached_dir.c index 02791ec3c5a1..88d5e9a32f28 100644 --- a/fs/smb/client/cached_dir.c +++ b/fs/smb/client/cached_dir.c @@ -286,6 +286,14 @@ int open_cached_dir(unsigned int xid, struct cifs_tcon *tcon, &rqst[0], &oplock, &oparms, utf16_path); if (rc) goto oshr_free; + + if (oplock != SMB2_OPLOCK_LEVEL_II) { + rc = -EINVAL; + cifs_dbg(FYI, "%s: Oplock level %d not suitable for cached directory\n", + __func__, oplock); + goto oshr_free; + } + smb2_set_next_command(tcon, &rqst[0]); memset(&qi_iov, 0, sizeof(qi_iov)); From 5e489c6c47a2ac15edbaca153b9348e42c1eacab Mon Sep 17 00:00:00 2001 From: Bjoern Doebel Date: Thu, 30 Apr 2026 08:57:17 +0000 Subject: [PATCH 059/333] smb: client: use kzalloc to zero-initialize security descriptor buffer Commit 62e7dd0a39c2d ("smb: common: change the data type of num_aces to le16") split struct smb_acl's __le32 num_aces field into __le16 num_aces and __le16 reserved. The reserved field corresponds to Sbz2 in the MS-DTYP ACL wire format, which must be zero [1]. When building an ACL descriptor in build_sec_desc(), we are using a kmalloc()'ed descriptor buffer and writing the fields explicitly using le16() writes now. This never writes to the 2 byte reserved field, leaving it as uninitialized heap data. When the reserved field happens to contain non-zero slab garbage, Samba rejects the security descriptor with "ndr_pull_security_descriptor failed: Range Error", causing chmod to fail with EINVAL. Change kmalloc() to kzalloc() to ensure the entire buffer is zero-initialized. Fixes: 62e7dd0a39c2d ("smb: common: change the data type of num_aces to le16") Cc: stable@vger.kernel.org Signed-off-by: Bjoern Doebel Assisted-by: Kiro:claude-opus-4.6 [1] https://learn.microsoft.com/en-us/openspecs/windows_protocols/ms-dtyp/20233ed8-a6c6-4097-aafa-dd545ed24428 Signed-off-by: Steve French --- fs/smb/client/cifsacl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/smb/client/cifsacl.c b/fs/smb/client/cifsacl.c index ec5d47779304..a2750f1e3d90 100644 --- a/fs/smb/client/cifsacl.c +++ b/fs/smb/client/cifsacl.c @@ -1732,7 +1732,7 @@ id_mode_to_cifs_acl(struct inode *inode, const char *path, __u64 *pnmode, * descriptor parameters, and security descriptor itself */ nsecdesclen = max_t(u32, nsecdesclen, DEFAULT_SEC_DESC_LEN); - pnntsd = kmalloc(nsecdesclen, GFP_KERNEL); + pnntsd = kzalloc(nsecdesclen, GFP_KERNEL); if (!pnntsd) { kfree(pntsd); cifs_put_tlink(tlink); From 84d5d76c4e8e2750fa17869b7272f189d2bdd40b Mon Sep 17 00:00:00 2001 From: Matthew Brost Date: Fri, 1 May 2026 23:53:38 -0700 Subject: [PATCH 060/333] drm/ttm: Fix GPU MM stats during pool shrinking TTM pool shrinking frees pages by calling __free_pages() directly, which bypasses updates to NR_GPU_ACTIVE and leaves GPU MM accounting out of sync. Introduce a helper, __free_pages_gpu_account(), and use it for all page frees in ttm_pool.c so GPU MM statistics are updated consistently. Reported-by: Kenneth Crudup Fixes: ae80122f3896 ("drm/ttm: use gpu mm stats to track gpu memory allocations. (v4)") Cc: Christian Koenig Cc: Huang Rui Cc: Matthew Auld Cc: David Airlie Cc: dri-devel@lists.freedesktop.org Signed-off-by: Matthew Brost Tested-by: Kenneth Crudup Reviewed-by: Dave Airlie Link: https://patch.msgid.link/20260502065338.2720646-1-matthew.brost@intel.com --- drivers/gpu/drm/ttm/ttm_pool.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/ttm/ttm_pool.c b/drivers/gpu/drm/ttm/ttm_pool.c index 26a3689e5fd9..278bbe7a11ad 100644 --- a/drivers/gpu/drm/ttm/ttm_pool.c +++ b/drivers/gpu/drm/ttm/ttm_pool.c @@ -206,6 +206,14 @@ static struct page *ttm_pool_alloc_page(struct ttm_pool *pool, gfp_t gfp_flags, return NULL; } +static void __free_pages_gpu_account(struct page *p, unsigned int order, + bool reclaim) +{ + mod_lruvec_page_state(p, reclaim ? NR_GPU_RECLAIM : NR_GPU_ACTIVE, + -(1 << order)); + __free_pages(p, order); +} + /* Reset the caching and pages of size 1 << order */ static void ttm_pool_free_page(struct ttm_pool *pool, enum ttm_caching caching, unsigned int order, struct page *p, bool reclaim) @@ -223,9 +231,7 @@ static void ttm_pool_free_page(struct ttm_pool *pool, enum ttm_caching caching, #endif if (!pool || !ttm_pool_uses_dma_alloc(pool)) { - mod_lruvec_page_state(p, reclaim ? NR_GPU_RECLAIM : NR_GPU_ACTIVE, - -(1 << order)); - __free_pages(p, order); + __free_pages_gpu_account(p, order, reclaim); return; } @@ -606,7 +612,7 @@ static int ttm_pool_restore_commit(struct ttm_pool_tt_restore *restore, */ ttm_pool_split_for_swap(restore->pool, p); copy_highpage(restore->alloced_page + i, p); - __free_pages(p, 0); + __free_pages_gpu_account(p, 0, false); } restore->restored_pages++; @@ -1068,7 +1074,7 @@ long ttm_pool_backup(struct ttm_pool *pool, struct ttm_tt *tt, if (flags->purge) { shrunken += num_pages; page->private = 0; - __free_pages(page, order); + __free_pages_gpu_account(page, order, false); memset(tt->pages + i, 0, num_pages * sizeof(*tt->pages)); } @@ -1109,7 +1115,7 @@ long ttm_pool_backup(struct ttm_pool *pool, struct ttm_tt *tt, } handle = shandle; tt->pages[i] = ttm_backup_handle_to_page_ptr(handle); - put_page(page); + __free_pages_gpu_account(page, 0, false); shrunken++; } From b8c2e9e27636b92dc96c12f16894cbc60c58a306 Mon Sep 17 00:00:00 2001 From: Yufan Chen Date: Mon, 4 May 2026 01:56:10 +0800 Subject: [PATCH 061/333] io_uring/napi: clear tracked NAPI entries on unregister IORING_UNREGISTER_NAPI disables NAPI busy polling, but it currently leaves any previously tracked NAPI IDs on the ring context. The normal wait path only checks whether the list is empty before entering the busy poll helper, so an unregistered ring can still observe stale entries and run an unexpected busy poll pass. Make unregister switch the context to inactive and free the tracked entries. Do the same inactive transition while changing the tracking strategy, and recheck the expected tracking mode under napi_lock before inserting a newly learned NAPI ID. This prevents a racing poll path from repopulating the list after unregister or reconfiguration. Also make the busy poll dispatcher ignore inactive mode explicitly. Signed-off-by: Yufan Chen Fixes: 6bf90bd8c58a ("io_uring/napi: add static napi tracking strategy") Link: https://patch.msgid.link/20260503175610.35521-1-yufan.chen@linux.dev Signed-off-by: Jens Axboe --- io_uring/napi.c | 27 ++++++++++++++++++++------- io_uring/napi.h | 8 +++++--- 2 files changed, 25 insertions(+), 10 deletions(-) diff --git a/io_uring/napi.c b/io_uring/napi.c index 8d68366a4b90..bfc771445912 100644 --- a/io_uring/napi.c +++ b/io_uring/napi.c @@ -38,7 +38,8 @@ static inline ktime_t net_to_ktime(unsigned long t) return ns_to_ktime(t << 10); } -int __io_napi_add_id(struct io_ring_ctx *ctx, unsigned int napi_id) +int __io_napi_add_id(struct io_ring_ctx *ctx, unsigned int napi_id, + unsigned int mode) { struct hlist_head *hash_list; struct io_napi_entry *e; @@ -69,6 +70,11 @@ int __io_napi_add_id(struct io_ring_ctx *ctx, unsigned int napi_id) * kfree() */ spin_lock(&ctx->napi_lock); + if (unlikely(READ_ONCE(ctx->napi_track_mode) != mode)) { + spin_unlock(&ctx->napi_lock); + kfree(e); + return -EINVAL; + } if (unlikely(io_napi_hash_find(hash_list, napi_id))) { spin_unlock(&ctx->napi_lock); kfree(e); @@ -196,9 +202,14 @@ __io_napi_do_busy_loop(struct io_ring_ctx *ctx, bool (*loop_end)(void *, unsigned long), void *loop_end_arg) { - if (READ_ONCE(ctx->napi_track_mode) == IO_URING_NAPI_TRACKING_STATIC) + switch (READ_ONCE(ctx->napi_track_mode)) { + case IO_URING_NAPI_TRACKING_STATIC: return static_tracking_do_busy_loop(ctx, loop_end, loop_end_arg); - return dynamic_tracking_do_busy_loop(ctx, loop_end, loop_end_arg); + case IO_URING_NAPI_TRACKING_DYNAMIC: + return dynamic_tracking_do_busy_loop(ctx, loop_end, loop_end_arg); + default: + return false; + } } static void io_napi_blocking_busy_loop(struct io_ring_ctx *ctx, @@ -273,13 +284,13 @@ static int io_napi_register_napi(struct io_ring_ctx *ctx, default: return -EINVAL; } - /* clean the napi list for new settings */ + WRITE_ONCE(ctx->napi_track_mode, IO_URING_NAPI_TRACKING_INACTIVE); io_napi_free(ctx); - WRITE_ONCE(ctx->napi_track_mode, napi->op_param); /* cap NAPI at 10 msec of spin time */ napi->busy_poll_to = min(10000, napi->busy_poll_to); WRITE_ONCE(ctx->napi_busy_poll_dt, napi->busy_poll_to * NSEC_PER_USEC); WRITE_ONCE(ctx->napi_prefer_busy_poll, !!napi->prefer_busy_poll); + WRITE_ONCE(ctx->napi_track_mode, napi->op_param); return 0; } @@ -315,7 +326,8 @@ int io_register_napi(struct io_ring_ctx *ctx, void __user *arg) case IO_URING_NAPI_STATIC_ADD_ID: if (curr.op_param != IO_URING_NAPI_TRACKING_STATIC) return -EINVAL; - return __io_napi_add_id(ctx, napi.op_param); + return __io_napi_add_id(ctx, napi.op_param, + IO_URING_NAPI_TRACKING_STATIC); case IO_URING_NAPI_STATIC_DEL_ID: if (curr.op_param != IO_URING_NAPI_TRACKING_STATIC) return -EINVAL; @@ -343,9 +355,10 @@ int io_unregister_napi(struct io_ring_ctx *ctx, void __user *arg) if (arg && copy_to_user(arg, &curr, sizeof(curr))) return -EFAULT; + WRITE_ONCE(ctx->napi_track_mode, IO_URING_NAPI_TRACKING_INACTIVE); WRITE_ONCE(ctx->napi_busy_poll_dt, 0); WRITE_ONCE(ctx->napi_prefer_busy_poll, false); - WRITE_ONCE(ctx->napi_track_mode, IO_URING_NAPI_TRACKING_INACTIVE); + io_napi_free(ctx); return 0; } diff --git a/io_uring/napi.h b/io_uring/napi.h index fa742f42e09b..e0aecccc5065 100644 --- a/io_uring/napi.h +++ b/io_uring/napi.h @@ -15,7 +15,8 @@ void io_napi_free(struct io_ring_ctx *ctx); int io_register_napi(struct io_ring_ctx *ctx, void __user *arg); int io_unregister_napi(struct io_ring_ctx *ctx, void __user *arg); -int __io_napi_add_id(struct io_ring_ctx *ctx, unsigned int napi_id); +int __io_napi_add_id(struct io_ring_ctx *ctx, unsigned int napi_id, + unsigned int mode); void __io_napi_busy_loop(struct io_ring_ctx *ctx, struct io_wait_queue *iowq); int io_napi_sqpoll_busy_poll(struct io_ring_ctx *ctx); @@ -43,13 +44,14 @@ static inline void io_napi_add(struct io_kiocb *req) { struct io_ring_ctx *ctx = req->ctx; struct socket *sock; + unsigned int mode = IO_URING_NAPI_TRACKING_DYNAMIC; - if (READ_ONCE(ctx->napi_track_mode) != IO_URING_NAPI_TRACKING_DYNAMIC) + if (READ_ONCE(ctx->napi_track_mode) != mode) return; sock = sock_from_file(req->file); if (sock && sock->sk) - __io_napi_add_id(ctx, READ_ONCE(sock->sk->sk_napi_id)); + __io_napi_add_id(ctx, READ_ONCE(sock->sk->sk_napi_id), mode); } #else From 04fe9aeb4f3c0999e6715385664c677469dfd8f4 Mon Sep 17 00:00:00 2001 From: Yufan Chen Date: Mon, 4 May 2026 01:57:10 +0800 Subject: [PATCH 062/333] io_uring/eventfd: reset deferred signal state Recursive eventfd wakeups must defer io_uring eventfd signaling because eventfd_signal_mask() rejects reentry from eventfd wakeup handlers. The io_ev_fd ops bit tracks an outstanding deferred signal so that the same rcu_head is not queued twice. That bit is only set today. Once the first deferred callback runs, later recursive notifications still see the bit set and skip queueing another deferred signal. This can leave new completions without a matching eventfd wake after the first recursive deferral. Clear the pending bit before issuing the deferred signal. If the wakeup path recurses while the callback runs, a new signal can be queued for the next RCU grace period while the current callback keeps its reference until it returns. Signed-off-by: Yufan Chen Fixes: 60b6c075e8eb ("io_uring/eventfd: move to more idiomatic RCU free usage") Link: https://patch.msgid.link/20260503175710.37209-1-yufan.chen@linux.dev Signed-off-by: Jens Axboe --- io_uring/eventfd.c | 1 + 1 file changed, 1 insertion(+) diff --git a/io_uring/eventfd.c b/io_uring/eventfd.c index 3da028500f76..d656cc2a0b9b 100644 --- a/io_uring/eventfd.c +++ b/io_uring/eventfd.c @@ -43,6 +43,7 @@ static void io_eventfd_do_signal(struct rcu_head *rcu) { struct io_ev_fd *ev_fd = container_of(rcu, struct io_ev_fd, rcu); + atomic_andnot(BIT(IO_EVENTFD_OP_SIGNAL_BIT), &ev_fd->ops); eventfd_signal_mask(ev_fd->cq_ev_fd, EPOLL_URING_WAKE); io_eventfd_put(ev_fd); } From 8bc6b14aab669f0c301febcf4aa5249b9393bf51 Mon Sep 17 00:00:00 2001 From: Marco Crivellari Date: Thu, 30 Apr 2026 11:08:10 +0200 Subject: [PATCH 063/333] i2c: testunit: Replace system_long_wq with system_dfl_long_wq Currently the code enqueue work items using {queue|mod}_delayed_work(), using system_long_wq. This workqueue should be used when long works are expected, but it is a per-cpu workqueue. This is important because queue_delayed_work() queue the work using: queue_delayed_work_on(WORK_CPU_UNBOUND, ...); Note that WORK_CPU_UNBOUND = NR_CPUS. This would end up calling __queue_delayed_work() that does: if (housekeeping_enabled(HK_TYPE_TIMER)) { // [....] } else { if (likely(cpu == WORK_CPU_UNBOUND)) add_timer_global(timer); else add_timer_on(timer, cpu); } So when cpu == WORK_CPU_UNBOUND the timer is global and is not using a specific CPU. Later, when __queue_work() is called: if (req_cpu == WORK_CPU_UNBOUND) { if (wq->flags & WQ_UNBOUND) cpu = wq_select_unbound_cpu(raw_smp_processor_id()); else cpu = raw_smp_processor_id(); } Because the wq is not unbound, it takes the CPU where the timer fired and enqueue the work on that CPU. The consequence of all of this is that the work can run anywhere, depending on where the timer fired. Recently, a new unbound workqueue specific for long running work has been added: c116737e972e ("workqueue: Add system_dfl_long_wq for long unbound works") So change system_long_wq with system_dfl_long_wq so that the work may benefit from scheduler task placement. Signed-off-by: Marco Crivellari [wsa: remove FIXME as well] Signed-off-by: Wolfram Sang --- drivers/i2c/i2c-slave-testunit.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/i2c/i2c-slave-testunit.c b/drivers/i2c/i2c-slave-testunit.c index 6de4307050dd..871c58461ebc 100644 --- a/drivers/i2c/i2c-slave-testunit.c +++ b/drivers/i2c/i2c-slave-testunit.c @@ -15,7 +15,7 @@ #include #include #include -#include /* FIXME: is system_long_wq the best choice? */ +#include #define TU_VERSION_MAX_LENGTH 128 @@ -124,7 +124,7 @@ static int i2c_slave_testunit_slave_cb(struct i2c_client *client, case I2C_SLAVE_STOP: if (tu->reg_idx == TU_NUM_REGS) { set_bit(TU_FLAG_IN_PROCESS, &tu->flags); - queue_delayed_work(system_long_wq, &tu->worker, + queue_delayed_work(system_dfl_long_wq, &tu->worker, msecs_to_jiffies(10 * tu->regs[TU_REG_DELAY])); } From 9a937ca22741eebe2bf10b18657b8b9aed9c009b Mon Sep 17 00:00:00 2001 From: Ronald Claveau Date: Fri, 24 Apr 2026 16:17:33 +0200 Subject: [PATCH 064/333] dt-bindings: i2c: amlogic: Add compatible for T7 SOC Add the T7 SOC compatible which fallback to AXG compatible. Acked-by: Rob Herring (Arm) Signed-off-by: Ronald Claveau Signed-off-by: Wolfram Sang --- .../devicetree/bindings/i2c/amlogic,meson6-i2c.yaml | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/Documentation/devicetree/bindings/i2c/amlogic,meson6-i2c.yaml b/Documentation/devicetree/bindings/i2c/amlogic,meson6-i2c.yaml index c4cc8af18280..7b59b60b62e5 100644 --- a/Documentation/devicetree/bindings/i2c/amlogic,meson6-i2c.yaml +++ b/Documentation/devicetree/bindings/i2c/amlogic,meson6-i2c.yaml @@ -16,10 +16,15 @@ allOf: properties: compatible: - enum: - - amlogic,meson6-i2c # Meson6, Meson8 and compatible SoCs - - amlogic,meson-gxbb-i2c # GXBB and compatible SoCs - - amlogic,meson-axg-i2c # AXG and compatible SoCs + oneOf: + - items: + - enum: + - amlogic,t7-i2c + - const: amlogic,meson-axg-i2c + - enum: + - amlogic,meson6-i2c # Meson6, Meson8 and compatible SoCs + - amlogic,meson-gxbb-i2c # GXBB and compatible SoCs + - amlogic,meson-axg-i2c # AXG and compatible SoCs reg: maxItems: 1 From 02b7cd683892d8d8464815d69a3a76579909a726 Mon Sep 17 00:00:00 2001 From: Marek Vasut Date: Sat, 2 May 2026 17:31:54 +0200 Subject: [PATCH 065/333] i2c: stm32f7: reinit_completion() per transfer not per msg Currently, the driver may repeatedly call reinit_completion() during transfer which contains multiple messages, while another thread is waiting for the completion. This happens during transfer with more than 1 message, invoked via stm32f7_i2c_xfer_core() -> stm32f7_i2c_xfer_msg(). After invoking the stm32f7_i2c_xfer_msg() to start transfer, stm32f7_i2c_xfer_core() calls wait_for_completion_timeout() to wait for completion of the transfer of all messages. When the first message transfer completes, the hard IRQ handler triggers, and detects transfer completion, which leads to stm32f7_i2c_isr_event_thread() IRQ thread being started. The stm32f7_i2c_isr_event_thread() calls stm32f7_i2c_xfer_msg() in case there are more messages. Without this change, the second and later stm32f7_i2c_xfer_msg() would call reinit_completion() on the completion which is still being waited for in stm32f7_i2c_xfer_core(). Fix this by moving the reinit_completion() into stm32f7_i2c_xfer_core(), together with wait_for_completion_timeout(). Since stm32f7_i2c_xfer_core() now waits for completion of the entire transfer, increase the default timeout. This fixes sporadic transfer timeouts on STM32MP25xx during kernel boot. Fixes: aeb068c57214 ("i2c: i2c-stm32f7: add driver") Signed-off-by: Marek Vasut [wsa: reworded commit subject] Signed-off-by: Wolfram Sang --- drivers/i2c/busses/i2c-stm32f7.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/i2c/busses/i2c-stm32f7.c b/drivers/i2c/busses/i2c-stm32f7.c index 70cb5822bf17..53d9df70ebe4 100644 --- a/drivers/i2c/busses/i2c-stm32f7.c +++ b/drivers/i2c/busses/i2c-stm32f7.c @@ -895,8 +895,6 @@ static void stm32f7_i2c_xfer_msg(struct stm32f7_i2c_dev *i2c_dev, f7_msg->result = 0; f7_msg->stop = (i2c_dev->msg_id >= i2c_dev->msg_num - 1); - reinit_completion(&i2c_dev->complete); - cr1 = readl_relaxed(base + STM32F7_I2C_CR1); cr2 = readl_relaxed(base + STM32F7_I2C_CR2); @@ -1728,6 +1726,8 @@ static int stm32f7_i2c_xfer_core(struct i2c_adapter *i2c_adap, if (ret) goto pm_free; + reinit_completion(&i2c_dev->complete); + stm32f7_i2c_xfer_msg(i2c_dev, msgs); if (!i2c_dev->atomic) @@ -2253,7 +2253,7 @@ static int stm32f7_i2c_probe(struct platform_device *pdev) snprintf(adap->name, sizeof(adap->name), "STM32F7 I2C(%pa)", &res->start); adap->owner = THIS_MODULE; - adap->timeout = 2 * HZ; + adap->timeout = 8 * HZ; adap->retries = 3; adap->algo = &stm32f7_i2c_algo; adap->dev.parent = &pdev->dev; From 10161b4a791d5c4b7ea16512c1ddb133c3f8f953 Mon Sep 17 00:00:00 2001 From: Weinan Liu Date: Thu, 30 Apr 2026 23:28:51 +0000 Subject: [PATCH 066/333] iommu/amd: Fix precedence order in set_dte_passthrough() Bitwise OR | operator has a higher precedence than the ternary ?: operatior. It will be incorrectly evaluated as: new->data[1] |= (FIELD_PREP(...) | dev_data->ats_enabled) ? DTE_FLAG_IOTLB : 0; Wrap the conditional operation in parentheses to enforce the correct evaluation order. Fixes: 93eee2a49c1b ("iommu/amd: Refactor logic to program the host page table in DTE") Signed-off-by: Weinan Liu Reviewed-by: Jason Gunthorpe Reviewed-by: Vasant Hegde Signed-off-by: Joerg Roedel --- drivers/iommu/amd/iommu.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/iommu/amd/iommu.c b/drivers/iommu/amd/iommu.c index 157505d96fdd..f78e23f03938 100644 --- a/drivers/iommu/amd/iommu.c +++ b/drivers/iommu/amd/iommu.c @@ -2149,7 +2149,8 @@ static void set_dte_passthrough(struct iommu_dev_data *dev_data, new->data[0] |= DTE_FLAG_TV | DTE_FLAG_IR | DTE_FLAG_IW; new->data[1] |= FIELD_PREP(DTE_DOMID_MASK, domain->id) | - (dev_data->ats_enabled) ? DTE_FLAG_IOTLB : 0; + (dev_data->ats_enabled ? DTE_FLAG_IOTLB : 0); + } static void set_dte_entry(struct amd_iommu *iommu, From c5f25f5800f56f1754d9eeb3ced7c1e08c29119a Mon Sep 17 00:00:00 2001 From: Janne Grunau Date: Fri, 20 Mar 2026 13:23:24 +0100 Subject: [PATCH 067/333] dt-bindings: i2c: apple,i2c: Add t8122 compatible The i2c block on the Apple silicon t8122 (M3) SoC is compatible with the existing driver. Add "apple,t8122-i2c" as SoC specific compatible under "apple,t8103-i2c" used by the deriver. Signed-off-by: Janne Grunau Acked-by: Andi Shyti Acked-by: Rob Herring (Arm) Signed-off-by: Wolfram Sang --- Documentation/devicetree/bindings/i2c/apple,i2c.yaml | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/Documentation/devicetree/bindings/i2c/apple,i2c.yaml b/Documentation/devicetree/bindings/i2c/apple,i2c.yaml index 500a965bdb7a..9e59200ad37b 100644 --- a/Documentation/devicetree/bindings/i2c/apple,i2c.yaml +++ b/Documentation/devicetree/bindings/i2c/apple,i2c.yaml @@ -22,7 +22,9 @@ properties: compatible: oneOf: - items: - - const: apple,t6020-i2c + - enum: + - apple,t6020-i2c + - apple,t8122-i2c - const: apple,t8103-i2c - items: - enum: From 9998e388be9930c106eb5904c23ecf2162407527 Mon Sep 17 00:00:00 2001 From: Niels Franke Date: Sat, 18 Apr 2026 07:37:19 +0200 Subject: [PATCH 068/333] i2c: acpi: Add ELAN0678 to i2c_acpi_force_100khz_device_ids The ELAN0678 touchpad (04F3:3195) found in the Lenovo ThinkPad X13 exhibits excessive smoothing when the I2C bus runs at 400KHz, making the touchpad feel sluggish when plugged into AC power. This is the same issue previously fixed for ELAN06FA. The device's ACPI table (Lenovo TP-R22) specifies 0x00061A80 (400KHz) for the I2cSerialBusV2 descriptor. Forcing the bus to 100KHz eliminates the sluggish behavior. Signed-off-by: Niels Franke Acked-by: Mika Westerberg [wsa: kept the sorting] Signed-off-by: Wolfram Sang --- drivers/i2c/i2c-core-acpi.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/i2c/i2c-core-acpi.c b/drivers/i2c/i2c-core-acpi.c index 2cbd31f77667..28c0e4884a7f 100644 --- a/drivers/i2c/i2c-core-acpi.c +++ b/drivers/i2c/i2c-core-acpi.c @@ -371,6 +371,7 @@ static const struct acpi_device_id i2c_acpi_force_100khz_device_ids[] = { * a 400KHz frequency. The root cause of the issue is not known. */ { "DLL0945", 0 }, + { "ELAN0678", 0 }, { "ELAN06FA", 0 }, {} }; From 32c91e8ee039777d0b95b914633fc6a42607959c Mon Sep 17 00:00:00 2001 From: Johan Hovold Date: Fri, 24 Apr 2026 12:49:10 +0200 Subject: [PATCH 069/333] staging: vme_user: fix root device leak on init failure Make sure to deregister and free the root device in case module initialisation fails. Fixes: 658bcdae9c67 ("vme: Adding Fake VME driver") Cc: stable@vger.kernel.org # 4.9 Cc: Martyn Welch Signed-off-by: Johan Hovold Link: https://patch.msgid.link/20260424104910.2619349-1-johan@kernel.org Signed-off-by: Greg Kroah-Hartman --- drivers/staging/vme_user/vme_fake.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/staging/vme_user/vme_fake.c b/drivers/staging/vme_user/vme_fake.c index be4ad47ed526..8abaa3165fbb 100644 --- a/drivers/staging/vme_user/vme_fake.c +++ b/drivers/staging/vme_user/vme_fake.c @@ -1230,6 +1230,8 @@ static int __init fake_init(void) err_driver: kfree(fake_bridge); err_struct: + root_device_unregister(vme_root); + return retval; } From 617eb7c0961a8dfcfc811844a6396e406b2923ea Mon Sep 17 00:00:00 2001 From: Mingyu Wang <25181214217@stu.xidian.edu.cn> Date: Mon, 27 Apr 2026 10:57:45 +0800 Subject: [PATCH 070/333] i2c: dev: prevent integer overflow in I2C_TIMEOUT ioctl While fuzzing with Syzkaller, a persistent `schedule_timeout: wrong timeout value` warning was observed, accompanied by SMBus controller state machine corruption. The I2C_TIMEOUT ioctl accepts a user-provided timeout in multiples of 10 ms. The user argument is checked against INT_MAX, but it is subsequently multiplied by 10 before being passed to msecs_to_jiffies(). A malicious user can pass a large value (e.g., 429496729) that passes the `arg > INT_MAX` check but overflows when multiplied by 10. This results in a truncated 32-bit unsigned value that bypasses the internal `(int)m < 0` check in `msecs_to_jiffies()`. The truncated value is then assigned to `client->adapter->timeout` (a signed 32-bit int), which is reinterpreted as a negative number. When passed to wait_for_completion_timeout(), this negative value undergoes sign extension to a 64-bit unsigned long, triggering the `schedule_timeout` warning and causing premature returns. This leaves the SMBus state machine in an unrecoverable state, constituting a local Denial of Service (DoS). Fix this by bounding the user argument to `INT_MAX / 10`. Signed-off-by: Mingyu Wang <25181214217@stu.xidian.edu.cn> [wsa: move the comment as well] Signed-off-by: Wolfram Sang --- drivers/i2c/i2c-dev.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/i2c/i2c-dev.c b/drivers/i2c/i2c-dev.c index 7bbe0263411e..ccaac5e29f90 100644 --- a/drivers/i2c/i2c-dev.c +++ b/drivers/i2c/i2c-dev.c @@ -487,12 +487,13 @@ static long i2cdev_ioctl(struct file *file, unsigned int cmd, unsigned long arg) client->adapter->retries = arg; break; case I2C_TIMEOUT: - if (arg > INT_MAX) + /* + * For historical reasons, user-space sets the timeout value in + * units of 10 ms. + */ + if (arg > INT_MAX / 10) return -EINVAL; - /* For historical reasons, user-space sets the timeout - * value in units of 10 ms. - */ client->adapter->timeout = msecs_to_jiffies(arg * 10); break; default: From bc851db06045a40c18233dd76ef0562d7f8bb6db Mon Sep 17 00:00:00 2001 From: Shyam Sunder Reddy Padira Date: Tue, 14 Apr 2026 12:43:06 +0530 Subject: [PATCH 071/333] staging: rtl8723bs: os_dep: avoid NULL pointer dereference in rtw_cbuf_alloc The return value of kzalloc_flex() is used without ensuring that the allocation succeeded, and the pointer is dereferenced unconditionally. Guard the access to the allocated structure to avoid a potential NULL pointer dereference if the allocation fails. Fixes: 980cd426a257 ("staging: rtl8723bs: replace rtw_zmalloc() with kzalloc()") Cc: stable Signed-off-by: Shyam Sunder Reddy Padira Reviewed-by: Dan Carpenter Link: https://patch.msgid.link/20260414071308.4781-2-shyamsunderreddypadira@gmail.com Signed-off-by: Greg Kroah-Hartman --- drivers/staging/rtl8723bs/os_dep/osdep_service.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/staging/rtl8723bs/os_dep/osdep_service.c b/drivers/staging/rtl8723bs/os_dep/osdep_service.c index 7959daeabc6f..4cfdf7c62344 100644 --- a/drivers/staging/rtl8723bs/os_dep/osdep_service.c +++ b/drivers/staging/rtl8723bs/os_dep/osdep_service.c @@ -194,7 +194,8 @@ struct rtw_cbuf *rtw_cbuf_alloc(u32 size) struct rtw_cbuf *cbuf; cbuf = kzalloc_flex(*cbuf, bufs, size); - cbuf->size = size; + if (cbuf) + cbuf->size = size; return cbuf; } From b47bc7c022ddab7c79a84dd5f3f0d07fe09ec786 Mon Sep 17 00:00:00 2001 From: "Nikola Z. Ivanov" Date: Wed, 15 Apr 2026 23:50:21 +0300 Subject: [PATCH 072/333] i2c: Compare the return value of gpiod_get_direction against GPIO_LINE_DIRECTION_OUT The GPIO_LINE_DIRECTION_* definitions have just recently been exposed to gpio consumers.h by breaking them out in a separate defs.h file. Use this to validate the gpio direction instead of the hard-coded literal. Signed-off-by: Nikola Z. Ivanov Signed-off-by: Wolfram Sang --- drivers/i2c/i2c-core-base.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/i2c/i2c-core-base.c b/drivers/i2c/i2c-core-base.c index 9c46147e3506..a2132d70fb36 100644 --- a/drivers/i2c/i2c-core-base.c +++ b/drivers/i2c/i2c-core-base.c @@ -445,8 +445,7 @@ static int i2c_init_recovery(struct i2c_adapter *adap) bri->set_scl = set_scl_gpio_value; if (bri->sda_gpiod) { bri->get_sda = get_sda_gpio_value; - /* FIXME: add proper flag instead of '0' once available */ - if (gpiod_get_direction(bri->sda_gpiod) == 0) + if (gpiod_get_direction(bri->sda_gpiod) == GPIO_LINE_DIRECTION_OUT) bri->set_sda = set_sda_gpio_value; } } else if (bri->recover_bus == i2c_generic_scl_recovery) { From 6036b5067a8199ba7a2dc7b377d4b9dd276d5f9e Mon Sep 17 00:00:00 2001 From: Weiming Shi Date: Wed, 15 Apr 2026 01:23:39 +0800 Subject: [PATCH 073/333] i2c: stub: Reject I2C block transfers with invalid length The I2C_SMBUS_I2C_BLOCK_DATA case in stub_xfer() uses data->block[0] as the transfer length. The existing check only clamps it to avoid overrunning the chip->words[256] register array, but does not validate it against I2C_SMBUS_BLOCK_MAX (32), which is the limit of the union i2c_smbus_data.block buffer (34 bytes total). The driver is a development/test tool (CONFIG_I2C_STUB=m, not built by default) that must be loaded with a chip_addr= parameter. A local user with access to /dev/i2c-* can issue an I2C_SMBUS ioctl with I2C_SMBUS_I2C_BLOCK_DATA and data->block[0] > 32, causing stub_xfer() to read or write past the end of the union i2c_smbus_data.block buffer: BUG: KASAN: stack-out-of-bounds in stub_xfer (drivers/i2c/i2c-stub.c:223) Read of size 1 at addr ffff88800abcfd92 by task exploit/81 Call Trace: stub_xfer (drivers/i2c/i2c-stub.c:223) __i2c_smbus_xfer (drivers/i2c/i2c-core-smbus.c:593) i2c_smbus_xfer (drivers/i2c/i2c-core-smbus.c:536) i2cdev_ioctl_smbus (drivers/i2c/i2c-dev.c:391) i2cdev_ioctl (drivers/i2c/i2c-dev.c:478) __x64_sys_ioctl (fs/ioctl.c:583) do_syscall_64 (arch/x86/entry/syscall_64.c:94) entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130) The bug exists because i2c-stub implements .smbus_xfer directly, bypassing the I2C_SMBUS_BLOCK_MAX validation in i2c_smbus_xfer_emulated(). The I2C_SMBUS_BLOCK_DATA case in the same function correctly validates against I2C_SMBUS_BLOCK_MAX, but the I2C_SMBUS_I2C_BLOCK_DATA case does not. Fix by rejecting transfers with data->block[0] == 0 or data->block[0] > I2C_SMBUS_BLOCK_MAX with -EINVAL, consistent with both the I2C_SMBUS_BLOCK_DATA case in the same function and the I2C_SMBUS_I2C_BLOCK_DATA validation in i2c_smbus_xfer_emulated(). Fixes: 4710317891e4 ("i2c-stub: Implement I2C block support") Reported-by: Xiang Mei Signed-off-by: Weiming Shi Reviewed-by: Jean Delvare Signed-off-by: Wolfram Sang --- drivers/i2c/i2c-stub.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/i2c/i2c-stub.c b/drivers/i2c/i2c-stub.c index fbb0db41b10e..04314e3ed24c 100644 --- a/drivers/i2c/i2c-stub.c +++ b/drivers/i2c/i2c-stub.c @@ -214,6 +214,11 @@ static s32 stub_xfer(struct i2c_adapter *adap, u16 addr, unsigned short flags, * We ignore banks here, because banked chips don't use I2C * block transfers */ + if (data->block[0] == 0 || + data->block[0] > I2C_SMBUS_BLOCK_MAX) { + ret = -EINVAL; + break; + } if (data->block[0] > 256 - command) /* Avoid overrun */ data->block[0] = 256 - command; len = data->block[0]; From 8acd2d7e0889ac62bc102bd7b648cd7bee04f902 Mon Sep 17 00:00:00 2001 From: Myeonghun Pak Date: Fri, 24 Apr 2026 20:25:18 +0900 Subject: [PATCH 074/333] drm/qxl: Fix missing KMS poll cleanup drm_kms_helper_poll_init() initializes the output polling work and enables polling for the DRM device. qxl enables polling before calling drm_dev_register(), but the drm_dev_register() failure path tears down the modeset and device state without disabling the polling helper. The remove path also unregisters and shuts down the DRM device without first disabling the polling helper. Add matching drm_kms_helper_poll_fini() calls in both paths so the delayed polling work is cancelled before qxl tears down the associated modeset/device state. Signed-off-by: Myeonghun Pak Reviewed-by: Thomas Zimmermann Fixes: 5ff91e442652 ("qxl: use drm helper hotplug support") Signed-off-by: Thomas Zimmermann Link: https://patch.msgid.link/20260424112543.57819-1-mhun512@gmail.com --- drivers/gpu/drm/qxl/qxl_drv.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/qxl/qxl_drv.c b/drivers/gpu/drm/qxl/qxl_drv.c index 2bbb1168a3ff..1e6a2392d7c6 100644 --- a/drivers/gpu/drm/qxl/qxl_drv.c +++ b/drivers/gpu/drm/qxl/qxl_drv.c @@ -118,12 +118,13 @@ qxl_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) /* Complete initialization. */ ret = drm_dev_register(&qdev->ddev, ent->driver_data); if (ret) - goto modeset_cleanup; + goto poll_fini; drm_client_setup(&qdev->ddev, NULL); return 0; -modeset_cleanup: +poll_fini: + drm_kms_helper_poll_fini(&qdev->ddev); qxl_modeset_fini(qdev); unload: qxl_device_fini(qdev); @@ -154,6 +155,7 @@ qxl_pci_remove(struct pci_dev *pdev) { struct drm_device *dev = pci_get_drvdata(pdev); + drm_kms_helper_poll_fini(dev); drm_dev_unregister(dev); drm_atomic_helper_shutdown(dev); if (pci_is_vga(pdev) && pdev->revision < 5) From c28c22c8cfbd43f2ad71a157324d9fbebc0d0f2e Mon Sep 17 00:00:00 2001 From: Francesco Lavra Date: Tue, 10 Feb 2026 18:35:45 +0100 Subject: [PATCH 075/333] drm/fb-helper: Fix clipping when damage area spans a single scanline When the damage area resulting from a dirty memory range spans a single scanline, the width of the rectangle is calculated dynamically because it may not coincide with the framebuffer width. If the dirty range ends exactly at the end of the scanline, the `bit_end` variable is incorrectly assigned a 0 value, which results in a bogus clip rectangle where the x2 coordinate is 0. This prevents the dirty scanline from being flushed to the hardware. Change the calculation of the `bit_end` value to fix the x2 coordinate value in the above edge case. Fixes: ded74cafeea9 ("drm/fb-helper: Clip damage area horizontally") Signed-off-by: Francesco Lavra Reviewed-by: Thomas Zimmermann Signed-off-by: Thomas Zimmermann Link: https://patch.msgid.link/20260210173545.733937-1-flavra@baylibre.com --- drivers/gpu/drm/drm_fb_helper.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index a80a335f4148..1541fc8a9ac2 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -490,7 +490,7 @@ static void drm_fb_helper_memory_range_to_clip(struct fb_info *info, off_t off, * the number of horizontal pixels that need an update. */ off_t bit_off = (off % line_length) * 8; - off_t bit_end = (end % line_length) * 8; + off_t bit_end = bit_off + len * 8; x1 = bit_off / info->var.bits_per_pixel; x2 = DIV_ROUND_UP(bit_end, info->var.bits_per_pixel); From 24e0fd8b852062d5e8a740f7945eaa26818adce8 Mon Sep 17 00:00:00 2001 From: John Madieu Date: Fri, 1 May 2026 13:59:49 +0000 Subject: [PATCH 076/333] spi: imx: Fix precedence bug in spi_imx_dma_max_wml_find() The watermark search in spi_imx_dma_max_wml_find() reads: if (!dma_data->dma_len % (i * bytes_per_word)) break; The unary ! binds tighter than %, so this parses as: if ((!dma_data->dma_len) % (i * bytes_per_word)) break; !dma_data->dma_len is 0 or 1, and `0 % x == 0` for any x; `1 % x` is 0 unless x == 1. The condition is therefore false in every case except dma_len != 0 with i * bytes_per_word == 1, i.e. i == 1 and bytes_per_word == 1. The loop almost always falls through to its end, leaving i == 0, which the post-loop fallback rewrites to 1: if (i == 0) i = 1; So spi_imx->wml ends up at 1 for essentially every DMA transfer, defeating the entire purpose of the function. The DMA engine then requests service after every single FIFO word instead of using multi-word bursts, hurting throughput on every DMA-capable variant. Add the missing parentheses so the modulo is computed first, then negated: if (!(dma_data->dma_len % (i * bytes_per_word))) break; Fixes: faa8e404ad8e ("spi: imx: support dynamic burst length for ECSPI DMA mode") Signed-off-by: John Madieu Reviewed-by: Frank Li Link: https://patch.msgid.link/20260501135951.2416527-2-john.madieu@gmail.com Signed-off-by: Mark Brown --- drivers/spi/spi-imx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/spi/spi-imx.c b/drivers/spi/spi-imx.c index e5c907c45b87..7ae8078c10ef 100644 --- a/drivers/spi/spi-imx.c +++ b/drivers/spi/spi-imx.c @@ -1836,7 +1836,7 @@ static void spi_imx_dma_max_wml_find(struct spi_imx_data *spi_imx, unsigned int i; for (i = spi_imx->devtype_data->fifo_size / 2; i > 0; i--) { - if (!dma_data->dma_len % (i * bytes_per_word)) + if (!(dma_data->dma_len % (i * bytes_per_word))) break; } /* Use 1 as wml in case no available burst length got */ From f5b5548255040ec3bef05bcb1e9c9c3614dfa7db Mon Sep 17 00:00:00 2001 From: John Madieu Date: Fri, 1 May 2026 13:59:50 +0000 Subject: [PATCH 077/333] spi: imx: Fix UAF on package-1 prepare failure in spi_imx_dma_data_prepare() When transfer->len exceeds MX51_ECSPI_CTRL_MAX_BURST and is not a multiple of it, spi_imx_dma_data_prepare() splits the transfer into two DMA packages. If preparing the second package fails: ret = spi_imx_dma_tx_data_handle(spi_imx, &spi_imx->dma_data[1], transfer->tx_buf + spi_imx->dma_data[0].data_len, false); if (ret) { kfree(spi_imx->dma_data[0].dma_tx_buf); kfree(spi_imx->dma_data[0].dma_rx_buf); kfree(spi_imx->dma_data); } } return 0; the function frees the package-0 buffers and the dma_data array, then falls through to `return 0`, telling the caller the prepare succeeded. The caller then dereferences the freed dma_data array, producing a use-after-free. Return the error from the failure path so the caller takes its existing failure branch. Fixes: faa8e404ad8e ("spi: imx: support dynamic burst length for ECSPI DMA mode") Signed-off-by: John Madieu Reviewed-by: Frank Li Link: https://patch.msgid.link/20260501135951.2416527-3-john.madieu@gmail.com Signed-off-by: Mark Brown --- drivers/spi/spi-imx.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/spi/spi-imx.c b/drivers/spi/spi-imx.c index 7ae8078c10ef..4e3dbd01d619 100644 --- a/drivers/spi/spi-imx.c +++ b/drivers/spi/spi-imx.c @@ -1709,6 +1709,7 @@ static int spi_imx_dma_data_prepare(struct spi_imx_data *spi_imx, kfree(spi_imx->dma_data[0].dma_tx_buf); kfree(spi_imx->dma_data[0].dma_rx_buf); kfree(spi_imx->dma_data); + return ret; } } From 894e04b7116297a6529e0c4ed90e3eb160939805 Mon Sep 17 00:00:00 2001 From: John Madieu Date: Fri, 1 May 2026 13:59:51 +0000 Subject: [PATCH 078/333] spi: imx: Propagate prepare_transfer() error from spi_imx_setupxfer() spi_imx_setupxfer() calls the per-variant prepare_transfer() callback and returns 0 unconditionally: spi_imx->devtype_data->prepare_transfer(spi_imx, spi, t); return 0; mx51_ecspi_prepare_transfer() can return -EINVAL when the requested word_delay does not fit in MX51_ECSPI_PERIOD_MASK. The error is detected after a partial set of register writes (CTRL: BL, clkdiv, SMC), so the controller is left in a partially-configured state and the transfer is then submitted as if setup succeeded. Propagate the return value. The other variants' prepare_transfer callbacks all return 0, so this is a no-op for them. Fixes: a3bb4e663df3 ("spi: imx: support word delay") Signed-off-by: John Madieu Reviewed-by: Frank Li Link: https://patch.msgid.link/20260501135951.2416527-4-john.madieu@gmail.com Signed-off-by: Mark Brown --- drivers/spi/spi-imx.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/spi/spi-imx.c b/drivers/spi/spi-imx.c index 4e3dbd01d619..480d1e8b281f 100644 --- a/drivers/spi/spi-imx.c +++ b/drivers/spi/spi-imx.c @@ -1382,9 +1382,7 @@ static int spi_imx_setupxfer(struct spi_device *spi, spi_imx->target_burst = t->len; } - spi_imx->devtype_data->prepare_transfer(spi_imx, spi, t); - - return 0; + return spi_imx->devtype_data->prepare_transfer(spi_imx, spi, t); } static void spi_imx_sdma_exit(struct spi_imx_data *spi_imx) From 7672749e1496215e8683ce57cf323119033954cf Mon Sep 17 00:00:00 2001 From: Conor Dooley Date: Thu, 30 Apr 2026 11:10:18 +0100 Subject: [PATCH 079/333] spi: microchip-core-qspi: control built-in cs manually The coreQSPI IP supports only a single chip select, which is automagically operated by the hardware - set low when the transmit buffer first gets written to and set high when the number of bytes written to the TOTALBYTES field of the FRAMES register have been sent on the bus. Additional devices must use GPIOs for their chip selects. It was reported to me that if there are two devices attached to this QSPI controller that the in-built chip select is set low while linux tries to access the device attached to the GPIO. This went undetected as the boards that connected multiple devices to the SPI controller all exclusively used GPIOs for chip selects, not relying on the built-in chip select at all. It turns out that this was because the built-in chip select, when controlled automagically, is set low when active and high when inactive, thereby ruling out its use for active-high devices or devices that need to transmit with the chip select disabled. Modify the driver so that it controls chip select directly, retaining the behaviour for mem_ops of setting the chip select active for the entire duration of the transfer in the exec_op callback. For regular transfers, implement the set_cs callback for the core to use. As part of this, the existing setup callback, mchp_coreqspi_setup_op(), is removed. Modifying the CLKIDLE field is not safe to do during operation when there are multiple devices, so this code is removed entirely. Setting the MASTER and ENABLE fields is something that can be done once at probe, it doesn't need to be re-run for each device. Instead the new setup callback sets the built-in chip select to its inactive state for active-low devices, as the reset value of the chip select in software controlled mode is low. Fixes: 8f9cf02c88528 ("spi: microchip-core-qspi: Add regular transfers") Fixes: 8596124c4c1bc ("spi: microchip-core-qspi: Add support for microchip fpga qspi controllers") CC: stable@vger.kernel.org Signed-off-by: Conor Dooley Link: https://patch.msgid.link/20260430-hamstring-busload-f941d0347b5e@spud Signed-off-by: Mark Brown --- drivers/spi/spi-microchip-core-qspi.c | 79 ++++++++++++++++++++++----- 1 file changed, 64 insertions(+), 15 deletions(-) diff --git a/drivers/spi/spi-microchip-core-qspi.c b/drivers/spi/spi-microchip-core-qspi.c index eab059fb0bc2..ffa0f33a0ae0 100644 --- a/drivers/spi/spi-microchip-core-qspi.c +++ b/drivers/spi/spi-microchip-core-qspi.c @@ -74,6 +74,13 @@ #define STATUS_FLAGSX4 BIT(8) #define STATUS_MASK GENMASK(8, 0) +/* + * QSPI Direct Access register defines + */ +#define DIRECT_ACCESS_EN_SSEL BIT(0) +#define DIRECT_ACCESS_OP_SSEL BIT(1) +#define DIRECT_ACCESS_OP_SSEL_SHIFT 1 + #define BYTESUPPER_MASK GENMASK(31, 16) #define BYTESLOWER_MASK GENMASK(15, 0) @@ -158,6 +165,38 @@ static int mchp_coreqspi_set_mode(struct mchp_coreqspi *qspi, const struct spi_m return 0; } +static void mchp_coreqspi_set_cs(struct spi_device *spi, bool enable) +{ + struct mchp_coreqspi *qspi = spi_controller_get_devdata(spi->controller); + u32 val; + + val = readl(qspi->regs + REG_DIRECT_ACCESS); + + val &= ~DIRECT_ACCESS_OP_SSEL; + val |= !enable << DIRECT_ACCESS_OP_SSEL_SHIFT; + + writel(val, qspi->regs + REG_DIRECT_ACCESS); +} + +static int mchp_coreqspi_setup(struct spi_device *spi) +{ + struct mchp_coreqspi *qspi = spi_controller_get_devdata(spi->controller); + u32 val; + + /* + * Active low devices need to be specifically set to their inactive + * states during probe. + */ + if (spi->mode & SPI_CS_HIGH) + return 0; + + val = readl(qspi->regs + REG_DIRECT_ACCESS); + val |= DIRECT_ACCESS_OP_SSEL; + writel(val, qspi->regs + REG_DIRECT_ACCESS); + + return 0; +} + static inline void mchp_coreqspi_read_op(struct mchp_coreqspi *qspi) { u32 control, data; @@ -380,19 +419,6 @@ static int mchp_coreqspi_setup_clock(struct mchp_coreqspi *qspi, struct spi_devi return 0; } -static int mchp_coreqspi_setup_op(struct spi_device *spi_dev) -{ - struct spi_controller *ctlr = spi_dev->controller; - struct mchp_coreqspi *qspi = spi_controller_get_devdata(ctlr); - u32 control = readl_relaxed(qspi->regs + REG_CONTROL); - - control |= (CONTROL_MASTER | CONTROL_ENABLE); - control &= ~CONTROL_CLKIDLE; - writel_relaxed(control, qspi->regs + REG_CONTROL); - - return 0; -} - static inline void mchp_coreqspi_config_op(struct mchp_coreqspi *qspi, const struct spi_mem_op *op) { u32 idle_cycles = 0; @@ -483,6 +509,7 @@ static int mchp_coreqspi_exec_op(struct spi_mem *mem, const struct spi_mem_op *o reinit_completion(&qspi->data_completion); mchp_coreqspi_config_op(qspi, op); + mchp_coreqspi_set_cs(mem->spi, true); if (op->cmd.opcode) { qspi->txbuf = &opcode; qspi->rxbuf = NULL; @@ -523,6 +550,7 @@ static int mchp_coreqspi_exec_op(struct spi_mem *mem, const struct spi_mem_op *o err = -ETIMEDOUT; error: + mchp_coreqspi_set_cs(mem->spi, false); mutex_unlock(&qspi->op_lock); mchp_coreqspi_disable_ints(qspi); @@ -686,6 +714,7 @@ static int mchp_coreqspi_probe(struct platform_device *pdev) struct device *dev = &pdev->dev; struct device_node *np = dev->of_node; int ret; + u32 num_cs, val; ctlr = devm_spi_alloc_host(&pdev->dev, sizeof(*qspi)); if (!ctlr) @@ -718,10 +747,18 @@ static int mchp_coreqspi_probe(struct platform_device *pdev) return ret; } + /* + * The IP core only has a single CS, any more have to be provided via + * gpios + */ + if (of_property_read_u32(pdev->dev.of_node, "num-cs", &num_cs)) + num_cs = 1; + + ctlr->num_chipselect = num_cs; + ctlr->bits_per_word_mask = SPI_BPW_MASK(8); ctlr->mem_ops = &mchp_coreqspi_mem_ops; ctlr->mem_caps = &mchp_coreqspi_mem_caps; - ctlr->setup = mchp_coreqspi_setup_op; ctlr->mode_bits = SPI_CPOL | SPI_CPHA | SPI_RX_DUAL | SPI_RX_QUAD | SPI_TX_DUAL | SPI_TX_QUAD; ctlr->dev.of_node = np; @@ -729,9 +766,21 @@ static int mchp_coreqspi_probe(struct platform_device *pdev) ctlr->prepare_message = mchp_coreqspi_prepare_message; ctlr->unprepare_message = mchp_coreqspi_unprepare_message; ctlr->transfer_one = mchp_coreqspi_transfer_one; - ctlr->num_chipselect = 2; + ctlr->setup = mchp_coreqspi_setup; + ctlr->set_cs = mchp_coreqspi_set_cs; ctlr->use_gpio_descriptors = true; + val = readl_relaxed(qspi->regs + REG_CONTROL); + val |= (CONTROL_MASTER | CONTROL_ENABLE); + writel_relaxed(val, qspi->regs + REG_CONTROL); + + /* + * Put cs into software controlled mode + */ + val = readl_relaxed(qspi->regs + REG_DIRECT_ACCESS); + val |= DIRECT_ACCESS_EN_SSEL; + writel(val, qspi->regs + REG_DIRECT_ACCESS); + ret = spi_register_controller(ctlr); if (ret) return dev_err_probe(&pdev->dev, ret, From eb56deaabf127e8985fc91fa6c97bf8a3b062844 Mon Sep 17 00:00:00 2001 From: Conor Dooley Date: Thu, 30 Apr 2026 11:10:19 +0100 Subject: [PATCH 080/333] spi: microchip-core-qspi: don't attempt to transmit during emulated read-only dual/quad operations The core will deal with reads by creating clock cycles itself, there's no need to generate clock cycles by transmitting garbage data at the driver level. Further, transmitting garbage data just bricks the transfer since QSPI doesn't have a dedicated master-out line like MOSI in regular SPI. I'm not entirely sure if the transfer is bricked because of the garbage data being transmitted on the bus or because the core loses track of whether it is supposed to be sending or receiving data. Fixes: 8f9cf02c88528 ("spi: microchip-core-qspi: Add regular transfers") CC: stable@vger.kernel.org Signed-off-by: Conor Dooley Link: https://patch.msgid.link/20260430-freezing-saloon-95b1f3d9dad0@spud Signed-off-by: Mark Brown --- drivers/spi/spi-microchip-core-qspi.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) diff --git a/drivers/spi/spi-microchip-core-qspi.c b/drivers/spi/spi-microchip-core-qspi.c index ffa0f33a0ae0..70215a407b5a 100644 --- a/drivers/spi/spi-microchip-core-qspi.c +++ b/drivers/spi/spi-microchip-core-qspi.c @@ -690,18 +690,28 @@ static int mchp_coreqspi_transfer_one(struct spi_controller *ctlr, struct spi_de struct spi_transfer *t) { struct mchp_coreqspi *qspi = spi_controller_get_devdata(ctlr); + bool dual_quad = false; qspi->tx_len = t->len; + if (t->tx_nbits == SPI_NBITS_QUAD || t->rx_nbits == SPI_NBITS_QUAD || + t->tx_nbits == SPI_NBITS_DUAL || + t->rx_nbits == SPI_NBITS_DUAL) + dual_quad = true; + if (t->tx_buf) qspi->txbuf = (u8 *)t->tx_buf; if (!t->rx_buf) { mchp_coreqspi_write_op(qspi); - } else { + } else if (!dual_quad) { qspi->rxbuf = (u8 *)t->rx_buf; qspi->rx_len = t->len; mchp_coreqspi_write_read_op(qspi); + } else { + qspi->rxbuf = (u8 *)t->rx_buf; + qspi->rx_len = t->len; + mchp_coreqspi_read_op(qspi); } return 0; From 0b2eb1f8473eddeff5317e521498329581432f89 Mon Sep 17 00:00:00 2001 From: Conor Dooley Date: Thu, 30 Apr 2026 11:10:20 +0100 Subject: [PATCH 081/333] spi: microchip-core-qspi: remove some inline markings Remove inline markings from a number of functions that are called as part of mem ops callbacks. None of them are either particularly trivial or sensitive to overhead of a function call. Just let the compiler decide what to do with them. Signed-off-by: Conor Dooley Link: https://patch.msgid.link/20260430-serpent-stimulate-59fb860ef429@spud Signed-off-by: Mark Brown --- drivers/spi/spi-microchip-core-qspi.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/spi/spi-microchip-core-qspi.c b/drivers/spi/spi-microchip-core-qspi.c index 70215a407b5a..4dee0fea1df8 100644 --- a/drivers/spi/spi-microchip-core-qspi.c +++ b/drivers/spi/spi-microchip-core-qspi.c @@ -197,7 +197,7 @@ static int mchp_coreqspi_setup(struct spi_device *spi) return 0; } -static inline void mchp_coreqspi_read_op(struct mchp_coreqspi *qspi) +static void mchp_coreqspi_read_op(struct mchp_coreqspi *qspi) { u32 control, data; @@ -233,7 +233,7 @@ static inline void mchp_coreqspi_read_op(struct mchp_coreqspi *qspi) } } -static inline void mchp_coreqspi_write_op(struct mchp_coreqspi *qspi) +static void mchp_coreqspi_write_op(struct mchp_coreqspi *qspi) { u32 control, data; @@ -261,7 +261,7 @@ static inline void mchp_coreqspi_write_op(struct mchp_coreqspi *qspi) } } -static inline void mchp_coreqspi_write_read_op(struct mchp_coreqspi *qspi) +static void mchp_coreqspi_write_read_op(struct mchp_coreqspi *qspi) { u32 control, data; @@ -419,7 +419,7 @@ static int mchp_coreqspi_setup_clock(struct mchp_coreqspi *qspi, struct spi_devi return 0; } -static inline void mchp_coreqspi_config_op(struct mchp_coreqspi *qspi, const struct spi_mem_op *op) +static void mchp_coreqspi_config_op(struct mchp_coreqspi *qspi, const struct spi_mem_op *op) { u32 idle_cycles = 0; int total_bytes, cmd_bytes, frames, ctrl; From d581fc99d3b958cb6e363104e9aab57f36aee6f3 Mon Sep 17 00:00:00 2001 From: David Carlier Date: Thu, 23 Apr 2026 13:56:47 +0100 Subject: [PATCH 082/333] mm/memfd_luo: reject memfds whose page count exceeds UINT_MAX memfd_luo_preserve_folios() declares max_folios as unsigned int and computes it from the inode size, then passes it to memfd_pin_folios() which itself caps max_folios at unsigned int. For files whose base-page count exceeds UINT_MAX (larger than 16 TiB with 4 KiB pages), the assignment truncates silently: only a prefix of the file gets pinned and preserved, while memfd_luo_preserve() still records the full inode size in ser->size. On retrieve the inode is restored to the full size but only the preserved prefix repopulates the page cache, so the tail comes back as holes and user data is silently lost across the live update. Reject such files at preserve time with -EFBIG rather than chunk the pin loop, which would also require enlarging the preserved folios array well beyond what is practical. Fixes: b3749f174d68 ("mm: memfd_luo: allow preserving memfd") Signed-off-by: David Carlier Reviewed-by: Pasha Tatashin Reviewed-by: Pratyush Yadav Link: https://patch.msgid.link/20260423125648.152113-1-devnexen@gmail.com Signed-off-by: Pasha Tatashin --- mm/memfd_luo.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/mm/memfd_luo.c b/mm/memfd_luo.c index 35d1247281e0..94ae113f68f6 100644 --- a/mm/memfd_luo.c +++ b/mm/memfd_luo.c @@ -259,7 +259,7 @@ static int memfd_luo_preserve(struct liveupdate_file_op_args *args) struct inode *inode = file_inode(args->file); struct memfd_luo_folio_ser *folios_ser; struct memfd_luo_ser *ser; - u64 nr_folios; + u64 nr_folios, inode_size; int err = 0, seals; inode_lock(inode); @@ -285,7 +285,18 @@ static int memfd_luo_preserve(struct liveupdate_file_op_args *args) } ser->pos = args->file->f_pos; - ser->size = i_size_read(inode); + inode_size = i_size_read(inode); + + /* + * memfd_pin_folios() caps at UINT_MAX folios; refuse larger + * files to avoid silently preserving only a prefix. + */ + if (DIV_ROUND_UP_ULL(inode_size, PAGE_SIZE) > UINT_MAX) { + err = -EFBIG; + goto err_free_ser; + } + + ser->size = inode_size; ser->seals = seals; err = memfd_luo_preserve_folios(args->file, &ser->folios, From 7b0b68b2b95606e65594958686833e53423f58f2 Mon Sep 17 00:00:00 2001 From: David Carlier Date: Thu, 23 Apr 2026 13:56:48 +0100 Subject: [PATCH 083/333] mm/memfd_luo: document preservation of file seals Commit 8a552d68a86e ("mm: memfd_luo: preserve file seals") started preserving file seals across live update and restoring them via memfd_add_seals() on retrieve, but the DOC header was not updated and still listed seals under "Non-Preserved Properties" as being unsealed on restore. Move the Seals entry to the "Preserved Properties" section and describe the actual behavior, including the MEMFD_LUO_ALL_SEALS restriction that both preserve and retrieve enforce. Fixes: 8a552d68a86e ("mm: memfd_luo: preserve file seals") Signed-off-by: David Carlier Reviewed-by: Pratyush Yadav Reviewed-by: Pasha Tatashin Link: https://patch.msgid.link/20260423125648.152113-2-devnexen@gmail.com Signed-off-by: Pasha Tatashin --- mm/memfd_luo.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/mm/memfd_luo.c b/mm/memfd_luo.c index 94ae113f68f6..59de210bee5f 100644 --- a/mm/memfd_luo.c +++ b/mm/memfd_luo.c @@ -50,6 +50,11 @@ * memfds are always opened with ``O_RDWR`` and ``O_LARGEFILE``. This property * is maintained. * + * Seals + * File seals set on the memfd are preserved and re-applied on restore. + * Only seals known to this LUO version (see ``MEMFD_LUO_ALL_SEALS``) may + * be present; preservation fails with ``-EOPNOTSUPP`` otherwise. + * * Non-Preserved Properties * ======================== * @@ -61,10 +66,6 @@ * A memfd can be created with the ``MFD_CLOEXEC`` flag that sets the * ``FD_CLOEXEC`` on the file. This flag is not preserved and must be set * again after restore via ``fcntl()``. - * - * Seals - * File seals are not preserved. The file is unsealed on restore and if - * needed, must be sealed again via ``fcntl()``. */ #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt From 2a46a9356ba7b1bdd741c8b41e5374edcd960557 Mon Sep 17 00:00:00 2001 From: "Kory Maincent (TI)" Date: Tue, 28 Apr 2026 11:04:56 +0200 Subject: [PATCH 084/333] drm/bridge: tda998x: Use __be32 for audio port OF property pointer of_get_property() returns a pointer to big-endian (__be32) data, but port_data in tda998x_get_audio_ports() was declared as const u32 *, causing a sparse endianness type mismatch warning. Fix the declaration to use const __be32 *. Fixes: 7e567624dc5a4 ("drm/i2c: tda998x: Register ASoC hdmi-codec and add audio DT binding") Cc: stable@vger.kernel.org Signed-off-by: Kory Maincent (TI) Reviewed-by: Russell King (Oracle) Link: https://patch.msgid.link/20260428090457.121894-1-kory.maincent@bootlin.com Signed-off-by: Luca Ceresoli --- drivers/gpu/drm/bridge/tda998x_drv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/bridge/tda998x_drv.c b/drivers/gpu/drm/bridge/tda998x_drv.c index d9b388165de1..779b976f601c 100644 --- a/drivers/gpu/drm/bridge/tda998x_drv.c +++ b/drivers/gpu/drm/bridge/tda998x_drv.c @@ -1762,7 +1762,7 @@ static const struct drm_bridge_funcs tda998x_bridge_funcs = { static int tda998x_get_audio_ports(struct tda998x_priv *priv, struct device_node *np) { - const u32 *port_data; + const __be32 *port_data; u32 size; int i; From b5d0ad616ca8dd8c7b6b24dc13012e342278a085 Mon Sep 17 00:00:00 2001 From: "Kory Maincent (TI)" Date: Fri, 17 Apr 2026 17:54:45 +0200 Subject: [PATCH 085/333] drm/bridge: tda998x: Return NULL instead of 0 in tda998x_edid_read() tda998x_edid_read() returns a const struct drm_edid pointer, but when tda998x_edid_delay_wait() fails (process killed while waiting for the HPD timeout), the integer literal 0 is returned instead of NULL, triggering a sparse warning: "Using plain integer as NULL pointer" Replace 0 with NULL to fix the sparse warning. Fixes: c76a8be4feec ("drm/bridge: tda998x: Add support for DRM_BRIDGE_ATTACH_NO_CONNECTOR") Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202604172257.Imo6GOH9-lkp@intel.com/ Signed-off-by: Kory Maincent (TI) Reviewed-by: Luca Ceresoli Link: https://patch.msgid.link/20260417155446.1068893-1-kory.maincent@bootlin.com Signed-off-by: Luca Ceresoli --- drivers/gpu/drm/bridge/tda998x_drv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/bridge/tda998x_drv.c b/drivers/gpu/drm/bridge/tda998x_drv.c index 779b976f601c..6c427bc75896 100644 --- a/drivers/gpu/drm/bridge/tda998x_drv.c +++ b/drivers/gpu/drm/bridge/tda998x_drv.c @@ -1293,7 +1293,7 @@ static const struct drm_edid *tda998x_edid_read(struct tda998x_priv *priv, * can't handle signals gracefully. */ if (tda998x_edid_delay_wait(priv)) - return 0; + return NULL; if (priv->rev == TDA19988) reg_clear(priv, REG_TX4, TX4_PD_RAM); From 17e4c68ff35090d8cb743e3c82c09f92fda1ebda Mon Sep 17 00:00:00 2001 From: David Gow Date: Sat, 25 Apr 2026 11:41:53 +0800 Subject: [PATCH 086/333] kunit: config: Enable KUNIT_DEBUGFS by default The KUNIT_DEBUGFS option is currently enabled based on the value of KUNIT_ALL_TESTS, but it really doesn't have anything to do with the set of enabled tests, so just enable it by default anyway. In particular, this shouldn't be only visible if KUNIT_ALL_TESTS is set, which is quite confusing. Link: https://lore.kernel.org/r/20260425034155.53913-1-david@davidgow.net Fixes: beaed42c427d ("kunit: default KUNIT_* fragments to KUNIT_ALL_TESTS") Signed-off-by: David Gow Signed-off-by: Shuah Khan --- lib/kunit/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/lib/kunit/Kconfig b/lib/kunit/Kconfig index 498cc51e493d..f80ca3aeedb0 100644 --- a/lib/kunit/Kconfig +++ b/lib/kunit/Kconfig @@ -16,8 +16,8 @@ menuconfig KUNIT if KUNIT config KUNIT_DEBUGFS - bool "KUnit - Enable /sys/kernel/debug/kunit debugfs representation" if !KUNIT_ALL_TESTS - default KUNIT_ALL_TESTS + bool "KUnit - Enable /sys/kernel/debug/kunit debugfs representation" + default y help Enable debugfs representation for kunit. Currently this consists of /sys/kernel/debug/kunit//results files for each From 8f80b5b227ef9ea422080487715c841856339aed Mon Sep 17 00:00:00 2001 From: David Gow Date: Sat, 25 Apr 2026 11:41:54 +0800 Subject: [PATCH 087/333] kunit: config: KUNIT_DEBUGFS should depend on DEBUG_FS CONFIG_KUNIT_DEBUGFS is totally useless without debugfs, so it should depend on CONFIG_DEBUG_FS. Link: https://lore.kernel.org/r/20260425034155.53913-2-david@davidgow.net Fixes: e2219db280e3 ("kunit: add debugfs /sys/kernel/debug/kunit//results display") Signed-off-by: David Gow Signed-off-by: Shuah Khan --- lib/kunit/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/lib/kunit/Kconfig b/lib/kunit/Kconfig index f80ca3aeedb0..94ff8e4089bf 100644 --- a/lib/kunit/Kconfig +++ b/lib/kunit/Kconfig @@ -17,6 +17,7 @@ if KUNIT config KUNIT_DEBUGFS bool "KUnit - Enable /sys/kernel/debug/kunit debugfs representation" + depends on DEBUG_FS default y help Enable debugfs representation for kunit. Currently this consists From 360f61bc29085d65a24dbaaf707c802553a239fd Mon Sep 17 00:00:00 2001 From: Peter Rosin Date: Thu, 30 Apr 2026 06:09:58 +0200 Subject: [PATCH 088/333] MAINTAINERS: Update mail for Peter Rosin I'm resigning from my position at Axentia. Signed-off-by: Peter Rosin Signed-off-by: Wolfram Sang --- .mailmap | 1 + MAINTAINERS | 24 +++++++++++------------- 2 files changed, 12 insertions(+), 13 deletions(-) diff --git a/.mailmap b/.mailmap index b78aa092b4bb..eec4a740f7ca 100644 --- a/.mailmap +++ b/.mailmap @@ -682,6 +682,7 @@ Peter A Jonsson Peter Hilber Peter Oruba Peter Oruba +Peter Rosin Pierre-Louis Bossart Pratyush Anand Pratyush Yadav diff --git a/MAINTAINERS b/MAINTAINERS index 882214b0e7db..1b0c453bca46 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4299,18 +4299,16 @@ F: Documentation/devicetree/bindings/leds/backlight/awinic,aw99706.yaml F: drivers/video/backlight/aw99706.c AXENTIA ARM DEVICES -M: Peter Rosin L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) -S: Maintained +S: Orphan F: arch/arm/boot/dts/microchip/at91-linea.dtsi F: arch/arm/boot/dts/microchip/at91-natte.dtsi F: arch/arm/boot/dts/microchip/at91-nattis-2-natte-2.dts F: arch/arm/boot/dts/microchip/at91-tse850-3.dts AXENTIA ASOC DRIVERS -M: Peter Rosin L: linux-sound@vger.kernel.org -S: Maintained +S: Orphan F: Documentation/devicetree/bindings/sound/axentia,* F: sound/soc/atmel/tse850-pcm5142.c @@ -12046,7 +12044,7 @@ F: Documentation/i2c/busses/i2c-nvidia-gpu.rst F: drivers/i2c/busses/i2c-nvidia-gpu.c I2C MUXES -M: Peter Rosin +M: Peter Rosin L: linux-i2c@vger.kernel.org S: Maintained F: Documentation/devicetree/bindings/i2c/i2c-arb* @@ -12447,7 +12445,7 @@ F: drivers/iio/industrialio-backend.c F: include/linux/iio/backend.h IIO DIGITAL POTENTIOMETER DAC -M: Peter Rosin +M: Peter Rosin L: linux-iio@vger.kernel.org S: Maintained F: Documentation/ABI/testing/sysfs-bus-iio-dac-dpot-dac @@ -12455,7 +12453,7 @@ F: Documentation/devicetree/bindings/iio/dac/dpot-dac.yaml F: drivers/iio/dac/dpot-dac.c IIO ENVELOPE DETECTOR -M: Peter Rosin +M: Peter Rosin L: linux-iio@vger.kernel.org S: Maintained F: Documentation/ABI/testing/sysfs-bus-iio-adc-envelope-detector @@ -12471,7 +12469,7 @@ F: include/linux/iio/iio-gts-helper.h F: drivers/iio/test/iio-test-gts.c IIO MULTIPLEXER -M: Peter Rosin +M: Peter Rosin L: linux-iio@vger.kernel.org S: Maintained F: Documentation/devicetree/bindings/iio/multiplexer/io-channel-mux.yaml @@ -12502,7 +12500,7 @@ F: include/linux/iio/ F: tools/iio/ IIO UNIT CONVERTER -M: Peter Rosin +M: Peter Rosin L: linux-iio@vger.kernel.org S: Maintained F: Documentation/devicetree/bindings/iio/afe/current-sense-amplifier.yaml @@ -15718,7 +15716,7 @@ F: Documentation/devicetree/bindings/media/i2c/maxim,max96717.yaml F: drivers/media/i2c/max96717.c MAX9860 MONO AUDIO VOICE CODEC DRIVER -M: Peter Rosin +M: Peter Rosin L: linux-sound@vger.kernel.org S: Maintained F: Documentation/devicetree/bindings/sound/max9860.txt @@ -15933,7 +15931,7 @@ F: Documentation/devicetree/bindings/net/can/microchip,mcp251xfd.yaml F: drivers/net/can/spi/mcp251xfd/ MCP4018 AND MCP4531 MICROCHIP DIGITAL POTENTIOMETER DRIVERS -M: Peter Rosin +M: Peter Rosin L: linux-iio@vger.kernel.org S: Maintained F: Documentation/ABI/testing/sysfs-bus-iio-potentiometer-mcp4531 @@ -18238,7 +18236,7 @@ F: include/linux/mmc/ F: include/uapi/linux/mmc/ MULTIPLEXER SUBSYSTEM -M: Peter Rosin +M: Peter Rosin S: Odd Fixes F: Documentation/ABI/testing/sysfs-class-mux* F: Documentation/devicetree/bindings/mux/ @@ -19347,7 +19345,7 @@ F: include/dt-bindings/display/tda998x.h K: "nxp,tda998x" NXP TFA9879 DRIVER -M: Peter Rosin +M: Peter Rosin L: linux-sound@vger.kernel.org S: Maintained F: Documentation/devicetree/bindings/sound/trivial-codec.yaml From 84ae1840260fece9b6b70d3872b79384bbe5a90b Mon Sep 17 00:00:00 2001 From: Osama Abdelkader Date: Thu, 23 Apr 2026 22:06:19 +0200 Subject: [PATCH 089/333] drm/sti: remove bridge when sti_hda component_add fails MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use devm_drm_bridge_add() so the bridge is released if probe fails after registration, and drop the manual drm_bridge_remove() in remove(). Check the return value of devm_drm_bridge_add(). Signed-off-by: Osama Abdelkader Fixes: d28726efc637 ("drm/sti: hda: add bridge before attaching") Cc: stable@vger.kernel.org Reviewed-by: Luca Ceresoli Acked-by: Raphaël Gallais-Pou Link: https://patch.msgid.link/20260423200622.325076-1-osama.abdelkader@gmail.com Signed-off-by: Raphael Gallais-Pou --- drivers/gpu/drm/sti/sti_hda.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/sti/sti_hda.c b/drivers/gpu/drm/sti/sti_hda.c index b7397827889c..360a88ca8f0c 100644 --- a/drivers/gpu/drm/sti/sti_hda.c +++ b/drivers/gpu/drm/sti/sti_hda.c @@ -741,6 +741,7 @@ static int sti_hda_probe(struct platform_device *pdev) struct device *dev = &pdev->dev; struct sti_hda *hda; struct resource *res; + int ret; DRM_INFO("%s\n", __func__); @@ -779,7 +780,9 @@ static int sti_hda_probe(struct platform_device *pdev) return PTR_ERR(hda->clk_hddac); } - drm_bridge_add(&hda->bridge); + ret = devm_drm_bridge_add(dev, &hda->bridge); + if (ret) + return ret; platform_set_drvdata(pdev, hda); @@ -788,10 +791,7 @@ static int sti_hda_probe(struct platform_device *pdev) static void sti_hda_remove(struct platform_device *pdev) { - struct sti_hda *hda = platform_get_drvdata(pdev); - component_del(&pdev->dev, &sti_hda_ops); - drm_bridge_remove(&hda->bridge); } static const struct of_device_id hda_of_match[] = { From aab3d205a086233c612fee86009265451793e0c2 Mon Sep 17 00:00:00 2001 From: Juha-Pekka Heikkila Date: Mon, 27 Apr 2026 19:57:15 +0300 Subject: [PATCH 090/333] drm/i915/display: enable ccs modifiers on dg2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since Xe driver aux ccs enablement dg2 ccs modifiers have been disabled on i915 driver. Here allow dg2 to use ccs again for framebuffers. Fixes: 6a99e91a6ca8 ("drm/i915/display: Detect AuxCCS support via display parent interface") Signed-off-by: Juha-Pekka Heikkila Reviewed-by: Ville Syrjälä Signed-off-by: Mika Kahola Link: https://patch.msgid.link/20260427165715.864721-1-juhapekka.heikkila@gmail.com (cherry picked from commit aee13ba1448213975f36942ba5d1ce693eb5c002) Signed-off-by: Tvrtko Ursulin --- drivers/gpu/drm/i915/i915_driver.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c index 385a634c3ed0..d9be7a5a239c 100644 --- a/drivers/gpu/drm/i915/i915_driver.c +++ b/drivers/gpu/drm/i915/i915_driver.c @@ -750,9 +750,8 @@ static bool has_auxccs(struct drm_device *drm) { struct drm_i915_private *i915 = to_i915(drm); - return IS_GRAPHICS_VER(i915, 9, 12) || - IS_ALDERLAKE_P(i915) || - IS_METEORLAKE(i915); + return IS_GRAPHICS_VER(i915, 9, 12) && + !HAS_FLAT_CCS(i915); } static bool has_fenced_regions(struct drm_device *drm) From 212ec34e4e726e8cd4af7bea4740db24de8a9dab Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Mon, 4 May 2026 08:34:32 -0600 Subject: [PATCH 091/333] block: only read from sqe on initial invocation of blkdev_uring_cmd() This passthrough helper currently only supports discards. Part of that command is the start and length, which is read from the SQE. It does so on every invocation, where it really should just make it stable on the first invocation. This avoids needing to copy the SQE upfront, as we only really need those two 8b values stored in our per-req payload. Cc: stable@vger.kernel.org # 6.17+ Signed-off-by: Jens Axboe --- block/ioctl.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/block/ioctl.c b/block/ioctl.c index fc3be0549aa7..ab2c9ed79946 100644 --- a/block/ioctl.c +++ b/block/ioctl.c @@ -857,6 +857,8 @@ long compat_blkdev_ioctl(struct file *file, unsigned cmd, unsigned long arg) #endif struct blk_iou_cmd { + u64 start; + u64 len; int res; bool nowait; }; @@ -946,23 +948,27 @@ int blkdev_uring_cmd(struct io_uring_cmd *cmd, unsigned int issue_flags) { struct block_device *bdev = I_BDEV(cmd->file->f_mapping->host); struct blk_iou_cmd *bic = io_uring_cmd_to_pdu(cmd, struct blk_iou_cmd); - const struct io_uring_sqe *sqe = cmd->sqe; u32 cmd_op = cmd->cmd_op; - uint64_t start, len; - if (unlikely(sqe->ioprio || sqe->__pad1 || sqe->len || - sqe->rw_flags || sqe->file_index)) - return -EINVAL; + /* Read what we need from the SQE on the first issue */ + if (!(issue_flags & IORING_URING_CMD_REISSUE)) { + const struct io_uring_sqe *sqe = cmd->sqe; + + if (unlikely(sqe->ioprio || sqe->__pad1 || sqe->len || + sqe->rw_flags || sqe->file_index)) + return -EINVAL; + + bic->start = READ_ONCE(sqe->addr); + bic->len = READ_ONCE(sqe->addr3); + } bic->res = 0; bic->nowait = issue_flags & IO_URING_F_NONBLOCK; - start = READ_ONCE(sqe->addr); - len = READ_ONCE(sqe->addr3); - switch (cmd_op) { case BLOCK_URING_CMD_DISCARD: - return blkdev_cmd_discard(cmd, bdev, start, len, bic->nowait); + return blkdev_cmd_discard(cmd, bdev, bic->start, bic->len, + bic->nowait); } return -EINVAL; } From 5f8719945244dd65b5fa06195f4600db62581610 Mon Sep 17 00:00:00 2001 From: Juergen Gross Date: Tue, 5 May 2026 10:06:53 +0200 Subject: [PATCH 092/333] x86/xen: Fix a potential problem in xen_e820_resolve_conflicts() When fixing a conflict in xen_e820_resolve_conflicts(), the loop over the E820 map entries needs to be restarted, as the E820 map will have been modified by the fix. Otherwise entries might be skipped by accident. Fixes: be35d91c8880 ("xen: tolerate ACPI NVS memory overlapping with Xen allocated memory") Signed-off-by: Juergen Gross Signed-off-by: Ingo Molnar Cc: xen-devel@lists.xenproject.org Link: https://patch.msgid.link/20260505080653.197775-1-jgross@suse.com --- arch/x86/xen/setup.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/arch/x86/xen/setup.c b/arch/x86/xen/setup.c index ac8021c3a997..bb95a05259b8 100644 --- a/arch/x86/xen/setup.c +++ b/arch/x86/xen/setup.c @@ -695,17 +695,22 @@ static void __init xen_e820_resolve_conflicts(phys_addr_t start, return; end = start + size; - entry = xen_e820_table.entries; + mapcnt = 0; - for (mapcnt = 0; mapcnt < xen_e820_table.nr_entries; mapcnt++) { + while (mapcnt < xen_e820_table.nr_entries) { + entry = xen_e820_table.entries + mapcnt; if (entry->addr >= end) return; if (entry->addr + entry->size > start && - entry->type == E820_TYPE_NVS) + entry->type == E820_TYPE_NVS) { xen_e820_swap_entry_with_ram(entry); + /* E820 map has been changed, restart loop! */ + mapcnt = 0; + continue; + } - entry++; + mapcnt++; } } From 3780c41460a9ad6d5d4c09a416765c6cc285033b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ma=C3=ADra=20Canal?= Date: Thu, 2 Apr 2026 16:32:35 -0300 Subject: [PATCH 093/333] drm/etnaviv: Fix armed job not being pushed to the DRM scheduler MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When xa_alloc_cyclic() failed in etnaviv_sched_push_job(), the error path skipped drm_sched_entity_push_job(). This is a violation of the DRM scheduler contract, as once a job has been armed with drm_sched_job_arm(), it must be pushed with drm_sched_entity_push_job(). From the DRM scheduler documentation, """ drm_sched_job_arm() is a point of no return since it initializes the fences and their sequence number etc. Once that function has been called, you *must* submit it with drm_sched_entity_push_job() and cannot simply abort it by calling drm_sched_job_cleanup(). """ Fix this by splitting the fence ID allocation into two phases: first, alloc an xarray slot before arming the job (which can fail), then fill in the actual fence with xa_store() after arming. This way, allocation failures are handled before the job is armed, and once armed, the job is always pushed to the scheduler. This also fixes a double call to drm_sched_job_cleanup(), as both etnaviv_sched_push_job() and its caller would call it on failure. Fixes: 764be12345c3 ("drm/etnaviv: convert user fence tracking to XArray") Signed-off-by: Maíra Canal Link: https://patch.msgid.link/20260402193424.2023318-1-mcanal@igalia.com Signed-off-by: Christian Gmeiner --- drivers/gpu/drm/etnaviv/etnaviv_sched.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/etnaviv/etnaviv_sched.c b/drivers/gpu/drm/etnaviv/etnaviv_sched.c index df4232d7e135..3cc50d697c89 100644 --- a/drivers/gpu/drm/etnaviv/etnaviv_sched.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_sched.c @@ -116,16 +116,18 @@ int etnaviv_sched_push_job(struct etnaviv_gem_submit *submit) */ mutex_lock(&gpu->sched_lock); + ret = xa_alloc_cyclic(&gpu->user_fences, &submit->out_fence_id, + NULL, xa_limit_32b, &gpu->next_user_fence, + GFP_KERNEL); + if (ret < 0) + goto out_unlock; + drm_sched_job_arm(&submit->sched_job); submit->out_fence = dma_fence_get(&submit->sched_job.s_fence->finished); - ret = xa_alloc_cyclic(&gpu->user_fences, &submit->out_fence_id, - submit->out_fence, xa_limit_32b, - &gpu->next_user_fence, GFP_KERNEL); - if (ret < 0) { - drm_sched_job_cleanup(&submit->sched_job); - goto out_unlock; - } + + xa_store(&gpu->user_fences, submit->out_fence_id, + submit->out_fence, GFP_KERNEL); /* the scheduler holds on to the job now */ kref_get(&submit->refcount); From c69df06e4e26e50611190ce04eab92c5cc261b61 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Thu, 26 Mar 2026 12:28:21 +0100 Subject: [PATCH 094/333] perf/core: Fix deadlock in perf_mmap() failure path Ian noted that commit 77de62ad3de3 ("perf/core: Fix refcount bug and potential UAF in perf_mmap") would cause a deadlock due to event->mmap_mutex recursion. This happens because we're now calling perf_mmap_close() under mmap_mutex, while that function itself can also take mmap_mutex. Solve this by noting that perf_mmap_close() is far more complicated than we need at this particular point, since it deals with scenarios that cannot happen in this particular case. Replace the call to perf_mmap_close() with a very narrow undo for the case of first-exposure. If this is not the first mmap(), there is no race and it is fine to drop the lock and call perf_mmap_close() to handle to more complicated scenarios. Note: move the rb->mmap_user (namespace) handling into the rb init/free code such that it does not complicate the mmap handling. Fixes: 77de62ad3de3 ("perf/core: Fix refcount bug and potential UAF in perf_mmap") Reported-by: Ian Rogers Closes: https://patch.msgid.link/CAP-5%3DfVJyVMZw%3DDqP53Kxg58nUmJ_0bxoaeOKAbC03BVc11HaA%40mail.gmail.com Signed-off-by: Peter Zijlstra (Intel) Link: https://patch.msgid.link/20260326112821.GK3738786@noisy.programming.kicks-ass.net --- kernel/events/core.c | 70 +++++++++++++++++++++++++++++-------- kernel/events/internal.h | 1 + kernel/events/ring_buffer.c | 2 ++ 3 files changed, 58 insertions(+), 15 deletions(-) diff --git a/kernel/events/core.c b/kernel/events/core.c index 6d1f8bad7e1c..7935d5663944 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -7006,6 +7006,7 @@ static void perf_mmap_open(struct vm_area_struct *vma) } static void perf_pmu_output_stop(struct perf_event *event); +static void perf_mmap_unaccount(struct vm_area_struct *vma, struct perf_buffer *rb); /* * A buffer can be mmap()ed multiple times; either directly through the same @@ -7021,8 +7022,6 @@ static void perf_mmap_close(struct vm_area_struct *vma) mapped_f unmapped = get_mapped(event, event_unmapped); struct perf_buffer *rb = ring_buffer_get(event); struct user_struct *mmap_user = rb->mmap_user; - int mmap_locked = rb->mmap_locked; - unsigned long size = perf_data_size(rb); bool detach_rest = false; /* FIXIES vs perf_pmu_unregister() */ @@ -7117,11 +7116,7 @@ static void perf_mmap_close(struct vm_area_struct *vma) * Aside from that, this buffer is 'fully' detached and unmapped, * undo the VM accounting. */ - - atomic_long_sub((size >> PAGE_SHIFT) + 1 - mmap_locked, - &mmap_user->locked_vm); - atomic64_sub(mmap_locked, &vma->vm_mm->pinned_vm); - free_uid(mmap_user); + perf_mmap_unaccount(vma, rb); out_put: ring_buffer_put(rb); /* could be last */ @@ -7261,6 +7256,15 @@ static void perf_mmap_account(struct vm_area_struct *vma, long user_extra, long atomic64_add(extra, &vma->vm_mm->pinned_vm); } +static void perf_mmap_unaccount(struct vm_area_struct *vma, struct perf_buffer *rb) +{ + struct user_struct *user = rb->mmap_user; + + atomic_long_sub((perf_data_size(rb) >> PAGE_SHIFT) + 1 - rb->mmap_locked, + &user->locked_vm); + atomic64_sub(rb->mmap_locked, &vma->vm_mm->pinned_vm); +} + static int perf_mmap_rb(struct vm_area_struct *vma, struct perf_event *event, unsigned long nr_pages) { @@ -7323,8 +7327,6 @@ static int perf_mmap_rb(struct vm_area_struct *vma, struct perf_event *event, if (!rb) return -ENOMEM; - refcount_set(&rb->mmap_count, 1); - rb->mmap_user = get_current_user(); rb->mmap_locked = extra; ring_buffer_attach(event, rb); @@ -7474,16 +7476,54 @@ static int perf_mmap(struct file *file, struct vm_area_struct *vma) mapped(event, vma->vm_mm); /* - * Try to map it into the page table. On fail, invoke - * perf_mmap_close() to undo the above, as the callsite expects - * full cleanup in this case and therefore does not invoke - * vmops::close(). + * Try to map it into the page table. On fail undo the above, + * as the callsite expects full cleanup in this case and + * therefore does not invoke vmops::close(). */ ret = map_range(event->rb, vma); - if (ret) - perf_mmap_close(vma); + if (likely(!ret)) + return 0; + + /* Error path */ + + /* + * If this is the first mmap(), then event->mmap_count should + * be stable at 1. It is only modified by: + * perf_mmap_{open,close}() and perf_mmap(). + * + * The former are not possible because this mmap() hasn't been + * successful yet, and the latter is serialized by + * event->mmap_mutex which we still hold (note that mmap_lock + * is not strictly sufficient here, because the event fd can + * be passed to another process through trivial means like + * fork(), leading to concurrent mmap() from different mm). + * + * Make sure to remove event->rb before releasing + * event->mmap_mutex, such that any concurrent mmap() will not + * attempt use this failed buffer. + */ + if (refcount_read(&event->mmap_count) == 1) { + /* + * Minimal perf_mmap_close(); there can't be AUX or + * other events on account of this being the first. + */ + mapped = get_mapped(event, event_unmapped); + if (mapped) + mapped(event, vma->vm_mm); + perf_mmap_unaccount(vma, event->rb); + ring_buffer_attach(event, NULL); /* drops last rb->refcount */ + refcount_set(&event->mmap_count, 0); + return ret; + } + + /* + * Otherwise this is an already existing buffer, and there is + * no race vs first exposure, so fall-through and call + * perf_mmap_close(). + */ } + perf_mmap_close(vma); return ret; } diff --git a/kernel/events/internal.h b/kernel/events/internal.h index d9cc57083091..c03c4f2eea57 100644 --- a/kernel/events/internal.h +++ b/kernel/events/internal.h @@ -67,6 +67,7 @@ static inline void rb_free_rcu(struct rcu_head *rcu_head) struct perf_buffer *rb; rb = container_of(rcu_head, struct perf_buffer, rcu_head); + free_uid(rb->mmap_user); rb_free(rb); } diff --git a/kernel/events/ring_buffer.c b/kernel/events/ring_buffer.c index 3e7de2661417..9fe92161715e 100644 --- a/kernel/events/ring_buffer.c +++ b/kernel/events/ring_buffer.c @@ -340,6 +340,8 @@ ring_buffer_init(struct perf_buffer *rb, long watermark, int flags) rb->paused = 1; mutex_init(&rb->aux_mutex); + rb->mmap_user = get_current_user(); + refcount_set(&rb->mmap_count, 1); } void perf_aux_output_flag(struct perf_output_handle *handle, u64 flags) From 5ad732a56be46aabf158c16aa0c095291727aaef Mon Sep 17 00:00:00 2001 From: Dapeng Mi Date: Thu, 30 Apr 2026 08:25:54 +0800 Subject: [PATCH 095/333] perf/x86/intel: Improve validation and configuration of ACR masks Currently there are several issues on the user space ACR mask validation and configuration. - The validation for user space ACR mask (attr.config2) is incomplete, e.g., the ACR mask could include the index which belongs to another ACR events group, but it's not validated. - An early return on an invalid ACR mask caused all subsequent ACR groups to be skipped. - The stale hardware ACR mask (hw.config1) is not cleared before setting new hardware ACR mask. The following changes address all of the above issues. - Figure out the event index group of an ACR group. Any bits in the user-space mask not present in the index group are now dropped. - Instead of an early return on invalid bits, drop only the invalid portions and continue iterating through all ACR events to ensure full configuration. - Explicitly clear the stale hardware ACR mask for each event prior to writing the new configuration. Besides, a non-leader event member of ACR group could be disabled in theory. This could cause bit-shifting errors in the acr_mask of remaining group members. But since ACR sampling requires all events to be active, this should not be a big concern in real use case. Add a "FIXME" comment to notice this risk. Fixes: ec980e4facef ("perf/x86/intel: Support auto counter reload") Signed-off-by: Dapeng Mi Signed-off-by: Peter Zijlstra (Intel) Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260430002558.712334-2-dapeng1.mi@linux.intel.com --- arch/x86/events/intel/core.c | 32 +++++++++++++++++++++++++------- 1 file changed, 25 insertions(+), 7 deletions(-) diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c index d9488ade0f8e..f8deb67b3c51 100644 --- a/arch/x86/events/intel/core.c +++ b/arch/x86/events/intel/core.c @@ -3332,23 +3332,41 @@ static void intel_pmu_enable_event(struct perf_event *event) static void intel_pmu_acr_late_setup(struct cpu_hw_events *cpuc) { struct perf_event *event, *leader; - int i, j, idx; + int i, j, k, bit, idx; + /* + * FIXME: ACR mask parsing relies on cpuc->event_list[] (active events only). + * Disabling an ACR event causes bit-shifting errors in the acr_mask of + * remaining group members. As ACR sampling requires all events to be active, + * this limitation is acceptable for now. Revisit if independent event toggling + * is required. + */ for (i = 0; i < cpuc->n_events; i++) { leader = cpuc->event_list[i]; if (!is_acr_event_group(leader)) continue; - /* The ACR events must be contiguous. */ + /* Find the last event of the ACR group. */ for (j = i; j < cpuc->n_events; j++) { event = cpuc->event_list[j]; if (event->group_leader != leader->group_leader) break; - for_each_set_bit(idx, (unsigned long *)&event->attr.config2, X86_PMC_IDX_MAX) { - if (i + idx >= cpuc->n_events || - !is_acr_event_group(cpuc->event_list[i + idx])) - return; - __set_bit(cpuc->assign[i + idx], (unsigned long *)&event->hw.config1); + } + + /* + * Translate the user-space ACR mask (attr.config2) into the physical + * counter bitmask (hw.config1) for each ACR event in the group. + * NOTE: ACR event contiguity is guaranteed by intel_pmu_hw_config(). + */ + for (k = i; k < j; k++) { + event = cpuc->event_list[k]; + event->hw.config1 = 0; + for_each_set_bit(bit, (unsigned long *)&event->attr.config2, X86_PMC_IDX_MAX) { + idx = i + bit; + /* Event index of ACR group must locate in [i, j). */ + if (idx >= j || !is_acr_event_group(cpuc->event_list[idx])) + continue; + __set_bit(cpuc->assign[idx], (unsigned long *)&event->hw.config1); } } i = j - 1; From 8ba0b706a485b1e607594cf4210786d517ad1611 Mon Sep 17 00:00:00 2001 From: Dapeng Mi Date: Thu, 30 Apr 2026 08:25:55 +0800 Subject: [PATCH 096/333] perf/x86/intel: Always reprogram ACR events to prevent stale masks Members of an ACR group are logically linked via a bitmask of their hardware counter indices. If some members of the group are assigned new hardware counters during rescheduling, even events that keep their original counter index must be updated with a new mask. Without this, an event will continue to use a stale acr_mask that references the old indices of its group peers. Ensure all ACR events are reprogrammed during the scheduling path to maintain consistency across the group. Fixes: ec980e4facef ("perf/x86/intel: Support auto counter reload") Signed-off-by: Dapeng Mi Signed-off-by: Peter Zijlstra (Intel) Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260430002558.712334-3-dapeng1.mi@linux.intel.com --- arch/x86/events/core.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) diff --git a/arch/x86/events/core.c b/arch/x86/events/core.c index 810ab21ffd99..4b9e105309c6 100644 --- a/arch/x86/events/core.c +++ b/arch/x86/events/core.c @@ -1294,13 +1294,16 @@ int x86_perf_rdpmc_index(struct perf_event *event) return event->hw.event_base_rdpmc; } -static inline int match_prev_assignment(struct hw_perf_event *hwc, +static inline int match_prev_assignment(struct perf_event *event, struct cpu_hw_events *cpuc, int i) { + struct hw_perf_event *hwc = &event->hw; + return hwc->idx == cpuc->assign[i] && - hwc->last_cpu == smp_processor_id() && - hwc->last_tag == cpuc->tags[i]; + hwc->last_cpu == smp_processor_id() && + hwc->last_tag == cpuc->tags[i] && + !is_acr_event_group(event); } static void x86_pmu_start(struct perf_event *event, int flags); @@ -1346,7 +1349,7 @@ static void x86_pmu_enable(struct pmu *pmu) * - no other event has used the counter since */ if (hwc->idx == -1 || - match_prev_assignment(hwc, cpuc, i)) + match_prev_assignment(event, cpuc, i)) continue; /* @@ -1367,7 +1370,7 @@ static void x86_pmu_enable(struct pmu *pmu) event = cpuc->event_list[i]; hwc = &event->hw; - if (!match_prev_assignment(hwc, cpuc, i)) + if (!match_prev_assignment(event, cpuc, i)) x86_assign_hw_event(event, cpuc, i); else if (i < n_running) continue; From 1271aeccc307066315b2d3b0d5af2510e27018b5 Mon Sep 17 00:00:00 2001 From: Dapeng Mi Date: Thu, 30 Apr 2026 08:25:56 +0800 Subject: [PATCH 097/333] perf/x86/intel: Disable PMI for self-reloaded ACR events On platforms with Auto Counter Reload (ACR) support, such as NVL, a "NMI received for unknown reason 30" warning is observed when running multiple events in a group with ACR enabled: $ perf record -e '{instructions/period=20000,acr_mask=0x2/u,\ cycles/period=40000,acr_mask=0x3/u}' ./test The warning occurs because the Performance Monitoring Interrupt (PMI) is enabled for the self-reloaded event (the cycles event in this case). According to the Intel SDM, the overflow bit (IA32_PERF_GLOBAL_STATUS.PMCn_OVF) is never set for self-reloaded events. Since the bit is not set, the perf NMI handler cannot identify the source of the interrupt, leading to the "unknown reason" message. Furthermore, enabling PMI for self-reloaded events is unnecessary and can lead to extraneous records that pollute the user's requested data. Disable the interrupt bit for all events configured with ACR self-reload. Fixes: ec980e4facef ("perf/x86/intel: Support auto counter reload") Reported-by: Andi Kleen Signed-off-by: Dapeng Mi Signed-off-by: Peter Zijlstra (Intel) Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260430002558.712334-4-dapeng1.mi@linux.intel.com --- arch/x86/events/intel/core.c | 17 +++++++++++++---- arch/x86/events/perf_event.h | 10 ++++++++++ 2 files changed, 23 insertions(+), 4 deletions(-) diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c index f8deb67b3c51..ead6d95cec6a 100644 --- a/arch/x86/events/intel/core.c +++ b/arch/x86/events/intel/core.c @@ -3118,11 +3118,11 @@ static void intel_pmu_enable_fixed(struct perf_event *event) intel_set_masks(event, idx); /* - * Enable IRQ generation (0x8), if not PEBS, - * and enable ring-3 counting (0x2) and ring-0 counting (0x1) - * if requested: + * Enable IRQ generation (0x8), if not PEBS or self-reloaded + * ACR event, and enable ring-3 counting (0x2) and ring-0 + * counting (0x1) if requested: */ - if (!event->attr.precise_ip) + if (!event->attr.precise_ip && !is_acr_self_reload_event(event)) bits |= INTEL_FIXED_0_ENABLE_PMI; if (hwc->config & ARCH_PERFMON_EVENTSEL_USR) bits |= INTEL_FIXED_0_USER; @@ -3306,6 +3306,15 @@ static void intel_pmu_enable_event(struct perf_event *event) intel_set_masks(event, idx); static_call_cond(intel_pmu_enable_acr_event)(event); static_call_cond(intel_pmu_enable_event_ext)(event); + /* + * For self-reloaded ACR event, don't enable PMI since + * HW won't set overflow bit in GLOBAL_STATUS. Otherwise, + * the PMI would be recognized as a suspicious NMI. + */ + if (is_acr_self_reload_event(event)) + hwc->config &= ~ARCH_PERFMON_EVENTSEL_INT; + else if (!event->attr.precise_ip) + hwc->config |= ARCH_PERFMON_EVENTSEL_INT; __x86_pmu_enable_event(hwc, enable_mask); break; case INTEL_PMC_IDX_FIXED ... INTEL_PMC_IDX_FIXED_BTS - 1: diff --git a/arch/x86/events/perf_event.h b/arch/x86/events/perf_event.h index fad87d3c8b2c..524668dcf4cc 100644 --- a/arch/x86/events/perf_event.h +++ b/arch/x86/events/perf_event.h @@ -137,6 +137,16 @@ static inline bool is_acr_event_group(struct perf_event *event) return check_leader_group(event->group_leader, PERF_X86_EVENT_ACR); } +static inline bool is_acr_self_reload_event(struct perf_event *event) +{ + struct hw_perf_event *hwc = &event->hw; + + if (hwc->idx < 0) + return false; + + return test_bit(hwc->idx, (unsigned long *)&hwc->config1); +} + struct amd_nb { int nb_id; /* NorthBridge id */ int refcnt; /* reference count */ From aa4384bc8f4360167f3c3d5322121fe892289ea2 Mon Sep 17 00:00:00 2001 From: Dapeng Mi Date: Thu, 30 Apr 2026 08:25:57 +0800 Subject: [PATCH 098/333] perf/x86/intel: Enable auto counter reload for DMR MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Panther cove µarch starts to support auto counter reload (ACR), but the static_call intel_pmu_enable_acr_event() is not updated for the Panther Cove µarch used by DMR. It leads to the auto counter reload is not really enabled on DMR. Update static_call intel_pmu_enable_acr_event() in intel_pmu_init_pnc(). Fixes: d345b6bb8860 ("perf/x86/intel: Add core PMU support for DMR") Signed-off-by: Dapeng Mi Signed-off-by: Peter Zijlstra (Intel) Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260430002558.712334-5-dapeng1.mi@linux.intel.com --- arch/x86/events/intel/core.c | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c index ead6d95cec6a..dd1e3aa75ee9 100644 --- a/arch/x86/events/intel/core.c +++ b/arch/x86/events/intel/core.c @@ -7531,6 +7531,7 @@ static __always_inline void intel_pmu_init_pnc(struct pmu *pmu) hybrid(pmu, event_constraints) = intel_pnc_event_constraints; hybrid(pmu, pebs_constraints) = intel_pnc_pebs_event_constraints; hybrid(pmu, extra_regs) = intel_pnc_extra_regs; + static_call_update(intel_pmu_enable_acr_event, intel_pmu_enable_acr); } static __always_inline void intel_pmu_init_skt(struct pmu *pmu) From 50987d4e6c55929aa2d4d3976e74ccbae22d5017 Mon Sep 17 00:00:00 2001 From: Chen Ni Date: Fri, 27 Mar 2026 10:17:28 +0800 Subject: [PATCH 099/333] drm/panel: himax-hx83121a: Fix incorrect error check for devm_drm_panel_alloc() Check devm_drm_panel_alloc() return value for ERR_PTR instead of NULL. devm_drm_panel_alloc() returns an ERR_PTR on failure, never NULL. Using a NULL check skips the error path and may cause a NULL pointer dereference. Fixes: a7c61963b727 ("drm/panel: Add Himax HX83121A panel driver") Signed-off-by: Chen Ni Reviewed-by: Pengyu Luo Signed-off-by: Neil Armstrong Link: https://patch.msgid.link/20260327021728.647182-1-nichen@iscas.ac.cn --- drivers/gpu/drm/panel/panel-himax-hx83121a.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/panel/panel-himax-hx83121a.c b/drivers/gpu/drm/panel/panel-himax-hx83121a.c index ebe643ba4184..bed79aa06f46 100644 --- a/drivers/gpu/drm/panel/panel-himax-hx83121a.c +++ b/drivers/gpu/drm/panel/panel-himax-hx83121a.c @@ -596,8 +596,8 @@ static int himax_probe(struct mipi_dsi_device *dsi) ctx = devm_drm_panel_alloc(dev, struct himax, panel, &himax_panel_funcs, DRM_MODE_CONNECTOR_DSI); - if (!ctx) - return -ENOMEM; + if (IS_ERR(ctx)) + return PTR_ERR(ctx); ret = devm_regulator_bulk_get_const(&dsi->dev, ARRAY_SIZE(himax_supplies), From defab7b01e0848e004077d7d8dcc04d305ea1a27 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Mon, 13 Apr 2026 09:10:19 +0200 Subject: [PATCH 100/333] drm/panel: hx83121a: select DRM_DISPLAY_DSC_HELPER Like a number of other panel drivers, this newly merged driver needs DRM_DISPLAY_DSC_HELPER to be enabled: arm-linux-gnueabi-ld: drivers/gpu/drm/panel/panel-himax-hx83121a.o: in function `himax_prepare': panel-himax-hx83121a.c:(.text+0x1024): undefined reference to `drm_dsc_pps_payload_pack' Fixes: a7c61963b727 ("drm/panel: Add Himax HX83121A panel driver") Signed-off-by: Arnd Bergmann Reviewed-by: Neil Armstrong Reviewed-by: David Heidelberg Signed-off-by: Neil Armstrong Link: https://patch.msgid.link/20260413071043.3829868-1-arnd@kernel.org --- drivers/gpu/drm/panel/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/panel/Kconfig b/drivers/gpu/drm/panel/Kconfig index d6863b28ddc5..d592f4f4b939 100644 --- a/drivers/gpu/drm/panel/Kconfig +++ b/drivers/gpu/drm/panel/Kconfig @@ -208,6 +208,7 @@ config DRM_PANEL_HIMAX_HX83121A depends on OF depends on DRM_MIPI_DSI depends on BACKLIGHT_CLASS_DEVICE + select DRM_DISPLAY_DSC_HELPER select DRM_KMS_HELPER help Say Y here if you want to enable support for Himax HX83121A-based From c67e8787f6743101c90c7a9c4bb7cf6f1f739f83 Mon Sep 17 00:00:00 2001 From: Christian Van Date: Sat, 25 Apr 2026 01:39:48 -0400 Subject: [PATCH 101/333] drm/panel: feiyang-fy07024di26a30d: return display-on error mipi_dsi_dcs_set_display_on() returns an error code, but feiyang_enable() currently ignores it and always reports success. Return the DCS command result so callers can observe enable failures. Signed-off-by: Christian Van Reviewed-by: Neil Armstrong Signed-off-by: Neil Armstrong Link: https://patch.msgid.link/20260425053948.117714-1-cvan20191@gmail.com --- drivers/gpu/drm/panel/panel-feiyang-fy07024di26a30d.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/gpu/drm/panel/panel-feiyang-fy07024di26a30d.c b/drivers/gpu/drm/panel/panel-feiyang-fy07024di26a30d.c index 4f8d6d8c07e4..dbdb7e3cb7b6 100644 --- a/drivers/gpu/drm/panel/panel-feiyang-fy07024di26a30d.c +++ b/drivers/gpu/drm/panel/panel-feiyang-fy07024di26a30d.c @@ -98,9 +98,7 @@ static int feiyang_enable(struct drm_panel *panel) /* T12 (video & logic signal rise + backlight rise) T12 >= 200ms */ msleep(200); - mipi_dsi_dcs_set_display_on(ctx->dsi); - - return 0; + return mipi_dsi_dcs_set_display_on(ctx->dsi); } static int feiyang_disable(struct drm_panel *panel) From 570cf799e87ae805eacfab3b4ba66676b5fccdb6 Mon Sep 17 00:00:00 2001 From: Icenowy Zheng Date: Sun, 3 May 2026 17:17:08 +0800 Subject: [PATCH 102/333] drm/panel: boe-tv101wum-nl6: restore MODE_LPM after sending disable cmds When preparing the panel, it seems that it always expects commands to be transferred in LP mode. However, the disable function removes the MIPI_DSI_MODE_LPM flag, and no other function re-adds it. As the unprepare function contains no DSI commands, re-adding the flag just after disabling the panel should be safe. Add the code re-adding the flag after the two commands for disabling the panel are sent. This fixes error messages shown in kernel log when unblanking on mt8183-kukui-kodama-sku32 device. Cc: stable@vger.kernel.org Fixes: a869b9db7adf ("drm/panel: support for boe tv101wum-nl6 wuxga dsi video mode panel") Signed-off-by: Icenowy Zheng Reviewed-by: Neil Armstrong Reviewed-by: Douglas Anderson Signed-off-by: Neil Armstrong Link: https://patch.msgid.link/20260503091708.1079962-1-zhengxingda@iscas.ac.cn --- drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c b/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c index d5fe105bdbdd..658ce64c71eb 100644 --- a/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c +++ b/drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c @@ -1324,6 +1324,8 @@ static int boe_panel_disable(struct drm_panel *panel) mipi_dsi_dcs_set_display_off_multi(&ctx); mipi_dsi_dcs_enter_sleep_mode_multi(&ctx); + boe->dsi->mode_flags |= MIPI_DSI_MODE_LPM; + mipi_dsi_msleep(&ctx, 150); return ctx.accum_err; From 2d4e80271f784aa0c7b17676e9762c7e8156be1c Mon Sep 17 00:00:00 2001 From: Icenowy Zheng Date: Sun, 26 Apr 2026 00:57:51 +0800 Subject: [PATCH 103/333] drm/panel: himax-hx83102: restore MODE_LPM after sending disable cmds When preparing the panel, it seems that it always expects commands to be transferred in LP mode. However, the disable function removes the MIPI_DSI_MODE_LPM flag, and no other function re-adds it. As the unprepare function contains no DSI commands, re-adding the flag just after disabling the panel should be safe. Add the code re-adding the flag after the two commands for disabling the panel are sent. This fixes screen unblanking (after blanking once) on mt8188-geralt-ciri-sku1 device. Cc: stable@vger.kernel.org # 6.11+ Fixes: 0ef94554dc40 ("drm/panel: himax-hx83102: Break out as separate driver") Signed-off-by: Icenowy Zheng Reviewed-by: Neil Armstrong Reviewed-by: Douglas Anderson Signed-off-by: Neil Armstrong Link: https://patch.msgid.link/20260425165751.1716569-1-zhengxingda@iscas.ac.cn --- drivers/gpu/drm/panel/panel-himax-hx83102.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/panel/panel-himax-hx83102.c b/drivers/gpu/drm/panel/panel-himax-hx83102.c index 8b2a68ee851e..a5e5c9ea7a73 100644 --- a/drivers/gpu/drm/panel/panel-himax-hx83102.c +++ b/drivers/gpu/drm/panel/panel-himax-hx83102.c @@ -937,6 +937,8 @@ static int hx83102_disable(struct drm_panel *panel) mipi_dsi_dcs_set_display_off_multi(&dsi_ctx); mipi_dsi_dcs_enter_sleep_mode_multi(&dsi_ctx); + dsi->mode_flags |= MIPI_DSI_MODE_LPM; + mipi_dsi_msleep(&dsi_ctx, 150); return dsi_ctx.accum_err; From 8cf5dd235eff6008cb04c3d8064d2acfa90616f1 Mon Sep 17 00:00:00 2001 From: Prasanna Kumar T S M Date: Wed, 1 Apr 2026 04:18:56 -0700 Subject: [PATCH 104/333] EDAC/versalnet: Fix device name memory leak The device name allocated via kzalloc() in init_one_mc() is assigned to dev->init_name but never freed on the normal removal path. device_register() copies init_name and then sets dev->init_name to NULL, so the name pointer becomes unreachable from the device. Thus leaking memory. Use a stack-local char array instead of using kzalloc() for name. Fixes: d5fe2fec6c40 ("EDAC: Add a driver for the AMD Versal NET DDR controller") Signed-off-by: Prasanna Kumar T S M Signed-off-by: Borislav Petkov (AMD) Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20260401111856.2342975-1-ptsm@linux.microsoft.com --- drivers/edac/versalnet_edac.c | 10 ++-------- 1 file changed, 2 insertions(+), 8 deletions(-) diff --git a/drivers/edac/versalnet_edac.c b/drivers/edac/versalnet_edac.c index ec1315582414..97ec05d68bbb 100644 --- a/drivers/edac/versalnet_edac.c +++ b/drivers/edac/versalnet_edac.c @@ -777,9 +777,9 @@ static int init_one_mc(struct mc_priv *priv, struct platform_device *pdev, int i u32 num_chans, rank, dwidth, config; struct edac_mc_layer layers[2]; struct mem_ctl_info *mci; + char name[MC_NAME_LEN]; struct device *dev; enum dev_type dt; - char *name; int rc; config = priv->adec[CONF + i * ADEC_NUM]; @@ -813,13 +813,9 @@ static int init_one_mc(struct mc_priv *priv, struct platform_device *pdev, int i layers[1].is_virt_csrow = false; rc = -ENOMEM; - name = kzalloc(MC_NAME_LEN, GFP_KERNEL); - if (!name) - return rc; - dev = kzalloc(sizeof(*dev), GFP_KERNEL); if (!dev) - goto err_name_free; + return rc; mci = edac_mc_alloc(i, ARRAY_SIZE(layers), layers, sizeof(struct mc_priv)); if (!mci) { @@ -858,8 +854,6 @@ static int init_one_mc(struct mc_priv *priv, struct platform_device *pdev, int i edac_mc_free(mci); err_dev_free: kfree(dev); -err_name_free: - kfree(name); return rc; } From b9eac6a9d93c952c4b7775a24d5c7a1bbf4c3c00 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 25 Apr 2026 00:47:54 +0200 Subject: [PATCH 105/333] rseq: Revert to historical performance killing behaviour The recent RSEQ optimization work broke the TCMalloc abuse of the RSEQ ABI as it not longer unconditionally updates the CPU, node, mm_cid fields, which are documented as read only for user space. Due to the observed behavior of the kernel it was possible for TCMalloc to overwrite the cpu_id_start field for their own purposes and rely on the kernel to update it unconditionally after each context switch and before signal delivery. The RSEQ ABI only guarantees that these fields are updated when the data changes, i.e. the task is migrated or the MMCID of the task changes due to switching from or to per CPU ownership mode. The optimization work eliminated the unconditional updates and reduced them to the documented ABI guarantees, which results in a massive performance win for syscall, scheduling heavy work loads, which in turn breaks the TCMalloc expectations. There have been several options discussed to restore the TCMalloc functionality while preserving the optimization benefits. They all end up in a series of hard to maintain workarounds, which in the worst case introduce overhead for everyone, e.g. in the scheduler. The requirements of TCMalloc and the optimization work are diametral and the required work arounds are a maintainence burden. They end up as fragile constructs, which are blocking further optimization work and are pretty much guaranteed to cause more subtle issues down the road. The optimization work heavily depends on the generic entry code, which is not used by all architectures yet. So the rework preserved the original mechanism moslty unmodified to keep the support for architectures, which handle rseq in their own exit to user space loop. That code is currently optimized out by the compiler on architectures which use the generic entry code. This allows to revert back to the original behaviour by replacing the compile time constant conditions with a runtime condition where required, which disables the optimization and the dependend time slice extension feature until the run-time condition can be enabled in the RSEQ registration code on a per task basis again. The following changes are required to restore the original behavior, which makes TCMalloc work again: 1) Replace the compile time constant conditionals with runtime conditionals where appropriate to prevent the compiler from optimizing the legacy mode out 2) Enforce unconditional update of IDs on context switch for the non-optimized v1 mode 3) Enforce update of IDs in the pre signal delivery path for the non-optimized v1 mode 4) Enforce update of IDs in the membarrier(RSEQ) IPI for the non-optimized v1 mode 5) Make time slice and future extensions depend on optimized v2 mode This brings back the full performance problems, but preserves the v2 optimization code and for generic entry code using architectures also the TIF_RSEQ optimization which avoids a full evaluation of the exit to user mode loop in many cases. Fixes: 566d8015f7ee ("rseq: Avoid CPU/MM CID updates when no event pending") Reported-by: Mathias Stearn Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Closes: https://lore.kernel.org/CAHnCjA25b+nO2n5CeifknSKHssJpPrjnf+dtr7UgzRw4Zgu=oA@mail.gmail.com Link: https://patch.msgid.link/20260428224427.517051752%40kernel.org Cc: stable@vger.kernel.org --- include/linux/rseq.h | 35 ++++++++++++++++++++++----------- include/linux/rseq_entry.h | 39 +++++++++++++++++++++++++++---------- include/linux/rseq_types.h | 9 ++++++++- kernel/rseq.c | 40 +++++++++++++++++++++++++++++++------- kernel/sched/membarrier.c | 11 ++++++++++- 5 files changed, 104 insertions(+), 30 deletions(-) diff --git a/include/linux/rseq.h b/include/linux/rseq.h index f446909551df..7ef79b25e714 100644 --- a/include/linux/rseq.h +++ b/include/linux/rseq.h @@ -9,6 +9,11 @@ void __rseq_handle_slowpath(struct pt_regs *regs); +static __always_inline bool rseq_v2(struct task_struct *t) +{ + return IS_ENABLED(CONFIG_GENERIC_IRQ_ENTRY) && likely(t->rseq.event.has_rseq > 1); +} + /* Invoked from resume_user_mode_work() */ static inline void rseq_handle_slowpath(struct pt_regs *regs) { @@ -16,8 +21,7 @@ static inline void rseq_handle_slowpath(struct pt_regs *regs) if (current->rseq.event.slowpath) __rseq_handle_slowpath(regs); } else { - /* '&' is intentional to spare one conditional branch */ - if (current->rseq.event.sched_switch & current->rseq.event.has_rseq) + if (current->rseq.event.sched_switch && current->rseq.event.has_rseq) __rseq_handle_slowpath(regs); } } @@ -30,9 +34,9 @@ void __rseq_signal_deliver(int sig, struct pt_regs *regs); */ static inline void rseq_signal_deliver(struct ksignal *ksig, struct pt_regs *regs) { - if (IS_ENABLED(CONFIG_GENERIC_IRQ_ENTRY)) { - /* '&' is intentional to spare one conditional branch */ - if (current->rseq.event.has_rseq & current->rseq.event.user_irq) + if (rseq_v2(current)) { + /* has_rseq is implied in rseq_v2() */ + if (current->rseq.event.user_irq) __rseq_signal_deliver(ksig->sig, regs); } else { if (current->rseq.event.has_rseq) @@ -50,15 +54,22 @@ static __always_inline void rseq_sched_switch_event(struct task_struct *t) { struct rseq_event *ev = &t->rseq.event; - if (IS_ENABLED(CONFIG_GENERIC_IRQ_ENTRY)) { + /* + * Only apply the user_irq optimization for RSEQ ABI V2 registrations. + * Legacy users like TCMalloc rely on the original ABI V1 behaviour + * which updates IDs on every context swtich. + */ + if (rseq_v2(t)) { /* - * Avoid a boat load of conditionals by using simple logic - * to determine whether NOTIFY_RESUME needs to be raised. + * Avoid a boat load of conditionals by using simple logic to + * determine whether TIF_NOTIFY_RESUME or TIF_RSEQ needs to be + * raised. * - * It's required when the CPU or MM CID has changed or - * the entry was from user space. + * It's required when the CPU or MM CID has changed or the entry + * was via interrupt from user space. ev->has_rseq does not have + * to be evaluated here because rseq_v2() implies has_rseq. */ - bool raise = (ev->user_irq | ev->ids_changed) & ev->has_rseq; + bool raise = ev->user_irq | ev->ids_changed; if (raise) { ev->sched_switch = true; @@ -66,6 +77,7 @@ static __always_inline void rseq_sched_switch_event(struct task_struct *t) } } else { if (ev->has_rseq) { + t->rseq.event.ids_changed = true; t->rseq.event.sched_switch = true; rseq_raise_notify_resume(t); } @@ -161,6 +173,7 @@ static inline unsigned int rseq_alloc_align(void) } #else /* CONFIG_RSEQ */ +static inline bool rseq_v2(struct task_struct *t) { return false; } static inline void rseq_handle_slowpath(struct pt_regs *regs) { } static inline void rseq_signal_deliver(struct ksignal *ksig, struct pt_regs *regs) { } static inline void rseq_sched_switch_event(struct task_struct *t) { } diff --git a/include/linux/rseq_entry.h b/include/linux/rseq_entry.h index f11ebd34f8b9..934db41ec782 100644 --- a/include/linux/rseq_entry.h +++ b/include/linux/rseq_entry.h @@ -111,6 +111,20 @@ static __always_inline void rseq_slice_clear_grant(struct task_struct *t) t->rseq.slice.state.granted = false; } +/* + * Open coded, so it can be invoked within a user access region. + * + * This clears the user space state of the time slice extensions field only when + * the task has registered the optimized RSEQ_ABI V2. Some legacy registrations, + * e.g. TCMalloc, have conflicting non-ABI fields in struct RSEQ, which would be + * overwritten by an unconditional write. + */ +#define rseq_slice_clear_user(rseq, efault) \ +do { \ + if (rseq_slice_extension_enabled()) \ + unsafe_put_user(0U, &rseq->slice_ctrl.all, efault); \ +} while (0) + static __always_inline bool __rseq_grant_slice_extension(bool work_pending) { struct task_struct *curr = current; @@ -230,6 +244,7 @@ static __always_inline bool rseq_slice_extension_enabled(void) { return false; } static __always_inline bool rseq_arm_slice_extension_timer(void) { return false; } static __always_inline void rseq_slice_clear_grant(struct task_struct *t) { } static __always_inline bool rseq_grant_slice_extension(unsigned long ti_work, unsigned long mask) { return false; } +#define rseq_slice_clear_user(rseq, efault) do { } while (0) #endif /* !CONFIG_RSEQ_SLICE_EXTENSION */ bool rseq_debug_update_user_cs(struct task_struct *t, struct pt_regs *regs, unsigned long csaddr); @@ -517,11 +532,9 @@ bool rseq_set_ids_get_csaddr(struct task_struct *t, struct rseq_ids *ids, if (csaddr) unsafe_get_user(*csaddr, &rseq->rseq_cs, efault); - /* Open coded, so it's in the same user access region */ - if (rseq_slice_extension_enabled()) { - /* Unconditionally clear it, no point in conditionals */ - unsafe_put_user(0U, &rseq->slice_ctrl.all, efault); - } + /* RSEQ ABI V2 only operations */ + if (rseq_v2(t)) + rseq_slice_clear_user(rseq, efault); } rseq_slice_clear_grant(t); @@ -612,6 +625,14 @@ static __always_inline bool rseq_exit_user_update(struct pt_regs *regs, struct t * interrupts disabled */ guard(pagefault)(); + /* + * This optimization is only valid when the task registered for the + * optimized RSEQ_ABI_V2 variant. Some legacy users rely on the original + * RSEQ implementation behaviour which unconditionally updated the IDs. + * rseq_sched_switch_event() ensures that legacy registrations always + * have both sched_switch and ids_changed set, which is compatible with + * the historical TIF_NOTIFY_RESUME behaviour. + */ if (likely(!t->rseq.event.ids_changed)) { struct rseq __user *rseq = t->rseq.usrptr; /* @@ -623,11 +644,9 @@ static __always_inline bool rseq_exit_user_update(struct pt_regs *regs, struct t scoped_user_rw_access(rseq, efault) { unsafe_get_user(csaddr, &rseq->rseq_cs, efault); - /* Open coded, so it's in the same user access region */ - if (rseq_slice_extension_enabled()) { - /* Unconditionally clear it, no point in conditionals */ - unsafe_put_user(0U, &rseq->slice_ctrl.all, efault); - } + /* RSEQ ABI V2 only operations */ + if (rseq_v2(t)) + rseq_slice_clear_user(rseq, efault); } rseq_slice_clear_grant(t); diff --git a/include/linux/rseq_types.h b/include/linux/rseq_types.h index 0b42045988db..a469c1870849 100644 --- a/include/linux/rseq_types.h +++ b/include/linux/rseq_types.h @@ -9,6 +9,12 @@ #ifdef CONFIG_RSEQ struct rseq; +/* + * rseq_event::has_rseq contains the ABI version number so preserving it + * in AND operations requires a mask. + */ +#define RSEQ_HAS_RSEQ_VERSION_MASK 0xff + /** * struct rseq_event - Storage for rseq related event management * @all: Compound to initialize and clear the data efficiently @@ -17,7 +23,8 @@ struct rseq; * exit to user * @ids_changed: Indicator that IDs need to be updated * @user_irq: True on interrupt entry from user mode - * @has_rseq: True if the task has a rseq pointer installed + * @has_rseq: Greater than 0 if the task has a rseq pointer installed. + * Contains the RSEQ version number * @error: Compound error code for the slow path to analyze * @fatal: User space data corrupted or invalid * @slowpath: Indicator that slow path processing via TIF_NOTIFY_RESUME diff --git a/kernel/rseq.c b/kernel/rseq.c index 586f58f652c6..aa25753ea135 100644 --- a/kernel/rseq.c +++ b/kernel/rseq.c @@ -253,11 +253,14 @@ static bool rseq_handle_cs(struct task_struct *t, struct pt_regs *regs) static void rseq_slowpath_update_usr(struct pt_regs *regs) { /* - * Preserve rseq state and user_irq state. The generic entry code - * clears user_irq on the way out, the non-generic entry - * architectures are not having user_irq. + * Preserve has_rseq and user_irq state. The generic entry code clears + * user_irq on the way out, the non-generic entry architectures are not + * setting user_irq. */ - const struct rseq_event evt_mask = { .has_rseq = true, .user_irq = true, }; + const struct rseq_event evt_mask = { + .has_rseq = RSEQ_HAS_RSEQ_VERSION_MASK, + .user_irq = true, + }; struct task_struct *t = current; struct rseq_ids ids; u32 node_id; @@ -330,8 +333,9 @@ void __rseq_handle_slowpath(struct pt_regs *regs) void __rseq_signal_deliver(int sig, struct pt_regs *regs) { rseq_stat_inc(rseq_stats.signal); + /* - * Don't update IDs, they are handled on exit to user if + * Don't update IDs yet, they are handled on exit to user if * necessary. The important thing is to abort a critical section of * the interrupted context as after this point the instruction * pointer in @regs points to the signal handler. @@ -344,6 +348,13 @@ void __rseq_signal_deliver(int sig, struct pt_regs *regs) current->rseq.event.error = 0; force_sigsegv(sig); } + + /* + * In legacy mode, force the update of IDs before returning to user + * space to stay compatible. + */ + if (!rseq_v2(current)) + rseq_force_update(); } /* @@ -408,6 +419,7 @@ static bool rseq_reset_ids(void) SYSCALL_DEFINE4(rseq, struct rseq __user *, rseq, u32, rseq_len, int, flags, u32, sig) { u32 rseqfl = 0; + u8 version = 1; if (flags & RSEQ_FLAG_UNREGISTER) { if (flags & ~RSEQ_FLAG_UNREGISTER) @@ -461,7 +473,11 @@ SYSCALL_DEFINE4(rseq, struct rseq __user *, rseq, u32, rseq_len, int, flags, u32 if (!access_ok(rseq, rseq_len)) return -EFAULT; - if (IS_ENABLED(CONFIG_RSEQ_SLICE_EXTENSION)) { + /* + * The version check effectivly disables time slice extensions until the + * RSEQ ABI V2 registration are implemented. + */ + if (IS_ENABLED(CONFIG_RSEQ_SLICE_EXTENSION) && version > 1) { if (rseq_slice_extension_enabled()) { rseqfl |= RSEQ_CS_FLAG_SLICE_EXT_AVAILABLE; if (flags & RSEQ_FLAG_SLICE_EXT_DEFAULT_ON) @@ -484,7 +500,15 @@ SYSCALL_DEFINE4(rseq, struct rseq __user *, rseq, u32, rseq_len, int, flags, u32 unsafe_put_user(RSEQ_CPU_ID_UNINITIALIZED, &rseq->cpu_id, efault); unsafe_put_user(0U, &rseq->node_id, efault); unsafe_put_user(0U, &rseq->mm_cid, efault); - unsafe_put_user(0U, &rseq->slice_ctrl.all, efault); + + /* + * All fields past mm_cid are only valid for non-legacy v2 + * registrations. + */ + if (version > 1) { + if (IS_ENABLED(CONFIG_RSEQ_SLICE_EXTENSION)) + unsafe_put_user(0U, &rseq->slice_ctrl.all, efault); + } } /* @@ -712,6 +736,8 @@ int rseq_slice_extension_prctl(unsigned long arg2, unsigned long arg3) return -ENOTSUPP; if (!current->rseq.usrptr) return -ENXIO; + if (!rseq_v2(current)) + return -ENOTSUPP; /* No change? */ if (enable == !!current->rseq.slice.state.enabled) diff --git a/kernel/sched/membarrier.c b/kernel/sched/membarrier.c index 623445603725..226a6329f3e9 100644 --- a/kernel/sched/membarrier.c +++ b/kernel/sched/membarrier.c @@ -199,7 +199,16 @@ static void ipi_rseq(void *info) * is negligible. */ smp_mb(); - rseq_sched_switch_event(current); + /* + * Legacy mode requires that IDs are written and the critical section is + * evaluated. V2 optimized mode handles the critical section and IDs are + * only updated if they change as a consequence of preemption after + * return from this IPI. + */ + if (rseq_v2(current)) + rseq_sched_switch_event(current); + else + rseq_force_update(); } static void ipi_sync_rq_state(void *info) From 02b44d943b3adddc3a15c1da97045e205b7d14c1 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 25 Apr 2026 15:46:06 +0200 Subject: [PATCH 106/333] selftests/rseq: Skip tests if time slice extensions are not available Don't fail, skip the test if the extensions are not enabled at compile or runtime. Fixes: 830969e7821a ("selftests/rseq: Implement time slice extension test") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224427.597838491%40kernel.org Cc: stable@vger.kernel.org --- tools/testing/selftests/rseq/slice_test.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/tools/testing/selftests/rseq/slice_test.c b/tools/testing/selftests/rseq/slice_test.c index 357122dcb487..77e668ff74d7 100644 --- a/tools/testing/selftests/rseq/slice_test.c +++ b/tools/testing/selftests/rseq/slice_test.c @@ -124,6 +124,13 @@ FIXTURE_SETUP(slice_ext) { cpu_set_t affinity; + if (rseq_register_current_thread()) + SKIP(return, "RSEQ not supported\n"); + + if (prctl(PR_RSEQ_SLICE_EXTENSION, PR_RSEQ_SLICE_EXTENSION_SET, + PR_RSEQ_SLICE_EXT_ENABLE, 0, 0)) + SKIP(return, "Time slice extension not supported\n"); + ASSERT_EQ(sched_getaffinity(0, sizeof(affinity), &affinity), 0); /* Pin it on a single CPU. Avoid CPU 0 */ @@ -137,11 +144,6 @@ FIXTURE_SETUP(slice_ext) break; } - ASSERT_EQ(rseq_register_current_thread(), 0); - - ASSERT_EQ(prctl(PR_RSEQ_SLICE_EXTENSION, PR_RSEQ_SLICE_EXTENSION_SET, - PR_RSEQ_SLICE_EXT_ENABLE, 0, 0), 0); - self->noise_params.noise_nsecs = variant->noise_nsecs; self->noise_params.sleep_nsecs = variant->sleep_nsecs; self->noise_params.run = 1; From d97cb2ef0b221b068e90b6058aa97faa0626bdab Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 26 Apr 2026 18:13:54 +0200 Subject: [PATCH 107/333] selftests/rseq: Make registration flexible for legacy and optimized mode rseq_register_current_thread() either uses the glibc registered RSEQ region or registers it's own region with the legacy size of 32 bytes. That worked so far, but becomes a problem when the kernel implements a distinction between legacy and performance optimized behavior based on the registration size as that does not allow to test both modes with the self test suite. Add two arguments to the function. One to enforce that the registration is not using libc provided mode and one to tell the registration to use the legacy size and not the kernel advertised size. Rename it and make the original one a inline wrapper which preserves the existing behavior. Fixes: 566d8015f7ee ("rseq: Avoid CPU/MM CID updates when no event pending") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224427.677889423%40kernel.org Cc: stable@vger.kernel.org --- tools/testing/selftests/rseq/rseq-abi.h | 7 ++++- tools/testing/selftests/rseq/rseq.c | 39 ++++++++++++------------- tools/testing/selftests/rseq/rseq.h | 8 ++++- 3 files changed, 31 insertions(+), 23 deletions(-) diff --git a/tools/testing/selftests/rseq/rseq-abi.h b/tools/testing/selftests/rseq/rseq-abi.h index ecef315204b2..5f4ea2152c2f 100644 --- a/tools/testing/selftests/rseq/rseq-abi.h +++ b/tools/testing/selftests/rseq/rseq-abi.h @@ -191,10 +191,15 @@ struct rseq_abi { */ struct rseq_abi_slice_ctrl slice_ctrl; + /* + * Place holder to push the size above 32 bytes. + */ + __u8 __reserved; + /* * Flexible array member at end of structure, after last feature field. */ char end[]; -} __attribute__((aligned(4 * sizeof(__u64)))); +} __attribute__((aligned(256))); #endif /* _RSEQ_ABI_H */ diff --git a/tools/testing/selftests/rseq/rseq.c b/tools/testing/selftests/rseq/rseq.c index a736727b83c1..be0d0a97031e 100644 --- a/tools/testing/selftests/rseq/rseq.c +++ b/tools/testing/selftests/rseq/rseq.c @@ -56,6 +56,7 @@ ptrdiff_t rseq_offset; * unsuccessful. */ unsigned int rseq_size = -1U; +static unsigned int rseq_alloc_size; /* Flags used during rseq registration. */ unsigned int rseq_flags; @@ -115,29 +116,17 @@ bool rseq_available(void) } } -/* The rseq areas need to be at least 32 bytes. */ -static -unsigned int get_rseq_min_alloc_size(void) -{ - unsigned int alloc_size = rseq_size; - - if (alloc_size < ORIG_RSEQ_ALLOC_SIZE) - alloc_size = ORIG_RSEQ_ALLOC_SIZE; - return alloc_size; -} - /* * Return the feature size supported by the kernel. * * Depending on the value returned by getauxval(AT_RSEQ_FEATURE_SIZE): * - * 0: Return ORIG_RSEQ_FEATURE_SIZE (20) + * 0: Return ORIG_RSEQ_FEATURE_SIZE (20) * > 0: Return the value from getauxval(AT_RSEQ_FEATURE_SIZE). * * It should never return a value below ORIG_RSEQ_FEATURE_SIZE. */ -static -unsigned int get_rseq_kernel_feature_size(void) +static unsigned int get_rseq_kernel_feature_size(void) { unsigned long auxv_rseq_feature_size, auxv_rseq_align; @@ -152,15 +141,24 @@ unsigned int get_rseq_kernel_feature_size(void) return ORIG_RSEQ_FEATURE_SIZE; } -int rseq_register_current_thread(void) +int __rseq_register_current_thread(bool nolibc, bool legacy) { + unsigned int size; int rc; if (!rseq_ownership) { /* Treat libc's ownership as a successful registration. */ - return 0; + return nolibc ? -EBUSY : 0; } - rc = sys_rseq(&__rseq.abi, get_rseq_min_alloc_size(), 0, RSEQ_SIG); + + /* The minimal allocation size is 32, which is the legacy allocation size */ + size = get_rseq_kernel_feature_size(); + if (legacy || size < ORIG_RSEQ_ALLOC_SIZE) + rseq_alloc_size = ORIG_RSEQ_ALLOC_SIZE; + else + rseq_alloc_size = size; + + rc = sys_rseq(&__rseq.abi, rseq_alloc_size, 0, RSEQ_SIG); if (rc) { /* * After at least one thread has registered successfully @@ -179,9 +177,8 @@ int rseq_register_current_thread(void) * The first thread to register sets the rseq_size to mimic the libc * behavior. */ - if (RSEQ_READ_ONCE(rseq_size) == 0) { - RSEQ_WRITE_ONCE(rseq_size, get_rseq_kernel_feature_size()); - } + if (RSEQ_READ_ONCE(rseq_size) == 0) + RSEQ_WRITE_ONCE(rseq_size, size); return 0; } @@ -194,7 +191,7 @@ int rseq_unregister_current_thread(void) /* Treat libc's ownership as a successful unregistration. */ return 0; } - rc = sys_rseq(&__rseq.abi, get_rseq_min_alloc_size(), RSEQ_ABI_FLAG_UNREGISTER, RSEQ_SIG); + rc = sys_rseq(&__rseq.abi, rseq_alloc_size, RSEQ_ABI_FLAG_UNREGISTER, RSEQ_SIG); if (rc) return -1; return 0; diff --git a/tools/testing/selftests/rseq/rseq.h b/tools/testing/selftests/rseq/rseq.h index f51a5fdb0444..c62ebb9290c0 100644 --- a/tools/testing/selftests/rseq/rseq.h +++ b/tools/testing/selftests/rseq/rseq.h @@ -8,6 +8,7 @@ #ifndef RSEQ_H #define RSEQ_H +#include #include #include #include @@ -142,7 +143,12 @@ static inline struct rseq_abi *rseq_get_abi(void) * succeed. A restartable sequence executed from a non-registered * thread will always fail. */ -int rseq_register_current_thread(void); +int __rseq_register_current_thread(bool nolibc, bool legacy); + +static inline int rseq_register_current_thread(void) +{ + return __rseq_register_current_thread(false, false); +} /* * Unregister rseq for current thread. From 9b4e3495d1bd2469bf94b74930c153c2d534ddb7 Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Mon, 20 Apr 2026 11:55:57 -0400 Subject: [PATCH 108/333] drm/amdkfd: Make all TLB-flushes heavy-weight With only one sequence number we cannot track the need for legacy vs heavy-weight flushes reliably. Always use heavy-weight. Signed-off-by: Felix Kuehling Reviewed-by: Philip Yang Signed-off-by: Alex Deucher (cherry picked from commit c1a3ff1d327820cd9a52bc1056b98681fc088949) Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 4 ++-- drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c | 6 +++--- drivers/gpu/drm/amd/amdkfd/kfd_priv.h | 6 +++--- drivers/gpu/drm/amd/amdkfd/kfd_svm.c | 4 ++-- 4 files changed, 10 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c index f829d65a79b4..f95bf6d95534 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c @@ -1360,7 +1360,7 @@ static int kfd_ioctl_map_memory_to_gpu(struct file *filep, peer_pdd = kfd_process_device_data_by_id(p, devices_arr[i]); if (WARN_ON_ONCE(!peer_pdd)) continue; - kfd_flush_tlb(peer_pdd, TLB_FLUSH_LEGACY); + kfd_flush_tlb(peer_pdd); } kfree(devices_arr); @@ -1455,7 +1455,7 @@ static int kfd_ioctl_unmap_memory_from_gpu(struct file *filep, if (WARN_ON_ONCE(!peer_pdd)) continue; if (flush_tlb) - kfd_flush_tlb(peer_pdd, TLB_FLUSH_HEAVYWEIGHT); + kfd_flush_tlb(peer_pdd); /* Remove dma mapping after tlb flush to avoid IO_PAGE_FAULT */ err = amdgpu_amdkfd_gpuvm_dmaunmap_mem(mem, peer_pdd->drm_priv); diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c index ab3b2e7be9bd..9185ebe4c079 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -572,7 +572,7 @@ static int allocate_vmid(struct device_queue_manager *dqm, qpd->vmid, qpd->page_table_base); /* invalidate the VM context after pasid and vmid mapping is set up */ - kfd_flush_tlb(qpd_to_pdd(qpd), TLB_FLUSH_LEGACY); + kfd_flush_tlb(qpd_to_pdd(qpd)); if (dqm->dev->kfd2kgd->set_scratch_backing_va) dqm->dev->kfd2kgd->set_scratch_backing_va(dqm->dev->adev, @@ -610,7 +610,7 @@ static void deallocate_vmid(struct device_queue_manager *dqm, if (flush_texture_cache_nocpsch(q->device, qpd)) dev_err(dev, "Failed to flush TC\n"); - kfd_flush_tlb(qpd_to_pdd(qpd), TLB_FLUSH_LEGACY); + kfd_flush_tlb(qpd_to_pdd(qpd)); /* Release the vmid mapping */ set_pasid_vmid_mapping(dqm, 0, qpd->vmid); @@ -1284,7 +1284,7 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm, dqm->dev->adev, qpd->vmid, qpd->page_table_base); - kfd_flush_tlb(pdd, TLB_FLUSH_LEGACY); + kfd_flush_tlb(pdd); } /* Take a safe reference to the mm_struct, which may otherwise diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index 163d665a6074..7b5b12206919 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -1554,13 +1554,13 @@ void kfd_signal_reset_event(struct kfd_node *dev); void kfd_signal_poison_consumed_event(struct kfd_node *dev, u32 pasid); void kfd_signal_process_terminate_event(struct kfd_process *p); -static inline void kfd_flush_tlb(struct kfd_process_device *pdd, - enum TLB_FLUSH_TYPE type) +static inline void kfd_flush_tlb(struct kfd_process_device *pdd) { struct amdgpu_device *adev = pdd->dev->adev; struct amdgpu_vm *vm = drm_priv_to_vm(pdd->drm_priv); - amdgpu_vm_flush_compute_tlb(adev, vm, type, pdd->dev->xcc_mask); + amdgpu_vm_flush_compute_tlb(adev, vm, TLB_FLUSH_HEAVYWEIGHT, + pdd->dev->xcc_mask); } static inline bool kfd_flush_tlb_after_unmap(struct kfd_dev *dev) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c index 38085a0a0f58..35ec67d9739b 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c @@ -1424,7 +1424,7 @@ svm_range_unmap_from_gpus(struct svm_range *prange, unsigned long start, if (r) break; } - kfd_flush_tlb(pdd, TLB_FLUSH_HEAVYWEIGHT); + kfd_flush_tlb(pdd); } return r; @@ -1571,7 +1571,7 @@ svm_range_map_to_gpus(struct svm_range *prange, unsigned long offset, } } - kfd_flush_tlb(pdd, TLB_FLUSH_LEGACY); + kfd_flush_tlb(pdd); } return r; From 7bbfb2559bcec39d1a4e1182d931a2046112c352 Mon Sep 17 00:00:00 2001 From: "John B. Moore" Date: Tue, 28 Apr 2026 11:35:12 -0500 Subject: [PATCH 109/333] drm/amdgpu/gfx9: drop unnecessary 64-bit fence flag check in KIQ MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Remove the BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT) assertion from gfx_v9_0_ring_emit_fence_kiq(). The KIQ hardware supports 64-bit fence writes; the 32-bit writeback address constraint is an upper-layer convention, not a hardware limitation. The check serves no purpose and should not be present. Found by code inspection while investigating related BUG_ON assertions in the GFX and compute ring emission paths. Reviewed-by: Christian König Signed-off-by: John B. Moore Signed-off-by: Alex Deucher (cherry picked from commit 1b1101a46a426bb4328116bb5273c326a2780389) Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index 95be105671ec..86c7c2a429b7 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -5660,9 +5660,6 @@ static void gfx_v9_0_ring_emit_fence_kiq(struct amdgpu_ring *ring, u64 addr, { struct amdgpu_device *adev = ring->adev; - /* we only allocate 32bit for each seq wb address */ - BUG_ON(flags & AMDGPU_FENCE_FLAG_64BIT); - /* write fence seq to the "addr" */ amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3)); amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(0) | From 2a561b361b7681509710f3cfc3d95d54c87ac69f Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 27 Apr 2026 11:38:58 -0400 Subject: [PATCH 110/333] drm/amdgpu/pm: add missing revision check for CI MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The ci_populate_all_memory_levels() workaround only applies to revision 0 SKUs. Link: https://gitlab.freedesktop.org/drm/amd/-/work_items/1816 Fixes: 9f4b35411cfe ("drm/amd/powerplay: add CI asics support to smumgr (v3)") Reviewed-by: Timur Kristóf Reviewed-by: Kent Russell Signed-off-by: Alex Deucher (cherry picked from commit 1db15ba8f72f400bbad8ae0ce24fafc43429d4bd) Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c index 731355bdb9bc..0a3a0722b5c9 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c @@ -1333,8 +1333,9 @@ static int ci_populate_all_memory_levels(struct pp_hwmgr *hwmgr) dev_id = adev->pdev->device; - if ((dpm_table->mclk_table.count >= 2) - && ((dev_id == 0x67B0) || (dev_id == 0x67B1))) { + if ((dpm_table->mclk_table.count >= 2) && + ((dev_id == 0x67B0) || (dev_id == 0x67B1)) && + (adev->pdev->revision == 0)) { smu_data->smc_state_table.MemoryLevel[1].MinVddci = smu_data->smc_state_table.MemoryLevel[0].MinVddci; smu_data->smc_state_table.MemoryLevel[1].MinMvdd = From 1987c79b4fe5789dfa14423e78b5c25f6acf3e9d Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Tue, 28 Apr 2026 10:42:49 -0400 Subject: [PATCH 111/333] drm/amdgpu/pm: align Hawaii mclk workaround with radeon MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Align the hawaii mclk workaround with radeon and windows. Link: https://gitlab.freedesktop.org/drm/amd/-/work_items/1816 Fixes: 9f4b35411cfe ("drm/amd/powerplay: add CI asics support to smumgr (v3)") Reviewed-by: Timur Kristóf Reviewed-by: Kent Russell Signed-off-by: Alex Deucher (cherry picked from commit 9649528b637f668c5af9f2b83ca4ad8576ae2121) Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c index 0a3a0722b5c9..3650e7beeb67 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/smumgr/ci_smumgr.c @@ -1336,10 +1336,10 @@ static int ci_populate_all_memory_levels(struct pp_hwmgr *hwmgr) if ((dpm_table->mclk_table.count >= 2) && ((dev_id == 0x67B0) || (dev_id == 0x67B1)) && (adev->pdev->revision == 0)) { - smu_data->smc_state_table.MemoryLevel[1].MinVddci = - smu_data->smc_state_table.MemoryLevel[0].MinVddci; - smu_data->smc_state_table.MemoryLevel[1].MinMvdd = - smu_data->smc_state_table.MemoryLevel[0].MinMvdd; + smu_data->smc_state_table.MemoryLevel[1].MinVddc = + smu_data->smc_state_table.MemoryLevel[0].MinVddc; + smu_data->smc_state_table.MemoryLevel[1].MinVddcPhases = + smu_data->smc_state_table.MemoryLevel[0].MinVddcPhases; } smu_data->smc_state_table.MemoryLevel[0].ActivityLevel = 0x1F; CONVERT_FROM_HOST_TO_SMC_US(smu_data->smc_state_table.MemoryLevel[0].ActivityLevel); From 17223816498f7b117d138d18eb0eba63604dc74e Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 27 Apr 2026 11:40:25 -0400 Subject: [PATCH 112/333] drm/radeon: add missing revision check for CI MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The memory level workarounds only apply to revision 0 SKUs. Link: https://gitlab.freedesktop.org/drm/amd/-/work_items/1816 Fixes: 127e056e2a82 ("drm/radeon: fix mclk vddc configuration for cards for hawaii") Fixes: 21b8a369046f ("drm/radeon: fix dram timing for certain hawaii boards") Fixes: 90b2fee35cb9 ("drm/radeon: fix dpm mc init for certain hawaii boards") Reviewed-by: Timur Kristóf Reviewed-by: Kent Russell Signed-off-by: Alex Deucher (cherry picked from commit 4d8dcc14311515077062b5740f39f427075de5c9) Cc: stable@vger.kernel.org --- drivers/gpu/drm/radeon/ci_dpm.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/radeon/ci_dpm.c b/drivers/gpu/drm/radeon/ci_dpm.c index 22321eb95b7d..703848fac189 100644 --- a/drivers/gpu/drm/radeon/ci_dpm.c +++ b/drivers/gpu/drm/radeon/ci_dpm.c @@ -2461,7 +2461,8 @@ static void ci_register_patching_mc_arb(struct radeon_device *rdev, if (patch && ((rdev->pdev->device == 0x67B0) || - (rdev->pdev->device == 0x67B1))) { + (rdev->pdev->device == 0x67B1)) && + (rdev->pdev->revision == 0)) { if ((memory_clock > 100000) && (memory_clock <= 125000)) { tmp2 = (((0x31 * engine_clock) / 125000) - 1) & 0xff; *dram_timimg2 &= ~0x00ff0000; @@ -3304,7 +3305,8 @@ static int ci_populate_all_memory_levels(struct radeon_device *rdev) pi->smc_state_table.MemoryLevel[0].EnabledForActivity = 1; if ((dpm_table->mclk_table.count >= 2) && - ((rdev->pdev->device == 0x67B0) || (rdev->pdev->device == 0x67B1))) { + ((rdev->pdev->device == 0x67B0) || (rdev->pdev->device == 0x67B1)) && + (rdev->pdev->revision == 0)) { pi->smc_state_table.MemoryLevel[1].MinVddc = pi->smc_state_table.MemoryLevel[0].MinVddc; pi->smc_state_table.MemoryLevel[1].MinVddcPhases = @@ -4493,7 +4495,8 @@ static int ci_register_patching_mc_seq(struct radeon_device *rdev, if (patch && ((rdev->pdev->device == 0x67B0) || - (rdev->pdev->device == 0x67B1))) { + (rdev->pdev->device == 0x67B1)) && + (rdev->pdev->revision == 0)) { for (i = 0; i < table->last; i++) { if (table->last >= SMU7_DISCRETE_MC_REGISTER_ARRAY_SIZE) return -EINVAL; From 78d2e624fa073c14970aa097adcf3ea31c157a66 Mon Sep 17 00:00:00 2001 From: "John B. Moore" Date: Mon, 27 Apr 2026 16:06:28 -0500 Subject: [PATCH 113/333] drm/amdgpu/sdma4: replace BUG_ON with WARN_ON in fence emission MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit sdma_v4_0_ring_emit_fence() contains two BUG_ON(addr & 0x3) assertions that verify fence writeback addresses are dword-aligned. These assertions can be reached from unprivileged userspace via crafted DRM_IOCTL_AMDGPU_CS submissions, causing a fatal kernel panic in a scheduler worker thread. Replace both BUG_ON() calls with WARN_ON() to log the condition without crashing the kernel. A misaligned fence address at this point indicates a driver bug, but crashing the kernel is never the correct response when the assertion is reachable from userspace. The CS IOCTL path is the correct place to filter invalid submissions; the ring emission callback is too late to do anything about it. Fixes: 2130f89ced2c ("drm/amdgpu: add SDMA v4.0 implementation (v2)") Reviewed-by: Christian König Signed-off-by: John B. Moore Signed-off-by: Alex Deucher (cherry picked from commit b90250bd933afd1ba94d86d6b13821997b22b18e) Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c index 44f0f23e1148..e64f2f6df9a9 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c @@ -889,7 +889,7 @@ static void sdma_v4_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se /* write the fence */ amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_FENCE)); /* zero in first two bits */ - BUG_ON(addr & 0x3); + WARN_ON(addr & 0x3); amdgpu_ring_write(ring, lower_32_bits(addr)); amdgpu_ring_write(ring, upper_32_bits(addr)); amdgpu_ring_write(ring, lower_32_bits(seq)); @@ -899,7 +899,7 @@ static void sdma_v4_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se addr += 4; amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_FENCE)); /* zero in first two bits */ - BUG_ON(addr & 0x3); + WARN_ON(addr & 0x3); amdgpu_ring_write(ring, lower_32_bits(addr)); amdgpu_ring_write(ring, upper_32_bits(addr)); amdgpu_ring_write(ring, upper_32_bits(seq)); From e6c2e6c2e1fa066968a16aca1cb66cd1bdde7741 Mon Sep 17 00:00:00 2001 From: Philip Yang Date: Mon, 27 Apr 2026 09:30:23 -0400 Subject: [PATCH 114/333] drm/amdgpu: zero-initialize GART table on allocation MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit GART TLB is flushed after unmapping but not after mapping. Since amdgpu_bo_create_kernel() does not zero-initialize the buffer, when a single PTE is written the TLB may speculatively load other uninitialized entries from the same cacheline. Those garbage entries can appear valid, and a subsequent write to another PTE in the same cacheline may cause the GPU to use a stale garbage PTE from the TLB. Fix this by calling memset_io() to zero-initialize the GART table with gart_pte_flags immediately after allocation. Using AMDGPU_GEM_CREATE_VRAM_CLEARED, SDMA-based clear will not work since SDMA needs GART to be initialized to work. Suggested-by: Felix Kuehling Signed-off-by: Philip Yang Reviewed-by: Christian König Signed-off-by: Alex Deucher (cherry picked from commit d9af8263b82b6eaa60c5718e0c6631c5037e4b24) Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index bc772ca3dab7..b6f849d51c2e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c @@ -262,12 +262,19 @@ void amdgpu_gart_table_ram_free(struct amdgpu_device *adev) */ int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev) { + int r; + if (adev->gart.bo != NULL) return 0; - return amdgpu_bo_create_kernel(adev, adev->gart.table_size, PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, &adev->gart.bo, - NULL, (void *)&adev->gart.ptr); + r = amdgpu_bo_create_kernel(adev, adev->gart.table_size, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM, &adev->gart.bo, + NULL, (void *)&adev->gart.ptr); + if (r) + return r; + + memset_io(adev->gart.ptr, adev->gart.gart_pte_flags, adev->gart.table_size); + return 0; } /** From 81665e35f143d93adef654f3be1360def9196e72 Mon Sep 17 00:00:00 2001 From: Xiaogang Chen Date: Fri, 24 Apr 2026 13:47:01 -0500 Subject: [PATCH 115/333] drm/amdkfd: Check if there are kfd porcesses using adev by kfd_processes_count During gpu hot-unplug need check if there are kfd porcesses still using the being removed gpu before clean resources of the device. Current driver checks if kfd_processes_table is empty. kfd processes are not terminated after removed from kfd_processes_table immediately. They are still alive and may access the device until kfd_process_wq work queue got ran. Check kfd->kfd_processes_count value that is updated after kfd process got uninitialized when its ref becomes zero. Fixes: 6cca686dfce7 ("drm/amdkfd: kfd driver supports hot unplug/replug amdgpu devices") Signed-off-by: Xiaogang Chen Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher (cherry picked from commit d12d05c4bc4c15585130af43e897923ff292df7b) --- drivers/gpu/drm/amd/amdkfd/kfd_device.c | 33 +------------------------ 1 file changed, 1 insertion(+), 32 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c index 8ff97bf7d95a..b7f8f7ff8198 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c @@ -1737,37 +1737,6 @@ bool kgd2kfd_vmfault_fast_path(struct amdgpu_device *adev, struct amdgpu_iv_entr return false; } -/* check if there is kfd process still uses adev */ -static bool kgd2kfd_check_device_idle(struct amdgpu_device *adev) -{ - struct kfd_process *p; - struct hlist_node *p_temp; - unsigned int temp; - struct kfd_node *dev; - - mutex_lock(&kfd_processes_mutex); - - if (hash_empty(kfd_processes_table)) { - mutex_unlock(&kfd_processes_mutex); - return true; - } - - /* check if there is device still use adev */ - hash_for_each_safe(kfd_processes_table, temp, p_temp, p, kfd_processes) { - for (int i = 0; i < p->n_pdds; i++) { - dev = p->pdds[i]->dev; - if (dev->adev == adev) { - mutex_unlock(&kfd_processes_mutex); - return false; - } - } - } - - mutex_unlock(&kfd_processes_mutex); - - return true; -} - /** kgd2kfd_teardown_processes - gracefully tear down existing * kfd processes that use adev * @@ -1800,7 +1769,7 @@ void kgd2kfd_teardown_processes(struct amdgpu_device *adev) mutex_unlock(&kfd_processes_mutex); /* wait all kfd processes use adev terminate */ - while (!kgd2kfd_check_device_idle(adev)) + while (!!atomic_read(&adev->kfd.dev->kfd_processes_count)) cond_resched(); } From 6da7b1242da4455b11c24ce667d1cab1a348c8ea Mon Sep 17 00:00:00 2001 From: Sunil Khatri Date: Mon, 4 May 2026 18:21:17 +0530 Subject: [PATCH 116/333] drm/amdgpu/userq: fix access to stale wptr mapping MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use drm_exec to take both locks i.e vm root bo and wptr_obj bo to access the mapping data properly. This fixes the security issue of unmap the wptr_obj while a queue creation is in progress and passing other bo at same address. Signed-off-by: Sunil Khatri Reviewed-by: Christian König Signed-off-by: Alex Deucher (cherry picked from commit 1fc6c8ab45dbee096469c08c13f6099d57a52d6c) Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/mes_userqueue.c | 97 +++++++++------------- 1 file changed, 38 insertions(+), 59 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c index 2fc39a6938f6..5b4121ddc78c 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c @@ -30,34 +30,6 @@ #define AMDGPU_USERQ_PROC_CTX_SZ PAGE_SIZE #define AMDGPU_USERQ_GANG_CTX_SZ PAGE_SIZE -static int -mes_userq_map_gtt_bo_to_gart(struct amdgpu_bo *bo) -{ - int ret; - - ret = amdgpu_bo_reserve(bo, true); - if (ret) { - DRM_ERROR("Failed to reserve bo. ret %d\n", ret); - goto err_reserve_bo_failed; - } - - ret = amdgpu_ttm_alloc_gart(&bo->tbo); - if (ret) { - DRM_ERROR("Failed to bind bo to GART. ret %d\n", ret); - goto err_map_bo_gart_failed; - } - - amdgpu_bo_unreserve(bo); - bo = amdgpu_bo_ref(bo); - - return 0; - -err_map_bo_gart_failed: - amdgpu_bo_unreserve(bo); -err_reserve_bo_failed: - return ret; -} - static int mes_userq_create_wptr_mapping(struct amdgpu_device *adev, struct amdgpu_userq_mgr *uq_mgr, @@ -65,55 +37,62 @@ mes_userq_create_wptr_mapping(struct amdgpu_device *adev, uint64_t wptr) { struct amdgpu_bo_va_mapping *wptr_mapping; - struct amdgpu_vm *wptr_vm; struct amdgpu_userq_obj *wptr_obj = &queue->wptr_obj; + struct amdgpu_bo *obj; + struct amdgpu_vm *vm = queue->vm; + struct drm_exec exec; int ret; - wptr_vm = queue->vm; - ret = amdgpu_bo_reserve(wptr_vm->root.bo, false); - if (ret) - return ret; - wptr &= AMDGPU_GMC_HOLE_MASK; - wptr_mapping = amdgpu_vm_bo_lookup_mapping(wptr_vm, wptr >> PAGE_SHIFT); - amdgpu_bo_unreserve(wptr_vm->root.bo); - if (!wptr_mapping) { - DRM_ERROR("Failed to lookup wptr bo\n"); - return -EINVAL; + + drm_exec_init(&exec, DRM_EXEC_IGNORE_DUPLICATES, 2); + drm_exec_until_all_locked(&exec) { + ret = amdgpu_vm_lock_pd(vm, &exec, 1); + drm_exec_retry_on_contention(&exec); + if (unlikely(ret)) + goto fail_lock; + + wptr_mapping = amdgpu_vm_bo_lookup_mapping(vm, wptr >> PAGE_SHIFT); + if (!wptr_mapping) { + ret = -EINVAL; + goto fail_lock; + } + + obj = wptr_mapping->bo_va->base.bo; + ret = drm_exec_lock_obj(&exec, &obj->tbo.base); + drm_exec_retry_on_contention(&exec); + if (unlikely(ret)) + goto fail_lock; } - wptr_obj->obj = wptr_mapping->bo_va->base.bo; + wptr_obj->obj = amdgpu_bo_ref(wptr_mapping->bo_va->base.bo); if (wptr_obj->obj->tbo.base.size > PAGE_SIZE) { - DRM_ERROR("Requested GART mapping for wptr bo larger than one page\n"); - return -EINVAL; - } - - ret = mes_userq_map_gtt_bo_to_gart(wptr_obj->obj); - if (ret) { - DRM_ERROR("Failed to map wptr bo to GART\n"); - return ret; - } - - ret = amdgpu_bo_reserve(wptr_obj->obj, true); - if (ret) { - DRM_ERROR("Failed to reserve wptr bo\n"); - return ret; + ret = -EINVAL; + goto fail_map; } /* TODO use eviction fence instead of pinning. */ ret = amdgpu_bo_pin(wptr_obj->obj, AMDGPU_GEM_DOMAIN_GTT); if (ret) { - drm_file_err(uq_mgr->file, "[Usermode queues] Failed to pin wptr bo\n"); - goto unresv_bo; + DRM_ERROR("Failed to pin wptr bo. ret %d\n", ret); + goto fail_map; + } + + ret = amdgpu_ttm_alloc_gart(&wptr_obj->obj->tbo); + if (ret) { + DRM_ERROR("Failed to bind bo to GART. ret %d\n", ret); + goto fail_map; } queue->wptr_obj.gpu_addr = amdgpu_bo_gpu_offset(wptr_obj->obj); - amdgpu_bo_unreserve(wptr_obj->obj); + drm_exec_fini(&exec); return 0; -unresv_bo: - amdgpu_bo_unreserve(wptr_obj->obj); +fail_map: + amdgpu_bo_unref(&wptr_obj->obj); +fail_lock: + drm_exec_fini(&exec); return ret; } From 4e02e0afa95f691dc7cc17538cdd648089a843f0 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Mon, 13 Oct 2025 15:26:02 +0200 Subject: [PATCH 117/333] drm/amdgpu: nuke amdgpu_userq_fence_slab v2 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As preparation for independent fences remove the extra slab, kmalloc should do just fine. v2: use GFP_KERNEL instead of GFP_ATOMIC Signed-off-by: Christian König Reviewed-by: Prike Liang Reviewed-by: Sunil Khatri Signed-off-by: Alex Deucher (cherry picked from commit 0d831487b5be0ae59cac865a0aa87b0acc3dc717) --- drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 13 ++------- .../gpu/drm/amd/amdgpu/amdgpu_userq_fence.c | 28 +++---------------- .../gpu/drm/amd/amdgpu/amdgpu_userq_fence.h | 3 -- 3 files changed, 7 insertions(+), 37 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 46aae3fad4bf..60debd543e44 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -3149,11 +3149,7 @@ static int __init amdgpu_init(void) r = amdgpu_sync_init(); if (r) - goto error_sync; - - r = amdgpu_userq_fence_slab_init(); - if (r) - goto error_fence; + return r; amdgpu_register_atpx_handler(); amdgpu_acpi_detect(); @@ -3161,7 +3157,7 @@ static int __init amdgpu_init(void) /* Ignore KFD init failures when CONFIG_HSA_AMD is not set. */ r = amdgpu_amdkfd_init(); if (r && r != -ENOENT) - goto error_fence; + goto error_fini_sync; if (amdgpu_pp_feature_mask & PP_OVERDRIVE_MASK) { add_taint(TAINT_CPU_OUT_OF_SPEC, LOCKDEP_STILL_OK); @@ -3172,10 +3168,8 @@ static int __init amdgpu_init(void) /* let modprobe override vga console setting */ return pci_register_driver(&amdgpu_kms_pci_driver); -error_fence: +error_fini_sync: amdgpu_sync_fini(); - -error_sync: return r; } @@ -3186,7 +3180,6 @@ static void __exit amdgpu_exit(void) amdgpu_unregister_atpx_handler(); amdgpu_acpi_release(); amdgpu_sync_fini(); - amdgpu_userq_fence_slab_fini(); mmu_notifier_synchronize(); amdgpu_xcp_drv_release(); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c index da39ac862f37..e2d5f04296e1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.c @@ -32,29 +32,9 @@ #include "amdgpu.h" #include "amdgpu_userq_fence.h" -static const struct dma_fence_ops amdgpu_userq_fence_ops; -static struct kmem_cache *amdgpu_userq_fence_slab; - #define AMDGPU_USERQ_MAX_HANDLES (1U << 16) -int amdgpu_userq_fence_slab_init(void) -{ - amdgpu_userq_fence_slab = kmem_cache_create("amdgpu_userq_fence", - sizeof(struct amdgpu_userq_fence), - 0, - SLAB_HWCACHE_ALIGN, - NULL); - if (!amdgpu_userq_fence_slab) - return -ENOMEM; - - return 0; -} - -void amdgpu_userq_fence_slab_fini(void) -{ - rcu_barrier(); - kmem_cache_destroy(amdgpu_userq_fence_slab); -} +static const struct dma_fence_ops amdgpu_userq_fence_ops; static inline struct amdgpu_userq_fence *to_amdgpu_userq_fence(struct dma_fence *f) { @@ -231,7 +211,7 @@ void amdgpu_userq_fence_driver_put(struct amdgpu_userq_fence_driver *fence_drv) static int amdgpu_userq_fence_alloc(struct amdgpu_userq_fence **userq_fence) { - *userq_fence = kmem_cache_alloc(amdgpu_userq_fence_slab, GFP_ATOMIC); + *userq_fence = kmalloc(sizeof(**userq_fence), GFP_KERNEL); return *userq_fence ? 0 : -ENOMEM; } @@ -342,7 +322,7 @@ static void amdgpu_userq_fence_free(struct rcu_head *rcu) amdgpu_userq_fence_driver_put(fence_drv); kvfree(userq_fence->fence_drv_array); - kmem_cache_free(amdgpu_userq_fence_slab, userq_fence); + kfree(userq_fence); } static void amdgpu_userq_fence_release(struct dma_fence *f) @@ -545,7 +525,7 @@ int amdgpu_userq_signal_ioctl(struct drm_device *dev, void *data, r = amdgpu_userq_fence_create(queue, userq_fence, wptr, &fence); if (r) { mutex_unlock(&userq_mgr->userq_mutex); - kmem_cache_free(amdgpu_userq_fence_slab, userq_fence); + kfree(userq_fence); goto put_gobj_write; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h index d56246ad8c26..d355a0eecc07 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq_fence.h @@ -58,9 +58,6 @@ struct amdgpu_userq_fence_driver { char timeline_name[TASK_COMM_LEN]; }; -int amdgpu_userq_fence_slab_init(void); -void amdgpu_userq_fence_slab_fini(void); - void amdgpu_userq_fence_driver_get(struct amdgpu_userq_fence_driver *fence_drv); void amdgpu_userq_fence_driver_put(struct amdgpu_userq_fence_driver *fence_drv); int amdgpu_userq_fence_driver_alloc(struct amdgpu_device *adev, From 26f6654a9a60eb4d241f42a0ec85412e8821480b Mon Sep 17 00:00:00 2001 From: Osama Abdelkader Date: Thu, 23 Apr 2026 22:06:20 +0200 Subject: [PATCH 118/333] drm/exynos: remove bridge when component_add fails MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Use devm_drm_bridge_add() so the bridge is released if probe fails after registration, and drop the manual drm_bridge_remove() in remove(). Check the return value of devm_drm_bridge_add(). Signed-off-by: Osama Abdelkader Fixes: 576d72fbfb45 ("drm/exynos: mic: add a bridge at probe") Cc: stable@vger.kernel.org Reviewed-by: Raphaël Gallais-Pou Reviewed-by: Luca Ceresoli Link: https://patch.msgid.link/20260423200622.325076-2-osama.abdelkader@gmail.com Signed-off-by: Luca Ceresoli --- drivers/gpu/drm/exynos/exynos_drm_mic.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/exynos/exynos_drm_mic.c b/drivers/gpu/drm/exynos/exynos_drm_mic.c index 29a8366513fa..e68c954ec3e6 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_mic.c +++ b/drivers/gpu/drm/exynos/exynos_drm_mic.c @@ -423,7 +423,9 @@ static int exynos_mic_probe(struct platform_device *pdev) mic->bridge.of_node = dev->of_node; - drm_bridge_add(&mic->bridge); + ret = devm_drm_bridge_add(dev, &mic->bridge); + if (ret) + goto err; pm_runtime_enable(dev); @@ -443,12 +445,8 @@ static int exynos_mic_probe(struct platform_device *pdev) static void exynos_mic_remove(struct platform_device *pdev) { - struct exynos_mic *mic = platform_get_drvdata(pdev); - component_del(&pdev->dev, &exynos_mic_component_ops); pm_runtime_disable(&pdev->dev); - - drm_bridge_remove(&mic->bridge); } static const struct of_device_id exynos_mic_of_match[] = { From 644132a48f4e28a1d949d162160869286f3e75de Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Tue, 5 May 2026 08:49:48 -0400 Subject: [PATCH 119/333] selinux: prune /sys/fs/selinux/checkreqprot commit a7e4676e8e2cb ("selinux: remove the 'checkreqprot' functionality") removed the ability to modify the checkreqprot setting but left everything except the updating of the checkreqprot value intact. Aside from unnecessary processing, this could produce a local DoS from log spam and incorrectly calls selinux_ima_measure_state() on each write even though no state has changed. Prune it to just log an error message once and return count (i.e. all bytes written successfully) so that userspace never breaks. Cc: stable@vger.kernel.org Signed-off-by: Stephen Smalley Signed-off-by: Paul Moore --- security/selinux/selinuxfs.c | 47 ++++++------------------------------ 1 file changed, 7 insertions(+), 40 deletions(-) diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index 83aa765a09f9..6f74f87cb2b0 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -689,46 +689,13 @@ static ssize_t sel_read_checkreqprot(struct file *filp, char __user *buf, static ssize_t sel_write_checkreqprot(struct file *file, const char __user *buf, size_t count, loff_t *ppos) { - char *page; - ssize_t length; - unsigned int new_value; - - length = avc_has_perm(current_sid(), SECINITSID_SECURITY, - SECCLASS_SECURITY, SECURITY__SETCHECKREQPROT, - NULL); - if (length) - return length; - - if (count >= PAGE_SIZE) - return -ENOMEM; - - /* No partial writes. */ - if (*ppos != 0) - return -EINVAL; - - page = memdup_user_nul(buf, count); - if (IS_ERR(page)) - return PTR_ERR(page); - - if (sscanf(page, "%u", &new_value) != 1) { - length = -EINVAL; - goto out; - } - length = count; - - if (new_value) { - char comm[sizeof(current->comm)]; - - strscpy(comm, current->comm); - pr_err("SELinux: %s (%d) set checkreqprot to 1. This is no longer supported.\n", - comm, current->pid); - } - - selinux_ima_measure_state(); - -out: - kfree(page); - return length; + /* + * Setting checkreqprot is no longer supported, see + * https://github.com/SELinuxProject/selinux-kernel/wiki/DEPRECATE-checkreqprot + */ + pr_err_once("SELinux: %s (%d) wrote to checkreqprot. This is no longer supported.\n", + current->comm, current->pid); + return count; } static const struct file_operations sel_checkreqprot_ops = { .read = sel_read_checkreqprot, From 19cfa0099024bb9cd40f6d950caa7f47ff8e77f6 Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Tue, 5 May 2026 08:49:49 -0400 Subject: [PATCH 120/333] selinux: prune /sys/fs/selinux/disable Commit f22f9aaf6c3d ("selinux: remove the runtime disable functionality") removed the underlying SELinux runtime disable functionality but left everything else intact and started logging an error message to warn any residual users. Prune it to just log an error message once and to return count (i.e. all bytes written successfully) to avoid breaking userspace. This also fixes a local DoS from logspam. Cc: stable@vger.kernel.org Signed-off-by: Stephen Smalley Signed-off-by: Paul Moore --- security/selinux/selinuxfs.c | 36 +++++++----------------------------- 1 file changed, 7 insertions(+), 29 deletions(-) diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index 6f74f87cb2b0..343303b73d6f 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -272,35 +272,13 @@ static ssize_t sel_write_disable(struct file *file, const char __user *buf, size_t count, loff_t *ppos) { - char *page; - ssize_t length; - int new_value; - - if (count >= PAGE_SIZE) - return -ENOMEM; - - /* No partial writes. */ - if (*ppos != 0) - return -EINVAL; - - page = memdup_user_nul(buf, count); - if (IS_ERR(page)) - return PTR_ERR(page); - - if (sscanf(page, "%d", &new_value) != 1) { - length = -EINVAL; - goto out; - } - length = count; - - if (new_value) { - pr_err("SELinux: https://github.com/SELinuxProject/selinux-kernel/wiki/DEPRECATE-runtime-disable\n"); - pr_err("SELinux: Runtime disable is not supported, use selinux=0 on the kernel cmdline.\n"); - } - -out: - kfree(page); - return length; + /* + * Setting disable is no longer supported, see + * https://github.com/SELinuxProject/selinux-kernel/wiki/DEPRECATE-runtime-disable + */ + pr_err_once("SELinux: %s (%d) wrote to disable. This is no longer supported.\n", + current->comm, current->pid); + return count; } static const struct file_operations sel_disable_ops = { From ad1ac3d740cc6b858a99ab9c45c8c0574be7d1d3 Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Tue, 5 May 2026 08:49:50 -0400 Subject: [PATCH 121/333] selinux: prune /sys/fs/selinux/user Remove the previously deprecated /sys/fs/selinux/user interface aside from a residual stub for userspace compatibility. Commit d7b6918e22c7 ("selinux: Deprecate /sys/fs/selinux/user") started the deprecation process for /sys/fs/selinux/user: The selinuxfs "user" node allows userspace to request a list of security contexts that can be reached for a given SELinux user from a given starting context. This was used by libselinux when various login-style programs requested contexts for users, but libselinux stopped using it in 2020. Kernel support will be removed no sooner than Dec 2025. A pr_warn() message has been in place since Linux v6.13, and a 5 second sleep was introduced since Linux v6.17 to help make it more noticeable. We are now past the stated deadline of Dec 2025, so remove the underlying functionality and replace it with a stub that returns a '0\0' buffer to avoid breaking userspace. This also avoids a local DoS from logspam and an uninterruptible sleep delay. Cc: stable@vger.kernel.org Signed-off-by: Stephen Smalley Signed-off-by: Paul Moore --- .../{obsolete => removed}/sysfs-selinux-user | 0 security/selinux/include/security.h | 2 - security/selinux/selinuxfs.c | 68 +--------- security/selinux/ss/services.c | 125 ------------------ 4 files changed, 5 insertions(+), 190 deletions(-) rename Documentation/ABI/{obsolete => removed}/sysfs-selinux-user (100%) diff --git a/Documentation/ABI/obsolete/sysfs-selinux-user b/Documentation/ABI/removed/sysfs-selinux-user similarity index 100% rename from Documentation/ABI/obsolete/sysfs-selinux-user rename to Documentation/ABI/removed/sysfs-selinux-user diff --git a/security/selinux/include/security.h b/security/selinux/include/security.h index d1f16d7f684d..0babb8992181 100644 --- a/security/selinux/include/security.h +++ b/security/selinux/include/security.h @@ -312,8 +312,6 @@ int security_context_to_sid_default(const char *scontext, u32 scontext_len, int security_context_to_sid_force(const char *scontext, u32 scontext_len, u32 *sid); -int security_get_user_sids(u32 fromsid, const char *username, u32 **sids, u32 *nel); - int security_port_sid(u8 protocol, u16 port, u32 *out_sid); int security_ib_pkey_sid(u64 subnet_prefix, u16 pkey_num, u32 *out_sid); diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index 343303b73d6f..6ed669309132 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -1018,69 +1018,11 @@ static ssize_t sel_write_relabel(struct file *file, char *buf, size_t size) static ssize_t sel_write_user(struct file *file, char *buf, size_t size) { - char *con = NULL, *user = NULL, *ptr; - u32 sid, *sids = NULL; - ssize_t length; - char *newcon; - int rc; - u32 i, len, nsids; - - pr_warn_ratelimited("SELinux: %s (%d) wrote to /sys/fs/selinux/user!" - " This will not be supported in the future; please update your" - " userspace.\n", current->comm, current->pid); - ssleep(5); - - length = avc_has_perm(current_sid(), SECINITSID_SECURITY, - SECCLASS_SECURITY, SECURITY__COMPUTE_USER, - NULL); - if (length) - goto out; - - length = -ENOMEM; - con = kzalloc(size + 1, GFP_KERNEL); - if (!con) - goto out; - - length = -ENOMEM; - user = kzalloc(size + 1, GFP_KERNEL); - if (!user) - goto out; - - length = -EINVAL; - if (sscanf(buf, "%s %s", con, user) != 2) - goto out; - - length = security_context_str_to_sid(con, &sid, GFP_KERNEL); - if (length) - goto out; - - length = security_get_user_sids(sid, user, &sids, &nsids); - if (length) - goto out; - - length = sprintf(buf, "%u", nsids) + 1; - ptr = buf + length; - for (i = 0; i < nsids; i++) { - rc = security_sid_to_context(sids[i], &newcon, &len); - if (rc) { - length = rc; - goto out; - } - if ((length + len) >= SIMPLE_TRANSACTION_LIMIT) { - kfree(newcon); - length = -ERANGE; - goto out; - } - memcpy(ptr, newcon, len); - kfree(newcon); - ptr += len; - length += len; - } -out: - kfree(sids); - kfree(user); - kfree(con); - return length; + pr_err_once("SELinux: %s (%d) wrote to user. This is no longer supported.\n", + current->comm, current->pid); + buf[0] = '0'; + buf[1] = 0; + return 2; } static ssize_t sel_write_member(struct file *file, char *buf, size_t size) diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c index e8e7ccbd1e44..143021c5e326 100644 --- a/security/selinux/ss/services.c +++ b/security/selinux/ss/services.c @@ -2746,131 +2746,6 @@ int security_node_sid(u16 domain, return rc; } -#define SIDS_NEL 25 - -/** - * security_get_user_sids - Obtain reachable SIDs for a user. - * @fromsid: starting SID - * @username: username - * @sids: array of reachable SIDs for user - * @nel: number of elements in @sids - * - * Generate the set of SIDs for legal security contexts - * for a given user that can be reached by @fromsid. - * Set *@sids to point to a dynamically allocated - * array containing the set of SIDs. Set *@nel to the - * number of elements in the array. - */ - -int security_get_user_sids(u32 fromsid, - const char *username, - u32 **sids, - u32 *nel) -{ - struct selinux_policy *policy; - struct policydb *policydb; - struct sidtab *sidtab; - struct context *fromcon, usercon; - u32 *mysids = NULL, *mysids2, sid; - u32 i, j, mynel, maxnel = SIDS_NEL; - struct user_datum *user; - struct role_datum *role; - struct ebitmap_node *rnode, *tnode; - int rc; - - *sids = NULL; - *nel = 0; - - if (!selinux_initialized()) - return 0; - - mysids = kcalloc(maxnel, sizeof(*mysids), GFP_KERNEL); - if (!mysids) - return -ENOMEM; - -retry: - mynel = 0; - rcu_read_lock(); - policy = rcu_dereference(selinux_state.policy); - policydb = &policy->policydb; - sidtab = policy->sidtab; - - context_init(&usercon); - - rc = -EINVAL; - fromcon = sidtab_search(sidtab, fromsid); - if (!fromcon) - goto out_unlock; - - rc = -EINVAL; - user = symtab_search(&policydb->p_users, username); - if (!user) - goto out_unlock; - - usercon.user = user->value; - - ebitmap_for_each_positive_bit(&user->roles, rnode, i) { - role = policydb->role_val_to_struct[i]; - usercon.role = i + 1; - ebitmap_for_each_positive_bit(&role->types, tnode, j) { - usercon.type = j + 1; - - if (mls_setup_user_range(policydb, fromcon, user, - &usercon)) - continue; - - rc = sidtab_context_to_sid(sidtab, &usercon, &sid); - if (rc == -ESTALE) { - rcu_read_unlock(); - goto retry; - } - if (rc) - goto out_unlock; - if (mynel < maxnel) { - mysids[mynel++] = sid; - } else { - rc = -ENOMEM; - maxnel += SIDS_NEL; - mysids2 = kcalloc(maxnel, sizeof(*mysids2), GFP_ATOMIC); - if (!mysids2) - goto out_unlock; - memcpy(mysids2, mysids, mynel * sizeof(*mysids2)); - kfree(mysids); - mysids = mysids2; - mysids[mynel++] = sid; - } - } - } - rc = 0; -out_unlock: - rcu_read_unlock(); - if (rc || !mynel) { - kfree(mysids); - return rc; - } - - rc = -ENOMEM; - mysids2 = kcalloc(mynel, sizeof(*mysids2), GFP_KERNEL); - if (!mysids2) { - kfree(mysids); - return rc; - } - for (i = 0, j = 0; i < mynel; i++) { - struct av_decision dummy_avd; - rc = avc_has_perm_noaudit(fromsid, mysids[i], - SECCLASS_PROCESS, /* kernel value */ - PROCESS__TRANSITION, AVC_STRICT, - &dummy_avd); - if (!rc) - mysids2[j++] = mysids[i]; - cond_resched(); - } - kfree(mysids); - *sids = mysids2; - *nel = j; - return 0; -} - /** * __security_genfs_sid - Helper to obtain a SID for a file in a filesystem * @policy: policy From a02cd6805562305f936e807da83e253b719dd965 Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Tue, 5 May 2026 10:06:38 -0400 Subject: [PATCH 122/333] selinux: allow multiple opens of /sys/fs/selinux/policy Currently there can only be a single open of /sys/fs/selinux/policy at any time. This allows any process to block any other process from reading the kernel policy. The original motivation seems to have been a mix of preventing an inconsistent view of the policy size and preventing userspace from allocating kernel memory without bound, but this is arguably equally bad. Eliminate the policy_opened flag and shrink the critical section that the policy mutex is held. While we are making changes here, drop a couple of extraneous BUG_ONs. Cc: stable@vger.kernel.org Link: https://lore.kernel.org/selinux/20100726193414.19538.64028.stgit@paris.rdu.redhat.com/ Signed-off-by: Stephen Smalley Signed-off-by: Paul Moore --- security/selinux/selinuxfs.c | 27 ++++----------------------- 1 file changed, 4 insertions(+), 23 deletions(-) diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index 6ed669309132..a43a38a3ae25 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -76,7 +76,6 @@ struct selinux_fs_info { int *bool_pending_values; struct dentry *class_dir; unsigned long last_class_ino; - bool policy_opened; unsigned long last_ino; struct super_block *sb; }; @@ -340,44 +339,31 @@ struct policy_load_memory { static int sel_open_policy(struct inode *inode, struct file *filp) { - struct selinux_fs_info *fsi = inode->i_sb->s_fs_info; struct policy_load_memory *plm = NULL; int rc; - BUG_ON(filp->private_data); - - mutex_lock(&selinux_state.policy_mutex); - rc = avc_has_perm(current_sid(), SECINITSID_SECURITY, SECCLASS_SECURITY, SECURITY__READ_POLICY, NULL); if (rc) - goto err; + return rc; - rc = -EBUSY; - if (fsi->policy_opened) - goto err; - - rc = -ENOMEM; plm = kzalloc_obj(*plm); if (!plm) - goto err; + return -ENOMEM; + mutex_lock(&selinux_state.policy_mutex); rc = security_read_policy(&plm->data, &plm->len); if (rc) goto err; - if ((size_t)i_size_read(inode) != plm->len) { inode_lock(inode); i_size_write(inode, plm->len); inode_unlock(inode); } - - fsi->policy_opened = 1; + mutex_unlock(&selinux_state.policy_mutex); filp->private_data = plm; - mutex_unlock(&selinux_state.policy_mutex); - return 0; err: mutex_unlock(&selinux_state.policy_mutex); @@ -390,13 +376,8 @@ static int sel_open_policy(struct inode *inode, struct file *filp) static int sel_release_policy(struct inode *inode, struct file *filp) { - struct selinux_fs_info *fsi = inode->i_sb->s_fs_info; struct policy_load_memory *plm = filp->private_data; - BUG_ON(!plm); - - fsi->policy_opened = 0; - vfree(plm->data); kfree(plm); From 868f31e4061eca8c3cd607d79d954d5e54f204aa Mon Sep 17 00:00:00 2001 From: Stephen Smalley Date: Thu, 30 Apr 2026 14:36:52 -0400 Subject: [PATCH 123/333] selinux: shrink critical section in sel_write_load() Currently sel_write_load() takes the policy mutex earlier than necessary. Move the taking of the mutex later. This avoids holding it unnecessarily across the vmalloc() and copy_from_user() of the policy data. Cc: stable@vger.kernel.org Signed-off-by: Stephen Smalley Signed-off-by: Paul Moore --- security/selinux/selinuxfs.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) diff --git a/security/selinux/selinuxfs.c b/security/selinux/selinuxfs.c index a43a38a3ae25..25ca7d714014 100644 --- a/security/selinux/selinuxfs.c +++ b/security/selinux/selinuxfs.c @@ -553,34 +553,31 @@ static ssize_t sel_write_load(struct file *file, const char __user *buf, if (!count) return -EINVAL; - mutex_lock(&selinux_state.policy_mutex); - length = avc_has_perm(current_sid(), SECINITSID_SECURITY, SECCLASS_SECURITY, SECURITY__LOAD_POLICY, NULL); if (length) - goto out; + return length; data = vmalloc(count); - if (!data) { - length = -ENOMEM; - goto out; - } + if (!data) + return -ENOMEM; if (copy_from_user(data, buf, count) != 0) { length = -EFAULT; goto out; } + mutex_lock(&selinux_state.policy_mutex); length = security_load_policy(data, count, &load_state); if (length) { pr_warn_ratelimited("SELinux: failed to load policy\n"); - goto out; + goto out_unlock; } fsi = file_inode(file)->i_sb->s_fs_info; length = sel_make_policy_nodes(fsi, load_state.policy); if (length) { pr_warn_ratelimited("SELinux: failed to initialize selinuxfs\n"); selinux_policy_cancel(&load_state); - goto out; + goto out_unlock; } selinux_policy_commit(&load_state); @@ -590,8 +587,9 @@ static ssize_t sel_write_load(struct file *file, const char __user *buf, from_kuid(&init_user_ns, audit_get_loginuid(current)), audit_get_sessionid(current)); -out: +out_unlock: mutex_unlock(&selinux_state.policy_mutex); +out: vfree(data); return length; } From 60a1e131a811b68703da58fd805ab359b704ab03 Mon Sep 17 00:00:00 2001 From: Gustavo Sousa Date: Thu, 16 Apr 2026 15:17:19 -0300 Subject: [PATCH 124/333] drm/xe/hdcp: Add NULL check for media_gt in intel_hdcp_gsc_check_status() When media GT is disabled via configfs, there is no allocation for media_gt, which is kept as NULL. In such scenario, intel_hdcp_gsc_check_status() results in a kernel pagefault error due to >->uc.gsc being evaluated as an invalid memory address. Fix that by introducing a NULL check on media_gt and bailing out early if so. While at it, also drop the NULL check for gsc, since it can't be NULL if media_gt is not NULL. v2: - Get address for gsc only after checking that gt is not NULL. (Shuicheng) - Drop the NULL check for gsc. (Shuicheng) v3: - Add "Fixes" and "Cc: " tags. (Matt) Fixes: 4af50beb4e0f ("drm/xe: Use gsc_proxy_init_done to check proxy status") Cc: # v6.10+ Reviewed-by: Matt Roper Reviewed-by: Shuicheng Lin Link: https://patch.msgid.link/20260416-check-for-null-media_gt-in-intel_hdcp_gsc_check_status-v2-1-9adb9fd3b621@intel.com Signed-off-by: Gustavo Sousa (cherry picked from commit bfaf87e84ca3ca3f6e275f9ae56da47a8b55ffd1) Signed-off-by: Matthew Brost --- drivers/gpu/drm/xe/display/xe_hdcp_gsc.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c index 29c72aa4b0d2..33494b86205d 100644 --- a/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c +++ b/drivers/gpu/drm/xe/display/xe_hdcp_gsc.c @@ -37,9 +37,17 @@ static bool intel_hdcp_gsc_check_status(struct drm_device *drm) struct xe_device *xe = to_xe_device(drm); struct xe_tile *tile = xe_device_get_root_tile(xe); struct xe_gt *gt = tile->media_gt; - struct xe_gsc *gsc = >->uc.gsc; + struct xe_gsc *gsc; - if (!gsc || !xe_uc_fw_is_available(&gsc->fw)) { + if (!gt) { + drm_dbg_kms(&xe->drm, + "not checking GSC status for HDCP2.x: media GT not present or disabled\n"); + return false; + } + + gsc = >->uc.gsc; + + if (!xe_uc_fw_is_available(&gsc->fw)) { drm_dbg_kms(&xe->drm, "GSC Components not ready for HDCP2.x\n"); return false; From d01012c740bbb298b957e30cc0848e482c6f486f Mon Sep 17 00:00:00 2001 From: Shuicheng Lin Date: Tue, 28 Apr 2026 20:14:48 +0000 Subject: [PATCH 125/333] drm/xe/pf: Fix EAGAIN sign in pf_migration_consume() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit PTR_ERR() returns a negative value, so comparing against the positive EAGAIN is always true for ERR_PTR(-EAGAIN), causing pf_migration_consume() to bail out instead of continuing to the remaining GTs. On multi-GT platforms this can skip GTs that already have data ready. Compare against -EAGAIN to match the intent (and the following line that correctly uses -EAGAIN). While at it, gate PTR_ERR() with IS_ERR(). v2: add IS_ERR() guard before PTR_ERR(). (Gustavo) Fixes: 67df4a5cbc58 ("drm/xe/pf: Add data structures and handlers for migration rings") Cc: Michał Winiarski Reviewed-by: Gustavo Sousa Link: https://patch.msgid.link/20260428201448.3999428-1-shuicheng.lin@intel.com Signed-off-by: Shuicheng Lin (cherry picked from commit 9d770e72e1edb54beacfce5f402edb51632811e3) Signed-off-by: Matthew Brost --- drivers/gpu/drm/xe/xe_sriov_pf_migration.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_sriov_pf_migration.c b/drivers/gpu/drm/xe/xe_sriov_pf_migration.c index 6c4b16409cc9..150a241110fb 100644 --- a/drivers/gpu/drm/xe/xe_sriov_pf_migration.c +++ b/drivers/gpu/drm/xe/xe_sriov_pf_migration.c @@ -149,10 +149,11 @@ pf_migration_consume(struct xe_device *xe, unsigned int vfid) for_each_gt(gt, xe, gt_id) { data = xe_gt_sriov_pf_migration_save_consume(gt, vfid); - if (data && PTR_ERR(data) != EAGAIN) + if (!data) + continue; + if (!IS_ERR(data) || PTR_ERR(data) != -EAGAIN) return data; - if (PTR_ERR(data) == -EAGAIN) - more_data = true; + more_data = true; } if (!more_data) From b87951a0ae9f95ca6590bf0939edced7d36929dd Mon Sep 17 00:00:00 2001 From: Shuicheng Lin Date: Wed, 29 Apr 2026 19:22:59 +0000 Subject: [PATCH 126/333] drm/xe/pf: Fix MMIO access using PF view instead of VF view during migration MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit pf_migration_mmio_save() and pf_migration_mmio_restore() initialize a local VF-specific MMIO view via xe_mmio_init_vf_view() but then pass >->mmio (the PF base) to all xe_mmio_read32()/xe_mmio_write32() calls instead of the local &mmio. This causes the PF own SW flag registers to be saved/restored rather than the target VF registers, silently corrupting migration state. Use the VF MMIO view for all register accesses, matching the correct pattern used in pf_clear_vf_scratch_regs(). Fixes: b7c1b990f719 ("drm/xe/pf: Handle MMIO migration data as part of PF control") Cc: Michał Winiarski Assisted-by: Claude:claude-opus-4.6 Reviewed-by: Michal Wajdeczko Reviewed-by: Stuart Summers Link: https://patch.msgid.link/20260429192259.4009211-1-shuicheng.lin@intel.com Signed-off-by: Shuicheng Lin (cherry picked from commit 7d9c39cfb31ff389490ca1308767c2807a9829a6) Signed-off-by: Matthew Brost --- drivers/gpu/drm/xe/xe_gt_sriov_pf_migration.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_migration.c b/drivers/gpu/drm/xe/xe_gt_sriov_pf_migration.c index 87a164efcc33..01fe03b9efe8 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_migration.c +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_migration.c @@ -385,10 +385,10 @@ static int pf_migration_mmio_save(struct xe_gt *gt, unsigned int vfid, void *buf if (xe_gt_is_media_type(gt)) for (n = 0; n < MED_VF_SW_FLAG_COUNT; n++) - regs[n] = xe_mmio_read32(>->mmio, MED_VF_SW_FLAG(n)); + regs[n] = xe_mmio_read32(&mmio, MED_VF_SW_FLAG(n)); else for (n = 0; n < VF_SW_FLAG_COUNT; n++) - regs[n] = xe_mmio_read32(>->mmio, VF_SW_FLAG(n)); + regs[n] = xe_mmio_read32(&mmio, VF_SW_FLAG(n)); return 0; } @@ -407,10 +407,10 @@ static int pf_migration_mmio_restore(struct xe_gt *gt, unsigned int vfid, if (xe_gt_is_media_type(gt)) for (n = 0; n < MED_VF_SW_FLAG_COUNT; n++) - xe_mmio_write32(>->mmio, MED_VF_SW_FLAG(n), regs[n]); + xe_mmio_write32(&mmio, MED_VF_SW_FLAG(n), regs[n]); else for (n = 0; n < VF_SW_FLAG_COUNT; n++) - xe_mmio_write32(>->mmio, VF_SW_FLAG(n), regs[n]); + xe_mmio_write32(&mmio, VF_SW_FLAG(n), regs[n]); return 0; } From b29987dfd943e655df6e3b641ecffad5cc1509c2 Mon Sep 17 00:00:00 2001 From: Satyanarayana K V P Date: Mon, 4 May 2026 09:49:26 +0000 Subject: [PATCH 127/333] drm/xe/guc: Exclude indirect ring state page from ADS engine state size The engine state size reported to GuC via ADS should only include the engine state portion and should not include the indirect ring state page that comes after it in the context image. The GuC uses this size to overwrite the engine state in the LRC on watchdog resets and we don't want it to overwrite the indirect ring state as well. Fixes: d6219e1cd5e3 ("drm/xe: Add Indirect Ring State support") Suggested-by: Daniele Ceraolo Spurio Signed-off-by: Satyanarayana K V P Cc: Michal Wajdeczko Cc: Matthew Brost Reviewed-by: Matthew Brost Reviewed-by: Daniele Ceraolo Spurio Signed-off-by: Daniele Ceraolo Spurio Link: https://patch.msgid.link/20260504094924.3760713-4-satyanarayana.k.v.p@intel.com (cherry picked from commit 3ec5f003f6c377beda8bd5438941f5a7795e1848) Signed-off-by: Matthew Brost --- drivers/gpu/drm/xe/xe_guc_ads.c | 5 +---- drivers/gpu/drm/xe/xe_lrc.c | 11 +++++++++-- drivers/gpu/drm/xe/xe_lrc.h | 2 +- 3 files changed, 11 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_guc_ads.c b/drivers/gpu/drm/xe/xe_guc_ads.c index 81b5f01b1f65..2b835d48b565 100644 --- a/drivers/gpu/drm/xe/xe_guc_ads.c +++ b/drivers/gpu/drm/xe/xe_guc_ads.c @@ -512,12 +512,9 @@ static void guc_golden_lrc_init(struct xe_guc_ads *ads) * that starts after the execlists LRC registers. This is * required to allow the GuC to restore just the engine state * when a watchdog reset occurs. - * We calculate the engine state size by removing the size of - * what comes before it in the context image (which is identical - * on all engines). */ ads_blob_write(ads, ads.eng_state_size[guc_class], - real_size - xe_lrc_skip_size(xe)); + xe_lrc_engine_state_size(gt, class)); ads_blob_write(ads, ads.golden_context_lrca[guc_class], addr_ggtt); diff --git a/drivers/gpu/drm/xe/xe_lrc.c b/drivers/gpu/drm/xe/xe_lrc.c index c725cde4508d..4af9f0d7c6f3 100644 --- a/drivers/gpu/drm/xe/xe_lrc.c +++ b/drivers/gpu/drm/xe/xe_lrc.c @@ -746,9 +746,16 @@ size_t xe_lrc_reg_size(struct xe_device *xe) return 80 * sizeof(u32); } -size_t xe_lrc_skip_size(struct xe_device *xe) +/** + * xe_lrc_engine_state_size() - Get size of the engine state within LRC + * @gt: the &xe_gt struct instance + * @class: Hardware engine class + * + * Returns: Size of the engine state + */ +size_t xe_lrc_engine_state_size(struct xe_gt *gt, enum xe_engine_class class) { - return LRC_PPHWSP_SIZE + xe_lrc_reg_size(xe); + return xe_gt_lrc_hang_replay_size(gt, class) - xe_lrc_reg_size(gt_to_xe(gt)); } static inline u32 __xe_lrc_seqno_offset(struct xe_lrc *lrc) diff --git a/drivers/gpu/drm/xe/xe_lrc.h b/drivers/gpu/drm/xe/xe_lrc.h index e7c975f9e2d9..5440663183f6 100644 --- a/drivers/gpu/drm/xe/xe_lrc.h +++ b/drivers/gpu/drm/xe/xe_lrc.h @@ -130,7 +130,7 @@ u32 xe_lrc_parallel_ggtt_addr(struct xe_lrc *lrc); struct iosys_map xe_lrc_parallel_map(struct xe_lrc *lrc); size_t xe_lrc_reg_size(struct xe_device *xe); -size_t xe_lrc_skip_size(struct xe_device *xe); +size_t xe_lrc_engine_state_size(struct xe_gt *gt, enum xe_engine_class class); void xe_lrc_dump_default(struct drm_printer *p, struct xe_gt *gt, From 54d54f33813d7911555226b4220737177a2ba8d6 Mon Sep 17 00:00:00 2001 From: Athira Rajeev Date: Sat, 14 Mar 2026 18:54:30 +0530 Subject: [PATCH 128/333] powerpc/pseries/htmdump: Free the global buffers in htmdump module exit htmdump modules uses global memory buffers to capture details like capabilities, status of specified HTM, read the trace buffer. These are initialized during module init and hence needs to be freed in module exit. Patch adds freeing of the memory in module exit. The change also includes minor clean up for the variable name. The read call back for the debugfs interface file saves filp->private_data to local variable name which is same as global variable name for the memory buffers. Rename these local variable names. Signed-off-by: Athira Rajeev Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260314132432.25581-1-atrajeev@linux.ibm.com --- arch/powerpc/platforms/pseries/htmdump.c | 31 +++++++++++++----------- 1 file changed, 17 insertions(+), 14 deletions(-) diff --git a/arch/powerpc/platforms/pseries/htmdump.c b/arch/powerpc/platforms/pseries/htmdump.c index 742ec52c9d4d..93f0cc2dc7fb 100644 --- a/arch/powerpc/platforms/pseries/htmdump.c +++ b/arch/powerpc/platforms/pseries/htmdump.c @@ -86,7 +86,7 @@ static ssize_t htm_return_check(long rc) static ssize_t htmdump_read(struct file *filp, char __user *ubuf, size_t count, loff_t *ppos) { - void *htm_buf = filp->private_data; + void *htm_buf_data = filp->private_data; unsigned long page, read_size, available; loff_t offset; long rc, ret; @@ -100,7 +100,7 @@ static ssize_t htmdump_read(struct file *filp, char __user *ubuf, * - last three values are address, size and offset */ rc = htm_hcall_wrapper(htmflags, nodeindex, nodalchipindex, coreindexonchip, - htmtype, H_HTM_OP_DUMP_DATA, virt_to_phys(htm_buf), + htmtype, H_HTM_OP_DUMP_DATA, virt_to_phys(htm_buf_data), PAGE_SIZE, page); ret = htm_return_check(rc); @@ -112,7 +112,7 @@ static ssize_t htmdump_read(struct file *filp, char __user *ubuf, available = PAGE_SIZE; read_size = min(count, available); *ppos += read_size; - return simple_read_from_buffer(ubuf, count, &offset, htm_buf, available); + return simple_read_from_buffer(ubuf, count, &offset, htm_buf_data, available); } static const struct file_operations htmdump_fops = { @@ -226,7 +226,7 @@ static int htmstart_get(void *data, u64 *val) static ssize_t htmstatus_read(struct file *filp, char __user *ubuf, size_t count, loff_t *ppos) { - void *htm_status_buf = filp->private_data; + void *htm_status_data = filp->private_data; long rc, ret; u64 *num_entries; u64 to_copy; @@ -238,7 +238,7 @@ static ssize_t htmstatus_read(struct file *filp, char __user *ubuf, * - last three values as addr, size and offset */ rc = htm_hcall_wrapper(htmflags, nodeindex, nodalchipindex, coreindexonchip, - htmtype, H_HTM_OP_STATUS, virt_to_phys(htm_status_buf), + htmtype, H_HTM_OP_STATUS, virt_to_phys(htm_status_data), PAGE_SIZE, 0); ret = htm_return_check(rc); @@ -255,13 +255,13 @@ static ssize_t htmstatus_read(struct file *filp, char __user *ubuf, * So total count to copy is: * 32 bytes (for first 7 fields) + (number of HTM entries * entry size) */ - num_entries = htm_status_buf + 0x10; + num_entries = htm_status_data + 0x10; if (htmtype == 0x2) htmstatus_flag = 0x8; else htmstatus_flag = 0x6; to_copy = 32 + (be64_to_cpu(*num_entries) * htmstatus_flag); - return simple_read_from_buffer(ubuf, count, ppos, htm_status_buf, to_copy); + return simple_read_from_buffer(ubuf, count, ppos, htm_status_data, to_copy); } static const struct file_operations htmstatus_fops = { @@ -273,7 +273,7 @@ static const struct file_operations htmstatus_fops = { static ssize_t htminfo_read(struct file *filp, char __user *ubuf, size_t count, loff_t *ppos) { - void *htm_info_buf = filp->private_data; + void *htm_info_data = filp->private_data; long rc, ret; u64 *num_entries; u64 to_copy; @@ -284,7 +284,7 @@ static ssize_t htminfo_read(struct file *filp, char __user *ubuf, * - last three values as addr, size and offset */ rc = htm_hcall_wrapper(htmflags, nodeindex, nodalchipindex, coreindexonchip, - htmtype, H_HTM_OP_DUMP_SYSPROC_CONF, virt_to_phys(htm_info_buf), + htmtype, H_HTM_OP_DUMP_SYSPROC_CONF, virt_to_phys(htm_info_data), PAGE_SIZE, 0); ret = htm_return_check(rc); @@ -301,15 +301,15 @@ static ssize_t htminfo_read(struct file *filp, char __user *ubuf, * So total count to copy is: * 32 bytes (for first 5 fields) + (number of HTM entries * entry size) */ - num_entries = htm_info_buf + 0x10; + num_entries = htm_info_data + 0x10; to_copy = 32 + (be64_to_cpu(*num_entries) * 16); - return simple_read_from_buffer(ubuf, count, ppos, htm_info_buf, to_copy); + return simple_read_from_buffer(ubuf, count, ppos, htm_info_data, to_copy); } static ssize_t htmcaps_read(struct file *filp, char __user *ubuf, size_t count, loff_t *ppos) { - void *htm_caps_buf = filp->private_data; + void *htm_caps_data = filp->private_data; long rc, ret; /* @@ -319,7 +319,7 @@ static ssize_t htmcaps_read(struct file *filp, char __user *ubuf, * and zero */ rc = htm_hcall_wrapper(htmflags, nodeindex, nodalchipindex, coreindexonchip, - htmtype, H_HTM_OP_CAPABILITIES, virt_to_phys(htm_caps_buf), + htmtype, H_HTM_OP_CAPABILITIES, virt_to_phys(htm_caps_data), 0x80, 0); ret = htm_return_check(rc); @@ -328,7 +328,7 @@ static ssize_t htmcaps_read(struct file *filp, char __user *ubuf, return ret; } - return simple_read_from_buffer(ubuf, count, ppos, htm_caps_buf, 0x80); + return simple_read_from_buffer(ubuf, count, ppos, htm_caps_data, 0x80); } static const struct file_operations htminfo_fops = { @@ -482,6 +482,9 @@ static void __exit htmdump_exit(void) { debugfs_remove_recursive(htmdump_debugfs_dir); kfree(htm_buf); + kfree(htm_status_buf); + kfree(htm_info_buf); + kfree(htm_caps_buf); } module_init(htmdump_init); From 4ec6ade73f1626a74d46e61ff247cf2b1b96446b Mon Sep 17 00:00:00 2001 From: Athira Rajeev Date: Sat, 14 Mar 2026 18:54:31 +0530 Subject: [PATCH 129/333] powerpc/pseries/htmdump: Fix the offset value used in processor configuration dump H_HTM call is invoked using three parameters specifying the address of the buffer, size of the buffer and offset where to read from. offset used was always zero. "offset" is value from output buffer header that points to next entry to dump. zero is the first entry to dump. next entry is read from the output bufferbyte offset 0x8. Update htminfo_read() function to use right offset. Return when offset points to -1 Fixes: dea7384e14e7 ("powerpc/pseries/htmdump: Add htm info support to htmdump module") Signed-off-by: Athira Rajeev Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260314132432.25581-2-atrajeev@linux.ibm.com --- arch/powerpc/platforms/pseries/htmdump.c | 17 +++++++++++++++-- 1 file changed, 15 insertions(+), 2 deletions(-) diff --git a/arch/powerpc/platforms/pseries/htmdump.c b/arch/powerpc/platforms/pseries/htmdump.c index 93f0cc2dc7fb..34978a794eba 100644 --- a/arch/powerpc/platforms/pseries/htmdump.c +++ b/arch/powerpc/platforms/pseries/htmdump.c @@ -277,15 +277,26 @@ static ssize_t htminfo_read(struct file *filp, char __user *ubuf, long rc, ret; u64 *num_entries; u64 to_copy; + loff_t offset = 0; + u64 info_offset = 0; /* * Invoke H_HTM call with: * - operation as htm status (H_HTM_OP_STATUS) * - last three values as addr, size and offset + * "offset" is value from output buffer header + * that points to next entry to dump. 0 is the first + * entry to dump. next entry is read from the output + * bufferbyte offset 0x8. */ + if (*ppos) { + info_offset = *(u64 *)(htm_info_data + 0x8); + if (info_offset == -1) + return 0; + } rc = htm_hcall_wrapper(htmflags, nodeindex, nodalchipindex, coreindexonchip, htmtype, H_HTM_OP_DUMP_SYSPROC_CONF, virt_to_phys(htm_info_data), - PAGE_SIZE, 0); + PAGE_SIZE, be64_to_cpu(info_offset)); ret = htm_return_check(rc); if (ret <= 0) { @@ -303,7 +314,9 @@ static ssize_t htminfo_read(struct file *filp, char __user *ubuf, */ num_entries = htm_info_data + 0x10; to_copy = 32 + (be64_to_cpu(*num_entries) * 16); - return simple_read_from_buffer(ubuf, count, ppos, htm_info_data, to_copy); + + *ppos += to_copy; + return simple_read_from_buffer(ubuf, count, &offset, htm_info_data, to_copy); } static ssize_t htmcaps_read(struct file *filp, char __user *ubuf, From 0031424cbc0a6eafc56a8f9201a4a69d3649981e Mon Sep 17 00:00:00 2001 From: Athira Rajeev Date: Sat, 14 Mar 2026 18:54:32 +0530 Subject: [PATCH 130/333] powerpc/pseries/htmdump: Fix the offset value used in htm status dump H_HTM call is invoked using three parameters specifying the address of the buffer, size of the buffer and offset where to read from. offset used was always zero. "offset" is value from output buffer header that points to next entry to dump. zero is the first entry to dump. next entry is read from the output bufferbyte offset 0x8. Update htmstatus_read() function to use right offset. Return when offset points to -1 Fixes: 627cf584f4c3 ("powerpc/pseries/htmdump: Add htm status support to htmdump module") Signed-off-by: Athira Rajeev Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260314132432.25581-3-atrajeev@linux.ibm.com --- arch/powerpc/platforms/pseries/htmdump.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/arch/powerpc/platforms/pseries/htmdump.c b/arch/powerpc/platforms/pseries/htmdump.c index 34978a794eba..76444c6c5cc1 100644 --- a/arch/powerpc/platforms/pseries/htmdump.c +++ b/arch/powerpc/platforms/pseries/htmdump.c @@ -231,15 +231,26 @@ static ssize_t htmstatus_read(struct file *filp, char __user *ubuf, u64 *num_entries; u64 to_copy; int htmstatus_flag; + loff_t offset = 0; + u64 status_offset = 0; /* * Invoke H_HTM call with: * - operation as htm status (H_HTM_OP_STATUS) - * - last three values as addr, size and offset + * - last three values as addr, size and offset. + * "offset" is value from output buffer header + * that points to next entry to dump. 0 is the first + * entry to dump. next entry is read from the output + * bufferbyte offset 0x8. */ + if (*ppos) { + status_offset = *(u64 *)(htm_status_data + 0x8); + if (status_offset == -1) + return 0; + } rc = htm_hcall_wrapper(htmflags, nodeindex, nodalchipindex, coreindexonchip, htmtype, H_HTM_OP_STATUS, virt_to_phys(htm_status_data), - PAGE_SIZE, 0); + PAGE_SIZE, be64_to_cpu(status_offset)); ret = htm_return_check(rc); if (ret <= 0) { @@ -261,7 +272,9 @@ static ssize_t htmstatus_read(struct file *filp, char __user *ubuf, else htmstatus_flag = 0x6; to_copy = 32 + (be64_to_cpu(*num_entries) * htmstatus_flag); - return simple_read_from_buffer(ubuf, count, ppos, htm_status_data, to_copy); + *ppos += to_copy; + + return simple_read_from_buffer(ubuf, count, &offset, htm_status_data, to_copy); } static const struct file_operations htmstatus_fops = { From 0342545a29bc2edfbe132c68c68b51c92a12444c Mon Sep 17 00:00:00 2001 From: Athira Rajeev Date: Sat, 14 Mar 2026 18:59:53 +0530 Subject: [PATCH 131/333] powerpc/pseries/htmdump: Add memory configuration dump support to htmdump module H_HTM (Hardware Trace Macro) hypervisor call has capability to capture SystemMemory Configuration. This information helps to understand the address mapping for the partitions in the system. Support dumping system memory configuration from Hardware Trace Macro (HTM) function via debugfs interface. Under debugfs folder "/sys/kernel/debug/powerpc/htmdump", add file "htmsystem_mem". The interface allows only read of this file which will present the content of HTM buffer from the hcall. The 16th offset of HTM buffer has value for the number of entries for array of processors. Use this information to copy data to the debugfs file Signed-off-by: Athira Rajeev Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260314132953.27269-1-atrajeev@linux.ibm.com --- arch/powerpc/platforms/pseries/htmdump.c | 70 ++++++++++++++++++++++++ 1 file changed, 70 insertions(+) diff --git a/arch/powerpc/platforms/pseries/htmdump.c b/arch/powerpc/platforms/pseries/htmdump.c index 76444c6c5cc1..489a80e87082 100644 --- a/arch/powerpc/platforms/pseries/htmdump.c +++ b/arch/powerpc/platforms/pseries/htmdump.c @@ -16,6 +16,7 @@ static void *htm_buf; static void *htm_status_buf; static void *htm_info_buf; static void *htm_caps_buf; +static void *htm_mem_buf; static u32 nodeindex; static u32 nodalchipindex; static u32 coreindexonchip; @@ -115,12 +116,72 @@ static ssize_t htmdump_read(struct file *filp, char __user *ubuf, return simple_read_from_buffer(ubuf, count, &offset, htm_buf_data, available); } +static ssize_t htmsystem_mem_read(struct file *filp, char __user *ubuf, + size_t count, loff_t *ppos) +{ + void *htm_mem_data = filp->private_data; + long rc, ret; + u64 *num_entries; + u64 to_copy = 0; + loff_t offset = 0; + u64 mem_offset = 0; + + /* + * Invoke H_HTM call with: + * - operation as htm status (H_HTM_OP_STATUS) + * - last three values as addr, size and offset. "offset" + * is value from output buffer header that points to next + * entry to dump. 0 is the first entry to dump. next entry + * is read from the output bufferbyte offset 0x8. + * + * When first time hcall is invoked, mem_offset should be + * zero because zero is the first entry. + * In the next hcall, offset of next entry to read from is + * picked from output buffer header itself. So don't fill + * mem_offset for first read. + * + * If there is no further data to read in next iteration, + * offset value from output buffer header will point to -1. + */ + if (*ppos) { + mem_offset = *(u64 *)(htm_mem_data + 0x8); + if (mem_offset == -1) + return 0; + } + rc = htm_hcall_wrapper(htmflags, nodeindex, nodalchipindex, coreindexonchip, + htmtype, H_HTM_OP_DUMP_SYSMEM_CONF, virt_to_phys(htm_mem_data), + PAGE_SIZE, be64_to_cpu(mem_offset)); + ret = htm_return_check(rc); + if (ret <= 0) { + pr_debug("H_HTM hcall returned for op: H_HTM_OP_DUMP_SYSMEM_CONF with hcall returning %ld\n", ret); + return ret; + } + + /* + * HTM system mem buffer, start of buffer + 0x10 gives the + * number of HTM entries in the buffer. + * So total count to copy is: + * 32 bytes (for first 5 fields) + (number of HTM entries * entry size) + */ + num_entries = htm_mem_data + 0x10; + to_copy = 32 + (be64_to_cpu(*num_entries) * 32); + + *ppos += to_copy; + return simple_read_from_buffer(ubuf, count, &offset, htm_mem_data, to_copy); +} + static const struct file_operations htmdump_fops = { .llseek = NULL, .read = htmdump_read, .open = simple_open, }; +static const struct file_operations htmsystem_mem_fops = { + .llseek = NULL, + .read = htmsystem_mem_read, + .open = simple_open, +}; + static int htmconfigure_set(void *data, u64 val) { long rc, ret; @@ -483,9 +544,17 @@ static int htmdump_init_debugfs(void) return -ENOMEM; } + /* Memory to present HTM system memory configuration */ + htm_mem_buf = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!htm_mem_buf) { + pr_err("Failed to allocate htm mem buf\n"); + return -ENOMEM; + } + debugfs_create_file("htmstatus", 0400, htmdump_debugfs_dir, htm_status_buf, &htmstatus_fops); debugfs_create_file("htminfo", 0400, htmdump_debugfs_dir, htm_info_buf, &htminfo_fops); debugfs_create_file("htmcaps", 0400, htmdump_debugfs_dir, htm_caps_buf, &htmcaps_fops); + debugfs_create_file("htmsystem_mem", 0400, htmdump_debugfs_dir, htm_mem_buf, &htmsystem_mem_fops); return 0; } @@ -511,6 +580,7 @@ static void __exit htmdump_exit(void) kfree(htm_status_buf); kfree(htm_info_buf); kfree(htm_caps_buf); + kfree(htm_mem_buf); } module_init(htmdump_init); From 7a4f0846ee6cc8cf44ae0046ed42e3259d1dd45b Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:40 +0530 Subject: [PATCH 132/333] pseries/papr-hvpipe: Fix race with interrupt handler While executing ->ioctl handler or ->release handler, if an interrupt fires on the same cpu, then we can enter into a deadlock. This patch fixes both these handlers to take spin_lock_irq{save|restore} versions of the lock to prevent this deadlock. Cc: stable@vger.kernel.org Fixes: 814ef095f12c9 ("powerpc/pseries: Add papr-hvpipe char driver for HVPIPE interfaces") Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/e4ed435c44fc191f2eb23c7907ba6f72f193e6aa.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index 14ae480d060a..c41d45e1986d 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -444,13 +444,14 @@ static int papr_hvpipe_handle_release(struct inode *inode, struct file *file) { struct hvpipe_source_info *src_info; + unsigned long flags; /* * Hold the lock, remove source from src_list, reset the * hvpipe status and release the lock to prevent any race * with message event IRQ. */ - spin_lock(&hvpipe_src_list_lock); + spin_lock_irqsave(&hvpipe_src_list_lock, flags); src_info = file->private_data; list_del(&src_info->list); file->private_data = NULL; @@ -461,10 +462,10 @@ static int papr_hvpipe_handle_release(struct inode *inode, */ if (src_info->hvpipe_status & HVPIPE_MSG_AVAILABLE) { src_info->hvpipe_status = 0; - spin_unlock(&hvpipe_src_list_lock); + spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); hvpipe_rtas_recv_msg(NULL, 0); } else - spin_unlock(&hvpipe_src_list_lock); + spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); kfree(src_info); return 0; @@ -480,20 +481,21 @@ static const struct file_operations papr_hvpipe_handle_ops = { static int papr_hvpipe_dev_create_handle(u32 srcID) { struct hvpipe_source_info *src_info __free(kfree) = NULL; + unsigned long flags; - spin_lock(&hvpipe_src_list_lock); + spin_lock_irqsave(&hvpipe_src_list_lock, flags); /* * Do not allow more than one process communicates with * each source. */ src_info = hvpipe_find_source(srcID); if (src_info) { - spin_unlock(&hvpipe_src_list_lock); + spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); pr_err("pid(%d) is already using the source(%d)\n", src_info->tsk->pid, srcID); return -EALREADY; } - spin_unlock(&hvpipe_src_list_lock); + spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); src_info = kzalloc_obj(*src_info, GFP_KERNEL_ACCOUNT); if (!src_info) @@ -510,18 +512,18 @@ static int papr_hvpipe_dev_create_handle(u32 srcID) return fdf.err; retain_and_null_ptr(src_info); - spin_lock(&hvpipe_src_list_lock); + spin_lock_irqsave(&hvpipe_src_list_lock, flags); /* * If two processes are executing ioctl() for the same * source ID concurrently, prevent the second process to * acquire FD. */ if (hvpipe_find_source(srcID)) { - spin_unlock(&hvpipe_src_list_lock); + spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); return -EALREADY; } list_add(&src_info->list, &hvpipe_src_list); - spin_unlock(&hvpipe_src_list_lock); + spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); return fd_publish(fdf); } From cefeed44296261173a806bef988b26bc565da4be Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:41 +0530 Subject: [PATCH 133/333] pseries/papr-hvpipe: Prevent kernel stack memory leak to userspace The hdr variable is allocated on the stack and only hdr.version and hdr.flags are initialized explicitly. Because the struct papr_hvpipe_hdr contains reserved padding bytes (reserved[3] and reserved2[40]), these could leak the uninitialized bytes to userspace after copy_to_user(). This patch fixes that by initializing the whole struct to 0. Cc: stable@vger.kernel.org Fixes: cebdb522fd3ed ("powerpc/pseries: Receive payload with ibm,receive-hvpipe-msg RTAS") Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/7bfe03b65a282c856ed8182d1871bb973c0b78f2.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index c41d45e1986d..3392874ebdf6 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -327,7 +327,7 @@ static ssize_t papr_hvpipe_handle_read(struct file *file, { struct hvpipe_source_info *src_info = file->private_data; - struct papr_hvpipe_hdr hdr; + struct papr_hvpipe_hdr hdr = {}; long ret; /* From 1b9f7aafa44f5ce852c00509104d10fd9eb0f402 Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:42 +0530 Subject: [PATCH 134/333] pseries/papr-hvpipe: Fix null ptr deref in papr_hvpipe_dev_create_handle() commit 6d3789d347a7 ("papr-hvpipe: convert papr_hvpipe_dev_create_handle() to FD_PREPARE()"), changed the create handle to FD_PREPARE(), but it caused kernel null-ptr-deref because after call to retain_and_null_ptr(src_info), src_info is re-used for adding it to the global list. Getting the following kernel panic in papr_hvpipe_dev_create_handle() when trying to add src_info to the list. Kernel attempted to write user page (0) - exploit attempt? (uid: 0) BUG: Kernel NULL pointer dereference on write at 0x00000000 Faulting instruction address: 0xc0000000001b44a0 Oops: Kernel access of bad area, sig: 11 [#1] ... Call Trace: papr_hvpipe_dev_ioctl+0x1f4/0x48c (unreliable) sys_ioctl+0x528/0x1064 system_call_exception+0x128/0x360 system_call_vectored_common+0x15c/0x2ec Now, the error handling with FD_PREPARE's file cleanup and __free(kfree) auto cleanup is getting too convoluted. This is mainly because we need to ensure only 1 user get the srcID handle. To simplify this, we allocate prepare the src_info in the beginning and add it to the global list under a spinlock after checking that no duplicates exist. This simplify the error handling where if the FD_ADD fails, we can simply remove the src_info from the list and consume any pending msg in hvpipe to be cleared, after src_info became visible in the global list. Cc: stable@vger.kernel.org Fixes: 6d3789d347a7 ("papr-hvpipe: convert papr_hvpipe_dev_create_handle() to FD_PREPARE()") Reported-by: Haren Myneni Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/31ad94bc89d44156ee700c5bd006cb47a748e3cb.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 57 ++++++++++---------- 1 file changed, 30 insertions(+), 27 deletions(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index 3392874ebdf6..402781299497 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -480,23 +480,10 @@ static const struct file_operations papr_hvpipe_handle_ops = { static int papr_hvpipe_dev_create_handle(u32 srcID) { - struct hvpipe_source_info *src_info __free(kfree) = NULL; + struct hvpipe_source_info *src_info; + int fd; unsigned long flags; - spin_lock_irqsave(&hvpipe_src_list_lock, flags); - /* - * Do not allow more than one process communicates with - * each source. - */ - src_info = hvpipe_find_source(srcID); - if (src_info) { - spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); - pr_err("pid(%d) is already using the source(%d)\n", - src_info->tsk->pid, srcID); - return -EALREADY; - } - spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); - src_info = kzalloc_obj(*src_info, GFP_KERNEL_ACCOUNT); if (!src_info) return -ENOMEM; @@ -505,26 +492,42 @@ static int papr_hvpipe_dev_create_handle(u32 srcID) src_info->tsk = current; init_waitqueue_head(&src_info->recv_wqh); - FD_PREPARE(fdf, O_RDONLY | O_CLOEXEC, - anon_inode_getfile("[papr-hvpipe]", &papr_hvpipe_handle_ops, - (void *)src_info, O_RDWR)); - if (fdf.err) - return fdf.err; - - retain_and_null_ptr(src_info); - spin_lock_irqsave(&hvpipe_src_list_lock, flags); /* - * If two processes are executing ioctl() for the same - * source ID concurrently, prevent the second process to - * acquire FD. + * Do not allow more than one process communicates with + * each source. */ + spin_lock_irqsave(&hvpipe_src_list_lock, flags); if (hvpipe_find_source(srcID)) { spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); + pr_err("pid(%d) could not get the source(%d)\n", + src_info->tsk->pid, srcID); + kfree(src_info); return -EALREADY; } list_add(&src_info->list, &hvpipe_src_list); spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); - return fd_publish(fdf); + + fd = FD_ADD(O_RDONLY | O_CLOEXEC, + anon_inode_getfile("[papr-hvpipe]", &papr_hvpipe_handle_ops, + (void *)src_info, O_RDWR)); + if (fd < 0) { + spin_lock_irqsave(&hvpipe_src_list_lock, flags); + list_del(&src_info->list); + spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); + /* + * if we fail to add FD, that means no userspace program is + * polling. In that case if there is a msg pending because the + * interrupt was fired after the src_info was added to the + * global list, then let's consume it here, to unblock the + * hvpipe + */ + if (src_info->hvpipe_status & HVPIPE_MSG_AVAILABLE) + hvpipe_rtas_recv_msg(NULL, 0); + kfree(src_info); + return fd; + } + + return fd; } /* From 713e468cdbc2277db6ce949c32c1acbd83501733 Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:43 +0530 Subject: [PATCH 135/333] pseries/papr-hvpipe: Fix & simplify error handling in papr_hvpipe_init() Remove such 3 levels of nesting patterns to check success return values from function calls. ret = enable_hvpipe_IRQ() if (!ret) ret = set_hvpipe_sys_param(1) if (!ret) ret = misc_register() Instead just bail out to "out*:" labels, in case of any error. This simplifies the init flow. While at it let's also fix the following error handling logic: We have already enabled interrupt sources and enabled hvpipe to received interrupts, if misc_register() fails, we will destroy the workqueue, but the HMC might send us a msg via hvpipe which will call, queue work on the workqueue which might be destroyed. So instead, let's reverse the order of enabling set_hvpipe_sys_param(1) and in case of an error let's remove the misc dev by calling misc_deregister(). Cc: stable@vger.kernel.org Fixes: 39a08a4f94980 ("powerpc/pseries: Enable hvpipe with ibm,set-system-parameter RTAS") Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/f2141eafb80e7780395e03aa9a22e8a37be80513.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 28 ++++++++++++-------- 1 file changed, 17 insertions(+), 11 deletions(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index 402781299497..800649f309a5 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -780,23 +780,29 @@ static int __init papr_hvpipe_init(void) } ret = enable_hvpipe_IRQ(); - if (!ret) { - ret = set_hvpipe_sys_param(1); - if (!ret) - ret = misc_register(&papr_hvpipe_dev); - } + if (ret) + goto out_wq; - if (!ret) { - pr_info("hvpipe feature is enabled\n"); - hvpipe_feature = true; - return 0; - } + ret = misc_register(&papr_hvpipe_dev); + if (ret) + goto out_wq; - pr_err("hvpipe feature is not enabled %d\n", ret); + ret = set_hvpipe_sys_param(1); + if (ret) + goto out_misc; + + pr_info("hvpipe feature is enabled\n"); + hvpipe_feature = true; + return 0; + +out_misc: + misc_deregister(&papr_hvpipe_dev); +out_wq: destroy_workqueue(papr_hvpipe_wq); out: kfree(papr_hvpipe_work); papr_hvpipe_work = NULL; + pr_err("hvpipe feature is not enabled %d\n", ret); return ret; } machine_device_initcall(pseries, papr_hvpipe_init); From d48654bd8b1a75f662e224d257db54de475120dc Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:44 +0530 Subject: [PATCH 136/333] pseries/papr-hvpipe: Fix the usage of copy_to_user() copy_to_user() return bytes_not_copied to the user buffer. If there was an error writing bytes into the user buffer, i.e. if copy_to_user returns a non-zero value, then we should simply return -EFAULT from the ->read() call. Otherwise, in the non-patched version, we may end up mixing "bytes_not_copied + bytes_copied (HVPIPE_HDR_LEN)" as the return value to the user in ->read() call Also let's make sure we clear the hvpipe_status flag, if we have consumed the hvpipe msg by making the rtas call. ret = -EFAULT means copy_to_user has failed but that still means that the msg was read from the hvpipe, hence for both cases, success & -EFAULT, we should clear the HVPIPE_MSG_AVAILABLE flag in hvpipe_status. Cc: stable@vger.kernel.org Fixes: cebdb522fd3edd1 ("powerpc/pseries: Receive payload with ibm,receive-hvpipe-msg RTAS") Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/8fda3212a1ad48879c174e92f67472d9b9f1c3b7.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 23 ++++++++++++-------- 1 file changed, 14 insertions(+), 9 deletions(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index 800649f309a5..c007560d2d8c 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -206,10 +206,11 @@ static int hvpipe_rtas_recv_msg(char __user *buf, int size) bytes_written, size); bytes_written = size; } - ret = copy_to_user(buf, + if (copy_to_user(buf, rtas_work_area_raw_buf(work_area), - bytes_written); - if (!ret) + bytes_written)) + ret = -EFAULT; + else ret = bytes_written; } } else { @@ -328,7 +329,7 @@ static ssize_t papr_hvpipe_handle_read(struct file *file, struct hvpipe_source_info *src_info = file->private_data; struct papr_hvpipe_hdr hdr = {}; - long ret; + ssize_t ret = 0; /* * Return -ENXIO during migration @@ -376,7 +377,7 @@ static ssize_t papr_hvpipe_handle_read(struct file *file, ret = copy_to_user(buf, &hdr, HVPIPE_HDR_LEN); if (ret) - return ret; + return -EFAULT; /* * Message event has payload, so get the payload with @@ -385,19 +386,23 @@ static ssize_t papr_hvpipe_handle_read(struct file *file, if (hdr.flags & HVPIPE_MSG_AVAILABLE) { ret = hvpipe_rtas_recv_msg(buf + HVPIPE_HDR_LEN, size - HVPIPE_HDR_LEN); - if (ret > 0) { + /* + * Always clear MSG_AVAILABLE once the RTAS call has drained + * the message, regardless of whether copy_to_user succeeded. + */ + if (ret >= 0 || ret == -EFAULT) src_info->hvpipe_status &= ~HVPIPE_MSG_AVAILABLE; - ret += HVPIPE_HDR_LEN; - } } else if (hdr.flags & HVPIPE_LOST_CONNECTION) { /* * Hypervisor is closing the pipe for the specific * source. So notify user space. */ src_info->hvpipe_status &= ~HVPIPE_LOST_CONNECTION; - ret = HVPIPE_HDR_LEN; } + if (ret >= 0) + ret += HVPIPE_HDR_LEN; + return ret; } From 2eeac577480848801b35885b3a8201aa35f46236 Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:45 +0530 Subject: [PATCH 137/333] pseries/papr-hvpipe: Simplify spin unlock usage in papr_hvpipe_handle_release() Once the src_info is removed from the global list, no one can access it. This simplies the usage of spin_unlock_irqrestore() in papr_hvpipe_handle_release() Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/4a980331557af3d10aada8576aaa16cddc691c65.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index c007560d2d8c..5aa37f6ad8c9 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -460,6 +460,7 @@ static int papr_hvpipe_handle_release(struct inode *inode, src_info = file->private_data; list_del(&src_info->list); file->private_data = NULL; + spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); /* * If the pipe for this specific source has any pending * payload, issue recv HVPIPE RTAS so that pipe will not @@ -467,10 +468,8 @@ static int papr_hvpipe_handle_release(struct inode *inode, */ if (src_info->hvpipe_status & HVPIPE_MSG_AVAILABLE) { src_info->hvpipe_status = 0; - spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); hvpipe_rtas_recv_msg(NULL, 0); - } else - spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); + } kfree(src_info); return 0; From 4e2d83c80495a9327141e8636f25dde13155f14f Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:46 +0530 Subject: [PATCH 138/333] pseries/papr-hvpipe: Kill task_struct pointer from struct hvpipe_source_info We don't really use task_struct pointer for anything meaningful. So just kill it for now, and we can bring back later if we need this for any future debug purposes. Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/895e061e45cdc95db36fa7f27aa1922b81eed867.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 5 ++--- arch/powerpc/platforms/pseries/papr-hvpipe.h | 1 - 2 files changed, 2 insertions(+), 4 deletions(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index 5aa37f6ad8c9..46159f1c1cf1 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -493,7 +493,6 @@ static int papr_hvpipe_dev_create_handle(u32 srcID) return -ENOMEM; src_info->srcID = srcID; - src_info->tsk = current; init_waitqueue_head(&src_info->recv_wqh); /* @@ -503,8 +502,8 @@ static int papr_hvpipe_dev_create_handle(u32 srcID) spin_lock_irqsave(&hvpipe_src_list_lock, flags); if (hvpipe_find_source(srcID)) { spin_unlock_irqrestore(&hvpipe_src_list_lock, flags); - pr_err("pid(%d) could not get the source(%d)\n", - src_info->tsk->pid, srcID); + pr_err("pid(%s:%d) could not get the source(%d)\n", + current->comm, task_pid_nr(current), srcID); kfree(src_info); return -EALREADY; } diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.h b/arch/powerpc/platforms/pseries/papr-hvpipe.h index c343f4230865..4bdf7bb2fc4d 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.h +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.h @@ -21,7 +21,6 @@ struct hvpipe_source_info { u32 srcID; u32 hvpipe_status; wait_queue_head_t recv_wqh; /* wake up poll() waitq */ - struct task_struct *tsk; }; /* From fe53d2ae82c06aa2d6402624af01e8f8ddfcd5b3 Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:47 +0530 Subject: [PATCH 139/333] pseries/papr-hvpipe: Refactor and simplify hvpipe_rtas_recv_msg() Simplify hvpipe_rtas_recv_msg() by removing three levels of nesting... if (!ret) if (buf) if (size < bytes_written) ... this refactoring of the function bails out to "out:" label first, in case of any error. This simplifies the init flow. Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/bbe7ddf8b8e25c9be8fc5e2c4aea9e5fca128bf4.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 52 ++++++++++---------- 1 file changed, 26 insertions(+), 26 deletions(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index 46159f1c1cf1..3688b2be0445 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -190,34 +190,34 @@ static int hvpipe_rtas_recv_msg(char __user *buf, int size) return -ENOMEM; } - ret = rtas_ibm_receive_hvpipe_msg(work_area, &srcID, - &bytes_written); - if (!ret) { - /* - * Recv HVPIPE RTAS is successful. - * When releasing FD or no one is waiting on the - * specific source, issue recv HVPIPE RTAS call - * so that pipe is not blocked - this func is called - * with NULL buf. - */ - if (buf) { - if (size < bytes_written) { - pr_err("Received the payload size = %d, but the buffer size = %d\n", - bytes_written, size); - bytes_written = size; - } - if (copy_to_user(buf, - rtas_work_area_raw_buf(work_area), - bytes_written)) - ret = -EFAULT; - else - ret = bytes_written; - } - } else { - pr_err("ibm,receive-hvpipe-msg failed with %d\n", - ret); + /* + * Recv HVPIPE RTAS is successful. + * When releasing FD or no one is waiting on the + * specific source, issue recv HVPIPE RTAS call + * so that pipe is not blocked - this func is called + * with NULL buf. + */ + ret = rtas_ibm_receive_hvpipe_msg(work_area, &srcID, &bytes_written); + if (ret) { + pr_err("ibm,receive-hvpipe-msg failed with %d\n", ret); + goto out; } + if (!buf) + goto out; + + if (size < bytes_written) { + pr_err("Received the payload size = %d, but the buffer size = %d\n", + bytes_written, size); + bytes_written = size; + } + + if (copy_to_user(buf, rtas_work_area_raw_buf(work_area), bytes_written)) + ret = -EFAULT; + else + ret = bytes_written; + +out: rtas_work_area_free(work_area); return ret; } From 629d1a901de57490d29a495273df11e64993ec04 Mon Sep 17 00:00:00 2001 From: "Ritesh Harjani (IBM)" Date: Fri, 1 May 2026 09:41:48 +0530 Subject: [PATCH 140/333] pseries/papr-hvpipe: Fix style and checkpatch issues in enable_hvpipe_IRQ() While at it let's also fix the similar style issue in enable_hvpipe_IRQ() function. This also fixes a minor checkpatch warning which I got due to an extra space before " ==". Signed-off-by: Ritesh Harjani (IBM) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/1174f60d0ae128e773dbefd11dd8d46d69e7f50e.1777606826.git.ritesh.list@gmail.com --- arch/powerpc/platforms/pseries/papr-hvpipe.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) diff --git a/arch/powerpc/platforms/pseries/papr-hvpipe.c b/arch/powerpc/platforms/pseries/papr-hvpipe.c index 3688b2be0445..0c40bdde45e2 100644 --- a/arch/powerpc/platforms/pseries/papr-hvpipe.c +++ b/arch/powerpc/platforms/pseries/papr-hvpipe.c @@ -693,20 +693,19 @@ static int __init enable_hvpipe_IRQ(void) struct device_node *np; hvpipe_check_exception_token = rtas_function_token(RTAS_FN_CHECK_EXCEPTION); - if (hvpipe_check_exception_token == RTAS_UNKNOWN_SERVICE) + if (hvpipe_check_exception_token == RTAS_UNKNOWN_SERVICE) return -ENODEV; /* hvpipe events */ np = of_find_node_by_path("/event-sources/ibm,hvpipe-msg-events"); - if (np != NULL) { - request_event_sources_irqs(np, hvpipe_event_interrupt, - "HPIPE_EVENT"); - of_node_put(np); - } else { - pr_err("Can not enable hvpipe event IRQ\n"); + if (!np) { + pr_err("No device node found, could not enable hvpipe event IRQ\n"); return -ENODEV; } + request_event_sources_irqs(np, hvpipe_event_interrupt, "HPIPE_EVENT"); + of_node_put(np); + return 0; } From b3a97f9484080c6e71db9e803e3cc1bb372a9bc7 Mon Sep 17 00:00:00 2001 From: Sourabh Jain Date: Tue, 7 Apr 2026 18:13:44 +0530 Subject: [PATCH 141/333] powerpc/kdump: fix KASAN sanitization flag for core_$(BITS).o MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit KASAN instrumentation is intended to be disabled for the kexec core code, but the existing Makefile entry misses the object suffix. As a result, the flag is not applied correctly to core_$(BITS).o. So when KASAN is enabled, kexec_copy_flush and copy_segments in kexec/core_64.c are instrumented, which can result in accesses to shadow memory via normal address translation paths. Since these run with the MMU disabled, such accesses may trigger page faults (bad_page_fault) that cannot be handled in the kdump path, ultimately causing a hang and preventing the kdump kernel from booting. The same is true for kexec as well, since the same functions are used there. Update the entry to include the “.o” suffix so that KASAN instrumentation is properly disabled for this object file. Fixes: 2ab2d5794f14 ("powerpc/kasan: Disable address sanitization in kexec paths") Reported-by: Venkat Rao Bagalkote Closes: https://lore.kernel.org/all/1dee8891-8bcc-46b4-93f3-fc3a774abd5b@linux.ibm.com/ Cc: stable@vger.kernel.org Reviewed-by: Ritesh Harjani (IBM) Tested-by: Venkat Rao Bagalkote Acked-by: Mahesh Salgaonkar Reviewed-by: Aboorva Devarajan Tested-by: Aboorva Devarajan Signed-off-by: Sourabh Jain Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260407124349.1698552-1-sourabhjain@linux.ibm.com --- arch/powerpc/kexec/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/powerpc/kexec/Makefile b/arch/powerpc/kexec/Makefile index 470eb0453e17..ec7a0eed75dc 100644 --- a/arch/powerpc/kexec/Makefile +++ b/arch/powerpc/kexec/Makefile @@ -16,4 +16,4 @@ GCOV_PROFILE_core_$(BITS).o := n KCOV_INSTRUMENT_core_$(BITS).o := n UBSAN_SANITIZE_core_$(BITS).o := n KASAN_SANITIZE_core.o := n -KASAN_SANITIZE_core_$(BITS) := n +KASAN_SANITIZE_core_$(BITS).o := n From 38e989d504fc52900a3786b7144fb53cd67e0389 Mon Sep 17 00:00:00 2001 From: Sourabh Jain Date: Tue, 7 Apr 2026 18:13:45 +0530 Subject: [PATCH 142/333] powerpc/vmx: avoid KASAN instrumentation in enter_vmx_ops() for kexec The kexec sequence invokes enter_vmx_ops() via copy_page() with the MMU disabled. In this context, code must not rely on normal virtual address translations or trigger page faults. With KASAN enabled, functions get instrumented and may access shadow memory using regular address translation. When executed with the MMU off, this can lead to page faults (bad_page_fault) from which the kernel cannot recover in the kexec path, resulting in a hang. The kexec path sets preempt_count to HARDIRQ_OFFSET before entering the MMU-off copy sequence. current_thread_info()->preempt_count = HARDIRQ_OFFSET kexec_sequence(..., copy_with_mmu_off = 1) -> kexec_copy_flush(image) copy_segments() -> copy_page(dest, addr) bl enter_vmx_ops() if (in_interrupt()) return 0 beq .Lnonvmx_copy Since kexec sets preempt_count to HARDIRQ_OFFSET, in_interrupt() evaluates to true and enter_vmx_ops() returns early. As in_interrupt() (and preempt_count()) are always inlined, mark enter_vmx_ops() with __no_sanitize_address to avoid KASAN instrumentation and shadow memory access with MMU disabled, helping kexec boot fine with KASAN enabled. Reported-by: Aboorva Devarajan Reviewed-by: Aboorva Devarajan Tested-by: Aboorva Devarajan Reviewed-by: Ritesh Harjani (IBM) Signed-off-by: Sourabh Jain Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260407124349.1698552-2-sourabhjain@linux.ibm.com --- arch/powerpc/lib/vmx-helper.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/arch/powerpc/lib/vmx-helper.c b/arch/powerpc/lib/vmx-helper.c index 554b248002b4..57e897b60db8 100644 --- a/arch/powerpc/lib/vmx-helper.c +++ b/arch/powerpc/lib/vmx-helper.c @@ -52,7 +52,14 @@ int exit_vmx_usercopy(void) } EXPORT_SYMBOL(exit_vmx_usercopy); -int enter_vmx_ops(void) +/* + * Can be called from kexec copy_page() path with MMU off. The kexec + * code sets preempt_count to HARDIRQ_OFFSET so we return early here. + * Since in_interrupt() is always inline, __no_sanitize_address on this + * function is sufficient to avoid KASAN shadow memory accesses in real + * mode. + */ +int __no_sanitize_address enter_vmx_ops(void) { if (in_interrupt()) return 0; From da107152c43915211e1fc0319b9526f4241abca2 Mon Sep 17 00:00:00 2001 From: "Christophe Leroy (CS GROUP)" Date: Tue, 21 Apr 2026 08:26:08 +0200 Subject: [PATCH 143/333] powerpc/8xx: Fix interrupt mask in cpm1_gpiochip_add16() Allthough fsl,cpm1-gpio-irq-mask always contains a 16 bits value, it is a standard u32 OF property as documented in Documentation/devicetree/bindings/soc/fsl/cpm_qe/gpio.txt The driver erroneously uses of_property_read_u16() leading to a mask which is always 0. Fix it by using of_property_read_u32() instead. Fixes: 726bd223105c ("powerpc/8xx: Adding support of IRQ in MPC8xx GPIO") Signed-off-by: Christophe Leroy (CS GROUP) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/bb0b6d6c4543238c38d5d29a776d0674a8c0c180.1776752750.git.chleroy@kernel.org --- arch/powerpc/platforms/8xx/cpm1.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/powerpc/platforms/8xx/cpm1.c b/arch/powerpc/platforms/8xx/cpm1.c index 7433be7d66ee..f00734f0590c 100644 --- a/arch/powerpc/platforms/8xx/cpm1.c +++ b/arch/powerpc/platforms/8xx/cpm1.c @@ -477,7 +477,7 @@ int cpm1_gpiochip_add16(struct device *dev) struct device_node *np = dev->of_node; struct cpm1_gpio16_chip *cpm1_gc; struct gpio_chip *gc; - u16 mask; + u32 mask; cpm1_gc = devm_kzalloc(dev, sizeof(*cpm1_gc), GFP_KERNEL); if (!cpm1_gc) @@ -485,7 +485,7 @@ int cpm1_gpiochip_add16(struct device *dev) spin_lock_init(&cpm1_gc->lock); - if (!of_property_read_u16(np, "fsl,cpm1-gpio-irq-mask", &mask)) { + if (!of_property_read_u32(np, "fsl,cpm1-gpio-irq-mask", &mask)) { int i, j; for (i = 0, j = 0; i < 16; i++) From 131717e656b379addb95af2dcb2d90c723bae24b Mon Sep 17 00:00:00 2001 From: Shivani Nittor Date: Tue, 21 Apr 2026 20:36:28 +0530 Subject: [PATCH 144/333] powerpc/perf: Update check for PERF_SAMPLE_DATA_SRC marked events The core-book3s PMU sampling code validates the SIER TYPE field when PERF_SAMPLE_DATA_SRC is requested. The SIER TYPE field indicates the instruction type and is only valid for random sampling (marked events). To handle cases observed where SIER TYPE could be zero even for marked events,validation was added to drop such samples and increment event->lost_samples. However, this validation was applied to all samples, including continuous sampling. In continuous sampling mode, the PMU does not set the SIER TYPE field, so it remains zero. As a result, valid continuous samples were incorrectly treated as invalid and dropped. Fixed this by gating the SIER TYPE validation with mark_event, so the check runs only for marked (random) events. Continuous samples now skip this check and are recorded normally in the final data recording path. Fixes: 2ffb26afa642 ("arch/powerpc/perf: Check the instruction type before creating sample with perf_mem_data_src") Signed-off-by: Shivani Nittor Reviewed-by: Mukesh Kumar Chaurasiya (IBM) Reviewed-by: Athira Rajeev [Maddy: Fixed reviewed-by tag] Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260421150628.96500-1-shivani@linux.ibm.com --- arch/powerpc/perf/core-book3s.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/arch/powerpc/perf/core-book3s.c b/arch/powerpc/perf/core-book3s.c index 8b0081441f85..2e6adf5b95c4 100644 --- a/arch/powerpc/perf/core-book3s.c +++ b/arch/powerpc/perf/core-book3s.c @@ -2242,6 +2242,7 @@ static void record_and_restart(struct perf_event *event, unsigned long val, const u64 last_period = event->hw.last_period; s64 prev, delta, left; int record = 0; + int mark_event = regs->dsisr & MMCRA_SAMPLE_ENABLE; if (event->hw.state & PERF_HES_STOPPED) { write_pmc(event->hw.idx, 0); @@ -2304,9 +2305,9 @@ static void record_and_restart(struct perf_event *event, unsigned long val, * In ISA v3.0 and before values "0" and "7" are considered reserved. * In ISA v3.1, value "7" has been used to indicate "larx/stcx". * Drop the sample if "type" has reserved values for this field with a - * ISA version check. + * ISA version check for marked events. */ - if (event->attr.sample_type & PERF_SAMPLE_DATA_SRC && + if (mark_event && event->attr.sample_type & PERF_SAMPLE_DATA_SRC && ppmu->get_mem_data_src) { val = (regs->dar & SIER_TYPE_MASK) >> SIER_TYPE_SHIFT; if (val == 0 || (val == 7 && !cpu_has_feature(CPU_FTR_ARCH_31))) { From 4052b932041614675fa2dbf48f725557c23ebf05 Mon Sep 17 00:00:00 2001 From: Thomas Zimmermann Date: Wed, 1 Apr 2026 10:30:03 +0200 Subject: [PATCH 145/333] arch/powerpc: Drop CONFIG_FIRMWARE_EDID from defconfig files CONFIG_FIRMWARE_EDID=y depends on X86 or EFI_GENERIC_STUB. Neither is true here, so drop the lines from the defconfig files. Signed-off-by: Thomas Zimmermann Reviewed-by: Christophe Leroy (CS GROUP) Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260401083023.214426-1-tzimmermann@suse.de --- arch/powerpc/configs/amigaone_defconfig | 1 - arch/powerpc/configs/chrp32_defconfig | 1 - arch/powerpc/configs/g5_defconfig | 1 - arch/powerpc/configs/pasemi_defconfig | 1 - arch/powerpc/configs/powernv_defconfig | 1 - arch/powerpc/configs/ppc64_defconfig | 1 - arch/powerpc/configs/ppc64e_defconfig | 1 - arch/powerpc/configs/skiroot_defconfig | 1 - 8 files changed, 8 deletions(-) diff --git a/arch/powerpc/configs/amigaone_defconfig b/arch/powerpc/configs/amigaone_defconfig index 69ef3dc31c4b..7a515390646b 100644 --- a/arch/powerpc/configs/amigaone_defconfig +++ b/arch/powerpc/configs/amigaone_defconfig @@ -76,7 +76,6 @@ CONFIG_SERIAL_8250_CONSOLE=y # CONFIG_HW_RANDOM is not set # CONFIG_HWMON is not set CONFIG_FB=y -CONFIG_FIRMWARE_EDID=y CONFIG_FB_TILEBLITTING=y CONFIG_FB_RADEON=y CONFIG_FB_3DFX=y diff --git a/arch/powerpc/configs/chrp32_defconfig b/arch/powerpc/configs/chrp32_defconfig index b799c95480ae..66eae5b7e16c 100644 --- a/arch/powerpc/configs/chrp32_defconfig +++ b/arch/powerpc/configs/chrp32_defconfig @@ -76,7 +76,6 @@ CONFIG_SERIAL_8250_CONSOLE=y CONFIG_NVRAM=y # CONFIG_HWMON is not set CONFIG_FB=y -CONFIG_FIRMWARE_EDID=y CONFIG_FB_OF=y CONFIG_FB_MATROX=y CONFIG_FB_MATROX_MILLENIUM=y diff --git a/arch/powerpc/configs/g5_defconfig b/arch/powerpc/configs/g5_defconfig index 04bbb37f5978..e9996711b362 100644 --- a/arch/powerpc/configs/g5_defconfig +++ b/arch/powerpc/configs/g5_defconfig @@ -121,7 +121,6 @@ CONFIG_I2C_CHARDEV=y CONFIG_AGP=m CONFIG_AGP_UNINORTH=m CONFIG_FB=y -CONFIG_FIRMWARE_EDID=y CONFIG_FB_TILEBLITTING=y CONFIG_FB_OF=y CONFIG_FB_NVIDIA=y diff --git a/arch/powerpc/configs/pasemi_defconfig b/arch/powerpc/configs/pasemi_defconfig index 8bbf51b38480..89bcbeb05067 100644 --- a/arch/powerpc/configs/pasemi_defconfig +++ b/arch/powerpc/configs/pasemi_defconfig @@ -98,7 +98,6 @@ CONFIG_SENSORS_LM85=y CONFIG_SENSORS_LM90=y CONFIG_DRM=y CONFIG_DRM_RADEON=y -CONFIG_FIRMWARE_EDID=y CONFIG_FB_TILEBLITTING=y CONFIG_FB_VGA16=y CONFIG_FB_NVIDIA=y diff --git a/arch/powerpc/configs/powernv_defconfig b/arch/powerpc/configs/powernv_defconfig index cc9802420237..5d32c2767a65 100644 --- a/arch/powerpc/configs/powernv_defconfig +++ b/arch/powerpc/configs/powernv_defconfig @@ -196,7 +196,6 @@ CONFIG_I2C_CHARDEV=y # CONFIG_PTP_1588_CLOCK is not set CONFIG_DRM=y CONFIG_DRM_AST=y -CONFIG_FIRMWARE_EDID=y CONFIG_FB_OF=y CONFIG_FB_MATROX=m CONFIG_FB_MATROX_MILLENIUM=y diff --git a/arch/powerpc/configs/ppc64_defconfig b/arch/powerpc/configs/ppc64_defconfig index 3bf518e3a573..6316ca4df25d 100644 --- a/arch/powerpc/configs/ppc64_defconfig +++ b/arch/powerpc/configs/ppc64_defconfig @@ -249,7 +249,6 @@ CONFIG_I2C_CHARDEV=y CONFIG_I2C_AMD8111=y CONFIG_I2C_PASEMI=y CONFIG_FB=y -CONFIG_FIRMWARE_EDID=y CONFIG_FB_OF=y CONFIG_FB_MATROX=y CONFIG_FB_MATROX_MILLENIUM=y diff --git a/arch/powerpc/configs/ppc64e_defconfig b/arch/powerpc/configs/ppc64e_defconfig index 0fd49f67331f..20cc17dce94d 100644 --- a/arch/powerpc/configs/ppc64e_defconfig +++ b/arch/powerpc/configs/ppc64e_defconfig @@ -118,7 +118,6 @@ CONFIG_SERIAL_8250_CONSOLE=y CONFIG_I2C_CHARDEV=y CONFIG_I2C_AMD8111=y CONFIG_FB=y -CONFIG_FIRMWARE_EDID=y CONFIG_FB_OF=y CONFIG_FB_MATROX=y CONFIG_FB_MATROX_MILLENIUM=y diff --git a/arch/powerpc/configs/skiroot_defconfig b/arch/powerpc/configs/skiroot_defconfig index ff1bed4b6d2c..005536ee75bb 100644 --- a/arch/powerpc/configs/skiroot_defconfig +++ b/arch/powerpc/configs/skiroot_defconfig @@ -214,7 +214,6 @@ CONFIG_SENSORS_IBMPOWERNV=m CONFIG_DRM=m CONFIG_DRM_AST=m CONFIG_FB=y -CONFIG_FIRMWARE_EDID=y # CONFIG_VGA_CONSOLE is not set CONFIG_FRAMEBUFFER_CONSOLE=y CONFIG_LOGO=y From 60c71369ee356d11ad845ddeb28ceb70ec6cd70e Mon Sep 17 00:00:00 2001 From: Nathan Chancellor Date: Wed, 11 Mar 2026 17:39:56 -0700 Subject: [PATCH 146/333] powerpc/vdso: Drop -DCC_USING_PATCHABLE_FUNCTION_ENTRY from 32-bit flags with clang After commit 73cdf24e81e4 ("powerpc64: make clang cross-build friendly"), building 64-bit little endian + CONFIG_COMPAT=y with clang results in many warnings along the lines of: $ cat arch/powerpc/configs/compat.config CONFIG_COMPAT=y $ make -skj"$(nproc)" ARCH=powerpc LLVM=1 ppc64le_defconfig compat.config arch/powerpc/kernel/vdso/ ... In file included from :4: In file included from lib/vdso/gettimeofday.c:6: In file included from include/vdso/datapage.h:15: In file included from include/vdso/cache.h:5: arch/powerpc/include/asm/cache.h:77:8: warning: unknown attribute 'patchable_function_entry' ignored [-Wunknown-attributes] 77 | static inline u32 l1_icache_bytes(void) | ^~~~~~ include/linux/compiler_types.h:235:58: note: expanded from macro 'inline' 235 | #define inline inline __gnu_inline __inline_maybe_unused notrace | ^~~~~~~ include/linux/compiler_types.h:215:34: note: expanded from macro 'notrace' 215 | #define notrace __attribute__((patchable_function_entry(0, 0))) | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ... arch/powerpc/Makefile adds -DCC_USING_PATCHABLE_FUNCTION_ENTRY to KBUILD_CPPFLAGS, which is inherited by the 32-bit vDSO. However, the 32-bit little endian target does not support '-fpatchable-function-entry', resulting in the warnings above. Remove -DCC_USING_PATCHABLE_FUNCTION_ENTRY from the 32-bit vDSO flags when building with clang to avoid the warnings. Fixes: 73cdf24e81e4 ("powerpc64: make clang cross-build friendly") Signed-off-by: Nathan Chancellor Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260311-ppc-vdso-drop-cc-using-pfe-define-clang-v1-1-66c790e22650@kernel.org --- arch/powerpc/kernel/vdso/Makefile | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/arch/powerpc/kernel/vdso/Makefile b/arch/powerpc/kernel/vdso/Makefile index 8834dfe9d727..368759f81708 100644 --- a/arch/powerpc/kernel/vdso/Makefile +++ b/arch/powerpc/kernel/vdso/Makefile @@ -62,6 +62,12 @@ CC32FLAGSREMOVE += -fno-stack-clash-protection # 32-bit one. clang validates the values passed to these arguments during # parsing, even when -fno-stack-protector is passed afterwards. CC32FLAGSREMOVE += -mstack-protector-guard% +# ftrace is disabled for the vdso but arch/powerpc/Makefile adds this define to +# KBUILD_CPPFLAGS, which enables use of the 'patchable_function_entry' +# attribute in the 'inline' define via 'notrace'. This attribute is not +# supported for the powerpcle target, resulting in many instances of +# -Wunknown-attributes. +CC32FLAGSREMOVE += -DCC_USING_PATCHABLE_FUNCTION_ENTRY endif LD32FLAGS := -Wl,-soname=linux-vdso32.so.1 AS32FLAGS := -D__VDSO32__ From 8333e4916040e529bd5b56b82d573aba51e88a14 Mon Sep 17 00:00:00 2001 From: Krzysztof Kozlowski Date: Tue, 17 Mar 2026 14:08:24 +0100 Subject: [PATCH 147/333] powerpc/ps3: Drop redundant result assignment Return value of ps3_start_probe_thread() is not used, so code can be simplified to fix W=1 clang warnings: arch/powerpc/platforms/ps3/device-init.c:953:6: error: variable 'result' set but not used [-Werror,-Wunused-but-set-variable] Signed-off-by: Krzysztof Kozlowski Reviewed-by: Geert Uytterhoeven Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260317130823.240279-3-krzysztof.kozlowski@oss.qualcomm.com --- arch/powerpc/platforms/ps3/device-init.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/arch/powerpc/platforms/ps3/device-init.c b/arch/powerpc/platforms/ps3/device-init.c index 12c473768c39..9109c218a060 100644 --- a/arch/powerpc/platforms/ps3/device-init.c +++ b/arch/powerpc/platforms/ps3/device-init.c @@ -950,8 +950,6 @@ static int __init ps3_start_probe_thread(enum ps3_bus_type bus_type) static int __init ps3_register_devices(void) { - int result; - if (!firmware_has_feature(FW_FEATURE_PS3_LV1)) return -ENODEV; @@ -959,7 +957,7 @@ static int __init ps3_register_devices(void) /* ps3_repository_dump_bus_info(); */ - result = ps3_start_probe_thread(PS3_BUS_TYPE_STORAGE); + ps3_start_probe_thread(PS3_BUS_TYPE_STORAGE); ps3_register_vuart_devices(); From f583bd5f64d40e083dde5bb22846c4d93e59d471 Mon Sep 17 00:00:00 2001 From: Krzysztof Kozlowski Date: Tue, 17 Mar 2026 14:08:25 +0100 Subject: [PATCH 148/333] powerpc/pasemi: Drop redundant res assignment Return value of pas_add_bridge() is not used, so code can be simplified to fix W=1 clang warnings: arch/powerpc/platforms/pasemi/pci.c:275:6: error: variable 'res' set but not used [-Werror,-Wunused-but-set-variable] Signed-off-by: Krzysztof Kozlowski Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20260317130823.240279-4-krzysztof.kozlowski@oss.qualcomm.com --- arch/powerpc/platforms/pasemi/pci.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/arch/powerpc/platforms/pasemi/pci.c b/arch/powerpc/platforms/pasemi/pci.c index 60f990a336c4..2df955274652 100644 --- a/arch/powerpc/platforms/pasemi/pci.c +++ b/arch/powerpc/platforms/pasemi/pci.c @@ -272,13 +272,12 @@ void __init pas_pci_init(void) { struct device_node *root = of_find_node_by_path("/"); struct device_node *np; - int res; pci_set_flags(PCI_SCAN_ALL_PCIE_DEVS); np = of_find_compatible_node(root, NULL, "pasemi,rootbus"); if (np) { - res = pas_add_bridge(np); + pas_add_bridge(np); of_node_put(np); } of_node_put(root); From bd3c45dd01283ada23b0a388c578dcf5600deb8a Mon Sep 17 00:00:00 2001 From: Frederic Weisbecker Date: Thu, 23 Apr 2026 18:53:49 +0200 Subject: [PATCH 149/333] timers/migration: Fix another hotplug activation race The hotplug control CPU is assumed to be active in the hierarchy but that doesn't imply that the root is active. If the current CPU is not the one that activated the current hierarchy, and the CPU performing this duty is still halfway through the tree, the root may still be observed inactive. And this can break the activation of a new root as in the following scenario: 1) Initially, the whole system has 64 CPUs and only CPU 63 is awake. [GRP1:0] active / | \ / | \ [GRP0:0] [...] [GRP0:7] idle idle active / | \ | CPU 0 CPU 1 ... CPU 63 idle idle active 2) CPU 63 goes idle _but_ due to a #VMEXIT it hasn't yet reached the [GRP1:0]->parent dereference (that would be NULL and stop the walk) in __walk_groups_from(). [GRP1:0] idle / | \ / | \ [GRP0:0] [...] [GRP0:7] idle idle idle / | \ | CPU 0 CPU 1 ... CPU 63 idle idle idle 3) CPU 1 wakes up, activates GRP0:0 but didn't yet manage to propagate up to GRP1:0 due to yet another #VMEXIT. [GRP1:0] idle / | \ / | \ [GRP0:0] [...] [GRP0:7] active idle idle / | \ | CPU 0 CPU 1 ... CPU 63 idle active idle 3) CPU 0 wakes up and doesn't need to walk above GRP0:0 as it's CPU 1 role. [GRP1:0] idle / | \ / | \ [GRP0:0] [...] [GRP0:7] active idle idle / | \ | CPU 0 CPU 1 ... CPU 63 active active idle 4) CPU 0 boots CPU 64. It creates a new root for it. [GRP2:0] idle / \ / \ [GRP1:0] [GRP1:1] idle idle / | \ \ / | \ \ [GRP0:0] [...] [GRP0:7] [GRP0:8] active idle idle idle / | \ | | CPU 0 CPU 1 ... CPU 63 CPU 64 active active idle offline 5) CPU 0 activates the new root, but note that GRP1:0 is still idle, waiting for CPU 1 to resume from #VMEXIT and activate it. [GRP2:0] active / \ / \ [GRP1:0] [GRP1:1] idle idle / | \ \ / | \ \ [GRP0:0] [...] [GRP0:7] [GRP0:8] active idle idle idle / | \ | | CPU 0 CPU 1 ... CPU 63 CPU 64 active active idle offline 6) CPU 63 resumes after #VMEXIT and sees the new GRP1:0 parent. Therefore it propagates the stale inactive state of GRP1:0 up to GRP2:0. [GRP2:0] idle / \ / \ [GRP1:0] [GRP1:1] idle idle / | \ \ / | \ \ [GRP0:0] [...] [GRP0:7] [GRP0:8] active idle idle idle / | \ | | CPU 0 CPU 1 ... CPU 63 CPU 64 active active idle offline 7) CPU 1 resumes after #VMEXIT and finally activates GRP1:0. But it doesn't observe its parent link because no ordering enforced that. Therefore GRP2:0 is spuriously left idle. [GRP2:0] idle / \ / \ [GRP1:0] [GRP1:1] active idle / | \ \ / | \ \ [GRP0:0] [...] [GRP0:7] [GRP0:8] active idle idle idle / | \ | | CPU 0 CPU 1 ... CPU 63 CPU 64 active active idle offline Such races are highly theoretical and the problem would solve itself once the old root ever becomes idle again. But it still leaves a taste of discomfort. Fix it with enforcing a fully ordered atomic read of the old root state before propagating the activate state up to the new root. It has a two directions ordering effect: * Acquire + release of the latest old root state: If the hotplug control CPU is not the one that woke up the old root, make sure to acquire its active state and propagate it upwards through the ordered chain of activation (the acquire pairs with the cmpxchg() in tmigr_active_up() and subsequent releases will pair with atomic_read_acquire() and smp_mb__after_atomic() in tmigr_inactive_up()). * Release: If the hotplug control CPU is not the one that must wake up the old root, but the CPU covering that is lagging behind its duty, publish the links from the old root to the new parents. This way the lagging CPU will propagate the active state itself. Fixes: 7ee988770326 ("timers: Implement the hierarchical pull model") Signed-off-by: Frederic Weisbecker Signed-off-by: Thomas Gleixner Link: https://patch.msgid.link/20260423165354.95152-2-frederic@kernel.org --- kernel/time/timer_migration.c | 40 +++++++++++++++++++++++++---------- 1 file changed, 29 insertions(+), 11 deletions(-) diff --git a/kernel/time/timer_migration.c b/kernel/time/timer_migration.c index 155eeaea4113..1d0d3a4058d5 100644 --- a/kernel/time/timer_migration.c +++ b/kernel/time/timer_migration.c @@ -1860,19 +1860,37 @@ static int tmigr_setup_groups(unsigned int cpu, unsigned int node, * child to the new parents. So tmigr_active_up() activates the * new parents while walking up from the old root to the new. * - * * It is ensured that @start is active, as this setup path is - * executed in hotplug prepare callback. This is executed by an - * already connected and !idle CPU. Even if all other CPUs go idle, - * the CPU executing the setup will be responsible up to current top - * level group. And the next time it goes inactive, it will release - * the new childmask and parent to subsequent walkers through this - * @child. Therefore propagate active state unconditionally. + * * It is ensured that @start is active, (or on the way to be activated + * by another CPU that woke up before the current one) as this setup path + * is executed in hotplug prepare callback. This is executed by an already + * connected and !idle CPU in the hierarchy. + * + * * The below RmW atomic operation ensures that: + * + * 1) If the old root has been completely activated, the latest state is + * acquired (the below implicit acquire pairs with the implicit release + * from cmpxchg() in tmigr_active_up()). + * + * 2) If the old root is still on the way to be activated, the lagging behind + * CPU performing the activation will acquire the links up to the new root. + * (The below implicit release pairs with the implicit acquire from cmpxchg() + * in tmigr_active_up()). + * + * 3) Every subsequent CPU below the old root will acquire the new links while + * walking through the old root (The below implicit release pairs with the + * implicit acquire from cmpxchg() in either tmigr_active_up()) or + * tmigr_inactive_up(). */ - state.state = atomic_read(&start->migr_state); - WARN_ON_ONCE(!state.active); + state.state = atomic_fetch_or(0, &start->migr_state); WARN_ON_ONCE(!start->parent); - data.childmask = start->groupmask; - __walk_groups_from(tmigr_active_up, &data, start, start->parent); + /* + * If the state of the old root is inactive, another CPU is on its way to activate + * it and propagate to the new root. + */ + if (state.active) { + data.childmask = start->groupmask; + __walk_groups_from(tmigr_active_up, &data, start, start->parent); + } } /* Root update */ From 86f33ca9bea30cf011f2b1edad4593faea9c6e98 Mon Sep 17 00:00:00 2001 From: Ming Lei Date: Wed, 6 May 2026 16:22:38 +0800 Subject: [PATCH 150/333] ublk: validate physical_bs_shift, io_min_shift and io_opt_shift ublk_validate_params() checks logical_bs_shift is within [9, PAGE_SHIFT] but has no upper bound for physical_bs_shift, io_min_shift, or io_opt_shift. A malicious userspace can set any of these to a large value (e.g., 44), causing undefined behavior from `1 << shift` in ublk_ctrl_start_dev() since the result is stored in 32-bit unsigned int. Cap all three at ilog2(SZ_256M) (28). 256M is big enough to cover all practical block sizes, and originates from the maximum physical block size possible in NVMe (lba_size * (1 + npwg), where npwg is 16-bit). Also zero out ub->params with memset() when copy_from_user() fails or ublk_validate_params() returns error, so that no stale or partial params survive for a subsequent START_DEV to consume. Fixes: 71f28f3136af ("ublk_drv: add io_uring based userspace block driver") Signed-off-by: Ming Lei Link: https://patch.msgid.link/20260506082238.22363-1-tom.leiming@gmail.com Signed-off-by: Jens Axboe --- drivers/block/ublk_drv.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/drivers/block/ublk_drv.c b/drivers/block/ublk_drv.c index d10460d29e4a..57ec900f0ce0 100644 --- a/drivers/block/ublk_drv.c +++ b/drivers/block/ublk_drv.c @@ -900,6 +900,20 @@ static int ublk_validate_params(const struct ublk_device *ub) if (p->logical_bs_shift > PAGE_SHIFT || p->logical_bs_shift < 9) return -EINVAL; + /* + * 256M is a reasonable upper bound for physical block size, + * io_min and io_opt; it aligns with the maximum physical + * block size possible in NVMe. + */ + if (p->physical_bs_shift > ilog2(SZ_256M)) + return -EINVAL; + + if (p->io_min_shift > ilog2(SZ_256M)) + return -EINVAL; + + if (p->io_opt_shift > ilog2(SZ_256M)) + return -EINVAL; + if (p->logical_bs_shift > p->physical_bs_shift) return -EINVAL; @@ -4992,13 +5006,15 @@ static int ublk_ctrl_set_params(struct ublk_device *ub, */ ret = -EACCES; } else if (copy_from_user(&ub->params, argp, ph.len)) { + /* zero out partial copy so no stale params survive */ + memset(&ub->params, 0, sizeof(ub->params)); ret = -EFAULT; } else { /* clear all we don't support yet */ ub->params.types &= UBLK_PARAM_TYPE_ALL; ret = ublk_validate_params(ub); if (ret) - ub->params.types = 0; + memset(&ub->params, 0, sizeof(ub->params)); } mutex_unlock(&ub->mutex); From 9cc6bac1bebf8310d2950d1411a91479e86d69a1 Mon Sep 17 00:00:00 2001 From: Maoyi Xie Date: Mon, 4 May 2026 23:37:54 +0800 Subject: [PATCH 151/333] io_uring/timeout: honour caller's time namespace for IORING_TIMEOUT_ABS io_uring's IORING_OP_TIMEOUT and IORING_OP_LINK_TIMEOUT accept a timespec from the caller via io_parse_user_time(). With IORING_TIMEOUT_ABS, the timestamp is an absolute deadline on the selected clock. The clock is CLOCK_MONOTONIC by default. CLOCK_BOOTTIME and CLOCK_REALTIME are also selectable. A submitter inside a CLONE_NEWTIME time namespace observes CLOCK_MONOTONIC and CLOCK_BOOTTIME shifted by the namespace's offsets relative to the host. Every other ABS timer interface in the kernel converts the caller's absolute time to host view via timens_ktime_to_host() before arming an hrtimer: kernel/time/posix-timers.c -- timer_settime(TIMER_ABSTIME) kernel/time/posix-stubs.c -- clock_nanosleep(TIMER_ABSTIME) kernel/time/alarmtimer.c -- alarm_timer_nsleep(TIMER_ABSTIME) fs/timerfd.c -- timerfd_settime(TFD_TIMER_ABSTIME) io_parse_user_time() does not. As a result, an absolute timeout submitted from within a time namespace is interpreted in host view. That is generally a different point in time. It may already be in the past, causing the timer to fire immediately, or far in the future, causing the timer not to fire when expected. Reproducer: in unshare --user --time, with a -10s monotonic offset, submit IORING_OP_TIMEOUT with IORING_TIMEOUT_ABS and deadline = now + 1s. The CQE is delivered after <1ms instead of the expected ~1s. Apply timens_ktime_to_host() to the parsed time when IORING_TIMEOUT_ABS is set. Split the existing clock id resolver in io_timeout_get_clock() into a flags only helper io_flags_to_clock(), so io_parse_user_time() can resolve the clock without a struct io_timeout_data. timens_ktime_to_host() is a no-op for clocks not affected by time namespaces, e.g. CLOCK_REALTIME. It is also a no-op for callers in the initial time namespace. The fast path is unchanged. SQPOLL is also covered. The SQPOLL kernel thread is created via create_io_thread() with CLONE_THREAD and no CLONE_NEW* flag. copy_namespaces() therefore shares the submitter's nsproxy by reference. Inside the SQPOLL kthread, current->nsproxy->time_ns is the submitter's time_ns. timens_ktime_to_host() resolves correctly. Suggested-by: Pavel Begunkov Suggested-by: Jens Axboe Signed-off-by: Maoyi Xie Link: https://patch.msgid.link/20260504153755.1293932-2-maoyi.xie@ntu.edu.sg Signed-off-by: Jens Axboe --- io_uring/timeout.c | 35 ++++++++++++++++++++++------------- 1 file changed, 22 insertions(+), 13 deletions(-) diff --git a/io_uring/timeout.c b/io_uring/timeout.c index 4cfdfc519770..e2595cae2b07 100644 --- a/io_uring/timeout.c +++ b/io_uring/timeout.c @@ -3,6 +3,7 @@ #include #include #include +#include #include @@ -35,6 +36,22 @@ struct io_timeout_rem { bool ltimeout; }; +static clockid_t io_flags_to_clock(unsigned flags) +{ + switch (flags & IORING_TIMEOUT_CLOCK_MASK) { + case IORING_TIMEOUT_BOOTTIME: + return CLOCK_BOOTTIME; + case IORING_TIMEOUT_REALTIME: + return CLOCK_REALTIME; + default: + /* can't happen, vetted at prep time */ + WARN_ON_ONCE(1); + fallthrough; + case 0: + return CLOCK_MONOTONIC; + } +} + static int io_parse_user_time(ktime_t *time, u64 arg, unsigned flags) { struct timespec64 ts; @@ -43,7 +60,7 @@ static int io_parse_user_time(ktime_t *time, u64 arg, unsigned flags) *time = ns_to_ktime(arg); if (*time < 0) return -EINVAL; - return 0; + goto out; } if (get_timespec64(&ts, u64_to_user_ptr(arg))) @@ -51,6 +68,9 @@ static int io_parse_user_time(ktime_t *time, u64 arg, unsigned flags) if (ts.tv_sec < 0 || ts.tv_nsec < 0) return -EINVAL; *time = timespec64_to_ktime(ts); +out: + if (flags & IORING_TIMEOUT_ABS) + *time = timens_ktime_to_host(io_flags_to_clock(flags), *time); return 0; } @@ -399,18 +419,7 @@ static enum hrtimer_restart io_link_timeout_fn(struct hrtimer *timer) static clockid_t io_timeout_get_clock(struct io_timeout_data *data) { - switch (data->flags & IORING_TIMEOUT_CLOCK_MASK) { - case IORING_TIMEOUT_BOOTTIME: - return CLOCK_BOOTTIME; - case IORING_TIMEOUT_REALTIME: - return CLOCK_REALTIME; - default: - /* can't happen, vetted at prep time */ - WARN_ON_ONCE(1); - fallthrough; - case 0: - return CLOCK_MONOTONIC; - } + return io_flags_to_clock(data->flags); } static int io_linked_timeout_update(struct io_ring_ctx *ctx, __u64 user_data, From 45d2b37a37ab98484693533496395c610a2cab96 Mon Sep 17 00:00:00 2001 From: Maoyi Xie Date: Mon, 4 May 2026 23:37:55 +0800 Subject: [PATCH 152/333] io_uring/wait: honour caller's time namespace for IORING_ENTER_ABS_TIMER io_uring_enter() with IORING_ENTER_ABS_TIMER takes an absolute timespec from the caller via ext_arg->ts. It arms an ABS mode hrtimer in __io_cqring_wait_schedule(). The conversion path in io_uring/wait.c parses ext_arg->ts inline rather than going through io_parse_user_time(). It therefore does not pick up the time namespace conversion added by the previous patch. Apply timens_ktime_to_host() to the parsed time on the IORING_ENTER_ABS_TIMER branch. This mirrors the IORING_TIMEOUT_ABS fix in io_parse_user_time(). Use ctx->clockid as the clock id. ctx->clockid is set either at ring creation or via IORING_REGISTER_CLOCK. timens_ktime_to_host() is a no-op for clocks not affected by time namespaces. It is also a no-op for callers in the initial time namespace. The fast path is unchanged. Reproducer: in unshare --user --time, with a -10s monotonic offset, call io_uring_enter with min_complete=1, IORING_ENTER_ABS_TIMER, and ts = now + 1s. The call returns -ETIME after <1ms instead of after the expected ~1s. Suggested-by: Pavel Begunkov Suggested-by: Jens Axboe Signed-off-by: Maoyi Xie Link: https://patch.msgid.link/20260504153755.1293932-3-maoyi.xie@ntu.edu.sg Signed-off-by: Jens Axboe --- io_uring/wait.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/io_uring/wait.c b/io_uring/wait.c index 91df86ce0d18..ec01e78a216d 100644 --- a/io_uring/wait.c +++ b/io_uring/wait.c @@ -5,6 +5,7 @@ #include #include #include +#include #include @@ -229,7 +230,10 @@ int io_cqring_wait(struct io_ring_ctx *ctx, int min_events, u32 flags, if (ext_arg->ts_set) { iowq.timeout = timespec64_to_ktime(ext_arg->ts); - if (!(flags & IORING_ENTER_ABS_TIMER)) + if (flags & IORING_ENTER_ABS_TIMER) + iowq.timeout = timens_ktime_to_host(ctx->clockid, + iowq.timeout); + else iowq.timeout = ktime_add(iowq.timeout, start_time); } From 5cbb61bf4168859d97c068d88d364f4f1f440325 Mon Sep 17 00:00:00 2001 From: Breno Leitao Date: Tue, 5 May 2026 09:02:13 -0700 Subject: [PATCH 153/333] arm64/fpsimd: ptrace: zero target's fpsimd_state, not the tracer's sve_set_common() is the backend for PTRACE_SETREGSET(NT_ARM_SVE) and PTRACE_SETREGSET(NT_ARM_SSVE). Every write in the function operates on the tracee (target) - except a single memset that uses current instead, zeroing the tracer's saved V0-V31 / FPSR / FPCR shadow on every ptrace SETREGSET call. The memset is meant to give the tracee a defined zero register image before the user-supplied payload is copied in (for partial writes, header-only writes, and FPSIMD<->SVE format switches). Aiming it at current both denies the tracee that clean slate and silently corrupts the tracer. The corruption of the tracer's saved FPSIMD state is not always observable. Where the tracer's state is live on a CPU, this may be reused without loading the corrupted state from memory, and will eventually be written back over the corrupted state. Where the tracer's state is saved in SVE_PT_REGS_SVE format, only the FPSR and FPCR are clobbered, and the effective copy of the vectors is in the task's sve_state. Reproducible on an arm64 kernel with SVE: a single-threaded tracer that loads a known pattern into V0-V31, issues PTRACE_SETREGSET(NT_ARM_SVE) on a child, and reads V0-V31 back observes them all zeroed within tens of thousands of iterations when a sibling thread keeps stealing the FPSIMD CPU binding. Fixes: 316283f276eb ("arm64/fpsimd: ptrace: Consistently handle partial writes to NT_ARM_(S)SVE") Cc: Signed-off-by: Breno Leitao Acked-by: Mark Rutland Signed-off-by: Catalin Marinas --- arch/arm64/kernel/ptrace.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c index ba5eab23fd90..4d08598e2891 100644 --- a/arch/arm64/kernel/ptrace.c +++ b/arch/arm64/kernel/ptrace.c @@ -983,8 +983,8 @@ static int sve_set_common(struct task_struct *target, } /* Always zero V regs, FPSR, and FPCR */ - memset(¤t->thread.uw.fpsimd_state, 0, - sizeof(current->thread.uw.fpsimd_state)); + memset(&target->thread.uw.fpsimd_state, 0, + sizeof(target->thread.uw.fpsimd_state)); /* Registers: FPSIMD-only case */ From 628497e6d925d43efb56e3ffecef0a9d217926b3 Mon Sep 17 00:00:00 2001 From: Fenglin Wu Date: Wed, 6 May 2026 02:28:51 -0700 Subject: [PATCH 154/333] regulator: qcom-rpmh: Fix index for pmh0101 ldo16 The wrong index is assigned to pmh0101 ldo16, which results incorrect rpmh resource being used when the regulator device is voted. Fix it. Fixes: 65efe5404d15 ("regulator: rpmh-regulator: Add RPMH regulator support for Glymur") Signed-off-by: Fenglin Wu Reviewed-by: Konrad Dybcio Link: https://patch.msgid.link/20260506-fix_pmh0101_ldo16_index-v1-1-cdc8708b01f4@oss.qualcomm.com Signed-off-by: Mark Brown --- drivers/regulator/qcom-rpmh-regulator.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/regulator/qcom-rpmh-regulator.c b/drivers/regulator/qcom-rpmh-regulator.c index 6e4cb2871fca..0dcb50bf5c35 100644 --- a/drivers/regulator/qcom-rpmh-regulator.c +++ b/drivers/regulator/qcom-rpmh-regulator.c @@ -1512,7 +1512,7 @@ static const struct rpmh_vreg_init_data pmh0101_vreg_data[] = { RPMH_VREG("ldo13", LDO, 13, &pmic5_pldo530_mvp150, "vdd-l2-l13-l14"), RPMH_VREG("ldo14", LDO, 14, &pmic5_pldo530_mvp150, "vdd-l2-l13-l14"), RPMH_VREG("ldo15", LDO, 15, &pmic5_nldo530, "vdd-l15"), - RPMH_VREG("ldo16", LDO, 15, &pmic5_pldo530_mvp600, "vdd-l5-l16"), + RPMH_VREG("ldo16", LDO, 16, &pmic5_pldo530_mvp600, "vdd-l5-l16"), RPMH_VREG("ldo17", LDO, 17, &pmic5_pldo515_mv, "vdd-l17"), RPMH_VREG("ldo18", LDO, 18, &pmic5_nldo530, "vdd-l18"), RPMH_VREG("bob1", BOB, 1, &pmic5_bob, "vdd-bob1"), From 4bacec2317527ba04b7172145848f1c206999ea1 Mon Sep 17 00:00:00 2001 From: Jiawei Liu Date: Wed, 6 May 2026 14:24:12 +0800 Subject: [PATCH 155/333] spi: ch341: correct company name in MODULE_DESCRIPTION The company name "QiHeng Electronics" is incorrect. The correct legal name is "Nanjing Qinheng Microelectronics Co., Ltd.". Update the module description accordingly. Signed-off-by: Jiawei Liu Link: https://patch.msgid.link/20260506062412.371034-1-ljw@wch.cn Signed-off-by: Mark Brown --- drivers/spi/spi-ch341.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/spi/spi-ch341.c b/drivers/spi/spi-ch341.c index 3eaa8f176f63..6448a44a8b67 100644 --- a/drivers/spi/spi-ch341.c +++ b/drivers/spi/spi-ch341.c @@ -250,5 +250,5 @@ static struct usb_driver ch341a_usb_driver = { module_usb_driver(ch341a_usb_driver); MODULE_AUTHOR("Johannes Thumshirn "); -MODULE_DESCRIPTION("QiHeng Electronics ch341 USB2SPI"); +MODULE_DESCRIPTION("Nanjing Qinheng Microelectronics CH341 USB2SPI driver"); MODULE_LICENSE("GPL v2"); From 0c8c88b8eb82a2a41bec5f17c076d6312dc40316 Mon Sep 17 00:00:00 2001 From: Colin Walters Date: Tue, 5 May 2026 15:42:57 -0700 Subject: [PATCH 156/333] ovl: fix verity lazy-load guard broken by fsverity_active() semantic change Commit f77f281b6118 ("fsverity: use a hashtable to find the fsverity_info") made fsverity_active() check whether the inode has the verity flag, rather than whether the inode's fsverity_info is loaded. This broke ovl_ensure_verity_loaded(), which wants to load the fsverity_info for any verity inodes that haven't had it loaded yet. Therefore, to check that the fsverity_info hasn't yet been loaded, use fsverity_get_info(inode) == NULL instead of !fsverity_active(inode). Also, since fsverity_get_info() now involves a hash table lookup, put the more lightweight IS_VERITY() flag check first. Fixes: f77f281b6118 ("fsverity: use a hashtable to find the fsverity_info") Cc: stable@vger.kernel.org Link: https://github.com/bootc-dev/bootc/issues/2174 Signed-off-by: Colin Walters Acked-by: Amir Goldstein Link: https://patch.msgid.link/20260505224257.23213-1-ebiggers@kernel.org Signed-off-by: Eric Biggers --- fs/overlayfs/util.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/overlayfs/util.c b/fs/overlayfs/util.c index 7b86a6bac644..b41f4788e4f0 100644 --- a/fs/overlayfs/util.c +++ b/fs/overlayfs/util.c @@ -1354,7 +1354,7 @@ int ovl_ensure_verity_loaded(const struct path *datapath) struct inode *inode = d_inode(datapath->dentry); struct file *filp; - if (!fsverity_active(inode) && IS_VERITY(inode)) { + if (IS_VERITY(inode) && fsverity_get_info(inode) == NULL) { /* * If this inode was not yet opened, the verity info hasn't been * loaded yet, so we need to do that here to force it into memory. From fdf4eb632683bfc2840acebe62716cb468d43e10 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 26 Apr 2026 17:51:07 +0200 Subject: [PATCH 157/333] selftests/rseq: Validate legacy behavior The RSEQ legacy mode behavior requires that the ID fields in the rseq region are unconditionally updated on every context switch and before signal delivery even if not required by the ABI specification. To ensure that this behavior is preserved for legacy users in the future, add a test which validates that with a sleep() and a signal sent to self. Provide a run script which prevents GLIBC from registering a RSEQ region, so that the test can register it's own legacy sized region. Fixes: 566d8015f7ee ("rseq: Avoid CPU/MM CID updates when no event pending") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224427.764705536%40kernel.org Cc: stable@vger.kernel.org --- tools/testing/selftests/rseq/Makefile | 5 +- tools/testing/selftests/rseq/legacy_check.c | 65 +++++++++++++++++++ .../selftests/rseq/run_legacy_check.sh | 4 ++ 3 files changed, 72 insertions(+), 2 deletions(-) create mode 100644 tools/testing/selftests/rseq/legacy_check.c create mode 100755 tools/testing/selftests/rseq/run_legacy_check.sh diff --git a/tools/testing/selftests/rseq/Makefile b/tools/testing/selftests/rseq/Makefile index 0d1947c0d623..0293a2f17f51 100644 --- a/tools/testing/selftests/rseq/Makefile +++ b/tools/testing/selftests/rseq/Makefile @@ -22,9 +22,10 @@ TEST_GEN_PROGS_EXTENDED = librseq.so \ param_test_compare_twice \ param_test_mm_cid \ param_test_mm_cid_compare_twice \ - syscall_errors_test + syscall_errors_test \ + legacy_check -TEST_PROGS = run_param_test.sh run_syscall_errors_test.sh +TEST_PROGS = run_param_test.sh run_syscall_errors_test.sh run_legacy_check.sh TEST_FILES := settings diff --git a/tools/testing/selftests/rseq/legacy_check.c b/tools/testing/selftests/rseq/legacy_check.c new file mode 100644 index 000000000000..3f7de4e28303 --- /dev/null +++ b/tools/testing/selftests/rseq/legacy_check.c @@ -0,0 +1,65 @@ +// SPDX-License-Identifier: GPL-2.0 +#ifndef _GNU_SOURCE +#define _GNU_SOURCE +#endif + +#include +#include +#include +#include + +#include "rseq.h" + +#include "../kselftest_harness.h" + +FIXTURE(legacy) +{ +}; + +static int cpu_id_in_sigfn = -1; + +static void sigfn(int sig) +{ + struct rseq_abi *rs = rseq_get_abi(); + + cpu_id_in_sigfn = rs->cpu_id_start; +} + +FIXTURE_SETUP(legacy) +{ + int res = __rseq_register_current_thread(true, true); + + switch (res) { + case -ENOSYS: + SKIP(return, "RSEQ not enabled\n"); + case -EBUSY: + SKIP(return, "GLIBC owns RSEQ. Disable GLIBC RSEQ registration\n"); + default: + ASSERT_EQ(res, 0); + } + + ASSERT_NE(signal(SIGUSR1, sigfn), SIG_ERR); +} + +FIXTURE_TEARDOWN(legacy) +{ +} + +TEST_F(legacy, legacy_test) +{ + struct rseq_abi *rs = rseq_get_abi(); + + ASSERT_NE(rs, NULL); + + /* Overwrite rs::cpu_id_start */ + rs->cpu_id_start = -1; + sleep(1); + ASSERT_NE(rs->cpu_id_start, -1); + + rs->cpu_id_start = -1; + ASSERT_EQ(raise(SIGUSR1), 0); + ASSERT_NE(rs->cpu_id_start, -1); + ASSERT_NE(cpu_id_in_sigfn, -1); +} + +TEST_HARNESS_MAIN diff --git a/tools/testing/selftests/rseq/run_legacy_check.sh b/tools/testing/selftests/rseq/run_legacy_check.sh new file mode 100755 index 000000000000..5577b46ea092 --- /dev/null +++ b/tools/testing/selftests/rseq/run_legacy_check.sh @@ -0,0 +1,4 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0 + +GLIBC_TUNABLES="${GLIBC_TUNABLES:-}:glibc.pthread.rseq=0" ./legacy_check From 82f572449cfe75f12ea985986da60e11f308f77d Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 26 Apr 2026 16:21:02 +0200 Subject: [PATCH 158/333] rseq: Implement read only ABI enforcement for optimized RSEQ V2 mode The optimized RSEQ V2 mode requires that user space adheres to the ABI specification and does not modify the read-only fields cpu_id_start, cpu_id, node_id and mm_cid behind the kernel's back. While the kernel does not rely on these fields, the adherence to this is a fundamental prerequisite to allow multiple entities, e.g. libraries, in an application to utilize the full potential of RSEQ without stepping on each other toes. Validate this adherence on every update of these fields. If the kernel detects that user space modified the fields, the application is force terminated. Fixes: d6200245c75e ("rseq: Allow registering RSEQ with slice extension") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224427.845230956%40kernel.org Cc: stable@vger.kernel.org --- include/linux/rseq_entry.h | 83 ++++++++++++++------------------------ include/linux/rseq_types.h | 4 +- kernel/rseq.c | 5 +-- 3 files changed, 35 insertions(+), 57 deletions(-) diff --git a/include/linux/rseq_entry.h b/include/linux/rseq_entry.h index 934db41ec782..2d0295df5107 100644 --- a/include/linux/rseq_entry.h +++ b/include/linux/rseq_entry.h @@ -248,7 +248,6 @@ static __always_inline bool rseq_grant_slice_extension(unsigned long ti_work, un #endif /* !CONFIG_RSEQ_SLICE_EXTENSION */ bool rseq_debug_update_user_cs(struct task_struct *t, struct pt_regs *regs, unsigned long csaddr); -bool rseq_debug_validate_ids(struct task_struct *t); static __always_inline void rseq_note_user_irq_entry(void) { @@ -368,43 +367,6 @@ bool rseq_debug_update_user_cs(struct task_struct *t, struct pt_regs *regs, return false; } -/* - * On debug kernels validate that user space did not mess with it if the - * debug branch is enabled. - */ -bool rseq_debug_validate_ids(struct task_struct *t) -{ - struct rseq __user *rseq = t->rseq.usrptr; - u32 cpu_id, uval, node_id; - - /* - * On the first exit after registering the rseq region CPU ID is - * RSEQ_CPU_ID_UNINITIALIZED and node_id in user space is 0! - */ - node_id = t->rseq.ids.cpu_id != RSEQ_CPU_ID_UNINITIALIZED ? - cpu_to_node(t->rseq.ids.cpu_id) : 0; - - scoped_user_read_access(rseq, efault) { - unsafe_get_user(cpu_id, &rseq->cpu_id_start, efault); - if (cpu_id != t->rseq.ids.cpu_id) - goto die; - unsafe_get_user(uval, &rseq->cpu_id, efault); - if (uval != cpu_id) - goto die; - unsafe_get_user(uval, &rseq->node_id, efault); - if (uval != node_id) - goto die; - unsafe_get_user(uval, &rseq->mm_cid, efault); - if (uval != t->rseq.ids.mm_cid) - goto die; - } - return true; -die: - t->rseq.event.fatal = true; -efault: - return false; -} - #endif /* RSEQ_BUILD_SLOW_PATH */ /* @@ -514,20 +476,32 @@ rseq_update_user_cs(struct task_struct *t, struct pt_regs *regs, unsigned long c * faults in task context are fatal too. */ static rseq_inline -bool rseq_set_ids_get_csaddr(struct task_struct *t, struct rseq_ids *ids, - u32 node_id, u64 *csaddr) +bool rseq_set_ids_get_csaddr(struct task_struct *t, struct rseq_ids *ids, u64 *csaddr) { struct rseq __user *rseq = t->rseq.usrptr; - if (static_branch_unlikely(&rseq_debug_enabled)) { - if (!rseq_debug_validate_ids(t)) - return false; - } - scoped_user_rw_access(rseq, efault) { + /* Validate the R/O fields for debug and optimized mode */ + if (static_branch_unlikely(&rseq_debug_enabled) || rseq_v2(t)) { + u32 cpu_id, uval; + + unsafe_get_user(cpu_id, &rseq->cpu_id_start, efault); + if (cpu_id != t->rseq.ids.cpu_id) + goto die; + unsafe_get_user(uval, &rseq->cpu_id, efault); + if (uval != cpu_id) + goto die; + unsafe_get_user(uval, &rseq->node_id, efault); + if (uval != t->rseq.ids.node_id) + goto die; + unsafe_get_user(uval, &rseq->mm_cid, efault); + if (uval != t->rseq.ids.mm_cid) + goto die; + } + unsafe_put_user(ids->cpu_id, &rseq->cpu_id_start, efault); unsafe_put_user(ids->cpu_id, &rseq->cpu_id, efault); - unsafe_put_user(node_id, &rseq->node_id, efault); + unsafe_put_user(ids->node_id, &rseq->node_id, efault); unsafe_put_user(ids->mm_cid, &rseq->mm_cid, efault); if (csaddr) unsafe_get_user(*csaddr, &rseq->rseq_cs, efault); @@ -539,10 +513,13 @@ bool rseq_set_ids_get_csaddr(struct task_struct *t, struct rseq_ids *ids, rseq_slice_clear_grant(t); /* Cache the new values */ - t->rseq.ids.cpu_cid = ids->cpu_cid; + t->rseq.ids = *ids; rseq_stat_inc(rseq_stats.ids); rseq_trace_update(t, ids); return true; + +die: + t->rseq.event.fatal = true; efault: return false; } @@ -552,11 +529,11 @@ bool rseq_set_ids_get_csaddr(struct task_struct *t, struct rseq_ids *ids, * is in a critical section. */ static rseq_inline bool rseq_update_usr(struct task_struct *t, struct pt_regs *regs, - struct rseq_ids *ids, u32 node_id) + struct rseq_ids *ids) { u64 csaddr; - if (!rseq_set_ids_get_csaddr(t, ids, node_id, &csaddr)) + if (!rseq_set_ids_get_csaddr(t, ids, &csaddr)) return false; /* @@ -659,12 +636,12 @@ static __always_inline bool rseq_exit_user_update(struct pt_regs *regs, struct t } struct rseq_ids ids = { - .cpu_id = task_cpu(t), - .mm_cid = task_mm_cid(t), + .cpu_id = task_cpu(t), + .mm_cid = task_mm_cid(t), + .node_id = cpu_to_node(ids.cpu_id), }; - u32 node_id = cpu_to_node(ids.cpu_id); - return rseq_update_usr(t, regs, &ids, node_id); + return rseq_update_usr(t, regs, &ids); efault: return false; } diff --git a/include/linux/rseq_types.h b/include/linux/rseq_types.h index a469c1870849..85739a63e85e 100644 --- a/include/linux/rseq_types.h +++ b/include/linux/rseq_types.h @@ -66,8 +66,9 @@ struct rseq_event { * compiler emit a single compare on 64-bit * @cpu_id: The CPU ID which was written last to user space * @mm_cid: The MM CID which was written last to user space + * @node_id: The node ID which was written last to user space * - * @cpu_id and @mm_cid are updated when the data is written to user space. + * @cpu_id, @mm_cid and @node_id are updated when the data is written to user space. */ struct rseq_ids { union { @@ -77,6 +78,7 @@ struct rseq_ids { u32 mm_cid; }; }; + u32 node_id; }; /** diff --git a/kernel/rseq.c b/kernel/rseq.c index aa25753ea135..101612027f6a 100644 --- a/kernel/rseq.c +++ b/kernel/rseq.c @@ -263,7 +263,6 @@ static void rseq_slowpath_update_usr(struct pt_regs *regs) }; struct task_struct *t = current; struct rseq_ids ids; - u32 node_id; bool event; if (unlikely(t->flags & PF_EXITING)) @@ -299,9 +298,9 @@ static void rseq_slowpath_update_usr(struct pt_regs *regs) if (!event) return; - node_id = cpu_to_node(ids.cpu_id); + ids.node_id = cpu_to_node(ids.cpu_id); - if (unlikely(!rseq_update_usr(t, regs, &ids, node_id))) { + if (unlikely(!rseq_update_usr(t, regs, &ids))) { /* * Clear the errors just in case this might survive magically, but * leave the rest intact. From 99428157dcf32fdac97355aa1cc1364dbc9e073c Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sun, 26 Apr 2026 10:01:56 +0200 Subject: [PATCH 159/333] rseq: Reenable performance optimizations conditionally Due to the incompatibility with TCMalloc the RSEQ optimizations and extended features (time slice extensions) have been disabled and made run-time conditional. The original RSEQ implementation, which TCMalloc depends on, registers a 32 byte region (ORIG_RSEG_SIZE). This region has a 32 byte alignment requirement. The extension safe newer variant exposes the kernel RSEQ feature size via getauxval(AT_RSEQ_FEATURE_SIZE) and the alignment requirement via getauxval(AT_RSEQ_ALIGN). The alignment requirement is that the registered RSEQ region is aligned to the next power of two of the feature size. The kernel currently has a feature size of 33 bytes, which means the alignment requirement is 64 bytes. The TCMalloc RSEQ region is embedded into a cache line aligned data structure starting at offset 32 bytes so that bytes 28-31 and the cpu_id_start field at bytes 32-35 form a 64-bit little endian pointer with the top-most bit (63 set) to check whether the kernel has overwritten cpu_id_start with an actual CPU id value, which is guaranteed to not have the top most bit set. As this is part of their performance tuned magic, it's a pretty safe assumption, that TCMalloc won't use a larger RSEQ size. This allows the kernel to declare that registrations with a size greater than the original size of 32 bytes, which is the cases since time slice extensions got introduced, as RSEQ ABI v2 with the following differences to the original behaviour: 1) Unconditional updates of the user read only fields (CPU, node, MMCID) are removed. Those fields are only updated on registration, task migration and MMCID changes. 2) Unconditional evaluation of the criticial section pointer is removed. It's only evaluated when user space was interrupted and was scheduled out or before delivering a signal in the interrupted context. 3) The read/only requirement of the ID fields is enforced. When the kernel detects that userspace manipulated the fields, the process is terminated. This ensures that multiple entities (libraries) can utilize RSEQ without interfering. 4) Todays extended RSEQ feature (time slice extensions) and future extensions are only enabled in the v2 enabled mode. Registrations with the original size of 32 bytes operate in backwards compatible legacy mode without performance improvements and extended features. Unfortunately that also affects users of older GLIBC versions which register the original size of 32 bytes and do not evaluate the kernel required size in the auxiliary vector AT_RSEQ_FEATURE_SIZE. That's the result of the lack of enforcement in the original implementation and the unwillingness of a single entity to cooperate with the larger ecosystem for many years. Implement the required registration changes by restructuring the spaghetti code and adding the size/version check. Also add documentation about the differences of legacy and optimized RSEQ V2 mode. Thanks to Mathieu for pointing out the ORIG_RSEQ_SIZE constraints! Fixes: d6200245c75e ("rseq: Allow registering RSEQ with slice extension") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224427.927160119%40kernel.org Cc: stable@vger.kernel.org --- Documentation/userspace-api/rseq.rst | 94 ++++++++++++++++- kernel/rseq.c | 144 ++++++++++++++++----------- 2 files changed, 178 insertions(+), 60 deletions(-) diff --git a/Documentation/userspace-api/rseq.rst b/Documentation/userspace-api/rseq.rst index 3cd27a3c7c7e..8549a6c61531 100644 --- a/Documentation/userspace-api/rseq.rst +++ b/Documentation/userspace-api/rseq.rst @@ -24,6 +24,97 @@ Quick access to CPU number, node ID Allows to implement per CPU data efficiently. Documentation is in code and selftests. :( +Optimized RSEQ V2 +----------------- + +On architectures which utilize the generic entry code and generic TIF bits +the kernel supports runtime optimizations for RSEQ, which also enable +enhanced features like scheduler time slice extensions. + +To enable them a task has to register the RSEQ region with at least the +length advertised by getauxval(AT_RSEQ_FEATURE_SIZE). + +If existing binaries register with RSEQ_ORIG_SIZE (32 bytes), the kernel +keeps the legacy low performance mode enabled to fulfil the expectations +of existing users regarding the original RSEQ implementation behaviour. + +The following table documents the ABI and behavioral guarantees of the +legacy and the optimized V2 mode. + +.. list-table:: RSEQ modes + :header-rows: 1 + + * - Nr + - What + + - Legacy + - Optimized V2 + + * - 1 + - The cpu_id_start, cpu_id, node_id and mm_cid fields (User mode read + only) + .. Legacy + - Updated by the kernel unconditionally after each context switch and + before signal delivery + .. Optimized V2 + - Updated by the kernel if and only if they change, i.e. if the task + is migrated or mm_cid changes + + * - 2 + - The rseq_cs critical section field + .. Legacy + - Evaluated and handled unconditionally after each context switch and + before signal delivery + .. Optimized V2 + - Evaluated and handled conditionally only when user space was + interrupted and was scheduled out or before delivering a signal in + the interrupted context. + + * - 3 + - Read only fields + .. Legacy + - No strict enforcement except in debug mode + .. Optimized V2 + - Strict enforcement + + * - 4 + - membarrier(...RSEQ) + .. Legacy + - All running threads of the process are interrupted and the ID fields + are rewritten and eventually active critical sections are aborted + before they return to user space. All threads which are scheduled + out whether voluntary or not are covered by #1/#2 above. + .. Optimized V2 + - All running threads of the process are interrupted and eventually + active critical sections are aborted before these threads return to + user space. The ID fields are only updated if changed as a + consequence of the interrupt. All threads which are scheduled out + whether voluntary or not are covered by #1/#2 above. + + * - 5 + - Time slice extensions + .. Legacy + - Not supported + .. Optimized V2 + - Supported + +The legacy mode is obviously less performant as it does unconditional +updates and critical section checks even if not strictly required by the +ABI contract. That can't be changed anymore as some users depend on that +observed behavior, which in turn enables them to violate the ABI and +overwrite the cpu_id_start field for their own purposes. This is obviously +discouraged as it renders RSEQ incompatible with the intended usage and +breaks the expectation of other libraries in the same application. + +The ABI compliant optimized v2 mode, which respects the read only fields, +does not require unconditional updates and therefore is way more +performant. The kernel validates the read only fields for compliance. If +user space modifies them, the process is killed. Compliant usage allows +multiple libraries in the same application to benefit from the RSEQ +functionality without disturbing each other. The ABI compliant optimized v2 +mode also enables extended RSEQ features like time slice extensions. + + Scheduler time slice extensions ------------------------------- @@ -37,7 +128,8 @@ The prerequisites for this functionality are: * Enabled at boot time (default is enabled) - * A rseq userspace pointer has been registered for the thread + * A rseq userspace pointer has been registered for the thread in + optimized V2 mode The thread has to enable the functionality via prctl(2):: diff --git a/kernel/rseq.c b/kernel/rseq.c index 101612027f6a..e75e3a5e312c 100644 --- a/kernel/rseq.c +++ b/kernel/rseq.c @@ -412,70 +412,23 @@ static bool rseq_reset_ids(void) /* The original rseq structure size (including padding) is 32 bytes. */ #define ORIG_RSEQ_SIZE 32 -/* - * sys_rseq - setup restartable sequences for caller thread. - */ -SYSCALL_DEFINE4(rseq, struct rseq __user *, rseq, u32, rseq_len, int, flags, u32, sig) +static long rseq_register(struct rseq __user * rseq, u32 rseq_len, int flags, u32 sig) { u32 rseqfl = 0; u8 version = 1; - if (flags & RSEQ_FLAG_UNREGISTER) { - if (flags & ~RSEQ_FLAG_UNREGISTER) - return -EINVAL; - /* Unregister rseq for current thread. */ - if (current->rseq.usrptr != rseq || !current->rseq.usrptr) - return -EINVAL; - if (rseq_len != current->rseq.len) - return -EINVAL; - if (current->rseq.sig != sig) - return -EPERM; - if (!rseq_reset_ids()) - return -EFAULT; - rseq_reset(current); - return 0; - } - - if (unlikely(flags & ~(RSEQ_FLAG_SLICE_EXT_DEFAULT_ON))) - return -EINVAL; - - if (current->rseq.usrptr) { - /* - * If rseq is already registered, check whether - * the provided address differs from the prior - * one. - */ - if (current->rseq.usrptr != rseq || rseq_len != current->rseq.len) - return -EINVAL; - if (current->rseq.sig != sig) - return -EPERM; - /* Already registered. */ - return -EBUSY; - } - - /* - * If there was no rseq previously registered, ensure the provided rseq - * is properly aligned, as communcated to user-space through the ELF - * auxiliary vector AT_RSEQ_ALIGN. If rseq_len is the original rseq - * size, the required alignment is the original struct rseq alignment. - * - * The rseq_len is required to be greater or equal to the original rseq - * size. In order to be valid, rseq_len is either the original rseq size, - * or large enough to contain all supported fields, as communicated to - * user-space through the ELF auxiliary vector AT_RSEQ_FEATURE_SIZE. - */ - if (rseq_len < ORIG_RSEQ_SIZE || - (rseq_len == ORIG_RSEQ_SIZE && !IS_ALIGNED((unsigned long)rseq, ORIG_RSEQ_SIZE)) || - (rseq_len != ORIG_RSEQ_SIZE && (!IS_ALIGNED((unsigned long)rseq, rseq_alloc_align()) || - rseq_len < offsetof(struct rseq, end)))) - return -EINVAL; if (!access_ok(rseq, rseq_len)) return -EFAULT; /* - * The version check effectivly disables time slice extensions until the - * RSEQ ABI V2 registration are implemented. + * Architectures, which use the generic IRQ entry code (at least) enable + * registrations with a size greater than the original v1 fixed sized + * @rseq_len, which has been validated already to utilize the optimized + * v2 ABI mode which also enables extended RSEQ features beyond MMCID. */ + if (IS_ENABLED(CONFIG_GENERIC_IRQ_ENTRY) && rseq_len > ORIG_RSEQ_SIZE) + version = 2; + if (IS_ENABLED(CONFIG_RSEQ_SLICE_EXTENSION) && version > 1) { if (rseq_slice_extension_enabled()) { rseqfl |= RSEQ_CS_FLAG_SLICE_EXT_AVAILABLE; @@ -523,11 +476,10 @@ SYSCALL_DEFINE4(rseq, struct rseq __user *, rseq, u32, rseq_len, int, flags, u32 #endif /* - * If rseq was previously inactive, and has just been - * registered, ensure the cpu_id_start and cpu_id fields - * are updated before returning to user-space. + * Ensure the cpu_id_start and cpu_id fields are updated before + * returning to user-space. */ - current->rseq.event.has_rseq = true; + current->rseq.event.has_rseq = version; rseq_force_update(); return 0; @@ -535,6 +487,80 @@ SYSCALL_DEFINE4(rseq, struct rseq __user *, rseq, u32, rseq_len, int, flags, u32 return -EFAULT; } +static long rseq_unregister(struct rseq __user * rseq, u32 rseq_len, int flags, u32 sig) +{ + if (flags & ~RSEQ_FLAG_UNREGISTER) + return -EINVAL; + if (current->rseq.usrptr != rseq || !current->rseq.usrptr) + return -EINVAL; + if (rseq_len != current->rseq.len) + return -EINVAL; + if (current->rseq.sig != sig) + return -EPERM; + if (!rseq_reset_ids()) + return -EFAULT; + rseq_reset(current); + return 0; +} + +static long rseq_reregister(struct rseq __user * rseq, u32 rseq_len, u32 sig) +{ + /* + * If rseq is already registered, check whether the provided address + * differs from the prior one. + */ + if (current->rseq.usrptr != rseq || rseq_len != current->rseq.len) + return -EINVAL; + if (current->rseq.sig != sig) + return -EPERM; + /* Already registered. */ + return -EBUSY; +} + +static bool rseq_length_valid(struct rseq __user *rseq, unsigned int rseq_len) +{ + /* + * Ensure the provided rseq is properly aligned, as communicated to + * user-space through the ELF auxiliary vector AT_RSEQ_ALIGN. If + * rseq_len is the original rseq size, the required alignment is the + * original struct rseq alignment. + * + * In order to be valid, rseq_len is either the original rseq size, or + * large enough to contain all supported fields, as communicated to + * user-space through the ELF auxiliary vector AT_RSEQ_FEATURE_SIZE. + */ + if (rseq_len < ORIG_RSEQ_SIZE) + return false; + + if (rseq_len == ORIG_RSEQ_SIZE) + return IS_ALIGNED((unsigned long)rseq, ORIG_RSEQ_SIZE); + + return IS_ALIGNED((unsigned long)rseq, rseq_alloc_align()) && + rseq_len >= offsetof(struct rseq, end); +} + +#define RSEQ_FLAGS_SUPPORTED (RSEQ_FLAG_SLICE_EXT_DEFAULT_ON) + +/* + * sys_rseq - Register or unregister restartable sequences for the caller thread. + */ +SYSCALL_DEFINE4(rseq, struct rseq __user *, rseq, u32, rseq_len, int, flags, u32, sig) +{ + if (flags & RSEQ_FLAG_UNREGISTER) + return rseq_unregister(rseq, rseq_len, flags, sig); + + if (unlikely(flags & ~RSEQ_FLAGS_SUPPORTED)) + return -EINVAL; + + if (current->rseq.usrptr) + return rseq_reregister(rseq, rseq_len, sig); + + if (!rseq_length_valid(rseq, rseq_len)) + return -EINVAL; + + return rseq_register(rseq, rseq_len, flags, sig); +} + #ifdef CONFIG_RSEQ_SLICE_EXTENSION struct slice_timer { struct hrtimer timer; From e744060076871eebc2647b24420b550ff44b2b65 Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Sat, 25 Apr 2026 14:48:23 +0200 Subject: [PATCH 160/333] selftests/rseq: Expand for optimized RSEQ ABI v2 Update the selftests so they are executed for legacy (32 bytes RSEQ region) and optimized RSEQ ABI v2 mode. Fixes: d6200245c75e ("rseq: Allow registering RSEQ with slice extension") Signed-off-by: Thomas Gleixner Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Dmitry Vyukov Tested-by: Dmitry Vyukov Link: https://patch.msgid.link/20260428224428.009121296%40kernel.org Cc: stable@vger.kernel.org --- tools/testing/selftests/rseq/Makefile | 11 ++++-- .../testing/selftests/rseq/check_optimized.c | 17 ++++++++ tools/testing/selftests/rseq/param_test.c | 25 +++++++----- .../testing/selftests/rseq/run_param_test.sh | 39 +++++++++++++++++++ .../selftests/rseq/run_timeslice_test.sh | 14 +++++++ tools/testing/selftests/rseq/slice_test.c | 2 +- 6 files changed, 95 insertions(+), 13 deletions(-) create mode 100644 tools/testing/selftests/rseq/check_optimized.c create mode 100755 tools/testing/selftests/rseq/run_timeslice_test.sh diff --git a/tools/testing/selftests/rseq/Makefile b/tools/testing/selftests/rseq/Makefile index 0293a2f17f51..50d69e22ee7a 100644 --- a/tools/testing/selftests/rseq/Makefile +++ b/tools/testing/selftests/rseq/Makefile @@ -15,7 +15,7 @@ LDLIBS += -lpthread -ldl OVERRIDE_TARGETS = 1 TEST_GEN_PROGS = basic_test basic_percpu_ops_test basic_percpu_ops_mm_cid_test \ - param_test_benchmark param_test_mm_cid_benchmark slice_test + param_test_benchmark param_test_mm_cid_benchmark TEST_GEN_PROGS_EXTENDED = librseq.so \ param_test \ @@ -23,9 +23,11 @@ TEST_GEN_PROGS_EXTENDED = librseq.so \ param_test_mm_cid \ param_test_mm_cid_compare_twice \ syscall_errors_test \ - legacy_check + legacy_check \ + slice_test \ + check_optimized -TEST_PROGS = run_param_test.sh run_syscall_errors_test.sh run_legacy_check.sh +TEST_PROGS = run_param_test.sh run_syscall_errors_test.sh run_legacy_check.sh run_timeslice_test.sh TEST_FILES := settings @@ -66,3 +68,6 @@ $(OUTPUT)/syscall_errors_test: syscall_errors_test.c $(TEST_GEN_PROGS_EXTENDED) $(OUTPUT)/slice_test: slice_test.c $(TEST_GEN_PROGS_EXTENDED) rseq.h rseq-*.h $(CC) $(CFLAGS) $< $(LDLIBS) -lrseq -o $@ + +$(OUTPUT)/check_optimized: check_optimized.c $(TEST_GEN_PROGS_EXTENDED) rseq.h rseq-*.h + $(CC) $(CFLAGS) $< $(LDLIBS) -lrseq -o $@ diff --git a/tools/testing/selftests/rseq/check_optimized.c b/tools/testing/selftests/rseq/check_optimized.c new file mode 100644 index 000000000000..a13e3f2c8fc6 --- /dev/null +++ b/tools/testing/selftests/rseq/check_optimized.c @@ -0,0 +1,17 @@ +// SPDX-License-Identifier: LGPL-2.1 +#define _GNU_SOURCE +#include +#include +#include +#include +#include +#include + +#include "rseq.h" + +int main(int argc, char **argv) +{ + if (__rseq_register_current_thread(true, false)) + return -1; + return 0; +} diff --git a/tools/testing/selftests/rseq/param_test.c b/tools/testing/selftests/rseq/param_test.c index 05d03e679e06..e1e98dbabe4b 100644 --- a/tools/testing/selftests/rseq/param_test.c +++ b/tools/testing/selftests/rseq/param_test.c @@ -38,7 +38,7 @@ static int opt_modulo, verbose; static int opt_yield, opt_signal, opt_sleep, opt_disable_rseq, opt_threads = 200, opt_disable_mod = 0, opt_test = 's'; - +static bool opt_rseq_legacy; static long long opt_reps = 5000; static __thread __attribute__((tls_model("initial-exec"))) @@ -281,9 +281,12 @@ unsigned int yield_mod_cnt, nr_abort; } \ } +#define rseq_no_glibc true + #else #define printf_verbose(fmt, ...) +#define rseq_no_glibc false #endif /* BENCHMARK */ @@ -481,7 +484,7 @@ void *test_percpu_spinlock_thread(void *arg) long long i, reps; if (!opt_disable_rseq && thread_data->reg && - rseq_register_current_thread()) + __rseq_register_current_thread(rseq_no_glibc, opt_rseq_legacy)) abort(); reps = thread_data->reps; for (i = 0; i < reps; i++) { @@ -558,7 +561,7 @@ void *test_percpu_inc_thread(void *arg) long long i, reps; if (!opt_disable_rseq && thread_data->reg && - rseq_register_current_thread()) + __rseq_register_current_thread(rseq_no_glibc, opt_rseq_legacy)) abort(); reps = thread_data->reps; for (i = 0; i < reps; i++) { @@ -712,7 +715,7 @@ void *test_percpu_list_thread(void *arg) long long i, reps; struct percpu_list *list = (struct percpu_list *)arg; - if (!opt_disable_rseq && rseq_register_current_thread()) + if (!opt_disable_rseq && __rseq_register_current_thread(rseq_no_glibc, opt_rseq_legacy)) abort(); reps = opt_reps; @@ -895,7 +898,7 @@ void *test_percpu_buffer_thread(void *arg) long long i, reps; struct percpu_buffer *buffer = (struct percpu_buffer *)arg; - if (!opt_disable_rseq && rseq_register_current_thread()) + if (!opt_disable_rseq && __rseq_register_current_thread(rseq_no_glibc, opt_rseq_legacy)) abort(); reps = opt_reps; @@ -1105,7 +1108,7 @@ void *test_percpu_memcpy_buffer_thread(void *arg) long long i, reps; struct percpu_memcpy_buffer *buffer = (struct percpu_memcpy_buffer *)arg; - if (!opt_disable_rseq && rseq_register_current_thread()) + if (!opt_disable_rseq && __rseq_register_current_thread(rseq_no_glibc, opt_rseq_legacy)) abort(); reps = opt_reps; @@ -1258,7 +1261,7 @@ void *test_membarrier_worker_thread(void *arg) const int iters = opt_reps; int i; - if (rseq_register_current_thread()) { + if (__rseq_register_current_thread(rseq_no_glibc, opt_rseq_legacy)) { fprintf(stderr, "Error: rseq_register_current_thread(...) failed(%d): %s\n", errno, strerror(errno)); abort(); @@ -1323,7 +1326,7 @@ void *test_membarrier_manager_thread(void *arg) intptr_t expect_a = 0, expect_b = 0; int cpu_a = 0, cpu_b = 0; - if (rseq_register_current_thread()) { + if (__rseq_register_current_thread(rseq_no_glibc, opt_rseq_legacy)) { fprintf(stderr, "Error: rseq_register_current_thread(...) failed(%d): %s\n", errno, strerror(errno)); abort(); @@ -1475,6 +1478,7 @@ static void show_usage(int argc, char **argv) printf(" [-D M] Disable rseq for each M threads\n"); printf(" [-T test] Choose test: (s)pinlock, (l)ist, (b)uffer, (m)emcpy, (i)ncrement, membarrie(r)\n"); printf(" [-M] Push into buffer and memcpy buffer with memory barriers.\n"); + printf(" [-O] Test with optimized RSEQ\n"); printf(" [-v] Verbose output.\n"); printf(" [-h] Show this help.\n"); printf("\n"); @@ -1602,6 +1606,9 @@ int main(int argc, char **argv) case 'M': opt_mo = RSEQ_MO_RELEASE; break; + case 'L': + opt_rseq_legacy = true; + break; default: show_usage(argc, argv); goto error; @@ -1618,7 +1625,7 @@ int main(int argc, char **argv) if (set_signal_handler()) goto error; - if (!opt_disable_rseq && rseq_register_current_thread()) + if (!opt_disable_rseq && __rseq_register_current_thread(rseq_no_glibc, opt_rseq_legacy)) goto error; if (!opt_disable_rseq && !rseq_validate_cpu_id()) { fprintf(stderr, "Error: cpu id getter unavailable\n"); diff --git a/tools/testing/selftests/rseq/run_param_test.sh b/tools/testing/selftests/rseq/run_param_test.sh index 8d31426ab41f..69a3fa049929 100755 --- a/tools/testing/selftests/rseq/run_param_test.sh +++ b/tools/testing/selftests/rseq/run_param_test.sh @@ -34,6 +34,11 @@ REPS=1000 SLOW_REPS=100 NR_THREADS=$((6*${NR_CPUS})) +# Prevent GLIBC from registering RSEQ so the selftest can run in legacy and +# performance optimized mode. +GLIBC_TUNABLES="${GLIBC_TUNABLES:-}:glibc.pthread.rseq=0" +export GLIBC_TUNABLES + function do_tests() { local i=0 @@ -103,6 +108,40 @@ function inject_blocking() NR_LOOPS= } +echo "Testing in legacy RSEQ mode" +echo "Yield injection (25%)" +inject_blocking -m 4 -y -L + +echo "Yield injection (50%)" +inject_blocking -m 2 -y -L + +echo "Yield injection (100%)" +inject_blocking -m 1 -y -L + +echo "Kill injection (25%)" +inject_blocking -m 4 -k -L + +echo "Kill injection (50%)" +inject_blocking -m 2 -k -L + +echo "Kill injection (100%)" +inject_blocking -m 1 -k -L + +echo "Sleep injection (1ms, 25%)" +inject_blocking -m 4 -s 1 -L + +echo "Sleep injection (1ms, 50%)" +inject_blocking -m 2 -s 1 -L + +echo "Sleep injection (1ms, 100%)" +inject_blocking -m 1 -s 1 -L + +./check_optimized || { + echo "Skipping optimized RSEQ mode test. Not supported"; + exit 0 +} + +echo "Testing in optimized RSEQ mode" echo "Yield injection (25%)" inject_blocking -m 4 -y diff --git a/tools/testing/selftests/rseq/run_timeslice_test.sh b/tools/testing/selftests/rseq/run_timeslice_test.sh new file mode 100755 index 000000000000..551ebed71ec6 --- /dev/null +++ b/tools/testing/selftests/rseq/run_timeslice_test.sh @@ -0,0 +1,14 @@ +#!/bin/bash +# SPDX-License-Identifier: GPL-2.0+ + +# Prevent GLIBC from registering RSEQ so the selftest can run in legacy +# and performance optimized mode. +GLIBC_TUNABLES="${GLIBC_TUNABLES:-}:glibc.pthread.rseq=0" +export GLIBC_TUNABLES + +./check_optimized || { + echo "Skipping optimized RSEQ mode test. Not supported"; + exit 0 +} + +./slice_test diff --git a/tools/testing/selftests/rseq/slice_test.c b/tools/testing/selftests/rseq/slice_test.c index 77e668ff74d7..e402d4440bc2 100644 --- a/tools/testing/selftests/rseq/slice_test.c +++ b/tools/testing/selftests/rseq/slice_test.c @@ -124,7 +124,7 @@ FIXTURE_SETUP(slice_ext) { cpu_set_t affinity; - if (rseq_register_current_thread()) + if (__rseq_register_current_thread(true, false)) SKIP(return, "RSEQ not supported\n"); if (prctl(PR_RSEQ_SLICE_EXTENSION, PR_RSEQ_SLICE_EXTENSION_SET, From b6eee96843e8d088200f01b035da98e72067c5fe Mon Sep 17 00:00:00 2001 From: Zhan Xusheng Date: Fri, 1 May 2026 12:40:06 +0200 Subject: [PATCH 161/333] sched/fair: Fix overflow in vruntime_eligible() Zhan Xusheng reported running into sporadic a s64 mult overflow in vruntime_eligible(). When constructing a worst case scenario: If you have cgroups, then you can have an entity of weight 2 (per calc_group_shares()), and its vlag should then be bounded by: (slice+TICK_NSEC) * NICE_0_LOAD, which is around 44 bits as per the comment on entity_key(). The other extreme is 100*NICE_0_LOAD, thus you get: {key, weight}[] := { puny: { (slice + TICK_NSEC) * NICE_0_LOAD, 2 }, max: { 0, 100*NICE_0_LOAD }, } The avg_vruntime() would end up being very close to 0 (which is zero_vruntime), so no real help making that more accurate. vruntime_eligible(puny) ends up with: avg = 2 * puny.key (+ 0) load = 2 + 100 * NICE_0_LOAD avg >= puny.key * load And that is: (slice + TICK_NSEC) * NICE_0_LOAD * NICE_0_LOAD * 100, which will overflow s64. Zhan suggested using __builtin_mul_overflow(), however after staring at compiler output for various architectures using godbolt, it seems that using an __int128 multiplication often results in better code. Specifically, a number of architectures already compute the __int128 product to determine the overflow. Eg. arm64 already has the 'smulh' instruction used. By explicitly doing an __int128 multiply, it will emit the 'mul; smulh' pattern, which modern cores can fuse (armv8-a clang-22.1.0). x86_64 has less branches (no OF handling). Since Linux has ARCH_SUPPORTS_INT128 to gate __int128 usage, also provide the __builtin_mul_overflow() variant as a fallback. [peterz: Changelog and __int128 bits] Fixes: 556146ce5e94 ("sched/fair: Avoid overflow in enqueue_entity()") Reported-by: Zhan Xusheng Closes: https://patch.msgid.link/20260415145742.10359-1-zhanxusheng%40xiaomi.com Signed-off-by: Zhan Xusheng Signed-off-by: Peter Zijlstra (Intel) Link: https://patch.msgid.link/20260505103155.GN3102924%40noisy.programming.kicks-ass.net --- kernel/sched/fair.c | 39 ++++++++++++++++++++++++++++++++++----- 1 file changed, 34 insertions(+), 5 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index 728965851842..b91c8b294229 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -882,11 +882,11 @@ bool update_entity_lag(struct cfs_rq *cfs_rq, struct sched_entity *se) * * lag_i >= 0 -> V >= v_i * - * \Sum (v_i - v)*w_i - * V = ------------------ + v + * \Sum (v_i - v0)*w_i + * V = ------------------- + v0 * \Sum w_i * - * lag_i >= 0 -> \Sum (v_i - v)*w_i >= (v_i - v)*(\Sum w_i) + * lag_i >= 0 -> \Sum (v_i - v0)*w_i >= (v_i - v0)*(\Sum w_i) * * Note: using 'avg_vruntime() > se->vruntime' is inaccurate due * to the loss in precision caused by the division. @@ -894,7 +894,7 @@ bool update_entity_lag(struct cfs_rq *cfs_rq, struct sched_entity *se) static int vruntime_eligible(struct cfs_rq *cfs_rq, u64 vruntime) { struct sched_entity *curr = cfs_rq->curr; - s64 avg = cfs_rq->sum_w_vruntime; + s64 key, avg = cfs_rq->sum_w_vruntime; long load = cfs_rq->sum_weight; if (curr && curr->on_rq) { @@ -904,7 +904,36 @@ static int vruntime_eligible(struct cfs_rq *cfs_rq, u64 vruntime) load += weight; } - return avg >= vruntime_op(vruntime, "-", cfs_rq->zero_vruntime) * load; + key = vruntime_op(vruntime, "-", cfs_rq->zero_vruntime); + + /* + * The worst case term for @key includes 'NSEC_TICK * NICE_0_LOAD' + * and @load obviously includes NICE_0_LOAD. NSEC_TICK is around 24 + * bits, while NICE_0_LOAD is 20 on 64bit and 10 otherwise. + * + * This gives that on 64bit the product will be at least 64bit which + * overflows s64, while on 32bit it will only be 44bits and should fit + * comfortably. + */ +#ifdef CONFIG_64BIT +#ifdef CONFIG_ARCH_SUPPORTS_INT128 + /* This often results in simpler code than __builtin_mul_overflow(). */ + return avg >= (__int128)key * load; +#else + s64 rhs; + /* + * On overflow, the sign of key tells us the correct answer: a large + * positive key means vruntime >> V, so not eligible; a large negative + * key means vruntime << V, so eligible. + */ + if (check_mul_overflow(key, load, &rhs)) + return key <= 0; + + return avg >= rhs; +#endif +#else /* 32bit */ + return avg >= key * load; +#endif } int entity_eligible(struct cfs_rq *cfs_rq, struct sched_entity *se) From 9f6d929ee2c6f0266edb564bcd2bd47fd6e884a8 Mon Sep 17 00:00:00 2001 From: Vincent Guittot Date: Sun, 3 May 2026 12:45:03 +0200 Subject: [PATCH 162/333] sched/fair: Fix wakeup_preempt_fair() for not waking up task Make sure to only call pick_next_entity() on an non-empty cfs_rq. The assumption that p is always enqueued and not delayed, is only true for wakeup. If p was moved while delayed, pick_next_entity() will dequeue it and the cfs might become empty. Test if there are still queued tasks before trying again to determine if p could be the next one to be picked. There are at least 2 cases: When cfs becomes idle, it tries to pull tasks but if those pulled tasks are delayed, they will be dequeued when attached to cfs. attach_tasks() -> attach_task() -> wakeup_preempt(rq, p, 0); A misfit task running on cfs A triggers a load balance to be pulled on a better cpu, the load balance on cfs B starts an active load balance to pulled the running misfit task. If there is a delayed dequeue task on cfs A, it can be pulled instead of the previously running misfit task. attach_one_task() -> attach_task() -> wakeup_preempt(rq, p, 0); Fixes: ac8e69e69363 ("sched/fair: Fix wakeup_preempt_fair() vs delayed dequeue") Signed-off-by: Vincent Guittot Signed-off-by: Peter Zijlstra (Intel) Link: https://patch.msgid.link/20260503104503.1732682-1-vincent.guittot@linaro.org --- kernel/sched/fair.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index b91c8b294229..3ebec186f982 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -9174,9 +9174,10 @@ static void wakeup_preempt_fair(struct rq *rq, struct task_struct *p, int wake_f /* * Because p is enqueued, nse being null can only mean that we - * dequeued a delayed task. + * dequeued a delayed task. If there are still entities queued in + * cfs, check if the next one will be p. */ - if (!nse) + if (!nse && cfs_rq->nr_queued) goto pick; if (sched_feat(RUN_TO_PARITY)) From 1f7305d87aa23db2579df222eba504a333c2c978 Mon Sep 17 00:00:00 2001 From: James Morse Date: Tue, 5 May 2026 17:52:03 +0100 Subject: [PATCH 163/333] KVM: arm64: Work around C1-Pro erratum 4193714 for protected guests C1-Pro cores with SME have an erratum where TLBI+DSB does not complete all outstanding SME accesses. Instead a DSB needs to be executed on the affected CPUs. The implication is that pages cannot be unmapped from the host Stage 2 and then provided to a protected guest or to the hypervisor. Host SME accesses may still complete after this point. This erratum breaks pKVM's guarantees, and the workaround is hard to implement as EL2 and EL1 share a security state meaning EL1 can mask IPIs sent by EL2, leading to interrupt blackouts. Instead, do this in EL3. This has the advantage of a separate security state, meaning lower EL cannot mask the IPI. It is also simpler for EL3 to know about CPUs that are off or in PSCI's CPU_SUSPEND. Add the needed hook to host_stage2_set_owner_metadata_locked(). This covers the cases where the host loses access to a page: __pkvm_host_donate_guest() __pkvm_guest_unshare_host() host_stage2_set_owner_locked() when owner_id == PKVM_ID_HYP Since pKVM relies on the firmware call for correctness, check for the firmware counterpart during protected KVM initialisation and fail the pKVM initialisation if it is missing. Signed-off-by: James Morse Co-developed-by: Catalin Marinas Signed-off-by: Catalin Marinas Cc: Mark Rutland Cc: Marc Zyngier Cc: Oliver Upton Cc: Will Deacon Cc: Vincent Donnefort Cc: Lorenzo Pieralisi Cc: Sudeep Holla Link: https://patch.msgid.link/20260505165205.2690919-1-catalin.marinas@arm.com Signed-off-by: Marc Zyngier --- arch/arm64/kvm/arm.c | 21 +++++++++++++++++++++ arch/arm64/kvm/hyp/nvhe/mem_protect.c | 23 ++++++++++++++++++++++- include/linux/arm-smccc.h | 6 ++++++ 3 files changed, 49 insertions(+), 1 deletion(-) diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index 8bb2c7422cc8..34c9950884d5 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -4,6 +4,7 @@ * Author: Christoffer Dall */ +#include #include #include #include @@ -2638,6 +2639,22 @@ static int init_pkvm_host_sve_state(void) return 0; } +static int pkvm_check_sme_dvmsync_fw_call(void) +{ + struct arm_smccc_res res; + + if (!cpus_have_final_cap(ARM64_WORKAROUND_4193714)) + return 0; + + arm_smccc_1_1_smc(ARM_SMCCC_CPU_WORKAROUND_4193714, &res); + if (res.a0) { + kvm_err("pKVM requires firmware support for C1-Pro erratum 4193714\n"); + return -ENODEV; + } + + return 0; +} + /* * Finalizes the initialization of hyp mode, once everything else is initialized * and the initialziation process cannot fail. @@ -2838,6 +2855,10 @@ static int __init init_hyp_mode(void) if (err) goto out_err; + err = pkvm_check_sme_dvmsync_fw_call(); + if (err) + goto out_err; + err = kvm_hyp_init_protection(hyp_va_bits); if (err) { kvm_err("Failed to init hyp memory protection\n"); diff --git a/arch/arm64/kvm/hyp/nvhe/mem_protect.c b/arch/arm64/kvm/hyp/nvhe/mem_protect.c index 28a471d1927c..a3050e2b65b1 100644 --- a/arch/arm64/kvm/hyp/nvhe/mem_protect.c +++ b/arch/arm64/kvm/hyp/nvhe/mem_protect.c @@ -5,6 +5,7 @@ */ #include + #include #include #include @@ -14,6 +15,7 @@ #include +#include #include #include #include @@ -29,6 +31,19 @@ static struct hyp_pool host_s2_pool; static DEFINE_PER_CPU(struct pkvm_hyp_vm *, __current_vm); #define current_vm (*this_cpu_ptr(&__current_vm)) +static void pkvm_sme_dvmsync_fw_call(void) +{ + if (alternative_has_cap_unlikely(ARM64_WORKAROUND_4193714)) { + struct arm_smccc_res res; + + /* + * Ignore the return value. Probing for the workaround + * availability took place in init_hyp_mode(). + */ + hyp_smccc_1_1_smc(ARM_SMCCC_CPU_WORKAROUND_4193714, &res); + } +} + static void guest_lock_component(struct pkvm_hyp_vm *vm) { hyp_spin_lock(&vm->lock); @@ -574,8 +589,14 @@ static int host_stage2_set_owner_metadata_locked(phys_addr_t addr, u64 size, ret = host_stage2_try(kvm_pgtable_stage2_annotate, &host_mmu.pgt, addr, size, &host_s2_pool, KVM_HOST_INVALID_PTE_TYPE_DONATION, annotation); - if (!ret) + if (!ret) { + /* + * After stage2 maintenance has happened, but before the page + * owner has changed. + */ + pkvm_sme_dvmsync_fw_call(); __host_update_page_state(addr, size, PKVM_NOPAGE); + } return ret; } diff --git a/include/linux/arm-smccc.h b/include/linux/arm-smccc.h index 50b47eba7d01..e7195750d21b 100644 --- a/include/linux/arm-smccc.h +++ b/include/linux/arm-smccc.h @@ -105,6 +105,12 @@ ARM_SMCCC_SMC_32, \ 0, 0x3fff) +/* C1-Pro erratum 4193714: SME DVMSync early acknowledgement */ +#define ARM_SMCCC_CPU_WORKAROUND_4193714 \ + ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL, \ + ARM_SMCCC_SMC_32, \ + ARM_SMCCC_OWNER_CPU, 0x10) + #define ARM_SMCCC_VENDOR_HYP_CALL_UID_FUNC_ID \ ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL, \ ARM_SMCCC_SMC_32, \ From 8d9b9d985ad3a81c751a6b97edaf1d3c0780af7c Mon Sep 17 00:00:00 2001 From: Wei-Lin Chang Date: Tue, 5 May 2026 15:47:35 +0100 Subject: [PATCH 164/333] KVM: arm64: nv: Consider the DS bit when translating TCR_EL2 When running an nVHE L1, TCR_EL2 is mapped to TCR_EL1. Writes to the register are trapped and written to TCR_EL1 after a translation. Booting an nVHE L1 with 52-bit VA isn't working because the translation was ignoring the DS bit set by the guest, hence causing repeating level 0 faults. Add it in the translation function. Signed-off-by: Wei-Lin Chang Link: https://patch.msgid.link/20260505144735.1496530-1-weilin.chang@arm.com Signed-off-by: Marc Zyngier --- arch/arm64/include/asm/kvm_nested.h | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/arm64/include/asm/kvm_nested.h b/arch/arm64/include/asm/kvm_nested.h index 091544e6af44..dc2957662ff2 100644 --- a/arch/arm64/include/asm/kvm_nested.h +++ b/arch/arm64/include/asm/kvm_nested.h @@ -23,6 +23,7 @@ static inline u64 tcr_el2_ps_to_tcr_el1_ips(u64 tcr_el2) static inline u64 translate_tcr_el2_to_tcr_el1(u64 tcr) { return TCR_EPD1_MASK | /* disable TTBR1_EL1 */ + ((tcr & TCR_EL2_DS) ? TCR_DS : 0) | ((tcr & TCR_EL2_TBI) ? TCR_TBI0 : 0) | tcr_el2_ps_to_tcr_el1_ips(tcr) | (tcr & TCR_EL2_TG0_MASK) | From 9be19df816dea9eb7dfe1661b3690bed6a2cb146 Mon Sep 17 00:00:00 2001 From: Alexandru Elisei Date: Tue, 5 May 2026 10:49:13 +0100 Subject: [PATCH 165/333] KVM: arm64: Handle permission faults with guest_memfd gmem_abort() calls kvm_pgtable_stage2_map() to make changes to stage 2. It does this for both relaxing permissions on an existing mapping and to install a missing mapping. kvm_pgtable_stage2_map() doesn't make changes to stage 2 if there is an existing, valid entry and the new entry modifies only the permissions. This is checked in: kvm_pgtable_stage2_map() stage2_map_walk_leaf() stage2_map_walker_try_leaf() stage2_pte_needs_update() and if only the permissions differ, kvm_pgtable_stage2_map() returns -EAGAIN and KVM returns to the guest to replay the instruction. The assumption is that a concurrent fault on a different VCPU already mapped the faulting IPA, and replaying the instruction will either succeed, or cause a permission fault, which should be handled with kvm_pgtable_stage2_relax_perms(). gmem_abort(), on a read or write fault on a system without DIC (instruction cache invalidation required for data to instruction coherence), installs a valid entry with read and write permissions, but without executable permissions. On an execution fault on the same page, gmem_abort() attempts to relax the permissions to allow execution, but calls kvm_pgtable_stage2_map() to change the existing, valid, entry. kvm_pgtable_stage2_map() returns -EAGAIN and KVM resumes execution from the faulting instruction, which leads to an infinite loop of permission faults on the same instruction. Allow the guest to make progress by using kvm_pgtable_stage2_relax_perms() to relax permissions. Fixes: a7b57e099592 ("KVM: arm64: Handle guest_memfd-backed guest page faults") Signed-off-by: Alexandru Elisei Reviewed-by: Fuad Tabba Link: https://patch.msgid.link/20260505094913.75317-1-alexandru.elisei@arm.com Signed-off-by: Marc Zyngier --- arch/arm64/kvm/mmu.c | 29 +++++++++++++++++++++-------- 1 file changed, 21 insertions(+), 8 deletions(-) diff --git a/arch/arm64/kvm/mmu.c b/arch/arm64/kvm/mmu.c index d089c107d9b7..4da9281312eb 100644 --- a/arch/arm64/kvm/mmu.c +++ b/arch/arm64/kvm/mmu.c @@ -1576,21 +1576,24 @@ struct kvm_s2_fault_desc { static int gmem_abort(const struct kvm_s2_fault_desc *s2fd) { bool write_fault, exec_fault; + bool perm_fault = kvm_vcpu_trap_is_permission_fault(s2fd->vcpu); enum kvm_pgtable_walk_flags flags = KVM_PGTABLE_WALK_SHARED; enum kvm_pgtable_prot prot = KVM_PGTABLE_PROT_R; struct kvm_pgtable *pgt = s2fd->vcpu->arch.hw_mmu->pgt; unsigned long mmu_seq; struct page *page; struct kvm *kvm = s2fd->vcpu->kvm; - void *memcache; + void *memcache = NULL; kvm_pfn_t pfn; gfn_t gfn; int ret; - memcache = get_mmu_memcache(s2fd->vcpu); - ret = topup_mmu_memcache(s2fd->vcpu, memcache); - if (ret) - return ret; + if (!perm_fault) { + memcache = get_mmu_memcache(s2fd->vcpu); + ret = topup_mmu_memcache(s2fd->vcpu, memcache); + if (ret) + return ret; + } if (s2fd->nested) gfn = kvm_s2_trans_output(s2fd->nested) >> PAGE_SHIFT; @@ -1631,9 +1634,19 @@ static int gmem_abort(const struct kvm_s2_fault_desc *s2fd) goto out_unlock; } - ret = KVM_PGT_FN(kvm_pgtable_stage2_map)(pgt, s2fd->fault_ipa, PAGE_SIZE, - __pfn_to_phys(pfn), prot, - memcache, flags); + if (perm_fault) { + /* + * Drop the SW bits in favour of those stored in the + * PTE, which will be preserved. + */ + prot &= ~KVM_NV_GUEST_MAP_SZ; + ret = KVM_PGT_FN(kvm_pgtable_stage2_relax_perms)(pgt, s2fd->fault_ipa, + prot, flags); + } else { + ret = KVM_PGT_FN(kvm_pgtable_stage2_map)(pgt, s2fd->fault_ipa, PAGE_SIZE, + __pfn_to_phys(pfn), prot, + memcache, flags); + } out_unlock: kvm_release_faultin_page(kvm, page, !!ret, prot & KVM_PGTABLE_PROT_W); From fc240715fc5003538ff530e3cfb985e7769b7171 Mon Sep 17 00:00:00 2001 From: Sebastian Ott Date: Mon, 4 May 2026 13:28:08 +0200 Subject: [PATCH 166/333] KVM: selftests: arm64: Fix steal_time test after UAPI refactoring Fix the following failure to the steal_time test on arm64 by making the timer address known to the guest. ==== Test Assertion Failure ==== steal_time.c:229: !ret pid=18514 tid=18514 errno=22 - Invalid argument 1 0x000000000040252f: check_steal_time_uapi at steal_time.c:229 (discriminator 20) 2 (inlined by) main at steal_time.c:537 (discriminator 20) 3 0x0000ffffa23d621b: ?? ??:0 4 0x0000ffffa23d62fb: ?? ??:0 5 0x0000000000402b6f: _start at ??:? KVM_SET_DEVICE_ATTR failed, rc: -1 errno: 22 (Invalid argument) Fixes: 40351ed924dd ("KVM: selftests: Refactor UAPI tests into dedicated function") Signed-off-by: Sebastian Ott Link: https://patch.msgid.link/20260504112808.21276-1-sebott@redhat.com Signed-off-by: Marc Zyngier --- tools/testing/selftests/kvm/steal_time.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tools/testing/selftests/kvm/steal_time.c b/tools/testing/selftests/kvm/steal_time.c index 7df2bc8eec02..76fcdd1fd3cb 100644 --- a/tools/testing/selftests/kvm/steal_time.c +++ b/tools/testing/selftests/kvm/steal_time.c @@ -220,6 +220,8 @@ static void check_steal_time_uapi(void) }; vcpu_ioctl(vcpu, KVM_HAS_DEVICE_ATTR, &dev); + vm_userspace_mem_region_add(vm, VM_MEM_SRC_ANONYMOUS, ST_GPA_BASE, 1, 1, 0); + virt_map(vm, ST_GPA_BASE, ST_GPA_BASE, 1); st_ipa = (ulong)ST_GPA_BASE | 1; ret = __vcpu_ioctl(vcpu, KVM_SET_DEVICE_ATTR, &dev); From 9a624ea3f26f40c76bd2c7f77cde30659d42efbd Mon Sep 17 00:00:00 2001 From: Mostafa Saleh Date: Thu, 30 Apr 2026 10:37:24 +0000 Subject: [PATCH 167/333] KVM: arm64: Remove potential UB on nvhe tracing clock update Sashiko(locally) reports possiblity of division by zero and out-of-bounds bitwise shift in trace_clock_update(). Although the clock update is untrusted, we should at least have some basic checks to avoid undefined behaviours. Reviewed-by: Vincent Donnefort Signed-off-by: Mostafa Saleh Link: https://patch.msgid.link/20260430103724.2151625-1-smostafa@google.com Signed-off-by: Marc Zyngier --- arch/arm64/kvm/hyp/nvhe/clock.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/arch/arm64/kvm/hyp/nvhe/clock.c b/arch/arm64/kvm/hyp/nvhe/clock.c index 32fc4313fe43..a7fc61976fd0 100644 --- a/arch/arm64/kvm/hyp/nvhe/clock.c +++ b/arch/arm64/kvm/hyp/nvhe/clock.c @@ -35,6 +35,9 @@ void trace_clock_update(u32 mult, u32 shift, u64 epoch_ns, u64 epoch_cyc) struct clock_data *clock = &trace_clock_data; u64 bank = clock->cur ^ 1; + if (!mult || shift >= 64) + return; + clock->data[bank].mult = mult; clock->data[bank].shift = shift; clock->data[bank].epoch_ns = epoch_ns; From 54ea22273eef67196f238263bc79f27da04d2114 Mon Sep 17 00:00:00 2001 From: Steffen Eiden Date: Tue, 28 Apr 2026 18:05:12 +0200 Subject: [PATCH 168/333] MAINTAINERS: Add Steffen as reviewer for KVM/arm64 KVM/arm64 and KVM/s390 will eventually share some code. Add me as a cross-reviewer from the s390 team to arm64 to help to keep both architectures in sync. Signed-off-by: Steffen Eiden Link: https://patch.msgid.link/20260428160527.1378085-16-seiden@linux.ibm.com [maz: rephrase commit message to use future tense, since this is merged ahead of the code] Signed-off-by: Marc Zyngier --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 882214b0e7db..3fe4ea59199c 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -14052,6 +14052,7 @@ KERNEL VIRTUAL MACHINE FOR ARM64 (KVM/arm64) M: Marc Zyngier M: Oliver Upton R: Joey Gouly +R: Steffen Eiden R: Suzuki K Poulose R: Zenghui Yu L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) From 5772f6535227ebd104065d80afa8ed3478d34c5c Mon Sep 17 00:00:00 2001 From: David Gow Date: Thu, 16 Apr 2026 14:57:43 +0800 Subject: [PATCH 169/333] x86/boot/e820: Re-enable BIOS fallback if e820 table is empty In commit: 157266edcc56 ("x86/boot/e820: Simplify append_e820_table() and remove restriction on single-entry tables") the check on the number of entries in the e820 table was removed. The intention was to support single-entry maps, but by removing the check entirely, we also skip the fallback (to, e.g., the BIOS 88h function). This means that if no E820 map is passed in from the bootloader (which is the case on some bootloaders, like linld), we end up with an empty memory map, and the kernel fails to boot (either by deadlocking on OOM, or by failing to allocate the real mode trampoline, or similar). Re-instate the check in append_e820_table(), but only check that nr_entries is non-zero. This allows e820__memory_setup_default() to fall back to other memory size sources, and doesn't affect e820__memory_setup_extended(), as the latter ignores the return value from append_e820_table(). In doing so, we also update the return values to be proper error codes, with -ENOENT for this case (there are no entries), and -EINVAL for the case where an entry appears invalid. Given none of the callers check the actual value -- just whether it's nonzero -- this is largely aesthetic in practice. Tested against linld, and the kernel boots again fine. [ mingo: Readability edits to the comment and the changelog. ] Fixes: 157266edcc56 ("x86/boot/e820: Simplify append_e820_table() and remove restriction on single-entry tables") Signed-off-by: David Gow Signed-off-by: Ingo Molnar Reviewed-by: Andy Shevchenko Cc: stable@vger.kernel.org Cc: Arnd Bergmann Cc: "H. Peter Anvin" Link: https://patch.msgid.link/20260416065746.1896647-1-david@davidgow.net --- arch/x86/kernel/e820.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/arch/x86/kernel/e820.c b/arch/x86/kernel/e820.c index 2a9992758933..eb72537bc0b1 100644 --- a/arch/x86/kernel/e820.c +++ b/arch/x86/kernel/e820.c @@ -450,6 +450,10 @@ __init static int append_e820_table(struct boot_e820_entry *entries, u32 nr_entr { struct boot_e820_entry *entry = entries; + /* If there aren't any entries, we'll want to fall back to another source: */ + if (!nr_entries) + return -ENOENT; + while (nr_entries) { u64 start = entry->addr; u64 size = entry->size; @@ -458,7 +462,7 @@ __init static int append_e820_table(struct boot_e820_entry *entries, u32 nr_entr /* Ignore the remaining entries on 64-bit overflow: */ if (start > end && likely(size)) - return -1; + return -EINVAL; e820__range_add(start, size, type); From b15838b03cd0c6cf35651cfde62d17f14bb1d566 Mon Sep 17 00:00:00 2001 From: Myeonghun Pak Date: Fri, 24 Apr 2026 21:34:28 +0900 Subject: [PATCH 170/333] drm/bochs: Drop manual put on probe error path bochs_pci_probe() allocates the DRM device with devm_drm_dev_alloc(), which registers a devres action to drop the initial DRM device reference on driver detach or probe failure. The error path currently calls drm_dev_put() manually. If probe then returns an error, devres will run the registered release action and put the same device again, after the first put may already have released it. Return the probe error directly and let devres own the final put. Signed-off-by: Myeonghun Pak Fixes: 04826f588682 ("drm/bochs: Allocate DRM device in struct bochs_device") Signed-off-by: Thomas Zimmermann Reviewed-by: Thomas Zimmermann Link: https://patch.msgid.link/20260424123506.32275-1-mhun512@gmail.com --- drivers/gpu/drm/tiny/bochs.c | 10 +++------- 1 file changed, 3 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/tiny/bochs.c b/drivers/gpu/drm/tiny/bochs.c index 222e4ae1abbd..5d8dc5efec77 100644 --- a/drivers/gpu/drm/tiny/bochs.c +++ b/drivers/gpu/drm/tiny/bochs.c @@ -761,25 +761,21 @@ static int bochs_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent ret = pcim_enable_device(pdev); if (ret) - goto err_free_dev; + return ret; pci_set_drvdata(pdev, dev); ret = bochs_load(bochs); if (ret) - goto err_free_dev; + return ret; ret = drm_dev_register(dev, 0); if (ret) - goto err_free_dev; + return ret; drm_client_setup(dev, NULL); return ret; - -err_free_dev: - drm_dev_put(dev); - return ret; } static void bochs_pci_remove(struct pci_dev *pdev) From 3051cd060fa496df42954291fa2306ed2eab4ecc Mon Sep 17 00:00:00 2001 From: Weiming Shi Date: Wed, 6 May 2026 01:55:11 +0800 Subject: [PATCH 171/333] i2c: smbus: reject oversized block transfers in the common path The SMBus block transfer length data->block[0] is validated in i2c_smbus_xfer_emulated() but that check runs too late for tracepoints and is skipped entirely when the adapter provides a native smbus_xfer implementation. This allows user-controlled oversized block lengths to reach tracepoint memcpy calls and driver callbacks unchecked. Add an early validation in __i2c_smbus_xfer() that rejects block transfers whose caller-supplied length is zero or exceeds I2C_SMBUS_BLOCK_MAX before any tracepoint fires or driver callback runs. data->block[0] is filled in by the device on SMBus block reads, so the check is scoped to operations where the length is actually supplied by the caller. This is consistent with the existing -EINVAL convention in the emulated path and protects all downstream consumers at once: the smbus_write tracepoint, all native smbus_xfer driver implementations, and the emulated path. Two distinct bugs are fixed by this change: Bug 1: smbus_write tracepoint OOB (include/trace/events/smbus.h) trace_smbus_write() fires before any validation and copies data->block[0]+1 bytes into a 34-byte event buffer. With block[0]=0xfe the tracepoint copies 255 bytes, overflowing by 221. BUG: KASAN: stack-out-of-bounds in trace_event_raw_event_smbus_write+0x27c/0x530 Read of size 255 at addr ffff88800d98fcf8 by task poc_smbus/91 Call Trace: __asan_memcpy+0x23/0x80 trace_event_raw_event_smbus_write+0x27c/0x530 __i2c_smbus_xfer+0x43a/0xa40 i2c_smbus_xfer+0x19e/0x340 i2cdev_ioctl_smbus+0x38f/0x7f0 i2cdev_ioctl+0x35e/0x680 __x64_sys_ioctl+0x147/0x1e0 do_syscall_64+0xcf/0x15a0 entry_SYSCALL_64_after_hwframe+0x76/0x7e Bug 2: i2c-stub I2C_SMBUS_I2C_BLOCK_DATA OOB (drivers/i2c/i2c-stub.c) stub_xfer() implements .smbus_xfer directly and only clamps block[0] against 256-command, not I2C_SMBUS_BLOCK_MAX. With block[0]=0xff and command=0 the loop accesses block[1+i] for i up to 254, far past the 34-byte union. UBSAN: array-index-out-of-bounds in drivers/i2c/i2c-stub.c:223:44 index 34 is out of range for type '__u8 [34]' Call Trace: __ubsan_handle_out_of_bounds+0xd7/0x120 stub_xfer+0x1971/0x198f [i2c_stub] __i2c_smbus_xfer+0x306/0xa40 i2c_smbus_xfer+0x19e/0x340 i2cdev_ioctl_smbus+0x38f/0x7f0 i2cdev_ioctl+0x35e/0x680 __x64_sys_ioctl+0x147/0x1e0 do_syscall_64+0xcf/0x15a0 entry_SYSCALL_64_after_hwframe+0x76/0x7e Both traces reproduced on v7.0-rc6+i2c/for-current with KASAN+UBSAN. Fixes: 8a325997d95d ("i2c: Add message transfer tracepoints for SMBUS [ver #2]") Fixes: 4710317891e4 ("i2c-stub: Implement I2C block support") Reported-by: Xiang Mei Signed-off-by: Weiming Shi Signed-off-by: Wolfram Sang --- drivers/i2c/i2c-core-smbus.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/drivers/i2c/i2c-core-smbus.c b/drivers/i2c/i2c-core-smbus.c index 71eb1ef56f0c..ad6acb5ebadc 100644 --- a/drivers/i2c/i2c-core-smbus.c +++ b/drivers/i2c/i2c-core-smbus.c @@ -566,6 +566,18 @@ s32 __i2c_smbus_xfer(struct i2c_adapter *adapter, u16 addr, if (res) return res; + /* Reject invalid caller-supplied block lengths before any + * tracepoint or native smbus_xfer callback runs. + */ + if (data && + (protocol == I2C_SMBUS_I2C_BLOCK_DATA || + protocol == I2C_SMBUS_BLOCK_PROC_CALL || + (protocol == I2C_SMBUS_BLOCK_DATA && + read_write == I2C_SMBUS_WRITE)) && + (data->block[0] == 0 || + data->block[0] > I2C_SMBUS_BLOCK_MAX)) + return -EINVAL; + /* If enabled, the following two tracepoints are conditional on * read_write and protocol. */ From d7396a72eae795d7f968fb451237b6ac1616d712 Mon Sep 17 00:00:00 2001 From: Fuad Tabba Date: Fri, 1 May 2026 12:21:44 +0100 Subject: [PATCH 172/333] KVM: arm64: Make EL2 exception entry and exit context-synchronization events SCTLR_EL2.EIS and SCTLR_EL2.EOS control whether exception entry and exit at EL2 are Context Synchronisation Events (CSEs). Per ARM DDI 0487 M.b D24.2.175 (p. D24-9754): - !FEAT_ExS: the bit is RES1, so the entry/exit is unconditionally a CSE. - FEAT_ExS: the reset value is architecturally UNKNOWN; software must set the bit to make the entry/exit a CSE. INIT_SCTLR_EL2_MMU_ON in arch/arm64/include/asm/sysreg.h sets neither bit. KVM/arm64 hot paths rely on ERET from EL2 being a CSE, and on synchronous EL1->EL2 entry being a CSE, to elide explicit ISBs after MSRs to context-switching system registers (HCR_EL2, ZCR_EL2, ptrauth keys, etc.). On FEAT_ExS hardware those reliances are not architecturally backed unless EOS=1 (and, for entry, EIS=1). Until commit 0a35bd285f43 ("arm64: Convert SCTLR_EL2 to sysreg infrastructure"), SCTLR_EL2_RES1 was a hand-rolled mask that included BIT(11) (EOS) and BIT(22) (EIS), so INIT_SCTLR_EL2_MMU_ON was setting both unconditionally. The conversion made SCTLR_EL2_RES1 auto-generated; because the sysreg tooling only models unconditionally-RES1 fields and EIS/EOS are RES1 only when FEAT_ExS is absent, the auto-generated mask is UL(0). The seven other bits dropped from the old mask (positions 4, 5, 16, 18, 23, 28, 29) are unconditionally RES1 in the E2H=0 SCTLR_EL2 layout per DDI 0487 M.b D24.2.175, so dropping them is harmless. EIS and EOS are the only bits whose semantics changed for FEAT_ExS hardware and where the kernel relies on the value being 1. Make the guarantee explicit: include SCTLR_ELx_EIS | SCTLR_ELx_EOS in INIT_SCTLR_EL2_MMU_ON so that EL2 exception entry and exit are unconditionally CSEs regardless of whether FEAT_ExS is implemented. This matches the pairing in arch/arm64/kvm/config.c which treats EIS and EOS together as RES1 under !FEAT_ExS. Fixes: 0a35bd285f43 ("arm64: Convert SCTLR_EL2 to sysreg infrastructure") Reviewed-by: Yuan Yao Assisted-by: Gemini:gemini-3.1-pro review-prompts Signed-off-by: Fuad Tabba Link: https://patch.msgid.link/20260501112149.2824881-2-tabba@google.com Signed-off-by: Marc Zyngier --- arch/arm64/include/asm/sysreg.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/arm64/include/asm/sysreg.h b/arch/arm64/include/asm/sysreg.h index 736561480f36..7aa08d59d494 100644 --- a/arch/arm64/include/asm/sysreg.h +++ b/arch/arm64/include/asm/sysreg.h @@ -844,7 +844,7 @@ #define INIT_SCTLR_EL2_MMU_ON \ (SCTLR_ELx_M | SCTLR_ELx_C | SCTLR_ELx_SA | SCTLR_ELx_I | \ SCTLR_ELx_IESB | SCTLR_ELx_WXN | ENDIAN_SET_EL2 | \ - SCTLR_ELx_ITFSB | SCTLR_EL2_RES1) + SCTLR_ELx_ITFSB | SCTLR_ELx_EIS | SCTLR_ELx_EOS | SCTLR_EL2_RES1) #define INIT_SCTLR_EL2_MMU_OFF \ (SCTLR_EL2_RES1 | ENDIAN_SET_EL2) From 300fac4cc266b7782d88602b6b6a7faf31ce6405 Mon Sep 17 00:00:00 2001 From: Fuad Tabba Date: Fri, 1 May 2026 12:21:45 +0100 Subject: [PATCH 173/333] KVM: arm64: Guard against NULL vcpu on VHE hyp panic path On VHE, __hyp_call_panic() unconditionally calls __deactivate_traps(vcpu) on the vcpu pointer read from host_ctxt->__hyp_running_vcpu. That pointer is cleared after every guest exit (and is never set when no guest is running), so an unexpected EL2 exception landing in _guest_exit_panic, e.g. via the el2t*_invalid / el2h_irq_invalid vectors - reaches this function with vcpu == NULL. __deactivate_traps() then dereferences vcpu via ___deactivate_traps() -> vserror_state_is_nested() -> vcpu_has_nv() -> vcpu->arch.features, faulting inside the panic handler and obscuring the original failure. The nVHE counterpart (hyp_panic() in arch/arm64/kvm/hyp/nvhe/switch.c) already guards its vcpu-using cleanup with "if (vcpu)"; mirror that here. sysreg_restore_host_state_vhe() does not depend on vcpu and continues to run unconditionally, preserving panic forensics. The trailing panic("...VCPU:%p", vcpu) prints "(null)" safely via printk's %p handling. Fixes: 6a0259ed29bb ("KVM: arm64: Remove hyp_panic arguments") Assisted-by: Gemini:gemini-3.1-pro review-prompts Signed-off-by: Fuad Tabba Link: https://patch.msgid.link/20260501112149.2824881-3-tabba@google.com Signed-off-by: Marc Zyngier --- arch/arm64/kvm/hyp/vhe/switch.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/arch/arm64/kvm/hyp/vhe/switch.c b/arch/arm64/kvm/hyp/vhe/switch.c index 9db3f11a4754..1e8995add14f 100644 --- a/arch/arm64/kvm/hyp/vhe/switch.c +++ b/arch/arm64/kvm/hyp/vhe/switch.c @@ -663,7 +663,8 @@ static void __noreturn __hyp_call_panic(u64 spsr, u64 elr, u64 par) host_ctxt = host_data_ptr(host_ctxt); vcpu = host_ctxt->__hyp_running_vcpu; - __deactivate_traps(vcpu); + if (vcpu) + __deactivate_traps(vcpu); sysreg_restore_host_state_vhe(host_ctxt); panic("HYP panic:\nPS:%08llx PC:%016llx ESR:%08llx\nFAR:%016llx HPFAR:%016llx PAR:%016llx\nVCPU:%p\n", From d4d215e5b81ba5acb17752cab12c514a8062bada Mon Sep 17 00:00:00 2001 From: Fuad Tabba Date: Fri, 1 May 2026 12:21:46 +0100 Subject: [PATCH 174/333] KVM: arm64: Fix __deactivate_fgt macro parameter typo __deactivate_fgt() declares its first parameter as "htcxt" but the body references "hctxt". The parameter is unused; the macro silently captures "hctxt" from the enclosing scope. Both existing callers (__deactivate_traps_hfgxtr() and __deactivate_traps_ich_hfgxtr()) happen to define a local "struct kvm_cpu_context *hctxt", so the macro works by coincidence. A future caller without an "hctxt" local in scope, or naming it differently, would compile but bind to the wrong context. Align the parameter name with the sibling __activate_fgt() macro. The "vcpu" parameter remains unused in the body, kept for API symmetry with __activate_fgt() (which uses it). Fixes: f5a5a406b4b8 ("KVM: arm64: Propagate and handle Fine-Grained UNDEF bits") Assisted-by: Gemini:gemini-3.1-pro review-prompts Signed-off-by: Fuad Tabba Link: https://patch.msgid.link/20260501112149.2824881-4-tabba@google.com Signed-off-by: Marc Zyngier --- arch/arm64/kvm/hyp/include/hyp/switch.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/arm64/kvm/hyp/include/hyp/switch.h b/arch/arm64/kvm/hyp/include/hyp/switch.h index 98b2976837b1..bf0eb5e43427 100644 --- a/arch/arm64/kvm/hyp/include/hyp/switch.h +++ b/arch/arm64/kvm/hyp/include/hyp/switch.h @@ -245,7 +245,7 @@ static inline void __activate_traps_ich_hfgxtr(struct kvm_vcpu *vcpu) __activate_fgt(hctxt, vcpu, ICH_HFGITR_EL2); } -#define __deactivate_fgt(htcxt, vcpu, reg) \ +#define __deactivate_fgt(hctxt, vcpu, reg) \ do { \ write_sysreg_s(ctxt_sys_reg(hctxt, reg), \ SYS_ ## reg); \ From 5130d450d1488e62e1b5310f41910a3c7320e827 Mon Sep 17 00:00:00 2001 From: Fuad Tabba Date: Fri, 1 May 2026 12:21:47 +0100 Subject: [PATCH 175/333] KVM: arm64: Seed pkvm_ownership_selftest vcpu memcache The hypercall handlers call pkvm_refill_memcache() to top up the hyp_vcpu memcache before invoking __pkvm_host_{share,donate}_guest(). pkvm_ownership_selftest invokes those functions directly with a static selftest_vcpu that has an empty memcache. Seed selftest_vcpu's memcache from the prepopulated selftest pages, leaving the remainder for selftest_vm.pool. Required by the memcache-sufficiency pre-check added in the following patches. Assisted-by: Gemini:gemini-3.1-pro review-prompts Signed-off-by: Fuad Tabba Link: https://patch.msgid.link/20260501112149.2824881-5-tabba@google.com Signed-off-by: Marc Zyngier --- arch/arm64/kvm/hyp/nvhe/pkvm.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/arch/arm64/kvm/hyp/nvhe/pkvm.c b/arch/arm64/kvm/hyp/nvhe/pkvm.c index e7496eb85628..eb1c10120f9f 100644 --- a/arch/arm64/kvm/hyp/nvhe/pkvm.c +++ b/arch/arm64/kvm/hyp/nvhe/pkvm.c @@ -752,16 +752,30 @@ static struct pkvm_hyp_vcpu selftest_vcpu = { struct pkvm_hyp_vcpu *init_selftest_vm(void *virt) { struct hyp_page *p = hyp_virt_to_page(virt); + unsigned long min_pages, seeded = 0; int i; selftest_vm.kvm.arch.mmu.vtcr = host_mmu.arch.mmu.vtcr; WARN_ON(kvm_guest_prepare_stage2(&selftest_vm, virt)); + /* + * Mirror pkvm_refill_memcache() for the share/donate pre-checks; + * the selftest invokes those functions directly and would + * otherwise see an empty memcache. + */ + min_pages = kvm_mmu_cache_min_pages(&selftest_vm.kvm.arch.mmu); + for (i = 0; i < pkvm_selftest_pages(); i++) { if (p[i].refcount) continue; p[i].refcount = 1; - hyp_put_page(&selftest_vm.pool, hyp_page_to_virt(&p[i])); + if (seeded < min_pages) { + push_hyp_memcache(&selftest_vcpu.vcpu.arch.pkvm_memcache, + hyp_page_to_virt(&p[i]), hyp_virt_to_phys); + seeded++; + } else { + hyp_put_page(&selftest_vm.pool, hyp_page_to_virt(&p[i])); + } } selftest_vm.kvm.arch.pkvm.handle = __pkvm_reserve_vm(); From 8234409ffb656970e2f5b29e416f041419980bef Mon Sep 17 00:00:00 2001 From: Fuad Tabba Date: Fri, 1 May 2026 12:21:48 +0100 Subject: [PATCH 176/333] KVM: arm64: Pre-check vcpu memcache for host->guest share __pkvm_host_share_guest() ends with kvm_pgtable_stage2_map() to install the guest stage-2 mapping, after a forward pass that mutates the host vmemmap (sets PKVM_PAGE_SHARED_OWNED and increments host_share_guest_count) for every page in the range. The map's return value is wrapped in WARN_ON() and otherwise discarded, asserting that the call cannot fail. WARN_ON() at nVHE EL2 panics, so this assertion is only correct if the call genuinely cannot fail. kvm_pgtable_stage2_map() can fail with -ENOMEM when the stage-2 walker exhausts the caller's memcache, and the host controls the vcpu memcache via the topup interface, so an under-provisioned share request would otherwise turn a recoverable -ENOMEM into a fatal hyp panic. Bound the worst-case walker allocation in the existing pre-check pass so that kvm_pgtable_stage2_map() cannot fail at the call site, using kvm_mmu_cache_min_pages() -- the same bound host EL1 uses for its own stage-2 maps. If the vcpu memcache holds fewer pages, return -ENOMEM before any state mutation. Fixes: d0bd3e6570ae ("KVM: arm64: Introduce __pkvm_host_share_guest()") Assisted-by: Gemini:gemini-3.1-pro review-prompts Signed-off-by: Fuad Tabba Link: https://patch.msgid.link/20260501112149.2824881-6-tabba@google.com Signed-off-by: Marc Zyngier --- arch/arm64/kvm/hyp/nvhe/mem_protect.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/arch/arm64/kvm/hyp/nvhe/mem_protect.c b/arch/arm64/kvm/hyp/nvhe/mem_protect.c index a3050e2b65b1..ffb123fc1145 100644 --- a/arch/arm64/kvm/hyp/nvhe/mem_protect.c +++ b/arch/arm64/kvm/hyp/nvhe/mem_protect.c @@ -1390,6 +1390,22 @@ int __pkvm_host_reclaim_page_guest(u64 gfn, struct pkvm_hyp_vm *vm) return ret && ret != -EHWPOISON ? ret : 0; } +/* + * share/donate install at most one stage-2 leaf (PAGE_SIZE, or one + * KVM_PGTABLE_LAST_LEVEL - 1 block for share). kvm_mmu_cache_min_pages() + * bounds the worst-case allocation: exact for the PAGE_SIZE leaf, + * conservative by one for the block. + */ +static int __guest_check_pgtable_memcache(struct pkvm_hyp_vcpu *vcpu) +{ + struct pkvm_hyp_vm *vm = pkvm_hyp_vcpu_to_hyp_vm(vcpu); + + if (vcpu->vcpu.arch.pkvm_memcache.nr_pages < kvm_mmu_cache_min_pages(vm->pgt.mmu)) + return -ENOMEM; + + return 0; +} + int __pkvm_host_donate_guest(u64 pfn, u64 gfn, struct pkvm_hyp_vcpu *vcpu) { struct pkvm_hyp_vm *vm = pkvm_hyp_vcpu_to_hyp_vm(vcpu); @@ -1474,6 +1490,10 @@ int __pkvm_host_share_guest(u64 pfn, u64 gfn, u64 nr_pages, struct pkvm_hyp_vcpu } } + ret = __guest_check_pgtable_memcache(vcpu); + if (ret) + goto unlock; + for_each_hyp_page(page, phys, size) { set_host_state(page, PKVM_PAGE_SHARED_OWNED); page->host_share_guest_count++; From effc0a39b8e0f30670fe24f51e44329d4324e566 Mon Sep 17 00:00:00 2001 From: Fuad Tabba Date: Fri, 1 May 2026 12:21:49 +0100 Subject: [PATCH 177/333] KVM: arm64: Pre-check vcpu memcache for host->guest donate __pkvm_host_donate_guest() flips the host stage-2 PTE for the donated page to a non-valid annotation via host_stage2_set_owner_metadata_locked() and then calls kvm_pgtable_stage2_map() to install the matching guest stage-2 mapping. The map's return value is wrapped in WARN_ON() and otherwise discarded, asserting that the call cannot fail. WARN_ON() at nVHE EL2 panics, so this assertion is only correct if the call genuinely cannot fail. kvm_pgtable_stage2_map() can fail with -ENOMEM even at PAGE_SIZE granularity: the donate path verifies PKVM_NOPAGE for the guest IPA before the map, so the walker must allocate fresh page-table pages from the vcpu memcache, and the host controls the vcpu memcache via the topup interface. An under-provisioned donation request would otherwise turn a recoverable -ENOMEM into a fatal hyp panic. Bound the worst-case walker allocation alongside the existing __host_check_page_state_range() / __guest_check_page_state_range() pre-checks, using the helper introduced for host->guest share. If the vcpu memcache holds fewer pages than kvm_mmu_cache_min_pages(), return -ENOMEM before any state mutation. Fixes: 1e579adca177 ("KVM: arm64: Introduce __pkvm_host_donate_guest()") Assisted-by: Gemini:gemini-3.1-pro review-prompts Signed-off-by: Fuad Tabba Link: https://patch.msgid.link/20260501112149.2824881-7-tabba@google.com Signed-off-by: Marc Zyngier --- arch/arm64/kvm/hyp/nvhe/mem_protect.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/arch/arm64/kvm/hyp/nvhe/mem_protect.c b/arch/arm64/kvm/hyp/nvhe/mem_protect.c index ffb123fc1145..25f04629014e 100644 --- a/arch/arm64/kvm/hyp/nvhe/mem_protect.c +++ b/arch/arm64/kvm/hyp/nvhe/mem_protect.c @@ -1425,6 +1425,10 @@ int __pkvm_host_donate_guest(u64 pfn, u64 gfn, struct pkvm_hyp_vcpu *vcpu) if (ret) goto unlock; + ret = __guest_check_pgtable_memcache(vcpu); + if (ret) + goto unlock; + meta = host_stage2_encode_gfn_meta(vm, gfn); WARN_ON(host_stage2_set_owner_metadata_locked(phys, PAGE_SIZE, PKVM_ID_GUEST, meta)); From 5be7a0cef3229fb3b63a07c0d289daf752545424 Mon Sep 17 00:00:00 2001 From: Piyush Sachdeva Date: Thu, 7 May 2026 22:22:13 +0530 Subject: [PATCH 178/333] smb: client: Use FullSessionKey for AES-256 encryption key derivation When Kerberos authentication is used with AES-256 encryption (AES-256-CCM or AES-256-GCM), the SMB3 encryption and decryption keys must be derived using the full session key (Session.FullSessionKey) rather than just the first 16 bytes (Session.SessionKey). Per MS-SMB2 section 3.2.5.3.1, when Connection.Dialect is "3.1.1" and Connection.CipherId is AES-256-CCM or AES-256-GCM, Session.FullSessionKey must be set to the full cryptographic key from the GSS authentication context. The encryption and decryption key derivation (SMBC2SCipherKey, SMBS2CCipherKey) must use this FullSessionKey as the KDF input. The signing key derivation continues to use Session.SessionKey (first 16 bytes) in all cases. Previously, generate_key() hardcoded SMB2_NTLMV2_SESSKEY_SIZE (16) as the HMAC-SHA256 key input length for all derivations. When Kerberos with AES-256 provides a 32-byte session key, the KDF for encryption/decryption was using only the first 16 bytes, producing keys that did not match the server's, causing mount failures with sec=krb5 and require_gcm_256=1. Add a full_key_size parameter to generate_key() and pass the appropriate size from generate_smb3signingkey(): - Signing: always SMB2_NTLMV2_SESSKEY_SIZE (16 bytes) - Encryption/Decryption: ses->auth_key.len when AES-256, otherwise 16 Also fix cifs_dump_full_key() to report the actual session key length for AES-256 instead of hardcoded CIFS_SESS_KEY_SIZE, so that userspace tools like Wireshark receive the correct key for decryption. Cc: Reviewed-by: Bharath SM Signed-off-by: Piyush Sachdeva Signed-off-by: Piyush Sachdeva Signed-off-by: Steve French --- fs/smb/client/ioctl.c | 2 +- fs/smb/client/smb2transport.c | 35 ++++++++++++++++++++++++++--------- 2 files changed, 27 insertions(+), 10 deletions(-) diff --git a/fs/smb/client/ioctl.c b/fs/smb/client/ioctl.c index 9afab3237e54..17408bb8ab65 100644 --- a/fs/smb/client/ioctl.c +++ b/fs/smb/client/ioctl.c @@ -296,7 +296,7 @@ static int cifs_dump_full_key(struct cifs_tcon *tcon, struct smb3_full_key_debug break; case SMB2_ENCRYPTION_AES256_CCM: case SMB2_ENCRYPTION_AES256_GCM: - out.session_key_length = CIFS_SESS_KEY_SIZE; + out.session_key_length = ses->auth_key.len; out.server_in_key_length = out.server_out_key_length = SMB3_GCM256_CRYPTKEY_SIZE; break; default: diff --git a/fs/smb/client/smb2transport.c b/fs/smb/client/smb2transport.c index 41009039b4cb..e8eeff9e50d6 100644 --- a/fs/smb/client/smb2transport.c +++ b/fs/smb/client/smb2transport.c @@ -251,7 +251,8 @@ smb2_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server) } static void generate_key(struct cifs_ses *ses, struct kvec label, - struct kvec context, __u8 *key, unsigned int key_size) + struct kvec context, __u8 *key, unsigned int key_size, + unsigned int full_key_size) { unsigned char zero = 0x0; __u8 i[4] = {0, 0, 0, 1}; @@ -265,7 +266,7 @@ static void generate_key(struct cifs_ses *ses, struct kvec label, memset(key, 0x0, key_size); hmac_sha256_init_usingrawkey(&hmac_ctx, ses->auth_key.response, - SMB2_NTLMV2_SESSKEY_SIZE); + full_key_size); hmac_sha256_update(&hmac_ctx, i, 4); hmac_sha256_update(&hmac_ctx, label.iov_base, label.iov_len); hmac_sha256_update(&hmac_ctx, &zero, 1); @@ -298,6 +299,7 @@ generate_smb3signingkey(struct cifs_ses *ses, struct TCP_Server_Info *server, const struct derivation_triplet *ptriplet) { + unsigned int full_key_size = SMB2_NTLMV2_SESSKEY_SIZE; bool is_binding = false; int chan_index = 0; @@ -330,12 +332,24 @@ generate_smb3signingkey(struct cifs_ses *ses, if (is_binding) { generate_key(ses, ptriplet->signing.label, ptriplet->signing.context, - ses->chans[chan_index].signkey, - SMB3_SIGN_KEY_SIZE); + ses->chans[chan_index].signkey, SMB3_SIGN_KEY_SIZE, + SMB2_NTLMV2_SESSKEY_SIZE); } else { generate_key(ses, ptriplet->signing.label, - ptriplet->signing.context, - ses->smb3signingkey, SMB3_SIGN_KEY_SIZE); + ptriplet->signing.context, ses->smb3signingkey, + SMB3_SIGN_KEY_SIZE, SMB2_NTLMV2_SESSKEY_SIZE); + + /* + * Per MS-SMB2 3.2.5.3.1, signing key always uses Session.SessionKey + * (first 16 bytes). Encryption/decryption keys use + * Session.FullSessionKey when dialect is 3.1.1 and cipher is + * AES-256-CCM or AES-256-GCM, otherwise Session.SessionKey. + */ + + if (server->dialect == SMB311_PROT_ID && + (server->cipher_type == SMB2_ENCRYPTION_AES256_CCM || + server->cipher_type == SMB2_ENCRYPTION_AES256_GCM)) + full_key_size = ses->auth_key.len; /* safe to access primary channel, since it will never go away */ spin_lock(&ses->chan_lock); @@ -345,10 +359,13 @@ generate_smb3signingkey(struct cifs_ses *ses, generate_key(ses, ptriplet->encryption.label, ptriplet->encryption.context, - ses->smb3encryptionkey, SMB3_ENC_DEC_KEY_SIZE); + ses->smb3encryptionkey, SMB3_ENC_DEC_KEY_SIZE, + full_key_size); + generate_key(ses, ptriplet->decryption.label, ptriplet->decryption.context, - ses->smb3decryptionkey, SMB3_ENC_DEC_KEY_SIZE); + ses->smb3decryptionkey, SMB3_ENC_DEC_KEY_SIZE, + full_key_size); } #ifdef CONFIG_CIFS_DEBUG_DUMP_KEYS @@ -361,7 +378,7 @@ generate_smb3signingkey(struct cifs_ses *ses, &ses->Suid); cifs_dbg(VFS, "Cipher type %d\n", server->cipher_type); cifs_dbg(VFS, "Session Key %*ph\n", - SMB2_NTLMV2_SESSKEY_SIZE, ses->auth_key.response); + (int)ses->auth_key.len, ses->auth_key.response); cifs_dbg(VFS, "Signing Key %*ph\n", SMB3_SIGN_KEY_SIZE, ses->smb3signingkey); if ((server->cipher_type == SMB2_ENCRYPTION_AES256_CCM) || From 8cb6fc3231500233ddaf63cb7fd5435008d9ed5f Mon Sep 17 00:00:00 2001 From: Piyush Sachdeva Date: Thu, 7 May 2026 22:22:14 +0530 Subject: [PATCH 179/333] smb: client: Zero-pad short GSS session keys per MS-SMB2 Per MS-SMB2 section 3.2.5.3, Session.SessionKey is the first 16 bytes of the GSS cryptographic key, right-padded with zero bytes if the key is shorter than 16 bytes. SMB2_auth_kerberos() copies the GSS session key from the cifs.upcall response using kmemdup(msg->data, msg->sesskey_len, ...) and stores the GSS-reported length verbatim in ses->auth_key.len. generate_key() reads SMB2_NTLMV2_SESSKEY_SIZE bytes from this buffer when feeding the HMAC-SHA256 KDF for signing key derivation. If a GSS mechanism returns a session key shorter than 16 bytes (e.g. a deprecated single-DES Kerberos enctype with an 8-byte session key), the KDF call performs an out-of-bounds slab read and derives keys that do not match the server, which pads per the spec. Modern KDCs disable short-key enctypes by default, so this is latent rather than reachable in production, but it is still a kernel heap over-read. Allocate auth_key.response with kzalloc() at a length of max(msg->sesskey_len, SMB2_NTLMV2_SESSKEY_SIZE), copy the GSS key in, and rely on kzalloc()'s zero initialization for the spec-mandated padding. Set ses->auth_key.len to the padded length. Larger GSS keys (e.g. the 32-byte aes256-cts-hmac-sha1-96 session key) continue to be stored at their natural length, preserving the FullSessionKey path. Emit a cifs_dbg(VFS, ...) message when a short key is encountered to surface deprecated-enctype usage. NTLMv2 and NTLMSSP code paths produce a 16-byte session key by construction and are unaffected. Signed-off-by: Piyush Sachdeva Signed-off-by: Piyush Sachdeva Signed-off-by: Steve French --- fs/smb/client/smb2pdu.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-) diff --git a/fs/smb/client/smb2pdu.c b/fs/smb/client/smb2pdu.c index cb61051f9af3..995fcdd30681 100644 --- a/fs/smb/client/smb2pdu.c +++ b/fs/smb/client/smb2pdu.c @@ -1713,17 +1713,30 @@ SMB2_auth_kerberos(struct SMB2_sess_data *sess_data) is_binding = (ses->ses_status == SES_GOOD); spin_unlock(&ses->ses_lock); + /* + * Per MS-SMB2 3.2.5.3, Session.SessionKey is the first 16 bytes of the + * GSS cryptographic key, right-padded with zero bytes if shorter. + * Allocate at least SMB2_NTLMV2_SESSKEY_SIZE bytes (zeroed) so the KDF + * input buffer is always valid for HMAC-SHA256 even with deprecated + * Kerberos enctypes that return a short session key. + */ + if (unlikely(msg->sesskey_len < SMB2_NTLMV2_SESSKEY_SIZE)) + cifs_dbg(VFS, + "short GSS session key (%u bytes); zero-padding per MS-SMB2 3.2.5.3\n", + msg->sesskey_len); + kfree_sensitive(ses->auth_key.response); - ses->auth_key.response = kmemdup(msg->data, - msg->sesskey_len, - GFP_KERNEL); + ses->auth_key.len = max_t(unsigned int, msg->sesskey_len, + SMB2_NTLMV2_SESSKEY_SIZE); + ses->auth_key.response = kzalloc(ses->auth_key.len, GFP_KERNEL); if (!ses->auth_key.response) { cifs_dbg(VFS, "%s: can't allocate (%u bytes) memory\n", - __func__, msg->sesskey_len); + __func__, ses->auth_key.len); + ses->auth_key.len = 0; rc = -ENOMEM; goto out_put_spnego_key; } - ses->auth_key.len = msg->sesskey_len; + memcpy(ses->auth_key.response, msg->data, msg->sesskey_len); sess_data->iov[1].iov_base = msg->data + msg->sesskey_len; sess_data->iov[1].iov_len = msg->secblob_len; From d62b8d236fab503c6fec1d3e9a38bea71feaca20 Mon Sep 17 00:00:00 2001 From: Zisen Ye Date: Sat, 2 May 2026 18:48:36 +0800 Subject: [PATCH 180/333] smb/client: fix out-of-bounds read in symlink_data() Since smb2_check_message() returns success without length validation for the symlink error response, in symlink_data() it is possible for iov->iov_len to be smaller than sizeof(struct smb2_err_rsp). If the buffer only contains the base SMB2 header (64 bytes), accessing err->ErrorContextCount (at offset 66) or err->ByteCount later in symlink_data() will cause an out-of-bounds read. Link: https://lore.kernel.org/linux-cifs/297d8d9b-adf7-42fd-a1c2-5b1f230032bc@chenxiaosong.com/ Fixes: 76894f3e2f71 ("cifs: improve symlink handling for smb2+") Cc: Stable@vger.kernel.org Signed-off-by: Zisen Ye Reviewed-by: ChenXiaoSong Signed-off-by: Steve French --- fs/smb/client/smb2misc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/smb/client/smb2misc.c b/fs/smb/client/smb2misc.c index 973fce3c959c..2a7355ce1a07 100644 --- a/fs/smb/client/smb2misc.c +++ b/fs/smb/client/smb2misc.c @@ -241,7 +241,8 @@ smb2_check_message(char *buf, unsigned int pdu_len, unsigned int len, if (len != calc_len) { /* create failed on symlink */ if (command == SMB2_CREATE_HE && - shdr->Status == STATUS_STOPPED_ON_SYMLINK) + shdr->Status == STATUS_STOPPED_ON_SYMLINK && + len > calc_len) return 0; /* Windows 7 server returns 24 bytes more */ if (calc_len + 24 == len && command == SMB2_OPLOCK_BREAK_HE) From 8d09328dfda089675e4c049f3f256064a1d1996b Mon Sep 17 00:00:00 2001 From: Zisen Ye Date: Wed, 6 May 2026 11:49:08 +0800 Subject: [PATCH 181/333] smb/client: fix out-of-bounds read in smb2_compound_op() If a server sends a truncated response but a large OutputBufferLength, and terminates the EA list early, check_wsl_eas() returns success without validating that the entire OutputBufferLength fits within iov_len. Then smb2_compound_op() does: memcpy(idata->wsl.eas, data[0], size[0]); Where size[0] is OutputBufferLength. If iov_len is smaller than size[0], memcpy can read beyond the end of the rsp_iov allocation and leak adjacent kernel heap memory. Link: https://lore.kernel.org/linux-cifs/d998240c-aca9-420d-9dbd-f5ba24af19e0@chenxiaosong.com/ Fixes: ea41367b2a60 ("smb: client: introduce SMB2_OP_QUERY_WSL_EA") Cc: stable@vger.kernel.org Signed-off-by: Zisen Ye Reviewed-by: ChenXiaoSong Signed-off-by: Steve French --- fs/smb/client/smb2inode.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/fs/smb/client/smb2inode.c b/fs/smb/client/smb2inode.c index 286912616c73..6c9c229b91f6 100644 --- a/fs/smb/client/smb2inode.c +++ b/fs/smb/client/smb2inode.c @@ -111,7 +111,7 @@ static int check_wsl_eas(struct kvec *rsp_iov) u32 outlen, next; u16 vlen; u8 nlen; - u8 *end; + u8 *ea_end, *iov_end; outlen = le32_to_cpu(rsp->OutputBufferLength); if (outlen < SMB2_WSL_MIN_QUERY_EA_RESP_SIZE || @@ -120,15 +120,19 @@ static int check_wsl_eas(struct kvec *rsp_iov) ea = (void *)((u8 *)rsp_iov->iov_base + le16_to_cpu(rsp->OutputBufferOffset)); - end = (u8 *)rsp_iov->iov_base + rsp_iov->iov_len; + ea_end = (u8 *)ea + outlen; + iov_end = (u8 *)rsp_iov->iov_base + rsp_iov->iov_len; + if (ea_end > iov_end) + return -EINVAL; + for (;;) { - if ((u8 *)ea > end - sizeof(*ea)) + if ((u8 *)ea > ea_end - sizeof(*ea)) return -EINVAL; nlen = ea->ea_name_length; vlen = le16_to_cpu(ea->ea_value_length); if (nlen != SMB2_WSL_XATTR_NAME_LEN || - (u8 *)ea->ea_data + nlen + 1 + vlen > end) + (u8 *)ea->ea_data + nlen + 1 + vlen > ea_end) return -EINVAL; switch (vlen) { From f98b48151cc502ada59d9778f0112d21f2586ca3 Mon Sep 17 00:00:00 2001 From: Michael Bommarito Date: Mon, 20 Apr 2026 10:47:47 -0400 Subject: [PATCH 182/333] smb: client: validate dacloffset before building DACL pointers parse_sec_desc(), build_sec_desc(), and the chown path in id_mode_to_cifs_acl() all add the server-supplied dacloffset to pntsd before proving a DACL header fits inside the returned security descriptor. On 32-bit builds a malicious server can return dacloffset near U32_MAX, wrap the derived DACL pointer below end_of_acl, and then slip past the later pointer-based bounds checks. build_sec_desc() and id_mode_to_cifs_acl() can then dereference DACL fields from the wrapped pointer in the chmod/chown rewrite paths. Validate dacloffset numerically before building any DACL pointer and reuse the same helper at the three DACL entry points. Fixes: bc3e9dd9d104 ("cifs: Change SIDs in ACEs while transferring file ownership.") Cc: stable@vger.kernel.org Assisted-by: Claude:claude-opus-4-6 Signed-off-by: Michael Bommarito Signed-off-by: Steve French --- fs/smb/client/cifsacl.c | 35 ++++++++++++++++++++++++++++++++--- 1 file changed, 32 insertions(+), 3 deletions(-) diff --git a/fs/smb/client/cifsacl.c b/fs/smb/client/cifsacl.c index a2750f1e3d90..786dbbc43c5b 100644 --- a/fs/smb/client/cifsacl.c +++ b/fs/smb/client/cifsacl.c @@ -1264,6 +1264,17 @@ static int parse_sid(struct smb_sid *psid, char *end_of_acl) return 0; } +static bool dacl_offset_valid(unsigned int acl_len, __u32 dacloffset) +{ + if (acl_len < sizeof(struct smb_acl)) + return false; + + if (dacloffset < sizeof(struct smb_ntsd)) + return false; + + return dacloffset <= acl_len - sizeof(struct smb_acl); +} + /* Convert CIFS ACL to POSIX form */ static int parse_sec_desc(struct cifs_sb_info *cifs_sb, @@ -1284,7 +1295,6 @@ static int parse_sec_desc(struct cifs_sb_info *cifs_sb, group_sid_ptr = (struct smb_sid *)((char *)pntsd + le32_to_cpu(pntsd->gsidoffset)); dacloffset = le32_to_cpu(pntsd->dacloffset); - dacl_ptr = (struct smb_acl *)((char *)pntsd + dacloffset); cifs_dbg(NOISY, "revision %d type 0x%x ooffset 0x%x goffset 0x%x sacloffset 0x%x dacloffset 0x%x\n", pntsd->revision, pntsd->type, le32_to_cpu(pntsd->osidoffset), le32_to_cpu(pntsd->gsidoffset), @@ -1315,11 +1325,18 @@ static int parse_sec_desc(struct cifs_sb_info *cifs_sb, return rc; } - if (dacloffset) + if (dacloffset) { + if (!dacl_offset_valid(acl_len, dacloffset)) { + cifs_dbg(VFS, "Server returned illegal DACL offset\n"); + return -EINVAL; + } + + dacl_ptr = (struct smb_acl *)((char *)pntsd + dacloffset); parse_dacl(dacl_ptr, end_of_acl, owner_sid_ptr, group_sid_ptr, fattr, get_mode_from_special_sid); - else + } else { cifs_dbg(FYI, "no ACL\n"); /* BB grant all or default perms? */ + } return rc; } @@ -1342,6 +1359,11 @@ static int build_sec_desc(struct smb_ntsd *pntsd, struct smb_ntsd *pnntsd, dacloffset = le32_to_cpu(pntsd->dacloffset); if (dacloffset) { + if (!dacl_offset_valid(secdesclen, dacloffset)) { + cifs_dbg(VFS, "Server returned illegal DACL offset\n"); + return -EINVAL; + } + dacl_ptr = (struct smb_acl *)((char *)pntsd + dacloffset); rc = validate_dacl(dacl_ptr, end_of_acl); if (rc) @@ -1710,6 +1732,12 @@ id_mode_to_cifs_acl(struct inode *inode, const char *path, __u64 *pnmode, nsecdesclen = sizeof(struct smb_ntsd) + (sizeof(struct smb_sid) * 2); dacloffset = le32_to_cpu(pntsd->dacloffset); if (dacloffset) { + if (!dacl_offset_valid(secdesclen, dacloffset)) { + cifs_dbg(VFS, "Server returned illegal DACL offset\n"); + rc = -EINVAL; + goto id_mode_to_cifs_acl_exit; + } + dacl_ptr = (struct smb_acl *)((char *)pntsd + dacloffset); rc = validate_dacl(dacl_ptr, (char *)pntsd + secdesclen); if (rc) { @@ -1752,6 +1780,7 @@ id_mode_to_cifs_acl(struct inode *inode, const char *path, __u64 *pnmode, rc = ops->set_acl(pnntsd, nsecdesclen, inode, path, aclflag); cifs_dbg(NOISY, "set_cifs_acl rc: %d\n", rc); } +id_mode_to_cifs_acl_exit: cifs_put_tlink(tlink); kfree(pnntsd); From 4616a9c36be7e2e051ef53b0e8fd729da0277abf Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 7 May 2026 11:05:31 -1000 Subject: [PATCH 183/333] sched_ext: Move scx_error() out of scx_link_sched()'s lock region scx_link_sched() holds scx_sched_lock. The scx_error() calls inside take the same lock through scx_claim_exit() and deadlock. Move them out of the guard. Fixes: 6b4576b09714 ("sched_ext: Reject sub-sched attachment to a disabled parent") Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- kernel/sched/ext.c | 22 +++++++++++++++++----- 1 file changed, 17 insertions(+), 5 deletions(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 3f0d8aeaed81..7d367c140a36 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -5585,10 +5585,12 @@ static void refresh_watchdog(void) static s32 scx_link_sched(struct scx_sched *sch) { + const char *err_msg; + s32 ret = 0; + scoped_guard(raw_spinlock_irq, &scx_sched_lock) { #ifdef CONFIG_EXT_SUB_SCHED struct scx_sched *parent = scx_parent(sch); - s32 ret; if (parent) { /* @@ -5598,15 +5600,16 @@ static s32 scx_link_sched(struct scx_sched *sch) * parent can shoot us down. */ if (atomic_read(&parent->exit_kind) != SCX_EXIT_NONE) { - scx_error(sch, "parent disabled"); - return -ENOENT; + err_msg = "parent disabled"; + ret = -ENOENT; + break; } ret = rhashtable_lookup_insert_fast(&scx_sched_hash, &sch->hash_node, scx_sched_hash_params); if (ret) { - scx_error(sch, "failed to insert into scx_sched_hash (%d)", ret); - return ret; + err_msg = "failed to insert into scx_sched_hash"; + break; } list_add_tail(&sch->sibling, &parent->children); @@ -5616,6 +5619,15 @@ static s32 scx_link_sched(struct scx_sched *sch) list_add_tail_rcu(&sch->all, &scx_sched_all); } + /* + * scx_error() takes scx_sched_lock via scx_claim_exit(), so it must run after + * the guard above is released. + */ + if (ret) { + scx_error(sch, "%s (%d)", err_msg, ret); + return ret; + } + refresh_watchdog(); return 0; } From dde2f938d02f2c740d49bb5113dea941f941026a Mon Sep 17 00:00:00 2001 From: Chen Wandun Date: Thu, 7 May 2026 18:54:34 +0800 Subject: [PATCH 184/333] cgroup/cpuset: move PF_EXITING check before __GFP_HARDWALL in cpuset_current_node_allowed() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since prepare_alloc_pages() unconditionally adds __GFP_HARDWALL for the fast path when cpusets are enabled, the __GFP_HARDWALL check in cpuset_current_node_allowed() causes the PF_EXITING escape path to be skipped on the first allocation attempt. This makes it unreachable in the common case, so dying tasks can get stuck in direct reclaim or even trigger OOM while trying to exit, despite being allowed to allocate from any node. Move the PF_EXITING check before __GFP_HARDWALL so that dying tasks can allocate memory from any node to exit quickly, even when cpusets are enabled. Also update the function comment to reflect the actual behavior of prepare_alloc_pages() and the corrected check ordering. Signed-off-by: Chen Wandun Acked-by: Michal Koutný Acked-by: Waiman Long Signed-off-by: Tejun Heo --- kernel/cgroup/cpuset.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index e3a081a07c6d..a48901a0416a 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -4176,11 +4176,11 @@ static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs) * current's mems_allowed, yes. If it's not a __GFP_HARDWALL request and this * node is set in the nearest hardwalled cpuset ancestor to current's cpuset, * yes. If current has access to memory reserves as an oom victim, yes. - * Otherwise, no. + * If the current task is PF_EXITING, yes. Otherwise, no. * * GFP_USER allocations are marked with the __GFP_HARDWALL bit, * and do not allow allocations outside the current tasks cpuset - * unless the task has been OOM killed. + * unless the task has been OOM killed or is exiting. * GFP_KERNEL allocations are not so marked, so can escape to the * nearest enclosing hardwalled ancestor cpuset. * @@ -4194,7 +4194,9 @@ static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs) * The first call here from mm/page_alloc:get_page_from_freelist() * has __GFP_HARDWALL set in gfp_mask, enforcing hardwall cpusets, * so no allocation on a node outside the cpuset is allowed (unless - * in interrupt, of course). + * in interrupt, of course). The PF_EXITING check must therefore + * come before the __GFP_HARDWALL check, otherwise a dying task + * would be blocked on the fast path. * * The second pass through get_page_from_freelist() doesn't even call * here for GFP_ATOMIC calls. For those calls, the __alloc_pages() @@ -4204,6 +4206,7 @@ static struct cpuset *nearest_hardwall_ancestor(struct cpuset *cs) * in_interrupt - any node ok (current task context irrelevant) * GFP_ATOMIC - any node ok * tsk_is_oom_victim - any node ok + * PF_EXITING - any node ok (let dying task exit quickly) * GFP_KERNEL - any node in enclosing hardwalled cpuset ok * GFP_USER - only nodes in current tasks mems allowed ok. */ @@ -4223,11 +4226,10 @@ bool cpuset_current_node_allowed(int node, gfp_t gfp_mask) */ if (unlikely(tsk_is_oom_victim(current))) return true; - if (gfp_mask & __GFP_HARDWALL) /* If hardwall request, stop here */ - return false; - if (current->flags & PF_EXITING) /* Let dying task have memory */ return true; + if (gfp_mask & __GFP_HARDWALL) /* If hardwall request, stop here */ + return false; /* Not hardwall and node outside mems_allowed: scan up cpusets */ spin_lock_irqsave(&callback_lock, flags); From 363a53749cc483409498e8f6e1525fe081f1d9d2 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 7 May 2026 12:09:21 -1000 Subject: [PATCH 185/333] sched_ext: Drop unused scx_find_sub_sched() stub scx_find_sub_sched()'s only caller, scx_bpf_sub_dispatch(), is gated on CONFIG_EXT_SUB_SCHED. When CONFIG_EXT_SUB_SCHED=n the caller compiles out and the stub becomes dead code, tripping -Wunused-function on randconfigs. Drop the stub. Fixes: 25037af712eb ("sched_ext: Add rhashtable lookup for sub-schedulers") Reported-by: kernel test robot Closes: https://lore.kernel.org/all/202605080556.42PXw8U9-lkp@intel.com/ Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 1 - 1 file changed, 1 deletion(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 7d367c140a36..48b4834c7027 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -297,7 +297,6 @@ static void scx_set_task_sched(struct task_struct *p, struct scx_sched *sch) #else /* CONFIG_EXT_SUB_SCHED */ static struct scx_sched *scx_parent(struct scx_sched *sch) { return NULL; } static struct scx_sched *scx_next_descendant_pre(struct scx_sched *pos, struct scx_sched *root) { return pos ? NULL : root; } -static struct scx_sched *scx_find_sub_sched(u64 cgroup_id) { return NULL; } static void scx_set_task_sched(struct task_struct *p, struct scx_sched *sch) {} #endif /* CONFIG_EXT_SUB_SCHED */ From 3607422cdeebd1f0c259964bf33aaa9792e21930 Mon Sep 17 00:00:00 2001 From: Markus Stockhausen Date: Sat, 2 May 2026 19:32:06 +0200 Subject: [PATCH 186/333] hwmon: (lm75) Fix AS6200 and TMP112 setup and alarm handling The initialization of the AS6200 has two shortcomings - The device-add-commit states "Conversion mode: continuous" but the the lm75_params structure uses set_mask = 0x94c0. This activates single shot mode (bit 15). According to the datasheet "The device features a single shot measurement mode if the device is in sleep mode (SM=1)". This is quite contradictionary. - It is the only device that activates polarity active-high (bit 10) All this is paired with a undefined clear mask bug in function lm75_write_config() that was introduced with a later refactoring commit. [as6200] = { .config_reg_16bits = true, .set_mask = 0x94C0, -> .clr_mask not defined here .default_resolution = 12, ... static inline int lm75_write_config(struct lm75_data *data, u16 set_mask, u16 clr_mask) { return regmap_update_bits(data->regmap, LM75_REG_CONF, clr_mask | LM75_SHUTDOWN, set_mask); } regmap_update_bits() requires clr_mask to be a superset of set_mask. So basically all sensors with "wrong" masks like the AS6200 are not initialized as intended. Fix that by - Change the set_mask to 0xc010 to reflect the current active-low setup properly and to drive the sensor in continous mode. This takes into account that the config register is little endian and the first byte sent to the chip is the LSB. - Adapt the alarm handling so it can report the alarm correctly even if it is high active. This is done by comparing config register bit 5 and 10 (translated to 2 and 13). This commit does not introduce any ABI breakage as the mutliple bugs effectly drive the AS6200 in standard active-low mode. Fixes: 4b6358e1fe46 ("hwmon: (lm75) Add AMS AS6200 temperature sensor") Suggested-by: Guenter Roeck Signed-off-by: Markus Stockhausen Link: https://lore.kernel.org/r/20260502173207.3567876-2-markus.stockhausen@gmx.de [groeck: Update set_mask for as6200 further: As modeled, the upper bits contain the conversion rate, so the config register needs to be set to 0xc010 instead of 0x10c0 to reflect 8 samples/s and 4 consecutive faults. Fix the same problem for TMP112.] Signed-off-by: Guenter Roeck --- drivers/hwmon/lm75.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/hwmon/lm75.c b/drivers/hwmon/lm75.c index f1a1e5b888f6..a1bf4e9813ed 100644 --- a/drivers/hwmon/lm75.c +++ b/drivers/hwmon/lm75.c @@ -137,7 +137,7 @@ static const struct lm75_params device_params[] = { }, [as6200] = { .config_reg_16bits = true, - .set_mask = 0x94C0, /* 8 sample/s, 4 CF, positive polarity */ + .set_mask = 0xC010, /* 8 sample/s, 4 CF */ .default_resolution = 12, .default_sample_time = 125, .num_sample_times = 4, @@ -286,8 +286,8 @@ static const struct lm75_params device_params[] = { }, [tmp112] = { .config_reg_16bits = true, - .set_mask = 0x60C0, /* 12-bit mode, 8 samples / second */ - .clr_mask = 1 << 15, /* no one-shot mode*/ + .set_mask = 0xC060, /* 12-bit mode, 8 samples / second */ + .clr_mask = 1 << 7, /* no one-shot mode*/ .default_resolution = 12, .default_sample_time = 125, .num_sample_times = 4, @@ -416,7 +416,7 @@ static int lm75_read(struct device *dev, enum hwmon_sensor_types type, switch (data->kind) { case as6200: case tmp112: - *val = (regval >> 13) & 0x1; + *val = !!(regval & BIT(13)) == !!(regval & BIT(2)); break; default: return -EINVAL; From 05aaac8746c5786eaa779b163fae4ddcd5172707 Mon Sep 17 00:00:00 2001 From: Markus Stockhausen Date: Sat, 2 May 2026 19:32:07 +0200 Subject: [PATCH 187/333] hwmon: (lm75) Fix configuration register writes. Sensors configurations are defined by set and clear masks. These do not follow a consistent "clear mask is a superset of set mask" rule. This relaxed definition breaks lm75_write_config() static inline int lm75_write_config(struct lm75_data *data, u16 set_mask, u16 clr_mask) { return regmap_update_bits(data->regmap, LM75_REG_CONF, clr_mask | LM75_SHUTDOWN, set_mask); } Basically all bits from set_mask that are not defined in clr_mask are dropped. Fix that by enhancing the helper to always combine clr_mask and set_mask into the mask bits of regmap_update_bits(). Fixes: 6da24a25f766 ("hwmon: (lm75) Hide register size differences in regmap access functions") Suggested-by: Guenter Roeck Signed-off-by: Markus Stockhausen Link: https://lore.kernel.org/r/20260502173207.3567876-3-markus.stockhausen@gmx.de Signed-off-by: Guenter Roeck --- drivers/hwmon/lm75.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/hwmon/lm75.c b/drivers/hwmon/lm75.c index a1bf4e9813ed..c283443e363b 100644 --- a/drivers/hwmon/lm75.c +++ b/drivers/hwmon/lm75.c @@ -353,7 +353,7 @@ static inline int lm75_write_config(struct lm75_data *data, u16 set_mask, u16 clr_mask) { return regmap_update_bits(data->regmap, LM75_REG_CONF, - clr_mask | LM75_SHUTDOWN, set_mask); + clr_mask | set_mask | LM75_SHUTDOWN, set_mask); } static irqreturn_t lm75_alarm_handler(int irq, void *private) From 99076a17a112ac43cbd37f6883898ae649166303 Mon Sep 17 00:00:00 2001 From: Tabrez Ahmed Date: Sat, 2 May 2026 07:38:42 +0530 Subject: [PATCH 188/333] hwmon: (ads7871) Fix endianness bug in 16-bit register reads The ads7871_read_reg16() function relies on spi_w8r16() to read the 16-bit sensor output. The ADS7871 device transmits the Least Significant Byte (LSB) first. On Little-Endian architectures, spi_w8r16() correctly reconstructs the 16-bit value. However, on Big-Endian architectures, the byte swapping causes the first received byte (LSB) to be placed in the most significant byte of the u16, resulting in corrupted voltage readings. To fix this, cast the integer result of spi_w8r16() to a restricted __le16 type and convert it to the host CPU's native byte order using le16_to_cpu(). Negative error codes returned by the SPI core are caught and returned prior to the conversion to avoid mangling the error status. Reported-by: Sashiko Closes: https://sashiko.dev/#/patchset/20260418034601.90226-1-tabreztalks@gmail.com Fixes: e0c70b8078629 ("hwmon: add TI ads7871 a/d converter driver") Suggested-by: David Laight Signed-off-by: Tabrez Ahmed Link: https://lore.kernel.org/r/20260502020844.110038-2-tabreztalks@gmail.com Signed-off-by: Guenter Roeck --- drivers/hwmon/ads7871.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/hwmon/ads7871.c b/drivers/hwmon/ads7871.c index 9bfdf9e6bcd7..9ee3ce01f130 100644 --- a/drivers/hwmon/ads7871.c +++ b/drivers/hwmon/ads7871.c @@ -77,9 +77,13 @@ static int ads7871_read_reg8(struct spi_device *spi, int reg) static int ads7871_read_reg16(struct spi_device *spi, int reg) { int ret; + reg = reg | INST_READ_BM | INST_16BIT_BM; ret = spi_w8r16(spi, reg); - return ret; + if (ret < 0) + return ret; + + return le16_to_cpu((__force __le16)ret); } static int ads7871_write_reg8(struct spi_device *spi, int reg, u8 val) From 8e72510db9fa2d41f2b06d5c01fe9020e076fee4 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 May 2026 12:07:13 +0200 Subject: [PATCH 189/333] netfilter: x_tables: allow initial table replace without emitting audit log message At the moment we emit the audit log a bit too early, which makes it necessary to also emit an unregister log in case we have to unwind errors after possible hook register failure. Followup patch will be slightly simpler if we can delay the register message until after the hooks have been wired up. Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/netfilter/x_tables.c | 29 ++++++++++++++++++++--------- 1 file changed, 20 insertions(+), 9 deletions(-) diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c index 2c67c2e6b132..bb0cb3959551 100644 --- a/net/netfilter/x_tables.c +++ b/net/netfilter/x_tables.c @@ -1472,11 +1472,9 @@ struct xt_counters *xt_counters_alloc(unsigned int counters) } EXPORT_SYMBOL(xt_counters_alloc); -struct xt_table_info * -xt_replace_table(struct xt_table *table, - unsigned int num_counters, - struct xt_table_info *newinfo, - int *error) +static struct xt_table_info * +do_replace_table(struct xt_table *table, unsigned int num_counters, + struct xt_table_info *newinfo, int *error) { struct xt_table_info *private; unsigned int cpu; @@ -1531,10 +1529,23 @@ xt_replace_table(struct xt_table *table, } } - audit_log_nfcfg(table->name, table->af, private->number, - !private->number ? AUDIT_XT_OP_REGISTER : - AUDIT_XT_OP_REPLACE, - GFP_KERNEL); + return private; +} + +struct xt_table_info * +xt_replace_table(struct xt_table *table, unsigned int num_counters, + struct xt_table_info *newinfo, + int *error) +{ + struct xt_table_info *private; + + private = do_replace_table(table, num_counters, newinfo, error); + if (private) + audit_log_nfcfg(table->name, table->af, private->number, + !private->number ? AUDIT_XT_OP_REGISTER : + AUDIT_XT_OP_REPLACE, + GFP_KERNEL); + return private; } EXPORT_SYMBOL_GPL(xt_replace_table); From b62eb8dcf2c47d4d676a434efbd57c4f776f7829 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 May 2026 12:07:14 +0200 Subject: [PATCH 190/333] netfilter: x_tables: allocate hook ops while under mutex arp/ip(6)t_register_table() add the table to the per-netns list via xt_register_table() before allocating the per-netns hook ops copy via kmemdup_array(). This leaves a window where the table is visible in the list with ops=NULL. If the pernet exit happens runs concurrently the pre_exit callback finds the table via xt_find_table() and passes the NULL ops pointer to nf_unregister_net_hooks(), causing a NULL dereference: general protection fault in nf_unregister_net_hooks+0xbc/0x150 RIP: nf_unregister_net_hooks (net/netfilter/core.c:613) Call Trace: ipt_unregister_table_pre_exit iptable_mangle_net_pre_exit ops_pre_exit_list cleanup_net Fix by moving the ops allocation into the xtables core so the table is never in the list without valid ops. Also ensure the table is no longer processing packets before its torn down on error unwind. nf_register_net_hooks might have published at least one hook; call synchronize_rcu() if there was an error. audit log register message gets deferred until all operations have passed, this avoids need to emit another ureg message in case of error unwinding. Based on earlier patch by Tristan Madani. Fixes: f9006acc8dfe5 ("netfilter: arp_tables: pass table pointer via nf_hook_ops") Fixes: ee177a54413a ("netfilter: ip6_tables: pass table pointer via nf_hook_ops") Fixes: ae689334225f ("netfilter: ip_tables: pass table pointer via nf_hook_ops") Link: https://lore.kernel.org/netfilter-devel/20260429175613.1459342-1-tristmd@gmail.com/ Signed-off-by: Tristan Madani Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- include/linux/netfilter/x_tables.h | 1 + net/ipv4/netfilter/arp_tables.c | 35 +++------------------ net/ipv4/netfilter/ip_tables.c | 41 +++--------------------- net/ipv6/netfilter/ip6_tables.c | 38 +++-------------------- net/netfilter/x_tables.c | 50 +++++++++++++++++++++++++----- 5 files changed, 55 insertions(+), 110 deletions(-) diff --git a/include/linux/netfilter/x_tables.h b/include/linux/netfilter/x_tables.h index a81b46af5118..cb4b694dd9e4 100644 --- a/include/linux/netfilter/x_tables.h +++ b/include/linux/netfilter/x_tables.h @@ -305,6 +305,7 @@ struct xt_counters *xt_counters_alloc(unsigned int counters); struct xt_table *xt_register_table(struct net *net, const struct xt_table *table, + const struct nf_hook_ops *template_ops, struct xt_table_info *bootstrap, struct xt_table_info *newinfo); void *xt_unregister_table(struct xt_table *table); diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c index 97ead883e4a1..c02e46a0271a 100644 --- a/net/ipv4/netfilter/arp_tables.c +++ b/net/ipv4/netfilter/arp_tables.c @@ -1522,13 +1522,11 @@ int arpt_register_table(struct net *net, const struct arpt_replace *repl, const struct nf_hook_ops *template_ops) { - struct nf_hook_ops *ops; - unsigned int num_ops; - int ret, i; - struct xt_table_info *newinfo; struct xt_table_info bootstrap = {0}; - void *loc_cpu_entry; + struct xt_table_info *newinfo; struct xt_table *new_table; + void *loc_cpu_entry; + int ret; newinfo = xt_alloc_table_info(repl->size); if (!newinfo) @@ -1543,7 +1541,7 @@ int arpt_register_table(struct net *net, return ret; } - new_table = xt_register_table(net, table, &bootstrap, newinfo); + new_table = xt_register_table(net, table, template_ops, &bootstrap, newinfo); if (IS_ERR(new_table)) { struct arpt_entry *iter; @@ -1553,31 +1551,6 @@ int arpt_register_table(struct net *net, return PTR_ERR(new_table); } - num_ops = hweight32(table->valid_hooks); - if (num_ops == 0) { - ret = -EINVAL; - goto out_free; - } - - ops = kmemdup_array(template_ops, num_ops, sizeof(*ops), GFP_KERNEL); - if (!ops) { - ret = -ENOMEM; - goto out_free; - } - - for (i = 0; i < num_ops; i++) - ops[i].priv = new_table; - - new_table->ops = ops; - - ret = nf_register_net_hooks(net, ops, num_ops); - if (ret != 0) - goto out_free; - - return ret; - -out_free: - __arpt_unregister_table(net, new_table); return ret; } diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c index 23c8deff8095..488c5945ebb2 100644 --- a/net/ipv4/netfilter/ip_tables.c +++ b/net/ipv4/netfilter/ip_tables.c @@ -1724,13 +1724,11 @@ int ipt_register_table(struct net *net, const struct xt_table *table, const struct ipt_replace *repl, const struct nf_hook_ops *template_ops) { - struct nf_hook_ops *ops; - unsigned int num_ops; - int ret, i; - struct xt_table_info *newinfo; struct xt_table_info bootstrap = {0}; - void *loc_cpu_entry; + struct xt_table_info *newinfo; struct xt_table *new_table; + void *loc_cpu_entry; + int ret; newinfo = xt_alloc_table_info(repl->size); if (!newinfo) @@ -1745,7 +1743,7 @@ int ipt_register_table(struct net *net, const struct xt_table *table, return ret; } - new_table = xt_register_table(net, table, &bootstrap, newinfo); + new_table = xt_register_table(net, table, template_ops, &bootstrap, newinfo); if (IS_ERR(new_table)) { struct ipt_entry *iter; @@ -1755,37 +1753,6 @@ int ipt_register_table(struct net *net, const struct xt_table *table, return PTR_ERR(new_table); } - /* No template? No need to do anything. This is used by 'nat' table, it registers - * with the nat core instead of the netfilter core. - */ - if (!template_ops) - return 0; - - num_ops = hweight32(table->valid_hooks); - if (num_ops == 0) { - ret = -EINVAL; - goto out_free; - } - - ops = kmemdup_array(template_ops, num_ops, sizeof(*ops), GFP_KERNEL); - if (!ops) { - ret = -ENOMEM; - goto out_free; - } - - for (i = 0; i < num_ops; i++) - ops[i].priv = new_table; - - new_table->ops = ops; - - ret = nf_register_net_hooks(net, ops, num_ops); - if (ret != 0) - goto out_free; - - return ret; - -out_free: - __ipt_unregister_table(net, new_table); return ret; } diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c index d585ac3c1113..dbe7c7acd702 100644 --- a/net/ipv6/netfilter/ip6_tables.c +++ b/net/ipv6/netfilter/ip6_tables.c @@ -1733,13 +1733,11 @@ int ip6t_register_table(struct net *net, const struct xt_table *table, const struct ip6t_replace *repl, const struct nf_hook_ops *template_ops) { - struct nf_hook_ops *ops; - unsigned int num_ops; - int ret, i; - struct xt_table_info *newinfo; struct xt_table_info bootstrap = {0}; - void *loc_cpu_entry; + struct xt_table_info *newinfo; struct xt_table *new_table; + void *loc_cpu_entry; + int ret; newinfo = xt_alloc_table_info(repl->size); if (!newinfo) @@ -1754,7 +1752,7 @@ int ip6t_register_table(struct net *net, const struct xt_table *table, return ret; } - new_table = xt_register_table(net, table, &bootstrap, newinfo); + new_table = xt_register_table(net, table, template_ops, &bootstrap, newinfo); if (IS_ERR(new_table)) { struct ip6t_entry *iter; @@ -1764,34 +1762,6 @@ int ip6t_register_table(struct net *net, const struct xt_table *table, return PTR_ERR(new_table); } - if (!template_ops) - return 0; - - num_ops = hweight32(table->valid_hooks); - if (num_ops == 0) { - ret = -EINVAL; - goto out_free; - } - - ops = kmemdup_array(template_ops, num_ops, sizeof(*ops), GFP_KERNEL); - if (!ops) { - ret = -ENOMEM; - goto out_free; - } - - for (i = 0; i < num_ops; i++) - ops[i].priv = new_table; - - new_table->ops = ops; - - ret = nf_register_net_hooks(net, ops, num_ops); - if (ret != 0) - goto out_free; - - return ret; - -out_free: - __ip6t_unregister_table(net, new_table); return ret; } diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c index bb0cb3959551..06f27bea9eed 100644 --- a/net/netfilter/x_tables.c +++ b/net/netfilter/x_tables.c @@ -1542,7 +1542,6 @@ xt_replace_table(struct xt_table *table, unsigned int num_counters, private = do_replace_table(table, num_counters, newinfo, error); if (private) audit_log_nfcfg(table->name, table->af, private->number, - !private->number ? AUDIT_XT_OP_REGISTER : AUDIT_XT_OP_REPLACE, GFP_KERNEL); @@ -1552,20 +1551,32 @@ EXPORT_SYMBOL_GPL(xt_replace_table); struct xt_table *xt_register_table(struct net *net, const struct xt_table *input_table, + const struct nf_hook_ops *template_ops, struct xt_table_info *bootstrap, struct xt_table_info *newinfo) { struct xt_pernet *xt_net = net_generic(net, xt_pernet_id); + struct xt_table *t, *table = NULL; + struct nf_hook_ops *ops = NULL; struct xt_table_info *private; - struct xt_table *t, *table; - int ret; + unsigned int num_ops; + int ret = -EINVAL; + + num_ops = hweight32(input_table->valid_hooks); + if (num_ops == 0) + goto out; + + ret = -ENOMEM; + if (template_ops) { + ops = kmemdup_array(template_ops, num_ops, sizeof(*ops), GFP_KERNEL); + if (!ops) + goto out; + } /* Don't add one object to multiple lists. */ table = kmemdup(input_table, sizeof(struct xt_table), GFP_KERNEL); - if (!table) { - ret = -ENOMEM; + if (!table) goto out; - } mutex_lock(&xt[table->af].mutex); /* Don't autoload: we'd eat our tail... */ @@ -1579,7 +1590,7 @@ struct xt_table *xt_register_table(struct net *net, /* Simplifies replace_table code. */ table->private = bootstrap; - if (!xt_replace_table(table, 0, newinfo, &ret)) + if (!do_replace_table(table, 0, newinfo, &ret)) goto unlock; private = table->private; @@ -1588,14 +1599,37 @@ struct xt_table *xt_register_table(struct net *net, /* save number of initial entries */ private->initial_entries = private->number; + if (ops) { + int i; + + for (i = 0; i < num_ops; i++) + ops[i].priv = table; + + ret = nf_register_net_hooks(net, ops, num_ops); + if (ret != 0) { + mutex_unlock(&xt[table->af].mutex); + /* nf_register_net_hooks() might have published a + * base chain before internal error unwind. + */ + synchronize_rcu(); + goto out; + } + + table->ops = ops; + } + + audit_log_nfcfg(table->name, table->af, private->number, + AUDIT_XT_OP_REGISTER, GFP_KERNEL); + list_add(&table->list, &xt_net->tables[table->af]); mutex_unlock(&xt[table->af].mutex); return table; unlock: mutex_unlock(&xt[table->af].mutex); - kfree(table); out: + kfree(table); + kfree(ops); return ERR_PTR(ret); } EXPORT_SYMBOL_GPL(xt_register_table); From 527d6931473b75d90e38942aae6537d1a527f1fd Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 May 2026 12:07:15 +0200 Subject: [PATCH 191/333] netfilter: x_tables: add and use xt_unregister_table_pre_exit Remove the copypasted variants of _pre_exit and add one single function in the xtables core. ebtables is not compatible with x_tables and therefore unchanged. This is a preparation patch to reduce noise in the followup bug fixes. Reviewed-by: Tristan Madani Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- include/linux/netfilter/x_tables.h | 1 + include/linux/netfilter_arp/arp_tables.h | 1 - include/linux/netfilter_ipv4/ip_tables.h | 1 - include/linux/netfilter_ipv6/ip6_tables.h | 1 - net/ipv4/netfilter/arp_tables.c | 9 ------- net/ipv4/netfilter/arptable_filter.c | 2 +- net/ipv4/netfilter/ip_tables.c | 9 ------- net/ipv4/netfilter/iptable_filter.c | 2 +- net/ipv4/netfilter/iptable_mangle.c | 2 +- net/ipv4/netfilter/iptable_nat.c | 1 + net/ipv4/netfilter/iptable_raw.c | 2 +- net/ipv4/netfilter/iptable_security.c | 2 +- net/ipv6/netfilter/ip6_tables.c | 9 ------- net/ipv6/netfilter/ip6table_filter.c | 2 +- net/ipv6/netfilter/ip6table_mangle.c | 2 +- net/ipv6/netfilter/ip6table_nat.c | 1 + net/ipv6/netfilter/ip6table_raw.c | 2 +- net/ipv6/netfilter/ip6table_security.c | 2 +- net/netfilter/x_tables.c | 29 +++++++++++++++++++++++ 19 files changed, 41 insertions(+), 39 deletions(-) diff --git a/include/linux/netfilter/x_tables.h b/include/linux/netfilter/x_tables.h index cb4b694dd9e4..74486714ae20 100644 --- a/include/linux/netfilter/x_tables.h +++ b/include/linux/netfilter/x_tables.h @@ -309,6 +309,7 @@ struct xt_table *xt_register_table(struct net *net, struct xt_table_info *bootstrap, struct xt_table_info *newinfo); void *xt_unregister_table(struct xt_table *table); +void xt_unregister_table_pre_exit(struct net *net, u8 af, const char *name); struct xt_table_info *xt_replace_table(struct xt_table *table, unsigned int num_counters, diff --git a/include/linux/netfilter_arp/arp_tables.h b/include/linux/netfilter_arp/arp_tables.h index a40aaf645fa4..05631a25e622 100644 --- a/include/linux/netfilter_arp/arp_tables.h +++ b/include/linux/netfilter_arp/arp_tables.h @@ -53,7 +53,6 @@ int arpt_register_table(struct net *net, const struct xt_table *table, const struct arpt_replace *repl, const struct nf_hook_ops *ops); void arpt_unregister_table(struct net *net, const char *name); -void arpt_unregister_table_pre_exit(struct net *net, const char *name); extern unsigned int arpt_do_table(void *priv, struct sk_buff *skb, const struct nf_hook_state *state); diff --git a/include/linux/netfilter_ipv4/ip_tables.h b/include/linux/netfilter_ipv4/ip_tables.h index 132b0e4a6d4d..13593391d605 100644 --- a/include/linux/netfilter_ipv4/ip_tables.h +++ b/include/linux/netfilter_ipv4/ip_tables.h @@ -26,7 +26,6 @@ int ipt_register_table(struct net *net, const struct xt_table *table, const struct ipt_replace *repl, const struct nf_hook_ops *ops); -void ipt_unregister_table_pre_exit(struct net *net, const char *name); void ipt_unregister_table_exit(struct net *net, const char *name); /* Standard entry. */ diff --git a/include/linux/netfilter_ipv6/ip6_tables.h b/include/linux/netfilter_ipv6/ip6_tables.h index 8b8885a73c76..c6d5b927830d 100644 --- a/include/linux/netfilter_ipv6/ip6_tables.h +++ b/include/linux/netfilter_ipv6/ip6_tables.h @@ -27,7 +27,6 @@ extern void *ip6t_alloc_initial_table(const struct xt_table *); int ip6t_register_table(struct net *net, const struct xt_table *table, const struct ip6t_replace *repl, const struct nf_hook_ops *ops); -void ip6t_unregister_table_pre_exit(struct net *net, const char *name); void ip6t_unregister_table_exit(struct net *net, const char *name); extern unsigned int ip6t_do_table(void *priv, struct sk_buff *skb, const struct nf_hook_state *state); diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c index c02e46a0271a..bd348b7bad2c 100644 --- a/net/ipv4/netfilter/arp_tables.c +++ b/net/ipv4/netfilter/arp_tables.c @@ -1554,15 +1554,6 @@ int arpt_register_table(struct net *net, return ret; } -void arpt_unregister_table_pre_exit(struct net *net, const char *name) -{ - struct xt_table *table = xt_find_table(net, NFPROTO_ARP, name); - - if (table) - nf_unregister_net_hooks(net, table->ops, hweight32(table->valid_hooks)); -} -EXPORT_SYMBOL(arpt_unregister_table_pre_exit); - void arpt_unregister_table(struct net *net, const char *name) { struct xt_table *table = xt_find_table(net, NFPROTO_ARP, name); diff --git a/net/ipv4/netfilter/arptable_filter.c b/net/ipv4/netfilter/arptable_filter.c index 78cd5ee24448..393d9a8c7739 100644 --- a/net/ipv4/netfilter/arptable_filter.c +++ b/net/ipv4/netfilter/arptable_filter.c @@ -43,7 +43,7 @@ static int arptable_filter_table_init(struct net *net) static void __net_exit arptable_filter_net_pre_exit(struct net *net) { - arpt_unregister_table_pre_exit(net, "filter"); + xt_unregister_table_pre_exit(net, NFPROTO_ARP, "filter"); } static void __net_exit arptable_filter_net_exit(struct net *net) diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c index 488c5945ebb2..864489928fb5 100644 --- a/net/ipv4/netfilter/ip_tables.c +++ b/net/ipv4/netfilter/ip_tables.c @@ -1756,14 +1756,6 @@ int ipt_register_table(struct net *net, const struct xt_table *table, return ret; } -void ipt_unregister_table_pre_exit(struct net *net, const char *name) -{ - struct xt_table *table = xt_find_table(net, NFPROTO_IPV4, name); - - if (table) - nf_unregister_net_hooks(net, table->ops, hweight32(table->valid_hooks)); -} - void ipt_unregister_table_exit(struct net *net, const char *name) { struct xt_table *table = xt_find_table(net, NFPROTO_IPV4, name); @@ -1854,7 +1846,6 @@ static void __exit ip_tables_fini(void) } EXPORT_SYMBOL(ipt_register_table); -EXPORT_SYMBOL(ipt_unregister_table_pre_exit); EXPORT_SYMBOL(ipt_unregister_table_exit); EXPORT_SYMBOL(ipt_do_table); module_init(ip_tables_init); diff --git a/net/ipv4/netfilter/iptable_filter.c b/net/ipv4/netfilter/iptable_filter.c index 3ab908b74795..b2fbd9651d61 100644 --- a/net/ipv4/netfilter/iptable_filter.c +++ b/net/ipv4/netfilter/iptable_filter.c @@ -61,7 +61,7 @@ static int __net_init iptable_filter_net_init(struct net *net) static void __net_exit iptable_filter_net_pre_exit(struct net *net) { - ipt_unregister_table_pre_exit(net, "filter"); + xt_unregister_table_pre_exit(net, NFPROTO_IPV4, "filter"); } static void __net_exit iptable_filter_net_exit(struct net *net) diff --git a/net/ipv4/netfilter/iptable_mangle.c b/net/ipv4/netfilter/iptable_mangle.c index 385d945d8ebe..a99e61996197 100644 --- a/net/ipv4/netfilter/iptable_mangle.c +++ b/net/ipv4/netfilter/iptable_mangle.c @@ -96,7 +96,7 @@ static int iptable_mangle_table_init(struct net *net) static void __net_exit iptable_mangle_net_pre_exit(struct net *net) { - ipt_unregister_table_pre_exit(net, "mangle"); + xt_unregister_table_pre_exit(net, NFPROTO_IPV4, "mangle"); } static void __net_exit iptable_mangle_net_exit(struct net *net) diff --git a/net/ipv4/netfilter/iptable_nat.c b/net/ipv4/netfilter/iptable_nat.c index 625a1ca13b1b..8fc4912e790d 100644 --- a/net/ipv4/netfilter/iptable_nat.c +++ b/net/ipv4/netfilter/iptable_nat.c @@ -129,6 +129,7 @@ static int iptable_nat_table_init(struct net *net) static void __net_exit iptable_nat_net_pre_exit(struct net *net) { ipt_nat_unregister_lookups(net); + xt_unregister_table_pre_exit(net, NFPROTO_IPV4, "nat"); } static void __net_exit iptable_nat_net_exit(struct net *net) diff --git a/net/ipv4/netfilter/iptable_raw.c b/net/ipv4/netfilter/iptable_raw.c index 0e7f53964d0a..42511721e538 100644 --- a/net/ipv4/netfilter/iptable_raw.c +++ b/net/ipv4/netfilter/iptable_raw.c @@ -53,7 +53,7 @@ static int iptable_raw_table_init(struct net *net) static void __net_exit iptable_raw_net_pre_exit(struct net *net) { - ipt_unregister_table_pre_exit(net, "raw"); + xt_unregister_table_pre_exit(net, NFPROTO_IPV4, "raw"); } static void __net_exit iptable_raw_net_exit(struct net *net) diff --git a/net/ipv4/netfilter/iptable_security.c b/net/ipv4/netfilter/iptable_security.c index d885443cb267..4646bf6d7d2b 100644 --- a/net/ipv4/netfilter/iptable_security.c +++ b/net/ipv4/netfilter/iptable_security.c @@ -50,7 +50,7 @@ static int iptable_security_table_init(struct net *net) static void __net_exit iptable_security_net_pre_exit(struct net *net) { - ipt_unregister_table_pre_exit(net, "security"); + xt_unregister_table_pre_exit(net, NFPROTO_IPV4, "security"); } static void __net_exit iptable_security_net_exit(struct net *net) diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c index dbe7c7acd702..edf50bc7787e 100644 --- a/net/ipv6/netfilter/ip6_tables.c +++ b/net/ipv6/netfilter/ip6_tables.c @@ -1765,14 +1765,6 @@ int ip6t_register_table(struct net *net, const struct xt_table *table, return ret; } -void ip6t_unregister_table_pre_exit(struct net *net, const char *name) -{ - struct xt_table *table = xt_find_table(net, NFPROTO_IPV6, name); - - if (table) - nf_unregister_net_hooks(net, table->ops, hweight32(table->valid_hooks)); -} - void ip6t_unregister_table_exit(struct net *net, const char *name) { struct xt_table *table = xt_find_table(net, NFPROTO_IPV6, name); @@ -1864,7 +1856,6 @@ static void __exit ip6_tables_fini(void) } EXPORT_SYMBOL(ip6t_register_table); -EXPORT_SYMBOL(ip6t_unregister_table_pre_exit); EXPORT_SYMBOL(ip6t_unregister_table_exit); EXPORT_SYMBOL(ip6t_do_table); diff --git a/net/ipv6/netfilter/ip6table_filter.c b/net/ipv6/netfilter/ip6table_filter.c index e8992693e14a..f05a9e4b2c67 100644 --- a/net/ipv6/netfilter/ip6table_filter.c +++ b/net/ipv6/netfilter/ip6table_filter.c @@ -60,7 +60,7 @@ static int __net_init ip6table_filter_net_init(struct net *net) static void __net_exit ip6table_filter_net_pre_exit(struct net *net) { - ip6t_unregister_table_pre_exit(net, "filter"); + xt_unregister_table_pre_exit(net, NFPROTO_IPV6, "filter"); } static void __net_exit ip6table_filter_net_exit(struct net *net) diff --git a/net/ipv6/netfilter/ip6table_mangle.c b/net/ipv6/netfilter/ip6table_mangle.c index 8dd4cd0c47bd..afa4a5703e43 100644 --- a/net/ipv6/netfilter/ip6table_mangle.c +++ b/net/ipv6/netfilter/ip6table_mangle.c @@ -89,7 +89,7 @@ static int ip6table_mangle_table_init(struct net *net) static void __net_exit ip6table_mangle_net_pre_exit(struct net *net) { - ip6t_unregister_table_pre_exit(net, "mangle"); + xt_unregister_table_pre_exit(net, NFPROTO_IPV6, "mangle"); } static void __net_exit ip6table_mangle_net_exit(struct net *net) diff --git a/net/ipv6/netfilter/ip6table_nat.c b/net/ipv6/netfilter/ip6table_nat.c index 5be723232df8..bb8aa3fc42b4 100644 --- a/net/ipv6/netfilter/ip6table_nat.c +++ b/net/ipv6/netfilter/ip6table_nat.c @@ -131,6 +131,7 @@ static int ip6table_nat_table_init(struct net *net) static void __net_exit ip6table_nat_net_pre_exit(struct net *net) { ip6t_nat_unregister_lookups(net); + xt_unregister_table_pre_exit(net, NFPROTO_IPV6, "nat"); } static void __net_exit ip6table_nat_net_exit(struct net *net) diff --git a/net/ipv6/netfilter/ip6table_raw.c b/net/ipv6/netfilter/ip6table_raw.c index fc9f6754028f..32d2da81c52a 100644 --- a/net/ipv6/netfilter/ip6table_raw.c +++ b/net/ipv6/netfilter/ip6table_raw.c @@ -52,7 +52,7 @@ static int ip6table_raw_table_init(struct net *net) static void __net_exit ip6table_raw_net_pre_exit(struct net *net) { - ip6t_unregister_table_pre_exit(net, "raw"); + xt_unregister_table_pre_exit(net, NFPROTO_IPV6, "raw"); } static void __net_exit ip6table_raw_net_exit(struct net *net) diff --git a/net/ipv6/netfilter/ip6table_security.c b/net/ipv6/netfilter/ip6table_security.c index 4df14a9bae78..3dfd8d6ea4b9 100644 --- a/net/ipv6/netfilter/ip6table_security.c +++ b/net/ipv6/netfilter/ip6table_security.c @@ -49,7 +49,7 @@ static int ip6table_security_table_init(struct net *net) static void __net_exit ip6table_security_net_pre_exit(struct net *net) { - ip6t_unregister_table_pre_exit(net, "security"); + xt_unregister_table_pre_exit(net, NFPROTO_IPV6, "security"); } static void __net_exit ip6table_security_net_exit(struct net *net) diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c index 06f27bea9eed..9c1e896c7b03 100644 --- a/net/netfilter/x_tables.c +++ b/net/netfilter/x_tables.c @@ -1650,6 +1650,35 @@ void *xt_unregister_table(struct xt_table *table) return private; } EXPORT_SYMBOL_GPL(xt_unregister_table); + +/** + * xt_unregister_table_pre_exit - pre-shutdown unregister of a table + * @net: network namespace + * @af: address family (e.g., NFPROTO_IPV4, NFPROTO_IPV6) + * @name: name of the table to unregister + * + * Unregisters the specified netfilter table from the given network namespace + * and also unregisters the hooks from netfilter core: no new packets will be + * processed. + */ +void xt_unregister_table_pre_exit(struct net *net, u8 af, const char *name) +{ + struct xt_pernet *xt_net = net_generic(net, xt_pernet_id); + struct xt_table *t; + + mutex_lock(&xt[af].mutex); + list_for_each_entry(t, &xt_net->tables[af], list) { + if (strcmp(t->name, name) == 0) { + mutex_unlock(&xt[af].mutex); + + if (t->ops) /* nat table registers with nat core, t->ops is NULL. */ + nf_unregister_net_hooks(net, t->ops, hweight32(t->valid_hooks)); + return; + } + } + mutex_unlock(&xt[af].mutex); +} +EXPORT_SYMBOL(xt_unregister_table_pre_exit); #endif #ifdef CONFIG_PROC_FS From d338693d778579b676a61346849bebd892427158 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 May 2026 12:07:16 +0200 Subject: [PATCH 192/333] netfilter: x_tables: unregister the templates first When the module is going away we need to zap the template first. Else there is a small race window where userspace could instantiate a new table after the pernet exit function has removed the current table. Fixes: fdacd57c79b7 ("netfilter: x_tables: never register tables by default") Reported-by: Tristan Madani Reviewed-by: Tristan Madani Closes: https://lore.kernel.org/netfilter-devel/20260429175613.1459342-1-tristmd@gmail.com/ Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/ipv4/netfilter/arptable_filter.c | 2 +- net/ipv4/netfilter/iptable_filter.c | 2 +- net/ipv4/netfilter/iptable_mangle.c | 2 +- net/ipv4/netfilter/iptable_raw.c | 2 +- net/ipv4/netfilter/iptable_security.c | 2 +- net/ipv6/netfilter/ip6table_filter.c | 2 +- net/ipv6/netfilter/ip6table_mangle.c | 2 +- net/ipv6/netfilter/ip6table_raw.c | 2 +- net/ipv6/netfilter/ip6table_security.c | 2 +- 9 files changed, 9 insertions(+), 9 deletions(-) diff --git a/net/ipv4/netfilter/arptable_filter.c b/net/ipv4/netfilter/arptable_filter.c index 393d9a8c7739..382345567a60 100644 --- a/net/ipv4/netfilter/arptable_filter.c +++ b/net/ipv4/netfilter/arptable_filter.c @@ -82,8 +82,8 @@ static int __init arptable_filter_init(void) static void __exit arptable_filter_fini(void) { - unregister_pernet_subsys(&arptable_filter_net_ops); xt_unregister_template(&packet_filter); + unregister_pernet_subsys(&arptable_filter_net_ops); kfree(arpfilter_ops); } diff --git a/net/ipv4/netfilter/iptable_filter.c b/net/ipv4/netfilter/iptable_filter.c index b2fbd9651d61..0dea754a9120 100644 --- a/net/ipv4/netfilter/iptable_filter.c +++ b/net/ipv4/netfilter/iptable_filter.c @@ -101,8 +101,8 @@ static int __init iptable_filter_init(void) static void __exit iptable_filter_fini(void) { - unregister_pernet_subsys(&iptable_filter_net_ops); xt_unregister_template(&packet_filter); + unregister_pernet_subsys(&iptable_filter_net_ops); kfree(filter_ops); } diff --git a/net/ipv4/netfilter/iptable_mangle.c b/net/ipv4/netfilter/iptable_mangle.c index a99e61996197..4d3b12492308 100644 --- a/net/ipv4/netfilter/iptable_mangle.c +++ b/net/ipv4/netfilter/iptable_mangle.c @@ -135,8 +135,8 @@ static int __init iptable_mangle_init(void) static void __exit iptable_mangle_fini(void) { - unregister_pernet_subsys(&iptable_mangle_net_ops); xt_unregister_template(&packet_mangler); + unregister_pernet_subsys(&iptable_mangle_net_ops); kfree(mangle_ops); } diff --git a/net/ipv4/netfilter/iptable_raw.c b/net/ipv4/netfilter/iptable_raw.c index 42511721e538..6f7afec7954b 100644 --- a/net/ipv4/netfilter/iptable_raw.c +++ b/net/ipv4/netfilter/iptable_raw.c @@ -100,9 +100,9 @@ static int __init iptable_raw_init(void) static void __exit iptable_raw_fini(void) { + xt_unregister_template(&packet_raw); unregister_pernet_subsys(&iptable_raw_net_ops); kfree(rawtable_ops); - xt_unregister_template(&packet_raw); } module_init(iptable_raw_init); diff --git a/net/ipv4/netfilter/iptable_security.c b/net/ipv4/netfilter/iptable_security.c index 4646bf6d7d2b..81175c20ccbe 100644 --- a/net/ipv4/netfilter/iptable_security.c +++ b/net/ipv4/netfilter/iptable_security.c @@ -89,9 +89,9 @@ static int __init iptable_security_init(void) static void __exit iptable_security_fini(void) { + xt_unregister_template(&security_table); unregister_pernet_subsys(&iptable_security_net_ops); kfree(sectbl_ops); - xt_unregister_template(&security_table); } module_init(iptable_security_init); diff --git a/net/ipv6/netfilter/ip6table_filter.c b/net/ipv6/netfilter/ip6table_filter.c index f05a9e4b2c67..cf561919bde8 100644 --- a/net/ipv6/netfilter/ip6table_filter.c +++ b/net/ipv6/netfilter/ip6table_filter.c @@ -100,8 +100,8 @@ static int __init ip6table_filter_init(void) static void __exit ip6table_filter_fini(void) { - unregister_pernet_subsys(&ip6table_filter_net_ops); xt_unregister_template(&packet_filter); + unregister_pernet_subsys(&ip6table_filter_net_ops); kfree(filter_ops); } diff --git a/net/ipv6/netfilter/ip6table_mangle.c b/net/ipv6/netfilter/ip6table_mangle.c index afa4a5703e43..1a758f2bc537 100644 --- a/net/ipv6/netfilter/ip6table_mangle.c +++ b/net/ipv6/netfilter/ip6table_mangle.c @@ -128,8 +128,8 @@ static int __init ip6table_mangle_init(void) static void __exit ip6table_mangle_fini(void) { - unregister_pernet_subsys(&ip6table_mangle_net_ops); xt_unregister_template(&packet_mangler); + unregister_pernet_subsys(&ip6table_mangle_net_ops); kfree(mangle_ops); } diff --git a/net/ipv6/netfilter/ip6table_raw.c b/net/ipv6/netfilter/ip6table_raw.c index 32d2da81c52a..923455921c1d 100644 --- a/net/ipv6/netfilter/ip6table_raw.c +++ b/net/ipv6/netfilter/ip6table_raw.c @@ -98,8 +98,8 @@ static int __init ip6table_raw_init(void) static void __exit ip6table_raw_fini(void) { - unregister_pernet_subsys(&ip6table_raw_net_ops); xt_unregister_template(&packet_raw); + unregister_pernet_subsys(&ip6table_raw_net_ops); kfree(rawtable_ops); } diff --git a/net/ipv6/netfilter/ip6table_security.c b/net/ipv6/netfilter/ip6table_security.c index 3dfd8d6ea4b9..c44834d93fc7 100644 --- a/net/ipv6/netfilter/ip6table_security.c +++ b/net/ipv6/netfilter/ip6table_security.c @@ -88,8 +88,8 @@ static int __init ip6table_security_init(void) static void __exit ip6table_security_fini(void) { - unregister_pernet_subsys(&ip6table_security_net_ops); xt_unregister_template(&security_table); + unregister_pernet_subsys(&ip6table_security_net_ops); kfree(sectbl_ops); } From b4597d5fd7d2f8cebfffd40dffb5e003cc78964c Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 May 2026 12:07:17 +0200 Subject: [PATCH 193/333] netfilter: x_tables: add and use xtables_unregister_table_exit Previous change added xtables_unregister_table_pre_exit to detach the table from the packetpath and to unlink it from the active table list. In case of rmmod, userspace that is doing set/getsockopt for this table will not be able to re-instantiate the table: 1. The larval table has been removed already 2. existing instantiated table is no longer on the xt pernet table list. This adds the second stage helper: unlink the table from the dying list, free the hook ops (if any) and do the audit notification. It replaces xt_unregister_table(). Fixes: fdacd57c79b7 ("netfilter: x_tables: never register tables by default") Reported-by: Tristan Madani Reviewed-by: Tristan Madani Closes: https://lore.kernel.org/netfilter-devel/20260429175613.1459342-1-tristmd@gmail.com/ Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- include/linux/netfilter/x_tables.h | 2 +- net/ipv4/netfilter/arp_tables.c | 9 ++-- net/ipv4/netfilter/ip_tables.c | 9 ++-- net/ipv4/netfilter/iptable_nat.c | 5 +- net/ipv6/netfilter/ip6_tables.c | 9 ++-- net/ipv6/netfilter/ip6table_nat.c | 5 +- net/netfilter/x_tables.c | 81 +++++++++++++++++++++++------- 7 files changed, 83 insertions(+), 37 deletions(-) diff --git a/include/linux/netfilter/x_tables.h b/include/linux/netfilter/x_tables.h index 74486714ae20..5a1c5c336fa4 100644 --- a/include/linux/netfilter/x_tables.h +++ b/include/linux/netfilter/x_tables.h @@ -308,8 +308,8 @@ struct xt_table *xt_register_table(struct net *net, const struct nf_hook_ops *template_ops, struct xt_table_info *bootstrap, struct xt_table_info *newinfo); -void *xt_unregister_table(struct xt_table *table); void xt_unregister_table_pre_exit(struct net *net, u8 af, const char *name); +struct xt_table *xt_unregister_table_exit(struct net *net, u8 af, const char *name); struct xt_table_info *xt_replace_table(struct xt_table *table, unsigned int num_counters, diff --git a/net/ipv4/netfilter/arp_tables.c b/net/ipv4/netfilter/arp_tables.c index bd348b7bad2c..ad2259678c78 100644 --- a/net/ipv4/netfilter/arp_tables.c +++ b/net/ipv4/netfilter/arp_tables.c @@ -1501,13 +1501,11 @@ static int do_arpt_get_ctl(struct sock *sk, int cmd, void __user *user, int *len static void __arpt_unregister_table(struct net *net, struct xt_table *table) { - struct xt_table_info *private; - void *loc_cpu_entry; + struct xt_table_info *private = table->private; struct module *table_owner = table->me; + void *loc_cpu_entry; struct arpt_entry *iter; - private = xt_unregister_table(table); - /* Decrease module usage counts and free resources */ loc_cpu_entry = private->entries; xt_entry_foreach(iter, loc_cpu_entry, private->size) @@ -1515,6 +1513,7 @@ static void __arpt_unregister_table(struct net *net, struct xt_table *table) if (private->number > private->initial_entries) module_put(table_owner); xt_free_table_info(private); + kfree(table); } int arpt_register_table(struct net *net, @@ -1556,7 +1555,7 @@ int arpt_register_table(struct net *net, void arpt_unregister_table(struct net *net, const char *name) { - struct xt_table *table = xt_find_table(net, NFPROTO_ARP, name); + struct xt_table *table = xt_unregister_table_exit(net, NFPROTO_ARP, name); if (table) __arpt_unregister_table(net, table); diff --git a/net/ipv4/netfilter/ip_tables.c b/net/ipv4/netfilter/ip_tables.c index 864489928fb5..5cbdb0815857 100644 --- a/net/ipv4/netfilter/ip_tables.c +++ b/net/ipv4/netfilter/ip_tables.c @@ -1704,12 +1704,10 @@ do_ipt_get_ctl(struct sock *sk, int cmd, void __user *user, int *len) static void __ipt_unregister_table(struct net *net, struct xt_table *table) { - struct xt_table_info *private; - void *loc_cpu_entry; + struct xt_table_info *private = table->private; struct module *table_owner = table->me; struct ipt_entry *iter; - - private = xt_unregister_table(table); + void *loc_cpu_entry; /* Decrease module usage counts and free resources */ loc_cpu_entry = private->entries; @@ -1718,6 +1716,7 @@ static void __ipt_unregister_table(struct net *net, struct xt_table *table) if (private->number > private->initial_entries) module_put(table_owner); xt_free_table_info(private); + kfree(table); } int ipt_register_table(struct net *net, const struct xt_table *table, @@ -1758,7 +1757,7 @@ int ipt_register_table(struct net *net, const struct xt_table *table, void ipt_unregister_table_exit(struct net *net, const char *name) { - struct xt_table *table = xt_find_table(net, NFPROTO_IPV4, name); + struct xt_table *table = xt_unregister_table_exit(net, NFPROTO_IPV4, name); if (table) __ipt_unregister_table(net, table); diff --git a/net/ipv4/netfilter/iptable_nat.c b/net/ipv4/netfilter/iptable_nat.c index 8fc4912e790d..a0df72554025 100644 --- a/net/ipv4/netfilter/iptable_nat.c +++ b/net/ipv4/netfilter/iptable_nat.c @@ -119,8 +119,11 @@ static int iptable_nat_table_init(struct net *net) } ret = ipt_nat_register_lookups(net); - if (ret < 0) + if (ret < 0) { + xt_unregister_table_pre_exit(net, NFPROTO_IPV4, "nat"); + synchronize_rcu(); ipt_unregister_table_exit(net, "nat"); + } kfree(repl); return ret; diff --git a/net/ipv6/netfilter/ip6_tables.c b/net/ipv6/netfilter/ip6_tables.c index edf50bc7787e..9d9c3763f2f5 100644 --- a/net/ipv6/netfilter/ip6_tables.c +++ b/net/ipv6/netfilter/ip6_tables.c @@ -1713,12 +1713,10 @@ do_ip6t_get_ctl(struct sock *sk, int cmd, void __user *user, int *len) static void __ip6t_unregister_table(struct net *net, struct xt_table *table) { - struct xt_table_info *private; - void *loc_cpu_entry; + struct xt_table_info *private = table->private; struct module *table_owner = table->me; struct ip6t_entry *iter; - - private = xt_unregister_table(table); + void *loc_cpu_entry; /* Decrease module usage counts and free resources */ loc_cpu_entry = private->entries; @@ -1727,6 +1725,7 @@ static void __ip6t_unregister_table(struct net *net, struct xt_table *table) if (private->number > private->initial_entries) module_put(table_owner); xt_free_table_info(private); + kfree(table); } int ip6t_register_table(struct net *net, const struct xt_table *table, @@ -1767,7 +1766,7 @@ int ip6t_register_table(struct net *net, const struct xt_table *table, void ip6t_unregister_table_exit(struct net *net, const char *name) { - struct xt_table *table = xt_find_table(net, NFPROTO_IPV6, name); + struct xt_table *table = xt_unregister_table_exit(net, NFPROTO_IPV6, name); if (table) __ip6t_unregister_table(net, table); diff --git a/net/ipv6/netfilter/ip6table_nat.c b/net/ipv6/netfilter/ip6table_nat.c index bb8aa3fc42b4..c2394e2c94b5 100644 --- a/net/ipv6/netfilter/ip6table_nat.c +++ b/net/ipv6/netfilter/ip6table_nat.c @@ -121,8 +121,11 @@ static int ip6table_nat_table_init(struct net *net) } ret = ip6t_nat_register_lookups(net); - if (ret < 0) + if (ret < 0) { + xt_unregister_table_pre_exit(net, NFPROTO_IPV6, "nat"); + synchronize_rcu(); ip6t_unregister_table_exit(net, "nat"); + } kfree(repl); return ret; diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c index 9c1e896c7b03..4e6708c23922 100644 --- a/net/netfilter/x_tables.c +++ b/net/netfilter/x_tables.c @@ -55,6 +55,9 @@ static struct list_head xt_templates[NFPROTO_NUMPROTO]; struct xt_pernet { struct list_head tables[NFPROTO_NUMPROTO]; + + /* stash area used during netns exit */ + struct list_head dead_tables[NFPROTO_NUMPROTO]; }; struct compat_delta { @@ -1634,23 +1637,6 @@ struct xt_table *xt_register_table(struct net *net, } EXPORT_SYMBOL_GPL(xt_register_table); -void *xt_unregister_table(struct xt_table *table) -{ - struct xt_table_info *private; - - mutex_lock(&xt[table->af].mutex); - private = table->private; - list_del(&table->list); - mutex_unlock(&xt[table->af].mutex); - audit_log_nfcfg(table->name, table->af, private->number, - AUDIT_XT_OP_UNREGISTER, GFP_KERNEL); - kfree(table->ops); - kfree(table); - - return private; -} -EXPORT_SYMBOL_GPL(xt_unregister_table); - /** * xt_unregister_table_pre_exit - pre-shutdown unregister of a table * @net: network namespace @@ -1660,6 +1646,14 @@ EXPORT_SYMBOL_GPL(xt_unregister_table); * Unregisters the specified netfilter table from the given network namespace * and also unregisters the hooks from netfilter core: no new packets will be * processed. + * + * This must be called prior to xt_unregister_table_exit() from the pernet + * .pre_exit callback. After this call, the table is no longer visible to + * the get/setsockopt path. In case of rmmod, module exit path must have + * called xt_unregister_template() prior to unregistering pernet ops to + * prevent re-instantiation of the table. + * + * See also: xt_unregister_table_exit() */ void xt_unregister_table_pre_exit(struct net *net, u8 af, const char *name) { @@ -1669,6 +1663,7 @@ void xt_unregister_table_pre_exit(struct net *net, u8 af, const char *name) mutex_lock(&xt[af].mutex); list_for_each_entry(t, &xt_net->tables[af], list) { if (strcmp(t->name, name) == 0) { + list_move(&t->list, &xt_net->dead_tables[af]); mutex_unlock(&xt[af].mutex); if (t->ops) /* nat table registers with nat core, t->ops is NULL. */ @@ -1679,6 +1674,50 @@ void xt_unregister_table_pre_exit(struct net *net, u8 af, const char *name) mutex_unlock(&xt[af].mutex); } EXPORT_SYMBOL(xt_unregister_table_pre_exit); + +/** + * xt_unregister_table_exit - remove a table during namespace teardown + * @net: the network namespace from which to unregister the table + * @af: address family (e.g., NFPROTO_IPV4, NFPROTO_IPV6) + * @name: name of the table to unregister + * + * Completes the unregister process for a table. This must be called from + * the pernet ops .exit callback. This is the second stage after + * xt_unregister_table_pre_exit(). + * + * pair with xt_unregister_table_pre_exit() during namespace shutdown. + * + * Return: the unregistered table or NULL if the table was never + * instantiated. The caller needs to kfree() the table after it + * has removed the family specific matches/targets. + */ +struct xt_table *xt_unregister_table_exit(struct net *net, u8 af, const char *name) +{ + struct xt_pernet *xt_net = net_generic(net, xt_pernet_id); + struct xt_table *table; + + mutex_lock(&xt[af].mutex); + list_for_each_entry(table, &xt_net->dead_tables[af], list) { + struct nf_hook_ops *ops = NULL; + + if (strcmp(table->name, name) != 0) + continue; + + list_del(&table->list); + + audit_log_nfcfg(table->name, table->af, table->private->number, + AUDIT_XT_OP_UNREGISTER, GFP_KERNEL); + swap(table->ops, ops); + mutex_unlock(&xt[af].mutex); + + kfree(ops); + return table; + } + mutex_unlock(&xt[af].mutex); + + return NULL; +} +EXPORT_SYMBOL_GPL(xt_unregister_table_exit); #endif #ifdef CONFIG_PROC_FS @@ -2125,8 +2164,10 @@ static int __net_init xt_net_init(struct net *net) struct xt_pernet *xt_net = net_generic(net, xt_pernet_id); int i; - for (i = 0; i < NFPROTO_NUMPROTO; i++) + for (i = 0; i < NFPROTO_NUMPROTO; i++) { INIT_LIST_HEAD(&xt_net->tables[i]); + INIT_LIST_HEAD(&xt_net->dead_tables[i]); + } return 0; } @@ -2135,8 +2176,10 @@ static void __net_exit xt_net_exit(struct net *net) struct xt_pernet *xt_net = net_generic(net, xt_pernet_id); int i; - for (i = 0; i < NFPROTO_NUMPROTO; i++) + for (i = 0; i < NFPROTO_NUMPROTO; i++) { WARN_ON_ONCE(!list_empty(&xt_net->tables[i])); + WARN_ON_ONCE(!list_empty(&xt_net->dead_tables[i])); + } } static struct pernet_operations xt_net_ops = { From b7f0544d86d439cb946515d2ef6a0a75e8626710 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 May 2026 12:07:18 +0200 Subject: [PATCH 194/333] netfilter: ebtables: move to two-stage removal scheme Like previous patches for x_tables, follow same pattern in ebtables. We can't reuse xt helpers: ebt_table struct layout is incompatible. table->ops assignment is now done while still holding the ebt mutex to make sure we never expose partially-filled table struct. Fixes: 87663c39f898 ("netfilter: ebtables: do not hook tables by default") Reviewed-by: Tristan Madani Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/bridge/netfilter/ebtable_broute.c | 2 +- net/bridge/netfilter/ebtable_filter.c | 2 +- net/bridge/netfilter/ebtable_nat.c | 2 +- net/bridge/netfilter/ebtables.c | 60 +++++++++++++++++---------- 4 files changed, 40 insertions(+), 26 deletions(-) diff --git a/net/bridge/netfilter/ebtable_broute.c b/net/bridge/netfilter/ebtable_broute.c index 741360219552..e6f9e343b41f 100644 --- a/net/bridge/netfilter/ebtable_broute.c +++ b/net/bridge/netfilter/ebtable_broute.c @@ -128,8 +128,8 @@ static int __init ebtable_broute_init(void) static void __exit ebtable_broute_fini(void) { - unregister_pernet_subsys(&broute_net_ops); ebt_unregister_template(&broute_table); + unregister_pernet_subsys(&broute_net_ops); } module_init(ebtable_broute_init); diff --git a/net/bridge/netfilter/ebtable_filter.c b/net/bridge/netfilter/ebtable_filter.c index dacd81b12e62..02b6501c15a5 100644 --- a/net/bridge/netfilter/ebtable_filter.c +++ b/net/bridge/netfilter/ebtable_filter.c @@ -109,8 +109,8 @@ static int __init ebtable_filter_init(void) static void __exit ebtable_filter_fini(void) { - unregister_pernet_subsys(&frame_filter_net_ops); ebt_unregister_template(&frame_filter); + unregister_pernet_subsys(&frame_filter_net_ops); } module_init(ebtable_filter_init); diff --git a/net/bridge/netfilter/ebtable_nat.c b/net/bridge/netfilter/ebtable_nat.c index 0f2a8c6118d4..9985a82555c4 100644 --- a/net/bridge/netfilter/ebtable_nat.c +++ b/net/bridge/netfilter/ebtable_nat.c @@ -109,8 +109,8 @@ static int __init ebtable_nat_init(void) static void __exit ebtable_nat_fini(void) { - unregister_pernet_subsys(&frame_nat_net_ops); ebt_unregister_template(&frame_nat); + unregister_pernet_subsys(&frame_nat_net_ops); } module_init(ebtable_nat_init); diff --git a/net/bridge/netfilter/ebtables.c b/net/bridge/netfilter/ebtables.c index aea3e19875c6..3578ffbc14ae 100644 --- a/net/bridge/netfilter/ebtables.c +++ b/net/bridge/netfilter/ebtables.c @@ -42,6 +42,7 @@ struct ebt_pernet { struct list_head tables; + struct list_head dead_tables; }; struct ebt_template { @@ -1162,11 +1163,6 @@ static int do_replace(struct net *net, sockptr_t arg, unsigned int len) static void __ebt_unregister_table(struct net *net, struct ebt_table *table) { - mutex_lock(&ebt_mutex); - list_del(&table->list); - mutex_unlock(&ebt_mutex); - audit_log_nfcfg(table->name, AF_BRIDGE, table->private->nentries, - AUDIT_XT_OP_UNREGISTER, GFP_KERNEL); EBT_ENTRY_ITERATE(table->private->entries, table->private->entries_size, ebt_cleanup_entry, net, NULL); if (table->private->nentries) @@ -1267,13 +1263,15 @@ int ebt_register_table(struct net *net, const struct ebt_table *input_table, for (i = 0; i < num_ops; i++) ops[i].priv = table; - list_add(&table->list, &ebt_net->tables); - mutex_unlock(&ebt_mutex); - table->ops = ops; ret = nf_register_net_hooks(net, ops, num_ops); - if (ret) + if (ret) { + synchronize_rcu(); __ebt_unregister_table(net, table); + } else { + list_add(&table->list, &ebt_net->tables); + } + mutex_unlock(&ebt_mutex); audit_log_nfcfg(repl->name, AF_BRIDGE, repl->nentries, AUDIT_XT_OP_REGISTER, GFP_KERNEL); @@ -1339,7 +1337,7 @@ void ebt_unregister_template(const struct ebt_table *t) } EXPORT_SYMBOL(ebt_unregister_template); -static struct ebt_table *__ebt_find_table(struct net *net, const char *name) +void ebt_unregister_table_pre_exit(struct net *net, const char *name) { struct ebt_pernet *ebt_net = net_generic(net, ebt_pernet_id); struct ebt_table *t; @@ -1348,30 +1346,36 @@ static struct ebt_table *__ebt_find_table(struct net *net, const char *name) list_for_each_entry(t, &ebt_net->tables, list) { if (strcmp(t->name, name) == 0) { + list_move(&t->list, &ebt_net->dead_tables); mutex_unlock(&ebt_mutex); - return t; + nf_unregister_net_hooks(net, t->ops, hweight32(t->valid_hooks)); + return; } } mutex_unlock(&ebt_mutex); - return NULL; -} - -void ebt_unregister_table_pre_exit(struct net *net, const char *name) -{ - struct ebt_table *table = __ebt_find_table(net, name); - - if (table) - nf_unregister_net_hooks(net, table->ops, hweight32(table->valid_hooks)); } EXPORT_SYMBOL(ebt_unregister_table_pre_exit); void ebt_unregister_table(struct net *net, const char *name) { - struct ebt_table *table = __ebt_find_table(net, name); + struct ebt_pernet *ebt_net = net_generic(net, ebt_pernet_id); + struct ebt_table *t; - if (table) - __ebt_unregister_table(net, table); + mutex_lock(&ebt_mutex); + + list_for_each_entry(t, &ebt_net->dead_tables, list) { + if (strcmp(t->name, name) == 0) { + list_del(&t->list); + audit_log_nfcfg(t->name, AF_BRIDGE, t->private->nentries, + AUDIT_XT_OP_UNREGISTER, GFP_KERNEL); + __ebt_unregister_table(net, t); + mutex_unlock(&ebt_mutex); + return; + } + } + + mutex_unlock(&ebt_mutex); } /* userspace just supplied us with counters */ @@ -2556,11 +2560,21 @@ static int __net_init ebt_pernet_init(struct net *net) struct ebt_pernet *ebt_net = net_generic(net, ebt_pernet_id); INIT_LIST_HEAD(&ebt_net->tables); + INIT_LIST_HEAD(&ebt_net->dead_tables); return 0; } +static void __net_exit ebt_pernet_exit(struct net *net) +{ + struct ebt_pernet *ebt_net = net_generic(net, ebt_pernet_id); + + WARN_ON_ONCE(!list_empty(&ebt_net->tables)); + WARN_ON_ONCE(!list_empty(&ebt_net->dead_tables)); +} + static struct pernet_operations ebt_net_ops = { .init = ebt_pernet_init, + .exit = ebt_pernet_exit, .id = &ebt_pernet_id, .size = sizeof(struct ebt_pernet), }; From 92c603fa07bc0d6a17345de3ad7954730b8de44b Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 May 2026 12:07:19 +0200 Subject: [PATCH 195/333] netfilter: ebtables: close dangling table module init race sashiko reported for a related patch: In modules like iptable_raw.c, [..], if register_pernet_subsys() fails, the rollback might call kfree(rawtable_ops) before [..] During this window, could a concurrent userspace process find the globally visible template, trigger table_init(), [..] The table init functions must always register the template last. Otherwise, set/getsockopt can instantiate a table in a namespace while the required pernet ops (contain the destructor) isn't available. This change is also required in x_tables, handled in followup change. Fixes: 87663c39f898 ("netfilter: ebtables: do not hook tables by default") Reviewed-by: Tristan Madani Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/bridge/netfilter/ebtable_broute.c | 12 +++++------- net/bridge/netfilter/ebtable_filter.c | 12 +++++------- net/bridge/netfilter/ebtable_nat.c | 10 ++++------ 3 files changed, 14 insertions(+), 20 deletions(-) diff --git a/net/bridge/netfilter/ebtable_broute.c b/net/bridge/netfilter/ebtable_broute.c index e6f9e343b41f..f05c79f215ea 100644 --- a/net/bridge/netfilter/ebtable_broute.c +++ b/net/bridge/netfilter/ebtable_broute.c @@ -112,18 +112,16 @@ static struct pernet_operations broute_net_ops = { static int __init ebtable_broute_init(void) { - int ret = ebt_register_template(&broute_table, broute_table_init); + int ret = register_pernet_subsys(&broute_net_ops); if (ret) return ret; - ret = register_pernet_subsys(&broute_net_ops); - if (ret) { - ebt_unregister_template(&broute_table); - return ret; - } + ret = ebt_register_template(&broute_table, broute_table_init); + if (ret) + unregister_pernet_subsys(&broute_net_ops); - return 0; + return ret; } static void __exit ebtable_broute_fini(void) diff --git a/net/bridge/netfilter/ebtable_filter.c b/net/bridge/netfilter/ebtable_filter.c index 02b6501c15a5..0fc03b07e62a 100644 --- a/net/bridge/netfilter/ebtable_filter.c +++ b/net/bridge/netfilter/ebtable_filter.c @@ -93,18 +93,16 @@ static struct pernet_operations frame_filter_net_ops = { static int __init ebtable_filter_init(void) { - int ret = ebt_register_template(&frame_filter, frame_filter_table_init); + int ret = register_pernet_subsys(&frame_filter_net_ops); if (ret) return ret; - ret = register_pernet_subsys(&frame_filter_net_ops); - if (ret) { - ebt_unregister_template(&frame_filter); - return ret; - } + ret = ebt_register_template(&frame_filter, frame_filter_table_init); + if (ret) + unregister_pernet_subsys(&frame_filter_net_ops); - return 0; + return ret; } static void __exit ebtable_filter_fini(void) diff --git a/net/bridge/netfilter/ebtable_nat.c b/net/bridge/netfilter/ebtable_nat.c index 9985a82555c4..8a10375d8909 100644 --- a/net/bridge/netfilter/ebtable_nat.c +++ b/net/bridge/netfilter/ebtable_nat.c @@ -93,16 +93,14 @@ static struct pernet_operations frame_nat_net_ops = { static int __init ebtable_nat_init(void) { - int ret = ebt_register_template(&frame_nat, frame_nat_table_init); + int ret = register_pernet_subsys(&frame_nat_net_ops); if (ret) return ret; - ret = register_pernet_subsys(&frame_nat_net_ops); - if (ret) { - ebt_unregister_template(&frame_nat); - return ret; - } + ret = ebt_register_template(&frame_nat, frame_nat_table_init); + if (ret) + unregister_pernet_subsys(&frame_nat_net_ops); return ret; } From 16bc4b6686b2c112c10e67d6b493adc3607256d3 Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Wed, 6 May 2026 12:07:20 +0200 Subject: [PATCH 196/333] netfilter: x_tables: close dangling table module init race Similar to the previous ebtables patch: template add exposes the table to userspace, we must do this last to rnsure the pernet ops are set up (contain the destructors). Fixes: fdacd57c79b7 ("netfilter: x_tables: never register tables by default") Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/ipv4/netfilter/arptable_filter.c | 23 ++++++++++++----------- net/ipv4/netfilter/iptable_filter.c | 23 ++++++++++++----------- net/ipv4/netfilter/iptable_mangle.c | 25 +++++++++++++------------ net/ipv4/netfilter/iptable_raw.c | 22 +++++++++++----------- net/ipv4/netfilter/iptable_security.c | 23 ++++++++++++----------- net/ipv6/netfilter/ip6table_filter.c | 22 +++++++++++----------- net/ipv6/netfilter/ip6table_mangle.c | 23 ++++++++++++----------- net/ipv6/netfilter/ip6table_raw.c | 20 ++++++++++---------- net/ipv6/netfilter/ip6table_security.c | 23 ++++++++++++----------- 9 files changed, 105 insertions(+), 99 deletions(-) diff --git a/net/ipv4/netfilter/arptable_filter.c b/net/ipv4/netfilter/arptable_filter.c index 382345567a60..370b635e3523 100644 --- a/net/ipv4/netfilter/arptable_filter.c +++ b/net/ipv4/netfilter/arptable_filter.c @@ -58,25 +58,26 @@ static struct pernet_operations arptable_filter_net_ops = { static int __init arptable_filter_init(void) { - int ret = xt_register_template(&packet_filter, - arptable_filter_table_init); - - if (ret < 0) - return ret; + int ret; arpfilter_ops = xt_hook_ops_alloc(&packet_filter, arpt_do_table); - if (IS_ERR(arpfilter_ops)) { - xt_unregister_template(&packet_filter); + if (IS_ERR(arpfilter_ops)) return PTR_ERR(arpfilter_ops); - } ret = register_pernet_subsys(&arptable_filter_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(&packet_filter, + arptable_filter_table_init); if (ret < 0) { - xt_unregister_template(&packet_filter); - kfree(arpfilter_ops); - return ret; + unregister_pernet_subsys(&arptable_filter_net_ops); + goto err_free; } + return 0; +err_free: + kfree(arpfilter_ops); return ret; } diff --git a/net/ipv4/netfilter/iptable_filter.c b/net/ipv4/netfilter/iptable_filter.c index 0dea754a9120..672d7da1071d 100644 --- a/net/ipv4/netfilter/iptable_filter.c +++ b/net/ipv4/netfilter/iptable_filter.c @@ -77,26 +77,27 @@ static struct pernet_operations iptable_filter_net_ops = { static int __init iptable_filter_init(void) { - int ret = xt_register_template(&packet_filter, - iptable_filter_table_init); - - if (ret < 0) - return ret; + int ret; filter_ops = xt_hook_ops_alloc(&packet_filter, ipt_do_table); - if (IS_ERR(filter_ops)) { - xt_unregister_template(&packet_filter); + if (IS_ERR(filter_ops)) return PTR_ERR(filter_ops); - } ret = register_pernet_subsys(&iptable_filter_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(&packet_filter, + iptable_filter_table_init); if (ret < 0) { - xt_unregister_template(&packet_filter); - kfree(filter_ops); - return ret; + unregister_pernet_subsys(&iptable_filter_net_ops); + goto err_free; } return 0; +err_free: + kfree(filter_ops); + return ret; } static void __exit iptable_filter_fini(void) diff --git a/net/ipv4/netfilter/iptable_mangle.c b/net/ipv4/netfilter/iptable_mangle.c index 4d3b12492308..13d25d9a4610 100644 --- a/net/ipv4/netfilter/iptable_mangle.c +++ b/net/ipv4/netfilter/iptable_mangle.c @@ -111,25 +111,26 @@ static struct pernet_operations iptable_mangle_net_ops = { static int __init iptable_mangle_init(void) { - int ret = xt_register_template(&packet_mangler, - iptable_mangle_table_init); - if (ret < 0) - return ret; + int ret; mangle_ops = xt_hook_ops_alloc(&packet_mangler, iptable_mangle_hook); - if (IS_ERR(mangle_ops)) { - xt_unregister_template(&packet_mangler); - ret = PTR_ERR(mangle_ops); - return ret; - } + if (IS_ERR(mangle_ops)) + return PTR_ERR(mangle_ops); ret = register_pernet_subsys(&iptable_mangle_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(&packet_mangler, + iptable_mangle_table_init); if (ret < 0) { - xt_unregister_template(&packet_mangler); - kfree(mangle_ops); - return ret; + unregister_pernet_subsys(&iptable_mangle_net_ops); + goto err_free; } + return 0; +err_free: + kfree(mangle_ops); return ret; } diff --git a/net/ipv4/netfilter/iptable_raw.c b/net/ipv4/netfilter/iptable_raw.c index 6f7afec7954b..2745c22f4034 100644 --- a/net/ipv4/netfilter/iptable_raw.c +++ b/net/ipv4/netfilter/iptable_raw.c @@ -77,24 +77,24 @@ static int __init iptable_raw_init(void) pr_info("Enabling raw table before defrag\n"); } - ret = xt_register_template(table, - iptable_raw_table_init); - if (ret < 0) - return ret; - rawtable_ops = xt_hook_ops_alloc(table, ipt_do_table); - if (IS_ERR(rawtable_ops)) { - xt_unregister_template(table); + if (IS_ERR(rawtable_ops)) return PTR_ERR(rawtable_ops); - } ret = register_pernet_subsys(&iptable_raw_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(table, + iptable_raw_table_init); if (ret < 0) { - xt_unregister_template(table); - kfree(rawtable_ops); - return ret; + unregister_pernet_subsys(&iptable_raw_net_ops); + goto err_free; } + return 0; +err_free: + kfree(rawtable_ops); return ret; } diff --git a/net/ipv4/netfilter/iptable_security.c b/net/ipv4/netfilter/iptable_security.c index 81175c20ccbe..491894511c54 100644 --- a/net/ipv4/netfilter/iptable_security.c +++ b/net/ipv4/netfilter/iptable_security.c @@ -65,25 +65,26 @@ static struct pernet_operations iptable_security_net_ops = { static int __init iptable_security_init(void) { - int ret = xt_register_template(&security_table, - iptable_security_table_init); - - if (ret < 0) - return ret; + int ret; sectbl_ops = xt_hook_ops_alloc(&security_table, ipt_do_table); - if (IS_ERR(sectbl_ops)) { - xt_unregister_template(&security_table); + if (IS_ERR(sectbl_ops)) return PTR_ERR(sectbl_ops); - } ret = register_pernet_subsys(&iptable_security_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(&security_table, + iptable_security_table_init); if (ret < 0) { - xt_unregister_template(&security_table); - kfree(sectbl_ops); - return ret; + unregister_pernet_subsys(&iptable_security_net_ops); + goto err_free; } + return 0; +err_free: + kfree(sectbl_ops); return ret; } diff --git a/net/ipv6/netfilter/ip6table_filter.c b/net/ipv6/netfilter/ip6table_filter.c index cf561919bde8..b074fc477676 100644 --- a/net/ipv6/netfilter/ip6table_filter.c +++ b/net/ipv6/netfilter/ip6table_filter.c @@ -76,25 +76,25 @@ static struct pernet_operations ip6table_filter_net_ops = { static int __init ip6table_filter_init(void) { - int ret = xt_register_template(&packet_filter, - ip6table_filter_table_init); - - if (ret < 0) - return ret; + int ret; filter_ops = xt_hook_ops_alloc(&packet_filter, ip6t_do_table); - if (IS_ERR(filter_ops)) { - xt_unregister_template(&packet_filter); + if (IS_ERR(filter_ops)) return PTR_ERR(filter_ops); - } ret = register_pernet_subsys(&ip6table_filter_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(&packet_filter, ip6table_filter_table_init); if (ret < 0) { - xt_unregister_template(&packet_filter); - kfree(filter_ops); - return ret; + unregister_pernet_subsys(&ip6table_filter_net_ops); + goto err_free; } + return 0; +err_free: + kfree(filter_ops); return ret; } diff --git a/net/ipv6/netfilter/ip6table_mangle.c b/net/ipv6/netfilter/ip6table_mangle.c index 1a758f2bc537..e6ee036a9b2c 100644 --- a/net/ipv6/netfilter/ip6table_mangle.c +++ b/net/ipv6/netfilter/ip6table_mangle.c @@ -104,25 +104,26 @@ static struct pernet_operations ip6table_mangle_net_ops = { static int __init ip6table_mangle_init(void) { - int ret = xt_register_template(&packet_mangler, - ip6table_mangle_table_init); - - if (ret < 0) - return ret; + int ret; mangle_ops = xt_hook_ops_alloc(&packet_mangler, ip6table_mangle_hook); - if (IS_ERR(mangle_ops)) { - xt_unregister_template(&packet_mangler); + if (IS_ERR(mangle_ops)) return PTR_ERR(mangle_ops); - } ret = register_pernet_subsys(&ip6table_mangle_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(&packet_mangler, + ip6table_mangle_table_init); if (ret < 0) { - xt_unregister_template(&packet_mangler); - kfree(mangle_ops); - return ret; + unregister_pernet_subsys(&ip6table_mangle_net_ops); + goto err_free; } + return 0; +err_free: + kfree(mangle_ops); return ret; } diff --git a/net/ipv6/netfilter/ip6table_raw.c b/net/ipv6/netfilter/ip6table_raw.c index 923455921c1d..3b161ee875bc 100644 --- a/net/ipv6/netfilter/ip6table_raw.c +++ b/net/ipv6/netfilter/ip6table_raw.c @@ -75,24 +75,24 @@ static int __init ip6table_raw_init(void) pr_info("Enabling raw table before defrag\n"); } - ret = xt_register_template(table, ip6table_raw_table_init); - if (ret < 0) - return ret; - /* Register hooks */ rawtable_ops = xt_hook_ops_alloc(table, ip6t_do_table); - if (IS_ERR(rawtable_ops)) { - xt_unregister_template(table); + if (IS_ERR(rawtable_ops)) return PTR_ERR(rawtable_ops); - } ret = register_pernet_subsys(&ip6table_raw_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(table, ip6table_raw_table_init); if (ret < 0) { - kfree(rawtable_ops); - xt_unregister_template(table); - return ret; + unregister_pernet_subsys(&ip6table_raw_net_ops); + goto err_free; } + return 0; +err_free: + kfree(rawtable_ops); return ret; } diff --git a/net/ipv6/netfilter/ip6table_security.c b/net/ipv6/netfilter/ip6table_security.c index c44834d93fc7..4bd5d97b8ab6 100644 --- a/net/ipv6/netfilter/ip6table_security.c +++ b/net/ipv6/netfilter/ip6table_security.c @@ -64,25 +64,26 @@ static struct pernet_operations ip6table_security_net_ops = { static int __init ip6table_security_init(void) { - int ret = xt_register_template(&security_table, - ip6table_security_table_init); - - if (ret < 0) - return ret; + int ret; sectbl_ops = xt_hook_ops_alloc(&security_table, ip6t_do_table); - if (IS_ERR(sectbl_ops)) { - xt_unregister_template(&security_table); + if (IS_ERR(sectbl_ops)) return PTR_ERR(sectbl_ops); - } ret = register_pernet_subsys(&ip6table_security_net_ops); + if (ret < 0) + goto err_free; + + ret = xt_register_template(&security_table, + ip6table_security_table_init); if (ret < 0) { - kfree(sectbl_ops); - xt_unregister_template(&security_table); - return ret; + unregister_pernet_subsys(&ip6table_security_net_ops); + goto err_free; } + return 0; +err_free: + kfree(sectbl_ops); return ret; } From 27414ff1b287ea9a2a11675149ec28e05539f3cc Mon Sep 17 00:00:00 2001 From: Florian Westphal Date: Thu, 7 May 2026 11:19:22 +0200 Subject: [PATCH 197/333] netfilter: bridge: eb_tables: close module init race sashiko reports for unrelated patch: Does the core ebtables initialization in ebtables.c suffer from a similar race? Once nf_register_sockopt() completes, the sockopts are exposed globally. sockopt has to be registered last, just like in ip/ip6/arptables. Fixes: 5b53951cfc85 ("netfilter: ebtables: use net_generic infra") Signed-off-by: Florian Westphal Signed-off-by: Pablo Neira Ayuso --- net/bridge/netfilter/ebtables.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/net/bridge/netfilter/ebtables.c b/net/bridge/netfilter/ebtables.c index 3578ffbc14ae..b9f4daac09af 100644 --- a/net/bridge/netfilter/ebtables.c +++ b/net/bridge/netfilter/ebtables.c @@ -2583,19 +2583,20 @@ static int __init ebtables_init(void) { int ret; - ret = xt_register_target(&ebt_standard_target); + ret = register_pernet_subsys(&ebt_net_ops); if (ret < 0) return ret; - ret = nf_register_sockopt(&ebt_sockopts); + + ret = xt_register_target(&ebt_standard_target); if (ret < 0) { - xt_unregister_target(&ebt_standard_target); + unregister_pernet_subsys(&ebt_net_ops); return ret; } - ret = register_pernet_subsys(&ebt_net_ops); + ret = nf_register_sockopt(&ebt_sockopts); if (ret < 0) { - nf_unregister_sockopt(&ebt_sockopts); xt_unregister_target(&ebt_standard_target); + unregister_pernet_subsys(&ebt_net_ops); return ret; } From dcb0f9aefdd604d36710fda53c25bd7cf4a3e37a Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Thu, 7 May 2026 13:00:28 +0200 Subject: [PATCH 198/333] netfilter: nf_conntrack_expect: restore helper propagation via expectation A recent series to fix expectations broke helper propagation via expectation, this mechanism is used by the sip and h323 helper. This also propagates the conntrack helper to expected connections. I changed semantics of exp->helper which now tells us the actual helper that created the expectation. Add an explicit assign_helper field to expectations for this purpose and update helpers to use it. Restore this feature for userspace conntrack helper via ctnetlink nfqueue integration so it is again possible to attach a helper to an expectation, where it makes sense. This is not restored via ctnetlink expectation creation as there is no client for such feature. Use the expectation layer 4 protocol number for the helper lookup for consistency. Make sure the expectation using this helper propagation mechanism also go away when the helper is unregistered. Fixes: 9c42bc9db90a ("netfilter: nf_conntrack_expect: honor expectation helper field") Fixes: 917b61fa2042 ("netfilter: ctnetlink: ignore explicit helper on new expectations") Reported-by: Ilya Maximets Tested-by: Ilya Maximets Signed-off-by: Pablo Neira Ayuso --- include/net/netfilter/nf_conntrack_expect.h | 5 ++++- net/netfilter/nf_conntrack_broadcast.c | 1 + net/netfilter/nf_conntrack_core.c | 7 +++++-- net/netfilter/nf_conntrack_expect.c | 1 + net/netfilter/nf_conntrack_h323_main.c | 12 ++++++------ net/netfilter/nf_conntrack_helper.c | 5 +++++ net/netfilter/nf_conntrack_netlink.c | 18 ++++++++++++++++-- net/netfilter/nf_conntrack_sip.c | 2 +- 8 files changed, 39 insertions(+), 12 deletions(-) diff --git a/include/net/netfilter/nf_conntrack_expect.h b/include/net/netfilter/nf_conntrack_expect.h index e9a8350e7ccf..80f50fd0f7ad 100644 --- a/include/net/netfilter/nf_conntrack_expect.h +++ b/include/net/netfilter/nf_conntrack_expect.h @@ -45,9 +45,12 @@ struct nf_conntrack_expect { void (*expectfn)(struct nf_conn *new, struct nf_conntrack_expect *this); - /* Helper to assign to new connection */ + /* Helper that created this expectation */ struct nf_conntrack_helper __rcu *helper; + /* Helper to assign to new connection */ + struct nf_conntrack_helper __rcu *assign_helper; + /* The conntrack of the master connection */ struct nf_conn *master; diff --git a/net/netfilter/nf_conntrack_broadcast.c b/net/netfilter/nf_conntrack_broadcast.c index 4f39bf7c843f..75e53fde6b29 100644 --- a/net/netfilter/nf_conntrack_broadcast.c +++ b/net/netfilter/nf_conntrack_broadcast.c @@ -72,6 +72,7 @@ int nf_conntrack_broadcast_help(struct sk_buff *skb, exp->flags = NF_CT_EXPECT_PERMANENT; exp->class = NF_CT_EXPECT_CLASS_DEFAULT; rcu_assign_pointer(exp->helper, helper); + rcu_assign_pointer(exp->assign_helper, NULL); write_pnet(&exp->net, net); #ifdef CONFIG_NF_CONNTRACK_ZONES exp->zone = ct->zone; diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c index b08189226320..8ba5b22a1eef 100644 --- a/net/netfilter/nf_conntrack_core.c +++ b/net/netfilter/nf_conntrack_core.c @@ -1811,14 +1811,17 @@ init_conntrack(struct net *net, struct nf_conn *tmpl, spin_lock_bh(&nf_conntrack_expect_lock); exp = nf_ct_find_expectation(net, zone, tuple, !tmpl || nf_ct_is_confirmed(tmpl)); if (exp) { + struct nf_conntrack_helper *assign_helper; + /* Welcome, Mr. Bond. We've been expecting you... */ __set_bit(IPS_EXPECTED_BIT, &ct->status); /* exp->master safe, refcnt bumped in nf_ct_find_expectation */ ct->master = exp->master; - if (exp->helper) { + assign_helper = rcu_dereference(exp->assign_helper); + if (assign_helper) { help = nf_ct_helper_ext_add(ct, GFP_ATOMIC); if (help) - rcu_assign_pointer(help->helper, exp->helper); + rcu_assign_pointer(help->helper, assign_helper); } #ifdef CONFIG_NF_CONNTRACK_MARK diff --git a/net/netfilter/nf_conntrack_expect.c b/net/netfilter/nf_conntrack_expect.c index 24d0576d84b7..8e943efbdf0a 100644 --- a/net/netfilter/nf_conntrack_expect.c +++ b/net/netfilter/nf_conntrack_expect.c @@ -344,6 +344,7 @@ void nf_ct_expect_init(struct nf_conntrack_expect *exp, unsigned int class, helper = rcu_dereference(help->helper); rcu_assign_pointer(exp->helper, helper); + rcu_assign_pointer(exp->assign_helper, NULL); write_pnet(&exp->net, net); #ifdef CONFIG_NF_CONNTRACK_ZONES exp->zone = ct->zone; diff --git a/net/netfilter/nf_conntrack_h323_main.c b/net/netfilter/nf_conntrack_h323_main.c index 3f5c50455b71..b2fe6554b9cf 100644 --- a/net/netfilter/nf_conntrack_h323_main.c +++ b/net/netfilter/nf_conntrack_h323_main.c @@ -643,7 +643,7 @@ static int expect_h245(struct sk_buff *skb, struct nf_conn *ct, &ct->tuplehash[!dir].tuple.src.u3, &ct->tuplehash[!dir].tuple.dst.u3, IPPROTO_TCP, NULL, &port); - rcu_assign_pointer(exp->helper, &nf_conntrack_helper_h245); + rcu_assign_pointer(exp->assign_helper, &nf_conntrack_helper_h245); nathook = rcu_dereference(nfct_h323_nat_hook); if (memcmp(&ct->tuplehash[dir].tuple.src.u3, @@ -767,7 +767,7 @@ static int expect_callforwarding(struct sk_buff *skb, nf_ct_expect_init(exp, NF_CT_EXPECT_CLASS_DEFAULT, nf_ct_l3num(ct), &ct->tuplehash[!dir].tuple.src.u3, &addr, IPPROTO_TCP, NULL, &port); - rcu_assign_pointer(exp->helper, nf_conntrack_helper_q931); + rcu_assign_pointer(exp->assign_helper, nf_conntrack_helper_q931); nathook = rcu_dereference(nfct_h323_nat_hook); if (memcmp(&ct->tuplehash[dir].tuple.src.u3, @@ -1234,7 +1234,7 @@ static int expect_q931(struct sk_buff *skb, struct nf_conn *ct, &ct->tuplehash[!dir].tuple.src.u3 : NULL, &ct->tuplehash[!dir].tuple.dst.u3, IPPROTO_TCP, NULL, &port); - rcu_assign_pointer(exp->helper, nf_conntrack_helper_q931); + rcu_assign_pointer(exp->assign_helper, nf_conntrack_helper_q931); exp->flags = NF_CT_EXPECT_PERMANENT; /* Accept multiple calls */ nathook = rcu_dereference(nfct_h323_nat_hook); @@ -1306,7 +1306,7 @@ static int process_gcf(struct sk_buff *skb, struct nf_conn *ct, nf_ct_expect_init(exp, NF_CT_EXPECT_CLASS_DEFAULT, nf_ct_l3num(ct), &ct->tuplehash[!dir].tuple.src.u3, &addr, IPPROTO_UDP, NULL, &port); - rcu_assign_pointer(exp->helper, nf_conntrack_helper_ras); + rcu_assign_pointer(exp->assign_helper, nf_conntrack_helper_ras); if (nf_ct_expect_related(exp, 0) == 0) { pr_debug("nf_ct_ras: expect RAS "); @@ -1523,7 +1523,7 @@ static int process_acf(struct sk_buff *skb, struct nf_conn *ct, &ct->tuplehash[!dir].tuple.src.u3, &addr, IPPROTO_TCP, NULL, &port); exp->flags = NF_CT_EXPECT_PERMANENT; - rcu_assign_pointer(exp->helper, nf_conntrack_helper_q931); + rcu_assign_pointer(exp->assign_helper, nf_conntrack_helper_q931); if (nf_ct_expect_related(exp, 0) == 0) { pr_debug("nf_ct_ras: expect Q.931 "); @@ -1577,7 +1577,7 @@ static int process_lcf(struct sk_buff *skb, struct nf_conn *ct, &ct->tuplehash[!dir].tuple.src.u3, &addr, IPPROTO_TCP, NULL, &port); exp->flags = NF_CT_EXPECT_PERMANENT; - rcu_assign_pointer(exp->helper, nf_conntrack_helper_q931); + rcu_assign_pointer(exp->assign_helper, nf_conntrack_helper_q931); if (nf_ct_expect_related(exp, 0) == 0) { pr_debug("nf_ct_ras: expect Q.931 "); diff --git a/net/netfilter/nf_conntrack_helper.c b/net/netfilter/nf_conntrack_helper.c index a715304a53d8..b594cd244fe1 100644 --- a/net/netfilter/nf_conntrack_helper.c +++ b/net/netfilter/nf_conntrack_helper.c @@ -400,6 +400,11 @@ static bool expect_iter_me(struct nf_conntrack_expect *exp, void *data) this = rcu_dereference_protected(exp->helper, lockdep_is_held(&nf_conntrack_expect_lock)); + if (this == me) + return true; + + this = rcu_dereference_protected(exp->assign_helper, + lockdep_is_held(&nf_conntrack_expect_lock)); return this == me; } diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index eda5fe4a75c8..d7209d124111 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -2634,6 +2634,7 @@ static const struct nla_policy exp_nla_policy[CTA_EXPECT_MAX+1] = { static struct nf_conntrack_expect * ctnetlink_alloc_expect(const struct nlattr *const cda[], struct nf_conn *ct, + const struct nf_conntrack_helper *assign_helper, struct nf_conntrack_tuple *tuple, struct nf_conntrack_tuple *mask); @@ -2860,6 +2861,7 @@ static int ctnetlink_glue_attach_expect(const struct nlattr *attr, struct nf_conn *ct, u32 portid, u32 report) { + struct nf_conntrack_helper *assign_helper = NULL; struct nlattr *cda[CTA_EXPECT_MAX+1]; struct nf_conntrack_tuple tuple, mask; struct nf_conntrack_expect *exp; @@ -2875,8 +2877,18 @@ ctnetlink_glue_attach_expect(const struct nlattr *attr, struct nf_conn *ct, if (err < 0) return err; + if (cda[CTA_EXPECT_HELP_NAME]) { + const char *helpname = nla_data(cda[CTA_EXPECT_HELP_NAME]); + + assign_helper = __nf_conntrack_helper_find(helpname, + nf_ct_l3num(ct), + tuple.dst.protonum); + if (!assign_helper) + return -EOPNOTSUPP; + } + exp = ctnetlink_alloc_expect((const struct nlattr * const *)cda, ct, - &tuple, &mask); + assign_helper, &tuple, &mask); if (IS_ERR(exp)) return PTR_ERR(exp); @@ -3515,6 +3527,7 @@ ctnetlink_parse_expect_nat(const struct nlattr *attr, static struct nf_conntrack_expect * ctnetlink_alloc_expect(const struct nlattr * const cda[], struct nf_conn *ct, + const struct nf_conntrack_helper *assign_helper, struct nf_conntrack_tuple *tuple, struct nf_conntrack_tuple *mask) { @@ -3568,6 +3581,7 @@ ctnetlink_alloc_expect(const struct nlattr * const cda[], struct nf_conn *ct, exp->zone = ct->zone; #endif rcu_assign_pointer(exp->helper, helper); + rcu_assign_pointer(exp->assign_helper, assign_helper); exp->tuple = *tuple; exp->mask.src.u3 = mask->src.u3; exp->mask.src.u.all = mask->src.u.all; @@ -3623,7 +3637,7 @@ ctnetlink_create_expect(struct net *net, ct = nf_ct_tuplehash_to_ctrack(h); rcu_read_lock(); - exp = ctnetlink_alloc_expect(cda, ct, &tuple, &mask); + exp = ctnetlink_alloc_expect(cda, ct, NULL, &tuple, &mask); if (IS_ERR(exp)) { err = PTR_ERR(exp); goto err_rcu; diff --git a/net/netfilter/nf_conntrack_sip.c b/net/netfilter/nf_conntrack_sip.c index 1eb55907d470..d24bfa9e8234 100644 --- a/net/netfilter/nf_conntrack_sip.c +++ b/net/netfilter/nf_conntrack_sip.c @@ -1383,7 +1383,7 @@ static int process_register_request(struct sk_buff *skb, unsigned int protoff, nf_ct_expect_init(exp, SIP_EXPECT_SIGNALLING, nf_ct_l3num(ct), saddr, &daddr, proto, NULL, &port); exp->timeout.expires = sip_timeout * HZ; - rcu_assign_pointer(exp->helper, helper); + rcu_assign_pointer(exp->assign_helper, helper); exp->flags = NF_CT_EXPECT_PERMANENT | NF_CT_EXPECT_INACTIVE; hooks = rcu_dereference(nf_nat_sip_hooks); From d8ef54c83ad70b81735b506431affadd2f720aa1 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Thu, 7 May 2026 23:57:55 +0200 Subject: [PATCH 199/333] netfilter: ctnetlink: check tuple and mask in expectations created via nfqueue Ensure the expectation tuple and mask attributes are present in netlink message, otherwise null-ptr-deref is possible. Fixes: bd0779370588 ("netfilter: nfnetlink_queue: allow to attach expectations to conntracks") Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_conntrack_netlink.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c index d7209d124111..befa7e83ee49 100644 --- a/net/netfilter/nf_conntrack_netlink.c +++ b/net/netfilter/nf_conntrack_netlink.c @@ -2872,6 +2872,9 @@ ctnetlink_glue_attach_expect(const struct nlattr *attr, struct nf_conn *ct, if (err < 0) return err; + if (!cda[CTA_EXPECT_TUPLE] || !cda[CTA_EXPECT_MASK]) + return -EINVAL; + err = ctnetlink_glue_exp_parse((const struct nlattr * const *)cda, ct, &tuple, &mask); if (err < 0) From eb6317739b1ea3ab28791e1f91b24781905fa815 Mon Sep 17 00:00:00 2001 From: Li Xiasong Date: Thu, 7 May 2026 22:04:22 +0800 Subject: [PATCH 200/333] netfilter: nf_conntrack_sip: get helper before allocating expectation process_register_request() allocates an expectation and then checks whether a conntrack helper is available. If helper lookup fails, the function returns early and the allocated expectation is left behind. Reorder the code to fetch and validate helper before calling nf_ct_expect_alloc(). This keeps the logic simpler and removes the leak path while preserving existing behavior. Fixes: e14575fa7529 ("netfilter: nf_conntrack: use rcu accessors where needed") Cc: stable@vger.kernel.org Signed-off-by: Li Xiasong Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nf_conntrack_sip.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/net/netfilter/nf_conntrack_sip.c b/net/netfilter/nf_conntrack_sip.c index d24bfa9e8234..e69941f1a101 100644 --- a/net/netfilter/nf_conntrack_sip.c +++ b/net/netfilter/nf_conntrack_sip.c @@ -1366,6 +1366,10 @@ static int process_register_request(struct sk_buff *skb, unsigned int protoff, goto store_cseq; } + helper = rcu_dereference(nfct_help(ct)->helper); + if (!helper) + return NF_DROP; + exp = nf_ct_expect_alloc(ct); if (!exp) { nf_ct_helper_log(skb, ct, "cannot alloc expectation"); @@ -1376,10 +1380,6 @@ static int process_register_request(struct sk_buff *skb, unsigned int protoff, if (sip_direct_signalling) saddr = &ct->tuplehash[!dir].tuple.src.u3; - helper = rcu_dereference(nfct_help(ct)->helper); - if (!helper) - return NF_DROP; - nf_ct_expect_init(exp, SIP_EXPECT_SIGNALLING, nf_ct_l3num(ct), saddr, &daddr, proto, NULL, &port); exp->timeout.expires = sip_timeout * HZ; From 19f94b6fee75b3ef7fbc06f3745b9a771a8a19a4 Mon Sep 17 00:00:00 2001 From: Li Xiasong Date: Thu, 7 May 2026 22:04:23 +0800 Subject: [PATCH 201/333] netfilter: nft_ct: fix missing expect put in obj eval nft_ct_expect_obj_eval() allocates an expectation and may call nf_ct_expect_related(), but never drops its local reference. Add nf_ct_expect_put(exp) before return to balance allocation. Fixes: 857b46027d6f ("netfilter: nft_ct: add ct expectations support") Cc: stable@vger.kernel.org Signed-off-by: Li Xiasong Signed-off-by: Pablo Neira Ayuso --- net/netfilter/nft_ct.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/netfilter/nft_ct.c b/net/netfilter/nft_ct.c index 60ee8d932fcb..fa2cc556331c 100644 --- a/net/netfilter/nft_ct.c +++ b/net/netfilter/nft_ct.c @@ -1334,6 +1334,8 @@ static void nft_ct_expect_obj_eval(struct nft_object *obj, if (nf_ct_expect_related(exp, 0) != 0) regs->verdict.code = NF_DROP; + + nf_ct_expect_put(exp); } static const struct nla_policy nft_ct_expect_policy[NFTA_CT_EXPECT_MAX + 1] = { From 1f91d0d5827512816789f74f4d72d16269bde1ec Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 7 May 2026 14:16:59 -1000 Subject: [PATCH 202/333] sched_ext: Fix !CONFIG_EXT_SUB_SCHED build warnings W=1 with CONFIG_EXT_SUB_SCHED=n flags 'err_msg' uninitialized and 'err_free_lb_resched' unused. Initialize err_msg and gate the label. Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 48b4834c7027..f4e2db8e56be 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -5584,7 +5584,7 @@ static void refresh_watchdog(void) static s32 scx_link_sched(struct scx_sched *sch) { - const char *err_msg; + const char *err_msg = ""; s32 ret = 0; scoped_guard(raw_spinlock_irq, &scx_sched_lock) { @@ -6652,8 +6652,10 @@ static struct scx_sched *scx_alloc_and_add_sched(struct sched_ext_ops *ops, #endif /* CONFIG_EXT_SUB_SCHED */ return sch; +#ifdef CONFIG_EXT_SUB_SCHED err_free_lb_resched: free_cpumask_var(sch->bypass_lb_resched_cpumask); +#endif err_free_lb_cpumask: free_cpumask_var(sch->bypass_lb_donee_cpumask); err_stop_helper: From 307abfac04a254c09c5705d816b33354acee97a0 Mon Sep 17 00:00:00 2001 From: Jianpeng Chang Date: Fri, 8 May 2026 09:56:36 +0900 Subject: [PATCH 203/333] kprobes: skip non-symbol addresses in kprobe_add_ksym_blacklist() When kprobe_add_area_blacklist() iterates through a section like .kprobes.text, the start address may not correspond to a named symbol. On ARM64 with CONFIG_DYNAMIC_FTRACE_WITH_CALL_OPS=y (introduced by commit baaf553d3bc3 ("arm64: Implement HAVE_DYNAMIC_FTRACE_WITH_CALL_OPS")), the compiler flag -fpatchable-function-entry=4,2 inserts 2 NOPs before each function entry point for ftrace call_ops. These pre-function NOPs sit at the section base address, before the first named function symbol. The compiler emits a $x mapping symbol at offset 0x00 to mark the start of code, but find_kallsyms_symbol() ignores mapping symbols. Without CONFIG_DYNAMIC_FTRACE_WITH_CALL_OPS (e.g. defconfig), no pre-function NOPs are inserted, the first function starts at offset 0x00, and the bug does not trigger. This only affects modules that have a .kprobes.text section (i.e. those using the __kprobes annotation). Modules using NOKPROBE_SYMBOL() instead (like kretprobe_example.ko) blacklist exact function addresses via the _kprobe_blacklist section and are not affected. For kprobe_example.ko on ARM64 with -fpatchable-function-entry=4,2, the .kprobes.text section layout is: offset 0x00: $x + 2 NOPs (mapping symbol + ftrace preamble) offset 0x08: handler_post (64 bytes) offset 0x50: handler_pre (68 bytes) kprobe_add_area_blacklist() starts iterating from the section base address (offset 0x00), which only has the $x mapping symbol. kprobe_add_ksym_blacklist() then calls kallsyms_lookup_size_offset() for this address, which goes through: kallsyms_lookup_size_offset() -> module_address_lookup() -> find_kallsyms_symbol() find_kallsyms_symbol() scans all module symbols to find the closest preceding symbol. Since no named text symbol exists at offset 0x00, find_kallsyms_symbol() picks __UNIQUE_ID_vermagic (a .modinfo symbol whose address is in the temporary image) as the "best" match. The computed "size" = next_text_symbol - modinfo_symbol spans across these two unrelated memory regions, creating a blacklist entry with a bogus range of tens of terabytes. Whether this causes a visible failure depends on address randomization, here is what happens on Raspberry Pi 4/5: - On RPi5, the bogus size was ~35 TB. start + size stayed within 64-bit range, so the blacklist entry covered the entire kernel text. register_kprobe() in the module's own init function failed with -EINVAL. - On RPi4, the bogus size was ~75 TB. start + size overflowed 64 bits and wrapped to a small address near zero. The range check (addr >= start && addr < end) then failed because end wrapped around, so the bogus entry was accidentally harmless and kprobes worked by luck. The same bug exists on both machines, but randomization determines whether the integer overflow masks it or not. Fix this by adding notrace to the __kprobes macro. Functions in .kprobes.text are kprobe infrastructure handlers that should never be traced by ftrace. With notrace, the compiler stops inserting them and the non-symbol gap at the section start disappears entirely. Link: https://lore.kernel.org/all/20260506012706.2785785-1-jianpeng.chang.cn@windriver.com/ Fixes: baaf553d3bc3 ("arm64: Implement HAVE_DYNAMIC_FTRACE_WITH_CALL_OPS") Signed-off-by: Jianpeng Chang Signed-off-by: Masami Hiramatsu (Google) --- include/asm-generic/kprobes.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/asm-generic/kprobes.h b/include/asm-generic/kprobes.h index 060eab094e5a..5290a2b2e15a 100644 --- a/include/asm-generic/kprobes.h +++ b/include/asm-generic/kprobes.h @@ -14,7 +14,7 @@ static unsigned long __used \ _kbl_addr_##fname = (unsigned long)fname; # define NOKPROBE_SYMBOL(fname) __NOKPROBE_SYMBOL(fname) /* Use this to forbid a kprobes attach on very low level functions */ -# define __kprobes __section(".kprobes.text") +# define __kprobes notrace __section(".kprobes.text") # define nokprobe_inline __always_inline #else # define NOKPROBE_SYMBOL(fname) From ef5581bb30efb939cc2bf093475c6cc85258e5cd Mon Sep 17 00:00:00 2001 From: Martin Kaiser Date: Fri, 8 May 2026 09:56:36 +0900 Subject: [PATCH 204/333] test_kprobes: clear kprobes between test runs Running the kprobes sanity tests twice makes all tests fail and eventually crashes the kernel. [root@martin-riscv-1 ~]# echo 1 > /sys/kernel/debug/kunit/kprobes_test/run ... # Totals: pass:5 fail:0 skip:0 total:5 ok 1 kprobes_test [root@martin-riscv-1 ~]# echo 1 > /sys/kernel/debug/kunit/kprobes_test/run ... # test_kprobe: EXPECTATION FAILED at lib/tests/test_kprobes.c:64 Expected 0 == register_kprobe(&kp), but register_kprobe(&kp) == -22 (0xffffffffffffffea) ... Unable to handle kernel paging request ... The testsuite defines several kprobes and kretprobes as static variables that are preserved across test runs. After register_kprobe and unregister_kprobe, a kprobe contains some leftover data that must be cleared before the kprobe can be registered again. The tests are setting symbol_name to define the probe location. Address and flags must be cleared. The existing code clears some of the probes between subsequent tests, but not between two test runs. The leftover data from a previous test run makes the registrations fail in the next run. Move the cleanups for all kprobes into kprobes_test_init, this function is called before each single test (including the first test of a test run). Link: https://lore.kernel.org/all/20260507134615.1010905-1-martin@kaiser.cx/ Fixes: e44e81c5b90f ("kprobes: convert tests to kunit") Signed-off-by: Martin Kaiser Signed-off-by: Masami Hiramatsu (Google) --- lib/tests/test_kprobes.c | 29 ++++++++++++++++++----------- 1 file changed, 18 insertions(+), 11 deletions(-) diff --git a/lib/tests/test_kprobes.c b/lib/tests/test_kprobes.c index b7582010125c..06e729e4de05 100644 --- a/lib/tests/test_kprobes.c +++ b/lib/tests/test_kprobes.c @@ -12,6 +12,12 @@ #define div_factor 3 +#define KP_CLEAR(_kp) \ +do { \ + (_kp).addr = NULL; \ + (_kp).flags = 0; \ +} while (0) + static u32 rand1, preh_val, posth_val; static u32 (*target)(u32 value); static u32 (*recursed_target)(u32 value); @@ -125,10 +131,6 @@ static void test_kprobes(struct kunit *test) current_test = test; - /* addr and flags should be cleard for reusing kprobe. */ - kp.addr = NULL; - kp.flags = 0; - KUNIT_EXPECT_EQ(test, 0, register_kprobes(kps, 2)); preh_val = 0; posth_val = 0; @@ -226,9 +228,6 @@ static void test_kretprobes(struct kunit *test) struct kretprobe *rps[2] = {&rp, &rp2}; current_test = test; - /* addr and flags should be cleard for reusing kprobe. */ - rp.kp.addr = NULL; - rp.kp.flags = 0; KUNIT_EXPECT_EQ(test, 0, register_kretprobes(rps, 2)); krph_val = 0; @@ -290,8 +289,6 @@ static void test_stacktrace_on_kretprobe(struct kunit *test) unsigned long myretaddr = (unsigned long)__builtin_return_address(0); current_test = test; - rp3.kp.addr = NULL; - rp3.kp.flags = 0; /* * Run the stacktrace_driver() to record correct return address in @@ -352,8 +349,6 @@ static void test_stacktrace_on_nested_kretprobe(struct kunit *test) struct kretprobe *rps[2] = {&rp3, &rp4}; current_test = test; - rp3.kp.addr = NULL; - rp3.kp.flags = 0; //KUNIT_ASSERT_NE(test, myretaddr, stacktrace_driver()); @@ -367,6 +362,18 @@ static void test_stacktrace_on_nested_kretprobe(struct kunit *test) static int kprobes_test_init(struct kunit *test) { + KP_CLEAR(kp); + KP_CLEAR(kp2); + KP_CLEAR(kp_missed); +#ifdef CONFIG_KRETPROBES + KP_CLEAR(rp.kp); + KP_CLEAR(rp2.kp); +#ifdef CONFIG_ARCH_CORRECT_STACKTRACE_ON_KRETPROBE + KP_CLEAR(rp3.kp); + KP_CLEAR(rp4.kp); +#endif +#endif + target = kprobe_target; target2 = kprobe_target2; recursed_target = kprobe_recursed_target; From 5e28b7b94408897e41c63477aabc9e1db439bc8c Mon Sep 17 00:00:00 2001 From: "Francis, David" Date: Tue, 28 Apr 2026 19:25:50 +0000 Subject: [PATCH 205/333] drm: Set old handle to NULL before prime swap in change_handle There was a potential race condition in change_handle. The ioctl briefly had a single object with two idr entries; a concurrent gem_close could delete the object and remove one of the handles while leaving the other one dangling, which could subsequently be dereferenced for a use-after-free. To fix this, do the same dance that gem_close itself does. (f6cd7daecff5 drm: Release driver references to handle before making it available again) First idr_replace the old handle to NULL. Later, if the prime operations are successful, actually close it. create_tail required a similar dance to avoid a similar problem. (bd46cece51a3 drm/gem: Fix race in drm_gem_handle_create_tail()) It idr_allocs the new handle with NULL, then swaps in the correct object later to avoid races. We don't need to do that here, since the only operations that could race are drm_prime, and change_handle holds the prime lock for the entire duration. v2: cleanups of error paths Signed-off-by: David Francis Co-authored-by: Dave Airlie Reported-by: Puttimet Thammasaeng Tested-by: Vitaly Prosyak Cc: Simona Vetter Cc: stable@vger.kernel.org Cc: Christian Koenig Fixes: 53096728b8910 ("drm: Add DRM prime interface to reassign GEM handle") Signed-off-by: Dave Airlie --- drivers/gpu/drm/drm_gem.c | 25 ++++++++++++++++++++++++- 1 file changed, 24 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c index d6424267260b..51a887cc7fd7 100644 --- a/drivers/gpu/drm/drm_gem.c +++ b/drivers/gpu/drm/drm_gem.c @@ -1019,7 +1019,7 @@ int drm_gem_change_handle_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { struct drm_gem_change_handle *args = data; - struct drm_gem_object *obj; + struct drm_gem_object *obj, *idrobj; int handle, ret; if (!drm_core_check_feature(dev, DRIVER_GEM)) @@ -1042,8 +1042,29 @@ int drm_gem_change_handle_ioctl(struct drm_device *dev, void *data, mutex_lock(&file_priv->prime.lock); spin_lock(&file_priv->table_lock); + + /* When create_tail allocs an obj idr, it needs to first alloc as NULL, + * then later replace with the correct object. This is not necessary + * here, because the only operations that could race are drm_prime + * bookkeeping, and we hold the prime lock. + */ ret = idr_alloc(&file_priv->object_idr, obj, handle, handle + 1, GFP_NOWAIT); + + if (ret < 0) { + spin_unlock(&file_priv->table_lock); + goto out_unlock; + } + + idrobj = idr_replace(&file_priv->object_idr, NULL, handle); + if (idrobj != obj) { + idr_replace(&file_priv->object_idr, idrobj, handle); + idr_remove(&file_priv->object_idr, args->new_handle); + spin_unlock(&file_priv->table_lock); + ret = -ENOENT; + goto out_unlock; + } + spin_unlock(&file_priv->table_lock); if (ret < 0) @@ -1055,6 +1076,8 @@ int drm_gem_change_handle_ioctl(struct drm_device *dev, void *data, if (ret < 0) { spin_lock(&file_priv->table_lock); idr_remove(&file_priv->object_idr, handle); + idrobj = idr_replace(&file_priv->object_idr, obj, handle); + WARN_ON(idrobj != NULL); spin_unlock(&file_priv->table_lock); goto out_unlock; } From f03e8583532941b07761c5429de7d50766fa3110 Mon Sep 17 00:00:00 2001 From: Jiexun Wang Date: Sun, 3 May 2026 12:28:58 +0800 Subject: [PATCH 206/333] batman-adv: stop caching unowned originator pointers in BAT IV BAT IV keeps the last-hop neighbor address in each neigh_node, but some paths also cache an originator pointer derived from a temporary lookup. That pointer is not owned by the neigh_node and may no longer refer to a live originator entry after purge handling runs. Stop storing the auxiliary originator pointer in the BAT IV neighbor state. When BAT IV needs the neighbor originator data, resolve it from the stored neighbor address and drop the reference again after use. Fixes: c6c8fea29769 ("net: Add batman-adv meshing protocol") Cc: stable@kernel.org Reported-by: Yuan Tan Reported-by: Yifan Wu Reported-by: Juefei Pu Reported-by: Xin Liu Signed-off-by: Jiexun Wang Signed-off-by: Ren Wei [sven: avoid bonding logic for outgoing OGM] Signed-off-by: Sven Eckelmann --- net/batman-adv/bat_iv_ogm.c | 83 ++++++++++++++++++++++++++----------- 1 file changed, 59 insertions(+), 24 deletions(-) diff --git a/net/batman-adv/bat_iv_ogm.c b/net/batman-adv/bat_iv_ogm.c index 618d1889c04e..74ef7dc2b2f9 100644 --- a/net/batman-adv/bat_iv_ogm.c +++ b/net/batman-adv/bat_iv_ogm.c @@ -173,19 +173,12 @@ batadv_iv_ogm_orig_get(struct batadv_priv *bat_priv, const u8 *addr) static struct batadv_neigh_node * batadv_iv_ogm_neigh_new(struct batadv_hard_iface *hard_iface, const u8 *neigh_addr, - struct batadv_orig_node *orig_node, - struct batadv_orig_node *orig_neigh) + struct batadv_orig_node *orig_node) { struct batadv_neigh_node *neigh_node; neigh_node = batadv_neigh_node_get_or_create(orig_node, hard_iface, neigh_addr); - if (!neigh_node) - goto out; - - neigh_node->orig_node = orig_neigh; - -out: return neigh_node; } @@ -906,6 +899,31 @@ static u8 batadv_iv_orig_ifinfo_sum(struct batadv_orig_node *orig_node, return sum; } +/** + * batadv_iv_ogm_neigh_ifinfo_sum() - Get bcast_own sum for a last-hop neighbor + * @bat_priv: the bat priv with all the mesh interface information + * @neigh_node: last-hop neighbor of an originator + * + * Return: Number of replied (rebroadcasted) OGMs for the originator currently + * announced by the neighbor. Returns 0 if the neighbor's originator entry is + * not available anymore. + */ +static u8 batadv_iv_ogm_neigh_ifinfo_sum(struct batadv_priv *bat_priv, + const struct batadv_neigh_node *neigh_node) +{ + struct batadv_orig_node *orig_neigh; + u8 sum; + + orig_neigh = batadv_orig_hash_find(bat_priv, neigh_node->addr); + if (!orig_neigh) + return 0; + + sum = batadv_iv_orig_ifinfo_sum(orig_neigh, neigh_node->if_incoming); + batadv_orig_node_put(orig_neigh); + + return sum; +} + /** * batadv_iv_ogm_orig_update() - use OGM to update corresponding data in an * originator @@ -975,17 +993,9 @@ batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv, } if (!neigh_node) { - struct batadv_orig_node *orig_tmp; - - orig_tmp = batadv_iv_ogm_orig_get(bat_priv, ethhdr->h_source); - if (!orig_tmp) - goto unlock; - neigh_node = batadv_iv_ogm_neigh_new(if_incoming, ethhdr->h_source, - orig_node, orig_tmp); - - batadv_orig_node_put(orig_tmp); + orig_node); if (!neigh_node) goto unlock; } else { @@ -1037,10 +1047,9 @@ batadv_iv_ogm_orig_update(struct batadv_priv *bat_priv, */ if (router_ifinfo && neigh_ifinfo->bat_iv.tq_avg == router_ifinfo->bat_iv.tq_avg) { - sum_orig = batadv_iv_orig_ifinfo_sum(router->orig_node, - router->if_incoming); - sum_neigh = batadv_iv_orig_ifinfo_sum(neigh_node->orig_node, - neigh_node->if_incoming); + sum_orig = batadv_iv_ogm_neigh_ifinfo_sum(bat_priv, router); + sum_neigh = batadv_iv_ogm_neigh_ifinfo_sum(bat_priv, + neigh_node); if (sum_orig >= sum_neigh) goto out; } @@ -1106,7 +1115,6 @@ static bool batadv_iv_ogm_calc_tq(struct batadv_orig_node *orig_node, if (!neigh_node) neigh_node = batadv_iv_ogm_neigh_new(if_incoming, orig_neigh_node->orig, - orig_neigh_node, orig_neigh_node); if (!neigh_node) @@ -1302,6 +1310,32 @@ batadv_iv_ogm_update_seqnos(const struct ethhdr *ethhdr, return ret; } +/** + * batadv_orig_to_direct_router() - get direct next hop neighbor to an orig address + * @bat_priv: the bat priv with all the mesh interface information + * @orig_addr: the originator MAC address to search the best next hop router for + * @if_outgoing: the interface where the OGM should be sent to + * + * Return: A neighbor node which is the best router towards the given originator + * address. Bonding candidates are ignored. + */ +static struct batadv_neigh_node * +batadv_orig_to_direct_router(struct batadv_priv *bat_priv, u8 *orig_addr, + struct batadv_hard_iface *if_outgoing) +{ + struct batadv_neigh_node *neigh_node; + struct batadv_orig_node *orig_node; + + orig_node = batadv_orig_hash_find(bat_priv, orig_addr); + if (!orig_node) + return NULL; + + neigh_node = batadv_orig_router_get(orig_node, if_outgoing); + batadv_orig_node_put(orig_node); + + return neigh_node; +} + /** * batadv_iv_ogm_process_per_outif() - process a batman iv OGM for an outgoing * interface @@ -1372,8 +1406,9 @@ batadv_iv_ogm_process_per_outif(const struct sk_buff *skb, int ogm_offset, router = batadv_orig_router_get(orig_node, if_outgoing); if (router) { - router_router = batadv_orig_router_get(router->orig_node, - if_outgoing); + router_router = batadv_orig_to_direct_router(bat_priv, + router->addr, + if_outgoing); router_ifinfo = batadv_neigh_ifinfo_get(router, if_outgoing); } From ce425dd05d0fe7594930a0fb103634f35ac47bb6 Mon Sep 17 00:00:00 2001 From: Sven Eckelmann Date: Wed, 6 May 2026 22:20:49 +0200 Subject: [PATCH 207/333] batman-adv: tp_meter: fix tp_num leak on kmalloc failure When batadv_tp_start() or batadv_tp_init_recv() fail to allocate a new tp_vars object, the previously incremented bat_priv->tp_num counter is never decremented. This causes tp_num to drift upward on each allocation failure. Since only BATADV_TP_MAX_NUM sessions can be started and the count is never reduced for these failed allocations, it causes to an exhaustion of throughput meter sessions. In worst case, no new throughput meter session can be started until the mesh interface is removed. The error handling must decrement tp_num releasing the lock and aborting the creation of an throughput meter session Cc: stable@kernel.org Fixes: 33a3bb4a3345 ("batman-adv: throughput meter implementation") Signed-off-by: Sven Eckelmann --- net/batman-adv/tp_meter.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/net/batman-adv/tp_meter.c b/net/batman-adv/tp_meter.c index 58ca59a2799e..066c76113fc4 100644 --- a/net/batman-adv/tp_meter.c +++ b/net/batman-adv/tp_meter.c @@ -994,6 +994,7 @@ void batadv_tp_start(struct batadv_priv *bat_priv, const u8 *dst, tp_vars = kmalloc_obj(*tp_vars, GFP_ATOMIC); if (!tp_vars) { + atomic_dec(&bat_priv->tp_num); spin_unlock_bh(&bat_priv->tp_list_lock); batadv_dbg(BATADV_DBG_TP_METER, bat_priv, "Meter: %s cannot allocate list elements\n", @@ -1366,8 +1367,10 @@ batadv_tp_init_recv(struct batadv_priv *bat_priv, } tp_vars = kmalloc_obj(*tp_vars, GFP_ATOMIC); - if (!tp_vars) + if (!tp_vars) { + atomic_dec(&bat_priv->tp_num); goto out_unlock; + } ether_addr_copy(tp_vars->other_end, icmp->orig); tp_vars->role = BATADV_TP_RECEIVER; From 4ae1709a314060a196981b344610d023ea841e57 Mon Sep 17 00:00:00 2001 From: Sven Eckelmann Date: Wed, 6 May 2026 22:20:50 +0200 Subject: [PATCH 208/333] batman-adv: bla: prevent use-after-free when deleting claims When batadv_bla_del_backbone_claims() removes all claims for a backbone, it does this by dropping the link entry in the hash list. This list entry itself was one of the references which need to be dropped at the same time via batadv_claim_put(). But the batadv_claim_put() must not be done before the last access to the claim object in this function. Otherwise the claim might be freed already by the batadv_claim_release() function before the list entry was dropped. Cc: stable@kernel.org Fixes: 23721387c409 ("batman-adv: add basic bridge loop avoidance code") Signed-off-by: Sven Eckelmann --- net/batman-adv/bridge_loop_avoidance.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/batman-adv/bridge_loop_avoidance.c b/net/batman-adv/bridge_loop_avoidance.c index 51fe028b9088..8b77dd2ecfa4 100644 --- a/net/batman-adv/bridge_loop_avoidance.c +++ b/net/batman-adv/bridge_loop_avoidance.c @@ -318,8 +318,8 @@ batadv_bla_del_backbone_claims(struct batadv_bla_backbone_gw *backbone_gw) if (claim->backbone_gw != backbone_gw) continue; - batadv_claim_put(claim); hlist_del_rcu(&claim->hash_entry); + batadv_claim_put(claim); } spin_unlock_bh(list_lock); } From cf6b604011591865ae39ac82de8978c1120d17af Mon Sep 17 00:00:00 2001 From: Sven Eckelmann Date: Wed, 6 May 2026 22:20:51 +0200 Subject: [PATCH 209/333] batman-adv: bla: only purge non-released claims When batadv_bla_purge_claims() goes through the list of claims, it is only traversing the hash list with an rcu_read_lock(). Due to a potential parallel batadv_claim_put(), it can happen that it encounters a claim which was actually in the process of being released+freed by batadv_claim_release(). In this case, backbone_gw is set to NULL before the delayed RCU kfree is started. Calling batadv_bla_claim_get_backbone_gw() is then no longer allowed because it would cause a NULL-ptr derefence. To avoid this, only claims with a valid reference counter must be purged. All others are already taken care of. Cc: stable@kernel.org Fixes: 23721387c409 ("batman-adv: add basic bridge loop avoidance code") Signed-off-by: Sven Eckelmann --- net/batman-adv/bridge_loop_avoidance.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/net/batman-adv/bridge_loop_avoidance.c b/net/batman-adv/bridge_loop_avoidance.c index 8b77dd2ecfa4..879ab043d57a 100644 --- a/net/batman-adv/bridge_loop_avoidance.c +++ b/net/batman-adv/bridge_loop_avoidance.c @@ -1288,6 +1288,13 @@ static void batadv_bla_purge_claims(struct batadv_priv *bat_priv, rcu_read_lock(); hlist_for_each_entry_rcu(claim, head, hash_entry) { + /* only purge claims not currently in the process of being released. + * Such claims could otherwise have a NULL-ptr backbone_gw set because + * they already went through batadv_claim_release() + */ + if (!kref_get_unless_zero(&claim->refcount)) + continue; + backbone_gw = batadv_bla_claim_get_backbone_gw(claim); if (now) goto purge_now; @@ -1313,6 +1320,7 @@ static void batadv_bla_purge_claims(struct batadv_priv *bat_priv, claim->addr, claim->vid); skip: batadv_backbone_gw_put(backbone_gw); + batadv_claim_put(claim); } rcu_read_unlock(); } From ba9d20ee9076dac32c371116bacbe72480eb356c Mon Sep 17 00:00:00 2001 From: Sven Eckelmann Date: Wed, 6 May 2026 22:20:52 +0200 Subject: [PATCH 210/333] batman-adv: bla: put backbone reference on failed claim hash insert When batadv_bla_add_claim() fails to insert a new claim into the hash, it leaked a reference to the backbone_gw for which the claim was intended. Call batadv_backbone_gw_put() on the error path to release the reference and avoid leaking the backbone_gw object. Cc: stable@kernel.org Fixes: 3db0decf1185 ("batman-adv: Fix non-atomic bla_claim::backbone_gw access") Signed-off-by: Sven Eckelmann --- net/batman-adv/bridge_loop_avoidance.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/batman-adv/bridge_loop_avoidance.c b/net/batman-adv/bridge_loop_avoidance.c index 879ab043d57a..cec11f1251d6 100644 --- a/net/batman-adv/bridge_loop_avoidance.c +++ b/net/batman-adv/bridge_loop_avoidance.c @@ -723,6 +723,7 @@ static void batadv_bla_add_claim(struct batadv_priv *bat_priv, if (unlikely(hash_added != 0)) { /* only local changes happened. */ + batadv_backbone_gw_put(backbone_gw); kfree(claim); return; } From f7700a4415afb3ac1767a556094e4ef8bd440e41 Mon Sep 17 00:00:00 2001 From: Ming Lei Date: Fri, 8 May 2026 20:37:46 +0800 Subject: [PATCH 211/333] ublk: fix use-after-free in ublk_cancel_cmd() When ublk_reset_ch_dev() clears io->cmd via ublk_queue_reinit() concurrently with ublk_cancel_cmd(), ublk_cancel_cmd() can read a stale pointer and pass it to io_uring_cmd_done(), causing a use-after-free. Fix by synchronizing the two paths with ubq->cancel_lock: - ublk_cancel_cmd(): read and clear io->cmd under cancel_lock, then call io_uring_cmd_done() on the saved local copy outside the lock. - ublk_reset_ch_dev(): hold cancel_lock across ublk_queue_reinit() so that io->cmd and io->flags are cleared atomically with respect to ublk_cancel_cmd(). Fixes: 216c8f5ef0f2 ("ublk: replace monitor with cancelable uring_cmd") Signed-off-by: Ming Lei Link: https://patch.msgid.link/20260508123746.242018-1-tom.leiming@gmail.com Signed-off-by: Jens Axboe --- drivers/block/ublk_drv.c | 20 +++++++++++++++----- 1 file changed, 15 insertions(+), 5 deletions(-) diff --git a/drivers/block/ublk_drv.c b/drivers/block/ublk_drv.c index 57ec900f0ce0..6d13f1481de0 100644 --- a/drivers/block/ublk_drv.c +++ b/drivers/block/ublk_drv.c @@ -2411,8 +2411,14 @@ static void ublk_reset_ch_dev(struct ublk_device *ub) { int i; - for (i = 0; i < ub->dev_info.nr_hw_queues; i++) - ublk_queue_reinit(ub, ublk_get_queue(ub, i)); + for (i = 0; i < ub->dev_info.nr_hw_queues; i++) { + struct ublk_queue *ubq = ublk_get_queue(ub, i); + + /* Sync with ublk_cancel_cmd() */ + spin_lock(&ubq->cancel_lock); + ublk_queue_reinit(ub, ubq); + spin_unlock(&ubq->cancel_lock); + } /* set to NULL, otherwise new tasks cannot mmap io_cmd_buf */ ub->mm = NULL; @@ -2753,6 +2759,7 @@ static void ublk_cancel_cmd(struct ublk_queue *ubq, unsigned tag, { struct ublk_io *io = &ubq->ios[tag]; struct ublk_device *ub = ubq->dev; + struct io_uring_cmd *cmd = NULL; struct request *req; bool done; @@ -2775,12 +2782,15 @@ static void ublk_cancel_cmd(struct ublk_queue *ubq, unsigned tag, spin_lock(&ubq->cancel_lock); done = !!(io->flags & UBLK_IO_FLAG_CANCELED); - if (!done) + if (!done) { io->flags |= UBLK_IO_FLAG_CANCELED; + cmd = io->cmd; + io->cmd = NULL; + } spin_unlock(&ubq->cancel_lock); - if (!done) - io_uring_cmd_done(io->cmd, UBLK_IO_RES_ABORT, issue_flags); + if (!done && cmd) + io_uring_cmd_done(cmd, UBLK_IO_RES_ABORT, issue_flags); } /* From ab28a0673daabe7f0fcbd7a5e36334f2f003f02f Mon Sep 17 00:00:00 2001 From: Zqiang Date: Fri, 8 May 2026 19:50:45 +0800 Subject: [PATCH 212/333] sched_ext: Use IRQ_WORK_INIT_HARD() to initialize sch->disable_irq_work For built with PREEMPT_RT kernels, the scx_disable_irq_workfn() is called from per-cpu irq_work kthreads context, this means that when call the scx_dump_state() in the scx_disable_irq_workfn() to output current->comm/pid, it always output current irq_work kthread's comm/pid. this commit therefore use the IRQ_WORK_INIT_HARD() to initialize sch->disable_irq_work to make scx_disable_irq_workfn() is called from hardirq context. Fixes: f4a6c506d118 ("sched_ext: Always bounce scx_disable() through irq_work") Signed-off-by: Zqiang Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index f4e2db8e56be..df305712a2d4 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -6589,7 +6589,7 @@ static struct scx_sched *scx_alloc_and_add_sched(struct sched_ext_ops *ops, sch->slice_dfl = SCX_SLICE_DFL; atomic_set(&sch->exit_kind, SCX_EXIT_NONE); - init_irq_work(&sch->disable_irq_work, scx_disable_irq_workfn); + sch->disable_irq_work = IRQ_WORK_INIT_HARD(scx_disable_irq_workfn); kthread_init_work(&sch->disable_work, scx_disable_workfn); timer_setup(&sch->bypass_lb_timer, scx_bypass_lb_timerfn, 0); From 481c2265286ef302327c93403a8cf7b3fe4506d0 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Mon, 4 May 2026 21:04:48 +0000 Subject: [PATCH 213/333] bpf: tcp: Fix type confusion in bpf_tcp_sock(). bpf_tcp_sock() only checks if sk->sk_protocol is IPPROTO_TCP, but RAW socket can bypass it: socket(AF_INET, SOCK_RAW, IPPROTO_TCP) Calling bpf_setsockopt() in SOCKOPT prog triggers out-of-bounds access to another slab object. [0] Let's use sk_is_tcp(). [0]: BUG: KASAN: slab-out-of-bounds in sol_tcp_sockopt (net/core/filter.c:5519) Read of size 8 at addr ffff88801083d760 by task test_progs/1259 CPU: 1 UID: 0 PID: 1259 Comm: test_progs Tainted: G OE 7.0.0-11175-gb5c111f4967b #1 PREEMPT(full) Tainted: [O]=OOT_MODULE, [E]=UNSIGNED_MODULE Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.17.0-debian-1.17.0-1 04/01/2014 Call Trace: dump_stack_lvl (lib/dump_stack.c:94 lib/dump_stack.c:120) print_report (mm/kasan/report.c:378 mm/kasan/report.c:482) kasan_report (mm/kasan/report.c:595) sol_tcp_sockopt (net/core/filter.c:5519) __bpf_getsockopt (net/core/filter.c:5633) bpf_sk_getsockopt (net/core/filter.c:5654) bpf_prog_629ba00a1601e9f2__setsockopt+0x86/0x22c __cgroup_bpf_run_filter_setsockopt (./include/linux/bpf.h:1402 ./include/linux/filter.h:722 ./include/linux/filter.h:729 kernel/bpf/cgroup.c:81 kernel/bpf/cgroup.c:2026) do_sock_setsockopt (net/socket.c:2363) __x64_sys_setsockopt (net/socket.c:2406) do_syscall_64 (arch/x86/entry/syscall_64.c:63) entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121) RIP: 0033:0x7f85f82fe7de Code: 55 48 63 c9 48 63 ff 45 89 c9 48 89 e5 48 83 ec 08 6a 2c e8 34 69 f7 ff c9 c3 66 90 f3 0f 1e fa 49 89 ca b8 36 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 0a c3 66 0f 1f 84 00 00 00 00 00 48 8b 15 e1 RSP: 002b:00007ffe59dcecd8 EFLAGS: 00000202 ORIG_RAX: 0000000000000036 RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007f85f82fe7de RDX: 000000000000001c RSI: 0000000000000006 RDI: 000000000000000d RBP: 00007ffe59dcef20 R08: 000000000000003c R09: 0000000000000000 R10: 00007ffe59dcef00 R11: 0000000000000202 R12: 00007ffe59dcf268 R13: 0000000000000003 R14: 00007f85f9da5000 R15: 000055b2f3201400 The buggy address belongs to the object at ffff88801083d280 which belongs to the cache RAW of size 1792 The buggy address is located 1248 bytes inside of allocated 1792-byte region [ffff88801083d280, ffff88801083d980) Fixes: 655a51e536c0 ("bpf: Add struct bpf_tcp_sock and BPF_FUNC_tcp_sock") Reported-by: Damiano Melotti Signed-off-by: Kuniyuki Iwashima Signed-off-by: Martin KaFai Lau Acked-by: Daniel Borkmann Link: https://patch.msgid.link/20260504210610.180150-2-kuniyu@google.com --- net/core/filter.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/core/filter.c b/net/core/filter.c index bc96c18df4e0..cd88633f8dc1 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -7475,7 +7475,7 @@ u32 bpf_tcp_sock_convert_ctx_access(enum bpf_access_type type, BPF_CALL_1(bpf_tcp_sock, struct sock *, sk) { - if (sk_fullsock(sk) && sk->sk_protocol == IPPROTO_TCP) + if (sk_fullsock(sk) && sk_is_tcp(sk)) return (unsigned long)sk; return (unsigned long)NULL; From a7488f089bdfa87c4fef1744d4dca9f4f8b46f8b Mon Sep 17 00:00:00 2001 From: Breno Leitao Date: Thu, 7 May 2026 04:04:46 -0700 Subject: [PATCH 214/333] workqueue: Release PENDING in __queue_work() drain/destroy reject path The caller of __queue_work() owns WORK_STRUCT_PENDING, won via test_and_set_bit() in queue_work_on()/__queue_delayed_work(). The state machine documented above __queue_work() requires that owner to either hand the token to a pwq (insert_work() -> set_work_pwq()), hand it to a timer, or release it via set_work_pool_and_clear_pending(). try_to_grab_pending() relies on this: when it observes "PENDING && off-queue" it busy-loops, trusting the current owner to make progress. The (__WQ_DESTROYING | __WQ_DRAINING) early-return path violates that contract. It WARN_ONCE()s and bare-returns, leaving work->data with PENDING set, WORK_STRUCT_PWQ clear, and work->entry empty. The path is reachable without explicit API abuse: queue_delayed_work() arms a timer with PENDING set; if drain_workqueue() runs while the timer is still pending, delayed_work_timer_fn() -> __queue_work() in softirq context hits the WARN, current is not a wq worker so is_chained_work() is false, and the work is silently dropped with PENDING leaked. Mirror what clear_pending_if_disabled() already does on its analogous reject path: unpack the off-queue data and call set_work_pool_and_clear_pending() to release the token before returning. I was able to reproduce this by queueing several slow works on a max_active=1 wq, arm a delayed_work whose timer fires while drain_workqueue() is blocked, then call cancel_delayed_work_sync(). Without this patch the cancel livelocks at 100% CPU; with it the cancel returns immediately. Signed-off-by: Breno Leitao Signed-off-by: Tejun Heo --- kernel/workqueue.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 3d2e3b2ec528..c0e58f877e08 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -2296,6 +2296,18 @@ static void __queue_work(int cpu, struct workqueue_struct *wq, if (unlikely(wq->flags & (__WQ_DESTROYING | __WQ_DRAINING) && WARN_ONCE(!is_chained_work(wq), "workqueue: cannot queue %ps on wq %s\n", work->func, wq->name))) { + struct work_offq_data offqd; + + /* + * State on entry: PENDING is set, work is off-queue (no + * insert_work() has run). + * + * Returning without clearing PENDING would leave the work + * in a weird state (PENDING=1, PWQ=0, entry empty) + */ + work_offqd_unpack(&offqd, *work_data_bits(work)); + set_work_pool_and_clear_pending(work, offqd.pool_id, + work_offqd_pack_flags(&offqd)); return; } rcu_read_lock(); From 0143033dc22cdff912cfc13419f5db92fea3b4cb Mon Sep 17 00:00:00 2001 From: Breno Leitao Date: Fri, 8 May 2026 09:22:03 -0700 Subject: [PATCH 215/333] workqueue: Fix wq->cpu_pwq leak in alloc_and_link_pwqs() WQ_UNBOUND path For WQ_UNBOUND workqueues, alloc_and_link_pwqs() allocates wq->cpu_pwq via alloc_percpu() and then calls apply_workqueue_attrs_locked(). On failure it returns the error directly, bypassing the enomem: label which holds the only free_percpu(wq->cpu_pwq) in this function. The caller's error path kfree()s wq without touching wq->cpu_pwq, leaking one percpu pointer table (nr_cpu_ids * sizeof(void *) bytes) per failed call. If kmemleak is enabled, we can see: unreferenced object (percpu) 0xc0fffa5b121048 (size 8): comm "insmod", pid 776, jiffies 4294682844 backtrace (crc 0): pcpu_alloc_noprof+0x665/0xac0 __alloc_workqueue+0x33f/0xa20 alloc_workqueue_noprof+0x60/0x100 Route the error through the existing enomem: cleanup and any error before this one. Cc: stable@kernel.org Fixes: 636b927eba5b ("workqueue: Make unbound workqueues to use per-cpu pool_workqueues") Signed-off-by: Breno Leitao Signed-off-by: Tejun Heo --- kernel/workqueue.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/kernel/workqueue.c b/kernel/workqueue.c index c0e58f877e08..33b721a9af02 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -5654,7 +5654,9 @@ static int alloc_and_link_pwqs(struct workqueue_struct *wq) ret = apply_workqueue_attrs_locked(wq, unbound_std_wq_attrs[highpri]); } - return ret; + if (ret) + goto enomem; + return 0; enomem: if (wq->cpu_pwq) { From d73549b8bb7fa6147666c579d66f72bf076f719f Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Mon, 4 May 2026 21:04:49 +0000 Subject: [PATCH 216/333] selftest: bpf: Add test for bpf_tcp_sock() and RAW socket. Let's extend sockopt_sk.c to cover bpf_tcp_sock() for the wrong socket type. Before: # ./test_progs -t sockopt_sk [ 151.948613] ================================================================== [ 151.951376] BUG: KASAN: slab-out-of-bounds in sol_tcp_sockopt+0xc7/0x8e0 [ 151.954159] Read of size 8 at addr ffff88801083d760 by task test_progs/1259 ... run_test:FAIL:getsetsockopt unexpected error: -1 (errno 0) #427 sockopt_sk:FAIL After: #427 sockopt_sk:OK While at it, missing free() is fixed up. Signed-off-by: Kuniyuki Iwashima Signed-off-by: Martin KaFai Lau Link: https://patch.msgid.link/20260504210610.180150-3-kuniyu@google.com --- .../selftests/bpf/prog_tests/sockopt_sk.c | 17 ++++++++++++++++- tools/testing/selftests/bpf/progs/sockopt_sk.c | 16 ++++++++++++++++ 2 files changed, 32 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/bpf/prog_tests/sockopt_sk.c b/tools/testing/selftests/bpf/prog_tests/sockopt_sk.c index 53637431ec5d..3a41c517b918 100644 --- a/tools/testing/selftests/bpf/prog_tests/sockopt_sk.c +++ b/tools/testing/selftests/bpf/prog_tests/sockopt_sk.c @@ -190,7 +190,7 @@ static int getsetsockopt(void) fd = socket(AF_NETLINK, SOCK_RAW, 0); if (fd < 0) { log_err("Failed to create AF_NETLINK socket"); - return -1; + goto err; } buf.u32 = 1; @@ -211,6 +211,21 @@ static int getsetsockopt(void) } ASSERT_EQ(optlen, 8, "Unexpected NETLINK_LIST_MEMBERSHIPS value"); + /* Trick bpf_tcp_sock() with IPPROTO_TCP */ + close(fd); + fd = socket(AF_INET, SOCK_RAW, IPPROTO_TCP); + if (!ASSERT_OK_FD(fd, "socket")) + goto err; + + /* The BPF prog intercepts this before the kernel sees it, any + * optlen works. Go with 4 bytes for simplicity. + */ + buf.u32 = 1; + optlen = sizeof(buf.u32); + err = setsockopt(fd, SOL_TCP, TCP_SAVED_SYN, &buf, optlen); + if (!ASSERT_ERR(err, "setsockopt(TCP_SAVED_SYN)")) + goto err; + free(big_buf); close(fd); return 0; diff --git a/tools/testing/selftests/bpf/progs/sockopt_sk.c b/tools/testing/selftests/bpf/progs/sockopt_sk.c index cb990a7d3d45..5e0b27e7855c 100644 --- a/tools/testing/selftests/bpf/progs/sockopt_sk.c +++ b/tools/testing/selftests/bpf/progs/sockopt_sk.c @@ -149,6 +149,20 @@ int _setsockopt(struct bpf_sockopt *ctx) if (sk && sk->family == AF_NETLINK) goto out; + if (sk && sk->family == AF_INET && sk->type == SOCK_RAW) { + struct bpf_tcp_sock *tp = bpf_tcp_sock(sk); + + if (tp) { + char saved_syn[60]; + + bpf_getsockopt(sk, SOL_TCP, TCP_SAVED_SYN, + &saved_syn, sizeof(saved_syn)); + goto consumed; + } + + goto out; + } + /* Make sure bpf_get_netns_cookie is callable. */ if (bpf_get_netns_cookie(NULL) == 0) @@ -224,6 +238,8 @@ int _setsockopt(struct bpf_sockopt *ctx) return 0; /* couldn't get sk storage */ storage->val = optval[0]; + +consumed: ctx->optlen = -1; /* BPF has consumed this option, don't call kernel * setsockopt handler. */ From 7995b216a731db657f356f6ae37a42f445b9a0ec Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Mon, 4 May 2026 21:04:50 +0000 Subject: [PATCH 217/333] mptcp: bpf: Fix type confusion in bpf_mptcp_sock_from_subflow() bpf_mptcp_sock_from_subflow() only checks if sk->sk_protocol is IPPROTO_TCP, but RAW socket can bypass it: socket(AF_INET, SOCK_RAW, IPPROTO_TCP) In this case, it would NOT be valid to call sk_is_mptcp() which will assume sk is a pointer to a struct tcp_sock, and wrongly checks for: tcp_sk(sk)->is_mptcp. Fixes: 3bc253c2e652 ("bpf: Add bpf_skc_to_mptcp_sock_proto") Signed-off-by: Matthieu Baerts (NGI0) Signed-off-by: Kuniyuki Iwashima Signed-off-by: Martin KaFai Lau Reviewed-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20260504210610.180150-4-kuniyu@google.com --- net/mptcp/bpf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/mptcp/bpf.c b/net/mptcp/bpf.c index 8a16672b94e2..4cc16cbeb328 100644 --- a/net/mptcp/bpf.c +++ b/net/mptcp/bpf.c @@ -14,7 +14,7 @@ struct mptcp_sock *bpf_mptcp_sock_from_subflow(struct sock *sk) { - if (sk && sk_fullsock(sk) && sk->sk_protocol == IPPROTO_TCP && sk_is_mptcp(sk)) + if (sk && sk_fullsock(sk) && sk_is_tcp(sk) && sk_is_mptcp(sk)) return mptcp_sk(mptcp_subflow_ctx(sk)->conn); return NULL; From decb84b8383ab7acff94db208ef7ed19f9c55e1f Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Mon, 4 May 2026 21:04:51 +0000 Subject: [PATCH 218/333] bpf: tcp: Fix type confusion in bpf_skc_to_tcp_sock(). bpf_skc_to_tcp_sock() only checks if sk->sk_protocol is IPPROTO_TCP, but RAW socket can bypass it: socket(AF_INET, SOCK_RAW, IPPROTO_TCP) Let's use sk_is_tcp(). Fixes: 478cfbdf5f13 ("bpf: Add bpf_skc_to_{tcp, tcp_timewait, tcp_request}_sock() helpers") Signed-off-by: Kuniyuki Iwashima Signed-off-by: Martin KaFai Lau Link: https://patch.msgid.link/20260504210610.180150-5-kuniyu@google.com --- net/core/filter.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/core/filter.c b/net/core/filter.c index cd88633f8dc1..7d945dc2cb92 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -11963,7 +11963,7 @@ const struct bpf_func_proto bpf_skc_to_tcp6_sock_proto = { BPF_CALL_1(bpf_skc_to_tcp_sock, struct sock *, sk) { - if (sk && sk_fullsock(sk) && sk->sk_protocol == IPPROTO_TCP) + if (sk && sk_fullsock(sk) && sk_is_tcp(sk)) return (unsigned long)sk; return (unsigned long)NULL; From 843064b0a77eed3d6d63ffc53aeaa359672b4e12 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Mon, 4 May 2026 21:04:52 +0000 Subject: [PATCH 219/333] bpf: tcp: Fix type confusion in bpf_skc_to_tcp6_sock(). bpf_skc_to_tcp6_sock() only checks if sk->sk_protocol is IPPROTO_TCP and sk->sk_family is AF_INET6, but RAW socket can bypass it: socket(AF_INET6, SOCK_RAW, IPPROTO_TCP) Let's check sk->sk_type too. Fixes: af7ec1383361 ("bpf: Add bpf_skc_to_tcp6_sock() helper") Signed-off-by: Kuniyuki Iwashima Signed-off-by: Martin KaFai Lau Link: https://patch.msgid.link/20260504210610.180150-6-kuniyu@google.com --- net/core/filter.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/core/filter.c b/net/core/filter.c index 7d945dc2cb92..684922efd481 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -11947,7 +11947,7 @@ BPF_CALL_1(bpf_skc_to_tcp6_sock, struct sock *, sk) */ BTF_TYPE_EMIT(struct tcp6_sock); if (sk && sk_fullsock(sk) && sk->sk_protocol == IPPROTO_TCP && - sk->sk_family == AF_INET6) + sk->sk_type == SOCK_STREAM && sk->sk_family == AF_INET6) return (unsigned long)sk; return (unsigned long)NULL; From 1c2958e4ab1ed4594db16425dbcab33c56ea8330 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Mon, 4 May 2026 21:04:53 +0000 Subject: [PATCH 220/333] bpf: tcp: Fix type confusion in sol_tcp_sockopt(). sol_tcp_sockopt() only checks if sk->sk_protocol is IPPROTO_TCP, but RAW socket can bypass it: socket(AF_INET, SOCK_RAW, IPPROTO_TCP) Let's use sk_is_tcp(). Note that initially sol_tcp_sockopt() checked sk->sk_prot->setsockopt. Fixes: 2ab42c7b871f ("bpf: Check the protocol of a sock to agree the calls to bpf_setsockopt().") Signed-off-by: Kuniyuki Iwashima Signed-off-by: Martin KaFai Lau Link: https://patch.msgid.link/20260504210610.180150-7-kuniyu@google.com --- net/core/filter.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/core/filter.c b/net/core/filter.c index 684922efd481..ef0877eefaa7 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -5481,7 +5481,7 @@ static int sol_tcp_sockopt(struct sock *sk, int optname, char *optval, int *optlen, bool getopt) { - if (sk->sk_protocol != IPPROTO_TCP) + if (!sk_is_tcp(sk)) return -EINVAL; switch (optname) { From 18fc650ccd7fe3376eca89203668cfb8268f60df Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Sun, 26 Apr 2026 01:26:43 +0000 Subject: [PATCH 221/333] bpf: Free reuseport cBPF prog after RCU grace period. Eulgyu Kim reported the splat below with a repro. [0] The repro sets up a UDP reuseport group with a cBPF prog and replaces it with a new one while another thread is sending a UDP packet to the group. The reuseport prog is freed by sk_reuseport_prog_free(). bpf_prog_put() is called for "e"BPF prog to destruct through multiple stages while cBPF prog is freed immediately by bpf_release_orig_filter() and bpf_prog_free(). If a reuseport prog is detached from the setsockopt() path (reuseport_attach_prog() or reuseport_detach_prog()), sk_reuseport_prog_free() is called without waiting for RCU readers to complete, resulting in various bugs. Let's defer freeing the reuseport cBPF prog after one RCU grace period. Note "e"BPF prog is safe as is unless the fast path starts to touch fields destroyed in bpf_prog_put_deferred() and __bpf_prog_put_noref(). [0]: BUG: KASAN: vmalloc-out-of-bounds in reuseport_select_sock+0xedc/0x1220 net/core/sock_reuseport.c:596 Read of size 4 at addr ffffc9000051e004 by task slowme/10208 CPU: 6 UID: 1000 PID: 10208 Comm: slowme Not tainted 7.0.0-geb7ac95ff75e #32 PREEMPT(full) Hardware name: QEMU Ubuntu 24.04 PC v2 (i440FX + PIIX, arch_caps fix, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/2014 Call Trace: dump_stack_lvl+0xe8/0x150 lib/dump_stack.c:120 print_address_description mm/kasan/report.c:378 [inline] print_report+0xca/0x240 mm/kasan/report.c:482 kasan_report+0x118/0x150 mm/kasan/report.c:595 reuseport_select_sock+0xedc/0x1220 net/core/sock_reuseport.c:596 udp4_lib_lookup2+0x3bc/0x950 net/ipv4/udp.c:495 __udp4_lib_lookup+0x768/0xe20 net/ipv4/udp.c:723 __udp4_lib_lookup_skb+0x297/0x390 net/ipv4/udp.c:752 __udp4_lib_rcv+0x1312/0x2620 net/ipv4/udp.c:2752 ip_protocol_deliver_rcu+0x282/0x440 net/ipv4/ip_input.c:207 ip_local_deliver_finish+0x3bb/0x6f0 net/ipv4/ip_input.c:241 NF_HOOK+0x30c/0x3a0 include/linux/netfilter.h:318 NF_HOOK+0x30c/0x3a0 include/linux/netfilter.h:318 __netif_receive_skb_one_core net/core/dev.c:6181 [inline] __netif_receive_skb net/core/dev.c:6294 [inline] process_backlog+0xaa4/0x1960 net/core/dev.c:6645 __napi_poll+0xae/0x340 net/core/dev.c:7709 napi_poll net/core/dev.c:7772 [inline] net_rx_action+0x5d7/0xf50 net/core/dev.c:7929 handle_softirqs+0x22b/0x870 kernel/softirq.c:622 do_softirq+0x76/0xd0 kernel/softirq.c:523 __local_bh_enable_ip+0xf8/0x130 kernel/softirq.c:450 local_bh_enable include/linux/bottom_half.h:33 [inline] rcu_read_unlock_bh include/linux/rcupdate.h:924 [inline] __dev_queue_xmit+0x1dd7/0x3710 net/core/dev.c:4890 neigh_output include/net/neighbour.h:556 [inline] ip_finish_output2+0xca9/0x1070 net/ipv4/ip_output.c:237 NF_HOOK_COND include/linux/netfilter.h:307 [inline] ip_output+0x29f/0x450 net/ipv4/ip_output.c:438 ip_send_skb+0x45/0xc0 net/ipv4/ip_output.c:1508 udp_send_skb+0xb04/0x1510 net/ipv4/udp.c:1195 udp_sendmsg+0x1a71/0x2350 net/ipv4/udp.c:1485 sock_sendmsg_nosec net/socket.c:727 [inline] __sock_sendmsg net/socket.c:742 [inline] __sys_sendto+0x554/0x680 net/socket.c:2206 __do_sys_sendto net/socket.c:2213 [inline] __se_sys_sendto net/socket.c:2209 [inline] __x64_sys_sendto+0xde/0x100 net/socket.c:2209 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0x160/0xf80 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0033:0x415a2d Code: b3 66 2e 0f 1f 84 00 00 00 00 00 66 90 f3 0f 1e fa 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 b8 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007f6bc31e41e8 EFLAGS: 00000212 ORIG_RAX: 000000000000002c RAX: ffffffffffffffda RBX: 00007f6bc31e4cdc RCX: 0000000000415a2d RDX: 0000000000000001 RSI: 00007f6bc31e421f RDI: 0000000000000003 RBP: 00007f6bc31e4240 R08: 00007f6bc31e4220 R09: 0000000000000010 R10: 0000000000000000 R11: 0000000000000212 R12: 00007f6bc31e46c0 R13: ffffffffffffffb8 R14: 0000000000000000 R15: 00007ffc9b0d70b0 Fixes: 538950a1b752 ("soreuseport: setsockopt SO_ATTACH_REUSEPORT_[CE]BPF") Reported-by: Eulgyu Kim Reported-by: Taeyang Lee <0wn@theori.io> Signed-off-by: Kuniyuki Iwashima Signed-off-by: Daniel Borkmann Acked-by: Daniel Borkmann Link: https://lore.kernel.org/bpf/20260426012647.3233119-1-kuniyu@google.com --- net/core/filter.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) diff --git a/net/core/filter.c b/net/core/filter.c index ef0877eefaa7..70eef14edb99 100644 --- a/net/core/filter.c +++ b/net/core/filter.c @@ -1654,15 +1654,24 @@ int sk_reuseport_attach_bpf(u32 ufd, struct sock *sk) return err; } +static void sk_reuseport_prog_free_rcu(struct rcu_head *rcu) +{ + struct bpf_prog_aux *aux = container_of(rcu, struct bpf_prog_aux, rcu); + struct bpf_prog *prog = aux->prog; + + bpf_release_orig_filter(prog); + bpf_prog_free(prog); +} + void sk_reuseport_prog_free(struct bpf_prog *prog) { if (!prog) return; - if (prog->type == BPF_PROG_TYPE_SK_REUSEPORT) - bpf_prog_put(prog); + if (bpf_prog_was_classic(prog)) + call_rcu(&prog->aux->rcu, sk_reuseport_prog_free_rcu); else - bpf_prog_destroy(prog); + bpf_prog_put(prog); } static inline int __bpf_try_make_writable(struct sk_buff *skb, From 909f7bf9b080c10df3c3b38533906dbf09ff1d8b Mon Sep 17 00:00:00 2001 From: Lukas Wunner Date: Wed, 15 Apr 2026 17:56:06 +0200 Subject: [PATCH 222/333] PCI: Update saved_config_space upon resource assignment Bernd reports passthrough failure of a Digital Devices Cine S2 V6 DVB adapter plugged into an ASRock X570S PG Riptide board with BIOS version P5.41 (09/07/2023): ddbridge 0000:05:00.0: detected Digital Devices Cine S2 V6 DVB adapter ddbridge 0000:05:00.0: cannot read registers ddbridge 0000:05:00.0: fail BIOS assigns an incorrect BAR to the DVB adapter which doesn't fit into the upstream bridge window. The kernel corrects the BAR assignment: pci 0000:07:00.0: BAR 0 [mem 0xfffffffffc500000-0xfffffffffc50ffff 64bit]: can't claim; no compatible bridge window pci 0000:07:00.0: BAR 0 [mem 0xfc500000-0xfc50ffff 64bit]: assigned Correction of the BAR assignment happens in an x86-specific fs_initcall, pcibios_assign_resources(), after device enumeration in a subsys_initcall. This order was introduced at the behest of Linus in 2004: https://git.kernel.org/tglx/history/c/a06a30144bbc No other architecture performs such a late BAR correction. Bernd bisected the issue to commit a2f1e22390ac ("PCI/ERR: Ensure error recoverability at all times"), but it only occurs in the absence of commit 4d4c10f763d7 ("PCI: Explicitly put devices into D0 when initializing"). This combination exists in stable kernel v6.12.70, but not in mainline, hence Bernd cannot reproduce the issue with mainline. Since a2f1e22390ac, config space is saved on enumeration, prior to BAR correction. Upon passthrough, the corrected BAR is overwritten with the incorrect saved value by: vfio_pci_core_register_device() vfio_pci_set_power_state() pci_restore_state() But only if the device's current_state is PCI_UNKNOWN, as it was prior to commit 4d4c10f763d7. Since the commit, it is PCI_D0, which changes the behavior of vfio_pci_set_power_state() to no longer restore the state without saving it first. Alexandre is reporting the same issue as Bernd, but in his case, mainline is affected as well. The difference is that on Alexandre's system, the host kernel binds a driver to the device which is unbound prior to passthrough, whereas on Bernd's system no driver gets bound by the host kernel. Unbinding sets current_state to PCI_UNKNOWN in pci_device_remove(), so when vfio-pci is subsequently bound to the device, pci_restore_state() is once again called without invoking pci_save_state() first. To robustly fix the issue, always update saved_config_space upon resource assignment. Reported-by: Bernd Schumacher Closes: https://lore.kernel.org/r/acfZrlP0Ua_5D3U4@eldamar.lan/ Reported-by: Alexandre N. Closes: https://lore.kernel.org/r/dd3c3358-de0f-4a56-9c81-04aceaab4058@mailo.com/ Fixes: a2f1e22390ac ("PCI/ERR: Ensure error recoverability at all times") Signed-off-by: Lukas Wunner Signed-off-by: Bjorn Helgaas Tested-by: Bernd Schumacher Tested-by: Alexandre N. Cc: stable@vger.kernel.org # v6.12+ Link: https://patch.msgid.link/febc3f354e0c1f5a9f5b3ee9ffddaa44caccf651.1776268054.git.lukas@wunner.de --- drivers/pci/setup-res.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/pci/setup-res.c b/drivers/pci/setup-res.c index fbc05cda96ee..991d3ed543f5 100644 --- a/drivers/pci/setup-res.c +++ b/drivers/pci/setup-res.c @@ -102,6 +102,7 @@ static void pci_std_update_resource(struct pci_dev *dev, int resno) } pci_write_config_dword(dev, reg, new); + dev->saved_config_space[reg / 4] = new; pci_read_config_dword(dev, reg, &check); if ((new ^ check) & mask) { @@ -112,6 +113,7 @@ static void pci_std_update_resource(struct pci_dev *dev, int resno) if (res->flags & IORESOURCE_MEM_64) { new = region.start >> 16 >> 16; pci_write_config_dword(dev, reg + 4, new); + dev->saved_config_space[(reg + 4) / 4] = new; pci_read_config_dword(dev, reg + 4, &check); if (check != new) { pci_err(dev, "%s: error updating (high %#010x != %#010x)\n", From f45a49a2380a47332817b7248c61a0ebbc6f0d00 Mon Sep 17 00:00:00 2001 From: Samiullah Khawaja Date: Tue, 5 May 2026 23:43:27 +0000 Subject: [PATCH 223/333] PCI: Initialize temporary device in new_id_store() When setting new_id of a PCI device driver using sysfs a lockdep splat occurs. This is because new_id_store() builds a temporary pci_dev for pci_match_device(), which calls device_match_driver_override(). That depends on the driver_override.lock added by cb3d1049f4ea ("driver core: generalize driver_override in struct device"). The new driver_override.lock was not initialized in the temporary pci_dev, resulting in this lockdep splat. Initialize the temporary pci_dev to fix this. Repro: Build with CONFIG_LOCKDEP=y, boot with QEMU, and add a new ID: # echo "8086 10f5" > /sys/bus/pci/drivers/e1000e/new_id INFO: trying to register non-static key. The code is fine but needs lockdep annotation, or maybe you didn't initialize this object before use? turning off the locking correctness validator. CPU: 2 UID: 0 PID: 177 Comm: liveupdate-iomm Not tainted 7.0.0+ #9 PREEMPT(full) Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014 Call Trace: dump_stack_lvl+0x5d/0x80 register_lock_class+0x77e/0x790 lock_acquire+0xbf/0x2e0 pci_match_device+0x24/0x180 new_id_store+0x189/0x1d0 kernfs_fop_write_iter+0x14f/0x210 vfs_write+0x263/0x5e0 ksys_write+0x79/0xf0 do_syscall_64+0x117/0xf80 Fixes: 10a4206a2401 ("PCI: use generic driver_override infrastructure") Fixes: 8895d3bcb8ba ("PCI: Fail new_id for vendor/device values already built into driver") Signed-off-by: Samiullah Khawaja [bhelgaas: add commit log details and repro, trim backtrace] Signed-off-by: Bjorn Helgaas Reviewed-by: Danilo Krummrich Link: https://patch.msgid.link/20260505234327.716630-1-skhawaja@google.com --- drivers/pci/pci-driver.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c index d10ece0889f0..e3f59001785a 100644 --- a/drivers/pci/pci-driver.c +++ b/drivers/pci/pci-driver.c @@ -179,6 +179,11 @@ static const struct pci_device_id *pci_match_device(struct pci_driver *drv, return NULL; } +static void _pci_free_device(struct device *dev) +{ + kfree(to_pci_dev(dev)); +} + /** * new_id_store - sysfs frontend to pci_add_dynid() * @driver: target device driver @@ -214,11 +219,13 @@ static ssize_t new_id_store(struct device_driver *driver, const char *buf, pdev->subsystem_vendor = subvendor; pdev->subsystem_device = subdevice; pdev->class = class; + pdev->dev.release = _pci_free_device; + device_initialize(&pdev->dev); if (pci_match_device(pdrv, pdev)) retval = -EEXIST; - kfree(pdev); + put_device(&pdev->dev); if (retval) return retval; From bf5421b3d8d3d9216d045af89d4f7e2f75375554 Mon Sep 17 00:00:00 2001 From: Marek Vasut Date: Tue, 28 Apr 2026 07:19:54 +0200 Subject: [PATCH 224/333] MAINTAINERS: Update Marek Vasut email for PCIe R-Car Use up to date address. No functional change. Signed-off-by: Marek Vasut Signed-off-by: Bjorn Helgaas Link: https://patch.msgid.link/20260428052030.51101-1-marek.vasut+renesas@mailbox.org --- MAINTAINERS | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index 2fb1c75afd16..db06875cfb46 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -20433,7 +20433,7 @@ F: drivers/pci/controller/plda/pcie-plda-host.c F: drivers/pci/controller/plda/pcie-plda.h PCI DRIVER FOR RENESAS R-CAR -M: Marek Vasut +M: Marek Vasut M: Yoshihiro Shimoda L: linux-pci@vger.kernel.org L: linux-renesas-soc@vger.kernel.org From 78e115d806b0eea83a0e4df182a689823aa80511 Mon Sep 17 00:00:00 2001 From: Hans Zhang <18255117159@163.com> Date: Fri, 8 May 2026 10:30:06 +0800 Subject: [PATCH 225/333] MAINTAINERS: Update Hans Zhang email for PCIe CIX Sky1 Update my email address as my work email account is no longer in use. Signed-off-by: Hans Zhang <18255117159@163.com> Signed-off-by: Bjorn Helgaas Link: https://patch.msgid.link/20260508023006.1787674-1-18255117159@163.com --- MAINTAINERS | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index db06875cfb46..b10ee9a3b5f5 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -20321,7 +20321,7 @@ F: Documentation/devicetree/bindings/pci/cdns,* F: drivers/pci/controller/cadence/*cadence* PCI DRIVER FOR CIX Sky1 -M: Hans Zhang +M: Hans Zhang <18255117159@163.com> L: linux-pci@vger.kernel.org S: Maintained F: Documentation/devicetree/bindings/pci/cix,sky1-pcie-*.yaml From 9ef40a09c5de18f0d275a451f5c2a7fd4f07158b Mon Sep 17 00:00:00 2001 From: Aksh Garg Date: Fri, 8 May 2026 11:39:51 +0530 Subject: [PATCH 226/333] MAINTAINERS: Add Aksh Garg as PCIe CADENCE reviewer I wish to contribute to the review process for Cadence PCIe IP drivers, hence add myself as a reviewer. Signed-off-by: Aksh Garg Signed-off-by: Bjorn Helgaas Link: https://patch.msgid.link/20260508060951.840233-1-a-garg7@ti.com --- MAINTAINERS | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index b10ee9a3b5f5..0cafa37693a6 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -20315,10 +20315,11 @@ F: Documentation/devicetree/bindings/pci/marvell,armada8k-pcie.yaml F: drivers/pci/controller/dwc/pcie-armada8k.c PCI DRIVER FOR CADENCE PCIE IP +R: Aksh Garg L: linux-pci@vger.kernel.org S: Orphan F: Documentation/devicetree/bindings/pci/cdns,* -F: drivers/pci/controller/cadence/*cadence* +F: drivers/pci/controller/cadence/ PCI DRIVER FOR CIX Sky1 M: Hans Zhang <18255117159@163.com> From 97c8a3c1f73d828de43a5a88e8a9a143efb2b661 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Wed, 6 May 2026 03:59:18 +0000 Subject: [PATCH 227/333] tcp: Fix potential UAF in reqsk_timer_handler(). When TCP socket migration fails at inet_ehash_insert() in reqsk_timer_handler(), we jump to the no_ownership: label and free the new reqsk immediately with __reqsk_free(). Thus, we must stop the new reqsk's timer before jumping to the label, but the timer might be missed since the cited commit, resulting in UAF. As we are in the original reqsk's timer context, we can safely call timer_delete_sync() for the new reqsk. Let's pass false to __inet_csk_reqsk_queue_drop() to stop the new reqsk's timer. Fixes: 83fccfc3940c ("inet: fix potential deadlock in reqsk_queue_unlink()") Reported-by: Damiano Melotti Signed-off-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20260506035954.1563147-2-kuniyu@google.com Signed-off-by: Jakub Kicinski --- net/ipv4/inet_connection_sock.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c index 928654c34156..971f9db2c586 100644 --- a/net/ipv4/inet_connection_sock.c +++ b/net/ipv4/inet_connection_sock.c @@ -1108,7 +1108,7 @@ static void reqsk_timer_handler(struct timer_list *t) if (!inet_ehash_insert(req_to_sk(nreq), req_to_sk(oreq), NULL)) { /* delete timer */ - __inet_csk_reqsk_queue_drop(sk_listener, nreq, true); + __inet_csk_reqsk_queue_drop(sk_listener, nreq, false); goto no_ownership; } From 7eca3292cac7c26dad4c236f51ba225c39a0523f Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Wed, 6 May 2026 03:59:19 +0000 Subject: [PATCH 228/333] tcp: Fix imbalanced icsk_accept_queue count. When TCP socket migration happens in reqsk_timer_handler(), @sk_listener will be updated with the new listener. When we call __inet_csk_reqsk_queue_drop(), the listener must be the one stored in req->rsk_listener. The cited commit accidentally replaced oreq->rsk_listener with sk_listener, leading to imbalanced icsk_accept_queue count. Let's pass the correct listener to __inet_csk_reqsk_queue_drop(). Fixes: e8c526f2bdf1 ("tcp/dccp: Don't use timer_pending() in reqsk_queue_unlink().") Reported-by: Damiano Melotti Signed-off-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20260506035954.1563147-3-kuniyu@google.com Signed-off-by: Jakub Kicinski --- net/ipv4/inet_connection_sock.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c index 971f9db2c586..dbcd37dfdc15 100644 --- a/net/ipv4/inet_connection_sock.c +++ b/net/ipv4/inet_connection_sock.c @@ -1134,7 +1134,7 @@ static void reqsk_timer_handler(struct timer_list *t) } drop: - __inet_csk_reqsk_queue_drop(sk_listener, oreq, true); + __inet_csk_reqsk_queue_drop(oreq->rsk_listener, oreq, true); reqsk_put(oreq); } From e539acf9f9c2550452914fb85aeb8fda67dd762f Mon Sep 17 00:00:00 2001 From: "Maciej W. Rozycki" Date: Mon, 27 Apr 2026 11:23:17 +0100 Subject: [PATCH 229/333] MAINTAINERS: Add self for the 3c509 network driver It appears there's a need for a maintainer for the 3Com EtherLink III family of Ethernet network adapters. There is documentation available and the driver is very mature so the task ought to be of little hassle, so I think I should be able to squeeze in any issues to be addressed. Signed-off-by: Maciej W. Rozycki Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/alpine.DEB.2.21.2604271056460.28583@angie.orcam.me.uk Signed-off-by: Jakub Kicinski --- MAINTAINERS | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/MAINTAINERS b/MAINTAINERS index 0650fa014f24..1a485549ee96 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -68,6 +68,12 @@ Maintainers List first. When adding to this list, please keep the entries in alphabetical order. +3C509 NETWORK DRIVER +M: "Maciej W. Rozycki" +L: netdev@vger.kernel.org +S: Maintained +F: drivers/net/ethernet/3com/3c509.c + 3C59X NETWORK DRIVER M: Steffen Klassert L: netdev@vger.kernel.org From 7ce5556f255a680d80daa31b1cedecf7f89e2c22 Mon Sep 17 00:00:00 2001 From: Maoyi Xie Date: Wed, 6 May 2026 16:24:15 +0800 Subject: [PATCH 230/333] ipv6: flowlabel: take ip6_fl_lock across mem_check and fl_intern mem_check() in net/ipv6/ip6_flowlabel.c reads fl_size without holding ip6_fl_lock. fl_intern() takes the lock immediately afterwards. The two checks therefore race against concurrent fl_intern, ip6_fl_gc and ip6_fl_purge writers, which makes the mem_check budget check approximate. Move spin_lock_bh(&ip6_fl_lock) and the matching unlock from fl_intern() into its only caller ipv6_flowlabel_get(). The mem_check() call now runs under the same critical section as the fl_intern() insert, so the budget check is exact. With all writers and the read of fl_size under ip6_fl_lock, convert fl_size from atomic_t to plain int. The four sites that update or read fl_size are fl_intern (insert path), ip6_fl_gc (garbage collector, the !sched check and the per-entry decrement), ip6_fl_purge (per-netns purge), and mem_check (budget check), and all four now run under ip6_fl_lock. This is a prerequisite for adding a per-netns budget alongside fl_size. The follow-up patch adds netns_ipv6::flowlabel_count and folds it into mem_check(). Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Suggested-by: Willem de Bruijn Reviewed-by: Willem de Bruijn Signed-off-by: Maoyi Xie Link: https://patch.msgid.link/20260506082416.2259567-2-maoyixie.tju@gmail.com Signed-off-by: Jakub Kicinski --- net/ipv6/ip6_flowlabel.c | 34 +++++++++++++++++++++------------- 1 file changed, 21 insertions(+), 13 deletions(-) diff --git a/net/ipv6/ip6_flowlabel.c b/net/ipv6/ip6_flowlabel.c index c92f98c6f6ec..a8974643195a 100644 --- a/net/ipv6/ip6_flowlabel.c +++ b/net/ipv6/ip6_flowlabel.c @@ -40,7 +40,7 @@ #define FL_HASH_MASK 255 #define FL_HASH(l) (ntohl(l)&FL_HASH_MASK) -static atomic_t fl_size = ATOMIC_INIT(0); +static int fl_size; static struct ip6_flowlabel __rcu *fl_ht[FL_HASH_MASK+1]; static void ip6_fl_gc(struct timer_list *unused); @@ -163,7 +163,7 @@ static void ip6_fl_gc(struct timer_list *unused) if (time_after_eq(now, ttd)) { *flp = fl->next; fl_free(fl); - atomic_dec(&fl_size); + fl_size--; continue; } if (!sched || time_before(ttd, sched)) @@ -172,7 +172,7 @@ static void ip6_fl_gc(struct timer_list *unused) flp = &fl->next; } } - if (!sched && atomic_read(&fl_size)) + if (!sched && fl_size) sched = now + FL_MAX_LINGER; if (sched) { mod_timer(&ip6_fl_gc_timer, sched); @@ -196,7 +196,7 @@ static void __net_exit ip6_fl_purge(struct net *net) atomic_read(&fl->users) == 0) { *flp = fl->next; fl_free(fl); - atomic_dec(&fl_size); + fl_size--; continue; } flp = &fl->next; @@ -210,10 +210,10 @@ static struct ip6_flowlabel *fl_intern(struct net *net, { struct ip6_flowlabel *lfl; + lockdep_assert_held(&ip6_fl_lock); + fl->label = label & IPV6_FLOWLABEL_MASK; - rcu_read_lock(); - spin_lock_bh(&ip6_fl_lock); if (label == 0) { for (;;) { fl->label = htonl(get_random_u32())&IPV6_FLOWLABEL_MASK; @@ -235,8 +235,6 @@ static struct ip6_flowlabel *fl_intern(struct net *net, lfl = __fl_lookup(net, fl->label); if (lfl) { atomic_inc(&lfl->users); - spin_unlock_bh(&ip6_fl_lock); - rcu_read_unlock(); return lfl; } } @@ -244,9 +242,7 @@ static struct ip6_flowlabel *fl_intern(struct net *net, fl->lastuse = jiffies; fl->next = fl_ht[FL_HASH(fl->label)]; rcu_assign_pointer(fl_ht[FL_HASH(fl->label)], fl); - atomic_inc(&fl_size); - spin_unlock_bh(&ip6_fl_lock); - rcu_read_unlock(); + fl_size++; return NULL; } @@ -464,10 +460,14 @@ fl_create(struct net *net, struct sock *sk, struct in6_flowlabel_req *freq, static int mem_check(struct sock *sk) { - int room = FL_MAX_SIZE - atomic_read(&fl_size); + int room; struct ipv6_fl_socklist *sfl; int count = 0; + lockdep_assert_held(&ip6_fl_lock); + + room = FL_MAX_SIZE - fl_size; + if (room > FL_MAX_SIZE - FL_MAX_PER_SOCK) return 0; @@ -692,11 +692,19 @@ static int ipv6_flowlabel_get(struct sock *sk, struct in6_flowlabel_req *freq, if (!sfl1) goto done; + rcu_read_lock(); + spin_lock_bh(&ip6_fl_lock); err = mem_check(sk); + if (err == 0) + fl1 = fl_intern(net, fl, freq->flr_label); + else + fl1 = NULL; + spin_unlock_bh(&ip6_fl_lock); + rcu_read_unlock(); + if (err != 0) goto done; - fl1 = fl_intern(net, fl, freq->flr_label); if (fl1) goto recheck; From e68eadffb724b36ffd3d5619e0efcaf29ec2a175 Mon Sep 17 00:00:00 2001 From: Maoyi Xie Date: Wed, 6 May 2026 16:24:16 +0800 Subject: [PATCH 231/333] ipv6: flowlabel: enforce per-netns limit for unprivileged callers fl_size, fl_ht and ip6_fl_lock in net/ipv6/ip6_flowlabel.c are file scope and shared across netns. mem_check() reads fl_size to decide whether to deny non-CAP_NET_ADMIN callers. capable() runs against init_user_ns, so an unprivileged user in any non-init userns can push fl_size past FL_MAX_SIZE - FL_MAX_SIZE / 4 and starve every other unprivileged userns on the host. Add struct netns_ipv6::flowlabel_count, bumped and decremented next to fl_size in fl_intern, ip6_fl_gc and ip6_fl_purge. The new field fills the existing 4-byte hole after ipmr_seq, so struct netns_ipv6 stays the same size on 64-bit builds. Bump FL_MAX_SIZE from 4096 to 8192. It has been 4096 since the file was added. Machines and connection counts have grown. mem_check() folds an extra per-netns ceiling into the existing non-CAP_NET_ADMIN conditional. The ceiling is half of the total budget that unprivileged callers have ever been able to use, i.e. (FL_MAX_SIZE - FL_MAX_SIZE / 4) / 2 = 3072 entries. With FL_MAX_SIZE doubled, this preserves the original per-user reach of 3K (what an unprivileged caller could already obtain before this change), while forcing an attacker to spread allocations across at least two netns to exhaust the global non-CAP_NET_ADMIN budget. CAP_NET_ADMIN against init_user_ns still bypasses both caps. The previous patch took ip6_fl_lock across mem_check and fl_intern, so the new flowlabel_count read in mem_check and the new flowlabel_count++ in fl_intern run under the same critical section. flowlabel_count is therefore plain int, like fl_size. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Suggested-by: Willem de Bruijn Reviewed-by: Willem de Bruijn Cc: stable@vger.kernel.org # v5.15+ Signed-off-by: Maoyi Xie Link: https://patch.msgid.link/20260506082416.2259567-3-maoyixie.tju@gmail.com Signed-off-by: Jakub Kicinski --- include/net/netns/ipv6.h | 1 + net/ipv6/ip6_flowlabel.c | 14 +++++++++++--- 2 files changed, 12 insertions(+), 3 deletions(-) diff --git a/include/net/netns/ipv6.h b/include/net/netns/ipv6.h index 499e4288170f..875916d60bfe 100644 --- a/include/net/netns/ipv6.h +++ b/include/net/netns/ipv6.h @@ -119,6 +119,7 @@ struct netns_ipv6 { struct fib_notifier_ops *notifier_ops; struct fib_notifier_ops *ip6mr_notifier_ops; atomic_t ipmr_seq; + int flowlabel_count; struct { struct hlist_head head; spinlock_t lock; diff --git a/net/ipv6/ip6_flowlabel.c b/net/ipv6/ip6_flowlabel.c index a8974643195a..b1ccdf0dc646 100644 --- a/net/ipv6/ip6_flowlabel.c +++ b/net/ipv6/ip6_flowlabel.c @@ -36,7 +36,7 @@ /* FL hash table */ #define FL_MAX_PER_SOCK 32 -#define FL_MAX_SIZE 4096 +#define FL_MAX_SIZE 8192 #define FL_HASH_MASK 255 #define FL_HASH(l) (ntohl(l)&FL_HASH_MASK) @@ -162,8 +162,9 @@ static void ip6_fl_gc(struct timer_list *unused) ttd = fl->expires; if (time_after_eq(now, ttd)) { *flp = fl->next; - fl_free(fl); fl_size--; + fl->fl_net->ipv6.flowlabel_count--; + fl_free(fl); continue; } if (!sched || time_before(ttd, sched)) @@ -197,6 +198,7 @@ static void __net_exit ip6_fl_purge(struct net *net) *flp = fl->next; fl_free(fl); fl_size--; + net->ipv6.flowlabel_count--; continue; } flp = &fl->next; @@ -243,6 +245,7 @@ static struct ip6_flowlabel *fl_intern(struct net *net, fl->next = fl_ht[FL_HASH(fl->label)]; rcu_assign_pointer(fl_ht[FL_HASH(fl->label)], fl); fl_size++; + net->ipv6.flowlabel_count++; return NULL; } @@ -460,6 +463,9 @@ fl_create(struct net *net, struct sock *sk, struct in6_flowlabel_req *freq, static int mem_check(struct sock *sk) { + const int unpriv_total_limit = FL_MAX_SIZE - (FL_MAX_SIZE / 4); + const int unpriv_user_limit = unpriv_total_limit / 2; + struct net *net = sock_net(sk); int room; struct ipv6_fl_socklist *sfl; int count = 0; @@ -478,7 +484,9 @@ static int mem_check(struct sock *sk) if (room <= 0 || ((count >= FL_MAX_PER_SOCK || - (count > 0 && room < FL_MAX_SIZE/2) || room < FL_MAX_SIZE/4) && + (count > 0 && room < FL_MAX_SIZE / 2) || + room < FL_MAX_SIZE / 4 || + net->ipv6.flowlabel_count >= unpriv_user_limit) && !capable(CAP_NET_ADMIN))) return -ENOBUFS; From 58e2330bd45572a6e3d46ea94cf7a9641f43591a Mon Sep 17 00:00:00 2001 From: Dragos Tatulea Date: Wed, 6 May 2026 09:08:08 +0000 Subject: [PATCH 232/333] net: napi: Avoid gro timer misfiring at end of busypoll When in irq deferral mode (defer-hard-irqs > 0), a short enough gro-flush timeout can trigger before NAPI_STATE_SCHED is cleared if the last poll in busy_poll_stop() takes too long. This can have the effect of leaving the queue stuck with interrupts disabled and no timer armed which results in a tx timeout if there is no subsequent busypoll cycle. To prevent this, defer the gro-flush timer arm after the last poll. Fixes: 7fd3253a7de6 ("net: Introduce preferred busy-polling") Co-developed-by: Martin Karsten Signed-off-by: Martin Karsten Signed-off-by: Dragos Tatulea Reviewed-by: Tariq Toukan Reviewed-by: Cosmin Ratiu Reviewed-by: Joe Damato Link: https://patch.msgid.link/20260506090808.820559-2-dtatulea@nvidia.com Signed-off-by: Jakub Kicinski --- net/core/dev.c | 21 ++++++++++++--------- 1 file changed, 12 insertions(+), 9 deletions(-) diff --git a/net/core/dev.c b/net/core/dev.c index 8bfa8313ef62..0c6c270d9f7d 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -6862,9 +6862,9 @@ static void skb_defer_free_flush(void) #if defined(CONFIG_NET_RX_BUSY_POLL) -static void __busy_poll_stop(struct napi_struct *napi, bool skip_schedule) +static void __busy_poll_stop(struct napi_struct *napi, unsigned long timeout) { - if (!skip_schedule) { + if (!timeout) { gro_normal_list(&napi->gro); __napi_schedule(napi); return; @@ -6874,6 +6874,8 @@ static void __busy_poll_stop(struct napi_struct *napi, bool skip_schedule) gro_flush_normal(&napi->gro, HZ >= 1000); clear_bit(NAPI_STATE_SCHED, &napi->state); + hrtimer_start(&napi->timer, ns_to_ktime(timeout), + HRTIMER_MODE_REL_PINNED); } enum { @@ -6885,8 +6887,7 @@ static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, unsigned flags, u16 budget) { struct bpf_net_context __bpf_net_ctx, *bpf_net_ctx; - bool skip_schedule = false; - unsigned long timeout; + unsigned long timeout = 0; int rc; /* Busy polling means there is a high chance device driver hard irq @@ -6906,10 +6907,12 @@ static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, if (flags & NAPI_F_PREFER_BUSY_POLL) { napi->defer_hard_irqs_count = napi_get_defer_hard_irqs(napi); - timeout = napi_get_gro_flush_timeout(napi); - if (napi->defer_hard_irqs_count && timeout) { - hrtimer_start(&napi->timer, ns_to_ktime(timeout), HRTIMER_MODE_REL_PINNED); - skip_schedule = true; + if (napi->defer_hard_irqs_count) { + /* A short enough gro flush timeout and long enough + * poll can result in timer firing too early. + * Timer will be armed later if necessary. + */ + timeout = napi_get_gro_flush_timeout(napi); } } @@ -6924,7 +6927,7 @@ static void busy_poll_stop(struct napi_struct *napi, void *have_poll_lock, trace_napi_poll(napi, rc, budget); netpoll_poll_unlock(have_poll_lock); if (rc == budget) - __busy_poll_stop(napi, skip_schedule); + __busy_poll_stop(napi, timeout); bpf_net_ctx_clear(bpf_net_ctx); local_bh_enable(); } From 0a549298f452a83ae57e6582e6ca389357f9355d Mon Sep 17 00:00:00 2001 From: Nicolas Ferre Date: Thu, 7 May 2026 14:04:44 +0200 Subject: [PATCH 233/333] MAINTAINERS: change maintainers for macb Ethernet driver MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit I would like to hand over the macb maintenance to Théo, as I'm unable to keep up with the recent flow of patches for this driver. After speaking with Claudiu, he indicated that he is in the same position as me. To help with this work, Conor has agreed to act as a reviewer. I was given responsibility for this driver years ago, and I'm glad to see it continue with talented developers. Signed-off-by: Nicolas Ferre Acked-by: Claudiu Beznea Acked-by: Conor Dooley Link: https://patch.msgid.link/20260507120444.9733-1-nicolas.ferre@microchip.com Signed-off-by: Jakub Kicinski --- MAINTAINERS | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index 1a485549ee96..8a134cf6e9bb 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -4187,8 +4187,8 @@ F: include/uapi/linux/sonet.h F: net/atm/ ATMEL MACB ETHERNET DRIVER -M: Nicolas Ferre -M: Claudiu Beznea +M: Théo Lebrun +R: Conor Dooley S: Maintained F: drivers/net/ethernet/cadence/ From 4908f1395fb1b832ceec11584af649874a2732ea Mon Sep 17 00:00:00 2001 From: Quan Sun <2022090917019@std.uestc.edu.cn> Date: Thu, 7 May 2026 21:17:38 +0800 Subject: [PATCH 234/333] net: ethtool: fix NULL pointer dereference in phy_reply_size In phy_prepare_data(), several strings such as 'name', 'drvname', 'upstream_sfp_name', and 'downstream_sfp_name' are allocated using kstrdup(). However, these allocations were not checked for failure. If kstrdup() fails for 'name', it returns NULL while the function continues. This leads to a kernel NULL pointer dereference and panic later in phy_reply_size() when it unconditionally calls strlen() on the NULL pointer. While other strings like 'upstream_sfp_name' might be checked before access in certain code paths, failing to handle these allocations consistently can lead to incomplete data reporting or hidden bugs. Fix this by adding proper NULL checks for all kstrdup() calls in phy_prepare_data() and implement a centralized error handling path using goto labels to ensure all previously allocated resources are freed on failure. Fixes: 9dd2ad5e92b9 ("net: ethtool: phy: Convert the PHY_GET command to generic phy dump") Signed-off-by: Quan Sun <2022090917019@std.uestc.edu.cn> Reviewed-by: Maxime Chevallier Link: https://patch.msgid.link/20260507131738.1173835-1-2022090917019@std.uestc.edu.cn Signed-off-by: Jakub Kicinski --- net/ethtool/phy.c | 32 ++++++++++++++++++++++++++++++-- 1 file changed, 30 insertions(+), 2 deletions(-) diff --git a/net/ethtool/phy.c b/net/ethtool/phy.c index d4e6887055ab..f76d94d848d6 100644 --- a/net/ethtool/phy.c +++ b/net/ethtool/phy.c @@ -76,6 +76,7 @@ static int phy_prepare_data(const struct ethnl_req_info *req_info, struct nlattr **tb = info->attrs; struct phy_device_node *pdn; struct phy_device *phydev; + int ret; /* RTNL is held by the caller */ phydev = ethnl_req_get_phydev(req_info, tb, ETHTOOL_A_PHY_HEADER, @@ -88,8 +89,17 @@ static int phy_prepare_data(const struct ethnl_req_info *req_info, return -EOPNOTSUPP; rep_data->phyindex = phydev->phyindex; + rep_data->name = kstrdup(dev_name(&phydev->mdio.dev), GFP_KERNEL); + if (!rep_data->name) + return -ENOMEM; + rep_data->drvname = kstrdup(phydev->drv->name, GFP_KERNEL); + if (!rep_data->drvname) { + ret = -ENOMEM; + goto err_free_name; + } + rep_data->upstream_type = pdn->upstream_type; if (pdn->upstream_type == PHY_UPSTREAM_PHY) { @@ -97,15 +107,33 @@ static int phy_prepare_data(const struct ethnl_req_info *req_info, rep_data->upstream_index = upstream->phyindex; } - if (pdn->parent_sfp_bus) + if (pdn->parent_sfp_bus) { rep_data->upstream_sfp_name = kstrdup(sfp_get_name(pdn->parent_sfp_bus), GFP_KERNEL); + if (!rep_data->upstream_sfp_name) { + ret = -ENOMEM; + goto err_free_drvname; + } + } - if (phydev->sfp_bus) + if (phydev->sfp_bus) { rep_data->downstream_sfp_name = kstrdup(sfp_get_name(phydev->sfp_bus), GFP_KERNEL); + if (!rep_data->downstream_sfp_name) { + ret = -ENOMEM; + goto err_free_upstream_sfp; + } + } return 0; + +err_free_upstream_sfp: + kfree(rep_data->upstream_sfp_name); +err_free_drvname: + kfree(rep_data->drvname); +err_free_name: + kfree(rep_data->name); + return ret; } static int phy_fill_reply(struct sk_buff *skb, From f2ab4fd02777c4081be38c35f939e4dc529b8952 Mon Sep 17 00:00:00 2001 From: Ilya Maximets Date: Thu, 7 May 2026 14:04:26 +0200 Subject: [PATCH 235/333] net: nsh: fix incorrect header length macros NSH header length is a 6-bit field that encodes the total length of the header in 4-byte words. So the maximum length is 0b111111 * 4, which is 252 and not 256. The maximum context length is the same number minus the length of the base header (8), so 244. These macros are used to validate push_nsh() action in openvswitch. Miscalculation here doesn't cause any real issues. In the worst case the oversized context is truncated while building the header, so we'll construct and send a broken packet, which is not a big problem, as any receiver should validate the fields. No invalid memory accesses will happen during the header push. But we should fix the macros to reject the incorrect actions in the first place. Using previously defined values and calculating the length instead of defining numbers directly, so it's easier to understand where they come from and harder to make a mistake. Fixes: 1f0b7744c505 ("net: add NSH header structures and helpers") Signed-off-by: Ilya Maximets Reviewed-by: Aaron Conole Link: https://patch.msgid.link/20260507120434.2962505-1-i.maximets@ovn.org Signed-off-by: Jakub Kicinski --- include/net/nsh.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/include/net/nsh.h b/include/net/nsh.h index 16a751093896..15a26c590815 100644 --- a/include/net/nsh.h +++ b/include/net/nsh.h @@ -247,10 +247,10 @@ struct nshhdr { #define NSH_M_TYPE1_LEN 24 /* NSH header maximum Length. */ -#define NSH_HDR_MAX_LEN 256 +#define NSH_HDR_MAX_LEN ((NSH_LEN_MASK >> NSH_LEN_SHIFT) * 4) /* NSH context headers maximum Length. */ -#define NSH_CTX_HDRS_MAX_LEN 248 +#define NSH_CTX_HDRS_MAX_LEN (NSH_HDR_MAX_LEN - NSH_BASE_HDR_LEN) static inline struct nshhdr *nsh_hdr(struct sk_buff *skb) { From efda25ee84325385f859d10872590e90ce837243 Mon Sep 17 00:00:00 2001 From: Alice Ryhl Date: Wed, 6 May 2026 20:07:13 +0000 Subject: [PATCH 236/333] genetlink: free the skb on 'group >= family->n_mcgrps' These methods generally consume ownership of the provided skb, so even if an error path is encountered, the skb is freed. This is because the very first thing they do after some initial setup is to unconditionally consume the skb via consume_skb(skb). Any subsequent errors lead to the core netlink layer freeing the skb. However, there is one check that occurs before ownership is passed, which is the check for the group index. So if this error condition is encountered, then the skb is leaked. This error condition is generally considered a violation of the netlink API, so it's not expected to occur under normal circumstances. For the same reason, no callers check for this error condition, and no callers need to be adjusted. However, we should still follow the same ownership semantics of the rest of the function. Thus, free the skb in this codepath. Suggested-by: Andrew Lunn Suggested-by: Matthew Maurer Fixes: 2a94fe48f32c ("genetlink: make multicast groups const, prevent abuse") Link: https://lore.kernel.org/r/845b36ba-7b3a-41f2-acb2-b284f253e2ca@lunn.ch Signed-off-by: Alice Ryhl Link: https://patch.msgid.link/20260506-genlmsg-return-v2-1-a63ee2a055d6@google.com Signed-off-by: Jakub Kicinski --- include/net/genetlink.h | 4 +++- net/netlink/genetlink.c | 8 ++++++-- 2 files changed, 9 insertions(+), 3 deletions(-) diff --git a/include/net/genetlink.h b/include/net/genetlink.h index 7b84f2cef8b1..d70510ac31ab 100644 --- a/include/net/genetlink.h +++ b/include/net/genetlink.h @@ -489,8 +489,10 @@ genlmsg_multicast_netns_filtered(const struct genl_family *family, netlink_filter_fn filter, void *filter_data) { - if (WARN_ON_ONCE(group >= family->n_mcgrps)) + if (WARN_ON_ONCE(group >= family->n_mcgrps)) { + nlmsg_free(skb); return -EINVAL; + } group = family->mcgrp_offset + group; return nlmsg_multicast_filtered(net->genl_sock, skb, portid, group, flags, filter, filter_data); diff --git a/net/netlink/genetlink.c b/net/netlink/genetlink.c index d251d894afd4..0da39eaed255 100644 --- a/net/netlink/genetlink.c +++ b/net/netlink/genetlink.c @@ -1972,8 +1972,10 @@ int genlmsg_multicast_allns(const struct genl_family *family, struct sk_buff *skb, u32 portid, unsigned int group) { - if (WARN_ON_ONCE(group >= family->n_mcgrps)) + if (WARN_ON_ONCE(group >= family->n_mcgrps)) { + kfree_skb(skb); return -EINVAL; + } group = family->mcgrp_offset + group; return genlmsg_mcast(skb, portid, group); @@ -1986,8 +1988,10 @@ void genl_notify(const struct genl_family *family, struct sk_buff *skb, struct net *net = genl_info_net(info); struct sock *sk = net->genl_sock; - if (WARN_ON_ONCE(group >= family->n_mcgrps)) + if (WARN_ON_ONCE(group >= family->n_mcgrps)) { + kfree_skb(skb); return; + } group = family->mcgrp_offset + group; nlmsg_notify(sk, skb, info->snd_portid, group, From a77d5a069d959dc45f5f472d48cba37d8cba0f1c Mon Sep 17 00:00:00 2001 From: Mohsin Bashir Date: Wed, 6 May 2026 16:37:45 -0700 Subject: [PATCH 237/333] net: shaper: Reject reparenting of existing nodes When an existing node-scope shaper is moved to a different parent via the group operation, the framework fails to update the leaves count on both the old and new parent shapers. Only newly created nodes (handle.id == NET_SHAPER_ID_UNSPEC) trigger the parent leaves increment at line 1039. This causes the parent's leaves counter to diverge from the actual number of children in the xarray. When the node is later deleted, pre_del_node() allocates an array sized by the stale leaves count, but the xarray iteration finds more children than expected, hitting the WARN_ON_ONCE guard and returning -EINVAL. Rather than adding reparenting support with complex leaves count bookkeeping, reject group calls that attempt to change an existing node's parent. Updates to an existing node's rate or leaves under the same parent remain permitted. We expect that for any modification of the topology user should always create new groups and let the kernel garbage collect the leaf-less nodes. Fixes: 5d5d4700e75d ("net-shapers: implement NL group operation") Signed-off-by: Mohsin Bashir Link: https://patch.msgid.link/20260506233745.111895-1-mohsin.bashr@gmail.com Signed-off-by: Jakub Kicinski --- net/shaper/shaper.c | 30 +++++++++++++++++++++++------- 1 file changed, 23 insertions(+), 7 deletions(-) diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index 94bc9c7382ea..1069fa4eb9f6 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -964,15 +964,22 @@ static int __net_shaper_group(struct net_shaper_binding *binding, int i, ret; if (node->handle.scope == NET_SHAPER_SCOPE_NODE) { + struct net_shaper *cur = NULL; + new_node = node->handle.id == NET_SHAPER_ID_UNSPEC; - if (!new_node && !net_shaper_lookup(binding, &node->handle)) { - /* The related attribute is not available when - * reaching here from the delete() op. - */ - NL_SET_ERR_MSG_FMT(extack, "Node shaper %d:%d does not exists", - node->handle.scope, node->handle.id); - return -ENOENT; + if (!new_node) { + cur = net_shaper_lookup(binding, &node->handle); + if (!cur) { + /* The related attribute is not available + * when reaching here from the delete() op. + */ + NL_SET_ERR_MSG_FMT(extack, + "Node shaper %d:%d does not exist", + node->handle.scope, + node->handle.id); + return -ENOENT; + } } /* When unspecified, the node parent scope is inherited from @@ -986,6 +993,15 @@ static int __net_shaper_group(struct net_shaper_binding *binding, return ret; } + if (cur && net_shaper_handle_cmp(&cur->parent, + &node->parent)) { + NL_SET_ERR_MSG_FMT(extack, + "Cannot reparent node shaper %d:%d", + node->handle.scope, + node->handle.id); + return -EOPNOTSUPP; + } + } else { net_shaper_default_parent(&node->handle, &node->parent); } From 1619553b0a6ba7a966b17b0226f3acb9dd4d5380 Mon Sep 17 00:00:00 2001 From: Matt Vollrath Date: Wed, 6 May 2026 14:48:10 -0700 Subject: [PATCH 238/333] i40e: Cleanup PTP registration on probe failure Fix two conditions which would leak PTP registration on probe failure: 1. i40e_setup_pf_switch can encounter an error in i40e_setup_pf_filter_control, call i40e_ptp_init, then return non-zero, sending i40e_probe to err_vsis. 2. i40e_setup_misc_vector can return non-zero, sending i40e_probe to err_vsis. Both of these conditions have been present since PTP was introduced in this driver. Found with coccinelle. Fixes: beb0dff1251db ("i40e: enable PTP") Signed-off-by: Matt Vollrath Tested-by: Sunitha Mekala Reviewed-by: Aleksandr Loktionov Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-1-a5ea4dc837a9@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/i40e/i40e_main.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c index 028bd500603a..f06fcef644e5 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_main.c +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c @@ -16108,6 +16108,7 @@ static int i40e_probe(struct pci_dev *pdev, const struct pci_device_id *ent) /* Unwind what we've done if something failed in the setup */ err_vsis: set_bit(__I40E_DOWN, pf->state); + i40e_ptp_stop(pf); i40e_clear_interrupt_scheme(pf); kfree(pf->vsi); err_switch_setup: From 678b713ece1e853f11e670a84cb887c35e1381b7 Mon Sep 17 00:00:00 2001 From: Matt Vollrath Date: Wed, 6 May 2026 14:48:11 -0700 Subject: [PATCH 239/333] i40e: Cleanup PTP pins on probe failure PTP pin structs are allocated early in probe, but never cleaned up. Fix this by calling i40e_ptp_free_pins in the error path. To support this, i40e_ptp_free_pins is added to the header and pin_config is correctly nullified after being freed. This has been an issue since i40e_ptp_alloc_pins was introduced. Fixes: 1050713026a08 ("i40e: add support for PTP external synchronization clock") Reported-by: Kohei Enju Cc: stable@vger.kernel.org Signed-off-by: Matt Vollrath Reviewed-by: Paul Menzel Reviewed-by: Aleksandr Loktionov Reviewed-by: Kohei Enju Tested-by: Sunitha Mekala Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-2-a5ea4dc837a9@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/i40e/i40e.h | 1 + drivers/net/ethernet/intel/i40e/i40e_main.c | 1 + drivers/net/ethernet/intel/i40e/i40e_ptp.c | 3 ++- 3 files changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/i40e/i40e.h b/drivers/net/ethernet/intel/i40e/i40e.h index dcb50c2e1aa2..83e780919ac9 100644 --- a/drivers/net/ethernet/intel/i40e/i40e.h +++ b/drivers/net/ethernet/intel/i40e/i40e.h @@ -1318,6 +1318,7 @@ void i40e_ptp_restore_hw_time(struct i40e_pf *pf); void i40e_ptp_init(struct i40e_pf *pf); void i40e_ptp_stop(struct i40e_pf *pf); int i40e_ptp_alloc_pins(struct i40e_pf *pf); +void i40e_ptp_free_pins(struct i40e_pf *pf); int i40e_update_adq_vsi_queues(struct i40e_vsi *vsi, int vsi_offset); int i40e_is_vsi_uplink_mode_veb(struct i40e_vsi *vsi); int i40e_get_partition_bw_setting(struct i40e_pf *pf); diff --git a/drivers/net/ethernet/intel/i40e/i40e_main.c b/drivers/net/ethernet/intel/i40e/i40e_main.c index f06fcef644e5..6d4f9218dc68 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_main.c +++ b/drivers/net/ethernet/intel/i40e/i40e_main.c @@ -16112,6 +16112,7 @@ static int i40e_probe(struct pci_dev *pdev, const struct pci_device_id *ent) i40e_clear_interrupt_scheme(pf); kfree(pf->vsi); err_switch_setup: + i40e_ptp_free_pins(pf); i40e_reset_interrupt_capability(pf); timer_shutdown_sync(&pf->service_timer); err_mac_addr: diff --git a/drivers/net/ethernet/intel/i40e/i40e_ptp.c b/drivers/net/ethernet/intel/i40e/i40e_ptp.c index 404a716db8da..7d07c389bb23 100644 --- a/drivers/net/ethernet/intel/i40e/i40e_ptp.c +++ b/drivers/net/ethernet/intel/i40e/i40e_ptp.c @@ -940,12 +940,13 @@ int i40e_ptp_hwtstamp_get(struct net_device *netdev, * * Release memory allocated for PTP pins. **/ -static void i40e_ptp_free_pins(struct i40e_pf *pf) +void i40e_ptp_free_pins(struct i40e_pf *pf) { if (i40e_is_ptp_pin_dev(&pf->hw)) { kfree(pf->ptp_pins); kfree(pf->ptp_caps.pin_config); pf->ptp_pins = NULL; + pf->ptp_caps.pin_config = NULL; } } From da4f76b6a84ede14a71282ef841768299ead0221 Mon Sep 17 00:00:00 2001 From: Emil Tantilov Date: Wed, 6 May 2026 14:48:12 -0700 Subject: [PATCH 240/333] idpf: fix read_dev_clk_lock spinlock init in idpf_ptp_init() In idpf_ptp_init(), read_dev_clk_lock is initialized after ptp_schedule_worker() had already been called (and after idpf_ptp_settime64() could reach the lock). The PTP aux worker fires immediately upon scheduling and can call into idpf_ptp_read_src_clk_reg_direct(), which takes spin_lock(&ptp->read_dev_clk_lock) on an uninitialized lock, triggering the lockdep "non-static key" warning: [12973.796587] idpf 0000:83:00.0: Device HW Reset initiated [12974.094507] INFO: trying to register non-static key. ... [12974.097208] Call Trace: [12974.097213] [12974.097218] dump_stack_lvl+0x93/0xe0 [12974.097234] register_lock_class+0x4c4/0x4e0 [12974.097249] ? __lock_acquire+0x427/0x2290 [12974.097259] __lock_acquire+0x98/0x2290 [12974.097272] lock_acquire+0xc6/0x310 [12974.097281] ? idpf_ptp_read_src_clk_reg+0xb7/0x150 [idpf] [12974.097311] ? lockdep_hardirqs_on_prepare+0xde/0x190 [12974.097318] ? finish_task_switch.isra.0+0xd2/0x350 [12974.097330] ? __pfx_ptp_aux_kworker+0x10/0x10 [ptp] [12974.097343] _raw_spin_lock+0x30/0x40 [12974.097353] ? idpf_ptp_read_src_clk_reg+0xb7/0x150 [idpf] [12974.097373] idpf_ptp_read_src_clk_reg+0xb7/0x150 [idpf] [12974.097391] ? kthread_worker_fn+0x88/0x3d0 [12974.097404] ? kthread_worker_fn+0x4e/0x3d0 [12974.097411] idpf_ptp_update_cached_phctime+0x26/0x120 [idpf] [12974.097428] ? _raw_spin_unlock_irq+0x28/0x50 [12974.097436] idpf_ptp_do_aux_work+0x15/0x20 [idpf] [12974.097454] ptp_aux_kworker+0x20/0x40 [ptp] [12974.097464] kthread_worker_fn+0xd5/0x3d0 [12974.097474] ? __pfx_kthread_worker_fn+0x10/0x10 [12974.097482] kthread+0xf4/0x130 [12974.097489] ? __pfx_kthread+0x10/0x10 [12974.097498] ret_from_fork+0x32c/0x410 [12974.097512] ? __pfx_kthread+0x10/0x10 [12974.097519] ret_from_fork_asm+0x1a/0x30 [12974.097540] Move the call to spin_lock_init() up a bit to make sure read_dev_clk_lock is not touched before it's been initialized. Fixes: 5cb8805d2366 ("idpf: negotiate PTP capabilities and get PTP clock") Signed-off-by: Emil Tantilov Reviewed-by: Madhu Chittim Reviewed-by: Aleksandr Loktionov Reviewed-by: Simon Horman Tested-by: Samuel Salin Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-3-a5ea4dc837a9@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/idpf/idpf_ptp.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/intel/idpf/idpf_ptp.c b/drivers/net/ethernet/intel/idpf/idpf_ptp.c index eec91c4f0a75..4a51d2727547 100644 --- a/drivers/net/ethernet/intel/idpf/idpf_ptp.c +++ b/drivers/net/ethernet/intel/idpf/idpf_ptp.c @@ -952,6 +952,8 @@ int idpf_ptp_init(struct idpf_adapter *adapter) goto free_ptp; } + spin_lock_init(&adapter->ptp->read_dev_clk_lock); + err = idpf_ptp_create_clock(adapter); if (err) goto free_ptp; @@ -977,8 +979,6 @@ int idpf_ptp_init(struct idpf_adapter *adapter) goto remove_clock; } - spin_lock_init(&adapter->ptp->read_dev_clk_lock); - pci_dbg(adapter->pdev, "PTP init successful\n"); return 0; From 6c77b9510829a424d1b74409b7db9456e3522871 Mon Sep 17 00:00:00 2001 From: Greg Kroah-Hartman Date: Wed, 6 May 2026 14:48:13 -0700 Subject: [PATCH 241/333] idpf: fix double free and use-after-free in aux device error paths When auxiliary_device_add() fails in idpf_plug_vport_aux_dev() or idpf_plug_core_aux_dev(), the err_aux_dev_add label calls auxiliary_device_uninit() and falls through to err_aux_dev_init. The uninit call will trigger put_device(), which invokes the release callback (idpf_vport_adev_release / idpf_core_adev_release) that frees iadev. The fall-through then reads adev->id from the freed iadev for ida_free() and double-frees iadev with kfree(). Free the IDA slot and clear the back-pointer before uninit, while adev is still valid, then return immediately. Commit 65637c3a1811 ("idpf: fix UAF in RDMA core aux dev deinitialization") fixed the same use-after-free in the matching unplug path in this file but missed both probe error paths. Cc: Tony Nguyen Cc: Przemek Kitszel Cc: Andrew Lunn Cc: stable@kernel.org Fixes: be91128c579c ("idpf: implement RDMA vport auxiliary dev create, init, and destroy") Fixes: f4312e6bfa2a ("idpf: implement core RDMA auxiliary dev create, init, and destroy") Signed-off-by: Greg Kroah-Hartman Reviewed-by: Aleksandr Loktionov Reviewed-by: Paul Menzel Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-4-a5ea4dc837a9@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/idpf/idpf_idc.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/net/ethernet/intel/idpf/idpf_idc.c b/drivers/net/ethernet/intel/idpf/idpf_idc.c index 7e4f4ac92653..b7d6b08fc89e 100644 --- a/drivers/net/ethernet/intel/idpf/idpf_idc.c +++ b/drivers/net/ethernet/intel/idpf/idpf_idc.c @@ -90,7 +90,10 @@ static int idpf_plug_vport_aux_dev(struct iidc_rdma_core_dev_info *cdev_info, return 0; err_aux_dev_add: + ida_free(&idpf_idc_ida, adev->id); + vdev_info->adev = NULL; auxiliary_device_uninit(adev); + return ret; err_aux_dev_init: ida_free(&idpf_idc_ida, adev->id); err_ida_alloc: @@ -228,7 +231,10 @@ static int idpf_plug_core_aux_dev(struct iidc_rdma_core_dev_info *cdev_info) return 0; err_aux_dev_add: + ida_free(&idpf_idc_ida, adev->id); + cdev_info->adev = NULL; auxiliary_device_uninit(adev); + return ret; err_aux_dev_init: ida_free(&idpf_idc_ida, adev->id); err_ida_alloc: From b3cda96feb60d91fe88d52b974ff110dcfa91239 Mon Sep 17 00:00:00 2001 From: Marcin Szycik Date: Wed, 6 May 2026 14:48:14 -0700 Subject: [PATCH 242/333] ice: fix setting RSS VSI hash for E830 ice_set_rss_hfunc() performs a VSI update, in which it sets hashing function, leaving other VSI options unchanged. However, ::q_opt_flags is mistakenly set to the value of another field, instead of its original value, probably due to a typo. What happens next is hardware-dependent: On E810, only the first bit is meaningful (see ICE_AQ_VSI_Q_OPT_PE_FLTR_EN) and can potentially end up in a different state than before VSI update. On E830, some of the remaining bits are not reserved. Setting them to some unrelated values can cause the firmware to reject the update because of invalid settings, or worse - succeed. Reproducer: sudo ethtool -X $PF1 equal 8 Output in dmesg: Failed to configure RSS hash for VSI 6, error -5 Fixes: 352e9bf23813 ("ice: enable symmetric-xor RSS for Toeplitz hash function") Reviewed-by: Aleksandr Loktionov Reviewed-by: Przemek Kitszel Signed-off-by: Marcin Szycik Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-5-a5ea4dc837a9@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ice/ice_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/intel/ice/ice_main.c b/drivers/net/ethernet/intel/ice/ice_main.c index 1d1947a7fe11..c52c465280f7 100644 --- a/drivers/net/ethernet/intel/ice/ice_main.c +++ b/drivers/net/ethernet/intel/ice/ice_main.c @@ -8046,7 +8046,7 @@ int ice_set_rss_hfunc(struct ice_vsi *vsi, u8 hfunc) ctx->info.q_opt_rss |= FIELD_PREP(ICE_AQ_VSI_Q_OPT_RSS_HASH_M, hfunc); ctx->info.q_opt_tc = vsi->info.q_opt_tc; - ctx->info.q_opt_flags = vsi->info.q_opt_rss; + ctx->info.q_opt_flags = vsi->info.q_opt_flags; err = ice_update_vsi(hw, vsi->idx, ctx, NULL); if (err) { From 0ded1f36ba4021cba50513e80be6b6e173710168 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Wed, 6 May 2026 14:48:15 -0700 Subject: [PATCH 243/333] ice: fix locking in ice_dcb_rebuild() Move the mutex_lock() call up to prevent that DCB settings change after the first ice_query_port_ets() call. The second ice_query_port_ets() call in ice_dcb_rebuild() is already protected by pf->tc_mutex. This also fixes a bug in an error path, as before taking the first "goto dcb_error" in the function jumped over mutex_lock() to mutex_unlock(). This bug has been detected by the clang thread-safety analyzer. Cc: intel-wired-lan@lists.osuosl.org Fixes: 242b5e068b25 ("ice: Fix DCB rebuild after reset") Signed-off-by: Bart Van Assche Reviewed-by: Aleksandr Loktionov Reviewed-by: Przemek Kitszel Tested-by: Arpana Arland Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-6-a5ea4dc837a9@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ice/ice_dcb_lib.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/intel/ice/ice_dcb_lib.c b/drivers/net/ethernet/intel/ice/ice_dcb_lib.c index 16aa25535152..0bc6dd375687 100644 --- a/drivers/net/ethernet/intel/ice/ice_dcb_lib.c +++ b/drivers/net/ethernet/intel/ice/ice_dcb_lib.c @@ -537,14 +537,14 @@ void ice_dcb_rebuild(struct ice_pf *pf) struct ice_dcbx_cfg *err_cfg; int ret; + mutex_lock(&pf->tc_mutex); + ret = ice_query_port_ets(pf->hw.port_info, &buf, sizeof(buf), NULL); if (ret) { dev_err(dev, "Query Port ETS failed\n"); goto dcb_error; } - mutex_lock(&pf->tc_mutex); - if (!pf->hw.port_info->qos_cfg.is_sw_lldp) ice_cfg_etsrec_defaults(pf->hw.port_info); From cce709d8df6ba6d2a0a0dbf34acc2cdd9e23bd46 Mon Sep 17 00:00:00 2001 From: Ivan Vecera Date: Wed, 6 May 2026 14:48:16 -0700 Subject: [PATCH 244/333] ice: dpll: fix rclk pin state get for E810 The refactoring of ice_dpll_rclk_state_on_pin_get() to use ice_dpll_pin_get_parent_idx() omitted the base_rclk_idx adjustment that was correctly added in the ice_dpll_rclk_state_on_pin_set() path. This breaks E810 devices where base_rclk_idx is non-zero, causing the wrong hardware index to be used for pin state lookup and incorrect recovered clock state to be reported via the DPLL subsystem. E825C is unaffected as its base_rclk_idx is 0. While at it, add bounds check against ICE_DPLL_RCLK_NUM_MAX on hw_idx after the base_rclk_idx subtraction in both ice_dpll_rclk_state_on_pin_{get,set}() to prevent out-of-bounds access on the pin state array. Fixes: ad1df4f2d591 ("ice: dpll: Support E825-C SyncE and dynamic pin discovery") Signed-off-by: Ivan Vecera Reviewed-by: Aleksandr Loktionov Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-7-a5ea4dc837a9@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ice/ice_dpll.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/net/ethernet/intel/ice/ice_dpll.c b/drivers/net/ethernet/intel/ice/ice_dpll.c index 27b460926bac..892bc7c2e28b 100644 --- a/drivers/net/ethernet/intel/ice/ice_dpll.c +++ b/drivers/net/ethernet/intel/ice/ice_dpll.c @@ -2523,6 +2523,8 @@ ice_dpll_rclk_state_on_pin_set(const struct dpll_pin *pin, void *pin_priv, if (hw_idx < 0) goto unlock; hw_idx -= pf->dplls.base_rclk_idx; + if (hw_idx >= ICE_DPLL_RCLK_NUM_MAX) + goto unlock; if ((enable && p->state[hw_idx] == DPLL_PIN_STATE_CONNECTED) || (!enable && p->state[hw_idx] == DPLL_PIN_STATE_DISCONNECTED)) { @@ -2586,6 +2588,9 @@ ice_dpll_rclk_state_on_pin_get(const struct dpll_pin *pin, void *pin_priv, hw_idx = ice_dpll_pin_get_parent_idx(p, parent_pin); if (hw_idx < 0) goto unlock; + hw_idx -= pf->dplls.base_rclk_idx; + if (hw_idx >= ICE_DPLL_RCLK_NUM_MAX) + goto unlock; ret = ice_dpll_pin_state_update(pf, p, ICE_DPLL_PIN_TYPE_RCLK_INPUT, extack); From 30f1658fc5387384c7a60b9d15c79cb959512c1a Mon Sep 17 00:00:00 2001 From: Ivan Vecera Date: Wed, 6 May 2026 14:48:17 -0700 Subject: [PATCH 245/333] ice: dpll: fix misplaced header macros The CGU register definitions (ICE_CGU_R10, ICE_CGU_R11 and related field masks) were placed after the #endif of the _ICE_DPLL_H_ include guard, leaving them unprotected. Move them inside the guard. Fixes: ad1df4f2d591 ("ice: dpll: Support E825-C SyncE and dynamic pin discovery") Signed-off-by: Ivan Vecera Reviewed-by: Aleksandr Loktionov Signed-off-by: Jacob Keller Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-8-a5ea4dc837a9@intel.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/intel/ice/ice_dpll.h | 32 +++++++++++------------ 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/drivers/net/ethernet/intel/ice/ice_dpll.h b/drivers/net/ethernet/intel/ice/ice_dpll.h index ae42cdea0ee1..8678575359b9 100644 --- a/drivers/net/ethernet/intel/ice/ice_dpll.h +++ b/drivers/net/ethernet/intel/ice/ice_dpll.h @@ -8,6 +8,22 @@ #define ICE_DPLL_RCLK_NUM_MAX 4 +#define ICE_CGU_R10 0x28 +#define ICE_CGU_R10_SYNCE_CLKO_SEL GENMASK(8, 5) +#define ICE_CGU_R10_SYNCE_CLKODIV_M1 GENMASK(13, 9) +#define ICE_CGU_R10_SYNCE_CLKODIV_LOAD BIT(14) +#define ICE_CGU_R10_SYNCE_DCK_RST BIT(15) +#define ICE_CGU_R10_SYNCE_ETHCLKO_SEL GENMASK(18, 16) +#define ICE_CGU_R10_SYNCE_ETHDIV_M1 GENMASK(23, 19) +#define ICE_CGU_R10_SYNCE_ETHDIV_LOAD BIT(24) +#define ICE_CGU_R10_SYNCE_DCK2_RST BIT(25) +#define ICE_CGU_R10_SYNCE_S_REF_CLK GENMASK(31, 27) + +#define ICE_CGU_R11 0x2C +#define ICE_CGU_R11_SYNCE_S_BYP_CLK GENMASK(6, 1) + +#define ICE_CGU_BYPASS_MUX_OFFSET_E825C 3 + /** * enum ice_dpll_pin_sw - enumerate ice software pin indices: * @ICE_DPLL_PIN_SW_1_IDX: index of first SW pin @@ -157,19 +173,3 @@ static inline void ice_dpll_deinit(struct ice_pf *pf) { } #endif #endif - -#define ICE_CGU_R10 0x28 -#define ICE_CGU_R10_SYNCE_CLKO_SEL GENMASK(8, 5) -#define ICE_CGU_R10_SYNCE_CLKODIV_M1 GENMASK(13, 9) -#define ICE_CGU_R10_SYNCE_CLKODIV_LOAD BIT(14) -#define ICE_CGU_R10_SYNCE_DCK_RST BIT(15) -#define ICE_CGU_R10_SYNCE_ETHCLKO_SEL GENMASK(18, 16) -#define ICE_CGU_R10_SYNCE_ETHDIV_M1 GENMASK(23, 19) -#define ICE_CGU_R10_SYNCE_ETHDIV_LOAD BIT(24) -#define ICE_CGU_R10_SYNCE_DCK2_RST BIT(25) -#define ICE_CGU_R10_SYNCE_S_REF_CLK GENMASK(31, 27) - -#define ICE_CGU_R11 0x2C -#define ICE_CGU_R11_SYNCE_S_BYP_CLK GENMASK(6, 1) - -#define ICE_CGU_BYPASS_MUX_OFFSET_E825C 3 From c4f3d6eb1fcf6cd9ce4644f604d5aad1ce594dfc Mon Sep 17 00:00:00 2001 From: Myeonghun Pak Date: Wed, 6 May 2026 21:43:11 +0900 Subject: [PATCH 246/333] net: lan966x: avoid unregistering netdev on register failure lan966x_probe_port() stores the newly allocated net_device in the port before calling register_netdev(). If register_netdev() fails, the probe error path calls lan966x_cleanup_ports(), which sees port->dev and calls unregister_netdev() for a device that was never registered. Destroy the phylink instance created for this port and clear port->dev before returning the registration error. The common cleanup path now skips ports without port->dev before reaching the registered netdev cleanup, so it only handles ports that reached the registered-netdev lifetime. This also avoids treating an uninitialized FDMA netdev and the failed port as a NULL == NULL match in the common cleanup path. Fixes: d28d6d2e37d1 ("net: lan966x: add port module support") Co-developed-by: Ijae Kim Signed-off-by: Ijae Kim Signed-off-by: Myeonghun Pak Link: https://patch.msgid.link/20260506124331.31945-1-mhun512@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/microchip/lan966x/lan966x_main.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/microchip/lan966x/lan966x_main.c b/drivers/net/ethernet/microchip/lan966x/lan966x_main.c index 47752d3fde0b..1179a6e127c5 100644 --- a/drivers/net/ethernet/microchip/lan966x/lan966x_main.c +++ b/drivers/net/ethernet/microchip/lan966x/lan966x_main.c @@ -749,11 +749,10 @@ static void lan966x_cleanup_ports(struct lan966x *lan966x) for (p = 0; p < lan966x->num_phys_ports; p++) { port = lan966x->ports[p]; - if (!port) + if (!port || !port->dev) continue; - if (port->dev) - unregister_netdev(port->dev); + unregister_netdev(port->dev); lan966x_xdp_port_deinit(port); if (lan966x->fdma && lan966x->fdma_ndev == port->dev) @@ -873,6 +872,9 @@ static int lan966x_probe_port(struct lan966x *lan966x, u32 p, err = register_netdev(dev); if (err) { dev_err(lan966x->dev, "register_netdev failed\n"); + phylink_destroy(phylink); + port->phylink = NULL; + port->dev = NULL; return err; } From 6635fa84403c3a59455b66007c019a7cc632db30 Mon Sep 17 00:00:00 2001 From: Shitalkumar Gandhi Date: Thu, 7 May 2026 01:28:13 +0530 Subject: [PATCH 247/333] net: ti: icssm-prueth: fix eth_ports_node leak in probe The error path on of_property_read_u32() failure inside icssm_prueth_probe() returns without putting eth_ports_node, which was acquired before the for_each_child_of_node() loop. Drop it before returning. Fixes: 511f6c1ae093 ("net: ti: icssm-prueth: Adds ICSSM Ethernet driver") Signed-off-by: Shitalkumar Gandhi Link: https://patch.msgid.link/20260506195813.641610-1-shitalkumar.gandhi@cambiumnetworks.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/ti/icssm/icssm_prueth.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/ti/icssm/icssm_prueth.c b/drivers/net/ethernet/ti/icssm/icssm_prueth.c index 53bbd9290904..b7e94244355a 100644 --- a/drivers/net/ethernet/ti/icssm/icssm_prueth.c +++ b/drivers/net/ethernet/ti/icssm/icssm_prueth.c @@ -1825,6 +1825,7 @@ static int icssm_prueth_probe(struct platform_device *pdev) dev_err(dev, "%pOF error reading port_id %d\n", eth_node, ret); of_node_put(eth_node); + of_node_put(eth_ports_node); return ret; } From abb5f36771cc4c05899b34000829a787572a8817 Mon Sep 17 00:00:00 2001 From: Ben Morris Date: Thu, 7 May 2026 17:14:55 -0700 Subject: [PATCH 248/333] sctp: revalidate list cursor after sctp_sendmsg_to_asoc() in SCTP_SENDALL The SCTP_SENDALL path in sctp_sendmsg() iterates ep->asocs with list_for_each_entry_safe(), which caches the next entry in @tmp before the loop body runs. The body calls sctp_sendmsg_to_asoc(), which may drop the socket lock inside sctp_wait_for_sndbuf(). While the lock is dropped, another thread can SCTP_SOCKOPT_PEELOFF the association cached in @tmp, migrating it to a new endpoint via sctp_sock_migrate() (list_del_init() + list_add_tail() to newep->asocs), and optionally close the new socket which frees the association via kfree_rcu(). The cached @tmp can also be freed by a network ABORT for that association, processed in softirq while the lock is dropped. sctp_wait_for_sndbuf() revalidates @asoc (the current entry) on re-lock via the "sk != asoc->base.sk" and "asoc->base.dead" checks, but nothing revalidates @tmp. After a successful return, the iterator advances to the stale @tmp, yielding either a use-after-free (if the peeled socket was closed) or a list-walk onto the new endpoint's list head (type confusion of &newep->asocs as a struct sctp_association *). Both are reachable from CapEff=0; the type-confusion path gives controlled indirect call via the outqueue.sched->init_sid pointer. Fix by re-deriving @tmp from @asoc after sctp_sendmsg_to_asoc() returns. @asoc is known to still be on ep->asocs at that point: the only callers that list_del an association from ep->asocs are sctp_association_free() (which sets asoc->base.dead) and sctp_assoc_migrate() (which changes asoc->base.sk), and sctp_wait_for_sndbuf() checks both under the lock before any successful return; a tripped check propagates as err < 0 and the loop bails before the re-derive. The SCTP_ABORT path in sctp_sendmsg_check_sflags() returns 0 and the loop hits 'continue' before sctp_sendmsg_to_asoc() is ever called, so the @tmp cached by list_for_each_entry_safe() still covers the lock-held free that ba59fb027307 ("sctp: walk the list of asoc safely") was added for. Fixes: 4910280503f3 ("sctp: add support for snd flag SCTP_SENDALL process in sendmsg") Cc: stable@vger.kernel.org Signed-off-by: Ben Morris Acked-by: Xin Long Link: https://patch.msgid.link/20260508001455.3137-1-joycathacker@gmail.com Signed-off-by: Jakub Kicinski --- net/sctp/socket.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/net/sctp/socket.c b/net/sctp/socket.c index 58d0d9747f0b..1d2568bb6bc2 100644 --- a/net/sctp/socket.c +++ b/net/sctp/socket.c @@ -1986,6 +1986,15 @@ static int sctp_sendmsg(struct sock *sk, struct msghdr *msg, size_t msg_len) goto out_unlock; iov_iter_revert(&msg->msg_iter, err); + + /* sctp_sendmsg_to_asoc() may have released the socket + * lock (sctp_wait_for_sndbuf), during which other + * associations on ep->asocs could have been peeled + * off or freed. @asoc itself is revalidated by the + * base.dead and base.sk checks in sctp_wait_for_sndbuf, + * so re-derive the cached cursor from it. + */ + tmp = list_next_entry(asoc, asocs); } goto out_unlock; From 496c0c4c53bbe1bad97e82cd12103df61a6e459d Mon Sep 17 00:00:00 2001 From: Holger Brunck Date: Thu, 7 May 2026 17:53:32 +0200 Subject: [PATCH 249/333] net: wan: fsl_ucc_hdlc: free tx_skbuff in uhdlc_memclean When the device is removed all allocated resources should be freed. In uhdlc_memclean the netdev transmit queue was already stopped. But at this point we may have pending skb in the transmit queue which must be freed. Therefore iterate over the tx_skbuff pointers and free all pending skb. The issue was discovered by sashiko. Tested on a ls1043a board running HDLC in bus mode on kernel 6.12. https: //sashiko.dev/#/patchset/20260429114208.941011-1-holger.brunck%40hitachienergy.com Fixes: c19b6d246a35 ("drivers/net: support hdlc function for QE-UCC") Signed-off-by: Holger Brunck Link: https://patch.msgid.link/20260507155332.3452319-1-holger.brunck@hitachienergy.com Signed-off-by: Jakub Kicinski --- drivers/net/wan/fsl_ucc_hdlc.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/net/wan/fsl_ucc_hdlc.c b/drivers/net/wan/fsl_ucc_hdlc.c index 15bfb78381d4..809f21fb93f5 100644 --- a/drivers/net/wan/fsl_ucc_hdlc.c +++ b/drivers/net/wan/fsl_ucc_hdlc.c @@ -740,6 +740,8 @@ static int uhdlc_open(struct net_device *dev) static void uhdlc_memclean(struct ucc_hdlc_private *priv) { + int i; + qe_muram_free(ioread16be(&priv->ucc_pram->riptr)); qe_muram_free(ioread16be(&priv->ucc_pram->tiptr)); @@ -770,6 +772,11 @@ static void uhdlc_memclean(struct ucc_hdlc_private *priv) kfree(priv->rx_skbuff); priv->rx_skbuff = NULL; + for (i = 0; i < TX_BD_RING_LEN; i++) { + dev_kfree_skb(priv->tx_skbuff[i]); + priv->tx_skbuff[i] = NULL; + } + kfree(priv->tx_skbuff); priv->tx_skbuff = NULL; From 512809bb8a370d071f66fc53abe67368e171dec5 Mon Sep 17 00:00:00 2001 From: Paul Chaignon Date: Thu, 7 May 2026 20:22:06 +0200 Subject: [PATCH 250/333] bpf: Don't run arg-tracking analysis twice on main subprog Because subprog 0, the main subprog, is considered a global function, we end up running the arg-tracking dataflow analysis twice on it. That results in slightly longer verification but mostly in more verbose verifier logs. This patch fixes it by keeping only the iteration over global subprogs. When running over all of Cilium's programs with BPF_LOG_LEVEL2, this reduces verbosity by ~20% on average. Fixes: bf0c571f7feb6 ("bpf: introduce forward arg-tracking dataflow analysis") Signed-off-by: Paul Chaignon Acked-by: Eduard Zingerman Link: https://lore.kernel.org/r/e4d7b53d4963ef520541a782f5fc8108a168877c.1778176504.git.paul.chaignon@gmail.com Signed-off-by: Alexei Starovoitov --- kernel/bpf/liveness.c | 25 +++++++------------------ 1 file changed, 7 insertions(+), 18 deletions(-) diff --git a/kernel/bpf/liveness.c b/kernel/bpf/liveness.c index 332e6e003f27..58197d73b120 100644 --- a/kernel/bpf/liveness.c +++ b/kernel/bpf/liveness.c @@ -1914,26 +1914,15 @@ int bpf_compute_subprog_arg_access(struct bpf_verifier_env *env) return -ENOMEM; } - instance = call_instance(env, NULL, 0, 0); - if (IS_ERR(instance)) { - err = PTR_ERR(instance); - goto out; - } - err = analyze_subprog(env, NULL, info, instance, callsites); - if (err) - goto out; - /* - * Subprogs and callbacks that don't receive FP-derived arguments - * cannot access ancestor stack frames, so they were skipped during - * the recursive walk above. Async callbacks (timer, workqueue) are - * also not reachable from the main program's call graph. Analyze - * all unvisited subprogs as independent roots at depth 0. + * Analyze every subprog in reverse topological order (callers + * before callees) so that each subprog is analyzed before its + * callees, allowing the recursive walk inside analyze_subprog() + * to naturally reach callees that receive FP-derived args. * - * Use reverse topological order (callers before callees) so that - * each subprog is analyzed before its callees, allowing the - * recursive walk inside analyze_subprog() to naturally - * reach nested callees that also lack FP-derived args. + * Subprogs and callbacks that don't receive FP-derived arguments + * cannot access ancestor stack frames are analyzed independently. + * Async callbacks (timer, workqueue) are handled the same way. */ for (k = env->subprog_cnt - 1; k >= 0; k--) { int sub = env->subprog_topo_order[k]; From bf6d507f7e3c65751d52fd8caf1ea4e003922624 Mon Sep 17 00:00:00 2001 From: Linpu Yu Date: Fri, 8 May 2026 22:43:43 +0800 Subject: [PATCH 251/333] xskmap: reject TX-only AF_XDP sockets XSKMAP entries are used as redirect targets for incoming XDP frames. A TX-only AF_XDP socket lacks an Rx ring and cannot handle redirected traffic, but xsk_map_update_elem() currently allows such sockets to be inserted into the map. Redirecting packets to such a socket on the veth generic-XDP path causes a kernel crash in xsk_generic_rcv(). This became possible after xsk_is_setup_for_bpf_map() was removed from the XSKMAP update path, which allowed bound TX-only sockets to be inserted into the map. Reject TX-only sockets during XSKMAP updates to avoid the crash. They remain fully operational for pure Tx purposes outside XSKMAP. Fixes: 968be23ceaca ("xsk: Fix possible segfault at xskmap entry insertion") Reported-by: Juefei Pu Reported-by: Yuan Tan Reported-by: Xin Liu Signed-off-by: Yifan Wu Signed-off-by: Linpu Yu Reviewed-by: Jason Xing Link: https://lore.kernel.org/r/20260508144344.694-1-linpu5433@gmail.com Signed-off-by: Alexei Starovoitov --- net/xdp/xskmap.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/xdp/xskmap.c b/net/xdp/xskmap.c index afa457506274..3bff346308d0 100644 --- a/net/xdp/xskmap.c +++ b/net/xdp/xskmap.c @@ -184,6 +184,10 @@ static long xsk_map_update_elem(struct bpf_map *map, void *key, void *value, } xs = (struct xdp_sock *)sock->sk; + if (!READ_ONCE(xs->rx)) { + sockfd_put(sock); + return -ENOBUFS; + } map_entry = &m->xsk_map[i]; node = xsk_map_node_alloc(m, map_entry); From 3ac1a467e37683f602221e243fa3c59b0de81165 Mon Sep 17 00:00:00 2001 From: Junyoung Jang Date: Mon, 27 Apr 2026 02:25:05 +0900 Subject: [PATCH 252/333] bpf: Fix off-by-one boundary validation in arena direct-value access BPF_MAP_TYPE_ARENA accepts BPF_PSEUDO_MAP_VALUE offsets at exactly the end of the arena mapping (off == arena_size). The boundary check in arena_map_direct_value_addr() uses `>` instead of `>=`, which incorrectly allows a one-past-end pointer to be accepted. Change the condition to `>=` to correctly reject offsets that fall outside the valid arena user_vm range. Fixes: 317460317a02 ("bpf: Introduce bpf_arena.") Signed-off-by: Junyoung Jang Reviewed-by: Emil Tsalapatis Link: https://lore.kernel.org/r/20260426172505.1947915-1-graypanda.inzag@gmail.com Signed-off-by: Alexei Starovoitov --- kernel/bpf/arena.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/bpf/arena.c b/kernel/bpf/arena.c index 802656c6fd3c..49a8f7b1beef 100644 --- a/kernel/bpf/arena.c +++ b/kernel/bpf/arena.c @@ -511,7 +511,7 @@ static int arena_map_direct_value_addr(const struct bpf_map *map, u64 *imm, u32 { struct bpf_arena *arena = container_of(map, struct bpf_arena, map); - if ((u64)off > arena->user_vm_end - arena->user_vm_start) + if ((u64)off >= arena->user_vm_end - arena->user_vm_start) return -ERANGE; *imm = (unsigned long)arena->user_vm_start; return 0; From aa54b1d27fe0c2b78e664a34fd0fdf7cd1960d71 Mon Sep 17 00:00:00 2001 From: Hyunwoo Kim Date: Fri, 8 May 2026 17:53:09 +0900 Subject: [PATCH 253/333] rxrpc: Also unshare DATA/RESPONSE packets when paged frags are present The DATA-packet handler in rxrpc_input_call_event() and the RESPONSE handler in rxrpc_verify_response() copy the skb to a linear one before calling into the security ops only when skb_cloned() is true. An skb that is not cloned but still carries externally-owned paged fragments (e.g. SKBFL_SHARED_FRAG set by splice() into a UDP socket via __ip_append_data, or a chained skb_has_frag_list()) falls through to the in-place decryption path, which binds the frag pages directly into the AEAD/skcipher SGL via skb_to_sgvec(). Extend the gate to also unshare when skb_has_frag_list() or skb_has_shared_frag() is true. This catches the splice-loopback vector and other externally-shared frag sources while preserving the zero-copy fast path for skbs whose frags are kernel-private (e.g. NIC page_pool RX, GRO). The OOM/trace handling already in place is reused. Fixes: d0d5c0cd1e71 ("rxrpc: Use skb_unshare() rather than skb_cow_data()") Cc: stable@vger.kernel.org Signed-off-by: Hyunwoo Kim Reviewed-by: Jiayuan Chen Acked-by: David Howells Signed-off-by: Linus Torvalds --- net/rxrpc/call_event.c | 4 +++- net/rxrpc/conn_event.c | 3 ++- 2 files changed, 5 insertions(+), 2 deletions(-) diff --git a/net/rxrpc/call_event.c b/net/rxrpc/call_event.c index fdd683261226..2b19b252225e 100644 --- a/net/rxrpc/call_event.c +++ b/net/rxrpc/call_event.c @@ -334,7 +334,9 @@ bool rxrpc_input_call_event(struct rxrpc_call *call) if (sp->hdr.type == RXRPC_PACKET_TYPE_DATA && sp->hdr.securityIndex != 0 && - skb_cloned(skb)) { + (skb_cloned(skb) || + skb_has_frag_list(skb) || + skb_has_shared_frag(skb))) { /* Unshare the packet so that it can be * modified by in-place decryption. */ diff --git a/net/rxrpc/conn_event.c b/net/rxrpc/conn_event.c index a2130d25aaa9..442414d90ba1 100644 --- a/net/rxrpc/conn_event.c +++ b/net/rxrpc/conn_event.c @@ -245,7 +245,8 @@ static int rxrpc_verify_response(struct rxrpc_connection *conn, { int ret; - if (skb_cloned(skb)) { + if (skb_cloned(skb) || skb_has_frag_list(skb) || + skb_has_shared_frag(skb)) { /* Copy the packet if shared so that we can do in-place * decryption. */ From 46e9b0224475abc739612ef72c35b7c90211a0c1 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Fri, 8 May 2026 13:41:13 -0700 Subject: [PATCH 254/333] tools/ynl: add missing uapi header deps in Makefile.deps ethtool.h includes linux/typelimits.h which is a relatively new header not yet shipped in most distro kernel-header packages. Without the explicit entry, the build silently falls through to -idirafter. dev_energymodel.h is a new YNL family whose uapi header is not in system paths at all and was missing a CFLAGS entry entirely. Signed-off-by: Stanislav Fomichev Link: https://patch.msgid.link/20260508204114.205896-2-sdf@fomichev.me Signed-off-by: Jakub Kicinski --- tools/net/ynl/Makefile.deps | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/tools/net/ynl/Makefile.deps b/tools/net/ynl/Makefile.deps index 08205f9fc525..cc53b2f21c44 100644 --- a/tools/net/ynl/Makefile.deps +++ b/tools/net/ynl/Makefile.deps @@ -15,9 +15,11 @@ UAPI_PATH:=../../../../include/uapi/ get_hdr_inc=-D$(1) -include $(UAPI_PATH)/linux/$(2) get_hdr_inc2=-D$(1) -D$(2) -include $(UAPI_PATH)/linux/$(3) +CFLAGS_dev-energymodel:=$(call get_hdr_inc,_LINUX_DEV_ENERGYMODEL_H,dev_energymodel.h) CFLAGS_devlink:=$(call get_hdr_inc,_LINUX_DEVLINK_H_,devlink.h) CFLAGS_dpll:=$(call get_hdr_inc,_LINUX_DPLL_H,dpll.h) -CFLAGS_ethtool:=$(call get_hdr_inc,_LINUX_ETHTOOL_H,ethtool.h) \ +CFLAGS_ethtool:=$(call get_hdr_inc,_LINUX_TYPELIMITS_H,typelimits.h) \ + $(call get_hdr_inc,_LINUX_ETHTOOL_H,ethtool.h) \ $(call get_hdr_inc,_LINUX_ETHTOOL_NETLINK_H_,ethtool_netlink.h) \ $(call get_hdr_inc,_LINUX_ETHTOOL_NETLINK_GENERATED_H,ethtool_netlink_generated.h) CFLAGS_handshake:=$(call get_hdr_inc,_LINUX_HANDSHAKE_H,handshake.h) From e42c755582f0960e684298762f0ab927b3778376 Mon Sep 17 00:00:00 2001 From: Arthur Kiyanovski Date: Fri, 8 May 2026 06:21:21 +0000 Subject: [PATCH 255/333] net: ena: PHC: Fix potential use-after-free in get_timestamp Move the phc->active check and resp pointer assignment to after acquiring the spinlock. Previously, phc->active was checked without holding the lock, and resp was cached from ena_dev->phc.virt_addr before the lock was acquired. If ena_com_phc_destroy() runs between the lockless active check and the lock acquisition, it sets active=false, releases the lock, frees the DMA memory, and sets virt_addr=NULL. The get_timestamp path would then read a NULL virt_addr and dereference it. With both the active check and the pointer read under the lock, destroy cannot free the memory while get_timestamp is using it. Fixes: e0ea34158ee8 ("net: ena: Add PHC support in the ENA driver") Cc: stable@vger.kernel.org Signed-off-by: Arthur Kiyanovski Reviewed-by: Vadim Fedorenko Link: https://patch.msgid.link/20260508062126.7273-1-akiyano@amazon.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/amazon/ena/ena_com.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/amazon/ena/ena_com.c b/drivers/net/ethernet/amazon/ena/ena_com.c index e67b592e5697..8c86789d867a 100644 --- a/drivers/net/ethernet/amazon/ena/ena_com.c +++ b/drivers/net/ethernet/amazon/ena/ena_com.c @@ -1782,20 +1782,23 @@ void ena_com_phc_destroy(struct ena_com_dev *ena_dev) int ena_com_phc_get_timestamp(struct ena_com_dev *ena_dev, u64 *timestamp) { - volatile struct ena_admin_phc_resp *resp = ena_dev->phc.virt_addr; const ktime_t zero_system_time = ktime_set(0, 0); struct ena_com_phc_info *phc = &ena_dev->phc; + volatile struct ena_admin_phc_resp *resp; ktime_t expire_time; ktime_t block_time; unsigned long flags = 0; int ret = 0; + spin_lock_irqsave(&phc->lock, flags); + if (!phc->active) { + spin_unlock_irqrestore(&phc->lock, flags); netdev_err(ena_dev->net_device, "PHC feature is not active in the device\n"); return -EOPNOTSUPP; } - spin_lock_irqsave(&phc->lock, flags); + resp = ena_dev->phc.virt_addr; /* Check if PHC is in blocked state */ if (unlikely(ktime_compare(phc->system_time, zero_system_time))) { From a450063ef86b9967234ca1f896c0d77400c74f11 Mon Sep 17 00:00:00 2001 From: Shitalkumar Gandhi Date: Thu, 7 May 2026 19:50:24 +0530 Subject: [PATCH 256/333] net: xgene: fix mdio_np leak in xgene_mdiobus_register() The for_each_child_of_node() loop captures mdio_np via break, holding the refcount. of_mdiobus_register() does not consume the reference, so it leaks on success. Put it after registration. Fixes: e6ad767305eb ("drivers: net: Add APM X-Gene SoC ethernet driver support.") Signed-off-by: Shitalkumar Gandhi Link: https://patch.msgid.link/20260507142024.811543-1-shitalkumar.gandhi@cambiumnetworks.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/apm/xgene/xgene_enet_hw.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/apm/xgene/xgene_enet_hw.c b/drivers/net/ethernet/apm/xgene/xgene_enet_hw.c index b854b6b42d77..2926e1e59941 100644 --- a/drivers/net/ethernet/apm/xgene/xgene_enet_hw.c +++ b/drivers/net/ethernet/apm/xgene/xgene_enet_hw.c @@ -910,7 +910,9 @@ static int xgene_mdiobus_register(struct xgene_enet_pdata *pdata, return -ENXIO; } - return of_mdiobus_register(mdio, mdio_np); + ret = of_mdiobus_register(mdio, mdio_np); + of_node_put(mdio_np); + return ret; } /* Mask out all PHYs from auto probing. */ From 6947bea4b79115f50138882512f85fa9c93b2827 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 10 May 2026 10:08:16 -1000 Subject: [PATCH 257/333] sched_ext: Cleanups in preparation for the SCX_TASK_INIT_BEGIN/DEAD work Cleanups in preparation for the state-machine work that follows: - Convert three sub-sched call sites that open-code rcu_assign_pointer(p->scx.sched, ...) to scx_set_task_sched(). - Move scx_get_task_state()/scx_set_task_state() above the SCX task iter section so scx_task_iter_next_locked() can use them without a forward declaration. No functional change. Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- kernel/sched/ext.c | 76 +++++++++++++++++++++++----------------------- 1 file changed, 38 insertions(+), 38 deletions(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index df305712a2d4..10c6e0261f11 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -711,6 +711,41 @@ struct bpf_iter_scx_dsq { } __attribute__((aligned(8))); +static u32 scx_get_task_state(const struct task_struct *p) +{ + return p->scx.flags & SCX_TASK_STATE_MASK; +} + +static void scx_set_task_state(struct task_struct *p, u32 state) +{ + u32 prev_state = scx_get_task_state(p); + bool warn = false; + + switch (state) { + case SCX_TASK_NONE: + break; + case SCX_TASK_INIT: + warn = prev_state != SCX_TASK_NONE; + break; + case SCX_TASK_READY: + warn = prev_state == SCX_TASK_NONE; + break; + case SCX_TASK_ENABLED: + warn = prev_state != SCX_TASK_READY; + break; + default: + WARN_ONCE(1, "sched_ext: Invalid task state %d -> %d for %s[%d]", + prev_state, state, p->comm, p->pid); + return; + } + + WARN_ONCE(warn, "sched_ext: Invalid task state transition 0x%x -> 0x%x for %s[%d]", + prev_state, state, p->comm, p->pid); + + p->scx.flags &= ~SCX_TASK_STATE_MASK; + p->scx.flags |= state; +} + /* * SCX task iterator. */ @@ -3499,41 +3534,6 @@ static struct cgroup *tg_cgrp(struct task_group *tg) #endif /* CONFIG_EXT_GROUP_SCHED */ -static u32 scx_get_task_state(const struct task_struct *p) -{ - return p->scx.flags & SCX_TASK_STATE_MASK; -} - -static void scx_set_task_state(struct task_struct *p, u32 state) -{ - u32 prev_state = scx_get_task_state(p); - bool warn = false; - - switch (state) { - case SCX_TASK_NONE: - break; - case SCX_TASK_INIT: - warn = prev_state != SCX_TASK_NONE; - break; - case SCX_TASK_READY: - warn = prev_state == SCX_TASK_NONE; - break; - case SCX_TASK_ENABLED: - warn = prev_state != SCX_TASK_READY; - break; - default: - WARN_ONCE(1, "sched_ext: Invalid task state %d -> %d for %s[%d]", - prev_state, state, p->comm, p->pid); - return; - } - - WARN_ONCE(warn, "sched_ext: Invalid task state transition 0x%x -> 0x%x for %s[%d]", - prev_state, state, p->comm, p->pid); - - p->scx.flags &= ~SCX_TASK_STATE_MASK; - p->scx.flags |= state; -} - static int __scx_init_task(struct scx_sched *sch, struct task_struct *p, bool fork) { int ret; @@ -5696,7 +5696,7 @@ static void scx_fail_parent(struct scx_sched *sch, scoped_guard (sched_change, p, DEQUEUE_SAVE | DEQUEUE_MOVE) { scx_disable_and_exit_task(sch, p); - rcu_assign_pointer(p->scx.sched, parent); + scx_set_task_sched(p, parent); } } scx_task_iter_stop(&sti); @@ -5783,7 +5783,7 @@ static void scx_sub_disable(struct scx_sched *sch) */ scx_disable_and_exit_task(sch, p); scx_set_task_state(p, SCX_TASK_INIT); - rcu_assign_pointer(p->scx.sched, parent); + scx_set_task_sched(p, parent); scx_set_task_state(p, SCX_TASK_READY); scx_enable_task(parent, p); } @@ -7212,7 +7212,7 @@ static void scx_sub_enable_workfn(struct kthread_work *work) * $p is now only initialized for @sch and READY, which * is what we want. Assign it to @sch and enable. */ - rcu_assign_pointer(p->scx.sched, sch); + scx_set_task_sched(p, sch); scx_enable_task(sch, p); p->scx.flags &= ~SCX_TASK_SUB_INIT; From 938dd9ab2bd7df0a7e58ce4249794156be9530b4 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 10 May 2026 10:08:16 -1000 Subject: [PATCH 258/333] sched_ext: Inline scx_init_task() and move RESET_RUNNABLE_AT into scx_set_task_state() Prepare for the SCX_TASK_INIT_BEGIN/DEAD work that follows by collapsing the scx_init_task() helper. Move the SCX_TASK_RESET_RUNNABLE_AT setting into scx_set_task_state() on the INIT transition (it was set unconditionally at every INIT site through the scx_init_task() helper), inline scx_init_task() into scx_fork() and scx_root_enable_workfn(), and drop the helper. As a side effect, scx_sub_disable() migration sequence now also sets RESET_RUNNABLE_AT (it previously wrote INIT directly without going through scx_init_task()). The flag triggers a runnable_at reset on the next set_task_runnable(), which is harmless on a task that has just been moved between scheds. On root-enable, p->scx.flags is written without the task's rq lock. The task isn't visible to scx yet, and a follow-up patch restores the lock-held write. v2: Note p->scx.flags rq-lock relaxation on root-enable path. (Andrea) Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- kernel/sched/ext.c | 31 +++++++++---------------------- 1 file changed, 9 insertions(+), 22 deletions(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 10c6e0261f11..81841277a54f 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -726,6 +726,7 @@ static void scx_set_task_state(struct task_struct *p, u32 state) break; case SCX_TASK_INIT: warn = prev_state != SCX_TASK_NONE; + p->scx.flags |= SCX_TASK_RESET_RUNNABLE_AT; break; case SCX_TASK_READY: warn = prev_state == SCX_TASK_NONE; @@ -3585,22 +3586,6 @@ static int __scx_init_task(struct scx_sched *sch, struct task_struct *p, bool fo return 0; } -static int scx_init_task(struct scx_sched *sch, struct task_struct *p, bool fork) -{ - int ret; - - ret = __scx_init_task(sch, p, fork); - if (!ret) { - /* - * While @p's rq is not locked. @p is not visible to the rest of - * SCX yet and it's safe to update the flags and state. - */ - p->scx.flags |= SCX_TASK_RESET_RUNNABLE_AT; - scx_set_task_state(p, SCX_TASK_INIT); - } - return ret; -} - static void __scx_enable_task(struct scx_sched *sch, struct task_struct *p) { struct rq *rq = task_rq(p); @@ -3763,10 +3748,11 @@ int scx_fork(struct task_struct *p, struct kernel_clone_args *kargs) #else struct scx_sched *sch = scx_root; #endif - ret = scx_init_task(sch, p, true); - if (!ret) - scx_set_task_sched(p, sch); - return ret; + ret = __scx_init_task(sch, p, true); + if (unlikely(ret)) + return ret; + scx_set_task_state(p, SCX_TASK_INIT); + scx_set_task_sched(p, sch); } return 0; @@ -6897,8 +6883,8 @@ static void scx_root_enable_workfn(struct kthread_work *work) scx_task_iter_unlock(&sti); - ret = scx_init_task(sch, p, false); - if (ret) { + ret = __scx_init_task(sch, p, false); + if (unlikely(ret)) { put_task_struct(p); scx_task_iter_stop(&sti); scx_error(sch, "ops.init_task() failed (%d) for %s[%d]", @@ -6906,6 +6892,7 @@ static void scx_root_enable_workfn(struct kthread_work *work) goto err_disable_unlock_all; } + scx_set_task_state(p, SCX_TASK_INIT); scx_set_task_sched(p, sch); scx_set_task_state(p, SCX_TASK_READY); From cceb8fa9cb2cf98e31d81ecf6353b6ba5ac57744 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 10 May 2026 10:08:16 -1000 Subject: [PATCH 259/333] sched_ext: Replace SCX_TASK_OFF_TASKS flag with SCX_TASK_DEAD state SCX_TASK_OFF_TASKS marked tasks already through sched_ext_dead() so cgroup task iteration would skip them. This can be expressed better with a task state. Replace the flag with SCX_TASK_DEAD. scx_disable_and_exit_task() resets state to NONE on its way out, so sched_ext_dead() now sets DEAD after the wrapper returns. The validation matrix grows NONE -> DEAD, warns on DEAD -> NONE, and tightens READY's predecessor to INIT or ENABLED so the new DEAD value cannot silently transition to READY. Prepares for the following enable vs dead race fix. Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- include/linux/sched/ext.h | 9 +++++---- kernel/sched/ext.c | 17 +++++++++++------ 2 files changed, 16 insertions(+), 10 deletions(-) diff --git a/include/linux/sched/ext.h b/include/linux/sched/ext.h index adb9a4de068a..9f1a326ad03e 100644 --- a/include/linux/sched/ext.h +++ b/include/linux/sched/ext.h @@ -101,24 +101,25 @@ enum scx_ent_flags { SCX_TASK_DEQD_FOR_SLEEP = 1 << 3, /* last dequeue was for SLEEP */ SCX_TASK_SUB_INIT = 1 << 4, /* task being initialized for a sub sched */ SCX_TASK_IMMED = 1 << 5, /* task is on local DSQ with %SCX_ENQ_IMMED */ - SCX_TASK_OFF_TASKS = 1 << 6, /* removed from scx_tasks by sched_ext_dead() */ /* - * Bits 8 and 9 are used to carry task state: + * Bits 8 to 10 are used to carry task state: * * NONE ops.init_task() not called yet * INIT ops.init_task() succeeded, but task can be cancelled * READY fully initialized, but not in sched_ext * ENABLED fully initialized and in sched_ext + * DEAD terminal state set by sched_ext_dead() */ - SCX_TASK_STATE_SHIFT = 8, /* bits 8 and 9 are used to carry task state */ - SCX_TASK_STATE_BITS = 2, + SCX_TASK_STATE_SHIFT = 8, + SCX_TASK_STATE_BITS = 3, SCX_TASK_STATE_MASK = ((1 << SCX_TASK_STATE_BITS) - 1) << SCX_TASK_STATE_SHIFT, SCX_TASK_NONE = 0 << SCX_TASK_STATE_SHIFT, SCX_TASK_INIT = 1 << SCX_TASK_STATE_SHIFT, SCX_TASK_READY = 2 << SCX_TASK_STATE_SHIFT, SCX_TASK_ENABLED = 3 << SCX_TASK_STATE_SHIFT, + SCX_TASK_DEAD = 4 << SCX_TASK_STATE_SHIFT, /* * Bits 12 and 13 are used to carry reenqueue reason. In addition to diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 81841277a54f..2fc4a12711f9 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -723,17 +723,22 @@ static void scx_set_task_state(struct task_struct *p, u32 state) switch (state) { case SCX_TASK_NONE: + warn = prev_state == SCX_TASK_DEAD; break; case SCX_TASK_INIT: warn = prev_state != SCX_TASK_NONE; p->scx.flags |= SCX_TASK_RESET_RUNNABLE_AT; break; case SCX_TASK_READY: - warn = prev_state == SCX_TASK_NONE; + warn = !(prev_state == SCX_TASK_INIT || + prev_state == SCX_TASK_ENABLED); break; case SCX_TASK_ENABLED: warn = prev_state != SCX_TASK_READY; break; + case SCX_TASK_DEAD: + warn = prev_state != SCX_TASK_NONE; + break; default: WARN_ONCE(1, "sched_ext: Invalid task state %d -> %d for %s[%d]", prev_state, state, p->comm, p->pid); @@ -972,11 +977,11 @@ static struct task_struct *scx_task_iter_next_locked(struct scx_task_iter *iter) /* * cgroup_task_dead() removes the dead tasks from cset->tasks * after sched_ext_dead() and cgroup iteration may see tasks - * which already finished sched_ext_dead(). %SCX_TASK_OFF_TASKS - * is set by sched_ext_dead() under @p's rq lock. Test it to + * which already finished sched_ext_dead(). %SCX_TASK_DEAD is + * set by sched_ext_dead() under @p's rq lock. Test it to * avoid visiting tasks which are already dead from SCX POV. */ - if (p->scx.flags & SCX_TASK_OFF_TASKS) { + if (scx_get_task_state(p) == SCX_TASK_DEAD) { __scx_task_iter_rq_unlock(iter); continue; } @@ -3847,7 +3852,7 @@ void sched_ext_dead(struct task_struct *p) * @p is off scx_tasks and wholly ours. scx_root_enable()'s READY -> * ENABLED transitions can't race us. Disable ops for @p. * - * %SCX_TASK_OFF_TASKS synchronizes against cgroup task iteration - see + * %SCX_TASK_DEAD synchronizes against cgroup task iteration - see * scx_task_iter_next_locked(). NONE tasks need no marking: cgroup * iteration is only used from sub-sched paths, which require root * enabled. Root enable transitions every live task to at least READY. @@ -3858,7 +3863,7 @@ void sched_ext_dead(struct task_struct *p) rq = task_rq_lock(p, &rf); scx_disable_and_exit_task(scx_task_sched(p), p); - p->scx.flags |= SCX_TASK_OFF_TASKS; + scx_set_task_state(p, SCX_TASK_DEAD); task_rq_unlock(rq, p, &rf); } } From c941d7391f258d5d06e0f7e962a52f99a547a83e Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 10 May 2026 10:08:16 -1000 Subject: [PATCH 260/333] sched_ext: Close root-enable vs sched_ext_dead() race with SCX_TASK_INIT_BEGIN scx_root_enable_workfn() drops the iter rq lock for ops.init_task() and a TASK_DEAD @p can fall through sched_ext_dead() in that window. The race hits when sched_ext_dead() observes SCX_TASK_INIT (the intermediate state before @p->scx.sched is published) and dereferences NULL via SCX_HAS_OP(NULL, exit_task), or observes SCX_TASK_NONE during the unlocked init window and skips cleanup so exit_task() never runs. Add SCX_TASK_INIT_BEGIN. The enable path writes NONE -> INIT_BEGIN under the iter rq lock, then takes the rq lock again after init to walk INIT_BEGIN -> INIT -> READY. sched_ext_dead() that wins the rq-lock race observes INIT_BEGIN and sets DEAD without calling into ops; the post-init recheck unwinds via scx_sub_init_cancel_task(). scx_fork() runs single-threaded against sched_ext_dead() (the task is not on scx_tasks until scx_post_fork() adds it) so its INIT_BEGIN -> INIT walk needs no rq-lock pairing; it rolls back to NONE on ops.init_task() failure. The validation matrix grows the INIT_BEGIN row and the INIT_BEGIN -> DEAD edge; INIT now requires INIT_BEGIN as the predecessor. scx_sub_disable()'s migration writes INIT_BEGIN as a synthetic predecessor to satisfy the tightened verification. The sub-sched paths still race with sched_ext_dead() during the unlocked init window. This will be fixed by the next patch. Reported-by: zhidao su Link: https://lore.kernel.org/all/20260429133155.3825247-1-suzhidao@xiaomi.com/ Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- include/linux/sched/ext.h | 10 ++++--- kernel/sched/ext.c | 56 ++++++++++++++++++++++++++++++++++----- 2 files changed, 55 insertions(+), 11 deletions(-) diff --git a/include/linux/sched/ext.h b/include/linux/sched/ext.h index 9f1a326ad03e..2129e18ada58 100644 --- a/include/linux/sched/ext.h +++ b/include/linux/sched/ext.h @@ -106,6 +106,7 @@ enum scx_ent_flags { * Bits 8 to 10 are used to carry task state: * * NONE ops.init_task() not called yet + * INIT_BEGIN ops.init_task() in flight; see sched_ext_dead() * INIT ops.init_task() succeeded, but task can be cancelled * READY fully initialized, but not in sched_ext * ENABLED fully initialized and in sched_ext @@ -116,10 +117,11 @@ enum scx_ent_flags { SCX_TASK_STATE_MASK = ((1 << SCX_TASK_STATE_BITS) - 1) << SCX_TASK_STATE_SHIFT, SCX_TASK_NONE = 0 << SCX_TASK_STATE_SHIFT, - SCX_TASK_INIT = 1 << SCX_TASK_STATE_SHIFT, - SCX_TASK_READY = 2 << SCX_TASK_STATE_SHIFT, - SCX_TASK_ENABLED = 3 << SCX_TASK_STATE_SHIFT, - SCX_TASK_DEAD = 4 << SCX_TASK_STATE_SHIFT, + SCX_TASK_INIT_BEGIN = 1 << SCX_TASK_STATE_SHIFT, + SCX_TASK_INIT = 2 << SCX_TASK_STATE_SHIFT, + SCX_TASK_READY = 3 << SCX_TASK_STATE_SHIFT, + SCX_TASK_ENABLED = 4 << SCX_TASK_STATE_SHIFT, + SCX_TASK_DEAD = 5 << SCX_TASK_STATE_SHIFT, /* * Bits 12 and 13 are used to carry reenqueue reason. In addition to diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 2fc4a12711f9..29fa9ffe7c7b 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -725,8 +725,11 @@ static void scx_set_task_state(struct task_struct *p, u32 state) case SCX_TASK_NONE: warn = prev_state == SCX_TASK_DEAD; break; - case SCX_TASK_INIT: + case SCX_TASK_INIT_BEGIN: warn = prev_state != SCX_TASK_NONE; + break; + case SCX_TASK_INIT: + warn = prev_state != SCX_TASK_INIT_BEGIN; p->scx.flags |= SCX_TASK_RESET_RUNNABLE_AT; break; case SCX_TASK_READY: @@ -737,7 +740,8 @@ static void scx_set_task_state(struct task_struct *p, u32 state) warn = prev_state != SCX_TASK_READY; break; case SCX_TASK_DEAD: - warn = prev_state != SCX_TASK_NONE; + warn = !(prev_state == SCX_TASK_NONE || + prev_state == SCX_TASK_INIT_BEGIN); break; default: WARN_ONCE(1, "sched_ext: Invalid task state %d -> %d for %s[%d]", @@ -3753,9 +3757,12 @@ int scx_fork(struct task_struct *p, struct kernel_clone_args *kargs) #else struct scx_sched *sch = scx_root; #endif + scx_set_task_state(p, SCX_TASK_INIT_BEGIN); ret = __scx_init_task(sch, p, true); - if (unlikely(ret)) + if (unlikely(ret)) { + scx_set_task_state(p, SCX_TASK_NONE); return ret; + } scx_set_task_state(p, SCX_TASK_INIT); scx_set_task_sched(p, sch); } @@ -3856,13 +3863,18 @@ void sched_ext_dead(struct task_struct *p) * scx_task_iter_next_locked(). NONE tasks need no marking: cgroup * iteration is only used from sub-sched paths, which require root * enabled. Root enable transitions every live task to at least READY. + * + * %INIT_BEGIN means ops.init_task() is running for @p. Don't call + * into ops; transition to %DEAD so the post-init recheck unwinds + * via scx_sub_init_cancel_task(). */ if (scx_get_task_state(p) != SCX_TASK_NONE) { struct rq_flags rf; struct rq *rq; rq = task_rq_lock(p, &rf); - scx_disable_and_exit_task(scx_task_sched(p), p); + if (scx_get_task_state(p) != SCX_TASK_INIT_BEGIN) + scx_disable_and_exit_task(scx_task_sched(p), p); scx_set_task_state(p, SCX_TASK_DEAD); task_rq_unlock(rq, p, &rf); } @@ -5773,6 +5785,7 @@ static void scx_sub_disable(struct scx_sched *sch) * $p having already been initialized, and then enable. */ scx_disable_and_exit_task(sch, p); + scx_set_task_state(p, SCX_TASK_INIT_BEGIN); scx_set_task_state(p, SCX_TASK_INIT); scx_set_task_sched(p, parent); scx_set_task_state(p, SCX_TASK_READY); @@ -6878,6 +6891,9 @@ static void scx_root_enable_workfn(struct kthread_work *work) scx_task_iter_start(&sti, NULL); while ((p = scx_task_iter_next_locked(&sti))) { + struct rq_flags rf; + struct rq *rq; + /* * @p may already be dead, have lost all its usages counts and * be waiting for RCU grace period before being freed. @p can't @@ -6886,10 +6902,26 @@ static void scx_root_enable_workfn(struct kthread_work *work) if (!tryget_task_struct(p)) continue; + /* + * Set %INIT_BEGIN under the iter's rq lock so that a concurrent + * sched_ext_dead() does not call ops.exit_task() on @p while + * ops.init_task() is running. If sched_ext_dead() runs before + * this store, it has already removed @p from scx_tasks and the + * iter won't visit @p; if it runs after, it observes + * %INIT_BEGIN and transitions to %DEAD without calling ops, + * leaving the post-init recheck below to unwind. + */ + scx_set_task_state(p, SCX_TASK_INIT_BEGIN); scx_task_iter_unlock(&sti); ret = __scx_init_task(sch, p, false); + + rq = task_rq_lock(p, &rf); + if (unlikely(ret)) { + if (scx_get_task_state(p) != SCX_TASK_DEAD) + scx_set_task_state(p, SCX_TASK_NONE); + task_rq_unlock(rq, p, &rf); put_task_struct(p); scx_task_iter_stop(&sti); scx_error(sch, "ops.init_task() failed (%d) for %s[%d]", @@ -6897,10 +6929,20 @@ static void scx_root_enable_workfn(struct kthread_work *work) goto err_disable_unlock_all; } - scx_set_task_state(p, SCX_TASK_INIT); - scx_set_task_sched(p, sch); - scx_set_task_state(p, SCX_TASK_READY); + if (scx_get_task_state(p) == SCX_TASK_DEAD) { + /* + * sched_ext_dead() observed %INIT_BEGIN and set %DEAD. + * ops.exit_task() is owed to the sched __scx_init_task() + * ran against; call it now. + */ + scx_sub_init_cancel_task(sch, p); + } else { + scx_set_task_state(p, SCX_TASK_INIT); + scx_set_task_sched(p, sch); + scx_set_task_state(p, SCX_TASK_READY); + } + task_rq_unlock(rq, p, &rf); put_task_struct(p); } scx_task_iter_stop(&sti); From cd6aab736702f981ac4d128e04a4e33105ea797d Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 10 May 2026 10:08:16 -1000 Subject: [PATCH 261/333] sched_ext: Close sub-sched init race with post-init DEAD recheck scx_sub_enable_workfn()'s init pass and scx_sub_disable() migration both drop the rq lock to call __scx_init_task() against the other sched. A TASK_DEAD @p can fall through sched_ext_dead() in that window. sched_ext_dead() runs ops.exit_task() on the sched @p was attached to, not on the sched whose init just completed, so the new allocation leaks. Reuse the DEAD signal set by sched_ext_dead(). After __scx_init_task() returns, take task_rq_lock(p) and check for DEAD; on hit, call scx_sub_init_cancel_task() against the sub sched the init ran for and drop @p; on miss, proceed as before. Reported-by: zhidao su Link: https://lore.kernel.org/all/20260429133155.3825247-1-suzhidao@xiaomi.com/ Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- kernel/sched/ext.c | 32 +++++++++++++++++++++++++++++++- 1 file changed, 31 insertions(+), 1 deletion(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 29fa9ffe7c7b..6fbe3160eccd 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -5777,6 +5777,21 @@ static void scx_sub_disable(struct scx_sched *sch) } rq = task_rq_lock(p, &rf); + + if (scx_get_task_state(p) == SCX_TASK_DEAD) { + /* + * sched_ext_dead() raced us between __scx_init_task() + * and this rq lock and ran exit_task() on @sch (the + * sched @p was on at that point), not on $parent. + * $parent's just-completed init is owed an exit_task() + * and we issue it here. + */ + scx_sub_init_cancel_task(parent, p); + task_rq_unlock(rq, p, &rf); + put_task_struct(p); + continue; + } + scoped_guard (sched_change, p, DEQUEUE_SAVE | DEQUEUE_MOVE) { /* * $p is initialized for $parent and still attached to @@ -5791,8 +5806,8 @@ static void scx_sub_disable(struct scx_sched *sch) scx_set_task_state(p, SCX_TASK_READY); scx_enable_task(parent, p); } - task_rq_unlock(rq, p, &rf); + task_rq_unlock(rq, p, &rf); put_task_struct(p); } scx_task_iter_stop(&sti); @@ -7212,6 +7227,21 @@ static void scx_sub_enable_workfn(struct kthread_work *work) goto abort; rq = task_rq_lock(p, &rf); + + if (scx_get_task_state(p) == SCX_TASK_DEAD) { + /* + * sched_ext_dead() raced us between __scx_init_task() + * and this rq lock and ran exit_task() on $parent (the + * sched @p was on at that point), not on @sch. @sch's + * just-completed init is owed an exit_task() and we + * issue it here. + */ + scx_sub_init_cancel_task(sch, p); + task_rq_unlock(rq, p, &rf); + put_task_struct(p); + continue; + } + p->scx.flags |= SCX_TASK_SUB_INIT; task_rq_unlock(rq, p, &rf); From d3e73a0808ddfb91ac36cd548643cbbeb00ad4db Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Sun, 10 May 2026 10:08:16 -1000 Subject: [PATCH 262/333] sched_ext: Handle SCX_TASK_NONE in disable/switched_from paths scx_fail_parent() leaves cgroup tasks at (state=NONE, sched=parent, sched_class=ext) until the parent itself is torn down by the scx_error() it raised. When the later root_disable iterates them, two paths trip on NONE. scx_disable_and_exit_task() re-enters the wrapper at NONE: the inner switch returns early but the trailing scx_set_task_sched(p, NULL) clobbers the parent sched left by scx_fail_parent(), and scx_set_task_state(p, NONE) wastes a write on an already-NONE task. switched_from_scx() then calls scx_disable_task(), which WARNs on non-ENABLED state and writes state=READY, producing a NONE -> READY transition the validation matrix rejects. Treat NONE as "nothing to do" in both paths. Add a NONE early-return at the top of scx_disable_and_exit_task() and a parallel NONE check in switched_from_scx() next to task_dead_and_done(). Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- kernel/sched/ext.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 6fbe3160eccd..4efe0099f79a 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -3703,6 +3703,15 @@ static void scx_sub_init_cancel_task(struct scx_sched *sch, struct task_struct * static void scx_disable_and_exit_task(struct scx_sched *sch, struct task_struct *p) { + /* + * %NONE means @p is already detached at the SCX level (e.g. handed + * back to the parent by scx_fail_parent() with no init to undo). + * Skip to avoid clobbering scx_task_sched() and writing %NONE again + * on a state that's already %NONE. + */ + if (scx_get_task_state(p) == SCX_TASK_NONE) + return; + __scx_disable_and_exit_task(sch, p); /* @@ -3921,6 +3930,16 @@ static void switched_from_scx(struct rq *rq, struct task_struct *p) if (task_dead_and_done(p)) return; + /* + * %NONE means SCX is no longer tracking @p at the task level (e.g. + * scx_fail_parent() handed @p back to the parent at NONE pending the + * parent's own teardown). There is nothing to disable; calling + * scx_disable_task() would WARN on the non-%ENABLED state and trigger a + * NONE -> READY validation failure. + */ + if (scx_get_task_state(p) == SCX_TASK_NONE) + return; + scx_disable_task(scx_task_sched(p), p); } From 5d6919055dec134de3c40167a490f33c74c12581 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 10 May 2026 14:08:09 -0700 Subject: [PATCH 263/333] Linux 7.1-rc3 --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 93eded47ccd2..b7b80e84e1eb 100644 --- a/Makefile +++ b/Makefile @@ -2,7 +2,7 @@ VERSION = 7 PATCHLEVEL = 1 SUBLEVEL = 0 -EXTRAVERSION = -rc2 +EXTRAVERSION = -rc3 NAME = Baby Opossum Posse # *DOCUMENTATION* From 796ad622040f7f955ccc3973085e953415920496 Mon Sep 17 00:00:00 2001 From: Guopeng Zhang Date: Mon, 11 May 2026 09:31:50 +0800 Subject: [PATCH 264/333] cgroup/dmem: Return -ENOMEM on failed pool preallocation get_cg_pool_unlocked() handles allocation failures under dmemcg_lock by dropping the lock, preallocating a pool with GFP_KERNEL, and retrying the locked lookup and creation path. If the fallback allocation fails too, pool remains NULL. Since the loop condition is while (!pool), the function can keep retrying instead of propagating the allocation failure to the caller. Set pool to ERR_PTR(-ENOMEM) when the fallback allocation fails so the loop exits through the existing common return path. The callers already handle ERR_PTR() from get_cg_pool_unlocked(), so this restores the expected error path. Fixes: b168ed458dde ("kernel/cgroup: Add "dmem" memory accounting cgroup") Cc: stable@vger.kernel.org # v6.14+ Signed-off-by: Guopeng Zhang Signed-off-by: Tejun Heo --- kernel/cgroup/dmem.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/cgroup/dmem.c b/kernel/cgroup/dmem.c index 1ab1fb47f271..4753a67d0f0f 100644 --- a/kernel/cgroup/dmem.c +++ b/kernel/cgroup/dmem.c @@ -602,6 +602,7 @@ get_cg_pool_unlocked(struct dmemcg_state *cg, struct dmem_cgroup_region *region) pool = NULL; continue; } + pool = ERR_PTR(-ENOMEM); } } From e32e6f02168f2ad7991eb5d160d312d2001520c8 Mon Sep 17 00:00:00 2001 From: Hongfu Li Date: Sat, 9 May 2026 16:03:28 +0800 Subject: [PATCH 265/333] selftests/cgroup: Fix cg_read_strcmp() empty string comparison cg_read_strcmp() allocated a buffer sized to strlen(expected) + 1, then passed it to read_text() which calls read(fd, buf, size-1). When comparing against an empty string (""), strlen("") = 0 gives a 1-byte buffer, and read() is asked to read 0 bytes. The file content is never actually read, so strcmp("", buf) always returns 0 regardless of the real content. This caused cg_test_proc_killed() to always report the cgroup as empty immediately, making OOM tests pass without verifying that processes were killed. Signed-off-by: Hongfu Li Signed-off-by: Tejun Heo --- tools/testing/selftests/cgroup/lib/cgroup_util.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/cgroup/lib/cgroup_util.c b/tools/testing/selftests/cgroup/lib/cgroup_util.c index 6a7295347e90..42f54936f4bb 100644 --- a/tools/testing/selftests/cgroup/lib/cgroup_util.c +++ b/tools/testing/selftests/cgroup/lib/cgroup_util.c @@ -106,8 +106,9 @@ int cg_read_strcmp(const char *cgroup, const char *control, /* Handle the case of comparing against empty string */ if (!expected) return -1; - else - size = strlen(expected) + 1; + + /* needs size > 1, otherwise cg_read() reads 0 bytes */ + size = (expected[0] == '\0') ? 2 : strlen(expected) + 1; buf = malloc(size); if (!buf) From 2a3d7256faf06d1a15bb5b07e851ac4e1680c26d Mon Sep 17 00:00:00 2001 From: Hongfu Li Date: Mon, 11 May 2026 09:39:57 +0800 Subject: [PATCH 266/333] selftests/cgroup: Fix string comparison in write_test Use string comparison (!=) instead of numeric comparison (-ne) for cpuset values like "0-1". For example: $ [[ "0-1" != "2-3" ]] && echo "true" || echo "false" true $ [[ "0-1" -ne "2-3" ]] && echo "true" || echo "false" false Signed-off-by: Hongfu Li Signed-off-by: Tejun Heo --- tools/testing/selftests/cgroup/test_cpuset_v1_base.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/testing/selftests/cgroup/test_cpuset_v1_base.sh b/tools/testing/selftests/cgroup/test_cpuset_v1_base.sh index 42a6628fb8bc..1c0444729e70 100755 --- a/tools/testing/selftests/cgroup/test_cpuset_v1_base.sh +++ b/tools/testing/selftests/cgroup/test_cpuset_v1_base.sh @@ -18,7 +18,7 @@ write_test() { echo "testing $interface $value" echo $value > $dir/$interface new=$(cat $dir/$interface) - [[ $value -ne $(cat $dir/$interface) ]] && { + [[ "$value" != "$new" ]] && { echo "$interface write $value failed: new:$new" exit 1 } From 3788e32516530dee66cf9186f846480a16799b05 Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Sun, 10 May 2026 19:52:11 +0200 Subject: [PATCH 267/333] selftests/sched_ext: Fix build error in dequeue selftest Building the dequeue selftest with newer compilers (e.g., gcc 16) triggers the following error: dequeue.c:28:22: error: variable 'sum' set but not used The 'volatile' qualifier prevents the writes from being optimized away, but does not silence the unused variable 'sum' is indeed only written and never read. Consume 'sum' via an empty asm() with a register input constraint. This forces the compiler to keep the accumulated value (preserving the CPU stress loop) and avoiding the build error. Fixes: 658ad2259b3e ("selftests/sched_ext: Add test to validate ops.dequeue() semantics") Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo --- tools/testing/selftests/sched_ext/dequeue.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/testing/selftests/sched_ext/dequeue.c b/tools/testing/selftests/sched_ext/dequeue.c index 4e93262703ca..383d06e972a4 100644 --- a/tools/testing/selftests/sched_ext/dequeue.c +++ b/tools/testing/selftests/sched_ext/dequeue.c @@ -33,6 +33,7 @@ static void worker_fn(int id) /* Do some work to trigger scheduling events */ for (j = 0; j < 10000; j++) sum += j; + asm volatile("" : : "r"(sum)); /* Sleep to trigger dequeue */ usleep(1000 + (id * 100)); From bbf30b383cf6e87f2fe57c292fbd640b1d88b4c3 Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Mon, 11 May 2026 08:18:12 +0200 Subject: [PATCH 268/333] sched_ext: Fix ops->priv clobber on concurrent attach/detach Under heavy concurrent attach/detach operations, scx_claim_exit() can trigger a NULL pointer dereference. This can be reproduced running the reload_loop kselftests inside a virtme-ng session: $ vng -v -- ./tools/testing/selftests/sched_ext/runner -t reload_loop ... BUG: kernel NULL pointer dereference, address: 0000000000000400 RIP: 0010:scx_claim_exit+0x3b/0x120 Call Trace: bpf_scx_unreg+0x45/0xb0 bpf_struct_ops_map_link_dealloc+0x39/0x50 bpf_link_release+0x18/0x20 __fput+0x10b/0x2e0 __x64_sys_close+0x47/0xa0 The underlying race (diagnosed by Tejun Heo) is a stomp of @ops->priv, not a missing NULL check: T2 unreg(K) T1 reg(K) ----------- --------- sch = ops->priv = sch_b800 scx_disable; flush_disable_work [scx_root_disable: scx_root=NULL, mutex_unlock, state=DISABLED] mutex_lock; state ok scx_alloc_and_add_sched: ops->priv = sch_a800 scx_root = sch_a800; init=0 state=ENABLED; mutex_unlock [flush returns] RCU_INIT_POINTER(ops->priv, NULL) <-- clobbers sch_a800 kobject_put(sch_b800) T1 acquires scx_enable_mutex inside scx_root_disable()'s mutex_unlock window and starts a fresh attach on the same kdata, assigning sch_a800 to @ops->priv. T2 then continues out of scx_disable()/flush_disable_work and clobbers @ops->priv to NULL, leaking sch_a800; the bpf_link is gone but state stays SCX_ENABLED, so all future attaches fail with -EBUSY permanently. The next bpf_scx_unreg() on that kdata then reads NULL @ops->priv and dereferences it in scx_claim_exit(). Make @ops->priv the lifecycle binding: in scx_root_enable_workfn() and scx_sub_enable_workfn(), after the existing state check and still under scx_enable_mutex, refuse with -EBUSY if @ops->priv is non-NULL. This rejects an attempt to reuse a kdata that is still bound to a previous scheduler instance, closing the race without changing the unreg side. Fixes: 105dcd005be2 ("sched_ext: Introduce scx_prog_sched()") Suggested-by: Tejun Heo Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 4efe0099f79a..8e06694094d7 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -6803,6 +6803,19 @@ static void scx_root_enable_workfn(struct kthread_work *work) goto err_unlock; } + /* + * @ops->priv binds @ops to its scx_sched instance. It is set here by + * scx_alloc_and_add_sched() and cleared at the tail of bpf_scx_unreg(), + * which runs after scx_root_disable() has dropped scx_enable_mutex. If + * it's still non-NULL here, a previous attachment on @ops has not + * finished tearing down; proceeding would let the in-flight unreg's + * RCU_INIT_POINTER(NULL) clobber the @ops->priv we are about to assign. + */ + if (rcu_access_pointer(ops->priv)) { + ret = -EBUSY; + goto err_unlock; + } + ret = alloc_kick_syncs(); if (ret) goto err_unlock; @@ -7120,6 +7133,12 @@ static void scx_sub_enable_workfn(struct kthread_work *work) goto out_unlock; } + /* See scx_root_enable_workfn() for the @ops->priv check. */ + if (rcu_access_pointer(ops->priv)) { + ret = -EBUSY; + goto out_unlock; + } + cgrp = cgroup_get_from_id(ops->sub_cgroup_id); if (IS_ERR(cgrp)) { ret = PTR_ERR(cgrp); From 4a39eda5fdd867fc39f3c039714dd432cee00268 Mon Sep 17 00:00:00 2001 From: Guopeng Zhang Date: Sat, 9 May 2026 18:20:30 +0800 Subject: [PATCH 269/333] cgroup/cpuset: Reset DL migration state on can_attach() failure cpuset_can_attach() accumulates temporary SCHED_DEADLINE migration state in the destination cpuset while walking the taskset. If a later task_can_attach() or security_task_setscheduler() check fails, cgroup_migrate_execute() treats cpuset as the failing subsystem and does not call cpuset_cancel_attach() for it. The partially accumulated state is then left behind and can be consumed by a later attach, corrupting cpuset DL task accounting and pending DL bandwidth accounting. Reset the pending DL migration state from the common error exit when ret is non-zero. Successful can_attach() keeps the state for cpuset_attach() or cpuset_cancel_attach(). Fixes: 2ef269ef1ac0 ("cgroup/cpuset: Free DL BW in case can_attach() fails") Cc: stable@vger.kernel.org # v6.10+ Signed-off-by: Guopeng Zhang Signed-off-by: Tejun Heo Reviewed-by: Chen Ridong Reviewed-by: Waiman Long --- kernel/cgroup/cpuset.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index a48901a0416a..3fbf6e7f68c3 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -3050,16 +3050,13 @@ static int cpuset_can_attach(struct cgroup_taskset *tset) int cpu = cpumask_any_and(cpu_active_mask, cs->effective_cpus); if (unlikely(cpu >= nr_cpu_ids)) { - reset_migrate_dl_data(cs); ret = -EINVAL; goto out_unlock; } ret = dl_bw_alloc(cpu, cs->sum_migrate_dl_bw); - if (ret) { - reset_migrate_dl_data(cs); + if (ret) goto out_unlock; - } cs->dl_bw_cpu = cpu; } @@ -3070,7 +3067,10 @@ static int cpuset_can_attach(struct cgroup_taskset *tset) * changes which zero cpus/mems_allowed. */ cs->attach_in_progress++; + out_unlock: + if (ret) + reset_migrate_dl_data(cs); mutex_unlock(&cpuset_mutex); return ret; } From 86ecb1c1a1f5c1bf4a45b91f54f8220c3121bd3b Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Mon, 11 May 2026 10:31:30 +0200 Subject: [PATCH 270/333] sched_ext: Clear ops->priv on scx_alloc_and_add_sched() error paths scx_alloc_and_add_sched() can fail after @sch has been assigned to ops->priv. In those cases @sch is torn down (either via kfree() through the err_free_* chain or via kobject_put() -> scx_kobj_release() -> RCU work), but @ops->priv is left pointing at the about-to-be-freed pointer. With the recent -EBUSY gate in scx_root_enable_workfn() and scx_sub_enable_workfn() that rejects an attach when @ops->priv is still non-NULL, see commit bbf30b383cf6 ("sched_ext: Fix ops->priv clobber on concurrent attach/detach"), a dangling @ops->priv permanently locks the kdata out: every future attach attempt sees a stale binding and returns -EBUSY even though no scheduler is actually attached. Clear @ops->priv on the post-assign failure paths so that the kdata returns to its pre-attach state when the function returns ERR_PTR(). Fixes: bbf30b383cf6 ("sched_ext: Fix ops->priv clobber on concurrent attach/detach") Suggested-by: Tejun Heo Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 8e06694094d7..1efd5d82b08b 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -6670,6 +6670,7 @@ static struct scx_sched *scx_alloc_and_add_sched(struct sched_ext_ops *ops, ret = kobject_init_and_add(&sch->kobj, &scx_ktype, NULL, "root"); if (ret < 0) { + RCU_INIT_POINTER(ops->priv, NULL); kobject_put(&sch->kobj); return ERR_PTR(ret); } @@ -6677,6 +6678,7 @@ static struct scx_sched *scx_alloc_and_add_sched(struct sched_ext_ops *ops, if (ops->sub_attach) { sch->sub_kset = kset_create_and_add("sub", NULL, &sch->kobj); if (!sch->sub_kset) { + RCU_INIT_POINTER(ops->priv, NULL); kobject_put(&sch->kobj); return ERR_PTR(-ENOMEM); } @@ -6684,6 +6686,7 @@ static struct scx_sched *scx_alloc_and_add_sched(struct sched_ext_ops *ops, #else /* CONFIG_EXT_SUB_SCHED */ ret = kobject_init_and_add(&sch->kobj, &scx_ktype, NULL, "root"); if (ret < 0) { + RCU_INIT_POINTER(ops->priv, NULL); kobject_put(&sch->kobj); return ERR_PTR(ret); } @@ -6692,6 +6695,7 @@ static struct scx_sched *scx_alloc_and_add_sched(struct sched_ext_ops *ops, #ifdef CONFIG_EXT_SUB_SCHED err_free_lb_resched: + RCU_INIT_POINTER(ops->priv, NULL); free_cpumask_var(sch->bypass_lb_resched_cpumask); #endif err_free_lb_cpumask: From 657b594b2084b39a4bc6d8493aa2140cb00cea49 Mon Sep 17 00:00:00 2001 From: "Masami Hiramatsu (Google)" Date: Thu, 7 May 2026 16:46:29 +0900 Subject: [PATCH 271/333] fprobe: Fix unregister_fprobe() to wait for RCU grace period Commit 4346ba1604093 ("fprobe: Rewrite fprobe on function-graph tracer") changed fprobe to register struct fprobe to an rcu-hlist, but it forgot to wait for RCU GP. Thus there can be use-after-free if the fprobe is released right after unregistering. This can be happened on fprobe event and sample module code. To fix this issue, add synchronize_rcu() in unregister_fprobe(). Note that BPF is OK because fprobe is used as a part of bpf_kprobe_multi_link. This unregisters its fprobe in bpf_kprobe_multi_link_release() and it is deallocated via bpf_kprobe_multi_link_dealloc(), which is invoked from bpf_link_defer_dealloc_rcu_gp() RCU callback. For BPF, this also introduced unregister_fprobe_async() which does NOT wait for RCU grace priod. Link: https://lore.kernel.org/all/177813998919.256460.2809243930741138224.stgit@mhiramat.tok.corp.google.com/ Fixes: 4346ba1604093 ("fprobe: Rewrite fprobe on function-graph tracer") Signed-off-by: Masami Hiramatsu (Google) --- include/linux/fprobe.h | 5 +++++ kernel/trace/bpf_trace.c | 3 ++- kernel/trace/fprobe.c | 23 +++++++++++++++++++++-- 3 files changed, 28 insertions(+), 3 deletions(-) diff --git a/include/linux/fprobe.h b/include/linux/fprobe.h index 0a3bcd1718f3..be1b38c981d4 100644 --- a/include/linux/fprobe.h +++ b/include/linux/fprobe.h @@ -94,6 +94,7 @@ int register_fprobe(struct fprobe *fp, const char *filter, const char *notfilter int register_fprobe_ips(struct fprobe *fp, unsigned long *addrs, int num); int register_fprobe_syms(struct fprobe *fp, const char **syms, int num); int unregister_fprobe(struct fprobe *fp); +int unregister_fprobe_async(struct fprobe *fp); bool fprobe_is_registered(struct fprobe *fp); int fprobe_count_ips_from_filter(const char *filter, const char *notfilter); #else @@ -113,6 +114,10 @@ static inline int unregister_fprobe(struct fprobe *fp) { return -EOPNOTSUPP; } +static inline int unregister_fprobe_async(struct fprobe *fp) +{ + return -EOPNOTSUPP; +} static inline bool fprobe_is_registered(struct fprobe *fp) { return false; diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c index af7079aa0f36..a02bd258677e 100644 --- a/kernel/trace/bpf_trace.c +++ b/kernel/trace/bpf_trace.c @@ -2384,7 +2384,8 @@ static void bpf_kprobe_multi_link_release(struct bpf_link *link) struct bpf_kprobe_multi_link *kmulti_link; kmulti_link = container_of(link, struct bpf_kprobe_multi_link, link); - unregister_fprobe(&kmulti_link->fp); + /* Don't wait for RCU GP here. */ + unregister_fprobe_async(&kmulti_link->fp); kprobe_multi_put_modules(kmulti_link->mods, kmulti_link->mods_cnt); } diff --git a/kernel/trace/fprobe.c b/kernel/trace/fprobe.c index cc49ebd2a773..f378613ad120 100644 --- a/kernel/trace/fprobe.c +++ b/kernel/trace/fprobe.c @@ -1093,14 +1093,15 @@ static int unregister_fprobe_nolock(struct fprobe *fp) } /** - * unregister_fprobe() - Unregister fprobe. + * unregister_fprobe_async() - Unregister fprobe without RCU GP wait * @fp: A fprobe data structure to be unregistered. * * Unregister fprobe (and remove ftrace hooks from the function entries). + * This function will NOT wait until the fprobe is no longer used. * * Return 0 if @fp is unregistered successfully, -errno if not. */ -int unregister_fprobe(struct fprobe *fp) +int unregister_fprobe_async(struct fprobe *fp) { guard(mutex)(&fprobe_mutex); if (!fp || !fprobe_registered(fp)) @@ -1108,6 +1109,24 @@ int unregister_fprobe(struct fprobe *fp) return unregister_fprobe_nolock(fp); } + +/** + * unregister_fprobe() - Unregister fprobe with RCU GP wait + * @fp: A fprobe data structure to be unregistered. + * + * Unregister fprobe (and remove ftrace hooks from the function entries). + * This function will block until the fprobe is no longer used. + * + * Return 0 if @fp is unregistered successfully, -errno if not. + */ +int unregister_fprobe(struct fprobe *fp) +{ + int ret = unregister_fprobe_async(fp); + + if (!ret) + synchronize_rcu(); + return ret; +} EXPORT_SYMBOL_GPL(unregister_fprobe); static int __init fprobe_initcall(void) From c1fa0bb633e4a6b11e83ffc57fa5abe8ebb87891 Mon Sep 17 00:00:00 2001 From: Jann Horn Date: Mon, 11 May 2026 08:55:11 -0700 Subject: [PATCH 272/333] exit: prevent preemption of oopsing TASK_DEAD task When an already-exiting task oopses, make_task_dead() currently calls do_task_dead() with preemption enabled. That is forbidden: do_task_dead() calls __schedule(), which has a comment saying "WARNING: must be called with preemption disabled!". If an oopsing task is preempted in do_task_dead(), between becoming TASK_DEAD and entering the scheduler explicitly, bad things happen: finish_task_switch() assumes that once the scheduler has switched away from a TASK_DEAD task, the task can never run again and its stack is no longer needed; but that assumption apparently doesn't hold if the dead task was preempted (the SM_PREEMPT case). This means that the scheduler ends up repeatedly dropping references on the dead task's stack, which can lead to use-after-free or double-free of the entire task stack; in other words, two tasks can end up running on the same stack, resulting in various kinds of memory corruption. (This does not just affect "recursively oopsing" tasks; it is enough to oops once during task exit, for example in a file_operations::release handler) Fixes: 7f80a2fd7db9 ("exit: Stop poorly open coding do_task_dead in make_task_dead") Cc: stable@kernel.org Signed-off-by: Jann Horn Acked-by: Peter Zijlstra Signed-off-by: Linus Torvalds --- kernel/exit.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/exit.c b/kernel/exit.c index 25e9cb6de7e7..9a909993ab1d 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -1073,6 +1073,7 @@ void __noreturn make_task_dead(int signr) futex_exit_recursive(tsk); tsk->exit_state = EXIT_DEAD; refcount_inc(&tsk->rcu_users); + preempt_disable(); do_task_dead(); } From 5dd74441cbf42c22e874450eb6a6bbb19390a216 Mon Sep 17 00:00:00 2001 From: Guopeng Zhang Date: Sat, 9 May 2026 18:20:31 +0800 Subject: [PATCH 273/333] cgroup/cpuset: Reserve DL bandwidth only for root-domain moves cpuset_can_attach() currently adds the bandwidth of all migrating SCHED_DEADLINE tasks to sum_migrate_dl_bw. If the source and destination cpuset effective CPU masks do not overlap, the whole sum is then reserved in the destination root domain. set_cpus_allowed_dl(), however, subtracts bandwidth from the source root domain only when the affinity change really moves the task between root domains. A DL task can move between cpusets that are still in the same root domain, so including that task in sum_migrate_dl_bw can reserve destination bandwidth without a matching source-side subtraction. Share the root-domain move test with set_cpus_allowed_dl(). Keep nr_migrate_dl_tasks counting all migrating deadline tasks for cpuset DL task accounting, but add to sum_migrate_dl_bw only for tasks that need a root-domain bandwidth move. Keep using the destination cpuset effective CPU mask and leave the broader can_attach()/attach() transaction model unchanged. Fixes: 2ef269ef1ac0 ("cgroup/cpuset: Free DL BW in case can_attach() fails") Cc: stable@vger.kernel.org # v6.10+ Signed-off-by: Guopeng Zhang Reviewed-by: Waiman Long Acked-by: Juri Lelli Tested-by: Juri Lelli Signed-off-by: Tejun Heo --- include/linux/sched/deadline.h | 9 +++++++++ kernel/cgroup/cpuset-internal.h | 1 + kernel/cgroup/cpuset.c | 35 ++++++++++++++++++--------------- kernel/sched/deadline.c | 13 +++++++++--- 4 files changed, 39 insertions(+), 19 deletions(-) diff --git a/include/linux/sched/deadline.h b/include/linux/sched/deadline.h index 1198138cb839..273538200a44 100644 --- a/include/linux/sched/deadline.h +++ b/include/linux/sched/deadline.h @@ -33,6 +33,15 @@ struct root_domain; extern void dl_add_task_root_domain(struct task_struct *p); extern void dl_clear_root_domain(struct root_domain *rd); extern void dl_clear_root_domain_cpu(int cpu); +/* + * Return whether moving DL task @p to @new_mask requires moving DL + * bandwidth accounting between root domains. This helper is specific to + * DL bandwidth move accounting semantics and is shared by + * cpuset_can_attach() and set_cpus_allowed_dl() so both paths use the + * same source root-domain test. + */ +extern bool dl_task_needs_bw_move(struct task_struct *p, + const struct cpumask *new_mask); extern u64 dl_cookie; extern bool dl_bw_visited(int cpu, u64 cookie); diff --git a/kernel/cgroup/cpuset-internal.h b/kernel/cgroup/cpuset-internal.h index bb4e692bea30..f7aaf01f7cd5 100644 --- a/kernel/cgroup/cpuset-internal.h +++ b/kernel/cgroup/cpuset-internal.h @@ -167,6 +167,7 @@ struct cpuset { */ int nr_deadline_tasks; int nr_migrate_dl_tasks; + /* DL bandwidth that needs destination reservation for this attach. */ u64 sum_migrate_dl_bw; /* * CPU used for temporary DL bandwidth allocation during attach; diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index 3fbf6e7f68c3..e84e801e22cf 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -2993,7 +2993,7 @@ static int cpuset_can_attach(struct cgroup_taskset *tset) struct cpuset *cs, *oldcs; struct task_struct *task; bool setsched_check; - int ret; + int cpu, ret; /* used later by cpuset_attach() */ cpuset_attach_old_cs = task_cs(cgroup_taskset_first(tset, &css)); @@ -3038,29 +3038,32 @@ static int cpuset_can_attach(struct cgroup_taskset *tset) } if (dl_task(task)) { + /* + * Count all migrating DL tasks for cpuset task accounting. + * Only tasks that need a root-domain bandwidth move + * contribute to sum_migrate_dl_bw. + */ cs->nr_migrate_dl_tasks++; - cs->sum_migrate_dl_bw += task->dl.dl_bw; + if (dl_task_needs_bw_move(task, cs->effective_cpus)) + cs->sum_migrate_dl_bw += task->dl.dl_bw; } } - if (!cs->nr_migrate_dl_tasks) + if (!cs->sum_migrate_dl_bw) goto out_success; - if (!cpumask_intersects(oldcs->effective_cpus, cs->effective_cpus)) { - int cpu = cpumask_any_and(cpu_active_mask, cs->effective_cpus); - - if (unlikely(cpu >= nr_cpu_ids)) { - ret = -EINVAL; - goto out_unlock; - } - - ret = dl_bw_alloc(cpu, cs->sum_migrate_dl_bw); - if (ret) - goto out_unlock; - - cs->dl_bw_cpu = cpu; + cpu = cpumask_any_and(cpu_active_mask, cs->effective_cpus); + if (unlikely(cpu >= nr_cpu_ids)) { + ret = -EINVAL; + goto out_unlock; } + ret = dl_bw_alloc(cpu, cs->sum_migrate_dl_bw); + if (ret) + goto out_unlock; + + cs->dl_bw_cpu = cpu; + out_success: /* * Mark attach is in progress. This makes validate_change() fail diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c index edca7849b165..7db4c87df83b 100644 --- a/kernel/sched/deadline.c +++ b/kernel/sched/deadline.c @@ -3107,20 +3107,18 @@ static void task_woken_dl(struct rq *rq, struct task_struct *p) static void set_cpus_allowed_dl(struct task_struct *p, struct affinity_context *ctx) { - struct root_domain *src_rd; struct rq *rq; WARN_ON_ONCE(!dl_task(p)); rq = task_rq(p); - src_rd = rq->rd; /* * Migrating a SCHED_DEADLINE task between exclusive * cpusets (different root_domains) entails a bandwidth * update. We already made space for us in the destination * domain (see cpuset_can_attach()). */ - if (!cpumask_intersects(src_rd->span, ctx->new_mask)) { + if (dl_task_needs_bw_move(p, ctx->new_mask)) { struct dl_bw *src_dl_b; src_dl_b = dl_bw_of(cpu_of(rq)); @@ -3137,6 +3135,15 @@ static void set_cpus_allowed_dl(struct task_struct *p, set_cpus_allowed_common(p, ctx); } +bool dl_task_needs_bw_move(struct task_struct *p, + const struct cpumask *new_mask) +{ + if (!dl_task(p)) + return false; + + return !cpumask_intersects(task_rq(p)->rd->span, new_mask); +} + /* Assumes rq->lock is held */ static void rq_online_dl(struct rq *rq) { From 9a415cc53711f2238e0f0ca8a6bcc796c003b127 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 11 May 2026 12:05:48 -1000 Subject: [PATCH 274/333] sched_ext: Avoid UAF in scx_root_enable_workfn() init failure path In scx_root_enable_workfn(), put_task_struct(p) is called before scx_error() dereferences p->comm and p->pid. If the iterator's reference is the last drop, the task is freed synchronously and the deref becomes a UAF. Move put_task_struct() past scx_error(). Reported-by: Sashiko Closes: https://lore.kernel.org/all/20260511214031.AF5E9C2BCB0@smtp.kernel.org/ Fixes: f0e1a0643a59 ("sched_ext: Implement BPF extensible scheduler class") Cc: stable@vger.kernel.org # v6.12+ Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 1efd5d82b08b..9354da79e162 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -6973,10 +6973,10 @@ static void scx_root_enable_workfn(struct kthread_work *work) if (scx_get_task_state(p) != SCX_TASK_DEAD) scx_set_task_state(p, SCX_TASK_NONE); task_rq_unlock(rq, p, &rf); - put_task_struct(p); scx_task_iter_stop(&sti); scx_error(sch, "ops.init_task() failed (%d) for %s[%d]", ret, p->comm, p->pid); + put_task_struct(p); goto err_disable_unlock_all; } From e174929793195e0cd6a4adb0cad731b39f9019b4 Mon Sep 17 00:00:00 2001 From: Allison Henderson Date: Tue, 5 May 2026 16:43:36 -0700 Subject: [PATCH 275/333] net/rds: reset op_nents when zerocopy page pin fails When iov_iter_get_pages2() fails in rds_message_zcopy_from_user(), the pinned pages are released with put_page(), and rm->data.op_mmp_znotifier is cleared. But we fail to properly clear rm->data.op_nents. Later when rds_message_purge() is called from rds_sendmsg() the cleanup loop iterates over the incorrectly non zero number of op_nents and frees them again. Fix this by properly resetting op_nents when it should be in rds_message_zcopy_from_user(). Fixes: 0cebaccef3ac ("rds: zerocopy Tx support.") Signed-off-by: Allison Henderson Reviewed-by: Simon Horman Link: https://patch.msgid.link/20260505234336.2132721-1-achender@kernel.org Signed-off-by: Jakub Kicinski --- net/rds/message.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/rds/message.c b/net/rds/message.c index 25fedcb3cd00..7feb0eb6537d 100644 --- a/net/rds/message.c +++ b/net/rds/message.c @@ -448,6 +448,7 @@ static int rds_message_zcopy_from_user(struct rds_message *rm, struct iov_iter * for (i = 0; i < rm->data.op_nents; i++) put_page(sg_page(&rm->data.op_sg[i])); + rm->data.op_nents = 0; mmp = &rm->data.op_mmp_znotifier->z_mmp; mm_unaccount_pinned_pages(mmp); ret = -EFAULT; From 24a08d7d6218d60c033015cf4870b6096446e734 Mon Sep 17 00:00:00 2001 From: Arthur Kiyanovski Date: Thu, 7 May 2026 00:35:15 +0000 Subject: [PATCH 276/333] net: ena: PHC: Check return code before setting timestamp output ena_phc_gettimex64() is setting the output parameter regardless of whether ena_com_phc_get_timestamp() succeeded or failed. When ena_com_phc_get_timestamp() returns an error, the timestamp parameter may contain uninitialized stack memory (e.g., when PHC is disabled or in blocked state) or invalid hardware values. Passing these to userspace via the PTP ioctl is both a security issue (information leak) and a correctness bug. Fix by checking the return code after releasing the lock and only setting the output timestamp on success. Fixes: e0ea34158ee8 ("net: ena: Add PHC support in the ENA driver") Cc: stable@vger.kernel.org Signed-off-by: Arthur Kiyanovski Reviewed-by: Vadim Fedorenko Link: https://patch.msgid.link/20260507003518.22554-1-akiyano@amazon.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/amazon/ena/ena_phc.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/amazon/ena/ena_phc.c b/drivers/net/ethernet/amazon/ena/ena_phc.c index 7867e893fd15..c2a3ff1ef645 100644 --- a/drivers/net/ethernet/amazon/ena/ena_phc.c +++ b/drivers/net/ethernet/amazon/ena/ena_phc.c @@ -46,9 +46,12 @@ static int ena_phc_gettimex64(struct ptp_clock_info *clock_info, spin_unlock_irqrestore(&phc_info->lock, flags); + if (rc) + return rc; + *ts = ns_to_timespec64(timestamp_nsec); - return rc; + return 0; } static int ena_phc_settime64(struct ptp_clock_info *clock_info, From 03cb001ef87b3f8d859cf7f96329acf3d6235d29 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 8 May 2026 12:08:46 +0000 Subject: [PATCH 277/333] tcp: Fix out-of-bounds access for twsk in tcp_ao_established_key(). lockdep_sock_is_held() was added in tcp_ao_established_key() by the cited commit. It can be called from tcp_v[46]_timewait_ack() with twsk. Since it does not have sk->sk_lock, the lockdep annotation results in out-of-bound access. $ pahole -C tcp_timewait_sock vmlinux | grep size /* size: 288, cachelines: 5, members: 8 */ $ pahole -C sock vmlinux | grep sk_lock socket_lock_t sk_lock; /* 440 192 */ Let's not use lockdep_sock_is_held() for TCP_TIME_WAIT. Fixes: 6b2d11e2d8fc ("net/tcp: Add missing lockdep annotations for TCP-AO hlist traversals") Reported-by: Damiano Melotti Signed-off-by: Kuniyuki Iwashima Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/20260508120853.4098365-1-kuniyu@google.com Signed-off-by: Jakub Kicinski --- net/ipv4/tcp_ao.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/net/ipv4/tcp_ao.c b/net/ipv4/tcp_ao.c index a97cdf3e6af4..0a4b38b315fe 100644 --- a/net/ipv4/tcp_ao.c +++ b/net/ipv4/tcp_ao.c @@ -116,7 +116,8 @@ struct tcp_ao_key *tcp_ao_established_key(const struct sock *sk, { struct tcp_ao_key *key; - hlist_for_each_entry_rcu(key, &ao->head, node, lockdep_sock_is_held(sk)) { + hlist_for_each_entry_rcu(key, &ao->head, node, + sk_fullsock(sk) && lockdep_sock_is_held(sk)) { if ((sndid >= 0 && key->sndid != sndid) || (rcvid >= 0 && key->rcvid != rcvid)) continue; From f8e64e956a635b054227f56e9586a1997add0646 Mon Sep 17 00:00:00 2001 From: Davide Caratti Date: Fri, 8 May 2026 19:05:10 +0200 Subject: [PATCH 278/333] net/sched: dualpi2: initialize timer earlier in dualpi2_init() 'pi2_timer' needs to be initialized in all error paths of dualpi2_init(): otherwise, a failure in qdisc_create_dflt() causes the following crash in dualpi2_destroy(): # tc qdisc add dev crash0 handle 1: root dualpi2 BUG: kernel NULL pointer dereference, address: 0000000000000010 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 0 P4D 0 Oops: Oops: 0000 [#1] SMP PTI CPU: 4 UID: 0 PID: 471 Comm: tc Tainted: G E 7.1.0-rc1-virtme #2 PREEMPT(full) Tainted: [E]=UNSIGNED_MODULE Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 RIP: 0010:hrtimer_active+0x39/0x60 Code: f9 eb 23 0f b6 41 38 3c 01 0f 87 87 64 c0 ff 83 e0 01 75 33 48 39 4a 50 74 28 44 3b 42 10 75 06 48 3b 51 30 74 21 48 8b 51 30 <44> 8b 42 10 41 f6 c0 01 74 cf f3 90 44 8b 42 10 41 f6 c0 01 74 c3 RSP: 0018:ffffd0db80b93620 EFLAGS: 00010282 RAX: ffffffffc0400320 RBX: ffff8cf24a4c86b8 RCX: ffff8cf24a4c86b8 RDX: 0000000000000000 RSI: ffff8cf2429c2ab0 RDI: ffff8cf24a4c86b8 RBP: 00000000fffffff4 R08: 0000000000000003 R09: 0000000000000000 R10: 0000000000000001 R11: ffff8cf24a39c500 R12: ffff8cf24822c000 R13: ffffd0db80b936c0 R14: ffffffffc02cf360 R15: 00000000ffffffff FS: 00007fbc01706580(0000) GS:ffff8cf2dc759000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000010 CR3: 0000000008e02003 CR4: 0000000000172ef0 Call Trace: hrtimer_cancel+0x15/0x40 dualpi2_destroy+0x20/0x40 [sch_dualpi2] qdisc_create+0x230/0x570 tc_modify_qdisc+0x716/0xc10 rtnetlink_rcv_msg+0x188/0x780 netlink_rcv_skb+0xcd/0x150 netlink_unicast+0x1ba/0x290 netlink_sendmsg+0x242/0x4d0 ____sys_sendmsg+0x39e/0x3e0 ___sys_sendmsg+0xe1/0x130 __sys_sendmsg+0xad/0x110 do_syscall_64+0x14f/0xf80 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0033:0x7fbc0188b08e Code: 4d 89 d8 e8 94 bd 00 00 4c 8b 5d f8 41 8b 93 08 03 00 00 59 5e 48 83 f8 fc 74 11 c9 c3 0f 1f 80 00 00 00 00 48 8b 45 10 0f 05 c3 83 e2 39 83 fa 08 75 e7 e8 03 ff ff ff 0f 1f 00 f3 0f 1e fa RSP: 002b:00007fff593260e0 EFLAGS: 00000202 ORIG_RAX: 000000000000002e RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fbc0188b08e RDX: 0000000000000000 RSI: 00007fff59326190 RDI: 0000000000000003 RBP: 00007fff593260f0 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000202 R12: 000055f06124f260 R13: 0000000069fca043 R14: 000055f061255640 R15: 000055f06124d3f8 Modules linked in: sch_dualpi2(E) CR2: 0000000000000010 [1] https://lore.kernel.org/netdev/2e78e01c504c633ebdff18d041833cf2e079a3a4.1607020450.git.dcaratti@redhat.com/ [2] https://lore.kernel.org/netdev/20200725201707.16909-1-xiyou.wangcong@gmail.com/ v2: - rebased on top of latest net.git Fixes: 320d031ad6e4 ("sched: Struct definition and parsing of dualpi2 qdisc") Signed-off-by: Davide Caratti Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/1faca91179702b31da5d87653e1e036543e32722.1778259798.git.dcaratti@redhat.com Signed-off-by: Jakub Kicinski --- net/sched/sch_dualpi2.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/net/sched/sch_dualpi2.c b/net/sched/sch_dualpi2.c index 241e6a46bd00..a22489c14458 100644 --- a/net/sched/sch_dualpi2.c +++ b/net/sched/sch_dualpi2.c @@ -938,6 +938,8 @@ static int dualpi2_init(struct Qdisc *sch, struct nlattr *opt, int err; sch->flags |= TCQ_F_DEQUEUE_DROPS; + hrtimer_setup(&q->pi2_timer, dualpi2_timer, CLOCK_MONOTONIC, + HRTIMER_MODE_ABS_PINNED_SOFT); q->l_queue = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, TC_H_MAKE(sch->handle, 1), extack); @@ -950,8 +952,6 @@ static int dualpi2_init(struct Qdisc *sch, struct nlattr *opt, q->sch = sch; dualpi2_reset_default(sch); - hrtimer_setup(&q->pi2_timer, dualpi2_timer, CLOCK_MONOTONIC, - HRTIMER_MODE_ABS_PINNED_SOFT); if (opt && nla_len(opt)) { err = dualpi2_change(sch, opt, extack); From be48e5fe51a5864566307998286a699d6b986934 Mon Sep 17 00:00:00 2001 From: Evgenii Burenchev Date: Thu, 7 May 2026 17:55:17 +0300 Subject: [PATCH 279/333] qed: fix division by zero in qed_init_wfq_param when all vports are configured In qed_init_wfq_param(), variable non_requested_count can become zero when the number of vports with the configured flag set (including the current vport being configured) equals total num_vports. This happens when configuring the last unconfigured vport or when re-configuring an already configured vport. The function then calculates left_rate_per_vp = total_left_rate / non_requested_count, which causes division by zero. Fix this by skipping the division when non_requested_count is zero. In that case, there is no remaining bandwidth to distribute, so just record the configuration for the current vport and return success. Fixes: bcd197c81f63 ("qed: Add vport WFQ configuration APIs") Signed-off-by: Evgenii Burenchev Link: https://patch.msgid.link/20260507145520.23106-1-evg28bur@yandex.ru Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/qlogic/qed/qed_dev.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/net/ethernet/qlogic/qed/qed_dev.c b/drivers/net/ethernet/qlogic/qed/qed_dev.c index 42c6dcfb1f0f..dd75c47758e1 100644 --- a/drivers/net/ethernet/qlogic/qed/qed_dev.c +++ b/drivers/net/ethernet/qlogic/qed/qed_dev.c @@ -5103,6 +5103,13 @@ static int qed_init_wfq_param(struct qed_hwfn *p_hwfn, return -EINVAL; } + /* All vports are already or become configured, nothing to distribute */ + if (non_requested_count == 0) { + p_hwfn->qm_info.wfq_data[vport_id].min_speed = req_rate; + p_hwfn->qm_info.wfq_data[vport_id].configured = true; + return 0; + } + total_left_rate = min_pf_rate - total_req_min_rate; left_rate_per_vp = total_left_rate / non_requested_count; From c21b90f77687075115d989e53a8ec5e2bb427ab1 Mon Sep 17 00:00:00 2001 From: Prathyushi Nangia Date: Tue, 9 Dec 2025 10:01:33 -0600 Subject: [PATCH 280/333] x86/CPU/AMD: Prevent improper isolation of shared resources in Zen2's op cache Make sure resources are not improperly shared in the op cache and cause instruction corruption this way. Signed-off-by: Prathyushi Nangia Co-developed-by: Borislav Petkov (AMD) Signed-off-by: Borislav Petkov (AMD) Cc: stable@vger.kernel.org Signed-off-by: Linus Torvalds --- arch/x86/include/asm/msr-index.h | 3 ++- arch/x86/kernel/cpu/amd.c | 3 +++ tools/arch/x86/include/asm/msr-index.h | 3 ++- 3 files changed, 7 insertions(+), 2 deletions(-) diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h index a14a0f43e04a..86554de9a3f5 100644 --- a/arch/x86/include/asm/msr-index.h +++ b/arch/x86/include/asm/msr-index.h @@ -803,9 +803,10 @@ #define MSR_AMD64_LBR_SELECT 0xc000010e /* Zen4 */ -#define MSR_ZEN4_BP_CFG 0xc001102e +#define MSR_ZEN4_BP_CFG 0xc001102e #define MSR_ZEN4_BP_CFG_BP_SPEC_REDUCE_BIT 4 #define MSR_ZEN4_BP_CFG_SHARED_BTB_FIX_BIT 5 +#define MSR_ZEN2_BP_CFG_BUG_FIX_BIT 33 /* Fam 19h MSRs */ #define MSR_F19H_UMC_PERF_CTL 0xc0010800 diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 2d9ae6ab1701..2f8e8ff2d000 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -989,6 +989,9 @@ static void init_amd_zen2(struct cpuinfo_x86 *c) /* Correct misconfigured CPUID on some clients. */ clear_cpu_cap(c, X86_FEATURE_INVLPGB); + + if (!cpu_has(c, X86_FEATURE_HYPERVISOR)) + msr_set_bit(MSR_ZEN4_BP_CFG, MSR_ZEN2_BP_CFG_BUG_FIX_BIT); } static void init_amd_zen3(struct cpuinfo_x86 *c) diff --git a/tools/arch/x86/include/asm/msr-index.h b/tools/arch/x86/include/asm/msr-index.h index 6673601246b3..eff29645719b 100644 --- a/tools/arch/x86/include/asm/msr-index.h +++ b/tools/arch/x86/include/asm/msr-index.h @@ -793,9 +793,10 @@ #define MSR_AMD64_LBR_SELECT 0xc000010e /* Zen4 */ -#define MSR_ZEN4_BP_CFG 0xc001102e +#define MSR_ZEN4_BP_CFG 0xc001102e #define MSR_ZEN4_BP_CFG_BP_SPEC_REDUCE_BIT 4 #define MSR_ZEN4_BP_CFG_SHARED_BTB_FIX_BIT 5 +#define MSR_ZEN2_BP_CFG_BUG_FIX_BIT 33 /* Fam 19h MSRs */ #define MSR_F19H_UMC_PERF_CTL 0xc0010800 From 911f54771ca97947cfdca360e9e9b4147a330740 Mon Sep 17 00:00:00 2001 From: Quan Sun <2022090917019@std.uestc.edu.cn> Date: Fri, 8 May 2026 20:46:36 +0800 Subject: [PATCH 281/333] net: hsr: fix NULL pointer dereference in hsr_get_node_data() In the HSR (High-availability Seamless Redundancy) protocol, node information is maintained in the node_db. When a supervision frame is received, node->addr_B_port is updated to track the receiving port type (e.g., HSR_PT_SLAVE_B). If the underlying physical interface associated with this slave port is removed (e.g., via `ip link del`), hsr_del_port() frees the hsr_port object. However, the stale node->addr_B_port reference is kept in the node_db until the node ages out. Subsequently, if userspace queries the node status via the Netlink command HSR_C_GET_NODE_STATUS, the kernel calls hsr_get_node_data(). This function unconditionally dereferences the pointer returned by hsr_port_get_hsr(): if (node->addr_B_port != HSR_PT_NONE) { port = hsr_port_get_hsr(hsr, node->addr_B_port); *addr_b_ifindex = port->dev->ifindex; // <-- NULL deref } If the slave port has been deleted, hsr_port_get_hsr() returns NULL, resulting in a kernel panic. Oops: general protection fault, probably for non-canonical address KASAN: null-ptr-deref in range [0x0000000000000010-0x0000000000000017] RIP: 0010:hsr_get_node_data+0x7b6/0x9e0 Call Trace: hsr_get_node_status+0x445/0xa40 Fix this by adding a proper NULL pointer check. If the port lookup fails due to a stale port type, gracefully treat it as if no valid port exists and assign -1 to the interface index. Steps to reproduce: 1. Create an HSR interface with two slave devices. 2. Receive a supervision frame to populate node_db with addr_B_port assigned to SLAVE_B. 3. Delete the underlying slave device B. 4. Send an HSR_C_GET_NODE_STATUS Netlink message. Fixes: c5a759117210 ("net/hsr: Use list_head (and rcu) instead of array for slave devices.") Signed-off-by: Quan Sun <2022090917019@std.uestc.edu.cn> Link: https://patch.msgid.link/20260508124636.1462346-1-2022090917019@std.uestc.edu.cn Signed-off-by: Paolo Abeni --- net/hsr/hsr_framereg.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/net/hsr/hsr_framereg.c b/net/hsr/hsr_framereg.c index d09875b33588..124619920d38 100644 --- a/net/hsr/hsr_framereg.c +++ b/net/hsr/hsr_framereg.c @@ -889,7 +889,10 @@ int hsr_get_node_data(struct hsr_priv *hsr, if (node->addr_B_port != HSR_PT_NONE) { port = hsr_port_get_hsr(hsr, node->addr_B_port); - *addr_b_ifindex = port->dev->ifindex; + if (port) + *addr_b_ifindex = port->dev->ifindex; + else + *addr_b_ifindex = -1; } else { *addr_b_ifindex = -1; } From 5f344d809e015fba3709e5219428c00b8ac5d7df Mon Sep 17 00:00:00 2001 From: Stefano Garzarella Date: Fri, 8 May 2026 18:44:10 +0200 Subject: [PATCH 282/333] vsock/virtio: fix length and offset in tap skb for split packets virtio_transport_build_skb() builds a new skb to be delivered to the vsockmon tap device. To build the new skb, it uses the original skb data length as payload length, but as the comment notes, the original packet stored in the skb may have been split in multiple packets, so we need to use the length in the header, which is correctly updated before the packet is delivered to the tap, and the offset for the data. This was also similar to what we did before commit 71dc9ec9ac7d ("virtio/vsock: replace virtio_vsock_pkt with sk_buff") where we probably missed something during the skb conversion. Also update the comment above, which was left stale by the skb conversion and still mentioned a buffer pointer that no longer exists. Fixes: 71dc9ec9ac7d ("virtio/vsock: replace virtio_vsock_pkt with sk_buff") Signed-off-by: Stefano Garzarella Reviewed-by: Bobby Eshleman Reviewed-by: Arseniy Krasnov Link: https://patch.msgid.link/20260508164411.261440-2-sgarzare@redhat.com Acked-by: Michael S. Tsirkin Signed-off-by: Paolo Abeni --- net/vmw_vsock/virtio_transport_common.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c index 9b8014516f4f..a678d5d75704 100644 --- a/net/vmw_vsock/virtio_transport_common.c +++ b/net/vmw_vsock/virtio_transport_common.c @@ -166,12 +166,12 @@ static struct sk_buff *virtio_transport_build_skb(void *opaque) struct sk_buff *skb; size_t payload_len; - /* A packet could be split to fit the RX buffer, so we can retrieve - * the payload length from the header and the buffer pointer taking - * care of the offset in the original packet. + /* A packet could be split to fit the RX buffer, so we use + * the payload length from the header, which has been updated + * by the sender to reflect the fragment size. */ pkt_hdr = virtio_vsock_hdr(pkt); - payload_len = pkt->len; + payload_len = le32_to_cpu(pkt_hdr->len); skb = alloc_skb(sizeof(*hdr) + sizeof(*pkt_hdr) + payload_len, GFP_ATOMIC); @@ -219,7 +219,8 @@ static struct sk_buff *virtio_transport_build_skb(void *opaque) virtio_transport_copy_nonlinear_skb(pkt, data, payload_len); } else { - skb_put_data(skb, pkt->data, payload_len); + skb_put_data(skb, pkt->data + VIRTIO_VSOCK_SKB_CB(pkt)->offset, + payload_len); } } From 3a3e3d90cbc79600544536723911657730759af3 Mon Sep 17 00:00:00 2001 From: Stefano Garzarella Date: Fri, 8 May 2026 18:44:11 +0200 Subject: [PATCH 283/333] vsock/virtio: fix empty payload in tap skb for non-linear buffers For non-linear skbs, virtio_transport_build_skb() goes through virtio_transport_copy_nonlinear_skb() to copy the original payload in the new skb to be delivered to the vsockmon tap device. This manually initializes an iov_iter but does not set iov_iter.count. Since the iov_iter is zero-initialized, the copy length is zero and no payload is actually copied to the monitor interface, leaving data un-initialized. Fix this by removing the linear vs non-linear split and using skb_copy_datagram_iter() with iov_iter_kvec() for all cases, as vhost-vsock already does. This handles both linear and non-linear skbs, properly initializes the iov_iter, and removes the now unused virtio_transport_copy_nonlinear_skb(). While touching this code, let's also check the return value of skb_copy_datagram_iter(), even though it's unlikely to fail. Fixes: 4b0bf10eb077 ("vsock/virtio: non-linear skb handling for tap") Reported-by: Yiqi Sun Signed-off-by: Stefano Garzarella Reviewed-by: Bobby Eshleman Reviewed-by: Arseniy Krasnov Link: https://patch.msgid.link/20260508164411.261440-3-sgarzare@redhat.com Acked-by: Michael S. Tsirkin Signed-off-by: Paolo Abeni --- net/vmw_vsock/virtio_transport_common.c | 38 +++++++------------------ 1 file changed, 11 insertions(+), 27 deletions(-) diff --git a/net/vmw_vsock/virtio_transport_common.c b/net/vmw_vsock/virtio_transport_common.c index a678d5d75704..989cc252d3d3 100644 --- a/net/vmw_vsock/virtio_transport_common.c +++ b/net/vmw_vsock/virtio_transport_common.c @@ -136,27 +136,6 @@ static void virtio_transport_init_hdr(struct sk_buff *skb, hdr->fwd_cnt = cpu_to_le32(0); } -static void virtio_transport_copy_nonlinear_skb(const struct sk_buff *skb, - void *dst, - size_t len) -{ - struct iov_iter iov_iter = { 0 }; - struct kvec kvec; - size_t to_copy; - - kvec.iov_base = dst; - kvec.iov_len = len; - - iov_iter.iter_type = ITER_KVEC; - iov_iter.kvec = &kvec; - iov_iter.nr_segs = 1; - - to_copy = min_t(size_t, len, skb->len); - - skb_copy_datagram_iter(skb, VIRTIO_VSOCK_SKB_CB(skb)->offset, - &iov_iter, to_copy); -} - /* Packet capture */ static struct sk_buff *virtio_transport_build_skb(void *opaque) { @@ -214,13 +193,18 @@ static struct sk_buff *virtio_transport_build_skb(void *opaque) skb_put_data(skb, pkt_hdr, sizeof(*pkt_hdr)); if (payload_len) { - if (skb_is_nonlinear(pkt)) { - void *data = skb_put(skb, payload_len); + struct iov_iter iov_iter; + struct kvec kvec; + void *data = skb_put(skb, payload_len); - virtio_transport_copy_nonlinear_skb(pkt, data, payload_len); - } else { - skb_put_data(skb, pkt->data + VIRTIO_VSOCK_SKB_CB(pkt)->offset, - payload_len); + kvec.iov_base = data; + kvec.iov_len = payload_len; + iov_iter_kvec(&iov_iter, ITER_DEST, &kvec, 1, payload_len); + + if (skb_copy_datagram_iter(pkt, VIRTIO_VSOCK_SKB_CB(pkt)->offset, + &iov_iter, payload_len)) { + kfree_skb(skb); + return NULL; } } From 2cb156213093a62b80cf40b1ec71738e93491971 Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Sat, 9 May 2026 00:13:36 +0200 Subject: [PATCH 284/333] net: ethernet: cortina: No mapping is a dropped rx Increase stats.rx_dropped++ even if this is the first fragment (skb == NULL) so we are doing proper accounting. Fixes: b266bacba796 ("net: ethernet: cortina: Drop half-assembled SKB") Link: https://sashiko.dev/#/patchset/20260505-gemini-ethernet-fix-v2-1-997c31d06079%40kernel.org Signed-off-by: Linus Walleij Link: https://patch.msgid.link/20260509-gemini-ethernet-fixes-v1-1-6c5d20ddc35b@kernel.org Signed-off-by: Paolo Abeni --- drivers/net/ethernet/cortina/gemini.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/cortina/gemini.c b/drivers/net/ethernet/cortina/gemini.c index 065cbbf52686..466445c9e08b 100644 --- a/drivers/net/ethernet/cortina/gemini.c +++ b/drivers/net/ethernet/cortina/gemini.c @@ -1491,9 +1491,9 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) gpage = gmac_get_queue_page(geth, port, mapping + PAGE_SIZE); if (!gpage) { dev_err(geth->dev, "could not find mapping\n"); + port->stats.rx_dropped++; if (skb) { napi_free_frags(&port->napi); - port->stats.rx_dropped++; skb = NULL; } continue; From 06937db21ee311ed07eba47954447245041a982d Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Sat, 9 May 2026 00:13:37 +0200 Subject: [PATCH 285/333] net: ethernet: cortina: Make RX SKB per-port The SKB used to assemble packets from fragments in gmac_rx() is static local, but the Gemini has two ethernet ports, meaning there can be races between the ports on a bad day if a device is using both. Make the RX SKB a per-port variable and carry it over between invocations in the port struct instead. Zero the pointer once we call napi_gro_frags(), on error (after calling napi_free_frags()) or if the port is stopped. Zero it in some place where not strictly necessary just to emphasize what is going on. This was found by Sashiko during normal patch review. Fixes: 4d5ae32f5e1e ("net: ethernet: Add a driver for Gemini gigabit ethernet") Link: https://sashiko.dev/#/patchset/20260505-gemini-ethernet-fix-v2-1-997c31d06079%40kernel.org Signed-off-by: Linus Walleij Link: https://patch.msgid.link/20260509-gemini-ethernet-fixes-v1-2-6c5d20ddc35b@kernel.org Signed-off-by: Paolo Abeni --- drivers/net/ethernet/cortina/gemini.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/cortina/gemini.c b/drivers/net/ethernet/cortina/gemini.c index 466445c9e08b..d5a56366fb43 100644 --- a/drivers/net/ethernet/cortina/gemini.c +++ b/drivers/net/ethernet/cortina/gemini.c @@ -122,6 +122,8 @@ struct gemini_ethernet_port { struct napi_struct napi; struct hrtimer rx_coalesce_timer; unsigned int rx_coalesce_nsecs; + struct sk_buff *rx_skb; + unsigned int freeq_refill; struct gmac_txq txq[TX_QUEUE_NUM]; unsigned int txq_order; @@ -1442,10 +1444,10 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) unsigned short m = (1 << port->rxq_order) - 1; struct gemini_ethernet *geth = port->geth; void __iomem *ptr_reg = port->rxq_rwptr; + struct sk_buff *skb = port->rx_skb; unsigned int frame_len, frag_len; struct gmac_rxdesc *rx = NULL; struct gmac_queue_page *gpage; - static struct sk_buff *skb; union gmac_rxdesc_0 word0; union gmac_rxdesc_1 word1; union gmac_rxdesc_3 word3; @@ -1504,6 +1506,7 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) if (skb) { napi_free_frags(&port->napi); port->stats.rx_dropped++; + skb = NULL; } skb = gmac_skb_if_good_frame(port, word0, frame_len); @@ -1554,6 +1557,7 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) port->stats.rx_dropped++; } + port->rx_skb = skb; writew(r, ptr_reg); return budget; } @@ -1881,6 +1885,7 @@ static int gmac_stop(struct net_device *netdev) gmac_disable_tx_rx(netdev); gmac_stop_dma(port); napi_disable(&port->napi); + port->rx_skb = NULL; gmac_enable_irq(netdev, 0); gmac_cleanup_rxq(netdev); From ebd8ec2b309e3a447851b456ccaf8fb39f3661e7 Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Sat, 9 May 2026 00:13:38 +0200 Subject: [PATCH 286/333] net: ethernet: cortina: Carry over frag counter The gmac_rx() NAPI poll function assembles packets in an SKB from a ring buffer. If the ring buffer gets completely emptied during a poll cycle, we exit gmac_rx(), but the packet is not yet completely assembled in the SKB, yet the fragment counter frag_nr is reset to zero on the next invocation. Solve this by making the RX fragment counter a part of the port struct, and carry it over between invocations. Reset the fragment counter only right after calling napi_gro_frags(), on error (after calling napi_free_frags()) or if stopping the port. Reset it in some place where not strictly necessary just to emphasize what is going on. This was found by Sashiko during normal patch review. Fixes: 4d5ae32f5e1e ("net: ethernet: Add a driver for Gemini gigabit ethernet") Link: https://sashiko.dev/#/patchset/20260505-gemini-ethernet-fix-v2-1-997c31d06079%40kernel.org Signed-off-by: Linus Walleij Link: https://patch.msgid.link/20260509-gemini-ethernet-fixes-v1-3-6c5d20ddc35b@kernel.org Signed-off-by: Paolo Abeni --- drivers/net/ethernet/cortina/gemini.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/cortina/gemini.c b/drivers/net/ethernet/cortina/gemini.c index d5a56366fb43..4c762229ce42 100644 --- a/drivers/net/ethernet/cortina/gemini.c +++ b/drivers/net/ethernet/cortina/gemini.c @@ -123,6 +123,7 @@ struct gemini_ethernet_port { struct hrtimer rx_coalesce_timer; unsigned int rx_coalesce_nsecs; struct sk_buff *rx_skb; + unsigned int rx_frag_nr; unsigned int freeq_refill; struct gmac_txq txq[TX_QUEUE_NUM]; @@ -1444,6 +1445,7 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) unsigned short m = (1 << port->rxq_order) - 1; struct gemini_ethernet *geth = port->geth; void __iomem *ptr_reg = port->rxq_rwptr; + unsigned int frag_nr = port->rx_frag_nr; struct sk_buff *skb = port->rx_skb; unsigned int frame_len, frag_len; struct gmac_rxdesc *rx = NULL; @@ -1457,7 +1459,6 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) unsigned short r, w; union dma_rwptr rw; dma_addr_t mapping; - int frag_nr = 0; spin_lock_irqsave(&geth->irq_lock, flags); rw.bits32 = readl(ptr_reg); @@ -1497,6 +1498,7 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) if (skb) { napi_free_frags(&port->napi); skb = NULL; + frag_nr = 0; } continue; } @@ -1507,6 +1509,7 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) napi_free_frags(&port->napi); port->stats.rx_dropped++; skb = NULL; + frag_nr = 0; } skb = gmac_skb_if_good_frame(port, word0, frame_len); @@ -1541,6 +1544,7 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) if (word3.bits32 & EOF_BIT) { napi_gro_frags(&port->napi); skb = NULL; + frag_nr = 0; --budget; } continue; @@ -1549,6 +1553,7 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) if (skb) { napi_free_frags(&port->napi); skb = NULL; + frag_nr = 0; } if (mapping) @@ -1558,6 +1563,7 @@ static unsigned int gmac_rx(struct net_device *netdev, unsigned int budget) } port->rx_skb = skb; + port->rx_frag_nr = frag_nr; writew(r, ptr_reg); return budget; } @@ -1886,6 +1892,7 @@ static int gmac_stop(struct net_device *netdev) gmac_stop_dma(port); napi_disable(&port->napi); port->rx_skb = NULL; + port->rx_frag_nr = 0; gmac_enable_irq(netdev, 0); gmac_cleanup_rxq(netdev); From 36a8d04a8293afcb9304cf0cd3741f67698f2a1a Mon Sep 17 00:00:00 2001 From: Ethan Nelson-Moore Date: Fri, 8 May 2026 19:37:28 -0700 Subject: [PATCH 287/333] net: ethernet: cs89x0: remove stale CONFIG_MACH_MX31ADS reference The legacy ARM board file for MACH_MX31ADS was removed in commit c93197b0041d ("ARM: imx: Remove i.MX31 board files"), but a reference to it remained in the cs89x0 driver. Drop this unused code. Signed-off-by: Ethan Nelson-Moore Fixes: c93197b0041d ("ARM: imx: Remove i.MX31 board files") Link: https://patch.msgid.link/20260509023732.42256-1-enelsonmoore@gmail.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/cirrus/cs89x0.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/net/ethernet/cirrus/cs89x0.c b/drivers/net/ethernet/cirrus/cs89x0.c index fa5857923db4..b4bfd6c174e7 100644 --- a/drivers/net/ethernet/cirrus/cs89x0.c +++ b/drivers/net/ethernet/cirrus/cs89x0.c @@ -1271,7 +1271,6 @@ static const struct net_device_ops net_ops = { static void __init reset_chip(struct net_device *dev) { -#if !defined(CONFIG_MACH_MX31ADS) struct net_local *lp = netdev_priv(dev); unsigned long reset_start_time; @@ -1298,7 +1297,6 @@ static void __init reset_chip(struct net_device *dev) while ((readreg(dev, PP_SelfST) & INIT_DONE) == 0 && time_before(jiffies, reset_start_time + 2)) ; -#endif /* !CONFIG_MACH_MX31ADS */ } /* This is the real probe routine. From 7cee43fcb0c3f71441d2faaa8c2202b6a88b6bef Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:28:55 -0700 Subject: [PATCH 288/333] net: shaper: flip the polarity of the valid flag The usual way of inserting entries which are not yet fully ready into XArray is to have a VALID flag. The shaper code has a NOT_VALID flag. Since XArray code does not let us create entries with marks already set - the creation of entries is currently not atomic. Flip the polarity of the VALID flag. This closes the tiny race in net_shaper_pre_insert() of entries being created without the NOT_VALID flag. Fixes: 93954b40f6a4 ("net-shapers: implement NL set and delete operations") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-2-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/shaper/shaper.c | 34 +++++++++++++++++----------------- 1 file changed, 17 insertions(+), 17 deletions(-) diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index 1069fa4eb9f6..d2b8f1f951b1 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -275,11 +275,13 @@ static void net_shaper_default_parent(const struct net_shaper_handle *handle, parent->id = 0; } -/* - * MARK_0 is already in use due to XA_FLAGS_ALLOC, can't reuse such flag as - * it's cleared by xa_store(). +/* MARK_0 is already in use due to XA_FLAGS_ALLOC. The VALID mark is set on + * an entry only after the device-side configuration has completed + * successfully (see net_shaper_commit()). Lookups and dumps must filter on + * this mark to avoid exposing tentative entries inserted by + * net_shaper_pre_insert() while the driver call is still in flight. */ -#define NET_SHAPER_NOT_VALID XA_MARK_1 +#define NET_SHAPER_VALID XA_MARK_1 static struct net_shaper * net_shaper_lookup(struct net_shaper_binding *binding, @@ -289,8 +291,8 @@ net_shaper_lookup(struct net_shaper_binding *binding, struct net_shaper_hierarchy *hierarchy; hierarchy = net_shaper_hierarchy_rcu(binding); - if (!hierarchy || xa_get_mark(&hierarchy->shapers, index, - NET_SHAPER_NOT_VALID)) + if (!hierarchy || !xa_get_mark(&hierarchy->shapers, index, + NET_SHAPER_VALID)) return NULL; return xa_load(&hierarchy->shapers, index); @@ -370,13 +372,10 @@ static int net_shaper_pre_insert(struct net_shaper_binding *binding, goto free_id; } - /* Mark 'tentative' shaper inside the hierarchy container. - * xa_set_mark is a no-op if the previous store fails. + /* Insert as 'tentative' (no VALID mark). The mark will be set by + * net_shaper_commit() once the driver-side configuration succeeds. */ - xa_lock(&hierarchy->shapers); - prev = __xa_store(&hierarchy->shapers, index, cur, GFP_KERNEL); - __xa_set_mark(&hierarchy->shapers, index, NET_SHAPER_NOT_VALID); - xa_unlock(&hierarchy->shapers); + prev = xa_store(&hierarchy->shapers, index, cur, GFP_KERNEL); if (xa_err(prev)) { NL_SET_ERR_MSG(extack, "Can't insert shaper into device store"); kfree_rcu(cur, rcu); @@ -413,8 +412,7 @@ static void net_shaper_commit(struct net_shaper_binding *binding, /* Successful update: drop the tentative mark * and update the hierarchy container. */ - __xa_clear_mark(&hierarchy->shapers, index, - NET_SHAPER_NOT_VALID); + __xa_set_mark(&hierarchy->shapers, index, NET_SHAPER_VALID); *cur = shapers[i]; } xa_unlock(&hierarchy->shapers); @@ -431,8 +429,9 @@ static void net_shaper_rollback(struct net_shaper_binding *binding) return; xa_lock(&hierarchy->shapers); - xa_for_each_marked(&hierarchy->shapers, index, cur, - NET_SHAPER_NOT_VALID) { + xa_for_each(&hierarchy->shapers, index, cur) { + if (xa_get_mark(&hierarchy->shapers, index, NET_SHAPER_VALID)) + continue; __xa_erase(&hierarchy->shapers, index); kfree(cur); } @@ -836,7 +835,8 @@ int net_shaper_nl_get_dumpit(struct sk_buff *skb, goto out_unlock; for (; (shaper = xa_find(&hierarchy->shapers, &ctx->start_index, - U32_MAX, XA_PRESENT)); ctx->start_index++) { + U32_MAX, NET_SHAPER_VALID)); + ctx->start_index++) { ret = net_shaper_fill_one(skb, binding, shaper, info); if (ret) break; From 235fb5376139c3419f2218349f1fa2f06f24f7ad Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:28:56 -0700 Subject: [PATCH 289/333] net: shaper: fix trivial ordering issue in net_shaper_commit() We should update the entry before we mark it as valid. Fixes: 93954b40f6a4 ("net-shapers: implement NL set and delete operations") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-3-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/shaper/shaper.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index d2b8f1f951b1..86319ddbf290 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -295,6 +295,10 @@ net_shaper_lookup(struct net_shaper_binding *binding, NET_SHAPER_VALID)) return NULL; + /* Pairs with smp_wmb() in net_shaper_commit(): if the entry is + * valid, its contents must be visible too. + */ + smp_rmb(); return xa_load(&hierarchy->shapers, index); } @@ -412,8 +416,9 @@ static void net_shaper_commit(struct net_shaper_binding *binding, /* Successful update: drop the tentative mark * and update the hierarchy container. */ - __xa_set_mark(&hierarchy->shapers, index, NET_SHAPER_VALID); *cur = shapers[i]; + smp_wmb(); + __xa_set_mark(&hierarchy->shapers, index, NET_SHAPER_VALID); } xa_unlock(&hierarchy->shapers); } @@ -837,6 +842,10 @@ int net_shaper_nl_get_dumpit(struct sk_buff *skb, for (; (shaper = xa_find(&hierarchy->shapers, &ctx->start_index, U32_MAX, NET_SHAPER_VALID)); ctx->start_index++) { + /* Pairs with smp_wmb() in net_shaper_commit(): the entry + * is marked VALID, so its contents must be visible too. + */ + smp_rmb(); ret = net_shaper_fill_one(skb, binding, shaper, info); if (ret) break; From a9a2fa1da619f276580b0d4c5d12efac89e8642b Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:28:57 -0700 Subject: [PATCH 290/333] net: shaper: reject duplicate leaves in GROUP request net_shaper_nl_group_doit() does not deduplicate NET_SHAPER_A_LEAVES entries. When userspace supplies the same leaf handle twice, the same old-parent pointer lands twice in old_nodes[]. The cleanup loop double frees the parent. Of course the same parent may still be in old_nodes[] twice if we are moving multiple of its leaves. Note that this patch also implicitly fixes the fact that the i >= leaves_count path forgets to set ret. Fixes: 5d5d4700e75d ("net-shapers: implement NL group operation") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-4-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/shaper/shaper.c | 60 +++++++++++++++++++++++++++++++++------------ 1 file changed, 45 insertions(+), 15 deletions(-) diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index 86319ddbf290..c8960821cf23 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -941,6 +941,46 @@ static int net_shaper_handle_cmp(const struct net_shaper_handle *a, return memcmp(a, b, sizeof(*a)); } +static int net_shaper_parse_leaves(struct net_shaper_binding *binding, + struct genl_info *info, + const struct net_shaper *node, + struct net_shaper *leaves, + int leaves_count) +{ + struct nlattr *attr; + int i, j, ret, rem; + + i = 0; + nla_for_each_attr_type(attr, NET_SHAPER_A_LEAVES, + genlmsg_data(info->genlhdr), + genlmsg_len(info->genlhdr), rem) { + if (WARN_ON_ONCE(i >= leaves_count)) + return -EINVAL; + + ret = net_shaper_parse_leaf(binding, attr, info, + node, &leaves[i]); + if (ret) + return ret; + + /* Reject duplicates */ + for (j = 0; j < i; j++) { + if (net_shaper_handle_cmp(&leaves[i].handle, + &leaves[j].handle)) + continue; + + NL_SET_ERR_MSG_ATTR_FMT(info->extack, attr, + "Duplicate leaf shaper %d:%d", + leaves[i].handle.scope, + leaves[i].handle.id); + return -EINVAL; + } + + i++; + } + + return 0; +} + static int net_shaper_parent_from_leaves(int leaves_count, const struct net_shaper *leaves, struct net_shaper *node, @@ -1197,10 +1237,9 @@ int net_shaper_nl_group_doit(struct sk_buff *skb, struct genl_info *info) struct net_shaper **old_nodes, *leaves, node = {}; struct net_shaper_hierarchy *hierarchy; struct net_shaper_binding *binding; - int i, ret, rem, leaves_count; + int i, ret, leaves_count; int old_nodes_count = 0; struct sk_buff *msg; - struct nlattr *attr; if (GENL_REQ_ATTR_CHECK(info, NET_SHAPER_A_LEAVES)) return -EINVAL; @@ -1228,19 +1267,10 @@ int net_shaper_nl_group_doit(struct sk_buff *skb, struct genl_info *info) if (ret) goto free_leaves; - i = 0; - nla_for_each_attr_type(attr, NET_SHAPER_A_LEAVES, - genlmsg_data(info->genlhdr), - genlmsg_len(info->genlhdr), rem) { - if (WARN_ON_ONCE(i >= leaves_count)) - goto free_leaves; - - ret = net_shaper_parse_leaf(binding, attr, info, - &node, &leaves[i]); - if (ret) - goto free_leaves; - i++; - } + ret = net_shaper_parse_leaves(binding, info, &node, + leaves, leaves_count); + if (ret) + goto free_leaves; /* Prepare the msg reply in advance, to avoid device operation * rollback on allocation failure. From 6e8ae9d805d4b9ecec49bb9e457d9bae0b21b540 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:28:58 -0700 Subject: [PATCH 291/333] selftests: drv-net: add shaper test for duplicate leaves Add test exercising duplicate leaves. Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-5-kuba@kernel.org Signed-off-by: Paolo Abeni --- tools/testing/selftests/drivers/net/shaper.py | 24 +++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/drivers/net/shaper.py b/tools/testing/selftests/drivers/net/shaper.py index 11310f19bfa0..e39d270e688d 100755 --- a/tools/testing/selftests/drivers/net/shaper.py +++ b/tools/testing/selftests/drivers/net/shaper.py @@ -1,7 +1,10 @@ #!/usr/bin/env python3 # SPDX-License-Identifier: GPL-2.0 -from lib.py import ksft_run, ksft_exit, ksft_eq, ksft_true, KsftSkipEx +import errno + +from lib.py import ksft_run, ksft_exit +from lib.py import ksft_eq, ksft_raises, ksft_true, KsftSkipEx from lib.py import EthtoolFamily, NetshaperFamily from lib.py import NetDrvEnv from lib.py import NlError @@ -438,6 +441,21 @@ def queue_update(cfg, nl_shaper) -> None: nl_shaper.delete({'ifindex': cfg.ifindex, 'handle': {'scope': 'queue', 'id': i}}) +def dup_leaves(cfg, nl_shaper) -> None: + """ Ensure that the kernel rejects duplicate leaves. """ + if not cfg.groups: + raise KsftSkipEx("device does not support node scope") + + with ksft_raises(NlError) as cm: + nl_shaper.group({ + 'ifindex': cfg.ifindex, + 'leaves':[{'handle': {'scope': 'queue', 'id': 0}}, + {'handle': {'scope': 'queue', 'id': 0}}], + 'handle': {'scope':'node'}, + 'metric': 'bps', + 'bw-max': 10000}) + ksft_eq(cm.exception.error, errno.EINVAL) + def main() -> None: with NetDrvEnv(__file__, queue_count=4) as cfg: cfg.queues = False @@ -453,7 +471,9 @@ def main() -> None: basic_groups, qgroups, delegation, - queue_update], args=(cfg, NetshaperFamily())) + dup_leaves, + queue_update], + args=(cfg, NetshaperFamily())) ksft_exit() From 8054f85b83f42a37d482fc77ea7c9ff06a9407d9 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:28:59 -0700 Subject: [PATCH 292/333] net: shaper: set ret to -ENOMEM when genlmsg_new() fails in group_doit genlmsg_new() alloc failure path in net_shaper_nl_group_doit() forgets to set ret before jumping to error handling. Fixes: 5d5d4700e75d ("net-shapers: implement NL group operation") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-6-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/shaper/shaper.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index c8960821cf23..12e5e0c18643 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -1276,8 +1276,10 @@ int net_shaper_nl_group_doit(struct sk_buff *skb, struct genl_info *info) * rollback on allocation failure. */ msg = genlmsg_new(net_shaper_handle_size(), GFP_KERNEL); - if (!msg) + if (!msg) { + ret = -ENOMEM; goto free_leaves; + } hierarchy = net_shaper_hierarchy_setup(binding); if (!hierarchy) { From 0f9a857e34d0f8c018a3e4435c6f0e92e8d2f38c Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:29:00 -0700 Subject: [PATCH 293/333] net: shaper: fix undersized reply skb allocation in GROUP command net_shaper_group_send_reply() writes both the NET_SHAPER_A_IFINDEX attribute (via net_shaper_fill_binding()) and the nested NET_SHAPER_A_HANDLE attribute (via net_shaper_fill_handle()), but the reply skb at the call site in net_shaper_nl_group_doit() is allocated using net_shaper_handle_size(), which only accounts for the nested handle. The allocation is therefore short by nla_total_size(sizeof(u32)) (8 bytes) for the IFINDEX attribute. In practice the slab allocator rounds up the small allocation so the bug is latent, but the size accounting is wrong and could bite if the reply grew further. Introduce net_shaper_group_reply_size() that accounts for the full reply payload and use it both at the genlmsg_new() call site and in the defensive WARN_ONCE message. Fixes: 5d5d4700e75d ("net-shapers: implement NL group operation") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-7-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/shaper/shaper.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index 12e5e0c18643..08fde2d9e8aa 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -90,6 +90,12 @@ static int net_shaper_handle_size(void) nla_total_size(sizeof(u32))); } +static int net_shaper_group_reply_size(void) +{ + return nla_total_size(sizeof(u32)) + /* NET_SHAPER_A_IFINDEX */ + net_shaper_handle_size(); /* NET_SHAPER_A_HANDLE */ +} + static int net_shaper_fill_binding(struct sk_buff *msg, const struct net_shaper_binding *binding, u32 type) @@ -1227,7 +1233,7 @@ static int net_shaper_group_send_reply(struct net_shaper_binding *binding, free_msg: /* Should never happen as msg is pre-allocated with enough space. */ WARN_ONCE(true, "calculated message payload length (%d)", - net_shaper_handle_size()); + net_shaper_group_reply_size()); nlmsg_free(msg); return -EMSGSIZE; } @@ -1275,7 +1281,7 @@ int net_shaper_nl_group_doit(struct sk_buff *skb, struct genl_info *info) /* Prepare the msg reply in advance, to avoid device operation * rollback on allocation failure. */ - msg = genlmsg_new(net_shaper_handle_size(), GFP_KERNEL); + msg = genlmsg_new(net_shaper_group_reply_size(), GFP_KERNEL); if (!msg) { ret = -ENOMEM; goto free_leaves; From fbf5df34a4dbcd09d433dd4f0916bf9b2ddb16de Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:29:01 -0700 Subject: [PATCH 294/333] tools: ynl: add scope qualifier for definitions Using definitions in kernel policies is awkward right now. On one hand we want defines for max values and such. On the other we don't have a way of adding kernel-only defines. Adding unnecessary defines to uAPI is a bad idea, we won't be able to delete them. And when it comes to policy user space should just query it via the policy dump, not use hard coded defines. Add a "scope" property to definitions, which will let us tell the codegen that a definition is for kernel use only. Support following values: - uapi: render into the uAPI header (default, today's behavior) - kernel: render to kernel header only - user: same as kernel but for the user-side generated header Definitions may have a header property (definition is "external", provided by existing header). Extend the scope to headers, too. If definition has both scope and header properties we will only generate the includes in the right scope. Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-8-kuba@kernel.org Signed-off-by: Paolo Abeni --- Documentation/netlink/genetlink-c.yaml | 9 ++++++ Documentation/netlink/genetlink-legacy.yaml | 9 ++++++ Documentation/netlink/genetlink.yaml | 9 ++++++ Documentation/netlink/netlink-raw.yaml | 9 ++++++ tools/net/ynl/pyynl/ynl_gen_c.py | 31 +++++++++++++++++++-- 5 files changed, 65 insertions(+), 2 deletions(-) diff --git a/Documentation/netlink/genetlink-c.yaml b/Documentation/netlink/genetlink-c.yaml index 57f59fe23e3f..4ea31e8fc4d1 100644 --- a/Documentation/netlink/genetlink-c.yaml +++ b/Documentation/netlink/genetlink-c.yaml @@ -69,6 +69,15 @@ properties: header: description: For C-compatible languages, header which already defines this value. type: string + scope: + description: | + Visibility of this definition. "uapi" (default) renders into + the uAPI header, "kernel" renders into the kernel-side + generated header, "user" renders into the user-side + generated header. When combined with `header:`, the + definition is not rendered, and the named header is + included only by code matching the scope. + enum: [ uapi, kernel, user ] type: enum: [ const, enum, flags ] doc: diff --git a/Documentation/netlink/genetlink-legacy.yaml b/Documentation/netlink/genetlink-legacy.yaml index 66fb8653a344..f9c44747729a 100644 --- a/Documentation/netlink/genetlink-legacy.yaml +++ b/Documentation/netlink/genetlink-legacy.yaml @@ -83,6 +83,15 @@ properties: header: description: For C-compatible languages, header which already defines this value. type: string + scope: + description: | + Visibility of this definition. "uapi" (default) renders into + the uAPI header, "kernel" renders into the kernel-side + generated header, "user" renders into the user-side + generated header. When combined with `header:`, the + definition is not rendered, and the named header is + included only by code matching the scope. + enum: [ uapi, kernel, user ] type: enum: [ const, enum, flags, struct ] # Trim doc: diff --git a/Documentation/netlink/genetlink.yaml b/Documentation/netlink/genetlink.yaml index a1194d5d93fc..d3f3f3399ddf 100644 --- a/Documentation/netlink/genetlink.yaml +++ b/Documentation/netlink/genetlink.yaml @@ -55,6 +55,15 @@ properties: header: description: For C-compatible languages, header which already defines this value. type: string + scope: + description: | + Visibility of this definition. "uapi" (default) renders into + the uAPI header, "kernel" renders into the kernel-side + generated header, "user" renders into the user-side + generated header. When combined with `header:`, the + definition is not rendered, and the named header is + included only by code matching the scope. + enum: [ uapi, kernel, user ] type: enum: [ const, enum, flags ] doc: diff --git a/Documentation/netlink/netlink-raw.yaml b/Documentation/netlink/netlink-raw.yaml index dd98dda55bd0..4c436b59a34b 100644 --- a/Documentation/netlink/netlink-raw.yaml +++ b/Documentation/netlink/netlink-raw.yaml @@ -87,6 +87,15 @@ properties: header: description: For C-compatible languages, header which already defines this value. type: string + scope: + description: | + Visibility of this definition. "uapi" (default) renders into + the uAPI header, "kernel" renders into the kernel-side + generated header, "user" renders into the user-side + generated header. When combined with `header:`, the + definition is not rendered, and the named header is + included only by code matching the scope. + enum: [ uapi, kernel, user ] type: enum: [ const, enum, flags, struct ] # Trim doc: diff --git a/tools/net/ynl/pyynl/ynl_gen_c.py b/tools/net/ynl/pyynl/ynl_gen_c.py index 0e1e486c1185..cdc3646f2642 100755 --- a/tools/net/ynl/pyynl/ynl_gen_c.py +++ b/tools/net/ynl/pyynl/ynl_gen_c.py @@ -3212,6 +3212,8 @@ def render_uapi(family, cw): for const in family['definitions']: if const.get('header'): continue + if const.get('scope', 'uapi') != 'uapi': + continue if const['type'] != 'const': cw.writes_defines(defines) @@ -3339,6 +3341,25 @@ def render_uapi(family, cw): cw.p(f'#endif /* {hdr_prot} */') +def render_scoped_consts(family, cw, scope): + defines = [] + for const in family['definitions']: + if const['type'] != 'const': + continue + if const.get('header'): + continue + if const.get('scope') != scope: + continue + name_pfx = const.get('name-prefix', f"{family.ident_name}-") + defines.append([ + c_upper(family.get('c-define-name', + f"{name_pfx}{const['name']}")), + const['value']]) + if defines: + cw.writes_defines(defines) + cw.nl() + + def _render_user_ntf_entry(ri, op): if not ri.family.is_classic(): ri.cw.block_start(line=f"[{op.enum_name}] = ") @@ -3504,8 +3525,12 @@ def main(): cw.p('#include "ynl.h"') headers = [] for definition in parsed['definitions'] + parsed['attribute-sets']: - if 'header' in definition: - headers.append(definition['header']) + if 'header' not in definition: + continue + scope = definition.get('scope', 'uapi') + if scope != 'uapi' and scope != args.mode: + continue + headers.append(definition['header']) if args.mode == 'user': headers.append(parsed.uapi_header) seen_header = [] @@ -3522,6 +3547,7 @@ def main(): for one in args.user_header: cw.p(f'#include "{one}"') else: + render_scoped_consts(parsed, cw, 'user') cw.p('struct ynl_sock;') cw.nl() render_user_family(parsed, cw, True) @@ -3529,6 +3555,7 @@ def main(): if args.mode == "kernel": if args.header: + render_scoped_consts(parsed, cw, 'kernel') for _, struct in sorted(parsed.pure_nested_structs.items()): if struct.request: cw.p('/* Common nested types */') From 8d5806c600fddb907ebe378f9c366d4b52ac3a39 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:29:02 -0700 Subject: [PATCH 295/333] net: shaper: reject handle IDs exceeding internal bit-width net_shaper_parse_handle() reads the user-supplied handle ID via nla_get_u32(), accepting the full u32 range. However, the xarray key is built by net_shaper_handle_to_index() using FIELD_PREP(NET_SHAPER_ID_MASK, handle->id), where NET_SHAPER_ID_MASK is GENMASK(25, 0) - only 26 bits wide. FIELD_PREP silently masks off the upper bits at runtime. A user-supplied NODE id like 0x04000123 becomes id 0x123. Additionally, a user-supplied id equal to NET_SHAPER_ID_UNSPEC (0x03FFFFFF, which is NET_SHAPER_ID_MASK itself) would collide with the sentinel used internally by the group operation to signal "allocate a new NODE id". Reject user-supplied IDs >= NET_SHAPER_ID_MASK (i.e., >= 0x03FFFFFF) in the policy. Fixes: 4b623f9f0f59 ("net-shapers: implement NL get operation") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-9-kuba@kernel.org Signed-off-by: Paolo Abeni --- Documentation/netlink/specs/net_shaper.yaml | 7 +++++++ net/shaper/shaper.c | 4 +++- net/shaper/shaper_nl_gen.c | 7 ++++++- net/shaper/shaper_nl_gen.h | 2 ++ 4 files changed, 18 insertions(+), 2 deletions(-) diff --git a/Documentation/netlink/specs/net_shaper.yaml b/Documentation/netlink/specs/net_shaper.yaml index 3f2ad772b64b..de01f922040a 100644 --- a/Documentation/netlink/specs/net_shaper.yaml +++ b/Documentation/netlink/specs/net_shaper.yaml @@ -33,6 +33,11 @@ doc: | @cap-get operation. definitions: + - + type: const + name: max-handle-id + value: 0x3fffffe + scope: kernel - type: enum name: scope @@ -140,6 +145,8 @@ attribute-sets: - name: id type: u32 + checks: + max: max-handle-id doc: | Numeric identifier of a shaper. The id semantic depends on the scope. For @queue scope it's the queue id and for @node diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index 08fde2d9e8aa..eb049847fed6 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -21,6 +21,8 @@ #define NET_SHAPER_ID_UNSPEC NET_SHAPER_ID_MASK +static_assert(NET_SHAPER_ID_UNSPEC == NET_SHAPER_MAX_HANDLE_ID + 1); + struct net_shaper_hierarchy { struct xarray shapers; }; @@ -360,7 +362,7 @@ static int net_shaper_pre_insert(struct net_shaper_binding *binding, handle->id == NET_SHAPER_ID_UNSPEC) { u32 min, max; - handle->id = NET_SHAPER_ID_MASK - 1; + handle->id = NET_SHAPER_MAX_HANDLE_ID; max = net_shaper_handle_to_index(handle); handle->id = 0; min = net_shaper_handle_to_index(handle); diff --git a/net/shaper/shaper_nl_gen.c b/net/shaper/shaper_nl_gen.c index 9b29be3ef19a..76eff85ec66d 100644 --- a/net/shaper/shaper_nl_gen.c +++ b/net/shaper/shaper_nl_gen.c @@ -11,10 +11,15 @@ #include +/* Integer value ranges */ +static const struct netlink_range_validation net_shaper_a_handle_id_range = { + .max = NET_SHAPER_MAX_HANDLE_ID, +}; + /* Common nested types */ const struct nla_policy net_shaper_handle_nl_policy[NET_SHAPER_A_HANDLE_ID + 1] = { [NET_SHAPER_A_HANDLE_SCOPE] = NLA_POLICY_MAX(NLA_U32, 3), - [NET_SHAPER_A_HANDLE_ID] = { .type = NLA_U32, }, + [NET_SHAPER_A_HANDLE_ID] = NLA_POLICY_FULL_RANGE(NLA_U32, &net_shaper_a_handle_id_range), }; const struct nla_policy net_shaper_leaf_info_nl_policy[NET_SHAPER_A_WEIGHT + 1] = { diff --git a/net/shaper/shaper_nl_gen.h b/net/shaper/shaper_nl_gen.h index 42c46c52c775..2406652a9014 100644 --- a/net/shaper/shaper_nl_gen.h +++ b/net/shaper/shaper_nl_gen.h @@ -12,6 +12,8 @@ #include +#define NET_SHAPER_MAX_HANDLE_ID 67108862 + /* Common nested types */ extern const struct nla_policy net_shaper_handle_nl_policy[NET_SHAPER_A_HANDLE_ID + 1]; extern const struct nla_policy net_shaper_leaf_info_nl_policy[NET_SHAPER_A_WEIGHT + 1]; From b62b29e6de6711f5918940aa6ff2bbab6d6af502 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:29:03 -0700 Subject: [PATCH 296/333] net: shaper: enforce singleton NETDEV scope with id 0 The NETDEV scope represents a singleton root shaper in the per-device hierarchy. All code assumes NETDEV shapers have id 0: net_shaper_default_parent() hardcodes parent->id = 0 when returning the NETDEV parent for QUEUE/NODE children, and the UAPI documentation describes NETDEV scope as "the main shaper" (singular, not plural). Make sure we reject non-0 IDs. Fixes: 4b623f9f0f59 ("net-shapers: implement NL get operation") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-10-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/shaper/shaper.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index eb049847fed6..4ae3ee6764a0 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -482,6 +482,12 @@ static int net_shaper_parse_handle(const struct nlattr *attr, else if (handle->scope == NET_SHAPER_SCOPE_NODE) id = NET_SHAPER_ID_UNSPEC; + if (id && handle->scope == NET_SHAPER_SCOPE_NETDEV) { + NL_SET_ERR_MSG_ATTR(info->extack, id_attr, + "Netdev scope is a singleton, must use ID 0"); + return -EINVAL; + } + handle->id = id; return 0; } From ce372e869f9f492f3d5aa9a0ae75ed52c61d2d6f Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Sun, 10 May 2026 12:29:04 -0700 Subject: [PATCH 297/333] net: shaper: reject QUEUE scope handle with missing id net_shaper_parse_handle() does not enforce that the user provides the handle ID. For NODE the ID defaults to UNSPEC for all other cases it defaults to 0. For NETDEV 0 is the only option. For QUEUE defaulting to 0 makes less intuitive sense. Specifically because the behavior should (IMHO) be the same for all cases where there may be more than one ID (QUEUE and NODE). We should either document this as intentional or reject. I picked the latter with no strong conviction. Fixes: 4b623f9f0f59 ("net-shapers: implement NL get operation") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260510192904.3987113-11-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/shaper/shaper.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/net/shaper/shaper.c b/net/shaper/shaper.c index 4ae3ee6764a0..b1c65110f04d 100644 --- a/net/shaper/shaper.c +++ b/net/shaper/shaper.c @@ -477,10 +477,15 @@ static int net_shaper_parse_handle(const struct nlattr *attr, * shaper (any other value). */ id_attr = tb[NET_SHAPER_A_HANDLE_ID]; - if (id_attr) + if (id_attr) { id = nla_get_u32(id_attr); - else if (handle->scope == NET_SHAPER_SCOPE_NODE) + } else if (handle->scope == NET_SHAPER_SCOPE_NODE) { id = NET_SHAPER_ID_UNSPEC; + } else if (handle->scope == NET_SHAPER_SCOPE_QUEUE) { + NL_SET_ERR_ATTR_MISS(info->extack, attr, + NET_SHAPER_A_HANDLE_ID); + return -EINVAL; + } if (id && handle->scope == NET_SHAPER_SCOPE_NETDEV) { NL_SET_ERR_MSG_ATTR(info->extack, id_attr, From e4a640475e43f406fdfd56d370b1f34b0cbbc18d Mon Sep 17 00:00:00 2001 From: Sergio Correia Date: Tue, 12 May 2026 14:28:33 +0100 Subject: [PATCH 298/333] audit: fix incorrect inheritable capability in CAPSET records __audit_log_capset() records the effective capability set into the inheritable field due to a copy-paste error. Every CAPSET audit record therefore reports cap_pi (process inheritable) with the value of cap_effective instead of cap_inheritable. This silently corrupts audit data used for compliance and forensic analysis: an attacker who modifies inheritable capabilities to prepare for a privilege-escalating exec would have the change masked in the audit trail. The bug has been present since the original introduction of CAPSET audit records in 2008. Cc: stable@vger.kernel.org Fixes: e68b75a027bb ("When the capset syscall is used it is not possible for audit to record the actual capbilities being added/removed. This patch adds a new record type which emits the target pid and the eff, inh, and perm cap sets.") Reviewed-by: Ricardo Robaina Assisted-by: Claude:claude-opus-4-6 Signed-off-by: Sergio Correia Signed-off-by: Paul Moore --- kernel/auditsc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/auditsc.c b/kernel/auditsc.c index ab54fccba215..abdf8da3be93 100644 --- a/kernel/auditsc.c +++ b/kernel/auditsc.c @@ -2786,7 +2786,7 @@ void __audit_log_capset(const struct cred *new, const struct cred *old) context->capset.pid = task_tgid_nr(current); context->capset.cap.effective = new->cap_effective; - context->capset.cap.inheritable = new->cap_effective; + context->capset.cap.inheritable = new->cap_inheritable; context->capset.cap.permitted = new->cap_permitted; context->capset.cap.ambient = new->cap_ambient; context->type = AUDIT_CAPSET; From f9e1c1324b4d98d591a6f7568fdebf5cf456dfc2 Mon Sep 17 00:00:00 2001 From: Sergio Correia Date: Tue, 12 May 2026 14:28:59 +0100 Subject: [PATCH 299/333] audit: enforce AUDIT_LOCKED for AUDIT_TRIM and AUDIT_MAKE_EQUIV AUDIT_ADD_RULE and AUDIT_DEL_RULE correctly check for AUDIT_LOCKED and return -EPERM, but AUDIT_TRIM and AUDIT_MAKE_EQUIV do not. This allows a process with CAP_AUDIT_CONTROL to modify directory tree watches and equivalence mappings even when the audit configuration has been locked, undermining the purpose of the lock. Add AUDIT_LOCKED checks to both commands. Cc: stable@vger.kernel.org Reviewed-by: Ricardo Robaina Assisted-by: Claude:claude-opus-4-6 Signed-off-by: Sergio Correia Signed-off-by: Paul Moore --- kernel/audit.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/kernel/audit.c b/kernel/audit.c index e1d489bc2dff..34dc7cb246ff 100644 --- a/kernel/audit.c +++ b/kernel/audit.c @@ -1468,6 +1468,8 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh, err = audit_list_rules_send(skb, seq); break; case AUDIT_TRIM: + if (audit_enabled == AUDIT_LOCKED) + return -EPERM; audit_trim_trees(); audit_log_common_recv_msg(audit_context(), &ab, AUDIT_CONFIG_CHANGE); @@ -1480,6 +1482,8 @@ static int audit_receive_msg(struct sk_buff *skb, struct nlmsghdr *nlh, size_t msglen = data_len; char *old, *new; + if (audit_enabled == AUDIT_LOCKED) + return -EPERM; err = -EINVAL; if (msglen < 2 * sizeof(u32)) break; From 577a8d3bae0531f0e5ccfac919cd8192f920a804 Mon Sep 17 00:00:00 2001 From: Aaron Sacks Date: Tue, 12 May 2026 02:07:42 -0400 Subject: [PATCH 300/333] KVM: Reject wrapped offset in kvm_reset_dirty_gfn() kvm_reset_dirty_gfn() guards the gfn range with if (!memslot || (offset + __fls(mask)) >= memslot->npages) return; but offset is u64 and the addition is unchecked. The check can be silently bypassed by a u64 wrap. The dirty ring backing those entries is MAP_SHARED at KVM_DIRTY_LOG_PAGE_OFFSET of the vcpu fd, so the VMM can rewrite the slot and offset fields of any entry between when the kernel pushes them and when KVM_RESET_DIRTY_RINGS consumes them. On reset, kvm_dirty_ring_reset() re-reads the values via READ_ONCE() and feeds them straight back into this check; only the flags handshake is treated as the handover, the slot/offset payload is taken on trust. Crafting two entries entry[i].offset = 0xffffffffffffffc1 entry[i+1].offset = 0 makes the coalescing loop in kvm_dirty_ring_reset() compute delta = (s64)(0 - 0xffffffffffffffc1) = 63 which falls in [0, BITS_PER_LONG), so it folds entry[i+1] into the existing mask by setting bit 63. The trailing kvm_reset_dirty_gfn() call then sees offset = 0xffffffffffffffc1 and __fls(mask) = 63; the sum is 0 in u64 and the bounds check passes. That offset propagates into kvm_arch_mmu_enable_log_dirty_pt_masked() unchanged. On the legacy MMU path -- kvm_memslots_have_rmaps() == true, i.e. shadow paging, any VM that has allocated shadow roots, or a write-tracked slot -- it reaches gfn_to_rmap(), which indexes slot->arch.rmap[0][] with a near-U64_MAX gfn. That is an out-of-bounds load of a kvm_rmap_head, followed by a conditional clear of PT_WRITABLE_MASK in whatever the loaded pointer points at. The path is reachable from any process holding /dev/kvm. Range-check offset on its own first, so the addition cannot wrap. memslot->npages is bounded well below U64_MAX, so once offset < npages holds, offset + __fls(mask) (with __fls(mask) < BITS_PER_LONG) stays in range. Fixes: fb04a1eddb1a ("KVM: X86: Implement ring-based dirty memory tracking") Cc: stable@vger.kernel.org Signed-off-by: Aaron Sacks Link: https://patch.msgid.link/20260512060742.1628959-1-contact@xchglabs.com/ Signed-off-by: Paolo Bonzini --- virt/kvm/dirty_ring.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/virt/kvm/dirty_ring.c b/virt/kvm/dirty_ring.c index 02bc6b00d76c..572b854edf74 100644 --- a/virt/kvm/dirty_ring.c +++ b/virt/kvm/dirty_ring.c @@ -63,7 +63,8 @@ static void kvm_reset_dirty_gfn(struct kvm *kvm, u32 slot, u64 offset, u64 mask) memslot = id_to_memslot(__kvm_memslots(kvm, as_id), id); - if (!memslot || (offset + __fls(mask)) >= memslot->npages) + if (!memslot || offset >= memslot->npages || + offset + __fls(mask) >= memslot->npages) return; KVM_MMU_LOCK(kvm); From 2b72f1674e427c56e3772c5ccf785fdda2138820 Mon Sep 17 00:00:00 2001 From: Qiang Ma Date: Tue, 12 May 2026 09:53:13 +0800 Subject: [PATCH 301/333] KVM: x86: Fix Xen hypercall tracepoint argument assignment TRACE_EVENT(kvm_xen_hypercall) stores a5 in __entry->a4 instead of __entry->a5. That overwrites the recorded a4 argument and leaves a5 unset in the trace entry. Fix the typo so both arguments are captured correctly. Signed-off-by: Qiang Ma Link: https://patch.msgid.link/20260512015313.1685784-1-maqianga@uniontech.com/ Cc: stable@vger.kernel.org Signed-off-by: Paolo Bonzini --- arch/x86/kvm/trace.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/kvm/trace.h b/arch/x86/kvm/trace.h index e7fdbe9efc90..0db25bba17f6 100644 --- a/arch/x86/kvm/trace.h +++ b/arch/x86/kvm/trace.h @@ -154,7 +154,7 @@ TRACE_EVENT(kvm_xen_hypercall, __entry->a2 = a2; __entry->a3 = a3; __entry->a4 = a4; - __entry->a4 = a5; + __entry->a5 = a5; ), TP_printk("cpl %d nr 0x%lx a0 0x%lx a1 0x%lx a2 0x%lx a3 0x%lx a4 0x%lx a5 %lx", From 5bd1ddb7911ba7e94b61cf429970963f1b22dd76 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 8 May 2026 14:33:21 -0700 Subject: [PATCH 302/333] KVM: nSVM: Never use L0's PAUSE loop exiting while L2 is running Never use L0's (KVM's) PAUSE loop exiting controls while L2 is running, and instead always configure vmcb02 according to L1's exact capabilities and desires. The purpose of intercepting PAUSE after N attempts is to detect when the vCPU may be stuck waiting on a lock, so that KVM can schedule in a different vCPU that may be holding said lock. Barring a very interesting setup, L1 and L2 do not share locks, and it's extremely unlikely that an L1 vCPU would hold a spinlock while running L2. I.e. having a vCPU executing in L1 yield to a vCPU running in L2 will not allow the L1 vCPU to make forward progress, and vice versa. While teaching KVM's "on spin" logic to only yield to other vCPUs in L2 is doable, in all likelihood it would do more harm than good for most setups. KVM has limited visibility into which L2 "vCPUs" belong to the same VM, and thus share a locking domain. And even if L2 vCPUs are in the same VM, KVM has no visilibity into L2 vCPU's that are scheduled out by the L1 hypervisor. Furthermore, KVM doesn't actually steal PAUSE exits from L1. If L1 is intercepting PAUSE, KVM will route PAUSE exits to L1, not L0, as nested_svm_intercept() gives priority to the vmcb12 intercept. As such, overriding the count/threshold fields in vmcb02 with vmcb01's values is nonsensical, as doing so clobbers all the training/learning that has been done in L1. Even worse, if L1 is not intercepting PAUSE, i.e. KVM is handling PAUSE exits, then KVM will adjust the PLE knobs based on L2 behavior, which could very well be detrimental to L1, e.g. due to essentially poisoning L1 PLE training with bad data. And copying the count from vmcb02 to vmcb01 on a nested VM-Exit makes even less sense, because again, the purpose of PLE is to detect spinning vCPUs. Whether or not a vCPU is spinning in L2 at the time of a nested VM-Exit has no relevance as to the behavior of the vCPU when it executes in L1. The only scenarios where any of this actually works is if at least one of KVM or L1 is NOT intercepting PAUSE for the guest. Per the original changelog, those were the only scenarios considered to be supported. Disabling KVM's use of PLE makes it so the VM is always in a "supported" mode. Last, but certainly not least, using KVM's count/threshold instead of the values provided by L1 is a blatant violation of the SVM architecture. Fixes: 74fd41ed16fd ("KVM: x86: nSVM: support PAUSE filtering when L0 doesn't intercept PAUSE") Cc: Maxim Levitsky Tested-by: David Kaplan Signed-off-by: Sean Christopherson Link: https://patch.msgid.link/20260508213321.373309-1-seanjc@google.com/ Signed-off-by: Paolo Bonzini --- arch/x86/kvm/svm/nested.c | 43 +++++++++++++-------------------------- arch/x86/kvm/svm/svm.c | 15 ++++++++++++-- 2 files changed, 27 insertions(+), 31 deletions(-) diff --git a/arch/x86/kvm/svm/nested.c b/arch/x86/kvm/svm/nested.c index 961804df5f45..b340dc9991ad 100644 --- a/arch/x86/kvm/svm/nested.c +++ b/arch/x86/kvm/svm/nested.c @@ -160,6 +160,16 @@ void nested_vmcb02_recalc_intercepts(struct vcpu_svm *svm) if (!intercept_smi) vmcb_clr_intercept(&vmcb02->control, INTERCEPT_SMI); + /* + * Intercept PAUSE if and only if L1 wants to. KVM intercepts PAUSE so + * that a vCPU that may be spinning waiting for a lock can be scheduled + * out in favor of the vCPU that holds said lock. KVM doesn't support + * yielding across L2 vCPUs, as KVM has limited visilibity into which + * L2 vCPUs are in the same L2 VM, i.e. may be contending for locks. + */ + if (!vmcb12_is_intercept(&svm->nested.ctl, INTERCEPT_PAUSE)) + vmcb_clr_intercept(&vmcb02->control, INTERCEPT_PAUSE); + if (nested_vmcb_needs_vls_intercept(svm)) { /* * If the virtual VMLOAD/VMSAVE is not enabled for the L2, @@ -819,7 +829,6 @@ static void nested_vmcb02_prepare_control(struct vcpu_svm *svm) struct vmcb *vmcb02 = svm->nested.vmcb02.ptr; struct vmcb *vmcb01 = svm->vmcb01.ptr; struct kvm_vcpu *vcpu = &svm->vcpu; - u32 pause_count12, pause_thresh12; nested_svm_transition_tlb_flush(vcpu); @@ -947,31 +956,13 @@ static void nested_vmcb02_prepare_control(struct vcpu_svm *svm) vmcb02->control.misc_ctl2 |= SVM_MISC2_ENABLE_V_VMLOAD_VMSAVE; if (guest_cpu_cap_has(vcpu, X86_FEATURE_PAUSEFILTER)) - pause_count12 = vmcb12_ctrl->pause_filter_count; + vmcb02->control.pause_filter_count = vmcb12_ctrl->pause_filter_count; else - pause_count12 = 0; + vmcb02->control.pause_filter_count = 0; if (guest_cpu_cap_has(vcpu, X86_FEATURE_PFTHRESHOLD)) - pause_thresh12 = vmcb12_ctrl->pause_filter_thresh; + vmcb02->control.pause_filter_thresh = vmcb12_ctrl->pause_filter_thresh; else - pause_thresh12 = 0; - if (kvm_pause_in_guest(svm->vcpu.kvm)) { - /* use guest values since host doesn't intercept PAUSE */ - vmcb02->control.pause_filter_count = pause_count12; - vmcb02->control.pause_filter_thresh = pause_thresh12; - - } else { - /* start from host values otherwise */ - vmcb02->control.pause_filter_count = vmcb01->control.pause_filter_count; - vmcb02->control.pause_filter_thresh = vmcb01->control.pause_filter_thresh; - - /* ... but ensure filtering is disabled if so requested. */ - if (vmcb12_is_intercept(vmcb12_ctrl, INTERCEPT_PAUSE)) { - if (!pause_count12) - vmcb02->control.pause_filter_count = 0; - if (!pause_thresh12) - vmcb02->control.pause_filter_thresh = 0; - } - } + vmcb02->control.pause_filter_thresh = 0; /* * Take ALLOW_LARGER_RAP from vmcb12 even though it should be safe to @@ -1298,12 +1289,6 @@ void nested_svm_vmexit(struct vcpu_svm *svm) /* in case we halted in L2 */ kvm_set_mp_state(vcpu, KVM_MP_STATE_RUNNABLE); - if (!kvm_pause_in_guest(vcpu->kvm)) { - vmcb01->control.pause_filter_count = vmcb02->control.pause_filter_count; - vmcb_mark_dirty(vmcb01, VMCB_INTERCEPTS); - - } - /* * Invalidate last_bus_lock_rip unless KVM is still waiting for the * guest to make forward progress before re-enabling bus lock detection. diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c index e7fdd7a9c280..e02a38da5296 100644 --- a/arch/x86/kvm/svm/svm.c +++ b/arch/x86/kvm/svm/svm.c @@ -913,7 +913,15 @@ static void grow_ple_window(struct kvm_vcpu *vcpu) struct vmcb_control_area *control = &svm->vmcb->control; int old = control->pause_filter_count; - if (kvm_pause_in_guest(vcpu->kvm)) + /* Adjusting pause_filter_count makes no sense if PLE is disabled. */ + WARN_ON_ONCE(kvm_pause_in_guest(vcpu->kvm)); + + /* + * While running L2, KVM should intercept PAUSE if and only if L1 wants + * to intercept PAUSE, and L1's intercept should take priority, i.e. + * KVM should never handle a PAUSE intercept from L2. + */ + if (WARN_ON_ONCE(is_guest_mode(vcpu))) return; control->pause_filter_count = __grow_ple_window(old, @@ -934,7 +942,10 @@ static void shrink_ple_window(struct kvm_vcpu *vcpu) struct vmcb_control_area *control = &svm->vmcb->control; int old = control->pause_filter_count; - if (kvm_pause_in_guest(vcpu->kvm)) + /* Adjusting pause_filter_count makes no sense if PLE is disabled. */ + WARN_ON_ONCE(kvm_pause_in_guest(vcpu->kvm)); + + if (is_guest_mode(vcpu)) return; control->pause_filter_count = From 80f4a7b8ce7513c203562191426e4d4cc635b095 Mon Sep 17 00:00:00 2001 From: Ninad Naik Date: Mon, 11 May 2026 23:13:02 +0530 Subject: [PATCH 303/333] Documentation: kvm: update links in the references section of AMD Memory Encryption Replace non-working links in the reference section with the working ones. Signed-off-by: Ninad Naik Link: https://patch.msgid.link/20260511174302.811918-1-ninadnaik07@gmail.com/ Reviewed-by: Liam Merwick Signed-off-by: Paolo Bonzini --- Documentation/virt/kvm/x86/amd-memory-encryption.rst | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/Documentation/virt/kvm/x86/amd-memory-encryption.rst b/Documentation/virt/kvm/x86/amd-memory-encryption.rst index b2395dd4769d..bd04a908a8db 100644 --- a/Documentation/virt/kvm/x86/amd-memory-encryption.rst +++ b/Documentation/virt/kvm/x86/amd-memory-encryption.rst @@ -656,8 +656,8 @@ References See [white-paper]_, [api-spec]_, [amd-apm]_, [kvm-forum]_, and [snp-fw-abi]_ for more info. -.. [white-paper] https://developer.amd.com/wordpress/media/2013/12/AMD_Memory_Encryption_Whitepaper_v7-Public.pdf -.. [api-spec] https://support.amd.com/TechDocs/55766_SEV-KM_API_Specification.pdf -.. [amd-apm] https://support.amd.com/TechDocs/24593.pdf (section 15.34) +.. [white-paper] https://docs.amd.com/v/u/en-US/memory-encryption-white-paper +.. [api-spec] https://docs.amd.com/v/u/en-US/55766_PUB_3.24_SEV_API +.. [amd-apm] https://docs.amd.com/v/u/en-US/24593_3.44_APM_Vol2 (section 15.34) .. [kvm-forum] https://www.linux-kvm.org/images/7/74/02x08A-Thomas_Lendacky-AMDs_Virtualizatoin_Memory_Encryption_Technology.pdf -.. [snp-fw-abi] https://www.amd.com/system/files/TechDocs/56860.pdf +.. [snp-fw-abi] https://www.amd.com/content/dam/amd/en/documents/developer/56860.pdf From 87c810160ed738cd983e4a65ebe9709927c702c9 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Tue, 12 May 2026 08:56:34 -0700 Subject: [PATCH 304/333] KVM: selftests: Ensure gmem file sizes are multiple of host page size When creating a guest_memfd file and associated memslot to validate shared guest memory, size the file+memslot to the maximum of the host or guest page size. Attempting to allocate a single guest page will fail if the host page size is greater than the guest page size, as KVM requires that the size of memslots and guest_memfd files are a multiple of the host page size. For simplicity, verify the entire file can be shared between guest and host, e.g. instead of trying to validate "partial" mappings. Fixes: 42188667be38 ("KVM: selftests: Add guest_memfd testcase to fault-in on !mmap()'d memory") Reported-by: Zenghui Yu Closes: https://lore.kernel.org/all/0064952b-048c-455d-ad89-e27e5cb82591@linux.dev Signed-off-by: Sean Christopherson Message-ID: <20260512155634.772602-1-seanjc@google.com> Signed-off-by: Paolo Bonzini --- tools/testing/selftests/kvm/guest_memfd_test.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index d6528c6f5e03..253e748c1d4a 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -510,7 +510,12 @@ static void test_guest_memfd_guest(void) "Default VM type should support INIT_SHARED, supported flags = 0x%x", vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS)); - size = vm->page_size; + /* + * Use the max of the host or guest page size for all operations, as + * KVM requires guest_memfd files and memslots to be sized to multiples + * of the host page size. + */ + size = max_t(size_t, vm->page_size, page_size); fd = vm_create_guest_memfd(vm, size, GUEST_MEMFD_FLAG_MMAP | GUEST_MEMFD_FLAG_INIT_SHARED); vm_set_user_memory_region2(vm, slot, KVM_MEM_GUEST_MEMFD, gpa, size, NULL, fd, 0); @@ -519,7 +524,7 @@ static void test_guest_memfd_guest(void) memset(mem, 0xaa, size); kvm_munmap(mem, size); - virt_pg_map(vm, gpa, gpa); + virt_map(vm, gpa, gpa, size / vm->page_size); vcpu_args_set(vcpu, 2, gpa, size); vcpu_run(vcpu); From 6b72d0578ca6f77b835d773d7c77c2f872d3e924 Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Sun, 3 May 2026 23:09:17 +0200 Subject: [PATCH 305/333] KVM: x86: use again the flush argument of __link_shadow_page() Except in the case of parentless nested-TDP pages, mmu_page_zap_pte() clears the SPTE but leaves the invalid_list empty. In this case, using kvm_flush_remote_tlbs() as kvm_mmu_remote_flush_or_zap() does is overkill. Avoid flushing the entirety of the remote TLBs unless the invalid_list was populated: instead, use a more efficient gfn-targeting flush (if available) and skip it altogether if the caller guarantees that a TLB flush is not necessary. Based-on: <20260503201029.106481-1-pbonzini@redhat.com> Signed-off-by: Paolo Bonzini Message-ID: <20260503210917.121840-1-pbonzini@redhat.com> Signed-off-by: Paolo Bonzini --- arch/x86/kvm/mmu/mmu.c | 23 +++++++++++++++++++++-- 1 file changed, 21 insertions(+), 2 deletions(-) diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c index 892246204435..f0144ae8d891 100644 --- a/arch/x86/kvm/mmu/mmu.c +++ b/arch/x86/kvm/mmu/mmu.c @@ -2526,6 +2526,23 @@ static void shadow_walk_next(struct kvm_shadow_walk_iterator *iterator) __shadow_walk_next(iterator, *iterator->sptep); } +/* + * Note: while normally KVM uses a "bool flush" return value to let + * the caller batch flushes, __link_shadow_page() flushes immediately + * before populating the parent PTE with the new shadow page. The + * typical callers, direct_map() and FNAME(fetch)(), are not going + * to zap more than one huge SPTE anyway. + * + * The only exception, where @flush can be false, is when a huge SPTE + * is replaced with a shadow page SPTE with a fully populated page table, + * which can happen from shadow_mmu_split_huge_page(). In this case, + * no memory is unmapped across the change to the page tables and no + * immediate flush is needed for correctness. + * + * Even in that case, calls to kvm_mmu_commit_zap_page() are not + * batched. Doing so would require adding an invalid_list argument + * all the way down to __walk_slot_rmaps(). + */ static void __link_shadow_page(struct kvm *kvm, struct kvm_mmu_memory_cache *cache, u64 *sptep, struct kvm_mmu_page *sp, bool flush) @@ -2541,8 +2558,10 @@ static void __link_shadow_page(struct kvm *kvm, parent_sp = sptep_to_sp(sptep); WARN_ON_ONCE(parent_sp->role.level == PG_LEVEL_4K); - mmu_page_zap_pte(kvm, parent_sp, sptep, &invalid_list); - kvm_mmu_remote_flush_or_zap(kvm, &invalid_list, true); + if (mmu_page_zap_pte(kvm, parent_sp, sptep, &invalid_list)) + kvm_mmu_commit_zap_page(kvm, &invalid_list); + else if (flush) + kvm_flush_remote_tlbs_sptep(kvm, sptep); } spte = make_nonleaf_spte(sp->spt, sp_ad_disabled(sp)); From 3098c076c83ea2913245cb915cdcba98eb24214c Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Wed, 6 May 2026 14:35:14 -0700 Subject: [PATCH 306/333] KVM: x86: Swap the dst and src operand for MOVNTDQA Swap the MOVNTDQA operands, as MOVNTDQA does NOT in fact have "the same characteristics as 0F E7 (MOVNTDQ)"; MOVNTDQA loads from memory and stores to registers, while MOVNTDQ loads from registers and stores to memory. Per the SDM: MOVNTDQ - Move packed integer values in xmm1 to m128 using non-temporal hint. MOVNTDQA - Move double quadword from m128 to xmm1 using non-temporal hint if WC memory type. Reported-by: Josh Eads Fixes: c57d9bafbd0b ("KVM: x86: Add support for emulating MOVNTDQA") Cc: stable@vger.kernel.org Signed-off-by: Sean Christopherson Message-ID: <20260506213514.2781948-1-seanjc@google.com> Signed-off-by: Paolo Bonzini --- arch/x86/kvm/emulate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c index c8c6cc0406d6..8013dccb3110 100644 --- a/arch/x86/kvm/emulate.c +++ b/arch/x86/kvm/emulate.c @@ -4481,7 +4481,7 @@ static const struct opcode opcode_map_0f_38[256] = { X16(N), X16(N), /* 0x20 - 0x2f */ X8(N), - X2(N), GP(SrcReg | DstMem | ModRM | Mov | Aligned, &pfx_0f_e7_0f_38_2a), N, N, N, N, N, + X2(N), GP(SrcMem | DstReg | ModRM | Mov | Aligned, &pfx_0f_e7_0f_38_2a), N, N, N, N, N, /* 0x30 - 0x7f */ X16(N), X16(N), X16(N), X16(N), X16(N), /* 0x80 - 0xef */ From 39e25a2100604320e8d9df54c6c31258f7a3df29 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 12 May 2026 10:30:00 -1000 Subject: [PATCH 307/333] sched_ext: Drop NONE early return in scx_disable_and_exit_task() d3e73a0808dd ("sched_ext: Handle SCX_TASK_NONE in disable/switched_from paths") skipped the trailing scx_set_task_sched(p, NULL) on NONE tasks. After scx_fail_parent() parks a task at NONE/sched=parent and the parent is later freed via queue_rcu_work() during root_disable, the preserved p->scx.sched dangles - print_scx_info() from sched_show_task() reads sch->ops.name from freed memory. Drop the early return. __scx_disable_and_exit_task() already short- circuits on NONE and the SUB_INIT block was cleared by scx_fail_parent()'s earlier call, so clearing p->scx.sched is the only work left - and the one thing the path actually needs. v2: Extend the SUB_INIT block comment to note that the flag is only set on the sub-enable path, so it's always clear on the NONE re-entry (Andrea). Fixes: d3e73a0808dd ("sched_ext: Handle SCX_TASK_NONE in disable/switched_from paths") Signed-off-by: Tejun Heo Reviewed-by: Andrea Righi --- kernel/sched/ext.c | 12 ++---------- 1 file changed, 2 insertions(+), 10 deletions(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 9354da79e162..68120f679178 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -3703,22 +3703,14 @@ static void scx_sub_init_cancel_task(struct scx_sched *sch, struct task_struct * static void scx_disable_and_exit_task(struct scx_sched *sch, struct task_struct *p) { - /* - * %NONE means @p is already detached at the SCX level (e.g. handed - * back to the parent by scx_fail_parent() with no init to undo). - * Skip to avoid clobbering scx_task_sched() and writing %NONE again - * on a state that's already %NONE. - */ - if (scx_get_task_state(p) == SCX_TASK_NONE) - return; - __scx_disable_and_exit_task(sch, p); /* * If set, @p exited between __scx_init_task() and scx_enable_task() in * scx_sub_enable() and is initialized for both the associated sched and * its parent. Exit for the child too - scx_enable_task() never ran for - * it, so undo only init_task. + * it, so undo only init_task. The flag is only set on the sub-enable + * path, so it's always clear when @p arrives here in %SCX_TASK_NONE. */ if (p->scx.flags & SCX_TASK_SUB_INIT) { if (!WARN_ON_ONCE(!scx_enabling_sub_sched)) From b273b75b8d677aea06dd06d80b61b3bb06e94680 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 11 May 2026 13:18:19 -1000 Subject: [PATCH 308/333] sched_ext: INIT_LIST_HEAD() &sch->all in scx_alloc_and_add_sched() On scx_link_sched() error paths (parent disabled, hash insert failure), &sch->all is never added to scx_sched_all. The cleanup path runs scx_unlink_sched() unconditionally, which calls list_del_rcu(&sch->all) on a list_head that was never initialized triggering a corruption warning. Initialize &sch->all. Fixes: 54be8de4236a ("sched_ext: Factor out scx_link_sched() and scx_unlink_sched()") Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 68120f679178..6d69ba29cfd7 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -6635,6 +6635,7 @@ static struct scx_sched *scx_alloc_and_add_sched(struct sched_ext_ops *ops, rcu_assign_pointer(ops->priv, sch); sch->kobj.kset = scx_kset; + INIT_LIST_HEAD(&sch->all); #ifdef CONFIG_EXT_SUB_SCHED char *buf = kzalloc(PATH_MAX, GFP_KERNEL); From cceb874eee46fe4b3d3c6c496f19125d9a3a9a8f Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Mon, 11 May 2026 13:18:23 -1000 Subject: [PATCH 309/333] sched_ext: Defer sub_kset base put to scx_sched_free_rcu_work scx_sub_enable_workfn() pins parent->kobj before dropping scx_sched_lock, but that does not pin parent->sub_kset. Concurrent disable can kset_unregister and free sub_kset before scx_alloc_and_add_sched() dereferences it. Split sub_kset teardown: kobject_del() at disable keeps sysfs removal; defer kobject_put() to scx_sched_free_rcu_work so the memory survives. A racing child sees state_in_sysfs=0 with valid memory, sysfs_create_dir() fails, and the existing exit_kind gate in scx_link_sched() turns it away with -ENOENT. Fixes: 411d3ef1a705 ("sched_ext: Unregister sub_kset on scheduler disable") Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 6d69ba29cfd7..23f7b3f63b09 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -4821,6 +4821,8 @@ static void scx_sched_free_rcu_work(struct work_struct *work) kfree(sch->cgrp_path); if (sch_cgroup(sch)) cgroup_put(sch_cgroup(sch)); + if (sch->sub_kset) + kobject_put(&sch->sub_kset->kobj); #endif /* CONFIG_EXT_SUB_SCHED */ for_each_possible_cpu(cpu) { @@ -5861,7 +5863,7 @@ static void scx_sub_disable(struct scx_sched *sch) if (sch->ops.exit) SCX_CALL_OP(sch, exit, NULL, sch->exit_info); if (sch->sub_kset) - kset_unregister(sch->sub_kset); + kobject_del(&sch->sub_kset->kobj); kobject_del(&sch->kobj); } #else /* CONFIG_EXT_SUB_SCHED */ @@ -5995,7 +5997,7 @@ static void scx_root_disable(struct scx_sched *sch) */ #ifdef CONFIG_EXT_SUB_SCHED if (sch->sub_kset) - kset_unregister(sch->sub_kset); + kobject_del(&sch->sub_kset->kobj); #endif kobject_del(&sch->kobj); From 2c308cf34284420963607d677d576a2b4124d8bd Mon Sep 17 00:00:00 2001 From: Zoran Ilievski Date: Mon, 11 May 2026 08:40:02 +0200 Subject: [PATCH 310/333] net: atlantic: preserve PCI wake-from-D3 on shutdown when WOL enabled The shutdown handler aq_pci_shutdown() unconditionally calls pci_wake_from_d3(pdev, false), clearing the PCI PME_En bit even when wake-on-LAN has been configured. While aq_nic_shutdown() correctly programs the NIC firmware via aq_nic_set_power() to listen for magic packets, the PCI subsystem will not propagate the resulting PME wake event from D3, so the system never wakes after poweroff. WOL from suspend (S3) is unaffected because aq_suspend_common() does not touch pci_wake_from_d3() and relies on the PM core's wake configuration via device_may_wakeup(). This affects all atlantic-supported NICs (AQC107/108/111/112/113); users have reported that WOL works if the atlantic driver is never loaded, but breaks once it has run its shutdown path. Pass the configured WOL state to pci_wake_from_d3() instead of a literal false, so the PCI PME_En bit is preserved when the user has armed WOL via ethtool. Fixes: 90869ddfefeb ("net: aquantia: Implement pci shutdown callback") Cc: stable@vger.kernel.org Signed-off-by: Zoran Ilievski Reviewed-by: Sukhdeep Singh Link: https://patch.msgid.link/20260511064002.1857-1-goodboy@rexbytes.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/aquantia/atlantic/aq_pci_func.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/aquantia/atlantic/aq_pci_func.c b/drivers/net/ethernet/aquantia/atlantic/aq_pci_func.c index e9e38af680c3..39e1b606a75a 100644 --- a/drivers/net/ethernet/aquantia/atlantic/aq_pci_func.c +++ b/drivers/net/ethernet/aquantia/atlantic/aq_pci_func.c @@ -371,7 +371,7 @@ static void aq_pci_shutdown(struct pci_dev *pdev) pci_disable_device(pdev); if (system_state == SYSTEM_POWER_OFF) { - pci_wake_from_d3(pdev, false); + pci_wake_from_d3(pdev, self->aq_hw->aq_nic_cfg->wol); pci_set_power_state(pdev, PCI_D3hot); } } From e3adf69f8eb121a9128c2b0029efd050d3649153 Mon Sep 17 00:00:00 2001 From: David Carlier Date: Sat, 9 May 2026 22:50:46 +0100 Subject: [PATCH 311/333] net: ethtool: phy: avoid NULL deref when PHY driver is unbound phydev->drv can become NULL while the phy_device is still attached to its net_device, namely after the PHY driver is unbound via sysfs: echo > /sys/bus/mdio_bus/drivers//unbind phy_remove() clears phydev->drv but doesn't call phy_detach(), so the phy_device stays in the link topology xarray and ethnl_req_get_phydev() still hands it back. ETHTOOL_MSG_PHY_GET then oopses on: rep_data->drvname = kstrdup(phydev->drv->name, GFP_KERNEL); drvname is already treated as optional by phy_reply_size(), phy_fill_reply() and phy_cleanup_data(), so just skip the allocation when there is no driver bound. Fixes: 9dd2ad5e92b9 ("net: ethtool: phy: Convert the PHY_GET command to generic phy dump") Cc: stable@vger.kernel.org # 6.13.x Signed-off-by: David Carlier Reviewed-by: Maxime Chevallier Tested-by: Maxime Chevallier Link: https://patch.msgid.link/20260509215046.107157-1-devnexen@gmail.com Signed-off-by: Jakub Kicinski --- net/ethtool/phy.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/net/ethtool/phy.c b/net/ethtool/phy.c index f76d94d848d6..ddc6eab701ed 100644 --- a/net/ethtool/phy.c +++ b/net/ethtool/phy.c @@ -94,10 +94,12 @@ static int phy_prepare_data(const struct ethnl_req_info *req_info, if (!rep_data->name) return -ENOMEM; - rep_data->drvname = kstrdup(phydev->drv->name, GFP_KERNEL); - if (!rep_data->drvname) { - ret = -ENOMEM; - goto err_free_name; + if (phydev->drv) { + rep_data->drvname = kstrdup(phydev->drv->name, GFP_KERNEL); + if (!rep_data->drvname) { + ret = -ENOMEM; + goto err_free_name; + } } rep_data->upstream_type = pdn->upstream_type; From f9e2342046ef1560d35bcd4a4b1197648ffd151d Mon Sep 17 00:00:00 2001 From: Wei Yang Date: Sat, 9 May 2026 20:23:58 +0800 Subject: [PATCH 312/333] net: atm: fix skb leak in sigd_send() default branch The default branch in sigd_send() calls sock_put() and returns -EINVAL without freeing the skb, while all other exit paths do so. Add the missing dev_kfree_skb() before sock_put() to fix the leak. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Wei Yang Link: https://patch.msgid.link/20260509122358.1102997-1-albin_yang@163.com Signed-off-by: Jakub Kicinski --- net/atm/signaling.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/atm/signaling.c b/net/atm/signaling.c index 358fbe5e4d1d..b991d937205a 100644 --- a/net/atm/signaling.c +++ b/net/atm/signaling.c @@ -179,6 +179,7 @@ static int sigd_send(struct atm_vcc *vcc, struct sk_buff *skb) break; default: pr_alert("bad message type %d\n", (int)msg->type); + dev_kfree_skb(skb); /* Paired with find_get_vcc(msg->vcc) above */ sock_put(sk); return -EINVAL; From a3fdd924d88c30b9f488636ce0e4696012cf5511 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Nicol=C3=B2=20Coccia?= Date: Sun, 10 May 2026 12:34:13 -0400 Subject: [PATCH 313/333] net/smc: fix sleep-inside-lock in __smc_setsockopt() causing local DoS MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit A logic flaw in __smc_setsockopt() allows a local unprivileged user to cause a Denial of Service (DoS) by holding the socket lock indefinitely. The function __smc_setsockopt() calls copy_from_sockptr() while holding lock_sock(sk). By passing a userfaultfd-monitored memory page (or FUSE-backed memory on systems where unprivileged userfaultfd is disabled) as the optval, an attacker can halt execution during the copy operation, keeping the lock held. Combined with asynchronous tear-down operations like shutdown(), this exhausts the kernel wq (kworkers) and triggers the hung task watchdog. [ 240.123456] INFO: task kworker/u8:2 blocked for more than 120 seconds. [ 240.123489] Call Trace: [ 240.123501] smc_shutdown+... [ 240.123512] lock_sock_nested+... This patch moves the user-space copy outside the lock_sock() critical section to prevent the issue. Fixes: a6a6fe27bab4 ("net/smc: Dynamic control handshake limitation by socket options") Signed-off-by: Nicolò Coccia Reviewed-by: Dust Li Tested-by: Dust Li Signed-off-by: Jakub Kicinski --- net/smc/af_smc.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c index 185dbed7de5d..da28652f6810 100644 --- a/net/smc/af_smc.c +++ b/net/smc/af_smc.c @@ -3054,18 +3054,17 @@ static int __smc_setsockopt(struct socket *sock, int level, int optname, smc = smc_sk(sk); + /* pre-fetch user data outside the lock */ + if (optname == SMC_LIMIT_HS) { + if (optlen < sizeof(int)) + return -EINVAL; + if (copy_from_sockptr(&val, optval, sizeof(int))) + return -EFAULT; + } + lock_sock(sk); switch (optname) { case SMC_LIMIT_HS: - if (optlen < sizeof(int)) { - rc = -EINVAL; - break; - } - if (copy_from_sockptr(&val, optval, sizeof(int))) { - rc = -EFAULT; - break; - } - smc->limit_smc_hs = !!val; rc = 0; break; From 7bf563badd37cb796df5477d2b78bb64148a1268 Mon Sep 17 00:00:00 2001 From: Xiang Mei Date: Sun, 10 May 2026 15:26:40 -0700 Subject: [PATCH 314/333] net/smc: avoid NULL deref of conn->lnk in smc_msg_event tracepoint The smc_msg_event tracepoint class, shared by smc_tx_sendmsg and smc_rx_recvmsg, unconditionally dereferences smc->conn.lnk: __string(name, smc->conn.lnk->ibname) conn->lnk is only set for SMC-R; for SMC-D it is NULL. Other code on these paths already handles this (e.g. !conn->lnk in SMC_STAT_RMB_TX_SIZE_SMALL()). With the tracepoint enabled, the first sendmsg()/recvmsg() on an SMC-D socket crashes: Oops: general protection fault, probably for non-canonical address KASAN: null-ptr-deref in range [...] RIP: 0010:strlen+0x1e/0xa0 Call Trace: trace_event_raw_event_smc_msg_event (net/smc/smc_tracepoint.h:44) smc_rx_recvmsg (net/smc/smc_rx.c:515) smc_recvmsg (net/smc/af_smc.c:2859) __sys_recvfrom (net/socket.c:2315) __x64_sys_recvfrom (net/socket.c:2326) do_syscall_64 The faulting address 0x3e0 is offsetof(struct smc_link, ibname), confirming the NULL ->lnk deref. Enabling the tracepoint requires root, but the trigger itself is unprivileged: socket(AF_SMC, ...) has no capability check, and SMC-D negotiation needs no admin step on s390 or on x86 with the loopback ISM device loaded. Log an empty device name for SMC-D instead of dereferencing NULL. Fixes: aff3083f10bf ("net/smc: Introduce tracepoints for tx and rx msg") Reported-by: Weiming Shi Signed-off-by: Xiang Mei Reviewed-by: Dust Li Reviewed-by: Sidraya Jayagond Signed-off-by: Jakub Kicinski --- net/smc/smc_tracepoint.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/smc/smc_tracepoint.h b/net/smc/smc_tracepoint.h index a9a6e3c1113a..53da84f57fd6 100644 --- a/net/smc/smc_tracepoint.h +++ b/net/smc/smc_tracepoint.h @@ -51,7 +51,7 @@ DECLARE_EVENT_CLASS(smc_msg_event, __field(const void *, smc) __field(u64, net_cookie) __field(size_t, len) - __string(name, smc->conn.lnk->ibname) + __string(name, smc->conn.lnk ? smc->conn.lnk->ibname : "") ), TP_fast_assign( From 3d042592ebd4c7e44974d556de0b727cb7db4dab Mon Sep 17 00:00:00 2001 From: Chenguang Zhao Date: Mon, 11 May 2026 09:43:43 +0800 Subject: [PATCH 315/333] ethtool: fix ethnl_bitmap32_not_zero() bit interval semantics ethnl_bitmap32_not_zero() should return true if some bit in [start, end) is set: - Fix inverted memchr_inv() sense: return true when the scan finds a non-zero byte, not when the middle words are all zero. - Return false for an empty interval (end <= start). - When end is 32-bit aligned, indices in [start, end) do not include any bits from map[end_word]; return false after earlier checks found no non-zero data. Fixes: 10b518d4e6dd ("ethtool: netlink bitset handling") Signed-off-by: Chenguang Zhao Signed-off-by: Jakub Kicinski --- net/ethtool/bitset.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/net/ethtool/bitset.c b/net/ethtool/bitset.c index 8bb98d3ea3db..a3a2cc6480a0 100644 --- a/net/ethtool/bitset.c +++ b/net/ethtool/bitset.c @@ -92,7 +92,7 @@ static bool ethnl_bitmap32_not_zero(const u32 *map, unsigned int start, u32 mask; if (end <= start) - return true; + return false; if (start % 32) { mask = ethnl_upper_bits(start); @@ -105,11 +105,11 @@ static bool ethnl_bitmap32_not_zero(const u32 *map, unsigned int start, start_word++; } - if (!memchr_inv(map + start_word, '\0', - (end_word - start_word) * sizeof(u32))) + if (memchr_inv(map + start_word, '\0', + (end_word - start_word) * sizeof(u32))) return true; if (end % 32 == 0) - return true; + return false; return map[end_word] & ethnl_lower_bits(end); } From f5b2772d14884f4be9e718644f1203d4d0e6f0d6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Niklas=20S=C3=B6derlund?= Date: Sun, 10 May 2026 12:30:17 +0200 Subject: [PATCH 316/333] net: ethernet: ravb: Do not check URAM suspension when WoL is active MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When updating the driver to match latest datasheet to suspend access to URAM when suspending DMA transfers a corner-case was missed, URAM access will not be suspended if WoL is enabled. This lead to the error message (correctly) being triggered as URAM access is not suspended even tho it's requested as part of stopping DMA. Avoid checking if URAM access is suspended and printing the error message if WoL is enabled when we suspend the system, as we know it will not be. Reported-by: Geert Uytterhoeven Closes: https://lore.kernel.org/all/CAMuHMdWnjV%3DHGE1o08zLhUfTgOSene5fYx1J5GG10mB%2BToq8qg@mail.gmail.com/ Fixes: 353d8e7989b6 ("net: ethernet: ravb: Suspend and resume the transmission flow") Signed-off-by: Niklas Söderlund Reviewed-by: Sai Krishna Tested-by: Geert Uytterhoeven Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/renesas/ravb_main.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/renesas/ravb_main.c b/drivers/net/ethernet/renesas/ravb_main.c index 1dbfadb2a881..5f88733094d0 100644 --- a/drivers/net/ethernet/renesas/ravb_main.c +++ b/drivers/net/ethernet/renesas/ravb_main.c @@ -1108,9 +1108,12 @@ static int ravb_stop_dma(struct net_device *ndev) /* Request for transmission suspension */ ravb_modify(ndev, CCC, CCC_DTSR, CCC_DTSR); - error = ravb_wait(ndev, CSR, CSR_DTS, CSR_DTS); - if (error) - netdev_err(ndev, "failed to stop AXI BUS\n"); + /* Access to URAM will not be suspended if WoL is enabled. */ + if (!priv->wol_enabled) { + error = ravb_wait(ndev, CSR, CSR_DTS, CSR_DTS); + if (error) + netdev_err(ndev, "failed to stop AXI BUS\n"); + } /* Stop AVB-DMAC process */ return ravb_set_opmode(ndev, CCC_OPC_CONFIG); From 2d5d3fc593c9b7e41bee86175d7b9e11f470072e Mon Sep 17 00:00:00 2001 From: Paolo Bonzini Date: Tue, 12 May 2026 16:58:48 +0200 Subject: [PATCH 317/333] KVM: VMX: introduce module parameter to disable CET There have been reports of host hangs caused by CET virtualization. Until these are analyzed further, introduce a module parameter that makes it possible to easily disable it. Link: https://lore.kernel.org/all/85548beb-1486-40f9-beb4-632c78e3360b@proxmox.com/ Cc: David Riley Signed-off-by: Paolo Bonzini --- arch/x86/kvm/vmx/capabilities.h | 1 + arch/x86/kvm/vmx/vmx.c | 17 +++++++++++++++-- 2 files changed, 16 insertions(+), 2 deletions(-) diff --git a/arch/x86/kvm/vmx/capabilities.h b/arch/x86/kvm/vmx/capabilities.h index 56cacc06225e..31568274d8bb 100644 --- a/arch/x86/kvm/vmx/capabilities.h +++ b/arch/x86/kvm/vmx/capabilities.h @@ -14,6 +14,7 @@ extern bool __read_mostly flexpriority_enabled; extern bool __read_mostly enable_ept; extern bool __read_mostly enable_unrestricted_guest; extern bool __read_mostly enable_ept_ad_bits; +extern bool __read_mostly enable_cet; extern bool __read_mostly enable_pml; extern int __read_mostly pt_mode; diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c index 5c2c33a5f7dc..49feecb286b2 100644 --- a/arch/x86/kvm/vmx/vmx.c +++ b/arch/x86/kvm/vmx/vmx.c @@ -108,6 +108,9 @@ module_param_named(unrestricted_guest, bool __read_mostly enable_ept_ad_bits = 1; module_param_named(eptad, enable_ept_ad_bits, bool, 0444); +bool __read_mostly enable_cet = 1; +module_param_named(cet, enable_cet, bool, 0444); + static bool __read_mostly emulate_invalid_guest_state = true; module_param(emulate_invalid_guest_state, bool, 0444); @@ -4476,7 +4479,7 @@ void vmx_set_constant_host_state(struct vcpu_vmx *vmx) * SSP is reloaded from IA32_PL3_SSP. Check SDM Vol.2A/B Chapter * 3 and 4 for details. */ - if (cpu_has_load_cet_ctrl()) { + if (enable_cet) { vmcs_writel(HOST_S_CET, kvm_host.s_cet); vmcs_writel(HOST_SSP, 0); vmcs_writel(HOST_INTR_SSP_TABLE, 0); @@ -4532,6 +4535,10 @@ static u32 vmx_get_initial_vmentry_ctrl(void) if (vmx_pt_mode_is_system()) vmentry_ctrl &= ~(VM_ENTRY_PT_CONCEAL_PIP | VM_ENTRY_LOAD_IA32_RTIT_CTL); + + if (!enable_cet) + vmentry_ctrl &= ~VM_ENTRY_LOAD_CET_STATE; + /* * IA32e mode, and loading of EFER and PERF_GLOBAL_CTRL are toggled dynamically. */ @@ -4546,6 +4553,9 @@ static u32 vmx_get_initial_vmexit_ctrl(void) { u32 vmexit_ctrl = vmcs_config.vmexit_ctrl; + if (!enable_cet) + vmexit_ctrl &= ~VM_EXIT_LOAD_CET_STATE; + /* * Not used by KVM and never set in vmcs01 or vmcs02, but emulated for * nested virtualization and thus allowed to be set in vmcs12. @@ -8155,7 +8165,7 @@ static __init void vmx_set_cpu_caps(void) * VMX_BASIC[bit56] == 0, inject #CP at VMX entry with error code * fails, so disable CET in this case too. */ - if (!cpu_has_load_cet_ctrl() || !enable_unrestricted_guest || + if (!enable_cet || !enable_unrestricted_guest || !cpu_has_vmx_basic_no_hw_errcode_cc()) { kvm_cpu_cap_clear(X86_FEATURE_SHSTK); kvm_cpu_cap_clear(X86_FEATURE_IBT); @@ -8630,6 +8640,9 @@ __init int vmx_hardware_setup(void) !cpu_has_vmx_invept_global()) enable_ept = 0; + if (!cpu_has_load_cet_ctrl()) + enable_cet = 0; + /* NX support is required for shadow paging. */ if (!enable_ept && !boot_cpu_has(X86_FEATURE_NX)) { pr_err_ratelimited("NX (Execute Disable) not supported\n"); From 7d8f3158a51cb40fc710d2a781549141a139b796 Mon Sep 17 00:00:00 2001 From: Yu Miao Date: Wed, 13 May 2026 10:39:07 +0800 Subject: [PATCH 318/333] selftests/cgroup: Fix error path leaks in test_percpu_basic When cg_name_indexed() returns NULL partway through the child creation loop, the code returned -1 without running cleanup_children and cleanup. That left the `parent` pathname allocation unreleased and did not remove child cgroup directories already created under the parent. Fix by jumping to cleanup_children instead of returning. When cg_create() fails, `child` (the pathname from cg_name_indexed()) was not freed before cleanup_children. Fix by freeing `child` before branching to cleanup_children. Fixes: 90631e1dea55 ("kselftests: cgroup: add perpcu memory accounting test") Signed-off-by: Yu Miao Signed-off-by: Tejun Heo --- tools/testing/selftests/cgroup/test_kmem.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/tools/testing/selftests/cgroup/test_kmem.c b/tools/testing/selftests/cgroup/test_kmem.c index eeabd34bf083..12f59925500b 100644 --- a/tools/testing/selftests/cgroup/test_kmem.c +++ b/tools/testing/selftests/cgroup/test_kmem.c @@ -368,11 +368,15 @@ static int test_percpu_basic(const char *root) for (i = 0; i < 1000; i++) { child = cg_name_indexed(parent, "child", i); - if (!child) - return -1; - - if (cg_create(child)) + if (!child) { + ret = -1; goto cleanup_children; + } + + if (cg_create(child)) { + free(child); + goto cleanup_children; + } free(child); } From 345f40166694e60db6d5cf02233814bb27ac5dec Mon Sep 17 00:00:00 2001 From: sunshaojie Date: Wed, 13 May 2026 18:37:38 +0800 Subject: [PATCH 319/333] cgroup/cpuset: Return only actually allocated CPUs during partition invalidation In update_parent_effective_cpumask() with partcmd_invalidate, the CPUs to return to the parent are computed as: adding = cpumask_and(tmp->addmask, xcpus, parent->effective_xcpus); where xcpus = user_xcpus(cs) which returns cs->exclusive_cpus (if set) or cs->cpus_allowed. When exclusive_cpus is not set, user_xcpus(cs) can contain CPUs that were never actually granted to the partition due to sibling exclusion in compute_excpus(). Consequently, the invalidation may return CPUs to the parent that remain in use by sibling partitions, causing overlapping effective_cpus and triggering the WARN_ON_ONCE(1) in generate_sched_domains(). Use cs->effective_xcpus instead, which reflects the CPUs actually granted to this partition. Reproducer (on a 4-CPU machine): cd /sys/fs/cgroup mkdir a1 b1 # a1 becomes partition root with CPUs 0-1 echo "0-1" > a1/cpuset.cpus echo "root" > a1/cpuset.cpus.partition # b1 becomes partition root with CPUs 1-2, but sibling exclusion # reduces its effective_xcpus to CPU 2 only echo "1-2" > b1/cpuset.cpus echo "root" > b1/cpuset.cpus.partition # b1 changes cpus_allowed to 0-1 -> partition invalidation echo "0-1" > b1/cpuset.cpus # Expected: CPUs 2-3 (only CPU 2 returned from b1) # Actual: CPUs 1-3 (CPU 0-1 returned, overlapping with a1) cat cpuset.cpus.effective dmesg will also show a WARNING from generate_sched_domains() reporting overlapping partition root effective_cpus. Fixes: 2a3602030d80 ("cgroup/cpuset: Don't invalidate sibling partitions on cpuset.cpus conflict") Cc: stable@vger.kernel.org # v7.0+ Signed-off-by: sunshaojie Tested-by: Chen Ridong Reviewed-by: Chen Ridong Reviewed-by: Waiman Long Signed-off-by: Tejun Heo --- kernel/cgroup/cpuset.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c index e84e801e22cf..5c33ab20cc20 100644 --- a/kernel/cgroup/cpuset.c +++ b/kernel/cgroup/cpuset.c @@ -1718,7 +1718,8 @@ static int update_parent_effective_cpumask(struct cpuset *cs, int cmd, */ if (is_partition_valid(parent)) adding = cpumask_and(tmp->addmask, - xcpus, parent->effective_xcpus); + cs->effective_xcpus, + parent->effective_xcpus); if (old_prs > 0) new_prs = -old_prs; From 6ae315d37924435516d697ea7dde0b799a5928e0 Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Wed, 13 May 2026 13:24:38 +0200 Subject: [PATCH 320/333] sched_ext: Use HK_TYPE_DOMAIN_BOOT to detect isolcpus= domain isolation scx_enable() refuses to attach a BPF scheduler when isolcpus=domain is in effect by comparing housekeeping_cpumask(HK_TYPE_DOMAIN) against cpu_possible_mask. Since commit 27c3a5967f05 ("sched/isolation: Convert housekeeping cpumasks to rcu pointers"), HK_TYPE_DOMAIN's cpumask is RCU protected and dereferencing it requires either RCU read lock, the cpu_hotplug write lock, or the cpuset lock; scx_enable() holds none of these, so booting with isolcpus=domain and attaching any BPF scheduler triggers the following lockdep splat: ============================= WARNING: suspicious RCU usage ----------------------------- kernel/sched/isolation.c:60 suspicious rcu_dereference_check() usage! 1 lock held by scx_flash/281: #0: ffffffff8379fce0 (update_mutex){+.+.}-{4:4}, at: bpf_struct_ops_link_create+0x134/0x1c0 Call Trace: dump_stack_lvl+0x6f/0xb0 lockdep_rcu_suspicious.cold+0x37/0x70 housekeeping_cpumask+0xcd/0xe0 scx_enable.isra.0+0x17/0x120 bpf_scx_reg+0x5e/0x80 bpf_struct_ops_link_create+0x151/0x1c0 __sys_bpf+0x1e4b/0x33c0 __x64_sys_bpf+0x21/0x30 do_syscall_64+0x117/0xf80 entry_SYSCALL_64_after_hwframe+0x77/0x7f In addition, commit 03ff73510169 ("cpuset: Update HK_TYPE_DOMAIN cpumask from cpuset") made HK_TYPE_DOMAIN include cpuset isolated partitions as well, which means the current check also rejects BPF schedulers when a cpuset partition is active. That contradicts the original intent of commit 9f391f94a173 ("sched_ext: Disallow loading BPF scheduler if isolcpus= domain isolation is in effect"), which explicitly noted that cpuset partitions are honored through per-task cpumasks and should not be rejected. Switch to housekeeping_enabled(HK_TYPE_DOMAIN_BOOT), which reads only the housekeeping flag bit (no RCU dereference) and reflects exactly the boot-time isolcpus= configuration that the error message refers to. Fixes: 27c3a5967f05 ("sched/isolation: Convert housekeeping cpumasks to rcu pointers") Cc: stable@vger.kernel.org # v7.0+ Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo Acked-by: Frederic Weisbecker --- kernel/sched/ext.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 23f7b3f63b09..a6d0a93d8174 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -7415,8 +7415,7 @@ static s32 scx_enable(struct sched_ext_ops *ops, struct bpf_link *link) static DEFINE_MUTEX(helper_mutex); struct scx_enable_cmd cmd; - if (!cpumask_equal(housekeeping_cpumask(HK_TYPE_DOMAIN), - cpu_possible_mask)) { + if (housekeeping_enabled(HK_TYPE_DOMAIN_BOOT)) { pr_err("sched_ext: Not compatible with \"isolcpus=\" domain isolation\n"); return -EINVAL; } From 4694efc4164123580f19467141cdcfb73f7a740a Mon Sep 17 00:00:00 2001 From: "Maciej W. Rozycki" Date: Sat, 9 May 2026 22:04:50 +0100 Subject: [PATCH 321/333] FDDI: defza: Sanitise the reset safety timer The reset actions of the DEFZA adapters are exceedingly slow, taking up to 30 seconds to complete by the device spec and typically in the range of 10 seconds in reality, as required for the device RTOS to boot, still quite a lot. Therefore a state machine is used that's interrupt driven, however a safety mechanism is required in case of adapter malfunction, so that if no state change interrupt has arrived in time, then the situation is taken care of. The safety mechanism depends on the origin of the reset. For regular adapter initialisation at the device probe time a sleep is requested. However a reset is also required by the device spec when the adapter has transitioned into the halted state, such as in response to a PC Trace event in the course of ring fault recovery, possibly a common network event. In that case no sleep is possible as a device halt is reported at the hardirq level. A timer is therefore set up to ensure progress in case no adapter state change interrupt has arrived in time, but as from commit 168f6b6ffbee ("timers: Use del_timer_sync() even on UP") a warning is issued as the timer is deleted in the hardirq handler upon an expected state change: defza: v.1.1.4 Oct 6 2018 Maciej W. Rozycki tc2: DEC FDDIcontroller 700 or 700-C at 0x18000000, irq 4 tc2: resetting the board... ------------[ cut here ]------------ WARNING: kernel/time/timer.c:1611 at __timer_delete_sync+0x104/0x120, CPU#0: swapper/0/0 Modules linked in: CPU: 0 UID: 0 PID: 0 Comm: swapper/0 Not tainted 7.0.0-dirty #2 VOLUNTARY Stack : 9800000002027d08 00000000140120e0 0000000000000000 ffffffff8089d468 0000000000000000 0000000000000000 ffffffff807ed6b8 ffffffff80897458 ffffffff80897400 9800000002027b88 0000000000000000 7070617773203a6d 0000000000000000 9800000002027ba4 0000000000001000 6465746e69617420 0000000000000000 ffffffff807ed6b8 00000000140120e0 0000000000000009 000000000000064b ffffffff800dd14c 0000000000000036 9800000002184000 0000000000000000 0000000000000020 0000000000000000 ffffffff80910000 ffffffff8085c000 9800000002027c70 0000000000000001 ffffffff80045fa0 0000000000000000 0000000000000000 0000000000000000 0000000000000009 000000000000064b ffffffff800502b8 ffffffff807ed6b8 ffffffff80045fa0 ... Call Trace: [] show_stack+0x28/0xf0 [] dump_stack_lvl+0x48/0x7c [] __warn+0xa0/0x128 [] warn_slowpath_fmt+0x64/0xa4 [] __timer_delete_sync+0x104/0x120 [] fza_interrupt+0xc74/0xeb8 [] __handle_irq_event_percpu+0x70/0x228 [] handle_irq_event_percpu+0x18/0x78 [] handle_percpu_irq+0x50/0x80 [] generic_handle_irq+0x90/0xd0 [] do_IRQ+0x1c/0x30 [] handle_int+0x148/0x154 [] do_idle+0x40/0x108 [] cpu_startup_entry+0x2c/0x38 [] kernel_init+0x0/0x108 ---[ end trace 0000000000000000 ]--- tc2: OK tc2: model 700 (DEFZA-AA), MMF PMD, address 08-00-2b-xx-xx-xx tc2: ROM rev. 1.0, firmware rev. 1.2, RMC rev. A, SMT ver. 1 tc2: link unavailable ------------[ cut here ]------------ WARNING: kernel/time/timer.c:1611 at __timer_delete_sync+0x104/0x120, CPU#0: swapper/0/0 Modules linked in: CPU: 0 UID: 0 PID: 0 Comm: swapper/0 Tainted: G W 7.0.0-dirty #2 VOLUNTARY Tainted: [W]=WARN Stack : 9800000002027d08 00000000140120e0 0000000000000000 ffffffff8089d468 0000000000000000 0000000000000000 ffffffff807ed6b8 ffffffff80897458 ffffffff80897400 9800000002027b88 0000000000000000 0000000000000000 0000000000000000 9800000002027ba4 0000000000001000 0000000000000000 0000000000000000 ffffffff807ed6b8 00000000140120e0 0000000000000009 000000000000064b ffffffff800dd14c 0000000000000036 9800000002184000 0000000000000000 0000000000000020 0000000000000000 ffffffff80910000 ffffffff8085c000 9800000002027c70 0000000000000001 ffffffff80045fa0 0000000000000000 0000000000000000 0000000000000000 0000000000000009 000000000000064b ffffffff800502b8 ffffffff807ed6b8 ffffffff80045fa0 ... Call Trace: [] show_stack+0x28/0xf0 [] dump_stack_lvl+0x48/0x7c [] __warn+0xa0/0x128 [] warn_slowpath_fmt+0x64/0xa4 [] __timer_delete_sync+0x104/0x120 [] fza_interrupt+0xc74/0xeb8 [] __handle_irq_event_percpu+0x70/0x228 [] handle_irq_event_percpu+0x18/0x78 [] handle_percpu_irq+0x50/0x80 [] generic_handle_irq+0x90/0xd0 [] do_IRQ+0x1c/0x30 [] handle_int+0x148/0x154 [] arch_local_irq_disable+0x4/0x28 [] do_idle+0x50/0x108 [] cpu_startup_entry+0x2c/0x38 [] kernel_init+0x0/0x108 ---[ end trace 0000000000000000 ]--- tc2: registered as fddi0 The immediate origin of the new warning is the switch away from aliasing del_timer_sync() to del_timer() (timer_delete_sync() to timer_delete() in terms of current function names) for UP configurations, which however is the only choice for this driver anyway as no SMP hardware supports the TURBOchannel bus this device interfaces to. Therefore there is a very remote issue only this is a sign of. Specifically if an adapter reset issued upon a transition to the halted state times out and first triggers fza_reset_timer() for another reset assertion, which then schedules fza_reset_timer() for reset deassertion and then that second call is pre-empted after poking at the hardware, but before the timer has been rearmed and owing to high system load causing exceedingly high scheduling latency control is not handed back before a transition to the uninitialised state has caused the timer to be deleted even before it has been started, then fza_reset_timer() will be called yet again and issue another reset even though by then the adapter has already recovered. Prevent this situation from happening by switching to timer_delete() for the transition to the halted state and protect the code region affected with a spinlock, also to make sure add_timer() has not been called twice in a row due to an execution race between the interrupt handler and the timer handler (though it could only happen on SMP, but let's keep the driver clean). It's a very unlikely sequence of events to happen and therefore there's no point in trying to be overly clever about it, such as by placing printk() calls outside the protection. For the transition to the uninitialised state switch to timer_delete_sync_try() instead, so that a timer isn't deleted that's just been rearmed by the timer handler and needs to watch for the device to come out of reset again (again, an SMP scenario only). Retain timer_delete_sync() invocations outside the hardirq context for a stray timer not to fire once device structures have been released. Fixes: 61414f5ec9834 ("FDDI: defza: Add support for DEC FDDIcontroller 700 TURBOchannel adapter") Signed-off-by: Maciej W. Rozycki Signed-off-by: Jakub Kicinski --- drivers/net/fddi/defza.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/drivers/net/fddi/defza.c b/drivers/net/fddi/defza.c index 064fa484f797..9bfecc87d6b2 100644 --- a/drivers/net/fddi/defza.c +++ b/drivers/net/fddi/defza.c @@ -984,7 +984,7 @@ static irqreturn_t fza_interrupt(int irq, void *dev_id) case FZA_STATE_UNINITIALIZED: netif_carrier_off(dev); - timer_delete_sync(&fp->reset_timer); + timer_delete_sync_try(&fp->reset_timer); fp->ring_cmd_index = 0; fp->ring_uns_index = 0; fp->ring_rmc_tx_index = 0; @@ -1018,7 +1018,9 @@ static irqreturn_t fza_interrupt(int irq, void *dev_id) fp->queue_active = 0; netif_stop_queue(dev); pr_debug("%s: queue stopped\n", fp->name); - timer_delete_sync(&fp->reset_timer); + + spin_lock(&fp->lock); + timer_delete(&fp->reset_timer); pr_warn("%s: halted, reason: %x\n", fp->name, FZA_STATUS_GET_HALT(status)); fza_regs_dump(fp); @@ -1027,6 +1029,8 @@ static irqreturn_t fza_interrupt(int irq, void *dev_id) fp->timer_state = 0; fp->reset_timer.expires = jiffies + 45 * HZ; add_timer(&fp->reset_timer); + spin_unlock(&fp->lock); + break; default: @@ -1046,7 +1050,9 @@ static irqreturn_t fza_interrupt(int irq, void *dev_id) static void fza_reset_timer(struct timer_list *t) { struct fza_private *fp = timer_container_of(fp, t, reset_timer); + unsigned long flags; + spin_lock_irqsave(&fp->lock, flags); if (!fp->timer_state) { pr_err("%s: RESET timed out!\n", fp->name); pr_info("%s: trying harder...\n", fp->name); @@ -1069,6 +1075,7 @@ static void fza_reset_timer(struct timer_list *t) fp->reset_timer.expires = jiffies + 45 * HZ; } add_timer(&fp->reset_timer); + spin_unlock_irqrestore(&fp->lock, flags); } static int fza_set_mac_address(struct net_device *dev, void *addr) From 320fb29ea23cfa1aeef32563da8748247db896ea Mon Sep 17 00:00:00 2001 From: Jamal Hadi Salim Date: Mon, 11 May 2026 14:30:57 -0400 Subject: [PATCH 322/333] net/sched: sch_cbs: Call qdisc_reset for child qdisc During a reset, CBS is not calling reset on its child qdisc, which might cause qlen/backlog accounting issues. For example, if we have CBS with a QFQ parent and a netem child with delay, we can create a scenario where the parent's qlen underflows. QFQ, specifically, uses qlen to check whether it should deference a pointer, so this scenario may cause a null-ptr deref in QFQ: [ 43.875639][ T319] Oops: general protection fault, probably for non-canonical address 0xdffffc0000000009: 0000 [#1] SMP KASAN NOPTI [ 43.876124][ T319] KASAN: null-ptr-deref in range [0x0000000000000048-0x000000000000004f] [ 43.876417][ T319] CPU: 10 UID: 0 PID: 319 Comm: ping Not tainted 7.0.0-13039-ge728258debd5 #773 PREEMPT(full) [ 43.876751][ T319] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 [ 43.876949][ T319] RIP: 0010:qfq_dequeue+0x35c/0x1650 [ 43.877123][ T319] Code: 00 fc ff df 80 3c 02 00 0f 85 17 0e 00 00 4c 8d 73 48 48 89 9d b8 02 00 00 48 b8 00 00 00 00 00 fc ff df 4c 89 f2 48 c1 ea 03 <80> 3c 02 00 0f 85 76 0c 00 00 48 b8 00 00 00 00 00 fc ff df 4c 8b [ 43.877648][ T319] RSP: 0018:ffff8881017ef4f0 EFLAGS: 00010216 [ 43.877845][ T319] RAX: dffffc0000000000 RBX: 0000000000000000 RCX: dffffc0000000000 [ 43.878073][ T319] RDX: 0000000000000009 RSI: 0000000c40000000 RDI: ffff88810eef02b0 [ 43.878306][ T319] RBP: ffff88810eef0000 R08: ffff88810eef0280 R09: 1ffff1102120fd63 [ 43.878523][ T319] R10: 1ffff1102120fd66 R11: 1ffff1102120fd67 R12: 0000000c40000000 [ 43.878742][ T319] R13: ffff88810eef02b8 R14: 0000000000000048 R15: 0000000020000000 [ 43.878959][ T319] FS: 00007f9c51c47c40(0000) GS:ffff88817a0be000(0000) knlGS:0000000000000000 [ 43.879214][ T319] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 43.879403][ T319] CR2: 000055e69a2230a8 CR3: 000000010c07a000 CR4: 0000000000750ef0 [ 43.879621][ T319] PKRU: 55555554 [ 43.879735][ T319] Call Trace: [ 43.879844][ T319] [ 43.879924][ T319] __qdisc_run+0x169/0x1900 [ 43.880075][ T319] ? dev_qdisc_enqueue+0x8b/0x210 [ 43.880222][ T319] __dev_queue_xmit+0x2346/0x37a0 [ 43.880376][ T319] ? register_lock_class+0x3f/0x800 [ 43.880531][ T319] ? srso_alias_return_thunk+0x5/0xfbef5 [ 43.880684][ T319] ? __pfx___dev_queue_xmit+0x10/0x10 [ 43.880834][ T319] ? srso_alias_return_thunk+0x5/0xfbef5 [ 43.880977][ T319] ? __lock_acquire+0x819/0x1df0 [ 43.881124][ T319] ? srso_alias_return_thunk+0x5/0xfbef5 [ 43.881275][ T319] ? srso_alias_return_thunk+0x5/0xfbef5 [ 43.881418][ T319] ? __asan_memcpy+0x3c/0x60 [ 43.881563][ T319] ? srso_alias_return_thunk+0x5/0xfbef5 [ 43.881708][ T319] ? eth_header+0x165/0x1a0 [ 43.881853][ T319] ? lockdep_hardirqs_on_prepare+0xdb/0x1a0 [ 43.882031][ T319] ? srso_alias_return_thunk+0x5/0xfbef5 [ 43.882174][ T319] ? neigh_resolve_output+0x3cc/0x7e0 [ 43.882325][ T319] ? srso_alias_return_thunk+0x5/0xfbef5 [ 43.882471][ T319] ip_finish_output2+0x6b6/0x1e10 Fix this by calling qdisc_reset for CBS' child qdisc. Sashiko caught an issue which could result in a null ptr deref if qdisc_create_dflt() is invoked on an unitialised cbs qdisc which is exposed by this patch. We add an early return if the qdisc is null to address this. This is a similar approach used by two other fixes[1][2]. The proper fix for this specific issue elucidated by sashiko is to remove the call to qdisc_reset when qdisc_create_dflt fails. Since the dflt qdisc isn't attached anywhere yet at that point, calling the reset callback doesn't make much sense (and as stated has been a source of two other bugs). We plan on submitting this fix in a later patch. [1] https://lore.kernel.org/netdev/20221018063201.306474-2-shaozhengchao@huawei.com/ [2] https://lore.kernel.org/netdev/20221018063201.306474-4-shaozhengchao@huawei.com/ Fixes: 585d763af09c ("net/sched: Introduce Credit Based Shaper (CBS) qdisc") Reported-by: Junyoung Jang Tested-by: Junyoung Jang Tested-by: Victor Nogueira Acked-by: Vinicius Costa Gomes Signed-off-by: Jamal Hadi Salim Signed-off-by: Jakub Kicinski --- net/sched/sch_cbs.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/net/sched/sch_cbs.c b/net/sched/sch_cbs.c index 8c9a0400c862..0f953bd46b58 100644 --- a/net/sched/sch_cbs.c +++ b/net/sched/sch_cbs.c @@ -243,6 +243,20 @@ static struct sk_buff *cbs_dequeue(struct Qdisc *sch) return q->dequeue(sch); } +static void cbs_reset(struct Qdisc *sch) +{ + struct cbs_sched_data *q = qdisc_priv(sch); + + /* Nothing to do if we couldn't create the underlying qdisc */ + if (!q->qdisc) + return; + + qdisc_reset(q->qdisc); + qdisc_watchdog_cancel(&q->watchdog); + q->credits = 0; + q->last = 0; +} + static const struct nla_policy cbs_policy[TCA_CBS_MAX + 1] = { [TCA_CBS_PARMS] = { .len = sizeof(struct tc_cbs_qopt) }, }; @@ -540,7 +554,7 @@ static struct Qdisc_ops cbs_qdisc_ops __read_mostly = { .dequeue = cbs_dequeue, .peek = qdisc_peek_dequeued, .init = cbs_init, - .reset = qdisc_reset_queue, + .reset = cbs_reset, .destroy = cbs_destroy, .change = cbs_change, .dump = cbs_dump, From 59afae20080a9681014bdc87897cbfd30bedd261 Mon Sep 17 00:00:00 2001 From: Victor Nogueira Date: Mon, 11 May 2026 14:30:58 -0400 Subject: [PATCH 323/333] selftests/tc-testing: Add QFQ/CBS qlen underflow test Since CBS was not calling reset for its child qdisc, there are scenarios where it could cause an underflow on its parent's qlen/backlog. When the parent is QFQ, a null-ptr deref could occur. Add a test case that reproduces the underflow followed by a null-ptr deref scenario. Acked-by: Jamal Hadi Salim Signed-off-by: Victor Nogueira Signed-off-by: Jakub Kicinski --- .../tc-testing/tc-tests/infra/qdiscs.json | 41 +++++++++++++++++++ 1 file changed, 41 insertions(+) diff --git a/tools/testing/selftests/tc-testing/tc-tests/infra/qdiscs.json b/tools/testing/selftests/tc-testing/tc-tests/infra/qdiscs.json index b1f856cf62c1..848696c373fc 100644 --- a/tools/testing/selftests/tc-testing/tc-tests/infra/qdiscs.json +++ b/tools/testing/selftests/tc-testing/tc-tests/infra/qdiscs.json @@ -1284,5 +1284,46 @@ "teardown": [ "$TC qdisc del dev $DUMMY handle 1: root" ] + }, + { + "id": "3a62", + "name": "Try to create a qlen underflow with QFQ/CBS", + "category": [ + "qdisc", + "qfq", + "cbs" + ], + "plugins": { + "requires": "nsPlugin" + }, + "setup": [ + "$IP link set dev $DUMMY up || true", + "$IP addr add 10.10.10.10/24 dev $DUMMY || true", + "$TC qdisc add dev $DUMMY root handle 1: qfq", + "$TC class add dev $DUMMY classid 1:1 parent 1: qfq", + "$TC class add dev $DUMMY classid 1:2 parent 1: qfq", + "$TC qdisc add dev $DUMMY handle 2: parent 1:1 cbs", + "$TC qdisc add dev $DUMMY handle 3: parent 2: netem delay 5000000000", + "$TC filter add dev $DUMMY parent 1: prio 1 u32 match ip dst 10.10.10.1 classid 1:1 action ok", + "$TC filter add dev $DUMMY parent 1: prio 2 u32 match ip dst 10.10.10.2 classid 1:2 action ok", + "ping -c 1 10.10.10.1 -W0.01 -I$DUMMY || true", + "$IP l set $DUMMY down", + "$IP l set $DUMMY up", + "$TC qdisc replace dev $DUMMY handle 4: parent 2: pfifo" + ], + "cmdUnderTest": "ping -c 1 10.10.10.2 -W0.01 -I$DUMMY", + "expExitCode": "1", + "verifyCmd": "$TC -s -j qdisc ls dev $DUMMY parent 1:1", + "matchJSON": [ + { + "kind": "cbs", + "handle": "2:", + "bytes": 0, + "packets": 0 + } + ], + "teardown": [ + "$TC qdisc del dev $DUMMY handle 1: root" + ] } ] From 9a390d34d55cb4ecbca4981c660dd95440827c70 Mon Sep 17 00:00:00 2001 From: Sukhdeep Singh Date: Tue, 12 May 2026 18:27:11 +0530 Subject: [PATCH 324/333] MAINTAINERS: update atlantic driver maintainer Igor Russkikh and Egor Pomozov have left Marvell. Take over maintenance of the atlantic driver and its PTP subsystem. Signed-off-by: Sukhdeep Singh Signed-off-by: Jakub Kicinski --- MAINTAINERS | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index 8a134cf6e9bb..7a4f2aab78ff 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -2021,7 +2021,7 @@ F: Documentation/hwmon/aquacomputer_d5next.rst F: drivers/hwmon/aquacomputer_d5next.c AQUANTIA ETHERNET DRIVER (atlantic) -M: Igor Russkikh +M: Sukhdeep Singh L: netdev@vger.kernel.org S: Maintained W: https://www.marvell.com/ @@ -2030,7 +2030,7 @@ F: Documentation/networking/device_drivers/ethernet/aquantia/atlantic.rst F: drivers/net/ethernet/aquantia/atlantic/ AQUANTIA ETHERNET DRIVER PTP SUBSYSTEM -M: Egor Pomozov +M: Sukhdeep Singh L: netdev@vger.kernel.org S: Maintained W: http://www.aquantia.com From b84c5632c7b31f8910167075a8128cfb9e50fcfe Mon Sep 17 00:00:00 2001 From: Faicker Mo Date: Mon, 11 May 2026 22:05:51 +0800 Subject: [PATCH 325/333] net: net_failover: Fix the deadlock in slave register There is netdev_lock_ops() before the NETDEV_REGISTER notifier in register_netdevice(), so use the non-locking functions in net_failover_slave_register(). failover_slave_register() in failover_existing_slave_register() adds lock and unlock ops too. Call Trace: __schedule+0x30d/0x7a0 schedule+0x27/0x90 schedule_preempt_disabled+0x15/0x30 __mutex_lock.constprop.0+0x538/0x9e0 __mutex_lock_slowpath+0x13/0x20 mutex_lock+0x3b/0x50 dev_set_mtu+0x40/0xe0 net_failover_slave_register+0x24/0x280 failover_slave_register+0x103/0x1b0 failover_event+0x15e/0x210 ? dropmon_net_event+0xac/0xe0 notifier_call_chain+0x5e/0xe0 raw_notifier_call_chain+0x16/0x30 call_netdevice_notifiers_info+0x52/0xa0 register_netdevice+0x5f4/0x7c0 register_netdev+0x1e/0x40 _mlx5e_probe+0xe2/0x370 [mlx5_core] mlx5e_probe+0x59/0x70 [mlx5_core] ? __pfx_mlx5e_probe+0x10/0x10 [mlx5_core] Fixes: 4c975fd70002 ("net: hold instance lock during NETDEV_REGISTER/UP") Signed-off-by: Faicker Mo Signed-off-by: Jakub Kicinski --- drivers/net/net_failover.c | 12 ++++++------ net/core/failover.c | 6 +++++- 2 files changed, 11 insertions(+), 7 deletions(-) diff --git a/drivers/net/net_failover.c b/drivers/net/net_failover.c index d0361aaf25ef..3f7d31033bae 100644 --- a/drivers/net/net_failover.c +++ b/drivers/net/net_failover.c @@ -502,7 +502,7 @@ static int net_failover_slave_register(struct net_device *slave_dev, /* Align MTU of slave with failover dev */ orig_mtu = slave_dev->mtu; - err = dev_set_mtu(slave_dev, failover_dev->mtu); + err = netif_set_mtu(slave_dev, failover_dev->mtu); if (err) { netdev_err(failover_dev, "unable to change mtu of %s to %u register failed\n", slave_dev->name, failover_dev->mtu); @@ -512,11 +512,11 @@ static int net_failover_slave_register(struct net_device *slave_dev, dev_hold(slave_dev); if (netif_running(failover_dev)) { - err = dev_open(slave_dev, NULL); + err = netif_open(slave_dev, NULL); if (err && (err != -EBUSY)) { netdev_err(failover_dev, "Opening slave %s failed err:%d\n", slave_dev->name, err); - goto err_dev_open; + goto err_netif_open; } } @@ -562,10 +562,10 @@ static int net_failover_slave_register(struct net_device *slave_dev, err_vlan_add: dev_uc_unsync(slave_dev, failover_dev); dev_mc_unsync(slave_dev, failover_dev); - dev_close(slave_dev); -err_dev_open: + netif_close(slave_dev); +err_netif_open: dev_put(slave_dev); - dev_set_mtu(slave_dev, orig_mtu); + netif_set_mtu(slave_dev, orig_mtu); done: return err; } diff --git a/net/core/failover.c b/net/core/failover.c index 11bb183c7a1b..e43c59cd6868 100644 --- a/net/core/failover.c +++ b/net/core/failover.c @@ -12,6 +12,7 @@ #include #include #include +#include #include static LIST_HEAD(failover_list); @@ -221,8 +222,11 @@ failover_existing_slave_register(struct net_device *failover_dev) for_each_netdev(net, dev) { if (netif_is_failover(dev)) continue; - if (ether_addr_equal(failover_dev->perm_addr, dev->perm_addr)) + if (ether_addr_equal(failover_dev->perm_addr, dev->perm_addr)) { + netdev_lock_ops(dev); failover_slave_register(dev); + netdev_unlock_ops(dev); + } } rtnl_unlock(); } From c6690a9030d784d3f099850800b6d5323771ca37 Mon Sep 17 00:00:00 2001 From: Jinliang Zheng Date: Mon, 11 May 2026 23:30:58 +0800 Subject: [PATCH 326/333] macsec: introduce dedicated workqueue for SA crypto cleanup Introduce a dedicated ordered workqueue, macsec_wq, which will be used by subsequent patches to defer SA crypto cleanup (crypto_free_aead and related teardown) out of softirq context. Using a dedicated workqueue instead of system_wq allows macsec_exit() to drain exactly the work items belonging to this module via destroy_workqueue(), without interfering with unrelated work items on system_wq or causing unexpected delays elsewhere. rcu_barrier() in macsec_exit() ensures all in-flight rcu_work callbacks have enqueued their work items before destroy_workqueue() drains and destroys the queue, making the two-step teardown correct and complete. The same sequence is kept in the error path of macsec_init() as a precaution, to mirror macsec_exit() and stay safe if work ever becomes queueable before this point in the future. While at it, rename the error labels in macsec_init() from the resource-named style (rtnl:, notifier:, wq:) to the err_xxx: style (err_rtnl:, err_notifier:, err_destroy_wq:) to align with the broader kernel convention. Signed-off-by: Jinliang Zheng Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20260511153102.2640368-2-alexjlzheng@tencent.com Signed-off-by: Jakub Kicinski --- drivers/net/macsec.c | 23 ++++++++++++++++++----- 1 file changed, 18 insertions(+), 5 deletions(-) diff --git a/drivers/net/macsec.c b/drivers/net/macsec.c index 6147ee8b1d78..ef5ac634f916 100644 --- a/drivers/net/macsec.c +++ b/drivers/net/macsec.c @@ -26,6 +26,8 @@ #include +static struct workqueue_struct *macsec_wq; + /* SecTAG length = macsec_eth_header without the optional SCI */ #define MACSEC_TAG_LEN 6 @@ -4505,25 +4507,35 @@ static int __init macsec_init(void) { int err; + macsec_wq = alloc_workqueue("macsec", WQ_UNBOUND, 0); + if (!macsec_wq) + return -ENOMEM; + pr_info("MACsec IEEE 802.1AE\n"); err = register_netdevice_notifier(&macsec_notifier); if (err) - return err; + goto err_destroy_wq; err = rtnl_link_register(&macsec_link_ops); if (err) - goto notifier; + goto err_notifier; err = genl_register_family(&macsec_fam); if (err) - goto rtnl; + goto err_rtnl; return 0; -rtnl: +err_rtnl: rtnl_link_unregister(&macsec_link_ops); -notifier: +err_notifier: unregister_netdevice_notifier(&macsec_notifier); +err_destroy_wq: + /* Precautionary, mirrors macsec_exit() to stay safe if work + * ever becomes queueable before this point in the future. + */ + rcu_barrier(); + destroy_workqueue(macsec_wq); return err; } @@ -4533,6 +4545,7 @@ static void __exit macsec_exit(void) rtnl_link_unregister(&macsec_link_ops); unregister_netdevice_notifier(&macsec_notifier); rcu_barrier(); + destroy_workqueue(macsec_wq); } module_init(macsec_init); From 6624bba469a325ecd699feae400b77cd11c76b98 Mon Sep 17 00:00:00 2001 From: Jinliang Zheng Date: Mon, 11 May 2026 23:30:59 +0800 Subject: [PATCH 327/333] macsec: use rcu_work to defer RX SA crypto cleanup out of softirq crypto_free_aead() can internally invoke vunmap() (e.g. via dma_free_attrs() in hardware crypto drivers such as hisi_sec2). vunmap() must not be called from softirq context, but free_rxsa() is an RCU callback that runs in softirq, leading to a kernel crash: vunmap+0x4c/0x70 __iommu_dma_free+0xd0/0x138 dma_free_attrs+0xf4/0x100 sec_aead_exit+0x64/0xb8 [hisi_sec2] crypto_destroy_tfm+0x98/0x110 free_rxsa+0x28/0x50 [macsec] rcu_do_batch+0x184/0x460 rcu_core+0xf4/0x1f8 handle_softirqs+0x118/0x330 Use rcu_work to defer the cleanup to a workqueue. rcu_work dispatches the worker asynchronously after the RCU grace period, so no thread blocks waiting, and concurrent releases of multiple SAs naturally share the same grace period. Fixes: c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver") Signed-off-by: Jinliang Zheng Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20260511153102.2640368-3-alexjlzheng@tencent.com Signed-off-by: Jakub Kicinski --- drivers/net/macsec.c | 8 +++++--- include/net/macsec.h | 4 +++- 2 files changed, 8 insertions(+), 4 deletions(-) diff --git a/drivers/net/macsec.c b/drivers/net/macsec.c index ef5ac634f916..e7ad24f1ea5b 100644 --- a/drivers/net/macsec.c +++ b/drivers/net/macsec.c @@ -176,9 +176,10 @@ static void macsec_rxsc_put(struct macsec_rx_sc *sc) call_rcu(&sc->rcu_head, free_rx_sc_rcu); } -static void free_rxsa(struct rcu_head *head) +static void free_rxsa_work(struct work_struct *work) { - struct macsec_rx_sa *sa = container_of(head, struct macsec_rx_sa, rcu); + struct macsec_rx_sa *sa = + container_of(to_rcu_work(work), struct macsec_rx_sa, destroy_work); crypto_free_aead(sa->key.tfm); free_percpu(sa->stats); @@ -188,7 +189,7 @@ static void free_rxsa(struct rcu_head *head) static void macsec_rxsa_put(struct macsec_rx_sa *sa) { if (refcount_dec_and_test(&sa->refcnt)) - call_rcu(&sa->rcu, free_rxsa); + queue_rcu_work(macsec_wq, &sa->destroy_work); } static struct macsec_tx_sa *macsec_txsa_get(struct macsec_tx_sa __rcu *ptr) @@ -1409,6 +1410,7 @@ static int init_rx_sa(struct macsec_rx_sa *rx_sa, char *sak, int key_len, rx_sa->next_pn = 1; refcount_set(&rx_sa->refcnt, 1); spin_lock_init(&rx_sa->lock); + INIT_RCU_WORK(&rx_sa->destroy_work, free_rxsa_work); return 0; } diff --git a/include/net/macsec.h b/include/net/macsec.h index bc7de5b53e54..0980ef36fbf0 100644 --- a/include/net/macsec.h +++ b/include/net/macsec.h @@ -9,6 +9,7 @@ #include #include +#include #include #include @@ -123,6 +124,7 @@ struct macsec_dev_stats { * @key: key structure * @ssci: short secure channel identifier * @stats: per-SA stats + * @destroy_work: deferred work to free the SA in process context after RCU grace period */ struct macsec_rx_sa { struct macsec_key key; @@ -136,7 +138,7 @@ struct macsec_rx_sa { bool active; struct macsec_rx_sa_stats __percpu *stats; struct macsec_rx_sc *sc; - struct rcu_head rcu; + struct rcu_work destroy_work; }; struct pcpu_rx_sc_stats { From 552cc2306c3d87632f44a655737d1d367c2a3295 Mon Sep 17 00:00:00 2001 From: Jinliang Zheng Date: Mon, 11 May 2026 23:31:00 +0800 Subject: [PATCH 328/333] macsec: use rcu_work to defer TX SA crypto cleanup out of softirq free_txsa() is an RCU callback running in softirq context, but calls crypto_free_aead() which can invoke vunmap() internally on hardware crypto drivers (e.g. hisi_sec2), triggering a kernel crash. Use rcu_work to defer the cleanup to a workqueue, for the same reasons as the analogous fix to free_rxsa() in the previous patch. Fixes: c09440f7dcb3 ("macsec: introduce IEEE 802.1AE driver") Signed-off-by: Jinliang Zheng Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20260511153102.2640368-4-alexjlzheng@tencent.com Signed-off-by: Jakub Kicinski --- drivers/net/macsec.c | 8 +++++--- include/net/macsec.h | 3 ++- 2 files changed, 7 insertions(+), 4 deletions(-) diff --git a/drivers/net/macsec.c b/drivers/net/macsec.c index e7ad24f1ea5b..f904f4d16b45 100644 --- a/drivers/net/macsec.c +++ b/drivers/net/macsec.c @@ -205,9 +205,10 @@ static struct macsec_tx_sa *macsec_txsa_get(struct macsec_tx_sa __rcu *ptr) return sa; } -static void free_txsa(struct rcu_head *head) +static void free_txsa_work(struct work_struct *work) { - struct macsec_tx_sa *sa = container_of(head, struct macsec_tx_sa, rcu); + struct macsec_tx_sa *sa = + container_of(to_rcu_work(work), struct macsec_tx_sa, destroy_work); crypto_free_aead(sa->key.tfm); free_percpu(sa->stats); @@ -217,7 +218,7 @@ static void free_txsa(struct rcu_head *head) static void macsec_txsa_put(struct macsec_tx_sa *sa) { if (refcount_dec_and_test(&sa->refcnt)) - call_rcu(&sa->rcu, free_txsa); + queue_rcu_work(macsec_wq, &sa->destroy_work); } static struct macsec_cb *macsec_skb_cb(struct sk_buff *skb) @@ -1510,6 +1511,7 @@ static int init_tx_sa(struct macsec_tx_sa *tx_sa, char *sak, int key_len, tx_sa->active = false; refcount_set(&tx_sa->refcnt, 1); spin_lock_init(&tx_sa->lock); + INIT_RCU_WORK(&tx_sa->destroy_work, free_txsa_work); return 0; } diff --git a/include/net/macsec.h b/include/net/macsec.h index 0980ef36fbf0..d962093ee923 100644 --- a/include/net/macsec.h +++ b/include/net/macsec.h @@ -176,6 +176,7 @@ struct macsec_rx_sc { * @key: key structure * @ssci: short secure channel identifier * @stats: per-SA stats + * @destroy_work: deferred work to free the SA in process context after RCU grace period */ struct macsec_tx_sa { struct macsec_key key; @@ -188,7 +189,7 @@ struct macsec_tx_sa { refcount_t refcnt; bool active; struct macsec_tx_sa_stats __percpu *stats; - struct rcu_head rcu; + struct rcu_work destroy_work; }; /** From 277740023def559a4a2ddc3e8e784ee37a0f16a9 Mon Sep 17 00:00:00 2001 From: Xiang Mei Date: Sun, 10 May 2026 23:21:38 -0700 Subject: [PATCH 329/333] net/smc: reject CHID-0 ACCEPT that matches an empty ism_dev slot On the SMC-D client, slot 0 of ini->ism_dev[]/ini->ism_chid[] is reserved for an SMC-Dv1 device. smc_find_ism_v2_device_clnt() populates V2 entries starting at index 1, so when no V1 device is selected slot 0 is left in its kzalloc()'ed state with ism_dev[0] == NULL and ism_chid[0] == 0. smc_v2_determine_accepted_chid() then matches the peer's CHID against the array starting from index 0 using the CHID alone. A malicious peer replying to a SMC-Dv2-only proposal with d1.chid == 0 matches the empty slot, ini->ism_selected becomes 0, and the subsequent ism_dev[0]->lgr_lock dereference in smc_conn_create() faults at offsetof(struct smcd_dev, lgr_lock) == 0x68: BUG: KASAN: null-ptr-deref in _raw_spin_lock_bh+0x79/0xe0 Write of size 4 at addr 0000000000000068 by task exploit/144 Call Trace: _raw_spin_lock_bh smc_conn_create (net/smc/smc_core.c:1997) __smc_connect (net/smc/af_smc.c:1447) smc_connect (net/smc/af_smc.c:1720) __sys_connect __x64_sys_connect do_syscall_64 Require ism_dev[i] to be non-NULL before accepting a CHID match. Fixes: a7c9c5f4af7f ("net/smc: CLC accept / confirm V2") Reported-by: Weiming Shi Assisted-by: Claude:claude-opus-4-7 Signed-off-by: Xiang Mei Link: https://patch.msgid.link/20260511062138.2839584-1-xmei5@asu.edu Signed-off-by: Paolo Abeni --- net/smc/af_smc.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c index da28652f6810..dffbd529762d 100644 --- a/net/smc/af_smc.c +++ b/net/smc/af_smc.c @@ -1400,7 +1400,8 @@ smc_v2_determine_accepted_chid(struct smc_clc_msg_accept_confirm *aclc, int i; for (i = 0; i < ini->ism_offered_cnt + 1; i++) { - if (ini->ism_chid[i] == ntohs(aclc->d1.chid)) { + if (ini->ism_dev[i] && + ini->ism_chid[i] == ntohs(aclc->d1.chid)) { ini->ism_selected = i; return 0; } From 285943c6e7ca309bbea84b253745154241d9788a Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Mon, 11 May 2026 10:49:17 -0700 Subject: [PATCH 330/333] net: tls: fix off-by-one in sg_chain entry count for wrapped sk_msg ring MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When an sk_msg scatterlist ring wraps (sg.end < sg.start), tls_push_record() chains the tail portion of the ring to the head using sg_chain(). An extra entry in the sg array is reserved for this: struct sk_msg_sg { [...] /* The extra two elements: * 1) used for chaining the front and sections when the list becomes * partitioned (e.g. end < start). The crypto APIs require the * chaining; * 2) to chain tailer SG entries after the message. */ struct scatterlist data[MAX_MSG_FRAGS + 2]; The current code uses MAX_SKB_FRAGS + 1 as the ring size: sg_chain(&msg_pl->sg.data[msg_pl->sg.start], MAX_SKB_FRAGS - msg_pl->sg.start + 1, msg_pl->sg.data); This places the chain pointer at sg_chain(data[start], (MAX_SKB_FRAGS - msg_start + 1) .. = &data[start] + (MAX_SKB_FRAGS - msg_start + 1) - 1 = data[start + (MAX_SKB_FRAGS - start + 1) - 1] = data[MAX_SKB_FRAGS] instead of the true last entry. This is likely due to a "race" of the commit under Fixes landing close to commit 031097d9e079 ("bpf: sk_msg, zap ingress queue on psock down") Convert to ARRAY_SIZE and drop the data[start] / - start (as suggested by Sabrina). Reported-by: 钱一铭 Fixes: 9aaaa56845a0 ("bpf: Sockmap/tls, skmsg can have wrapped skmsg that needs extra chaining") Signed-off-by: Jakub Kicinski Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20260511174920.433155-2-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/tls/tls_sw.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index 2590e855f6a5..2608b0c01849 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -800,11 +800,9 @@ static int tls_push_record(struct sock *sk, int flags, sg_mark_end(sk_msg_elem(msg_pl, i)); } - if (msg_pl->sg.end < msg_pl->sg.start) { - sg_chain(&msg_pl->sg.data[msg_pl->sg.start], - MAX_SKB_FRAGS - msg_pl->sg.start + 1, + if (msg_pl->sg.end < msg_pl->sg.start) + sg_chain(msg_pl->sg.data, ARRAY_SIZE(msg_pl->sg.data), msg_pl->sg.data); - } i = msg_pl->sg.start; sg_chain(rec->sg_aead_in, 2, &msg_pl->sg.data[i]); From ff26a0e8377dec07e4a7230db7675bed1b9a6d03 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Mon, 11 May 2026 10:49:18 -0700 Subject: [PATCH 331/333] net: tls: prevent chain-after-chain in plain text SG Sashiko points out that if end = 0 (start != 0) the current code will create a chain link to content type right after the wrap link: This would create a chain where the wrap link points directly to another chain link. The scatterlist API sg_next iterator does not recursively resolve consecutive chain links. meaning this is illegal input to crypto. The wrapping link is unnecessary if end = 0. end is the entry after the last one used so end = 0 means there's nothing pushed after the wrap: end start i v v v [ ]...[ ][ d ][ d ][ d ][ d ][rsv for wrap] Skip the wrapping in this case. TLS 1.3 can use the "wrapping slot" for it's chaining if end = 0. This avoids the chain-after-chain. Move the wrap chaining before marking END and chaining off content type, that feels like more logical ordering to me, but should not matter from functional perspective. Reported-by: Sashiko Fixes: 9aaaa56845a0 ("bpf: Sockmap/tls, skmsg can have wrapped skmsg that needs extra chaining") Signed-off-by: Jakub Kicinski Link: https://patch.msgid.link/20260511174920.433155-3-kuba@kernel.org Signed-off-by: Paolo Abeni --- net/tls/tls_sw.c | 24 ++++++++++++++++++------ 1 file changed, 18 insertions(+), 6 deletions(-) diff --git a/net/tls/tls_sw.c b/net/tls/tls_sw.c index 2608b0c01849..3bfdaf5e64f5 100644 --- a/net/tls/tls_sw.c +++ b/net/tls/tls_sw.c @@ -789,21 +789,33 @@ static int tls_push_record(struct sock *sk, int flags, i = msg_pl->sg.end; sk_msg_iter_var_prev(i); + /* msg_pl->sg.data is a ring; data[MAX+1] is reserved for the wrap + * link (frags won't use it). 'i' is now the last filled entry: + * + * i end start + * v v v [ rsv ] + * [ d ][ d ][ ][ ]...[ ][ d ][ d ][ d ][chain] + * ^ END v + * `-----------------------------------------' + * + * Note that SGL does not allow chain-after-chain, so for TLS 1.3, + * we must make sure we don't create the wrap entry and then chain + * link to content_type immediately at index 0. + */ + if (i < msg_pl->sg.start) + sg_chain(msg_pl->sg.data, ARRAY_SIZE(msg_pl->sg.data), + msg_pl->sg.data); + rec->content_type = record_type; if (prot->version == TLS_1_3_VERSION) { /* Add content type to end of message. No padding added */ sg_set_buf(&rec->sg_content_type, &rec->content_type, 1); sg_mark_end(&rec->sg_content_type); - sg_chain(msg_pl->sg.data, msg_pl->sg.end + 1, - &rec->sg_content_type); + sg_chain(msg_pl->sg.data, i + 2, &rec->sg_content_type); } else { sg_mark_end(sk_msg_elem(msg_pl, i)); } - if (msg_pl->sg.end < msg_pl->sg.start) - sg_chain(msg_pl->sg.data, ARRAY_SIZE(msg_pl->sg.data), - msg_pl->sg.data); - i = msg_pl->sg.start; sg_chain(rec->sg_aead_in, 2, &msg_pl->sg.data[i]); From c78bdba7b9666020c0832150a4fc4c0aebc7c6ac Mon Sep 17 00:00:00 2001 From: Sven Schuchmann Date: Tue, 12 May 2026 09:19:47 +0200 Subject: [PATCH 332/333] net: phy: DP83TC811: add reading of abilities At this time the driver is not listing any speeds it supports. This should be ETHTOOL_LINK_MODE_100baseT1_Full_BIT for DP83TC811. Add the missing call for phylib to read the abilities. Fixes: b753a9faaf9a ("net: phy: DP83TC811: Introduce support for the DP83TC811 phy") Suggested-by: Andrew Lunn Signed-off-by: Sven Schuchmann Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/20260512071949.6218-1-schuchmann@schleissheimer.de [pabeni@redhat.com: dropped revision history] Signed-off-by: Paolo Abeni --- drivers/net/phy/dp83tc811.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/phy/dp83tc811.c b/drivers/net/phy/dp83tc811.c index e480c2a07450..252fb12b3e68 100644 --- a/drivers/net/phy/dp83tc811.c +++ b/drivers/net/phy/dp83tc811.c @@ -393,6 +393,7 @@ static struct phy_driver dp83811_driver[] = { .config_init = dp83811_config_init, .config_aneg = dp83811_config_aneg, .soft_reset = dp83811_phy_reset, + .get_features = genphy_c45_pma_read_ext_abilities, .get_wol = dp83811_get_wol, .set_wol = dp83811_set_wol, .config_intr = dp83811_config_intr, From 31e62c2ebbfdc3fe3dbdf5e02c92a9dc67087a3a Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 13 May 2026 11:37:18 -0700 Subject: [PATCH 333/333] ptrace: slightly saner 'get_dumpable()' logic The 'dumpability' of a task is fundamentally about the memory image of the task - the concept comes from whether it can core dump or not - and makes no sense when you don't have an associated mm. And almost all users do in fact use it only for the case where the task has a mm pointer. But we have one odd special case: ptrace_may_access() uses 'dumpable' to check various other things entirely independently of the MM (typically explicitly using flags like PTRACE_MODE_READ_FSCREDS). Including for threads that no longer have a VM (and maybe never did, like most kernel threads). It's not what this flag was designed for, but it is what it is. The ptrace code does check that the uid/gid matches, so you do have to be uid-0 to see kernel thread details, but this means that the traditional "drop capabilities" model doesn't make any difference for this all. Make it all make a *bit* more sense by saying that if you don't have a MM pointer, we'll use a cached "last dumpability" flag if the thread ever had a MM (it will be zero for kernel threads since it is never set), and require a proper CAP_SYS_PTRACE capability to override. Reported-by: Qualys Security Advisory Cc: Oleg Nesterov Cc: Kees Cook Signed-off-by: Linus Torvalds --- include/linux/sched.h | 3 +++ kernel/exit.c | 1 + kernel/ptrace.c | 22 ++++++++++++++++------ 3 files changed, 20 insertions(+), 6 deletions(-) diff --git a/include/linux/sched.h b/include/linux/sched.h index 368c7b4d7cb5..ee06cba5c6f5 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -1002,6 +1002,9 @@ struct task_struct { unsigned sched_rt_mutex:1; #endif + /* Save user-dumpable when mm goes away */ + unsigned user_dumpable:1; + /* Bit to tell TOMOYO we're in execve(): */ unsigned in_execve:1; unsigned in_iowait:1; diff --git a/kernel/exit.c b/kernel/exit.c index 9a909993ab1d..f50d73c272d6 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -571,6 +571,7 @@ static void exit_mm(void) */ smp_mb__after_spinlock(); local_irq_disable(); + current->user_dumpable = (get_dumpable(mm) == SUID_DUMP_USER); current->mm = NULL; membarrier_update_current_mm(NULL); enter_lazy_tlb(mm, current); diff --git a/kernel/ptrace.c b/kernel/ptrace.c index 68c17daef8d4..130043bfc209 100644 --- a/kernel/ptrace.c +++ b/kernel/ptrace.c @@ -272,11 +272,24 @@ static bool ptrace_has_cap(struct user_namespace *ns, unsigned int mode) return ns_capable(ns, CAP_SYS_PTRACE); } +static bool task_still_dumpable(struct task_struct *task, unsigned int mode) +{ + struct mm_struct *mm = task->mm; + if (mm) { + if (get_dumpable(mm) == SUID_DUMP_USER) + return true; + return ptrace_has_cap(mm->user_ns, mode); + } + + if (task->user_dumpable) + return true; + return ptrace_has_cap(&init_user_ns, mode); +} + /* Returns 0 on success, -errno on denial. */ static int __ptrace_may_access(struct task_struct *task, unsigned int mode) { const struct cred *cred = current_cred(), *tcred; - struct mm_struct *mm; kuid_t caller_uid; kgid_t caller_gid; @@ -337,11 +350,8 @@ static int __ptrace_may_access(struct task_struct *task, unsigned int mode) * Pairs with a write barrier in commit_creds(). */ smp_rmb(); - mm = task->mm; - if (mm && - ((get_dumpable(mm) != SUID_DUMP_USER) && - !ptrace_has_cap(mm->user_ns, mode))) - return -EPERM; + if (!task_still_dumpable(task, mode)) + return -EPERM; return security_ptrace_access_check(task, mode); }