From 93a4b36ef3cf4ce5e6a7e7a7686181de76e246a1 Mon Sep 17 00:00:00 2001 From: Nirbhay Sharma Date: Fri, 3 Oct 2025 17:15:55 +0530 Subject: [PATCH 001/330] cgroup: Fix seqcount lockdep assertion in cgroup freezer MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The commit afa3701c0e45 ("cgroup: cgroup.stat.local time accounting") introduced a seqcount to track freeze timing but initialized it as a plain seqcount_t using seqcount_init(). However, the write-side critical section in cgroup_do_freeze() holds the css_set_lock spinlock while calling write_seqcount_begin(). On PREEMPT_RT kernels, spinlocks do not disable preemption, causing the lockdep assertion for a plain seqcount_t, which checks for preemption being disabled, to fail. This triggers the following warning: WARNING: CPU: 0 PID: 9692 at include/linux/seqlock.h:221 Fix this by changing the type to seqcount_spinlock_t and initializing it with seqcount_spinlock_init() to associate css_set_lock with the seqcount. This allows lockdep to correctly validate that the spinlock is held during write operations, resolving the assertion failure on all kernel configurations. Reported-by: syzbot+27a2519eb4dad86d0156@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=27a2519eb4dad86d0156 Fixes: afa3701c0e45 ("cgroup: cgroup.stat.local time accounting") Signed-off-by: Nirbhay Sharma Link: https://lore.kernel.org/r/20251002165510.KtY3IT--@linutronix.de/ Acked-by: Michal Koutný Signed-off-by: Tejun Heo --- include/linux/cgroup-defs.h | 2 +- kernel/cgroup/cgroup.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/include/linux/cgroup-defs.h b/include/linux/cgroup-defs.h index 93318fce31f3..b760a3c470a5 100644 --- a/include/linux/cgroup-defs.h +++ b/include/linux/cgroup-defs.h @@ -452,7 +452,7 @@ struct cgroup_freezer_state { int nr_frozen_tasks; /* Freeze time data consistency protection */ - seqcount_t freeze_seq; + seqcount_spinlock_t freeze_seq; /* * Most recent time the cgroup was requested to freeze. diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index 6ae5f48cf64e..fdee387f0d6b 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -5892,7 +5892,7 @@ static struct cgroup *cgroup_create(struct cgroup *parent, const char *name, * if the parent has to be frozen, the child has too. */ cgrp->freezer.e_freeze = parent->freezer.e_freeze; - seqcount_init(&cgrp->freezer.freeze_seq); + seqcount_spinlock_init(&cgrp->freezer.freeze_seq, &css_set_lock); if (cgrp->freezer.e_freeze) { /* * Set the CGRP_FREEZE flag, so when a process will be From 48b77733d0dbaf8cd0a122712072f92b2d95d894 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 1 Oct 2025 15:19:07 +0200 Subject: [PATCH 002/330] expfs: Fix exportfs_can_encode_fh() for EXPORT_FH_FID After commit 5402c4d4d200 ("exportfs: require ->fh_to_parent() to encode connectable file handles") we will fail to create non-decodable file handles for filesystems without export operations. Fix it. Fixes: 5402c4d4d200 ("exportfs: require ->fh_to_parent() to encode connectable file handles") Reviewed-by: Christian Brauner Reviewed-by: Amir Goldstein Signed-off-by: Jan Kara --- include/linux/exportfs.h | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/include/linux/exportfs.h b/include/linux/exportfs.h index d0cf10d5e0f7..f0cf2714ec52 100644 --- a/include/linux/exportfs.h +++ b/include/linux/exportfs.h @@ -320,9 +320,6 @@ static inline bool exportfs_can_decode_fh(const struct export_operations *nop) static inline bool exportfs_can_encode_fh(const struct export_operations *nop, int fh_flags) { - if (!nop) - return false; - /* * If a non-decodeable file handle was requested, we only need to make * sure that filesystem did not opt-out of encoding fid. @@ -330,6 +327,10 @@ static inline bool exportfs_can_encode_fh(const struct export_operations *nop, if (fh_flags & EXPORT_FH_FID) return exportfs_can_encode_fid(nop); + /* Normal file handles cannot be created without export ops */ + if (!nop) + return false; + /* * If a connectable file handle was requested, we need to make sure that * filesystem can also decode connected file handles. From a7c4bb43bfdc2b9f06ee9d036028ed13a83df42a Mon Sep 17 00:00:00 2001 From: Jakub Acs Date: Wed, 1 Oct 2025 10:09:55 +0000 Subject: [PATCH 003/330] fs/notify: call exportfs_encode_fid with s_umount Calling intotify_show_fdinfo() on fd watching an overlayfs inode, while the overlayfs is being unmounted, can lead to dereferencing NULL ptr. This issue was found by syzkaller. Race Condition Diagram: Thread 1 Thread 2 -------- -------- generic_shutdown_super() shrink_dcache_for_umount sb->s_root = NULL | | vfs_read() | inotify_fdinfo() | * inode get from mark * | show_mark_fhandle(m, inode) | exportfs_encode_fid(inode, ..) | ovl_encode_fh(inode, ..) | ovl_check_encode_origin(inode) | * deref i_sb->s_root * | | v fsnotify_sb_delete(sb) Which then leads to: [ 32.133461] Oops: general protection fault, probably for non-canonical address 0xdffffc0000000006: 0000 [#1] SMP DEBUG_PAGEALLOC KASAN NOPTI [ 32.134438] KASAN: null-ptr-deref in range [0x0000000000000030-0x0000000000000037] [ 32.135032] CPU: 1 UID: 0 PID: 4468 Comm: systemd-coredum Not tainted 6.17.0-rc6 #22 PREEMPT(none) [ 32.143353] Call Trace: [ 32.143732] ovl_encode_fh+0xd5/0x170 [ 32.144031] exportfs_encode_inode_fh+0x12f/0x300 [ 32.144425] show_mark_fhandle+0xbe/0x1f0 [ 32.145805] inotify_fdinfo+0x226/0x2d0 [ 32.146442] inotify_show_fdinfo+0x1c5/0x350 [ 32.147168] seq_show+0x530/0x6f0 [ 32.147449] seq_read_iter+0x503/0x12a0 [ 32.148419] seq_read+0x31f/0x410 [ 32.150714] vfs_read+0x1f0/0x9e0 [ 32.152297] ksys_read+0x125/0x240 IOW ovl_check_encode_origin derefs inode->i_sb->s_root, after it was set to NULL in the unmount path. Fix it by protecting calling exportfs_encode_fid() from show_mark_fhandle() with s_umount lock. This form of fix was suggested by Amir in [1]. [1]: https://lore.kernel.org/all/CAOQ4uxhbDwhb+2Brs1UdkoF0a3NSdBAOQPNfEHjahrgoKJpLEw@mail.gmail.com/ Fixes: c45beebfde34 ("ovl: support encoding fid from inode with no alias") Signed-off-by: Jakub Acs Cc: Jan Kara Cc: Amir Goldstein Cc: Miklos Szeredi Cc: Christian Brauner Cc: linux-unionfs@vger.kernel.org Cc: linux-fsdevel@vger.kernel.org Cc: linux-kernel@vger.kernel.org Cc: stable@vger.kernel.org Signed-off-by: Jan Kara --- fs/notify/fdinfo.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/fs/notify/fdinfo.c b/fs/notify/fdinfo.c index 1161eabf11ee..9cc7eb863643 100644 --- a/fs/notify/fdinfo.c +++ b/fs/notify/fdinfo.c @@ -17,6 +17,7 @@ #include "fanotify/fanotify.h" #include "fdinfo.h" #include "fsnotify.h" +#include "../internal.h" #if defined(CONFIG_PROC_FS) @@ -46,7 +47,12 @@ static void show_mark_fhandle(struct seq_file *m, struct inode *inode) size = f->handle_bytes >> 2; + if (!super_trylock_shared(inode->i_sb)) + return; + ret = exportfs_encode_fid(inode, (struct fid *)f->f_handle, &size); + up_read(&inode->i_sb->s_umount); + if ((ret == FILEID_INVALID) || (ret < 0)) return; From b69ffeaa0ae43892683113b3f4ddf156398738b9 Mon Sep 17 00:00:00 2001 From: Long Li Date: Wed, 1 Oct 2025 22:05:30 -0700 Subject: [PATCH 004/330] scsi: storvsc: Prefer returning channel with the same CPU as on the I/O issuing CPU When selecting an outgoing channel for I/O, storvsc tries to select a channel with a returning CPU that is not the same as issuing CPU. This worked well in the past, however it doesn't work well when the Hyper-V exposes a large number of channels (up to the number of all CPUs). Use a different CPU for returning channel is not efficient on Hyper-V. Change this behavior by preferring to the channel with the same CPU as the current I/O issuing CPU whenever possible. Tests have shown improvements in newer Hyper-V/Azure environment, and no regression with older Hyper-V/Azure environments. Tested-by: Raheel Abdul Faizy Signed-off-by: Long Li Message-Id: <1759381530-7414-1-git-send-email-longli@linux.microsoft.com> Signed-off-by: Martin K. Petersen --- drivers/scsi/storvsc_drv.c | 98 ++++++++++++++++++-------------------- 1 file changed, 46 insertions(+), 52 deletions(-) diff --git a/drivers/scsi/storvsc_drv.c b/drivers/scsi/storvsc_drv.c index 7449743930d2..7fb57dca86e2 100644 --- a/drivers/scsi/storvsc_drv.c +++ b/drivers/scsi/storvsc_drv.c @@ -1406,14 +1406,19 @@ static struct vmbus_channel *get_og_chn(struct storvsc_device *stor_device, } /* - * Our channel array is sparsley populated and we + * Our channel array could be sparsley populated and we * initiated I/O on a processor/hw-q that does not * currently have a designated channel. Fix this. * The strategy is simple: - * I. Ensure NUMA locality - * II. Distribute evenly (best effort) + * I. Prefer the channel associated with the current CPU + * II. Ensure NUMA locality + * III. Distribute evenly (best effort) */ + /* Prefer the channel on the I/O issuing processor/hw-q */ + if (cpumask_test_cpu(q_num, &stor_device->alloced_cpus)) + return stor_device->stor_chns[q_num]; + node_mask = cpumask_of_node(cpu_to_node(q_num)); num_channels = 0; @@ -1469,59 +1474,48 @@ static int storvsc_do_io(struct hv_device *device, /* See storvsc_change_target_cpu(). */ outgoing_channel = READ_ONCE(stor_device->stor_chns[q_num]); if (outgoing_channel != NULL) { - if (outgoing_channel->target_cpu == q_num) { - /* - * Ideally, we want to pick a different channel if - * available on the same NUMA node. - */ - node_mask = cpumask_of_node(cpu_to_node(q_num)); - for_each_cpu_wrap(tgt_cpu, - &stor_device->alloced_cpus, q_num + 1) { - if (!cpumask_test_cpu(tgt_cpu, node_mask)) - continue; - if (tgt_cpu == q_num) - continue; - channel = READ_ONCE( - stor_device->stor_chns[tgt_cpu]); - if (channel == NULL) - continue; - if (hv_get_avail_to_write_percent( - &channel->outbound) - > ring_avail_percent_lowater) { - outgoing_channel = channel; - goto found_channel; - } - } + if (hv_get_avail_to_write_percent(&outgoing_channel->outbound) + > ring_avail_percent_lowater) + goto found_channel; - /* - * All the other channels on the same NUMA node are - * busy. Try to use the channel on the current CPU - */ - if (hv_get_avail_to_write_percent( - &outgoing_channel->outbound) - > ring_avail_percent_lowater) + /* + * Channel is busy, try to find a channel on the same NUMA node + */ + node_mask = cpumask_of_node(cpu_to_node(q_num)); + for_each_cpu_wrap(tgt_cpu, &stor_device->alloced_cpus, + q_num + 1) { + if (!cpumask_test_cpu(tgt_cpu, node_mask)) + continue; + channel = READ_ONCE(stor_device->stor_chns[tgt_cpu]); + if (!channel) + continue; + if (hv_get_avail_to_write_percent(&channel->outbound) + > ring_avail_percent_lowater) { + outgoing_channel = channel; goto found_channel; - - /* - * If we reach here, all the channels on the current - * NUMA node are busy. Try to find a channel in - * other NUMA nodes - */ - for_each_cpu(tgt_cpu, &stor_device->alloced_cpus) { - if (cpumask_test_cpu(tgt_cpu, node_mask)) - continue; - channel = READ_ONCE( - stor_device->stor_chns[tgt_cpu]); - if (channel == NULL) - continue; - if (hv_get_avail_to_write_percent( - &channel->outbound) - > ring_avail_percent_lowater) { - outgoing_channel = channel; - goto found_channel; - } } } + + /* + * If we reach here, all the channels on the current + * NUMA node are busy. Try to find a channel in + * all NUMA nodes + */ + for_each_cpu_wrap(tgt_cpu, &stor_device->alloced_cpus, + q_num + 1) { + channel = READ_ONCE(stor_device->stor_chns[tgt_cpu]); + if (!channel) + continue; + if (hv_get_avail_to_write_percent(&channel->outbound) + > ring_avail_percent_lowater) { + outgoing_channel = channel; + goto found_channel; + } + } + /* + * If we reach here, all the channels are busy. Use the + * original channel found. + */ } else { spin_lock_irqsave(&stor_device->lock, flags); outgoing_channel = stor_device->stor_chns[q_num]; From 987da233b2982c686a8ea5cd4c76f0bd5e957ee3 Mon Sep 17 00:00:00 2001 From: Alok Tiwari Date: Mon, 29 Sep 2025 02:25:54 -0700 Subject: [PATCH 005/330] scsi: qla4xxx: Fix typos in comments Fix several spelling mistakes in qla4xxx driver comments: "Unfortunely" -> "Unfortunately" "becase" -> "because" "funcions" -> "functions" "targer_id" -> "target_id" Signed-off-by: Alok Tiwari Signed-off-by: Martin K. Petersen --- drivers/scsi/qla4xxx/ql4_os.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/drivers/scsi/qla4xxx/ql4_os.c b/drivers/scsi/qla4xxx/ql4_os.c index a39f1da4ce47..e4ab541561d0 100644 --- a/drivers/scsi/qla4xxx/ql4_os.c +++ b/drivers/scsi/qla4xxx/ql4_os.c @@ -4104,7 +4104,7 @@ void qla4xxx_srb_compl(struct kref *ref) * The mid-level driver tries to ensure that queuecommand never gets * invoked concurrently with itself or the interrupt handler (although * the interrupt handler may call this routine as part of request- - * completion handling). Unfortunely, it sometimes calls the scheduler + * completion handling). Unfortunately, it sometimes calls the scheduler * in interrupt context which is a big NO! NO!. **/ static int qla4xxx_queuecommand(struct Scsi_Host *host, struct scsi_cmnd *cmd) @@ -4647,7 +4647,7 @@ static int qla4xxx_cmd_wait(struct scsi_qla_host *ha) cmd = scsi_host_find_tag(ha->host, index); /* * We cannot just check if the index is valid, - * becase if we are run from the scsi eh, then + * because if we are run from the scsi eh, then * the scsi/block layer is going to prevent * the tag from being released. */ @@ -4952,7 +4952,7 @@ static int qla4xxx_recover_adapter(struct scsi_qla_host *ha) /* Upon successful firmware/chip reset, re-initialize the adapter */ if (status == QLA_SUCCESS) { /* For ISP-4xxx, force function 1 to always initialize - * before function 3 to prevent both funcions from + * before function 3 to prevent both functions from * stepping on top of the other */ if (is_qla40XX(ha) && (ha->mac_index == 3)) ssleep(6); @@ -6912,7 +6912,7 @@ static int qla4xxx_sess_conn_setup(struct scsi_qla_host *ha, struct ddb_entry *ddb_entry = NULL; /* Create session object, with INVALID_ENTRY, - * the targer_id would get set when we issue the login + * the target_id would get set when we issue the login */ cls_sess = iscsi_session_setup(&qla4xxx_iscsi_transport, ha->host, cmds_max, sizeof(struct ddb_entry), From 120642726ecb1b7a266f5c21bec90821e1154509 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Tue, 30 Sep 2025 15:38:09 +0300 Subject: [PATCH 006/330] scsi: libfc: Prevent integer overflow in fc_fcp_recv_data() The "offset" comes from the skb->data that we received. Here the code is verifying that "offset + len" is within bounds however it does not take integer overflows into account. Use size_add() to be safe. This would only be an issue on 32bit systems which are probably a very small percent of the users. Still, it's worth fixing just for correctness sake. Fixes: 42e9a92fe6a9 ("[SCSI] libfc: A modular Fibre Channel library") Signed-off-by: Dan Carpenter Message-Id: Signed-off-by: Martin K. Petersen --- drivers/scsi/libfc/fc_fcp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/libfc/fc_fcp.c b/drivers/scsi/libfc/fc_fcp.c index 16d0f02af1e4..31d08c115521 100644 --- a/drivers/scsi/libfc/fc_fcp.c +++ b/drivers/scsi/libfc/fc_fcp.c @@ -503,7 +503,7 @@ static void fc_fcp_recv_data(struct fc_fcp_pkt *fsp, struct fc_frame *fp) host_bcode = FC_ERROR; goto err; } - if (offset + len > fsp->data_len) { + if (size_add(offset, len) > fsp->data_len) { /* this should never happen */ if ((fr_flags(fp) & FCPHF_CRC_UNCHECKED) && fc_frame_crc_check(fp)) From 6dfc353af575e33c94f5d740f7b0569fa9b784d9 Mon Sep 17 00:00:00 2001 From: Nitin Rawat Date: Wed, 24 Sep 2025 16:29:00 -0700 Subject: [PATCH 007/330] scsi: ufs: qcom: dt-bindings: Document the Kaanapali UFS controller Document the UFS Controller on the Kaanapali Platform. Signed-off-by: Nitin Rawat Signed-off-by: Jingyi Wang Message-Id: <20250924-knp-ufs-v1-1-42e0955a1f7c@oss.qualcomm.com> Signed-off-by: Martin K. Petersen --- Documentation/devicetree/bindings/ufs/qcom,sm8650-ufshc.yaml | 2 ++ 1 file changed, 2 insertions(+) diff --git a/Documentation/devicetree/bindings/ufs/qcom,sm8650-ufshc.yaml b/Documentation/devicetree/bindings/ufs/qcom,sm8650-ufshc.yaml index aaa0bbb5bfe1..cea84ab2204f 100644 --- a/Documentation/devicetree/bindings/ufs/qcom,sm8650-ufshc.yaml +++ b/Documentation/devicetree/bindings/ufs/qcom,sm8650-ufshc.yaml @@ -15,6 +15,7 @@ select: compatible: contains: enum: + - qcom,kaanapali-ufshc - qcom,sm8650-ufshc - qcom,sm8750-ufshc required: @@ -24,6 +25,7 @@ properties: compatible: items: - enum: + - qcom,kaanapali-ufshc - qcom,sm8650-ufshc - qcom,sm8750-ufshc - const: qcom,ufshc From 7c3321f3d279eda7f7d622312ffdbb889f3bec97 Mon Sep 17 00:00:00 2001 From: Jingyi Wang Date: Wed, 24 Sep 2025 16:29:01 -0700 Subject: [PATCH 008/330] scsi: ufs: phy: dt-bindings: Add QMP UFS PHY compatible for Kaanapali Document the QMP UFS PHY compatible for Qualcomm Kaanapali to support physical layer functionality for UFS found on the SoC. Use fallback to indicate the compatibility of the QMP UFS PHY on the Kaanapali with that on the SM8750. Signed-off-by: Jingyi Wang Acked-by: Konrad Dybcio Signed-off-by: Martin K. Petersen --- .../devicetree/bindings/phy/qcom,sc8280xp-qmp-ufs-phy.yaml | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/Documentation/devicetree/bindings/phy/qcom,sc8280xp-qmp-ufs-phy.yaml b/Documentation/devicetree/bindings/phy/qcom,sc8280xp-qmp-ufs-phy.yaml index a58370a6a5d3..fba7b2549dde 100644 --- a/Documentation/devicetree/bindings/phy/qcom,sc8280xp-qmp-ufs-phy.yaml +++ b/Documentation/devicetree/bindings/phy/qcom,sc8280xp-qmp-ufs-phy.yaml @@ -24,6 +24,10 @@ properties: - enum: - qcom,qcs8300-qmp-ufs-phy - const: qcom,sa8775p-qmp-ufs-phy + - items: + - enum: + - qcom,kaanapali-qmp-ufs-phy + - const: qcom,sm8750-qmp-ufs-phy - enum: - qcom,msm8996-qmp-ufs-phy - qcom,msm8998-qmp-ufs-phy From 1bcc3f87912779f66cdf1789a066046536ca6ccc Mon Sep 17 00:00:00 2001 From: Yan Zhao Date: Wed, 24 Sep 2025 10:42:55 -0700 Subject: [PATCH 009/330] KVM: selftests: Test prefault memory during concurrent memslot removal Expand the prefault memory selftest to add a regression test for a KVM bug where KVM's retry logic would result in (breakable) deadlock due to the memslot deletion waiting on prefaulting to release SRCU, and prefaulting waiting on the memslot to fully disappear (KVM uses a two-step process to delete memslots, and KVM x86 retries page faults if a to-be-deleted, a.k.a. INVALID, memslot is encountered). To exercise concurrent memslot remove, spawn a second thread to initiate memslot removal at roughly the same time as prefaulting. Test memslot removal for all testcases, i.e. don't limit concurrent removal to only the success case. There are essentially three prefault scenarios (so far) that are of interest: 1. Success 2. ENOENT due to no memslot 3. EAGAIN due to INVALID memslot For all intents and purposes, #1 and #2 are mutually exclusive, or rather, easier to test via separate testcases since writing to non-existent memory is trivial. But for #3, making it mutually exclusive with #1 _or_ #2 is actually more complex than testing memslot removal for all scenarios. The only requirement to let memslot removal coexist with other scenarios is a way to guarantee a stable result, e.g. that the "no memslot" test observes ENOENT, not EAGAIN, for the final checks. So, rather than make memslot removal mutually exclusive with the ENOENT scenario, simply restore the memslot and retry prefaulting. For the "no memslot" case, KVM_PRE_FAULT_MEMORY should be idempotent, i.e. should always fail with ENOENT regardless of how many times userspace attempts prefaulting. Pass in both the base GPA and the offset (instead of the "full" GPA) so that the worker can recreate the memslot. Signed-off-by: Yan Zhao Co-developed-by: Sean Christopherson Link: https://lore.kernel.org/r/20250924174255.2141847-1-seanjc@google.com Signed-off-by: Sean Christopherson --- .../selftests/kvm/pre_fault_memory_test.c | 131 +++++++++++++++--- 1 file changed, 114 insertions(+), 17 deletions(-) diff --git a/tools/testing/selftests/kvm/pre_fault_memory_test.c b/tools/testing/selftests/kvm/pre_fault_memory_test.c index 0350a8896a2f..f04768c1d2e4 100644 --- a/tools/testing/selftests/kvm/pre_fault_memory_test.c +++ b/tools/testing/selftests/kvm/pre_fault_memory_test.c @@ -10,6 +10,7 @@ #include #include #include +#include /* Arbitrarily chosen values */ #define TEST_SIZE (SZ_2M + PAGE_SIZE) @@ -30,18 +31,66 @@ static void guest_code(uint64_t base_gpa) GUEST_DONE(); } -static void pre_fault_memory(struct kvm_vcpu *vcpu, u64 gpa, u64 size, - u64 left) +struct slot_worker_data { + struct kvm_vm *vm; + u64 gpa; + uint32_t flags; + bool worker_ready; + bool prefault_ready; + bool recreate_slot; +}; + +static void *delete_slot_worker(void *__data) +{ + struct slot_worker_data *data = __data; + struct kvm_vm *vm = data->vm; + + WRITE_ONCE(data->worker_ready, true); + + while (!READ_ONCE(data->prefault_ready)) + cpu_relax(); + + vm_mem_region_delete(vm, TEST_SLOT); + + while (!READ_ONCE(data->recreate_slot)) + cpu_relax(); + + vm_userspace_mem_region_add(vm, VM_MEM_SRC_ANONYMOUS, data->gpa, + TEST_SLOT, TEST_NPAGES, data->flags); + + return NULL; +} + +static void pre_fault_memory(struct kvm_vcpu *vcpu, u64 base_gpa, u64 offset, + u64 size, u64 expected_left, bool private) { struct kvm_pre_fault_memory range = { - .gpa = gpa, + .gpa = base_gpa + offset, .size = size, .flags = 0, }; - u64 prev; + struct slot_worker_data data = { + .vm = vcpu->vm, + .gpa = base_gpa, + .flags = private ? KVM_MEM_GUEST_MEMFD : 0, + }; + bool slot_recreated = false; + pthread_t slot_worker; int ret, save_errno; + u64 prev; - do { + /* + * Concurrently delete (and recreate) the slot to test KVM's handling + * of a racing memslot deletion with prefaulting. + */ + pthread_create(&slot_worker, NULL, delete_slot_worker, &data); + + while (!READ_ONCE(data.worker_ready)) + cpu_relax(); + + WRITE_ONCE(data.prefault_ready, true); + + for (;;) { prev = range.size; ret = __vcpu_ioctl(vcpu, KVM_PRE_FAULT_MEMORY, &range); save_errno = errno; @@ -49,18 +98,65 @@ static void pre_fault_memory(struct kvm_vcpu *vcpu, u64 gpa, u64 size, "%sexpecting range.size to change on %s", ret < 0 ? "not " : "", ret < 0 ? "failure" : "success"); - } while (ret >= 0 ? range.size : save_errno == EINTR); - TEST_ASSERT(range.size == left, - "Completed with %lld bytes left, expected %" PRId64, - range.size, left); + /* + * Immediately retry prefaulting if KVM was interrupted by an + * unrelated signal/event. + */ + if (ret < 0 && save_errno == EINTR) + continue; - if (left == 0) - __TEST_ASSERT_VM_VCPU_IOCTL(!ret, "KVM_PRE_FAULT_MEMORY", ret, vcpu->vm); + /* + * Tell the worker to recreate the slot in order to complete + * prefaulting (if prefault didn't already succeed before the + * slot was deleted) and/or to prepare for the next testcase. + * Wait for the worker to exit so that the next invocation of + * prefaulting is guaranteed to complete (assuming no KVM bugs). + */ + if (!slot_recreated) { + WRITE_ONCE(data.recreate_slot, true); + pthread_join(slot_worker, NULL); + slot_recreated = true; + + /* + * Retry prefaulting to get a stable result, i.e. to + * avoid seeing random EAGAIN failures. Don't retry if + * prefaulting already succeeded, as KVM disallows + * prefaulting with size=0, i.e. blindly retrying would + * result in test failures due to EINVAL. KVM should + * always return success if all bytes are prefaulted, + * i.e. there is no need to guard against EAGAIN being + * returned. + */ + if (range.size) + continue; + } + + /* + * All done if there are no remaining bytes to prefault, or if + * prefaulting failed (EINTR was handled above, and EAGAIN due + * to prefaulting a memslot that's being actively deleted should + * be impossible since the memslot has already been recreated). + */ + if (!range.size || ret < 0) + break; + } + + TEST_ASSERT(range.size == expected_left, + "Completed with %llu bytes left, expected %lu", + range.size, expected_left); + + /* + * Assert success if prefaulting the entire range should succeed, i.e. + * complete with no bytes remaining. Otherwise prefaulting should have + * failed due to ENOENT (due to RET_PF_EMULATE for emulated MMIO when + * no memslot exists). + */ + if (!expected_left) + TEST_ASSERT_VM_VCPU_IOCTL(!ret, KVM_PRE_FAULT_MEMORY, ret, vcpu->vm); else - /* No memory slot causes RET_PF_EMULATE. it results in -ENOENT. */ - __TEST_ASSERT_VM_VCPU_IOCTL(ret && save_errno == ENOENT, - "KVM_PRE_FAULT_MEMORY", ret, vcpu->vm); + TEST_ASSERT_VM_VCPU_IOCTL(ret && save_errno == ENOENT, + KVM_PRE_FAULT_MEMORY, ret, vcpu->vm); } static void __test_pre_fault_memory(unsigned long vm_type, bool private) @@ -97,9 +193,10 @@ static void __test_pre_fault_memory(unsigned long vm_type, bool private) if (private) vm_mem_set_private(vm, guest_test_phys_mem, TEST_SIZE); - pre_fault_memory(vcpu, guest_test_phys_mem, SZ_2M, 0); - pre_fault_memory(vcpu, guest_test_phys_mem + SZ_2M, PAGE_SIZE * 2, PAGE_SIZE); - pre_fault_memory(vcpu, guest_test_phys_mem + TEST_SIZE, PAGE_SIZE, PAGE_SIZE); + + pre_fault_memory(vcpu, guest_test_phys_mem, 0, SZ_2M, 0, private); + pre_fault_memory(vcpu, guest_test_phys_mem, SZ_2M, PAGE_SIZE * 2, PAGE_SIZE, private); + pre_fault_memory(vcpu, guest_test_phys_mem, TEST_SIZE, PAGE_SIZE, PAGE_SIZE, private); vcpu_args_set(vcpu, 1, guest_test_virt_mem); vcpu_run(vcpu); From e07e10ae83bdf429f59c8c149173a8c4f29c481e Mon Sep 17 00:00:00 2001 From: Ketil Johnsen Date: Wed, 8 Oct 2025 12:51:11 +0200 Subject: [PATCH 010/330] drm/panthor: Ensure MCU is disabled on suspend Currently the Panthor driver needs the GPU to be powered down between suspend and resume. If this is not done, then the MCU_CONTROL register will be preserved as AUTO, which again will cause a premature FW boot on resume. The FW will go directly into fatal state in this case. This case needs to be handled as there is no guarantee that the GPU will be powered down after the suspend callback on all platforms. The fix is to call panthor_fw_stop() in "pre-reset" path to ensure the MCU_CONTROL register is cleared (set DISABLE). This matches well with the already existing call to panthor_fw_start() from the "post-reset" path. Signed-off-by: Ketil Johnsen Acked-by: Boris Brezillon Reviewed-by: Steven Price Fixes: 2718d91816ee ("drm/panthor: Add the FW logical block") Signed-off-by: Steven Price Link: https://lore.kernel.org/r/20251008105112.4077015-1-ketil.johnsen@arm.com --- drivers/gpu/drm/panthor/panthor_fw.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/panthor/panthor_fw.c b/drivers/gpu/drm/panthor/panthor_fw.c index 36f1034839c2..44a995835188 100644 --- a/drivers/gpu/drm/panthor/panthor_fw.c +++ b/drivers/gpu/drm/panthor/panthor_fw.c @@ -1099,6 +1099,7 @@ void panthor_fw_pre_reset(struct panthor_device *ptdev, bool on_hang) } panthor_job_irq_suspend(&ptdev->fw->irq); + panthor_fw_stop(ptdev); } /** From bb642e2d300ee27dcede65cda7ffc47a7047bd69 Mon Sep 17 00:00:00 2001 From: Amit Chaudhary Date: Fri, 26 Sep 2025 12:08:22 -0700 Subject: [PATCH 011/330] nvme-multipath: Skip nr_active increments in RETRY disposition For queue-depth I/O policy, this patch fixes unbalanced I/Os across nvme multipaths. Issue Description: The RETRY disposition incorrectly increments ns->ctrl->nr_active counter and reinitializes iostat start-time. In such cases nr_active counter never goes back to zero until that path disconnects and reconnects. Such a path is not chosen for new I/Os if multiple RETRY cases on a given a path cause its queue-depth counter to be artificially higher compared to other paths. This leads to unbalanced I/Os across paths. The patch skips incrementing nr_active if NVME_MPATH_CNT_ACTIVE is already set. And it skips restarting io stats if NVME_MPATH_IO_STATS is already set. base-commit: e989a3da2d371a4b6597ee8dee5c72e407b4db7a Fixes: d4d957b53d91eeb ("nvme-multipath: support io stats on the mpath device") Signed-off-by: Amit Chaudhary Reviewed-by: Randy Jennings Signed-off-by: Keith Busch --- drivers/nvme/host/multipath.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/nvme/host/multipath.c b/drivers/nvme/host/multipath.c index 3da980dc60d9..543e17aead12 100644 --- a/drivers/nvme/host/multipath.c +++ b/drivers/nvme/host/multipath.c @@ -182,12 +182,14 @@ void nvme_mpath_start_request(struct request *rq) struct nvme_ns *ns = rq->q->queuedata; struct gendisk *disk = ns->head->disk; - if (READ_ONCE(ns->head->subsys->iopolicy) == NVME_IOPOLICY_QD) { + if ((READ_ONCE(ns->head->subsys->iopolicy) == NVME_IOPOLICY_QD) && + !(nvme_req(rq)->flags & NVME_MPATH_CNT_ACTIVE)) { atomic_inc(&ns->ctrl->nr_active); nvme_req(rq)->flags |= NVME_MPATH_CNT_ACTIVE; } - if (!blk_queue_io_stat(disk->queue) || blk_rq_is_passthrough(rq)) + if (!blk_queue_io_stat(disk->queue) || blk_rq_is_passthrough(rq) || + (nvme_req(rq)->flags & NVME_MPATH_IO_STATS)) return; nvme_req(rq)->flags |= NVME_MPATH_IO_STATS; From 812258ff4166bcd41c7d44707e0591f9ae32ac8c Mon Sep 17 00:00:00 2001 From: Conor Dooley Date: Mon, 8 Sep 2025 14:12:35 +0100 Subject: [PATCH 012/330] rust: cfi: only 64-bit arm and x86 support CFI_CLANG The kernel uses the standard rustc targets for non-x86 targets, and out of those only 64-bit arm's target has kcfi support enabled. For x86, the custom 64-bit target enables kcfi. The HAVE_CFI_ICALL_NORMALIZE_INTEGERS_RUSTC config option that allows CFI_CLANG to be used in combination with RUST does not check whether the rustc target supports kcfi. This breaks the build on riscv (and presumably 32-bit arm) when CFI_CLANG and RUST are enabled at the same time. Ordinarily, a rustc-option check would be used to detect target support but unfortunately rustc-option filters out the target for reasons given in commit 46e24a545cdb4 ("rust: kasan/kbuild: fix missing flags on first build"). As a result, if the host supports kcfi but the target does not, e.g. when building for riscv on x86_64, the build would remain broken. Instead, make HAVE_CFI_ICALL_NORMALIZE_INTEGERS_RUSTC depend on the only two architectures where the target used supports it to fix the build. CC: stable@vger.kernel.org Fixes: ca627e636551e ("rust: cfi: add support for CFI_CLANG with Rust") Signed-off-by: Conor Dooley Acked-by: Miguel Ojeda Reviewed-by: Alice Ryhl Link: https://lore.kernel.org/r/20250908-distill-lint-1ae78bcf777c@spud Signed-off-by: Paul Walmsley --- arch/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/Kconfig b/arch/Kconfig index ebe08b9186ad..74ff01133532 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -965,6 +965,7 @@ config HAVE_CFI_ICALL_NORMALIZE_INTEGERS_RUSTC def_bool y depends on HAVE_CFI_ICALL_NORMALIZE_INTEGERS depends on RUSTC_VERSION >= 107900 + depends on ARM64 || X86_64 # With GCOV/KASAN we need this fix: https://github.com/rust-lang/rust/pull/129373 depends on (RUSTC_LLVM_VERSION >= 190103 && RUSTC_VERSION >= 108200) || \ (!GCOV_KERNEL && !KASAN_GENERIC && !KASAN_SW_TAGS) From 781380d2cdef34559a0125ca6464b90bfc01594f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Miquel=20Sabat=C3=A9=20Sol=C3=A0?= Date: Mon, 15 Sep 2025 16:32:52 +0200 Subject: [PATCH 013/330] riscv: kgdb: Ensure that BUFMAX > NUMREGBYTES MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The current value of BUFMAX is similar as in other architectures, but as per documentation on KGDB (see 'Documentation/process/debugging/kgdb.rst'), BUFMAX has to be larger than NUMREGBYTES. Some NUMREGBYTES architectures (e.g. powerpc or hexagon) actually define BUFMAX in relation to NUMREGBYTES, and thus this condition is always guaranteed. Since 2048 is a value that is generally accepted on all architectures, and that is larger than the current value of NUMREGBYTES, we can keep this value in arch/riscv, but we can at least add an 'static_assert' as an extra measure just in case NUMREGBYTES changes in the future for some unforseen reason. Signed-off-by: Miquel Sabaté Solà Link: https://lore.kernel.org/r/20250915143252.154955-1-mikisabate@gmail.com Signed-off-by: Paul Walmsley --- arch/riscv/include/asm/kgdb.h | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/arch/riscv/include/asm/kgdb.h b/arch/riscv/include/asm/kgdb.h index 7559d728c5ff..78b18e2fd771 100644 --- a/arch/riscv/include/asm/kgdb.h +++ b/arch/riscv/include/asm/kgdb.h @@ -3,14 +3,18 @@ #ifndef __ASM_KGDB_H_ #define __ASM_KGDB_H_ +#include + #ifdef __KERNEL__ #define GDB_SIZEOF_REG sizeof(unsigned long) -#define DBG_MAX_REG_NUM (36) -#define NUMREGBYTES ((DBG_MAX_REG_NUM) * GDB_SIZEOF_REG) +#define DBG_MAX_REG_NUM 36 +#define NUMREGBYTES (DBG_MAX_REG_NUM * GDB_SIZEOF_REG) #define CACHE_FLUSH_IS_SAFE 1 #define BUFMAX 2048 +static_assert(BUFMAX > NUMREGBYTES, + "As per KGDB documentation, BUFMAX must be larger than NUMREGBYTES"); #ifdef CONFIG_RISCV_ISA_C #define BREAK_INSTR_SIZE 2 #else @@ -97,6 +101,7 @@ extern unsigned long kgdb_compiled_break; #define DBG_REG_STATUS_OFF 33 #define DBG_REG_BADADDR_OFF 34 #define DBG_REG_CAUSE_OFF 35 +/* NOTE: increase DBG_MAX_REG_NUM if you add more values here. */ extern const char riscv_gdb_stub_feature[64]; From ae9e9f3d67dcef7582a4524047b01e33c5185ddb Mon Sep 17 00:00:00 2001 From: Danil Skrebenkov Date: Fri, 19 Sep 2025 16:28:46 +0300 Subject: [PATCH 014/330] RISC-V: clear hot-unplugged cores from all task mm_cpumasks to avoid rfence errors openSBI v1.7 adds harts checks for ipi operations. Especially it adds comparison between hmask passed as an argument from linux and mask of online harts (from openSBI side). If they don't fit each other the error occurs. When cpu is offline, cpu_online_mask is explicitly cleared in __cpu_disable. However, there is no explicit clearing of mm_cpumask. mm_cpumask is used for rfence operations that call openSBI RFENCE extension which uses ipi to remote harts. If hart is offline there may be error if mask of linux is not as mask of online harts in openSBI. this patch adds explicit clearing of mm_cpumask for offline hart. Signed-off-by: Danil Skrebenkov Reviewed-by: Andrew Jones Link: https://lore.kernel.org/r/20250919132849.31676-1-danil.skrebenkov@cloudbear.ru [pjw@kernel.org: rewrote subject line for clarity] Signed-off-by: Paul Walmsley --- arch/riscv/kernel/cpu-hotplug.c | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/riscv/kernel/cpu-hotplug.c b/arch/riscv/kernel/cpu-hotplug.c index a1e38ecfc8be..3f50d3dd76c6 100644 --- a/arch/riscv/kernel/cpu-hotplug.c +++ b/arch/riscv/kernel/cpu-hotplug.c @@ -54,6 +54,7 @@ void arch_cpuhp_cleanup_dead_cpu(unsigned int cpu) pr_notice("CPU%u: off\n", cpu); + clear_tasks_mm_cpumask(cpu); /* Verify from the firmware if the cpu is really stopped*/ if (cpu_ops->cpu_is_stopped) ret = cpu_ops->cpu_is_stopped(cpu); From c199745d3ac3f836515a5734a6ca5c6f55a8809b Mon Sep 17 00:00:00 2001 From: Florian Schmaus Date: Mon, 6 Oct 2025 11:37:42 +0200 Subject: [PATCH 015/330] riscv: entry: fix typo in comment 'instruciton' -> 'instruction' Fix a typo in a comment in the RISC-V entry.S. Signed-off-by: Florian Schmaus Link: https://lore.kernel.org/r/20251006093742.53925-1-flo@geekplace.eu [pjw@kernel.org: wrote a basic patch description] Signed-off-by: Paul Walmsley --- arch/riscv/kernel/entry.S | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/riscv/kernel/entry.S b/arch/riscv/kernel/entry.S index d3d92a4becc7..9b9dec6893b8 100644 --- a/arch/riscv/kernel/entry.S +++ b/arch/riscv/kernel/entry.S @@ -455,7 +455,7 @@ SYM_DATA_START_LOCAL(excp_vect_table) RISCV_PTR do_trap_ecall_s RISCV_PTR do_trap_unknown RISCV_PTR do_trap_ecall_m - /* instruciton page fault */ + /* instruction page fault */ ALT_PAGE_FAULT(RISCV_PTR do_page_fault) RISCV_PTR do_page_fault /* load page fault */ RISCV_PTR do_trap_unknown From 9e68bd803fac49274fde914466fd3b07c4d602c8 Mon Sep 17 00:00:00 2001 From: Fabian Vogt Date: Wed, 10 Sep 2025 17:25:13 +0200 Subject: [PATCH 016/330] riscv: kprobes: Fix probe address validation When adding a kprobe such as "p:probe/tcp_sendmsg _text+15392192", arch_check_kprobe would start iterating all instructions starting from _text until the probed address. Not only is this very inefficient, but literal values in there (e.g. left by function patching) are misinterpreted in a way that causes a desync. Fix this by doing it like x86: start the iteration at the closest preceding symbol instead of the given starting point. Fixes: 87f48c7ccc73 ("riscv: kprobe: Fixup kernel panic when probing an illegal position") Signed-off-by: Fabian Vogt Signed-off-by: Marvin Friedrich Acked-by: Guo Ren Link: https://lore.kernel.org/r/6191817.lOV4Wx5bFT@fvogt-thinkpad Signed-off-by: Paul Walmsley --- arch/riscv/kernel/probes/kprobes.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) diff --git a/arch/riscv/kernel/probes/kprobes.c b/arch/riscv/kernel/probes/kprobes.c index c0738d6c6498..8723390c7cad 100644 --- a/arch/riscv/kernel/probes/kprobes.c +++ b/arch/riscv/kernel/probes/kprobes.c @@ -49,10 +49,15 @@ static void __kprobes arch_simulate_insn(struct kprobe *p, struct pt_regs *regs) post_kprobe_handler(p, kcb, regs); } -static bool __kprobes arch_check_kprobe(struct kprobe *p) +static bool __kprobes arch_check_kprobe(unsigned long addr) { - unsigned long tmp = (unsigned long)p->addr - p->offset; - unsigned long addr = (unsigned long)p->addr; + unsigned long tmp, offset; + + /* start iterating at the closest preceding symbol */ + if (!kallsyms_lookup_size_offset(addr, NULL, &offset)) + return false; + + tmp = addr - offset; while (tmp <= addr) { if (tmp == addr) @@ -71,7 +76,7 @@ int __kprobes arch_prepare_kprobe(struct kprobe *p) if ((unsigned long)insn & 0x1) return -EILSEQ; - if (!arch_check_kprobe(p)) + if (!arch_check_kprobe((unsigned long)p->addr)) return -EILSEQ; /* copy instruction */ From 69a8b62a7aa1e54ff7623064f6507fa29c1d0d4e Mon Sep 17 00:00:00 2001 From: Han Gao Date: Wed, 10 Sep 2025 19:24:01 +0800 Subject: [PATCH 017/330] riscv: acpi: avoid errors caused by probing DT devices when ACPI is used Similar to the ARM64 commit 3505f30fb6a9s ("ARM64 / ACPI: If we chose to boot from acpi then disable FDT"), let's not do DT hardware probing if ACPI is enabled in early boot. This avoids errors caused by repeated driver probing. Signed-off-by: Han Gao Link: https://lore.kernel.org/r/20250910112401.552987-1-rabenda.cn@gmail.com [pjw@kernel.org: cleaned up patch description and subject] Signed-off-by: Paul Walmsley --- arch/riscv/kernel/setup.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/arch/riscv/kernel/setup.c b/arch/riscv/kernel/setup.c index 14235e58c539..b5bc5fc65cea 100644 --- a/arch/riscv/kernel/setup.c +++ b/arch/riscv/kernel/setup.c @@ -331,11 +331,14 @@ void __init setup_arch(char **cmdline_p) /* Parse the ACPI tables for possible boot-time configuration */ acpi_boot_table_init(); + if (acpi_disabled) { #if IS_ENABLED(CONFIG_BUILTIN_DTB) - unflatten_and_copy_device_tree(); + unflatten_and_copy_device_tree(); #else - unflatten_device_tree(); + unflatten_device_tree(); #endif + } + misc_mem_init(); init_resources(); From 7882d2c45ccba538cddb0615a893a008dd2efcde Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Thomas=20Wei=C3=9Fschuh?= Date: Thu, 9 Oct 2025 11:18:48 +0200 Subject: [PATCH 018/330] riscv: Respect dependencies of ARCH_HAS_ELF_CORE_EFLAGS MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This kconfig symbol has dependencies and is only selectable if those dependencies are also enabled. Respect the dependencies. Fixes the following warning when configuring an 'allnoconfig': WARNING: unmet direct dependencies detected for ARCH_HAS_ELF_CORE_EFLAGS Depends on [n]: BINFMT_ELF [=n] && ELF_CORE [=y] Selected by [y]: - RISCV [=y] Fixes: 8c94db0ae97c ("binfmt_elf: preserve original ELF e_flags for core dumps") Signed-off-by: Thomas Weißschuh Link: https://lore.kernel.org/r/20251009-riscv-elf-core-eflags-v1-1-e9b45ab6b36d@linutronix.de Signed-off-by: Paul Walmsley --- arch/riscv/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig index 0c6038dc5dfd..22cda9c452d2 100644 --- a/arch/riscv/Kconfig +++ b/arch/riscv/Kconfig @@ -29,7 +29,7 @@ config RISCV select ARCH_HAS_DEBUG_VIRTUAL if MMU select ARCH_HAS_DEBUG_VM_PGTABLE select ARCH_HAS_DEBUG_WX - select ARCH_HAS_ELF_CORE_EFLAGS + select ARCH_HAS_ELF_CORE_EFLAGS if BINFMT_ELF && ELF_CORE select ARCH_HAS_FAST_MULTIPLIER select ARCH_HAS_FORTIFY_SOURCE select ARCH_HAS_GCOV_PROFILE_ALL From 034417c1439a533a315562a57bd340d963eaac6b Mon Sep 17 00:00:00 2001 From: Dapeng Mi Date: Fri, 10 Oct 2025 08:52:39 +0800 Subject: [PATCH 019/330] KVM: x86/pmu: Don't try to get perf capabilities for hybrid CPUs Explicitly zero kvm_host_pmu instead of attempting to get the perf PMU capabilities when running on a hybrid CPU to avoid running afoul of perf's sanity check. ------------[ cut here ]------------ WARNING: arch/x86/events/core.c:3089 at perf_get_x86_pmu_capability+0xd/0xc0, Call Trace: kvm_x86_vendor_init+0x1b0/0x1a40 [kvm] vmx_init+0xdb/0x260 [kvm_intel] vt_init+0x12/0x9d0 [kvm_intel] do_one_initcall+0x60/0x3f0 do_init_module+0x97/0x2b0 load_module+0x2d08/0x2e30 init_module_from_file+0x96/0xe0 idempotent_init_module+0x117/0x330 __x64_sys_finit_module+0x73/0xe0 Always read the capabilities for non-hybrid CPUs, i.e. don't entirely revert to reading capabilities if and only if KVM wants to use a PMU, as it may be useful to have the host PMU capabilities available, e.g. if only or debug. Reported-by: Chaitanya Kumar Borah Closes: https://lore.kernel.org/all/70b64347-2aca-4511-af78-a767d5fa8226@intel.com/ Fixes: 51f34b1e650f ("KVM: x86/pmu: Snapshot host (i.e. perf's) reported PMU capabilities") Suggested-by: Sean Christopherson Signed-off-by: Dapeng Mi Link: https://lore.kernel.org/r/20251010005239.146953-1-dapeng1.mi@linux.intel.com [sean: rework changelog, call out hybrid CPUs in shortlog] Signed-off-by: Sean Christopherson --- arch/x86/kvm/pmu.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/arch/x86/kvm/pmu.c b/arch/x86/kvm/pmu.c index 40ac4cb44ed2..487ad19a236e 100644 --- a/arch/x86/kvm/pmu.c +++ b/arch/x86/kvm/pmu.c @@ -108,16 +108,18 @@ void kvm_init_pmu_capability(const struct kvm_pmu_ops *pmu_ops) bool is_intel = boot_cpu_data.x86_vendor == X86_VENDOR_INTEL; int min_nr_gp_ctrs = pmu_ops->MIN_NR_GP_COUNTERS; - perf_get_x86_pmu_capability(&kvm_host_pmu); - /* * Hybrid PMUs don't play nice with virtualization without careful * configuration by userspace, and KVM's APIs for reporting supported * vPMU features do not account for hybrid PMUs. Disable vPMU support * for hybrid PMUs until KVM gains a way to let userspace opt-in. */ - if (cpu_feature_enabled(X86_FEATURE_HYBRID_CPU)) + if (cpu_feature_enabled(X86_FEATURE_HYBRID_CPU)) { enable_pmu = false; + memset(&kvm_host_pmu, 0, sizeof(kvm_host_pmu)); + } else { + perf_get_x86_pmu_capability(&kvm_host_pmu); + } if (enable_pmu) { /* From d2042d8f96ddefdeee823737f813efe3ab4b4e8d Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:25:54 -0700 Subject: [PATCH 020/330] KVM: Rework KVM_CAP_GUEST_MEMFD_MMAP into KVM_CAP_GUEST_MEMFD_FLAGS Rework the not-yet-released KVM_CAP_GUEST_MEMFD_MMAP into a more generic KVM_CAP_GUEST_MEMFD_FLAGS capability so that adding new flags doesn't require a new capability, and so that developers aren't tempted to bundle multiple flags into a single capability. Note, kvm_vm_ioctl_check_extension_generic() can only return a 32-bit value, but that limitation can be easily circumvented by adding e.g. KVM_CAP_GUEST_MEMFD_FLAGS2 in the unlikely event guest_memfd supports more than 32 flags. Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Reviewed-by: David Hildenbrand Link: https://lore.kernel.org/r/20251003232606.4070510-2-seanjc@google.com Signed-off-by: Sean Christopherson --- Documentation/virt/kvm/api.rst | 10 +++++++--- include/uapi/linux/kvm.h | 2 +- tools/testing/selftests/kvm/guest_memfd_test.c | 13 ++++++------- virt/kvm/kvm_main.c | 7 +++++-- 4 files changed, 19 insertions(+), 13 deletions(-) diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst index 6ae24c5ca559..7ba92f2ced38 100644 --- a/Documentation/virt/kvm/api.rst +++ b/Documentation/virt/kvm/api.rst @@ -6432,9 +6432,13 @@ most one mapping per page, i.e. binding multiple memory regions to a single guest_memfd range is not allowed (any number of memory regions can be bound to a single guest_memfd file, but the bound ranges must not overlap). -When the capability KVM_CAP_GUEST_MEMFD_MMAP is supported, the 'flags' field -supports GUEST_MEMFD_FLAG_MMAP. Setting this flag on guest_memfd creation -enables mmap() and faulting of guest_memfd memory to host userspace. +The capability KVM_CAP_GUEST_MEMFD_FLAGS enumerates the `flags` that can be +specified via KVM_CREATE_GUEST_MEMFD. Currently defined flags: + + ============================ ================================================ + GUEST_MEMFD_FLAG_MMAP Enable using mmap() on the guest_memfd file + descriptor. + ============================ ================================================ When the KVM MMU performs a PFN lookup to service a guest fault and the backing guest_memfd has the GUEST_MEMFD_FLAG_MMAP set, then the fault will always be diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h index 6efa98a57ec1..b1d52d0c56ec 100644 --- a/include/uapi/linux/kvm.h +++ b/include/uapi/linux/kvm.h @@ -962,7 +962,7 @@ struct kvm_enable_cap { #define KVM_CAP_ARM_EL2_E2H0 241 #define KVM_CAP_RISCV_MP_STATE_RESET 242 #define KVM_CAP_ARM_CACHEABLE_PFNMAP_SUPPORTED 243 -#define KVM_CAP_GUEST_MEMFD_MMAP 244 +#define KVM_CAP_GUEST_MEMFD_FLAGS 244 struct kvm_irq_routing_irqchip { __u32 irqchip; diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index b3ca6737f304..3e58bd496104 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -262,19 +262,17 @@ static void test_guest_memfd_flags(struct kvm_vm *vm, uint64_t valid_flags) static void test_guest_memfd(unsigned long vm_type) { - uint64_t flags = 0; struct kvm_vm *vm; size_t total_size; size_t page_size; + uint64_t flags; int fd; page_size = getpagesize(); total_size = page_size * 4; vm = vm_create_barebones_type(vm_type); - - if (vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_MMAP)) - flags |= GUEST_MEMFD_FLAG_MMAP; + flags = vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS); test_create_guest_memfd_multiple(vm); test_create_guest_memfd_invalid_sizes(vm, flags, page_size); @@ -328,13 +326,14 @@ static void test_guest_memfd_guest(void) size_t size; int fd, i; - if (!kvm_has_cap(KVM_CAP_GUEST_MEMFD_MMAP)) + if (!kvm_check_cap(KVM_CAP_GUEST_MEMFD_FLAGS)) return; vm = __vm_create_shape_with_one_vcpu(VM_SHAPE_DEFAULT, &vcpu, 1, guest_code); - TEST_ASSERT(vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_MMAP), - "Default VM type should always support guest_memfd mmap()"); + TEST_ASSERT(vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS) & GUEST_MEMFD_FLAG_MMAP, + "Default VM type should support MMAP, supported flags = 0x%x", + vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS)); size = vm->page_size; fd = vm_create_guest_memfd(vm, size, GUEST_MEMFD_FLAG_MMAP); diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 226faeaa8e56..e3a268757621 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -4928,8 +4928,11 @@ static int kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg) #ifdef CONFIG_KVM_GUEST_MEMFD case KVM_CAP_GUEST_MEMFD: return 1; - case KVM_CAP_GUEST_MEMFD_MMAP: - return !kvm || kvm_arch_supports_gmem_mmap(kvm); + case KVM_CAP_GUEST_MEMFD_FLAGS: + if (!kvm || kvm_arch_supports_gmem_mmap(kvm)) + return GUEST_MEMFD_FLAG_MMAP; + + return 0; #endif default: break; From fe2bf6234e947bf5544db6d386af1df2a8db80f3 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:25:55 -0700 Subject: [PATCH 021/330] KVM: guest_memfd: Add INIT_SHARED flag, reject user page faults if not set Add a guest_memfd flag to allow userspace to state that the underlying memory should be configured to be initialized as shared, and reject user page faults if the guest_memfd instance's memory isn't shared. Because KVM doesn't yet support in-place private<=>shared conversions, all guest_memfd memory effectively follows the initial state. Alternatively, KVM could deduce the initial state based on MMAP, which for all intents and purposes is what KVM currently does. However, implicitly deriving the default state based on MMAP will result in a messy ABI when support for in-place conversions is added. For x86 CoCo VMs, which don't yet support MMAP, memory is currently private by default (otherwise the memory would be unusable). If MMAP implies memory is shared by default, then the default state for CoCo VMs will vary based on MMAP, and from userspace's perspective, will change when in-place conversion support is added. I.e. to maintain guest<=>host ABI, userspace would need to immediately convert all memory from shared=>private, which is both ugly and inefficient. The inefficiency could be avoided by adding a flag to state that memory is _private_ by default, irrespective of MMAP, but that would lead to an equally messy and hard to document ABI. Bite the bullet and immediately add a flag to control the default state so that the effective behavior is explicit and straightforward. Fixes: 3d3a04fad25a ("KVM: Allow and advertise support for host mmap() on guest_memfd files") Cc: David Hildenbrand Reviewed-by: Fuad Tabba Tested-by: Fuad Tabba Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Reviewed-by: David Hildenbrand Link: https://lore.kernel.org/r/20251003232606.4070510-3-seanjc@google.com Signed-off-by: Sean Christopherson --- Documentation/virt/kvm/api.rst | 5 +++++ include/uapi/linux/kvm.h | 3 ++- tools/testing/selftests/kvm/guest_memfd_test.c | 15 ++++++++++++--- virt/kvm/guest_memfd.c | 6 +++++- virt/kvm/kvm_main.c | 3 ++- 5 files changed, 26 insertions(+), 6 deletions(-) diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst index 7ba92f2ced38..754b662a453c 100644 --- a/Documentation/virt/kvm/api.rst +++ b/Documentation/virt/kvm/api.rst @@ -6438,6 +6438,11 @@ specified via KVM_CREATE_GUEST_MEMFD. Currently defined flags: ============================ ================================================ GUEST_MEMFD_FLAG_MMAP Enable using mmap() on the guest_memfd file descriptor. + GUEST_MEMFD_FLAG_INIT_SHARED Make all memory in the file shared during + KVM_CREATE_GUEST_MEMFD (memory files created + without INIT_SHARED will be marked private). + Shared memory can be faulted into host userspace + page tables. Private memory cannot. ============================ ================================================ When the KVM MMU performs a PFN lookup to service a guest fault and the backing diff --git a/include/uapi/linux/kvm.h b/include/uapi/linux/kvm.h index b1d52d0c56ec..52f6000ab020 100644 --- a/include/uapi/linux/kvm.h +++ b/include/uapi/linux/kvm.h @@ -1599,7 +1599,8 @@ struct kvm_memory_attributes { #define KVM_MEMORY_ATTRIBUTE_PRIVATE (1ULL << 3) #define KVM_CREATE_GUEST_MEMFD _IOWR(KVMIO, 0xd4, struct kvm_create_guest_memfd) -#define GUEST_MEMFD_FLAG_MMAP (1ULL << 0) +#define GUEST_MEMFD_FLAG_MMAP (1ULL << 0) +#define GUEST_MEMFD_FLAG_INIT_SHARED (1ULL << 1) struct kvm_create_guest_memfd { __u64 size; diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index 3e58bd496104..0de56ce3c4e2 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -239,8 +239,9 @@ static void test_create_guest_memfd_multiple(struct kvm_vm *vm) close(fd1); } -static void test_guest_memfd_flags(struct kvm_vm *vm, uint64_t valid_flags) +static void test_guest_memfd_flags(struct kvm_vm *vm) { + uint64_t valid_flags = vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS); size_t page_size = getpagesize(); uint64_t flag; int fd; @@ -274,6 +275,10 @@ static void test_guest_memfd(unsigned long vm_type) vm = vm_create_barebones_type(vm_type); flags = vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS); + /* This test doesn't yet support testing mmap() on private memory. */ + if (!(flags & GUEST_MEMFD_FLAG_INIT_SHARED)) + flags &= ~GUEST_MEMFD_FLAG_MMAP; + test_create_guest_memfd_multiple(vm); test_create_guest_memfd_invalid_sizes(vm, flags, page_size); @@ -292,7 +297,7 @@ static void test_guest_memfd(unsigned long vm_type) test_fallocate(fd, page_size, total_size); test_invalid_punch_hole(fd, page_size, total_size); - test_guest_memfd_flags(vm, flags); + test_guest_memfd_flags(vm); close(fd); kvm_vm_free(vm); @@ -334,9 +339,13 @@ static void test_guest_memfd_guest(void) TEST_ASSERT(vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS) & GUEST_MEMFD_FLAG_MMAP, "Default VM type should support MMAP, supported flags = 0x%x", vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS)); + TEST_ASSERT(vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS) & GUEST_MEMFD_FLAG_INIT_SHARED, + "Default VM type should support INIT_SHARED, supported flags = 0x%x", + vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS)); size = vm->page_size; - fd = vm_create_guest_memfd(vm, size, GUEST_MEMFD_FLAG_MMAP); + 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); mem = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); diff --git a/virt/kvm/guest_memfd.c b/virt/kvm/guest_memfd.c index 94bafd6c558c..cf3afba23a6b 100644 --- a/virt/kvm/guest_memfd.c +++ b/virt/kvm/guest_memfd.c @@ -328,6 +328,9 @@ static vm_fault_t kvm_gmem_fault_user_mapping(struct vm_fault *vmf) if (((loff_t)vmf->pgoff << PAGE_SHIFT) >= i_size_read(inode)) return VM_FAULT_SIGBUS; + if (!((u64)inode->i_private & GUEST_MEMFD_FLAG_INIT_SHARED)) + return VM_FAULT_SIGBUS; + folio = kvm_gmem_get_folio(inode, vmf->pgoff); if (IS_ERR(folio)) { int err = PTR_ERR(folio); @@ -525,7 +528,8 @@ int kvm_gmem_create(struct kvm *kvm, struct kvm_create_guest_memfd *args) u64 valid_flags = 0; if (kvm_arch_supports_gmem_mmap(kvm)) - valid_flags |= GUEST_MEMFD_FLAG_MMAP; + valid_flags |= GUEST_MEMFD_FLAG_MMAP | + GUEST_MEMFD_FLAG_INIT_SHARED; if (flags & ~valid_flags) return -EINVAL; diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index e3a268757621..5f644ca54af3 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -4930,7 +4930,8 @@ static int kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg) return 1; case KVM_CAP_GUEST_MEMFD_FLAGS: if (!kvm || kvm_arch_supports_gmem_mmap(kvm)) - return GUEST_MEMFD_FLAG_MMAP; + return GUEST_MEMFD_FLAG_MMAP | + GUEST_MEMFD_FLAG_INIT_SHARED; return 0; #endif From 5d3341d684be80892d8f6f9812f90f9274b81177 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:25:56 -0700 Subject: [PATCH 022/330] KVM: guest_memfd: Invalidate SHARED GPAs if gmem supports INIT_SHARED When invalidating gmem ranges, e.g. in response to PUNCH_HOLE, process all possible range types (PRIVATE vs. SHARED) for the gmem instance. Since since guest_memfd doesn't yet support in-place conversions, simply pivot on INIT_SHARED as a gmem instance can currently only have private or shared memory, not both. Failure to mark shared GPAs for invalidation is benign in the current code base, as only x86's TDX consumes KVM_FILTER_{PRIVATE,SHARED}, and TDX doesn't yet support INIT_SHARED with guest_memfd. However, invalidating only private GPAs is conceptually wrong and a lurking bug, e.g. could result in missed invalidations if ARM starts filtering invalidations based on attributes. Fixes: 3d3a04fad25a ("KVM: Allow and advertise support for host mmap() on guest_memfd files") Reviewed-by: Ackerley Tng Reviewed-by: David Hildenbrand Link: https://lore.kernel.org/r/20251003232606.4070510-4-seanjc@google.com Signed-off-by: Sean Christopherson --- virt/kvm/guest_memfd.c | 64 +++++++++++++++++++++++++++++------------- 1 file changed, 44 insertions(+), 20 deletions(-) diff --git a/virt/kvm/guest_memfd.c b/virt/kvm/guest_memfd.c index cf3afba23a6b..e10d2c71e78c 100644 --- a/virt/kvm/guest_memfd.c +++ b/virt/kvm/guest_memfd.c @@ -102,8 +102,17 @@ static struct folio *kvm_gmem_get_folio(struct inode *inode, pgoff_t index) return filemap_grab_folio(inode->i_mapping, index); } -static void kvm_gmem_invalidate_begin(struct kvm_gmem *gmem, pgoff_t start, - pgoff_t end) +static enum kvm_gfn_range_filter kvm_gmem_get_invalidate_filter(struct inode *inode) +{ + if ((u64)inode->i_private & GUEST_MEMFD_FLAG_INIT_SHARED) + return KVM_FILTER_SHARED; + + return KVM_FILTER_PRIVATE; +} + +static void __kvm_gmem_invalidate_begin(struct kvm_gmem *gmem, pgoff_t start, + pgoff_t end, + enum kvm_gfn_range_filter attr_filter) { bool flush = false, found_memslot = false; struct kvm_memory_slot *slot; @@ -118,8 +127,7 @@ static void kvm_gmem_invalidate_begin(struct kvm_gmem *gmem, pgoff_t start, .end = slot->base_gfn + min(pgoff + slot->npages, end) - pgoff, .slot = slot, .may_block = true, - /* guest memfd is relevant to only private mappings. */ - .attr_filter = KVM_FILTER_PRIVATE, + .attr_filter = attr_filter, }; if (!found_memslot) { @@ -139,8 +147,21 @@ static void kvm_gmem_invalidate_begin(struct kvm_gmem *gmem, pgoff_t start, KVM_MMU_UNLOCK(kvm); } -static void kvm_gmem_invalidate_end(struct kvm_gmem *gmem, pgoff_t start, - pgoff_t end) +static void kvm_gmem_invalidate_begin(struct inode *inode, pgoff_t start, + pgoff_t end) +{ + struct list_head *gmem_list = &inode->i_mapping->i_private_list; + enum kvm_gfn_range_filter attr_filter; + struct kvm_gmem *gmem; + + attr_filter = kvm_gmem_get_invalidate_filter(inode); + + list_for_each_entry(gmem, gmem_list, entry) + __kvm_gmem_invalidate_begin(gmem, start, end, attr_filter); +} + +static void __kvm_gmem_invalidate_end(struct kvm_gmem *gmem, pgoff_t start, + pgoff_t end) { struct kvm *kvm = gmem->kvm; @@ -151,12 +172,20 @@ static void kvm_gmem_invalidate_end(struct kvm_gmem *gmem, pgoff_t start, } } -static long kvm_gmem_punch_hole(struct inode *inode, loff_t offset, loff_t len) +static void kvm_gmem_invalidate_end(struct inode *inode, pgoff_t start, + pgoff_t end) { struct list_head *gmem_list = &inode->i_mapping->i_private_list; + struct kvm_gmem *gmem; + + list_for_each_entry(gmem, gmem_list, entry) + __kvm_gmem_invalidate_end(gmem, start, end); +} + +static long kvm_gmem_punch_hole(struct inode *inode, loff_t offset, loff_t len) +{ pgoff_t start = offset >> PAGE_SHIFT; pgoff_t end = (offset + len) >> PAGE_SHIFT; - struct kvm_gmem *gmem; /* * Bindings must be stable across invalidation to ensure the start+end @@ -164,13 +193,11 @@ static long kvm_gmem_punch_hole(struct inode *inode, loff_t offset, loff_t len) */ filemap_invalidate_lock(inode->i_mapping); - list_for_each_entry(gmem, gmem_list, entry) - kvm_gmem_invalidate_begin(gmem, start, end); + kvm_gmem_invalidate_begin(inode, start, end); truncate_inode_pages_range(inode->i_mapping, offset, offset + len - 1); - list_for_each_entry(gmem, gmem_list, entry) - kvm_gmem_invalidate_end(gmem, start, end); + kvm_gmem_invalidate_end(inode, start, end); filemap_invalidate_unlock(inode->i_mapping); @@ -280,8 +307,9 @@ static int kvm_gmem_release(struct inode *inode, struct file *file) * Zap all SPTEs pointed at by this file. Do not free the backing * memory, as its lifetime is associated with the inode, not the file. */ - kvm_gmem_invalidate_begin(gmem, 0, -1ul); - kvm_gmem_invalidate_end(gmem, 0, -1ul); + __kvm_gmem_invalidate_begin(gmem, 0, -1ul, + kvm_gmem_get_invalidate_filter(inode)); + __kvm_gmem_invalidate_end(gmem, 0, -1ul); list_del(&gmem->entry); @@ -403,8 +431,6 @@ static int kvm_gmem_migrate_folio(struct address_space *mapping, static int kvm_gmem_error_folio(struct address_space *mapping, struct folio *folio) { - struct list_head *gmem_list = &mapping->i_private_list; - struct kvm_gmem *gmem; pgoff_t start, end; filemap_invalidate_lock_shared(mapping); @@ -412,8 +438,7 @@ static int kvm_gmem_error_folio(struct address_space *mapping, struct folio *fol start = folio->index; end = start + folio_nr_pages(folio); - list_for_each_entry(gmem, gmem_list, entry) - kvm_gmem_invalidate_begin(gmem, start, end); + kvm_gmem_invalidate_begin(mapping->host, start, end); /* * Do not truncate the range, what action is taken in response to the @@ -424,8 +449,7 @@ static int kvm_gmem_error_folio(struct address_space *mapping, struct folio *fol * error to userspace. */ - list_for_each_entry(gmem, gmem_list, entry) - kvm_gmem_invalidate_end(gmem, start, end); + kvm_gmem_invalidate_end(mapping->host, start, end); filemap_invalidate_unlock_shared(mapping); From 9aef71c892a55e004419923ba7129abe3e58d9f1 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:25:57 -0700 Subject: [PATCH 023/330] KVM: Explicitly mark KVM_GUEST_MEMFD as depending on KVM_GENERIC_MMU_NOTIFIER Add KVM_GENERIC_MMU_NOTIFIER as a dependency for selecting KVM_GUEST_MEMFD, as guest_memfd relies on kvm_mmu_invalidate_{begin,end}(), which are defined if and only if the generic mmu_notifier implementation is enabled. The missing dependency is currently benign as s390 is the only KVM arch that doesn't utilize the generic mmu_notifier infrastructure, and s390 doesn't currently support guest_memfd. Fixes: a7800aa80ea4 ("KVM: Add KVM_CREATE_GUEST_MEMFD ioctl() for guest-specific backing memory") Reviewed-by: David Hildenbrand Link: https://lore.kernel.org/r/20251003232606.4070510-5-seanjc@google.com Signed-off-by: Sean Christopherson --- virt/kvm/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/virt/kvm/Kconfig b/virt/kvm/Kconfig index 1b7d5be0b6c4..a01cc5743137 100644 --- a/virt/kvm/Kconfig +++ b/virt/kvm/Kconfig @@ -113,6 +113,7 @@ config KVM_GENERIC_MEMORY_ATTRIBUTES bool config KVM_GUEST_MEMFD + depends on KVM_GENERIC_MMU_NOTIFIER select XARRAY_MULTI bool From 44c6cb9fe9888b371e31165b2854bd0f4e2787d4 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:25:58 -0700 Subject: [PATCH 024/330] KVM: guest_memfd: Allow mmap() on guest_memfd for x86 VMs with private memory Allow mmap() on guest_memfd instances for x86 VMs with private memory as the need to track private vs. shared state in the guest_memfd instance is only pertinent to INIT_SHARED. Doing mmap() on private memory isn't terrible useful (yet!), but it's now possible, and will be desirable when guest_memfd gains support for other VMA-based syscalls, e.g. mbind() to set NUMA policy. Lift the restriction now, before MMAP support is officially released, so that KVM doesn't need to add another capability to enumerate support for mmap() on private memory. Fixes: 3d3a04fad25a ("KVM: Allow and advertise support for host mmap() on guest_memfd files") Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Reviewed-by: David Hildenbrand Link: https://lore.kernel.org/r/20251003232606.4070510-6-seanjc@google.com Signed-off-by: Sean Christopherson --- arch/x86/kvm/x86.c | 7 ++++--- include/linux/kvm_host.h | 12 +++++++++++- virt/kvm/guest_memfd.c | 9 ++------- virt/kvm/kvm_main.c | 6 +----- 4 files changed, 18 insertions(+), 16 deletions(-) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 4b8138bd4857..fe3dc3eb4331 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -13942,10 +13942,11 @@ bool kvm_arch_no_poll(struct kvm_vcpu *vcpu) #ifdef CONFIG_KVM_GUEST_MEMFD /* - * KVM doesn't yet support mmap() on guest_memfd for VMs with private memory - * (the private vs. shared tracking needs to be moved into guest_memfd). + * KVM doesn't yet support initializing guest_memfd memory as shared for VMs + * with private memory (the private vs. shared tracking needs to be moved into + * guest_memfd). */ -bool kvm_arch_supports_gmem_mmap(struct kvm *kvm) +bool kvm_arch_supports_gmem_init_shared(struct kvm *kvm) { return !kvm_arch_has_private_mem(kvm); } diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h index 19b8c4bebb9c..680ca838f018 100644 --- a/include/linux/kvm_host.h +++ b/include/linux/kvm_host.h @@ -729,7 +729,17 @@ static inline bool kvm_arch_has_private_mem(struct kvm *kvm) #endif #ifdef CONFIG_KVM_GUEST_MEMFD -bool kvm_arch_supports_gmem_mmap(struct kvm *kvm); +bool kvm_arch_supports_gmem_init_shared(struct kvm *kvm); + +static inline u64 kvm_gmem_get_supported_flags(struct kvm *kvm) +{ + u64 flags = GUEST_MEMFD_FLAG_MMAP; + + if (!kvm || kvm_arch_supports_gmem_init_shared(kvm)) + flags |= GUEST_MEMFD_FLAG_INIT_SHARED; + + return flags; +} #endif #ifndef kvm_arch_has_readonly_mem diff --git a/virt/kvm/guest_memfd.c b/virt/kvm/guest_memfd.c index e10d2c71e78c..fbca8c0972da 100644 --- a/virt/kvm/guest_memfd.c +++ b/virt/kvm/guest_memfd.c @@ -485,7 +485,7 @@ static const struct inode_operations kvm_gmem_iops = { .setattr = kvm_gmem_setattr, }; -bool __weak kvm_arch_supports_gmem_mmap(struct kvm *kvm) +bool __weak kvm_arch_supports_gmem_init_shared(struct kvm *kvm) { return true; } @@ -549,13 +549,8 @@ int kvm_gmem_create(struct kvm *kvm, struct kvm_create_guest_memfd *args) { loff_t size = args->size; u64 flags = args->flags; - u64 valid_flags = 0; - if (kvm_arch_supports_gmem_mmap(kvm)) - valid_flags |= GUEST_MEMFD_FLAG_MMAP | - GUEST_MEMFD_FLAG_INIT_SHARED; - - if (flags & ~valid_flags) + if (flags & ~kvm_gmem_get_supported_flags(kvm)) return -EINVAL; if (size <= 0 || !PAGE_ALIGNED(size)) diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c index 5f644ca54af3..b7a0ae2a7b20 100644 --- a/virt/kvm/kvm_main.c +++ b/virt/kvm/kvm_main.c @@ -4929,11 +4929,7 @@ static int kvm_vm_ioctl_check_extension_generic(struct kvm *kvm, long arg) case KVM_CAP_GUEST_MEMFD: return 1; case KVM_CAP_GUEST_MEMFD_FLAGS: - if (!kvm || kvm_arch_supports_gmem_mmap(kvm)) - return GUEST_MEMFD_FLAG_MMAP | - GUEST_MEMFD_FLAG_INIT_SHARED; - - return 0; + return kvm_gmem_get_supported_flags(kvm); #endif default: break; From 3a6c08538c742624c60cb6a53dd61eb025e0d1e1 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:25:59 -0700 Subject: [PATCH 025/330] KVM: selftests: Stash the host page size in a global in the guest_memfd test Use a global variable to track the host page size in the guest_memfd test so that the information doesn't need to be constantly passed around. The state is purely a reflection of the underlying system, i.e. can't be set by the test and is constant for a given invocation of the test, and thus explicitly passing the host page size to individual testcases adds no value, e.g. doesn't allow testing different combinations. Making page_size a global will simplify an upcoming change to create a new guest_memfd instance per testcase. No functional change intended. Reviewed-by: Fuad Tabba Tested-by: Fuad Tabba Reviewed-by: David Hildenbrand Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Link: https://lore.kernel.org/r/20251003232606.4070510-7-seanjc@google.com Signed-off-by: Sean Christopherson --- .../testing/selftests/kvm/guest_memfd_test.c | 37 +++++++++---------- 1 file changed, 18 insertions(+), 19 deletions(-) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index 0de56ce3c4e2..a7c9601bd31e 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -24,6 +24,8 @@ #include "test_util.h" #include "ucall_common.h" +static size_t page_size; + static void test_file_read_write(int fd) { char buf[64]; @@ -38,7 +40,7 @@ static void test_file_read_write(int fd) "pwrite on a guest_mem fd should fail"); } -static void test_mmap_supported(int fd, size_t page_size, size_t total_size) +static void test_mmap_supported(int fd, size_t total_size) { const char val = 0xaa; char *mem; @@ -78,7 +80,7 @@ void fault_sigbus_handler(int signum) siglongjmp(jmpbuf, 1); } -static void test_fault_overflow(int fd, size_t page_size, size_t total_size) +static void test_fault_overflow(int fd, size_t total_size) { struct sigaction sa_old, sa_new = { .sa_handler = fault_sigbus_handler, @@ -106,7 +108,7 @@ static void test_fault_overflow(int fd, size_t page_size, size_t total_size) TEST_ASSERT(!ret, "munmap() should succeed."); } -static void test_mmap_not_supported(int fd, size_t page_size, size_t total_size) +static void test_mmap_not_supported(int fd, size_t total_size) { char *mem; @@ -117,7 +119,7 @@ static void test_mmap_not_supported(int fd, size_t page_size, size_t total_size) TEST_ASSERT_EQ(mem, MAP_FAILED); } -static void test_file_size(int fd, size_t page_size, size_t total_size) +static void test_file_size(int fd, size_t total_size) { struct stat sb; int ret; @@ -128,7 +130,7 @@ static void test_file_size(int fd, size_t page_size, size_t total_size) TEST_ASSERT_EQ(sb.st_blksize, page_size); } -static void test_fallocate(int fd, size_t page_size, size_t total_size) +static void test_fallocate(int fd, size_t total_size) { int ret; @@ -165,7 +167,7 @@ static void test_fallocate(int fd, size_t page_size, size_t total_size) TEST_ASSERT(!ret, "fallocate to restore punched hole should succeed"); } -static void test_invalid_punch_hole(int fd, size_t page_size, size_t total_size) +static void test_invalid_punch_hole(int fd, size_t total_size) { struct { off_t offset; @@ -196,8 +198,7 @@ static void test_invalid_punch_hole(int fd, size_t page_size, size_t total_size) } static void test_create_guest_memfd_invalid_sizes(struct kvm_vm *vm, - uint64_t guest_memfd_flags, - size_t page_size) + uint64_t guest_memfd_flags) { size_t size; int fd; @@ -214,7 +215,6 @@ static void test_create_guest_memfd_multiple(struct kvm_vm *vm) { int fd1, fd2, ret; struct stat st1, st2; - size_t page_size = getpagesize(); fd1 = __vm_create_guest_memfd(vm, page_size, 0); TEST_ASSERT(fd1 != -1, "memfd creation should succeed"); @@ -242,7 +242,6 @@ static void test_create_guest_memfd_multiple(struct kvm_vm *vm) static void test_guest_memfd_flags(struct kvm_vm *vm) { uint64_t valid_flags = vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS); - size_t page_size = getpagesize(); uint64_t flag; int fd; @@ -265,11 +264,9 @@ static void test_guest_memfd(unsigned long vm_type) { struct kvm_vm *vm; size_t total_size; - size_t page_size; uint64_t flags; int fd; - page_size = getpagesize(); total_size = page_size * 4; vm = vm_create_barebones_type(vm_type); @@ -280,22 +277,22 @@ static void test_guest_memfd(unsigned long vm_type) flags &= ~GUEST_MEMFD_FLAG_MMAP; test_create_guest_memfd_multiple(vm); - test_create_guest_memfd_invalid_sizes(vm, flags, page_size); + test_create_guest_memfd_invalid_sizes(vm, flags); fd = vm_create_guest_memfd(vm, total_size, flags); test_file_read_write(fd); if (flags & GUEST_MEMFD_FLAG_MMAP) { - test_mmap_supported(fd, page_size, total_size); - test_fault_overflow(fd, page_size, total_size); + test_mmap_supported(fd, total_size); + test_fault_overflow(fd, total_size); } else { - test_mmap_not_supported(fd, page_size, total_size); + test_mmap_not_supported(fd, total_size); } - test_file_size(fd, page_size, total_size); - test_fallocate(fd, page_size, total_size); - test_invalid_punch_hole(fd, page_size, total_size); + test_file_size(fd, total_size); + test_fallocate(fd, total_size); + test_invalid_punch_hole(fd, total_size); test_guest_memfd_flags(vm); @@ -374,6 +371,8 @@ int main(int argc, char *argv[]) TEST_REQUIRE(kvm_has_cap(KVM_CAP_GUEST_MEMFD)); + page_size = getpagesize(); + /* * Not all architectures support KVM_CAP_VM_TYPES. However, those that * support guest_memfd have that support for the default VM type. From 21d602ed616aebae4de568be55db65a1f5a3be10 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:26:00 -0700 Subject: [PATCH 026/330] KVM: selftests: Create a new guest_memfd for each testcase Refactor the guest_memfd selftest to improve test isolation by creating a a new guest_memfd for each testcase. Currently, the test reuses a single guest_memfd instance for all testcases, and thus creates dependencies between tests, e.g. not truncating folios from the guest_memfd instance at the end of a test could lead to unexpected results (see the PUNCH_HOLE purging that needs to done by in-flight the NUMA testcases[1]). Invoke each test via a macro wrapper to create and close a guest_memfd to cut down on the boilerplate copy+paste needed to create a test. Link: https://lore.kernel.org/all/20250827175247.83322-10-shivankg@amd.com Reported-by: Ackerley Tng Reviewed-by: Fuad Tabba Tested-by: Fuad Tabba Reviewed-by: David Hildenbrand Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Link: https://lore.kernel.org/r/20251003232606.4070510-8-seanjc@google.com Signed-off-by: Sean Christopherson --- .../testing/selftests/kvm/guest_memfd_test.c | 31 ++++++++++--------- 1 file changed, 16 insertions(+), 15 deletions(-) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index a7c9601bd31e..afdc4d3a956d 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -26,7 +26,7 @@ static size_t page_size; -static void test_file_read_write(int fd) +static void test_file_read_write(int fd, size_t total_size) { char buf[64]; @@ -260,14 +260,18 @@ static void test_guest_memfd_flags(struct kvm_vm *vm) } } +#define gmem_test(__test, __vm, __flags) \ +do { \ + int fd = vm_create_guest_memfd(__vm, page_size * 4, __flags); \ + \ + test_##__test(fd, page_size * 4); \ + close(fd); \ +} while (0) + static void test_guest_memfd(unsigned long vm_type) { struct kvm_vm *vm; - size_t total_size; uint64_t flags; - int fd; - - total_size = page_size * 4; vm = vm_create_barebones_type(vm_type); flags = vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS); @@ -279,24 +283,21 @@ static void test_guest_memfd(unsigned long vm_type) test_create_guest_memfd_multiple(vm); test_create_guest_memfd_invalid_sizes(vm, flags); - fd = vm_create_guest_memfd(vm, total_size, flags); - - test_file_read_write(fd); + gmem_test(file_read_write, vm, flags); if (flags & GUEST_MEMFD_FLAG_MMAP) { - test_mmap_supported(fd, total_size); - test_fault_overflow(fd, total_size); + gmem_test(mmap_supported, vm, flags); + gmem_test(fault_overflow, vm, flags); } else { - test_mmap_not_supported(fd, total_size); + gmem_test(mmap_not_supported, vm, flags); } - test_file_size(fd, total_size); - test_fallocate(fd, total_size); - test_invalid_punch_hole(fd, total_size); + gmem_test(file_size, vm, flags); + gmem_test(fallocate, vm, flags); + gmem_test(invalid_punch_hole, vm, flags); test_guest_memfd_flags(vm); - close(fd); kvm_vm_free(vm); } From df0d9923f7055169cb01b050236e5a9a2b36db02 Mon Sep 17 00:00:00 2001 From: Ackerley Tng Date: Fri, 3 Oct 2025 16:26:01 -0700 Subject: [PATCH 027/330] KVM: selftests: Add test coverage for guest_memfd without GUEST_MEMFD_FLAG_MMAP If a VM type supports KVM_CAP_GUEST_MEMFD_MMAP, the guest_memfd test will run all test cases with GUEST_MEMFD_FLAG_MMAP set. This leaves the code path for creating a non-mmap()-able guest_memfd on a VM that supports mappable guest memfds untested. Refactor the test to run the main test suite with a given set of flags. Then, for VM types that support the mappable capability, invoke the test suite twice: once with no flags, and once with GUEST_MEMFD_FLAG_MMAP set. This ensures both creation paths are properly exercised on capable VMs. Run test_guest_memfd_flags() only once per VM type since it depends only on the set of valid/supported flags, i.e. iterating over an arbitrary set of flags is both unnecessary and wrong. Signed-off-by: Ackerley Tng [sean: use double-underscores for the inner helper] Reviewed-by: Fuad Tabba Tested-by: Fuad Tabba Reviewed-by: David Hildenbrand Link: https://lore.kernel.org/r/20251003232606.4070510-9-seanjc@google.com Signed-off-by: Sean Christopherson --- .../testing/selftests/kvm/guest_memfd_test.c | 27 +++++++++++-------- 1 file changed, 16 insertions(+), 11 deletions(-) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index afdc4d3a956d..9f98a067ab51 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -268,18 +268,8 @@ do { \ close(fd); \ } while (0) -static void test_guest_memfd(unsigned long vm_type) +static void __test_guest_memfd(struct kvm_vm *vm, uint64_t flags) { - struct kvm_vm *vm; - uint64_t flags; - - vm = vm_create_barebones_type(vm_type); - flags = vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS); - - /* This test doesn't yet support testing mmap() on private memory. */ - if (!(flags & GUEST_MEMFD_FLAG_INIT_SHARED)) - flags &= ~GUEST_MEMFD_FLAG_MMAP; - test_create_guest_memfd_multiple(vm); test_create_guest_memfd_invalid_sizes(vm, flags); @@ -295,9 +285,24 @@ static void test_guest_memfd(unsigned long vm_type) gmem_test(file_size, vm, flags); gmem_test(fallocate, vm, flags); gmem_test(invalid_punch_hole, vm, flags); +} + +static void test_guest_memfd(unsigned long vm_type) +{ + struct kvm_vm *vm = vm_create_barebones_type(vm_type); + uint64_t flags; test_guest_memfd_flags(vm); + __test_guest_memfd(vm, 0); + + flags = vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS); + + /* MMAP should always be supported if INIT_SHARED is supported. */ + if (flags & GUEST_MEMFD_FLAG_INIT_SHARED) + __test_guest_memfd(vm, GUEST_MEMFD_FLAG_MMAP | + GUEST_MEMFD_FLAG_INIT_SHARED); + kvm_vm_free(vm); } From 61cee97f40180312dcca9580a5be0b0aa2217f6e Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:26:02 -0700 Subject: [PATCH 028/330] KVM: selftests: Add wrappers for mmap() and munmap() to assert success Add and use wrappers for mmap() and munmap() that assert success to reduce a significant amount of boilerplate code, to ensure all tests assert on failure, and to provide consistent error messages on failure. No functional change intended. Reviewed-by: Fuad Tabba Tested-by: Fuad Tabba Reviewed-by: David Hildenbrand Reviewed-by: Ackerley Tng Link: https://lore.kernel.org/r/20251003232606.4070510-10-seanjc@google.com Signed-off-by: Sean Christopherson --- .../testing/selftests/kvm/guest_memfd_test.c | 21 +++------ .../testing/selftests/kvm/include/kvm_util.h | 25 +++++++++++ tools/testing/selftests/kvm/lib/kvm_util.c | 44 +++++++------------ tools/testing/selftests/kvm/mmu_stress_test.c | 5 +-- .../selftests/kvm/s390/ucontrol_test.c | 16 +++---- .../selftests/kvm/set_memory_region_test.c | 17 ++++--- 6 files changed, 64 insertions(+), 64 deletions(-) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index 9f98a067ab51..319fda4f5d53 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -50,8 +50,7 @@ static void test_mmap_supported(int fd, size_t total_size) mem = mmap(NULL, total_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0); TEST_ASSERT(mem == MAP_FAILED, "Copy-on-write not allowed by guest_memfd."); - mem = mmap(NULL, total_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); - TEST_ASSERT(mem != MAP_FAILED, "mmap() for guest_memfd should succeed."); + mem = kvm_mmap(total_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd); memset(mem, val, total_size); for (i = 0; i < total_size; i++) @@ -70,8 +69,7 @@ static void test_mmap_supported(int fd, size_t total_size) for (i = 0; i < total_size; i++) TEST_ASSERT_EQ(READ_ONCE(mem[i]), val); - ret = munmap(mem, total_size); - TEST_ASSERT(!ret, "munmap() should succeed."); + kvm_munmap(mem, total_size); } static sigjmp_buf jmpbuf; @@ -89,10 +87,8 @@ static void test_fault_overflow(int fd, size_t total_size) const char val = 0xaa; char *mem; size_t i; - int ret; - mem = mmap(NULL, map_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); - TEST_ASSERT(mem != MAP_FAILED, "mmap() for guest_memfd should succeed."); + mem = kvm_mmap(map_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd); sigaction(SIGBUS, &sa_new, &sa_old); if (sigsetjmp(jmpbuf, 1) == 0) { @@ -104,8 +100,7 @@ static void test_fault_overflow(int fd, size_t total_size) for (i = 0; i < total_size; i++) TEST_ASSERT_EQ(READ_ONCE(mem[i]), val); - ret = munmap(mem, map_size); - TEST_ASSERT(!ret, "munmap() should succeed."); + kvm_munmap(mem, map_size); } static void test_mmap_not_supported(int fd, size_t total_size) @@ -351,10 +346,9 @@ static void test_guest_memfd_guest(void) GUEST_MEMFD_FLAG_INIT_SHARED); vm_set_user_memory_region2(vm, slot, KVM_MEM_GUEST_MEMFD, gpa, size, NULL, fd, 0); - mem = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); - TEST_ASSERT(mem != MAP_FAILED, "mmap() on guest_memfd failed"); + mem = kvm_mmap(size, PROT_READ | PROT_WRITE, MAP_SHARED, fd); memset(mem, 0xaa, size); - munmap(mem, size); + kvm_munmap(mem, size); virt_pg_map(vm, gpa, gpa); vcpu_args_set(vcpu, 2, gpa, size); @@ -362,8 +356,7 @@ static void test_guest_memfd_guest(void) TEST_ASSERT_EQ(get_ucall(vcpu, NULL), UCALL_DONE); - mem = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); - TEST_ASSERT(mem != MAP_FAILED, "mmap() on guest_memfd failed"); + mem = kvm_mmap(size, PROT_READ | PROT_WRITE, MAP_SHARED, fd); for (i = 0; i < size; i++) TEST_ASSERT_EQ(mem[i], 0xff); diff --git a/tools/testing/selftests/kvm/include/kvm_util.h b/tools/testing/selftests/kvm/include/kvm_util.h index 26cc30290e76..ee60dbf5208a 100644 --- a/tools/testing/selftests/kvm/include/kvm_util.h +++ b/tools/testing/selftests/kvm/include/kvm_util.h @@ -286,6 +286,31 @@ static inline bool kvm_has_cap(long cap) #define __KVM_SYSCALL_ERROR(_name, _ret) \ "%s failed, rc: %i errno: %i (%s)", (_name), (_ret), errno, strerror(errno) +static inline void *__kvm_mmap(size_t size, int prot, int flags, int fd, + off_t offset) +{ + void *mem; + + mem = mmap(NULL, size, prot, flags, fd, offset); + TEST_ASSERT(mem != MAP_FAILED, __KVM_SYSCALL_ERROR("mmap()", + (int)(unsigned long)MAP_FAILED)); + + return mem; +} + +static inline void *kvm_mmap(size_t size, int prot, int flags, int fd) +{ + return __kvm_mmap(size, prot, flags, fd, 0); +} + +static inline void kvm_munmap(void *mem, size_t size) +{ + int ret; + + ret = munmap(mem, size); + TEST_ASSERT(!ret, __KVM_SYSCALL_ERROR("munmap()", ret)); +} + /* * Use the "inner", double-underscore macro when reporting errors from within * other macros so that the name of ioctl() and not its literal numeric value diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c index 6743fbd9bd67..83a721be7ec5 100644 --- a/tools/testing/selftests/kvm/lib/kvm_util.c +++ b/tools/testing/selftests/kvm/lib/kvm_util.c @@ -741,13 +741,11 @@ static void vm_vcpu_rm(struct kvm_vm *vm, struct kvm_vcpu *vcpu) int ret; if (vcpu->dirty_gfns) { - ret = munmap(vcpu->dirty_gfns, vm->dirty_ring_size); - TEST_ASSERT(!ret, __KVM_SYSCALL_ERROR("munmap()", ret)); + kvm_munmap(vcpu->dirty_gfns, vm->dirty_ring_size); vcpu->dirty_gfns = NULL; } - ret = munmap(vcpu->run, vcpu_mmap_sz()); - TEST_ASSERT(!ret, __KVM_SYSCALL_ERROR("munmap()", ret)); + kvm_munmap(vcpu->run, vcpu_mmap_sz()); ret = close(vcpu->fd); TEST_ASSERT(!ret, __KVM_SYSCALL_ERROR("close()", ret)); @@ -783,20 +781,16 @@ void kvm_vm_release(struct kvm_vm *vmp) static void __vm_mem_region_delete(struct kvm_vm *vm, struct userspace_mem_region *region) { - int ret; - rb_erase(®ion->gpa_node, &vm->regions.gpa_tree); rb_erase(®ion->hva_node, &vm->regions.hva_tree); hash_del(®ion->slot_node); sparsebit_free(®ion->unused_phy_pages); sparsebit_free(®ion->protected_phy_pages); - ret = munmap(region->mmap_start, region->mmap_size); - TEST_ASSERT(!ret, __KVM_SYSCALL_ERROR("munmap()", ret)); + kvm_munmap(region->mmap_start, region->mmap_size); if (region->fd >= 0) { /* There's an extra map when using shared memory. */ - ret = munmap(region->mmap_alias, region->mmap_size); - TEST_ASSERT(!ret, __KVM_SYSCALL_ERROR("munmap()", ret)); + kvm_munmap(region->mmap_alias, region->mmap_size); close(region->fd); } if (region->region.guest_memfd >= 0) @@ -1053,12 +1047,9 @@ void vm_mem_add(struct kvm_vm *vm, enum vm_mem_backing_src_type src_type, region->fd = kvm_memfd_alloc(region->mmap_size, src_type == VM_MEM_SRC_SHARED_HUGETLB); - region->mmap_start = mmap(NULL, region->mmap_size, - PROT_READ | PROT_WRITE, - vm_mem_backing_src_alias(src_type)->flag, - region->fd, 0); - TEST_ASSERT(region->mmap_start != MAP_FAILED, - __KVM_SYSCALL_ERROR("mmap()", (int)(unsigned long)MAP_FAILED)); + region->mmap_start = kvm_mmap(region->mmap_size, PROT_READ | PROT_WRITE, + vm_mem_backing_src_alias(src_type)->flag, + region->fd); TEST_ASSERT(!is_backing_src_hugetlb(src_type) || region->mmap_start == align_ptr_up(region->mmap_start, backing_src_pagesz), @@ -1129,12 +1120,10 @@ void vm_mem_add(struct kvm_vm *vm, enum vm_mem_backing_src_type src_type, /* If shared memory, create an alias. */ if (region->fd >= 0) { - region->mmap_alias = mmap(NULL, region->mmap_size, - PROT_READ | PROT_WRITE, - vm_mem_backing_src_alias(src_type)->flag, - region->fd, 0); - TEST_ASSERT(region->mmap_alias != MAP_FAILED, - __KVM_SYSCALL_ERROR("mmap()", (int)(unsigned long)MAP_FAILED)); + region->mmap_alias = kvm_mmap(region->mmap_size, + PROT_READ | PROT_WRITE, + vm_mem_backing_src_alias(src_type)->flag, + region->fd); /* Align host alias address */ region->host_alias = align_ptr_up(region->mmap_alias, alignment); @@ -1344,10 +1333,8 @@ struct kvm_vcpu *__vm_vcpu_add(struct kvm_vm *vm, uint32_t vcpu_id) TEST_ASSERT(vcpu_mmap_sz() >= sizeof(*vcpu->run), "vcpu mmap size " "smaller than expected, vcpu_mmap_sz: %zi expected_min: %zi", vcpu_mmap_sz(), sizeof(*vcpu->run)); - vcpu->run = (struct kvm_run *) mmap(NULL, vcpu_mmap_sz(), - PROT_READ | PROT_WRITE, MAP_SHARED, vcpu->fd, 0); - TEST_ASSERT(vcpu->run != MAP_FAILED, - __KVM_SYSCALL_ERROR("mmap()", (int)(unsigned long)MAP_FAILED)); + vcpu->run = kvm_mmap(vcpu_mmap_sz(), PROT_READ | PROT_WRITE, + MAP_SHARED, vcpu->fd); if (kvm_has_cap(KVM_CAP_BINARY_STATS_FD)) vcpu->stats.fd = vcpu_get_stats_fd(vcpu); @@ -1794,9 +1781,8 @@ void *vcpu_map_dirty_ring(struct kvm_vcpu *vcpu) page_size * KVM_DIRTY_LOG_PAGE_OFFSET); TEST_ASSERT(addr == MAP_FAILED, "Dirty ring mapped exec"); - addr = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_SHARED, vcpu->fd, - page_size * KVM_DIRTY_LOG_PAGE_OFFSET); - TEST_ASSERT(addr != MAP_FAILED, "Dirty ring map failed"); + addr = __kvm_mmap(size, PROT_READ | PROT_WRITE, MAP_SHARED, vcpu->fd, + page_size * KVM_DIRTY_LOG_PAGE_OFFSET); vcpu->dirty_gfns = addr; vcpu->dirty_gfns_count = size / sizeof(struct kvm_dirty_gfn); diff --git a/tools/testing/selftests/kvm/mmu_stress_test.c b/tools/testing/selftests/kvm/mmu_stress_test.c index 6a437d2be9fa..37b7e6524533 100644 --- a/tools/testing/selftests/kvm/mmu_stress_test.c +++ b/tools/testing/selftests/kvm/mmu_stress_test.c @@ -339,8 +339,7 @@ int main(int argc, char *argv[]) TEST_ASSERT(max_gpa > (4 * slot_size), "MAXPHYADDR <4gb "); fd = kvm_memfd_alloc(slot_size, hugepages); - mem = mmap(NULL, slot_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); - TEST_ASSERT(mem != MAP_FAILED, "mmap() failed"); + mem = kvm_mmap(slot_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd); TEST_ASSERT(!madvise(mem, slot_size, MADV_NOHUGEPAGE), "madvise() failed"); @@ -413,7 +412,7 @@ int main(int argc, char *argv[]) for (slot = (slot - 1) & ~1ull; slot >= first_slot; slot -= 2) vm_set_user_memory_region(vm, slot, 0, 0, 0, NULL); - munmap(mem, slot_size / 2); + kvm_munmap(mem, slot_size / 2); /* Sanity check that the vCPUs actually ran. */ for (i = 0; i < nr_vcpus; i++) diff --git a/tools/testing/selftests/kvm/s390/ucontrol_test.c b/tools/testing/selftests/kvm/s390/ucontrol_test.c index d265b34c54be..50bc1c38225a 100644 --- a/tools/testing/selftests/kvm/s390/ucontrol_test.c +++ b/tools/testing/selftests/kvm/s390/ucontrol_test.c @@ -142,19 +142,17 @@ FIXTURE_SETUP(uc_kvm) self->kvm_run_size = ioctl(self->kvm_fd, KVM_GET_VCPU_MMAP_SIZE, NULL); ASSERT_GE(self->kvm_run_size, sizeof(struct kvm_run)) TH_LOG(KVM_IOCTL_ERROR(KVM_GET_VCPU_MMAP_SIZE, self->kvm_run_size)); - self->run = (struct kvm_run *)mmap(NULL, self->kvm_run_size, - PROT_READ | PROT_WRITE, MAP_SHARED, self->vcpu_fd, 0); - ASSERT_NE(self->run, MAP_FAILED); + self->run = kvm_mmap(self->kvm_run_size, PROT_READ | PROT_WRITE, + MAP_SHARED, self->vcpu_fd); /** * For virtual cpus that have been created with S390 user controlled * virtual machines, the resulting vcpu fd can be memory mapped at page * offset KVM_S390_SIE_PAGE_OFFSET in order to obtain a memory map of * the virtual cpu's hardware control block. */ - self->sie_block = (struct kvm_s390_sie_block *)mmap(NULL, PAGE_SIZE, - PROT_READ | PROT_WRITE, MAP_SHARED, - self->vcpu_fd, KVM_S390_SIE_PAGE_OFFSET << PAGE_SHIFT); - ASSERT_NE(self->sie_block, MAP_FAILED); + self->sie_block = __kvm_mmap(PAGE_SIZE, PROT_READ | PROT_WRITE, + MAP_SHARED, self->vcpu_fd, + KVM_S390_SIE_PAGE_OFFSET << PAGE_SHIFT); TH_LOG("VM created %p %p", self->run, self->sie_block); @@ -186,8 +184,8 @@ FIXTURE_SETUP(uc_kvm) FIXTURE_TEARDOWN(uc_kvm) { - munmap(self->sie_block, PAGE_SIZE); - munmap(self->run, self->kvm_run_size); + kvm_munmap(self->sie_block, PAGE_SIZE); + kvm_munmap(self->run, self->kvm_run_size); close(self->vcpu_fd); close(self->vm_fd); close(self->kvm_fd); diff --git a/tools/testing/selftests/kvm/set_memory_region_test.c b/tools/testing/selftests/kvm/set_memory_region_test.c index ce3ac0fd6dfb..7fe427ff9b38 100644 --- a/tools/testing/selftests/kvm/set_memory_region_test.c +++ b/tools/testing/selftests/kvm/set_memory_region_test.c @@ -433,10 +433,10 @@ static void test_add_max_memory_regions(void) pr_info("Adding slots 0..%i, each memory region with %dK size\n", (max_mem_slots - 1), MEM_REGION_SIZE >> 10); - mem = mmap(NULL, (size_t)max_mem_slots * MEM_REGION_SIZE + alignment, - PROT_READ | PROT_WRITE, - MAP_PRIVATE | MAP_ANONYMOUS | MAP_NORESERVE, -1, 0); - TEST_ASSERT(mem != MAP_FAILED, "Failed to mmap() host"); + + mem = kvm_mmap((size_t)max_mem_slots * MEM_REGION_SIZE + alignment, + PROT_READ | PROT_WRITE, + MAP_PRIVATE | MAP_ANONYMOUS | MAP_NORESERVE, -1); mem_aligned = (void *)(((size_t) mem + alignment - 1) & ~(alignment - 1)); for (slot = 0; slot < max_mem_slots; slot++) @@ -446,9 +446,8 @@ static void test_add_max_memory_regions(void) mem_aligned + (uint64_t)slot * MEM_REGION_SIZE); /* Check it cannot be added memory slots beyond the limit */ - mem_extra = mmap(NULL, MEM_REGION_SIZE, PROT_READ | PROT_WRITE, - MAP_PRIVATE | MAP_ANONYMOUS, -1, 0); - TEST_ASSERT(mem_extra != MAP_FAILED, "Failed to mmap() host"); + mem_extra = kvm_mmap(MEM_REGION_SIZE, PROT_READ | PROT_WRITE, + MAP_PRIVATE | MAP_ANONYMOUS, -1); ret = __vm_set_user_memory_region(vm, max_mem_slots, 0, (uint64_t)max_mem_slots * MEM_REGION_SIZE, @@ -456,8 +455,8 @@ static void test_add_max_memory_regions(void) TEST_ASSERT(ret == -1 && errno == EINVAL, "Adding one more memory slot should fail with EINVAL"); - munmap(mem, (size_t)max_mem_slots * MEM_REGION_SIZE + alignment); - munmap(mem_extra, MEM_REGION_SIZE); + kvm_munmap(mem, (size_t)max_mem_slots * MEM_REGION_SIZE + alignment); + kvm_munmap(mem_extra, MEM_REGION_SIZE); kvm_vm_free(vm); } From 505c953009ec8260870d41ef8109bb4c7e208e6f Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:26:03 -0700 Subject: [PATCH 029/330] KVM: selftests: Isolate the guest_memfd Copy-on-Write negative testcase Move the guest_memfd Copy-on-Write (CoW) testcase to its own function to better separate positive testcases from negative testcases. No functional change intended. Suggested-by: Ackerley Tng Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Link: https://lore.kernel.org/r/20251003232606.4070510-11-seanjc@google.com Signed-off-by: Sean Christopherson --- tools/testing/selftests/kvm/guest_memfd_test.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index 319fda4f5d53..640636c76eb9 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -40,6 +40,14 @@ static void test_file_read_write(int fd, size_t total_size) "pwrite on a guest_mem fd should fail"); } +static void test_mmap_cow(int fd, size_t size) +{ + void *mem; + + mem = mmap(NULL, size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0); + TEST_ASSERT(mem == MAP_FAILED, "Copy-on-write not allowed by guest_memfd."); +} + static void test_mmap_supported(int fd, size_t total_size) { const char val = 0xaa; @@ -47,9 +55,6 @@ static void test_mmap_supported(int fd, size_t total_size) size_t i; int ret; - mem = mmap(NULL, total_size, PROT_READ | PROT_WRITE, MAP_PRIVATE, fd, 0); - TEST_ASSERT(mem == MAP_FAILED, "Copy-on-write not allowed by guest_memfd."); - mem = kvm_mmap(total_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd); memset(mem, val, total_size); @@ -272,6 +277,7 @@ static void __test_guest_memfd(struct kvm_vm *vm, uint64_t flags) if (flags & GUEST_MEMFD_FLAG_MMAP) { gmem_test(mmap_supported, vm, flags); + gmem_test(mmap_cow, vm, flags); gmem_test(fault_overflow, vm, flags); } else { gmem_test(mmap_not_supported, vm, flags); From f91187c0ecc6358ccecf533c5fcc6b7dbb4735cb Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:26:04 -0700 Subject: [PATCH 030/330] KVM: selftests: Add wrapper macro to handle and assert on expected SIGBUS Extract the guest_memfd test's SIGBUS handling functionality into a common TEST_EXPECT_SIGBUS() macro in anticipation of adding more SIGBUS testcases. Eating a SIGBUS isn't terrible difficult, but it requires a non-trivial amount of boilerplate code, and using a macro allows selftests to print out the exact action that failed to generate a SIGBUS without the developer needing to remember to add a useful error message. Explicitly mark the SIGBUS handler as "used", as gcc-14 at least likes to discard the function before linking. Opportunistically use TEST_FAIL(...) instead of TEST_ASSERT(false, ...), and fix the write path of the guest_memfd test to use the local "val" instead of hardcoding the literal value a second time. Suggested-by: Ackerley Tng Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Reviewed-by: Lisa Wang Tested-by: Lisa Wang Link: https://lore.kernel.org/r/20251003232606.4070510-12-seanjc@google.com Signed-off-by: Sean Christopherson --- .../testing/selftests/kvm/guest_memfd_test.c | 18 +----------------- .../testing/selftests/kvm/include/test_util.h | 19 +++++++++++++++++++ tools/testing/selftests/kvm/lib/test_util.c | 7 +++++++ 3 files changed, 27 insertions(+), 17 deletions(-) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index 640636c76eb9..73c2e54e7297 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -14,8 +14,6 @@ #include #include #include -#include -#include #include #include #include @@ -77,17 +75,8 @@ static void test_mmap_supported(int fd, size_t total_size) kvm_munmap(mem, total_size); } -static sigjmp_buf jmpbuf; -void fault_sigbus_handler(int signum) -{ - siglongjmp(jmpbuf, 1); -} - static void test_fault_overflow(int fd, size_t total_size) { - struct sigaction sa_old, sa_new = { - .sa_handler = fault_sigbus_handler, - }; size_t map_size = total_size * 4; const char val = 0xaa; char *mem; @@ -95,12 +84,7 @@ static void test_fault_overflow(int fd, size_t total_size) mem = kvm_mmap(map_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd); - sigaction(SIGBUS, &sa_new, &sa_old); - if (sigsetjmp(jmpbuf, 1) == 0) { - memset(mem, 0xaa, map_size); - TEST_ASSERT(false, "memset() should have triggered SIGBUS."); - } - sigaction(SIGBUS, &sa_old, NULL); + TEST_EXPECT_SIGBUS(memset(mem, val, map_size)); for (i = 0; i < total_size; i++) TEST_ASSERT_EQ(READ_ONCE(mem[i]), val); diff --git a/tools/testing/selftests/kvm/include/test_util.h b/tools/testing/selftests/kvm/include/test_util.h index c6ef895fbd9a..b4872ba8ed12 100644 --- a/tools/testing/selftests/kvm/include/test_util.h +++ b/tools/testing/selftests/kvm/include/test_util.h @@ -8,6 +8,8 @@ #ifndef SELFTEST_KVM_TEST_UTIL_H #define SELFTEST_KVM_TEST_UTIL_H +#include +#include #include #include #include @@ -78,6 +80,23 @@ do { \ __builtin_unreachable(); \ } while (0) +extern sigjmp_buf expect_sigbus_jmpbuf; +void expect_sigbus_handler(int signum); + +#define TEST_EXPECT_SIGBUS(action) \ +do { \ + struct sigaction sa_old, sa_new = { \ + .sa_handler = expect_sigbus_handler, \ + }; \ + \ + sigaction(SIGBUS, &sa_new, &sa_old); \ + if (sigsetjmp(expect_sigbus_jmpbuf, 1) == 0) { \ + action; \ + TEST_FAIL("'%s' should have triggered SIGBUS", #action); \ + } \ + sigaction(SIGBUS, &sa_old, NULL); \ +} while (0) + size_t parse_size(const char *size); int64_t timespec_to_ns(struct timespec ts); diff --git a/tools/testing/selftests/kvm/lib/test_util.c b/tools/testing/selftests/kvm/lib/test_util.c index 03eb99af9b8d..8a1848586a85 100644 --- a/tools/testing/selftests/kvm/lib/test_util.c +++ b/tools/testing/selftests/kvm/lib/test_util.c @@ -18,6 +18,13 @@ #include "test_util.h" +sigjmp_buf expect_sigbus_jmpbuf; + +void __attribute__((used)) expect_sigbus_handler(int signum) +{ + siglongjmp(expect_sigbus_jmpbuf, 1); +} + /* * Random number generator that is usable from guest code. This is the * Park-Miller LCG using standard constants. From 19942d4fd9cf022b28a9afb432a6338dcf96fc2f Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:26:05 -0700 Subject: [PATCH 031/330] KVM: selftests: Verify that faulting in private guest_memfd memory fails Add a guest_memfd testcase to verify that faulting in private memory gets a SIGBUS. For now, test only the case where memory is private by default since KVM doesn't yet support in-place conversion. Deliberately run the CoW test with and without INIT_SHARED set as KVM should disallow MAP_PRIVATE regardless of whether the memory itself is private from a CoCo perspective. Cc: Ackerley Tng Reviewed-by: Fuad Tabba Tested-by: Fuad Tabba Reviewed-by: David Hildenbrand Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Link: https://lore.kernel.org/r/20251003232606.4070510-13-seanjc@google.com Signed-off-by: Sean Christopherson --- .../testing/selftests/kvm/guest_memfd_test.c | 26 +++++++++++++++---- 1 file changed, 21 insertions(+), 5 deletions(-) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index 73c2e54e7297..f5372fdf096d 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -75,9 +75,8 @@ static void test_mmap_supported(int fd, size_t total_size) kvm_munmap(mem, total_size); } -static void test_fault_overflow(int fd, size_t total_size) +static void test_fault_sigbus(int fd, size_t accessible_size, size_t map_size) { - size_t map_size = total_size * 4; const char val = 0xaa; char *mem; size_t i; @@ -86,12 +85,22 @@ static void test_fault_overflow(int fd, size_t total_size) TEST_EXPECT_SIGBUS(memset(mem, val, map_size)); - for (i = 0; i < total_size; i++) + for (i = 0; i < accessible_size; i++) TEST_ASSERT_EQ(READ_ONCE(mem[i]), val); kvm_munmap(mem, map_size); } +static void test_fault_overflow(int fd, size_t total_size) +{ + test_fault_sigbus(fd, total_size, total_size * 4); +} + +static void test_fault_private(int fd, size_t total_size) +{ + test_fault_sigbus(fd, 0, total_size); +} + static void test_mmap_not_supported(int fd, size_t total_size) { char *mem; @@ -260,9 +269,14 @@ static void __test_guest_memfd(struct kvm_vm *vm, uint64_t flags) gmem_test(file_read_write, vm, flags); if (flags & GUEST_MEMFD_FLAG_MMAP) { - gmem_test(mmap_supported, vm, flags); + if (flags & GUEST_MEMFD_FLAG_INIT_SHARED) { + gmem_test(mmap_supported, vm, flags); + gmem_test(fault_overflow, vm, flags); + } else { + gmem_test(fault_private, vm, flags); + } + gmem_test(mmap_cow, vm, flags); - gmem_test(fault_overflow, vm, flags); } else { gmem_test(mmap_not_supported, vm, flags); } @@ -282,6 +296,8 @@ static void test_guest_memfd(unsigned long vm_type) __test_guest_memfd(vm, 0); flags = vm_check_cap(vm, KVM_CAP_GUEST_MEMFD_FLAGS); + if (flags & GUEST_MEMFD_FLAG_MMAP) + __test_guest_memfd(vm, GUEST_MEMFD_FLAG_MMAP); /* MMAP should always be supported if INIT_SHARED is supported. */ if (flags & GUEST_MEMFD_FLAG_INIT_SHARED) From 505f5224b197b77169c977e747cbc18b222f85f9 Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 3 Oct 2025 16:26:06 -0700 Subject: [PATCH 032/330] KVM: selftests: Verify that reads to inaccessible guest_memfd VMAs SIGBUS Expand the guest_memfd negative testcases for overflow and MAP_PRIVATE to verify that reads to inaccessible memory also get a SIGBUS. Reviewed-by: Ackerley Tng Tested-by: Ackerley Tng Reviewed-by: Lisa Wang Tested-by: Lisa Wang Link: https://lore.kernel.org/r/20251003232606.4070510-14-seanjc@google.com Signed-off-by: Sean Christopherson --- tools/testing/selftests/kvm/guest_memfd_test.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/testing/selftests/kvm/guest_memfd_test.c b/tools/testing/selftests/kvm/guest_memfd_test.c index f5372fdf096d..e7d9aeb418d3 100644 --- a/tools/testing/selftests/kvm/guest_memfd_test.c +++ b/tools/testing/selftests/kvm/guest_memfd_test.c @@ -84,6 +84,7 @@ static void test_fault_sigbus(int fd, size_t accessible_size, size_t map_size) mem = kvm_mmap(map_size, PROT_READ | PROT_WRITE, MAP_SHARED, fd); TEST_EXPECT_SIGBUS(memset(mem, val, map_size)); + TEST_EXPECT_SIGBUS((void)READ_ONCE(mem[accessible_size])); for (i = 0; i < accessible_size; i++) TEST_ASSERT_EQ(READ_ONCE(mem[i]), val); From 852947be66b826c3d0ba328e19a3559fdf7ac726 Mon Sep 17 00:00:00 2001 From: Paul Walmsley Date: Fri, 10 Oct 2025 15:50:24 -0600 Subject: [PATCH 033/330] riscv: kprobes: convert one final __ASSEMBLY__ to __ASSEMBLER__ Per the reasoning in commit f811f58597ac ("riscv: Replace __ASSEMBLY__ with __ASSEMBLER__ in non-uapi headers"), convert one last remaining instance of __ASSEMBLY__ in the arch/riscv kprobes code. This entered the tree from patches that were sent before Thomas' changes; and when I reviewed the kprobes patches before queuing them, I missed this instance. Cc: Nam Cao Cc: Thomas Huth Link: https://lore.kernel.org/linux-riscv/16b74b63-f223-4f0b-b6e5-31cea5e620b4@redhat.com/ Link: https://lore.kernel.org/linux-riscv/20250606070952.498274-1-thuth@redhat.com/ Signed-off-by: Paul Walmsley --- arch/riscv/kernel/tests/kprobes/test-kprobes.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/arch/riscv/kernel/tests/kprobes/test-kprobes.h b/arch/riscv/kernel/tests/kprobes/test-kprobes.h index 3886ab491ecb..537f44aa9d3f 100644 --- a/arch/riscv/kernel/tests/kprobes/test-kprobes.h +++ b/arch/riscv/kernel/tests/kprobes/test-kprobes.h @@ -11,7 +11,7 @@ #define KPROBE_TEST_MAGIC_LOWER 0x0000babe #define KPROBE_TEST_MAGIC_UPPER 0xcafe0000 -#ifndef __ASSEMBLY__ +#ifndef __ASSEMBLER__ /* array of addresses to install kprobes */ extern void *test_kprobes_addresses[]; @@ -19,6 +19,6 @@ extern void *test_kprobes_addresses[]; /* array of functions that return KPROBE_TEST_MAGIC */ extern long (*test_kprobes_functions[])(void); -#endif /* __ASSEMBLY__ */ +#endif /* __ASSEMBLER__ */ #endif /* TEST_KPROBES_H */ From 8527bbb33936340525a3504a00932b2f8fd75754 Mon Sep 17 00:00:00 2001 From: Denis Arefev Date: Tue, 7 Oct 2025 10:38:31 +0300 Subject: [PATCH 034/330] ALSA: hda: cs35l41: Fix NULL pointer dereference in cs35l41_get_acpi_mute_state() Return value of a function acpi_evaluate_dsm() is dereferenced without checking for NULL, but it is usually checked for this function. acpi_evaluate_dsm() may return NULL, when acpi_evaluate_object() returns acpi_status other than ACPI_SUCCESS, so add a check to prevent the crach. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: 447106e92a0c ("ALSA: hda: cs35l41: Support mute notifications for CS35L41 HDA") Cc: stable@vger.kernel.org Signed-off-by: Denis Arefev Signed-off-by: Takashi Iwai --- sound/hda/codecs/side-codecs/cs35l41_hda.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/sound/hda/codecs/side-codecs/cs35l41_hda.c b/sound/hda/codecs/side-codecs/cs35l41_hda.c index c04208e685a0..c0f2a3ff77a1 100644 --- a/sound/hda/codecs/side-codecs/cs35l41_hda.c +++ b/sound/hda/codecs/side-codecs/cs35l41_hda.c @@ -1410,6 +1410,8 @@ static int cs35l41_get_acpi_mute_state(struct cs35l41_hda *cs35l41, acpi_handle if (cs35l41_dsm_supported(handle, CS35L41_DSM_GET_MUTE)) { ret = acpi_evaluate_dsm(handle, &guid, 0, CS35L41_DSM_GET_MUTE, NULL); + if (!ret) + return -EINVAL; mute = *ret->buffer.pointer; dev_dbg(cs35l41->dev, "CS35L41_DSM_GET_MUTE: %d\n", mute); } From 1cf11d80db5df805b538c942269e05a65bcaf5bc Mon Sep 17 00:00:00 2001 From: Denis Arefev Date: Thu, 9 Oct 2025 13:50:47 +0300 Subject: [PATCH 035/330] ALSA: hda: Fix missing pointer check in hda_component_manager_init function The __component_match_add function may assign the 'matchptr' pointer the value ERR_PTR(-ENOMEM), which will subsequently be dereferenced. The call stack leading to the error looks like this: hda_component_manager_init |-> component_match_add |-> component_match_add_release |-> __component_match_add ( ... ,**matchptr, ... ) |-> *matchptr = ERR_PTR(-ENOMEM); // assign |-> component_master_add_with_match( ... match) |-> component_match_realloc(match, match->num); // dereference Add IS_ERR() check to prevent the crash. Found by Linux Verification Center (linuxtesting.org) with SVACE. Fixes: ae7abe36e352 ("ALSA: hda/realtek: Add CS35L41 support for Thinkpad laptops") Cc: stable@vger.kernel.org Signed-off-by: Denis Arefev Signed-off-by: Takashi Iwai --- sound/hda/codecs/side-codecs/hda_component.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/sound/hda/codecs/side-codecs/hda_component.c b/sound/hda/codecs/side-codecs/hda_component.c index bcf47a301697..8a2a200600a7 100644 --- a/sound/hda/codecs/side-codecs/hda_component.c +++ b/sound/hda/codecs/side-codecs/hda_component.c @@ -174,6 +174,10 @@ int hda_component_manager_init(struct hda_codec *cdc, sm->match_str = match_str; sm->index = i; component_match_add(dev, &match, hda_comp_match_dev_name, sm); + if (IS_ERR(match)) { + codec_err(cdc, "Fail to add component %ld\n", PTR_ERR(match)); + return PTR_ERR(match); + } } ret = component_master_add_with_match(dev, ops, match); From 30b3211aa24161856134b2c2ea2ab1c6eb534b36 Mon Sep 17 00:00:00 2001 From: Stuart Hayhurst Date: Sat, 11 Oct 2025 00:28:29 +0100 Subject: [PATCH 036/330] ALSA: hda/intel: Add MSI X870E Tomahawk to denylist This motherboard uses USB audio instead, causing this driver to complain about "no codecs found!". Add it to the denylist to silence the warning. Signed-off-by: Stuart Hayhurst Cc: Signed-off-by: Takashi Iwai --- sound/hda/controllers/intel.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/hda/controllers/intel.c b/sound/hda/controllers/intel.c index 48c52a207024..a19258c95886 100644 --- a/sound/hda/controllers/intel.c +++ b/sound/hda/controllers/intel.c @@ -2075,6 +2075,7 @@ static const struct pci_device_id driver_denylist[] = { { PCI_DEVICE_SUB(0x1022, 0x1487, 0x1043, 0x874f) }, /* ASUS ROG Zenith II / Strix */ { PCI_DEVICE_SUB(0x1022, 0x1487, 0x1462, 0xcb59) }, /* MSI TRX40 Creator */ { PCI_DEVICE_SUB(0x1022, 0x1487, 0x1462, 0xcb60) }, /* MSI TRX40 */ + { PCI_DEVICE_SUB(0x1022, 0x15e3, 0x1462, 0xee59) }, /* MSI X870E Tomahawk WiFi */ {} }; From 77908b81766781dfcd086878aefc29f5db8dae6a Mon Sep 17 00:00:00 2001 From: Cryolitia PukNgae Date: Sat, 11 Oct 2025 17:51:18 +0800 Subject: [PATCH 037/330] ALSA: usb-audio: apply quirk for Huawei Technologies Co., Ltd. CM-Q3 There're several different actual hardwares sold by Huawei, using the same USB ID 12d1:3a07. The first one we found, having a volume control named "Headset Playback Volume", reports a min value -15360, and will mute iff setting it to -15360. It can be simply fixed by quirk flag MIXER_PLAYBACK_MIN_MUTE, which we have already submitted previously.[1] The second one we found today, having a volume control named "PCM Playback Volume", reports its min -11520 and res 256, and will mute when less than -11008. Because of the already existing quirk flag, we can just set its min to -11264, and the new minimum value will still not be available to userspace, so that userspace's minimum will be the correct -11008. 1. https://lore.kernel.org/all/20250903-sound-v1-3-d4ca777b8512@uniontech.com/ Tested-by: Guoli An Signed-off-by: Cryolitia PukNgae Signed-off-by: Takashi Iwai --- sound/usb/mixer.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/sound/usb/mixer.c b/sound/usb/mixer.c index 34bcbfd8b54e..ae412e651faf 100644 --- a/sound/usb/mixer.c +++ b/sound/usb/mixer.c @@ -1189,6 +1189,13 @@ static void volume_control_quirks(struct usb_mixer_elem_info *cval, cval->min = -14208; /* Mute under it */ } break; + case USB_ID(0x12d1, 0x3a07): /* Huawei Technologies Co., Ltd. CM-Q3 */ + if (!strcmp(kctl->id.name, "PCM Playback Volume")) { + usb_audio_info(chip, + "set volume quirk for Huawei Technologies Co., Ltd. CM-Q3\n"); + cval->min = -11264; /* Mute under it */ + } + break; } } From db74b04edce1bc86b9a5acc724c7ca06f427ab60 Mon Sep 17 00:00:00 2001 From: Marek Vasut Date: Sat, 11 Oct 2025 12:59:53 +0200 Subject: [PATCH 038/330] drm/bridge: lt9211: Drop check for last nibble of version register There is now a new LT9211 rev. U5, which reports chip ID 0x18 0x01 0xe4 . The previous LT9211 reported chip ID 0x18 0x01 0xe3 , which is what the driver checks for right now. Since there is a possibility there will be yet another revision of the LT9211 in the future, drop the last version nibble check to allow all future revisions of the chip to work with this driver. This fix makes LT9211 rev. U5 work with this driver. Fixes: 8ce4129e3de4 ("drm/bridge: lt9211: Add Lontium LT9211 bridge driver") Signed-off-by: Marek Vasut Reviewed-by: Dmitry Baryshkov Link: https://lore.kernel.org/r/20251011110017.12521-1-marek.vasut@mailbox.org Signed-off-by: Dmitry Baryshkov --- drivers/gpu/drm/bridge/lontium-lt9211.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/gpu/drm/bridge/lontium-lt9211.c b/drivers/gpu/drm/bridge/lontium-lt9211.c index 399fa7eebd49..03fc8fd10f20 100644 --- a/drivers/gpu/drm/bridge/lontium-lt9211.c +++ b/drivers/gpu/drm/bridge/lontium-lt9211.c @@ -121,8 +121,7 @@ static int lt9211_read_chipid(struct lt9211 *ctx) } /* Test for known Chip ID. */ - if (chipid[0] != REG_CHIPID0_VALUE || chipid[1] != REG_CHIPID1_VALUE || - chipid[2] != REG_CHIPID2_VALUE) { + if (chipid[0] != REG_CHIPID0_VALUE || chipid[1] != REG_CHIPID1_VALUE) { dev_err(ctx->dev, "Unknown Chip ID: 0x%02x 0x%02x 0x%02x\n", chipid[0], chipid[1], chipid[2]); return -EINVAL; From 39dec6cd888bde4171bd4b8fcf45f73ab684404d Mon Sep 17 00:00:00 2001 From: Markus Elfring Date: Wed, 1 Oct 2025 21:09:00 +0200 Subject: [PATCH 039/330] smb: server: Use common error handling code in smb_direct_rdma_xmit() Add two jump targets so that a bit of exception handling can be better reused at the end of this function implementation. Signed-off-by: Markus Elfring Reviewed-by: Stefan Metzmacher Acked-by: Namjae Jeon Signed-off-by: Steve French --- fs/smb/server/transport_rdma.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/fs/smb/server/transport_rdma.c b/fs/smb/server/transport_rdma.c index b3077766d6ec..a201c5871a77 100644 --- a/fs/smb/server/transport_rdma.c +++ b/fs/smb/server/transport_rdma.c @@ -1574,18 +1574,14 @@ static int smb_direct_rdma_xmit(struct smb_direct_transport *t, get_buf_page_count(desc_buf, desc_buf_len), msg->sg_list, SG_CHUNK_SIZE); if (ret) { - kfree(msg); ret = -ENOMEM; - goto out; + goto free_msg; } ret = get_sg_list(desc_buf, desc_buf_len, msg->sgt.sgl, msg->sgt.orig_nents); - if (ret < 0) { - sg_free_table_chained(&msg->sgt, SG_CHUNK_SIZE); - kfree(msg); - goto out; - } + if (ret < 0) + goto free_table; ret = rdma_rw_ctx_init(&msg->rdma_ctx, sc->ib.qp, sc->ib.qp->port, msg->sgt.sgl, @@ -1596,9 +1592,7 @@ static int smb_direct_rdma_xmit(struct smb_direct_transport *t, is_read ? DMA_FROM_DEVICE : DMA_TO_DEVICE); if (ret < 0) { pr_err("failed to init rdma_rw_ctx: %d\n", ret); - sg_free_table_chained(&msg->sgt, SG_CHUNK_SIZE); - kfree(msg); - goto out; + goto free_table; } list_add_tail(&msg->list, &msg_list); @@ -1630,6 +1624,12 @@ static int smb_direct_rdma_xmit(struct smb_direct_transport *t, atomic_add(credits_needed, &sc->rw_io.credits.count); wake_up(&sc->rw_io.credits.wait_queue); return ret; + +free_table: + sg_free_table_chained(&msg->sgt, SG_CHUNK_SIZE); +free_msg: + kfree(msg); + goto out; } static int smb_direct_rdma_write(struct ksmbd_transport *t, From ef3e73a917ec7d080e0fb0e4015098a4fb0f1cff Mon Sep 17 00:00:00 2001 From: Nam Cao Date: Fri, 10 Oct 2025 12:03:07 +0000 Subject: [PATCH 040/330] powerpc/pseries/msi: Fix NULL pointer dereference at irq domain teardown pseries_msi_ops_teardown() reads pci_dev* from msi_alloc_info_t. However, pseries_msi_ops_prepare() does not populate this structure, thus it is all zeros. Consequently, pseries_msi_ops_teardown() triggers a NULL pointer dereference crash. struct pci_dev is available in struct irq_domain. Read it there instead. Reported-by: Venkat Rao Bagalkote Closes: https://lore.kernel.org/linuxppc-dev/878d7651-433a-46fe-a28b-1b7e893fcbe0@linux.ibm.com/ Tested-by: Venkat Rao Bagalkote Signed-off-by: Nam Cao Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20251010120307.3281720-1-namcao@linutronix.de --- arch/powerpc/platforms/pseries/msi.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/arch/powerpc/platforms/pseries/msi.c b/arch/powerpc/platforms/pseries/msi.c index 825f9432e03d..a82aaa786e9e 100644 --- a/arch/powerpc/platforms/pseries/msi.c +++ b/arch/powerpc/platforms/pseries/msi.c @@ -443,8 +443,7 @@ static int pseries_msi_ops_prepare(struct irq_domain *domain, struct device *dev */ static void pseries_msi_ops_teardown(struct irq_domain *domain, msi_alloc_info_t *arg) { - struct msi_desc *desc = arg->desc; - struct pci_dev *pdev = msi_desc_to_pci_dev(desc); + struct pci_dev *pdev = to_pci_dev(domain->dev); rtas_disable_msi(pdev); } From 2743cf75f7c92d2a0a4acabd7aef1c17d98fe123 Mon Sep 17 00:00:00 2001 From: Nam Cao Date: Wed, 8 Oct 2025 08:13:59 +0000 Subject: [PATCH 041/330] powerpc, ocxl: Fix extraction of struct xive_irq_data Commit cc0cc23babc9 ("powerpc/xive: Untangle xive from child interrupt controller drivers") changed xive_irq_data to be stashed to chip_data instead of handler_data. However, multiple places are still attempting to read xive_irq_data from handler_data and get a NULL pointer deference bug. Update them to read xive_irq_data from chip_data. Non-XIVE files which touch xive_irq_data seem quite strange to me, especially the ocxl driver. I think there ought to be an alternative platform-independent solution, instead of touching XIVE's data directly. Therefore, I think this whole thing should be cleaned up. But perhaps I just misunderstand something. In any case, this cleanup would not be trivial; for now, just get things working again. Fixes: cc0cc23babc9 ("powerpc/xive: Untangle xive from child interrupt controller drivers") Reported-by: Ritesh Harjani (IBM) Closes: https://lore.kernel.org/linuxppc-dev/68e48df8.170a0220.4b4b0.217d@mx.google.com/ Signed-off-by: Nam Cao Reviewed-by: Ganesh Goudar Acked-by: Andrew Donnellan # ocxl Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20251008081359.1382699-1-namcao@linutronix.de --- arch/powerpc/kvm/book3s_xive.c | 12 ++++-------- arch/powerpc/platforms/powernv/vas.c | 2 +- arch/powerpc/sysdev/xive/common.c | 2 +- drivers/misc/ocxl/afu_irq.c | 2 +- 4 files changed, 7 insertions(+), 11 deletions(-) diff --git a/arch/powerpc/kvm/book3s_xive.c b/arch/powerpc/kvm/book3s_xive.c index 1302b5ac5672..89a1b8c21ab4 100644 --- a/arch/powerpc/kvm/book3s_xive.c +++ b/arch/powerpc/kvm/book3s_xive.c @@ -916,8 +916,7 @@ int kvmppc_xive_attach_escalation(struct kvm_vcpu *vcpu, u8 prio, * it fires once. */ if (single_escalation) { - struct irq_data *d = irq_get_irq_data(xc->esc_virq[prio]); - struct xive_irq_data *xd = irq_data_get_irq_handler_data(d); + struct xive_irq_data *xd = irq_get_chip_data(xc->esc_virq[prio]); xive_vm_esb_load(xd, XIVE_ESB_SET_PQ_01); vcpu->arch.xive_esc_raddr = xd->eoi_page; @@ -1612,7 +1611,7 @@ int kvmppc_xive_set_mapped(struct kvm *kvm, unsigned long guest_irq, /* Grab info about irq */ state->pt_number = hw_irq; - state->pt_data = irq_data_get_irq_handler_data(host_data); + state->pt_data = irq_data_get_irq_chip_data(host_data); /* * Configure the IRQ to match the existing configuration of @@ -1787,8 +1786,7 @@ void kvmppc_xive_disable_vcpu_interrupts(struct kvm_vcpu *vcpu) */ void xive_cleanup_single_escalation(struct kvm_vcpu *vcpu, int irq) { - struct irq_data *d = irq_get_irq_data(irq); - struct xive_irq_data *xd = irq_data_get_irq_handler_data(d); + struct xive_irq_data *xd = irq_get_chip_data(irq); /* * This slightly odd sequence gives the right result @@ -2827,9 +2825,7 @@ int kvmppc_xive_debug_show_queues(struct seq_file *m, struct kvm_vcpu *vcpu) i0, i1); } if (xc->esc_virq[i]) { - struct irq_data *d = irq_get_irq_data(xc->esc_virq[i]); - struct xive_irq_data *xd = - irq_data_get_irq_handler_data(d); + struct xive_irq_data *xd = irq_get_chip_data(xc->esc_virq[i]); u64 pq = xive_vm_esb_load(xd, XIVE_ESB_GET); seq_printf(m, " ESC %d %c%c EOI @%llx", diff --git a/arch/powerpc/platforms/powernv/vas.c b/arch/powerpc/platforms/powernv/vas.c index b65256a63e87..9c9650319f3b 100644 --- a/arch/powerpc/platforms/powernv/vas.c +++ b/arch/powerpc/platforms/powernv/vas.c @@ -121,7 +121,7 @@ static int init_vas_instance(struct platform_device *pdev) return -EINVAL; } - xd = irq_get_handler_data(vinst->virq); + xd = irq_get_chip_data(vinst->virq); if (!xd) { pr_err("Inst%d: Invalid virq %d\n", vinst->vas_id, vinst->virq); diff --git a/arch/powerpc/sysdev/xive/common.c b/arch/powerpc/sysdev/xive/common.c index 625361a15424..8d0123b0ae84 100644 --- a/arch/powerpc/sysdev/xive/common.c +++ b/arch/powerpc/sysdev/xive/common.c @@ -1580,7 +1580,7 @@ static void xive_flush_cpu_queue(unsigned int cpu, struct xive_cpu *xc) cpu, irq); #endif raw_spin_lock(&desc->lock); - xd = irq_desc_get_handler_data(desc); + xd = irq_desc_get_chip_data(desc); /* * Clear saved_p to indicate that it's no longer pending diff --git a/drivers/misc/ocxl/afu_irq.c b/drivers/misc/ocxl/afu_irq.c index 36f7379b8e2d..f6b821fc274c 100644 --- a/drivers/misc/ocxl/afu_irq.c +++ b/drivers/misc/ocxl/afu_irq.c @@ -203,7 +203,7 @@ u64 ocxl_afu_irq_get_addr(struct ocxl_context *ctx, int irq_id) mutex_lock(&ctx->irq_lock); irq = idr_find(&ctx->irq_idr, irq_id); if (irq) { - xd = irq_get_handler_data(irq->virq); + xd = irq_get_chip_data(irq->virq); addr = xd ? xd->trig_page : 0; } mutex_unlock(&ctx->irq_lock); From 0843ba458439f38efdc14aa359c14ad0127edb01 Mon Sep 17 00:00:00 2001 From: Sourabh Jain Date: Wed, 8 Oct 2025 08:59:34 +0530 Subject: [PATCH 042/330] powerpc/fadump: skip parameter area allocation when fadump is disabled Fadump allocates memory to pass additional kernel command-line argument to the fadump kernel. However, this allocation is not needed when fadump is disabled. So avoid allocating memory for the additional parameter area in such cases. Fixes: f4892c68ecc1 ("powerpc/fadump: allocate memory for additional parameters early") Reviewed-by: Hari Bathini Signed-off-by: Sourabh Jain Fixes: f4892c68ecc1 ("powerpc/fadump: allocate memory for additional parameters early") Signed-off-by: Madhavan Srinivasan Link: https://patch.msgid.link/20251008032934.262683-1-sourabhjain@linux.ibm.com --- arch/powerpc/kernel/fadump.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/arch/powerpc/kernel/fadump.c b/arch/powerpc/kernel/fadump.c index 5782e743fd27..4ebc333dd786 100644 --- a/arch/powerpc/kernel/fadump.c +++ b/arch/powerpc/kernel/fadump.c @@ -1747,6 +1747,9 @@ void __init fadump_setup_param_area(void) { phys_addr_t range_start, range_end; + if (!fw_dump.fadump_enabled) + return; + if (!fw_dump.param_area_supported || fw_dump.dump_active) return; From 5ec6f9434225e18496a393f920b03eb46d67d71d Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Mon, 13 Oct 2025 09:26:49 +0200 Subject: [PATCH 043/330] ALSA: hda/realtek: Add quirk entry for HP ZBook 17 G6 HP ZBook 17 G6 with SSID 103c:860c requires a similar workaround as its 15-inch model in order to make the speaker and mute LED working. Add the corresponding quirk entry to address it. Link: https://bugzilla.kernel.org/show_bug.cgi?id=220372 Cc: Signed-off-by: Takashi Iwai --- sound/hda/codecs/realtek/alc269.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/hda/codecs/realtek/alc269.c b/sound/hda/codecs/realtek/alc269.c index 214eb9df6ef8..6b9e5c091bf3 100644 --- a/sound/hda/codecs/realtek/alc269.c +++ b/sound/hda/codecs/realtek/alc269.c @@ -6397,6 +6397,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = { SND_PCI_QUIRK(0x103c, 0x854a, "HP EliteBook 830 G6", ALC285_FIXUP_HP_GPIO_LED), SND_PCI_QUIRK(0x103c, 0x85c6, "HP Pavilion x360 Convertible 14-dy1xxx", ALC295_FIXUP_HP_MUTE_LED_COEFBIT11), SND_PCI_QUIRK(0x103c, 0x85de, "HP Envy x360 13-ar0xxx", ALC285_FIXUP_HP_ENVY_X360), + SND_PCI_QUIRK(0x103c, 0x860c, "HP ZBook 17 G6", ALC285_FIXUP_HP_GPIO_AMP_INIT), SND_PCI_QUIRK(0x103c, 0x860f, "HP ZBook 15 G6", ALC285_FIXUP_HP_GPIO_AMP_INIT), SND_PCI_QUIRK(0x103c, 0x861f, "HP Elite Dragonfly G1", ALC285_FIXUP_HP_GPIO_AMP_INIT), SND_PCI_QUIRK(0x103c, 0x869d, "HP", ALC236_FIXUP_HP_MUTE_LED), From 66233e583d1e00b1742d1ba36ae31568109ba6bd Mon Sep 17 00:00:00 2001 From: Shenghao Ding Date: Mon, 13 Oct 2025 12:35:55 +0800 Subject: [PATCH 044/330] ALSA: hda/tas2781: Set tas2781_hda::tasdevice_priv::chip_id as TAS5825 in case of tas5825 The software reset for TAS5825 is different form other chips, as it will set as 0x11 instead of 0x1 during reset in the tasdevice_reset(). So set tas2781_hda::tasdevice_priv::chip_id as TAS5825, tasdevice_reset() can work correctly. Fixes: 7ceb69ca82b1 ("ASoC: tas2781: Add tas2118, tas2x20, tas5825 support") Signed-off-by: Shenghao Ding Signed-off-by: Takashi Iwai --- sound/hda/codecs/side-codecs/tas2781_hda_i2c.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/hda/codecs/side-codecs/tas2781_hda_i2c.c b/sound/hda/codecs/side-codecs/tas2781_hda_i2c.c index a126f04c3ed7..0357401a6023 100644 --- a/sound/hda/codecs/side-codecs/tas2781_hda_i2c.c +++ b/sound/hda/codecs/side-codecs/tas2781_hda_i2c.c @@ -669,6 +669,7 @@ static int tas2781_hda_i2c_probe(struct i2c_client *clt) */ device_name = "TXNW5825"; hda_priv->hda_chip_id = HDA_TAS5825; + tas_hda->priv->chip_id = TAS5825; } else { return -ENODEV; } From 6079165e6e027c03e06556ff3df0ed03a34d68f0 Mon Sep 17 00:00:00 2001 From: Le Qi Date: Thu, 9 Oct 2025 17:06:18 +0800 Subject: [PATCH 045/330] ASoC: dt-bindings: qcom,sm8250: Add QCS615 sound card Add bindings for QCS615 sound card, which looks fully compatible with existing SM8250. Signed-off-by: Le Qi Acked-by: Krzysztof Kozlowski Link: https://patch.msgid.link/20251009090619.1097388-2-le.qi@oss.qualcomm.com Signed-off-by: Mark Brown --- Documentation/devicetree/bindings/sound/qcom,sm8250.yaml | 1 + 1 file changed, 1 insertion(+) diff --git a/Documentation/devicetree/bindings/sound/qcom,sm8250.yaml b/Documentation/devicetree/bindings/sound/qcom,sm8250.yaml index 8ac91625dce5..b49a920af704 100644 --- a/Documentation/devicetree/bindings/sound/qcom,sm8250.yaml +++ b/Documentation/devicetree/bindings/sound/qcom,sm8250.yaml @@ -33,6 +33,7 @@ properties: - qcom,apq8096-sndcard - qcom,glymur-sndcard - qcom,qcm6490-idp-sndcard + - qcom,qcs615-sndcard - qcom,qcs6490-rb3gen2-sndcard - qcom,qcs8275-sndcard - qcom,qcs9075-sndcard From dee4ef0ebe4dee655657ead30892aeca16462823 Mon Sep 17 00:00:00 2001 From: Le Qi Date: Thu, 9 Oct 2025 17:06:19 +0800 Subject: [PATCH 046/330] ASoC: qcom: sc8280xp: Add support for QCS615 Add compatible for sound card on QCS615 boards. Signed-off-by: Le Qi Reviewed-by: Krzysztof Kozlowski Link: https://patch.msgid.link/20251009090619.1097388-3-le.qi@oss.qualcomm.com Signed-off-by: Mark Brown --- sound/soc/qcom/sc8280xp.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/soc/qcom/sc8280xp.c b/sound/soc/qcom/sc8280xp.c index 78e327bc2f07..187f37ffe328 100644 --- a/sound/soc/qcom/sc8280xp.c +++ b/sound/soc/qcom/sc8280xp.c @@ -192,6 +192,7 @@ static int sc8280xp_platform_probe(struct platform_device *pdev) static const struct of_device_id snd_sc8280xp_dt_match[] = { {.compatible = "qcom,qcm6490-idp-sndcard", "qcm6490"}, + {.compatible = "qcom,qcs615-sndcard", "qcs615"}, {.compatible = "qcom,qcs6490-rb3gen2-sndcard", "qcs6490"}, {.compatible = "qcom,qcs8275-sndcard", "qcs8300"}, {.compatible = "qcom,qcs9075-sndcard", "sa8775p"}, From 53a3c6e222836a23e8e0693395584aefc456dca6 Mon Sep 17 00:00:00 2001 From: Baojun Xu Date: Thu, 2 Oct 2025 15:29:24 +0800 Subject: [PATCH 047/330] ASoC: tas2781: Support more newly-released amplifiers tas58xx in the driver TAS5802/TAS5815/TAS5828 has on-chip DSP without current/voltage feedback. Signed-off-by: Baojun Xu Link: https://patch.msgid.link/20251002072925.26242-1-baojun.xu@ti.com Signed-off-by: Mark Brown --- include/sound/tas2781.h | 3 +++ sound/soc/codecs/tas2781-i2c.c | 21 +++++++++++++++++++-- 2 files changed, 22 insertions(+), 2 deletions(-) diff --git a/include/sound/tas2781.h b/include/sound/tas2781.h index ddd997ac3216..0fbcdb15c74b 100644 --- a/include/sound/tas2781.h +++ b/include/sound/tas2781.h @@ -120,8 +120,11 @@ enum audio_device { TAS2570, TAS2572, TAS2781, + TAS5802, + TAS5815, TAS5825, TAS5827, + TAS5828, TAS_OTHERS, }; diff --git a/sound/soc/codecs/tas2781-i2c.c b/sound/soc/codecs/tas2781-i2c.c index 1539b70881d1..ba880b5de7e8 100644 --- a/sound/soc/codecs/tas2781-i2c.c +++ b/sound/soc/codecs/tas2781-i2c.c @@ -108,8 +108,11 @@ static const struct i2c_device_id tasdevice_id[] = { { "tas2570", TAS2570 }, { "tas2572", TAS2572 }, { "tas2781", TAS2781 }, + { "tas5802", TAS5802 }, + { "tas5815", TAS5815 }, { "tas5825", TAS5825 }, { "tas5827", TAS5827 }, + { "tas5828", TAS5828 }, {} }; MODULE_DEVICE_TABLE(i2c, tasdevice_id); @@ -124,8 +127,11 @@ static const struct of_device_id tasdevice_of_match[] = { { .compatible = "ti,tas2570" }, { .compatible = "ti,tas2572" }, { .compatible = "ti,tas2781" }, + { .compatible = "ti,tas5802" }, + { .compatible = "ti,tas5815" }, { .compatible = "ti,tas5825" }, { .compatible = "ti,tas5827" }, + { .compatible = "ti,tas5828" }, {}, }; MODULE_DEVICE_TABLE(of, tasdevice_of_match); @@ -1665,8 +1671,10 @@ static void tasdevice_fw_ready(const struct firmware *fmw, } tas_priv->fw_state = TASDEVICE_DSP_FW_ALL_OK; - /* There is no calibration required for TAS5825/TAS5827. */ - if (tas_priv->chip_id < TAS5825) { + /* There is no calibration required for + * TAS5802/TAS5815/TAS5825/TAS5827/TAS5828. + */ + if (tas_priv->chip_id < TAS5802) { ret = tasdevice_create_cali_ctrls(tas_priv); if (ret) { dev_err(tas_priv->dev, "cali controls error\n"); @@ -1720,8 +1728,11 @@ static void tasdevice_fw_ready(const struct firmware *fmw, switch (tas_priv->chip_id) { case TAS2563: case TAS2781: + case TAS5802: + case TAS5815: case TAS5825: case TAS5827: + case TAS5828: /* If DSP FW fail, DSP kcontrol won't be created. */ tasdevice_dsp_remove(tas_priv); } @@ -1882,8 +1893,11 @@ static int tasdevice_codec_probe(struct snd_soc_component *codec) p = (struct snd_kcontrol_new *)tas2781_snd_controls; size = ARRAY_SIZE(tas2781_snd_controls); break; + case TAS5802: + case TAS5815: case TAS5825: case TAS5827: + case TAS5828: p = (struct snd_kcontrol_new *)tas5825_snd_controls; size = ARRAY_SIZE(tas5825_snd_controls); break; @@ -2054,8 +2068,11 @@ static const struct acpi_device_id tasdevice_acpi_match[] = { { "TXNW2570", TAS2570 }, { "TXNW2572", TAS2572 }, { "TXNW2781", TAS2781 }, + { "TXNW5802", TAS5802 }, + { "TXNW5815", TAS5815 }, { "TXNW5825", TAS5825 }, { "TXNW5827", TAS5827 }, + { "TXNW5828", TAS5828 }, {}, }; From 7e6cfa3e94cf8278ee3170dc0b81fc6db1287e28 Mon Sep 17 00:00:00 2001 From: Baojun Xu Date: Thu, 2 Oct 2025 15:29:25 +0800 Subject: [PATCH 048/330] ASoC: tas2781: Update ti,tas2781.yaml for adding tas58xx Update ti,tas2781.yaml for adding TAS5802/TAS5815/TAS5828. Signed-off-by: Baojun Xu Link: https://patch.msgid.link/20251002072925.26242-2-baojun.xu@ti.com Signed-off-by: Mark Brown --- .../devicetree/bindings/sound/ti,tas2781.yaml | 43 ++++++++++++++++--- 1 file changed, 37 insertions(+), 6 deletions(-) diff --git a/Documentation/devicetree/bindings/sound/ti,tas2781.yaml b/Documentation/devicetree/bindings/sound/ti,tas2781.yaml index bd00afa47d62..7f84f506013c 100644 --- a/Documentation/devicetree/bindings/sound/ti,tas2781.yaml +++ b/Documentation/devicetree/bindings/sound/ti,tas2781.yaml @@ -24,10 +24,10 @@ description: | Instruments Smart Amp speaker protection algorithm. The integrated speaker voltage and current sense provides for real time monitoring of loudspeaker behavior. - The TAS5825/TAS5827 is a stereo, digital input Class-D audio - amplifier optimized for efficiently driving high peak power into - small loudspeakers. An integrated on-chip DSP supports Texas - Instruments Smart Amp speaker protection algorithm. + The TAS5802/TAS5815/TAS5825/TAS5827/TAS5828 is a stereo, digital input + Class-D audio amplifier optimized for efficiently driving high peak + power into small loudspeakers. An integrated on-chip DSP supports + Texas Instruments Smart Amp speaker protection algorithm. Specifications about the audio amplifier can be found at: https://www.ti.com/lit/gpn/tas2120 @@ -35,8 +35,10 @@ description: | https://www.ti.com/lit/gpn/tas2563 https://www.ti.com/lit/gpn/tas2572 https://www.ti.com/lit/gpn/tas2781 + https://www.ti.com/lit/gpn/tas5815 https://www.ti.com/lit/gpn/tas5825m https://www.ti.com/lit/gpn/tas5827 + https://www.ti.com/lit/gpn/tas5828m properties: compatible: @@ -65,11 +67,21 @@ properties: Protection and Audio Processing, 16/20/24/32bit stereo I2S or multichannel TDM. + ti,tas5802: 22-W, Inductor-Less, Digital Input, Closed-Loop Class-D + Audio Amplifier with 96-Khz Extended Processing and Low Idle Power + Dissipation. + + ti,tas5815: 30-W, Digital Input, Stereo, Closed-loop Class-D Audio + Amplifier with 96 kHz Enhanced Processing + ti,tas5825: 38-W Stereo, Inductor-Less, Digital Input, Closed-Loop 4.5V to 26.4V Class-D Audio Amplifier with 192-kHz Extended Audio Processing. - ti,tas5827: 47-W Stereo, Digital Input, High Efficiency Closed-Loop Class-D - Amplifier with Class-H Algorithm + ti,tas5827: 47-W Stereo, Digital Input, High Efficiency Closed-Loop + Class-D Amplifier with Class-H Algorithm + + ti,tas5828: 50-W Stereo, Digital Input, High Efficiency Closed-Loop + Class-D Amplifier with Hybrid-Pro Algorithm oneOf: - items: - enum: @@ -80,8 +92,11 @@ properties: - ti,tas2563 - ti,tas2570 - ti,tas2572 + - ti,tas5802 + - ti,tas5815 - ti,tas5825 - ti,tas5827 + - ti,tas5828 - const: ti,tas2781 - enum: - ti,tas2781 @@ -177,12 +192,28 @@ allOf: minimum: 0x38 maximum: 0x3f + - if: + properties: + compatible: + contains: + enum: + - ti,tas5802 + - ti,tas5815 + then: + properties: + reg: + maxItems: 4 + items: + minimum: 0x54 + maximum: 0x57 + - if: properties: compatible: contains: enum: - ti,tas5827 + - ti,tas5828 then: properties: reg: From 6370a996f308ea3276030769b7482b346e7cc7c1 Mon Sep 17 00:00:00 2001 From: Christophe Leroy Date: Thu, 9 Oct 2025 19:03:13 +0200 Subject: [PATCH 049/330] ASoC: codecs: Fix gain setting ranges for Renesas IDT821034 codec The gain ranges specified in Renesas IDT821034 codec documentation are [-3dB;+13dB] in the transmit path (ADC) and [-13dB;+3dB] in the receive path (DAC). Allthough the registers allow programming values outside those ranges, the signal S/N and distorsion are only guaranteed in the specified ranges. Set ranges to the specified ones. Fixes: e51166990e81 ("ASoC: codecs: Add support for the Renesas IDT821034 codec") Signed-off-by: Christophe Leroy Link: https://patch.msgid.link/2bd547194f3398e6182f770d7d6be711c702b4b2.1760029099.git.christophe.leroy@csgroup.eu Signed-off-by: Mark Brown --- sound/soc/codecs/idt821034.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/sound/soc/codecs/idt821034.c b/sound/soc/codecs/idt821034.c index a03d4e5e7d14..cab2f2eecdfb 100644 --- a/sound/soc/codecs/idt821034.c +++ b/sound/soc/codecs/idt821034.c @@ -548,14 +548,14 @@ static int idt821034_kctrl_mute_put(struct snd_kcontrol *kcontrol, return ret; } -static const DECLARE_TLV_DB_LINEAR(idt821034_gain_in, -6520, 1306); -#define IDT821034_GAIN_IN_MIN_RAW 1 /* -65.20 dB -> 10^(-65.2/20.0) * 1820 = 1 */ -#define IDT821034_GAIN_IN_MAX_RAW 8191 /* 13.06 dB -> 10^(13.06/20.0) * 1820 = 8191 */ +static const DECLARE_TLV_DB_LINEAR(idt821034_gain_in, -300, 1300); +#define IDT821034_GAIN_IN_MIN_RAW 1288 /* -3.0 dB -> 10^(-3.0/20.0) * 1820 = 1288 */ +#define IDT821034_GAIN_IN_MAX_RAW 8130 /* 13.0 dB -> 10^(13.0/20.0) * 1820 = 8130 */ #define IDT821034_GAIN_IN_INIT_RAW 1820 /* 0dB -> 10^(0/20) * 1820 = 1820 */ -static const DECLARE_TLV_DB_LINEAR(idt821034_gain_out, -6798, 1029); -#define IDT821034_GAIN_OUT_MIN_RAW 1 /* -67.98 dB -> 10^(-67.98/20.0) * 2506 = 1*/ -#define IDT821034_GAIN_OUT_MAX_RAW 8191 /* 10.29 dB -> 10^(10.29/20.0) * 2506 = 8191 */ +static const DECLARE_TLV_DB_LINEAR(idt821034_gain_out, -1300, 300); +#define IDT821034_GAIN_OUT_MIN_RAW 561 /* -13.0 dB -> 10^(-13.0/20.0) * 2506 = 561 */ +#define IDT821034_GAIN_OUT_MAX_RAW 3540 /* 3.0 dB -> 10^(3.0/20.0) * 2506 = 3540 */ #define IDT821034_GAIN_OUT_INIT_RAW 2506 /* 0dB -> 10^(0/20) * 2506 = 2506 */ static const struct snd_kcontrol_new idt821034_controls[] = { From 7e8242405b94ceac6db820de7d4fd9318cbc1219 Mon Sep 17 00:00:00 2001 From: Bean Huo Date: Wed, 1 Oct 2025 08:08:03 +0200 Subject: [PATCH 050/330] rpmb: move rpmb_frame struct and constants to common header Move struct rpmb_frame and RPMB operation constants from MMC block driver to include/linux/rpmb.h for reuse across different RPMB implementations (UFS, NVMe, etc.). Signed-off-by: Bean Huo Reviewed-by: Avri Altman Acked-by: Jens Wiklander Reviewed-by: Bart Van Assche Signed-off-by: Ulf Hansson --- drivers/mmc/core/block.c | 42 -------------------------------------- include/linux/rpmb.h | 44 ++++++++++++++++++++++++++++++++++++++++ 2 files changed, 44 insertions(+), 42 deletions(-) diff --git a/drivers/mmc/core/block.c b/drivers/mmc/core/block.c index 9399bf6c766a..c0ffe0817fd4 100644 --- a/drivers/mmc/core/block.c +++ b/drivers/mmc/core/block.c @@ -79,48 +79,6 @@ MODULE_ALIAS("mmc:block"); #define MMC_EXTRACT_INDEX_FROM_ARG(x) ((x & 0x00FF0000) >> 16) #define MMC_EXTRACT_VALUE_FROM_ARG(x) ((x & 0x0000FF00) >> 8) -/** - * struct rpmb_frame - rpmb frame as defined by eMMC 5.1 (JESD84-B51) - * - * @stuff : stuff bytes - * @key_mac : The authentication key or the message authentication - * code (MAC) depending on the request/response type. - * The MAC will be delivered in the last (or the only) - * block of data. - * @data : Data to be written or read by signed access. - * @nonce : Random number generated by the host for the requests - * and copied to the response by the RPMB engine. - * @write_counter: Counter value for the total amount of the successful - * authenticated data write requests made by the host. - * @addr : Address of the data to be programmed to or read - * from the RPMB. Address is the serial number of - * the accessed block (half sector 256B). - * @block_count : Number of blocks (half sectors, 256B) requested to be - * read/programmed. - * @result : Includes information about the status of the write counter - * (valid, expired) and result of the access made to the RPMB. - * @req_resp : Defines the type of request and response to/from the memory. - * - * The stuff bytes and big-endian properties are modeled to fit to the spec. - */ -struct rpmb_frame { - u8 stuff[196]; - u8 key_mac[32]; - u8 data[256]; - u8 nonce[16]; - __be32 write_counter; - __be16 addr; - __be16 block_count; - __be16 result; - __be16 req_resp; -} __packed; - -#define RPMB_PROGRAM_KEY 0x1 /* Program RPMB Authentication Key */ -#define RPMB_GET_WRITE_COUNTER 0x2 /* Read RPMB write counter */ -#define RPMB_WRITE_DATA 0x3 /* Write data to RPMB partition */ -#define RPMB_READ_DATA 0x4 /* Read data from RPMB partition */ -#define RPMB_RESULT_READ 0x5 /* Read result request (Internal) */ - #define RPMB_FRAME_SIZE sizeof(struct rpmb_frame) #define CHECK_SIZE_NEQ(val) ((val) != sizeof(struct rpmb_frame)) #define CHECK_SIZE_ALIGNED(val) IS_ALIGNED((val), sizeof(struct rpmb_frame)) diff --git a/include/linux/rpmb.h b/include/linux/rpmb.h index cccda73eea4d..ed3f8e431eff 100644 --- a/include/linux/rpmb.h +++ b/include/linux/rpmb.h @@ -61,6 +61,50 @@ struct rpmb_dev { #define to_rpmb_dev(x) container_of((x), struct rpmb_dev, dev) +/** + * struct rpmb_frame - RPMB frame structure for authenticated access + * + * @stuff : stuff bytes, a padding/reserved area of 196 bytes at the + * beginning of the RPMB frame. They don’t carry meaningful + * data but are required to make the frame exactly 512 bytes. + * @key_mac : The authentication key or the message authentication + * code (MAC) depending on the request/response type. + * The MAC will be delivered in the last (or the only) + * block of data. + * @data : Data to be written or read by signed access. + * @nonce : Random number generated by the host for the requests + * and copied to the response by the RPMB engine. + * @write_counter: Counter value for the total amount of the successful + * authenticated data write requests made by the host. + * @addr : Address of the data to be programmed to or read + * from the RPMB. Address is the serial number of + * the accessed block (half sector 256B). + * @block_count : Number of blocks (half sectors, 256B) requested to be + * read/programmed. + * @result : Includes information about the status of the write counter + * (valid, expired) and result of the access made to the RPMB. + * @req_resp : Defines the type of request and response to/from the memory. + * + * The stuff bytes and big-endian properties are modeled to fit to the spec. + */ +struct rpmb_frame { + u8 stuff[196]; + u8 key_mac[32]; + u8 data[256]; + u8 nonce[16]; + __be32 write_counter; + __be16 addr; + __be16 block_count; + __be16 result; + __be16 req_resp; +}; + +#define RPMB_PROGRAM_KEY 0x1 /* Program RPMB Authentication Key */ +#define RPMB_GET_WRITE_COUNTER 0x2 /* Read RPMB write counter */ +#define RPMB_WRITE_DATA 0x3 /* Write data to RPMB partition */ +#define RPMB_READ_DATA 0x4 /* Read data from RPMB partition */ +#define RPMB_RESULT_READ 0x5 /* Read result request (Internal) */ + #if IS_ENABLED(CONFIG_RPMB) struct rpmb_dev *rpmb_dev_get(struct rpmb_dev *rdev); void rpmb_dev_put(struct rpmb_dev *rdev); From 6e54919cb541fdf1063b16f3254c28d01bc9e5ff Mon Sep 17 00:00:00 2001 From: Cristian Ciocaltea Date: Fri, 3 Oct 2025 21:03:23 +0300 Subject: [PATCH 051/330] ASoC: nau8821: Cancel jdet_work before handling jack ejection The microphone detection work scheduled by a prior jack insertion interrupt may still be in a pending state or under execution when a jack ejection interrupt has been fired. This might lead to a racing condition or nau8821_jdet_work() completing after nau8821_eject_jack(), which will override the currently disconnected state of the jack and incorrectly report the headphone or the headset as being connected. Cancel any pending jdet_work or wait for its execution to finish before attempting to handle the ejection interrupt. Proceed similarly before launching the eject handler as a consequence of detecting an invalid insert interrupt. Fixes: aab1ad11d69f ("ASoC: nau8821: new driver") Signed-off-by: Cristian Ciocaltea Link: https://patch.msgid.link/20251003-nau8821-jdet-fixes-v1-1-f7b0e2543f09@collabora.com Signed-off-by: Mark Brown --- sound/soc/codecs/nau8821.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/sound/soc/codecs/nau8821.c b/sound/soc/codecs/nau8821.c index edb95f869a4a..fed5d51fa5c3 100644 --- a/sound/soc/codecs/nau8821.c +++ b/sound/soc/codecs/nau8821.c @@ -1185,6 +1185,7 @@ static irqreturn_t nau8821_interrupt(int irq, void *data) if ((active_irq & NAU8821_JACK_EJECT_IRQ_MASK) == NAU8821_JACK_EJECT_DETECTED) { + cancel_work_sync(&nau8821->jdet_work); regmap_update_bits(regmap, NAU8821_R71_ANALOG_ADC_1, NAU8821_MICDET_MASK, NAU8821_MICDET_DIS); nau8821_eject_jack(nau8821); @@ -1199,11 +1200,11 @@ static irqreturn_t nau8821_interrupt(int irq, void *data) clear_irq = NAU8821_KEY_RELEASE_IRQ; } else if ((active_irq & NAU8821_JACK_INSERT_IRQ_MASK) == NAU8821_JACK_INSERT_DETECTED) { + cancel_work_sync(&nau8821->jdet_work); regmap_update_bits(regmap, NAU8821_R71_ANALOG_ADC_1, NAU8821_MICDET_MASK, NAU8821_MICDET_EN); if (nau8821_is_jack_inserted(regmap)) { /* detect microphone and jack type */ - cancel_work_sync(&nau8821->jdet_work); schedule_work(&nau8821->jdet_work); /* Turn off insertion interruption at manual mode */ regmap_update_bits(regmap, From 9273aa85b35cc02d0953a1ba3b7bd694e5a2c10e Mon Sep 17 00:00:00 2001 From: Cristian Ciocaltea Date: Fri, 3 Oct 2025 21:03:24 +0300 Subject: [PATCH 052/330] ASoC: nau8821: Generalize helper to clear IRQ status Instead of adding yet another utility function for dealing with the interrupt clearing register, generalize nau8821_int_status_clear_all() by renaming it to nau8821_irq_status_clear(), whilst introducing a second parameter to allow restricting the operation scope to a single interrupt instead of the whole range of active IRQs. While at it, also fix a spelling typo in the comment block. Note this is mainly a prerequisite for subsequent patches aiming to address some deficiencies in the implementation of the interrupt handler. Thus the presence of the Fixes tag below is intentional, to facilitate backporting. Fixes: aab1ad11d69f ("ASoC: nau8821: new driver") Signed-off-by: Cristian Ciocaltea Link: https://patch.msgid.link/20251003-nau8821-jdet-fixes-v1-2-f7b0e2543f09@collabora.com Signed-off-by: Mark Brown --- sound/soc/codecs/nau8821.c | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) diff --git a/sound/soc/codecs/nau8821.c b/sound/soc/codecs/nau8821.c index fed5d51fa5c3..cefce9789312 100644 --- a/sound/soc/codecs/nau8821.c +++ b/sound/soc/codecs/nau8821.c @@ -1021,12 +1021,17 @@ static bool nau8821_is_jack_inserted(struct regmap *regmap) return active_high == is_high; } -static void nau8821_int_status_clear_all(struct regmap *regmap) +static void nau8821_irq_status_clear(struct regmap *regmap, int active_irq) { - int active_irq, clear_irq, i; + int clear_irq, i; - /* Reset the intrruption status from rightmost bit if the corres- - * ponding irq event occurs. + if (active_irq) { + regmap_write(regmap, NAU8821_R11_INT_CLR_KEY_STATUS, active_irq); + return; + } + + /* Reset the interruption status from rightmost bit if the + * corresponding irq event occurs. */ regmap_read(regmap, NAU8821_R10_IRQ_STATUS, &active_irq); for (i = 0; i < NAU8821_REG_DATA_LEN; i++) { @@ -1053,7 +1058,7 @@ static void nau8821_eject_jack(struct nau8821 *nau8821) snd_soc_dapm_sync(dapm); /* Clear all interruption status */ - nau8821_int_status_clear_all(regmap); + nau8821_irq_status_clear(regmap, 0); /* Enable the insertion interruption, disable the ejection inter- * ruption, and then bypass de-bounce circuit. @@ -1522,7 +1527,7 @@ static int nau8821_resume_setup(struct nau8821 *nau8821) nau8821_configure_sysclk(nau8821, NAU8821_CLK_DIS, 0); if (nau8821->irq) { /* Clear all interruption status */ - nau8821_int_status_clear_all(regmap); + nau8821_irq_status_clear(regmap, 0); /* Enable both insertion and ejection interruptions, and then * bypass de-bounce circuit. From a698679fe8b0fec41d1fb9547a53127a85c1be92 Mon Sep 17 00:00:00 2001 From: Cristian Ciocaltea Date: Fri, 3 Oct 2025 21:03:25 +0300 Subject: [PATCH 053/330] ASoC: nau8821: Consistently clear interrupts before unmasking The interrupt handler attempts to perform some IRQ status clear operations *after* rather than *before* unmasking and enabling interrupts. This is a rather fragile approach since it may generally lead to missing IRQ requests or causing spurious interrupts. Make use of the nau8821_irq_status_clear() helper instead of manipulating the related register directly and ensure any interrupt clearing is performed *after* the target interrupts are disabled/masked and *before* proceeding with additional interrupt unmasking/enablement operations. This also implicitly drops the redundant clear operation of the ejection IRQ in the interrupt handler, since nau8821_eject_jack() has been already responsible for clearing all active interrupts. Fixes: aab1ad11d69f ("ASoC: nau8821: new driver") Fixes: 2551b6e89936 ("ASoC: nau8821: Add headset button detection") Signed-off-by: Cristian Ciocaltea Link: https://patch.msgid.link/20251003-nau8821-jdet-fixes-v1-3-f7b0e2543f09@collabora.com Signed-off-by: Mark Brown --- sound/soc/codecs/nau8821.c | 58 ++++++++++++++++++++------------------ 1 file changed, 30 insertions(+), 28 deletions(-) diff --git a/sound/soc/codecs/nau8821.c b/sound/soc/codecs/nau8821.c index cefce9789312..56e5a0d80782 100644 --- a/sound/soc/codecs/nau8821.c +++ b/sound/soc/codecs/nau8821.c @@ -1057,20 +1057,24 @@ static void nau8821_eject_jack(struct nau8821 *nau8821) snd_soc_component_disable_pin(component, "MICBIAS"); snd_soc_dapm_sync(dapm); + /* Disable & mask both insertion & ejection IRQs */ + regmap_update_bits(regmap, NAU8821_R12_INTERRUPT_DIS_CTRL, + NAU8821_IRQ_INSERT_DIS | NAU8821_IRQ_EJECT_DIS, + NAU8821_IRQ_INSERT_DIS | NAU8821_IRQ_EJECT_DIS); + regmap_update_bits(regmap, NAU8821_R0F_INTERRUPT_MASK, + NAU8821_IRQ_INSERT_EN | NAU8821_IRQ_EJECT_EN, + NAU8821_IRQ_INSERT_EN | NAU8821_IRQ_EJECT_EN); + /* Clear all interruption status */ nau8821_irq_status_clear(regmap, 0); - /* Enable the insertion interruption, disable the ejection inter- - * ruption, and then bypass de-bounce circuit. - */ + /* Enable & unmask the insertion IRQ */ regmap_update_bits(regmap, NAU8821_R12_INTERRUPT_DIS_CTRL, - NAU8821_IRQ_EJECT_DIS | NAU8821_IRQ_INSERT_DIS, - NAU8821_IRQ_EJECT_DIS); - /* Mask unneeded IRQs: 1 - disable, 0 - enable */ + NAU8821_IRQ_INSERT_DIS, 0); regmap_update_bits(regmap, NAU8821_R0F_INTERRUPT_MASK, - NAU8821_IRQ_EJECT_EN | NAU8821_IRQ_INSERT_EN, - NAU8821_IRQ_EJECT_EN); + NAU8821_IRQ_INSERT_EN, 0); + /* Bypass de-bounce circuit */ regmap_update_bits(regmap, NAU8821_R0D_JACK_DET_CTRL, NAU8821_JACK_DET_DB_BYPASS, NAU8821_JACK_DET_DB_BYPASS); @@ -1094,7 +1098,6 @@ static void nau8821_eject_jack(struct nau8821 *nau8821) NAU8821_IRQ_KEY_RELEASE_DIS | NAU8821_IRQ_KEY_PRESS_DIS); } - } static void nau8821_jdet_work(struct work_struct *work) @@ -1151,6 +1154,15 @@ static void nau8821_setup_inserted_irq(struct nau8821 *nau8821) { struct regmap *regmap = nau8821->regmap; + /* Disable & mask insertion IRQ */ + regmap_update_bits(regmap, NAU8821_R12_INTERRUPT_DIS_CTRL, + NAU8821_IRQ_INSERT_DIS, NAU8821_IRQ_INSERT_DIS); + regmap_update_bits(regmap, NAU8821_R0F_INTERRUPT_MASK, + NAU8821_IRQ_INSERT_EN, NAU8821_IRQ_INSERT_EN); + + /* Clear insert IRQ status */ + nau8821_irq_status_clear(regmap, NAU8821_JACK_INSERT_DETECTED); + /* Enable internal VCO needed for interruptions */ if (nau8821->dapm->bias_level < SND_SOC_BIAS_PREPARE) nau8821_configure_sysclk(nau8821, NAU8821_CLK_INTERNAL, 0); @@ -1169,17 +1181,18 @@ static void nau8821_setup_inserted_irq(struct nau8821 *nau8821) regmap_update_bits(regmap, NAU8821_R0D_JACK_DET_CTRL, NAU8821_JACK_DET_DB_BYPASS, 0); + /* Unmask & enable the ejection IRQs */ regmap_update_bits(regmap, NAU8821_R0F_INTERRUPT_MASK, - NAU8821_IRQ_EJECT_EN, 0); + NAU8821_IRQ_EJECT_EN, 0); regmap_update_bits(regmap, NAU8821_R12_INTERRUPT_DIS_CTRL, - NAU8821_IRQ_EJECT_DIS, 0); + NAU8821_IRQ_EJECT_DIS, 0); } static irqreturn_t nau8821_interrupt(int irq, void *data) { struct nau8821 *nau8821 = (struct nau8821 *)data; struct regmap *regmap = nau8821->regmap; - int active_irq, clear_irq = 0, event = 0, event_mask = 0; + int active_irq, event = 0, event_mask = 0; if (regmap_read(regmap, NAU8821_R10_IRQ_STATUS, &active_irq)) { dev_err(nau8821->dev, "failed to read irq status\n"); @@ -1195,14 +1208,13 @@ static irqreturn_t nau8821_interrupt(int irq, void *data) NAU8821_MICDET_MASK, NAU8821_MICDET_DIS); nau8821_eject_jack(nau8821); event_mask |= SND_JACK_HEADSET; - clear_irq = NAU8821_JACK_EJECT_IRQ_MASK; } else if (active_irq & NAU8821_KEY_SHORT_PRESS_IRQ) { event |= NAU8821_BUTTON; event_mask |= NAU8821_BUTTON; - clear_irq = NAU8821_KEY_SHORT_PRESS_IRQ; + nau8821_irq_status_clear(regmap, NAU8821_KEY_SHORT_PRESS_IRQ); } else if (active_irq & NAU8821_KEY_RELEASE_IRQ) { event_mask = NAU8821_BUTTON; - clear_irq = NAU8821_KEY_RELEASE_IRQ; + nau8821_irq_status_clear(regmap, NAU8821_KEY_RELEASE_IRQ); } else if ((active_irq & NAU8821_JACK_INSERT_IRQ_MASK) == NAU8821_JACK_INSERT_DETECTED) { cancel_work_sync(&nau8821->jdet_work); @@ -1212,27 +1224,17 @@ static irqreturn_t nau8821_interrupt(int irq, void *data) /* detect microphone and jack type */ schedule_work(&nau8821->jdet_work); /* Turn off insertion interruption at manual mode */ - regmap_update_bits(regmap, - NAU8821_R12_INTERRUPT_DIS_CTRL, - NAU8821_IRQ_INSERT_DIS, - NAU8821_IRQ_INSERT_DIS); - regmap_update_bits(regmap, - NAU8821_R0F_INTERRUPT_MASK, - NAU8821_IRQ_INSERT_EN, - NAU8821_IRQ_INSERT_EN); nau8821_setup_inserted_irq(nau8821); } else { dev_warn(nau8821->dev, "Inserted IRQ fired but not connected\n"); nau8821_eject_jack(nau8821); } + } else { + /* Clear the rightmost interrupt */ + nau8821_irq_status_clear(regmap, active_irq); } - if (!clear_irq) - clear_irq = active_irq; - /* clears the rightmost interruption */ - regmap_write(regmap, NAU8821_R11_INT_CLR_KEY_STATUS, clear_irq); - if (event_mask) snd_soc_jack_report(nau8821->jack, event, event_mask); From 2b4eda7bf7d8a4e2f7575a98f55d8336dec0f302 Mon Sep 17 00:00:00 2001 From: Cristian Ciocaltea Date: Fri, 3 Oct 2025 21:03:26 +0300 Subject: [PATCH 054/330] ASoC: nau8821: Add DMI quirk to bypass jack debounce circuit Stress testing the audio jack hotplug handling on a few Steam Deck units revealed that the debounce circuit is responsible for having a negative impact on the detection reliability, e.g. in some cases the ejection interrupt is not fired, while in other instances it goes into a kind of invalid state and generates a flood of misleading interrupts. Add new entries to the DMI table introduced via commit 1bc40efdaf4a ("ASoC: nau8821: Add DMI quirk mechanism for active-high jack-detect") and extend the quirk logic to allow bypassing the debounce circuit used for jack detection on Valve Steam Deck LCD and OLED models. While at it, rename existing NAU8821_JD_ACTIVE_HIGH quirk bitfield to NAU8821_QUIRK_JD_ACTIVE_HIGH. This should help improve code readability by differentiating from similarly named register bits. Fixes: aab1ad11d69f ("ASoC: nau8821: new driver") Signed-off-by: Cristian Ciocaltea Link: https://patch.msgid.link/20251003-nau8821-jdet-fixes-v1-4-f7b0e2543f09@collabora.com Signed-off-by: Mark Brown --- sound/soc/codecs/nau8821.c | 33 +++++++++++++++++++++++++++------ 1 file changed, 27 insertions(+), 6 deletions(-) diff --git a/sound/soc/codecs/nau8821.c b/sound/soc/codecs/nau8821.c index 56e5a0d80782..a8ff2ce70be9 100644 --- a/sound/soc/codecs/nau8821.c +++ b/sound/soc/codecs/nau8821.c @@ -26,7 +26,8 @@ #include #include "nau8821.h" -#define NAU8821_JD_ACTIVE_HIGH BIT(0) +#define NAU8821_QUIRK_JD_ACTIVE_HIGH BIT(0) +#define NAU8821_QUIRK_JD_DB_BYPASS BIT(1) static int nau8821_quirk; static int quirk_override = -1; @@ -1177,9 +1178,10 @@ static void nau8821_setup_inserted_irq(struct nau8821 *nau8821) regmap_update_bits(regmap, NAU8821_R1D_I2S_PCM_CTRL2, NAU8821_I2S_MS_MASK, NAU8821_I2S_MS_SLAVE); - /* Not bypass de-bounce circuit */ - regmap_update_bits(regmap, NAU8821_R0D_JACK_DET_CTRL, - NAU8821_JACK_DET_DB_BYPASS, 0); + /* Do not bypass de-bounce circuit */ + if (!(nau8821_quirk & NAU8821_QUIRK_JD_DB_BYPASS)) + regmap_update_bits(regmap, NAU8821_R0D_JACK_DET_CTRL, + NAU8821_JACK_DET_DB_BYPASS, 0); /* Unmask & enable the ejection IRQs */ regmap_update_bits(regmap, NAU8821_R0F_INTERRUPT_MASK, @@ -1864,7 +1866,23 @@ static const struct dmi_system_id nau8821_quirk_table[] = { DMI_MATCH(DMI_SYS_VENDOR, "Positivo Tecnologia SA"), DMI_MATCH(DMI_BOARD_NAME, "CW14Q01P-V2"), }, - .driver_data = (void *)(NAU8821_JD_ACTIVE_HIGH), + .driver_data = (void *)(NAU8821_QUIRK_JD_ACTIVE_HIGH), + }, + { + /* Valve Steam Deck LCD */ + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Valve"), + DMI_MATCH(DMI_PRODUCT_NAME, "Jupiter"), + }, + .driver_data = (void *)(NAU8821_QUIRK_JD_DB_BYPASS), + }, + { + /* Valve Steam Deck OLED */ + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Valve"), + DMI_MATCH(DMI_PRODUCT_NAME, "Galileo"), + }, + .driver_data = (void *)(NAU8821_QUIRK_JD_DB_BYPASS), }, {} }; @@ -1906,9 +1924,12 @@ static int nau8821_i2c_probe(struct i2c_client *i2c) nau8821_check_quirks(); - if (nau8821_quirk & NAU8821_JD_ACTIVE_HIGH) + if (nau8821_quirk & NAU8821_QUIRK_JD_ACTIVE_HIGH) nau8821->jkdet_polarity = 0; + if (nau8821_quirk & NAU8821_QUIRK_JD_DB_BYPASS) + dev_dbg(dev, "Force bypassing jack detection debounce circuit\n"); + nau8821_print_device_properties(nau8821); nau8821_reset_chip(nau8821->regmap); From ee70bacef1c6050e4836409927294d744dbcfa72 Mon Sep 17 00:00:00 2001 From: Cristian Ciocaltea Date: Fri, 3 Oct 2025 21:03:27 +0300 Subject: [PATCH 055/330] ASoC: nau8821: Avoid unnecessary blocking in IRQ handler The interrupt handler offloads the microphone detection logic to nau8821_jdet_work(), which implies a sleep operation. However, before being able to process any subsequent hotplug event, the interrupt handler needs to wait for any prior scheduled work to complete. Move the sleep out of jdet_work by converting it to a delayed work. This eliminates the undesired blocking in the interrupt handler when attempting to cancel a recently scheduled work item and should help reducing transient input reports that might confuse user-space. Signed-off-by: Cristian Ciocaltea Link: https://patch.msgid.link/20251003-nau8821-jdet-fixes-v1-5-f7b0e2543f09@collabora.com Signed-off-by: Mark Brown --- sound/soc/codecs/nau8821.c | 22 ++++++++++++---------- sound/soc/codecs/nau8821.h | 2 +- 2 files changed, 13 insertions(+), 11 deletions(-) diff --git a/sound/soc/codecs/nau8821.c b/sound/soc/codecs/nau8821.c index a8ff2ce70be9..4fa9a785513e 100644 --- a/sound/soc/codecs/nau8821.c +++ b/sound/soc/codecs/nau8821.c @@ -1104,16 +1104,12 @@ static void nau8821_eject_jack(struct nau8821 *nau8821) static void nau8821_jdet_work(struct work_struct *work) { struct nau8821 *nau8821 = - container_of(work, struct nau8821, jdet_work); + container_of(work, struct nau8821, jdet_work.work); struct snd_soc_dapm_context *dapm = nau8821->dapm; struct snd_soc_component *component = snd_soc_dapm_to_component(dapm); struct regmap *regmap = nau8821->regmap; int jack_status_reg, mic_detected, event = 0, event_mask = 0; - snd_soc_component_force_enable_pin(component, "MICBIAS"); - snd_soc_dapm_sync(dapm); - msleep(20); - regmap_read(regmap, NAU8821_R58_I2C_DEVICE_ID, &jack_status_reg); mic_detected = !(jack_status_reg & NAU8821_KEYDET); if (mic_detected) { @@ -1146,6 +1142,7 @@ static void nau8821_jdet_work(struct work_struct *work) snd_soc_component_disable_pin(component, "MICBIAS"); snd_soc_dapm_sync(dapm); } + event_mask |= SND_JACK_HEADSET; snd_soc_jack_report(nau8821->jack, event, event_mask); } @@ -1194,6 +1191,7 @@ static irqreturn_t nau8821_interrupt(int irq, void *data) { struct nau8821 *nau8821 = (struct nau8821 *)data; struct regmap *regmap = nau8821->regmap; + struct snd_soc_component *component; int active_irq, event = 0, event_mask = 0; if (regmap_read(regmap, NAU8821_R10_IRQ_STATUS, &active_irq)) { @@ -1205,7 +1203,7 @@ static irqreturn_t nau8821_interrupt(int irq, void *data) if ((active_irq & NAU8821_JACK_EJECT_IRQ_MASK) == NAU8821_JACK_EJECT_DETECTED) { - cancel_work_sync(&nau8821->jdet_work); + cancel_delayed_work_sync(&nau8821->jdet_work); regmap_update_bits(regmap, NAU8821_R71_ANALOG_ADC_1, NAU8821_MICDET_MASK, NAU8821_MICDET_DIS); nau8821_eject_jack(nau8821); @@ -1219,12 +1217,15 @@ static irqreturn_t nau8821_interrupt(int irq, void *data) nau8821_irq_status_clear(regmap, NAU8821_KEY_RELEASE_IRQ); } else if ((active_irq & NAU8821_JACK_INSERT_IRQ_MASK) == NAU8821_JACK_INSERT_DETECTED) { - cancel_work_sync(&nau8821->jdet_work); + cancel_delayed_work_sync(&nau8821->jdet_work); regmap_update_bits(regmap, NAU8821_R71_ANALOG_ADC_1, NAU8821_MICDET_MASK, NAU8821_MICDET_EN); if (nau8821_is_jack_inserted(regmap)) { - /* detect microphone and jack type */ - schedule_work(&nau8821->jdet_work); + /* Detect microphone and jack type */ + component = snd_soc_dapm_to_component(nau8821->dapm); + snd_soc_component_force_enable_pin(component, "MICBIAS"); + snd_soc_dapm_sync(nau8821->dapm); + schedule_delayed_work(&nau8821->jdet_work, msecs_to_jiffies(20)); /* Turn off insertion interruption at manual mode */ nau8821_setup_inserted_irq(nau8821); } else { @@ -1661,7 +1662,8 @@ int nau8821_enable_jack_detect(struct snd_soc_component *component, nau8821->jack = jack; /* Initiate jack detection work queue */ - INIT_WORK(&nau8821->jdet_work, nau8821_jdet_work); + INIT_DELAYED_WORK(&nau8821->jdet_work, nau8821_jdet_work); + ret = devm_request_threaded_irq(nau8821->dev, nau8821->irq, NULL, nau8821_interrupt, IRQF_TRIGGER_LOW | IRQF_ONESHOT, "nau8821", nau8821); diff --git a/sound/soc/codecs/nau8821.h b/sound/soc/codecs/nau8821.h index f0935ffafcbe..88602923780d 100644 --- a/sound/soc/codecs/nau8821.h +++ b/sound/soc/codecs/nau8821.h @@ -561,7 +561,7 @@ struct nau8821 { struct regmap *regmap; struct snd_soc_dapm_context *dapm; struct snd_soc_jack *jack; - struct work_struct jdet_work; + struct delayed_work jdet_work; int irq; int clk_id; int micbias_voltage; From ed25dcfbc4327570b28f0328a8e17d121434c0ea Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Fri, 26 Sep 2025 12:41:07 -0700 Subject: [PATCH 056/330] KVM: arm64: nv: Don't treat ZCR_EL2 as a 'mapped' register Unlike the other mapped EL2 sysregs ZCR_EL2 isn't guaranteed to be resident when a vCPU is loaded as it actually follows the SVE context. As such, the contents of ZCR_EL1 may belong to another guest if the vCPU has been preempted before reaching sysreg emulation. Unconditionally use the in-memory value of ZCR_EL2 and switch to the memory-only accessors. The in-memory value is guaranteed to be valid as fpsimd_lazy_switch_to_{guest,host}() will restore/save the register appropriately. Signed-off-by: Oliver Upton Signed-off-by: Marc Zyngier --- arch/arm64/kvm/sys_regs.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index 91053aa832d0..4a75e5f0c259 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -203,7 +203,6 @@ static void locate_register(const struct kvm_vcpu *vcpu, enum vcpu_sysreg reg, MAPPED_EL2_SYSREG(AMAIR_EL2, AMAIR_EL1, NULL ); MAPPED_EL2_SYSREG(ELR_EL2, ELR_EL1, NULL ); MAPPED_EL2_SYSREG(SPSR_EL2, SPSR_EL1, NULL ); - MAPPED_EL2_SYSREG(ZCR_EL2, ZCR_EL1, NULL ); MAPPED_EL2_SYSREG(CONTEXTIDR_EL2, CONTEXTIDR_EL1, NULL ); MAPPED_EL2_SYSREG(SCTLR2_EL2, SCTLR2_EL1, NULL ); case CNTHCTL_EL2: @@ -2709,14 +2708,13 @@ static bool access_zcr_el2(struct kvm_vcpu *vcpu, } if (!p->is_write) { - p->regval = vcpu_read_sys_reg(vcpu, ZCR_EL2); + p->regval = __vcpu_sys_reg(vcpu, ZCR_EL2); return true; } vq = SYS_FIELD_GET(ZCR_ELx, LEN, p->regval) + 1; vq = min(vq, vcpu_sve_max_vq(vcpu)); - vcpu_write_sys_reg(vcpu, vq - 1, ZCR_EL2); - + __vcpu_assign_sys_reg(vcpu, ZCR_EL2, vq - 1); return true; } From 9a1950f97741a23fc68a7b2cfd487e059d389be5 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Fri, 26 Sep 2025 12:41:08 -0700 Subject: [PATCH 057/330] KVM: arm64: nv: Don't advance PC when pending an SVE exception Jan reports that running a nested guest on Neoverse-V2 leads to a WARN in the host due to simultaneously pending an exception and PC increment after an access to ZCR_EL2. Returning true from a sysreg accessor is an indication that the sysreg instruction has been retired. Of course this isn't the case when we've pended a synchronous SVE exception for the guest. Fix the return value and let the exception propagate to the guest as usual. Reported-by: Jan Kotas Closes: https://lore.kernel.org/kvmarm/865xd61tt5.wl-maz@kernel.org/ Signed-off-by: Oliver Upton Signed-off-by: Marc Zyngier --- arch/arm64/kvm/sys_regs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index 4a75e5f0c259..ee8a7033c85b 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -2704,7 +2704,7 @@ static bool access_zcr_el2(struct kvm_vcpu *vcpu, if (guest_hyp_sve_traps_enabled(vcpu)) { kvm_inject_nested_sve_trap(vcpu); - return true; + return false; } if (!p->is_write) { From a46c09b382eea3f9e3d16576096b987a2171fcca Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Fri, 26 Sep 2025 15:42:46 -0700 Subject: [PATCH 058/330] KVM: arm64: Use the in-context stage-1 in __kvm_find_s1_desc_level() Running the external_aborts selftest at EL2 leads to an ugly splat due to the stage-1 MMU being disabled for the walked context, owing to the fact that __kvm_find_s1_desc_level() is hardcoded to the EL1&0 regime. Select the appropriate translation regime for the stage-1 walk based on the current vCPU context. Fixes: b8e625167a32 ("KVM: arm64: Add S1 IPA to page table level walker") Signed-off-by: Oliver Upton Signed-off-by: Marc Zyngier --- arch/arm64/kvm/at.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/arch/arm64/kvm/at.c b/arch/arm64/kvm/at.c index 20bb9af125b1..e2e06ec8a67b 100644 --- a/arch/arm64/kvm/at.c +++ b/arch/arm64/kvm/at.c @@ -1602,13 +1602,17 @@ int __kvm_find_s1_desc_level(struct kvm_vcpu *vcpu, u64 va, u64 ipa, int *level) .fn = match_s1_desc, .priv = &dm, }, - .regime = TR_EL10, .as_el0 = false, .pan = false, }; struct s1_walk_result wr = {}; int ret; + if (is_hyp_ctxt(vcpu)) + wi.regime = vcpu_el2_e2h_is_set(vcpu) ? TR_EL20 : TR_EL2; + else + wi.regime = TR_EL10; + ret = setup_s1_walk(vcpu, &wi, &wr, va); if (ret) return ret; From 890c608b4d5e6a616693da92a2d4e7de4ab9e6c5 Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Fri, 26 Sep 2025 15:44:54 -0700 Subject: [PATCH 059/330] KVM: arm64: selftests: Test effective value of HCR_EL2.AMO A defect against the architecture now allows an implementation to treat AMO as 1 when HCR_EL2.{E2H, TGE} = {1, 0}. KVM now takes advantage of this interpretation to address a quality of emulation issue w.r.t. SError injection. Add a corresponding test case and expect a pending SError to be taken. Signed-off-by: Oliver Upton Signed-off-by: Marc Zyngier --- .../selftests/kvm/arm64/external_aborts.c | 43 +++++++++++++++++++ .../selftests/kvm/include/arm64/processor.h | 12 +++++- 2 files changed, 54 insertions(+), 1 deletion(-) diff --git a/tools/testing/selftests/kvm/arm64/external_aborts.c b/tools/testing/selftests/kvm/arm64/external_aborts.c index 592b26ded779..d8fe17a6cc59 100644 --- a/tools/testing/selftests/kvm/arm64/external_aborts.c +++ b/tools/testing/selftests/kvm/arm64/external_aborts.c @@ -359,6 +359,44 @@ static void test_mmio_ease(void) kvm_vm_free(vm); } +static void test_serror_amo_guest(void) +{ + /* + * The ISB is entirely unnecessary (and highlights how FEAT_NV2 is borked) + * since the write is redirected to memory. But don't write (intentionally) + * broken code! + */ + sysreg_clear_set(hcr_el2, HCR_EL2_AMO | HCR_EL2_TGE, 0); + isb(); + + GUEST_SYNC(0); + GUEST_ASSERT(read_sysreg(isr_el1) & ISR_EL1_A); + + /* + * KVM treats the effective value of AMO as 1 when + * HCR_EL2.{E2H,TGE} = {1, 0}, meaning the SError will be taken when + * unmasked. + */ + local_serror_enable(); + isb(); + local_serror_disable(); + + GUEST_FAIL("Should've taken pending SError exception"); +} + +static void test_serror_amo(void) +{ + struct kvm_vcpu *vcpu; + struct kvm_vm *vm = vm_create_with_dabt_handler(&vcpu, test_serror_amo_guest, + unexpected_dabt_handler); + + vm_install_exception_handler(vm, VECTOR_ERROR_CURRENT, expect_serror_handler); + vcpu_run_expect_sync(vcpu); + vcpu_inject_serror(vcpu); + vcpu_run_expect_done(vcpu); + kvm_vm_free(vm); +} + int main(void) { test_mmio_abort(); @@ -369,4 +407,9 @@ int main(void) test_serror_emulated(); test_mmio_ease(); test_s1ptw_abort(); + + if (!test_supports_el2()) + return 0; + + test_serror_amo(); } diff --git a/tools/testing/selftests/kvm/include/arm64/processor.h b/tools/testing/selftests/kvm/include/arm64/processor.h index 6f481475c135..ff928716574d 100644 --- a/tools/testing/selftests/kvm/include/arm64/processor.h +++ b/tools/testing/selftests/kvm/include/arm64/processor.h @@ -305,7 +305,17 @@ void test_wants_mte(void); void test_disable_default_vgic(void); bool vm_supports_el2(struct kvm_vm *vm); -static bool vcpu_has_el2(struct kvm_vcpu *vcpu) + +static inline bool test_supports_el2(void) +{ + struct kvm_vm *vm = vm_create(1); + bool supported = vm_supports_el2(vm); + + kvm_vm_free(vm); + return supported; +} + +static inline bool vcpu_has_el2(struct kvm_vcpu *vcpu) { return vcpu->init.features[0] & BIT(KVM_ARM_VCPU_HAS_EL2); } From cb49b7b8622e914171e9eb7197c006320889e0fc Mon Sep 17 00:00:00 2001 From: Sean Christopherson Date: Fri, 26 Sep 2025 08:58:38 -0700 Subject: [PATCH 060/330] KVM: arm64: selftests: Track width of timer counter as "int", not "uint64_t" Store the width of arm64's timer counter as an "int", not a "uint64_t". ilog2() returns an "int", and more importantly using what is an "unsigned long" under the hood makes clang unhappy due to a type mismatch when clamping the width to a sane value. arm64/arch_timer_edge_cases.c:1032:10: error: comparison of distinct pointer types ('typeof (width) *' (aka 'unsigned long *') and 'typeof (56) *' (aka 'int *')) [-Werror,-Wcompare-distinct-pointer-types] 1032 | width = clamp(width, 56, 64); | ^~~~~~~~~~~~~~~~~~~~ tools/include/linux/kernel.h:47:45: note: expanded from macro 'clamp' 47 | #define clamp(val, lo, hi) min((typeof(val))max(val, lo), hi) | ^~~~~~~~~~~~ tools/include/linux/kernel.h:33:17: note: expanded from macro 'max' 33 | (void) (&_max1 == &_max2); \ | ~~~~~~ ^ ~~~~~~ tools/include/linux/kernel.h:39:9: note: expanded from macro 'min' 39 | typeof(x) _min1 = (x); \ | ^ Fixes: fad4cf944839 ("KVM: arm64: selftests: Determine effective counter width in arch_timer_edge_cases") Cc: Sebastian Ott Signed-off-by: Sean Christopherson Reviewed-by: Oliver Upton Signed-off-by: Marc Zyngier --- tools/testing/selftests/kvm/arm64/arch_timer_edge_cases.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/testing/selftests/kvm/arm64/arch_timer_edge_cases.c b/tools/testing/selftests/kvm/arm64/arch_timer_edge_cases.c index 91906414a474..993c9e38e729 100644 --- a/tools/testing/selftests/kvm/arm64/arch_timer_edge_cases.c +++ b/tools/testing/selftests/kvm/arm64/arch_timer_edge_cases.c @@ -1020,7 +1020,7 @@ static void set_counter_defaults(void) { const uint64_t MIN_ROLLOVER_SECS = 40ULL * 365 * 24 * 3600; uint64_t freq = read_sysreg(CNTFRQ_EL0); - uint64_t width = ilog2(MIN_ROLLOVER_SECS * freq); + int width = ilog2(MIN_ROLLOVER_SECS * freq); width = clamp(width, 56, 64); CVAL_MAX = GENMASK_ULL(width - 1, 0); From 0aa1b76fe1429629215a7c79820e4b96233ac4a3 Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Tue, 30 Sep 2025 01:52:37 -0700 Subject: [PATCH 061/330] KVM: arm64: Prevent access to vCPU events before init Another day, another syzkaller bug. KVM erroneously allows userspace to pend vCPU events for a vCPU that hasn't been initialized yet, leading to KVM interpreting a bunch of uninitialized garbage for routing / injecting the exception. In one case the injection code and the hyp disagree on whether the vCPU has a 32bit EL1 and put the vCPU into an illegal mode for AArch64, tripping the BUG() in exception_target_el() during the next injection: kernel BUG at arch/arm64/kvm/inject_fault.c:40! Internal error: Oops - BUG: 00000000f2000800 [#1] SMP CPU: 3 UID: 0 PID: 318 Comm: repro Not tainted 6.17.0-rc4-00104-g10fd0285305d #6 PREEMPT Hardware name: linux,dummy-virt (DT) pstate: 21402009 (nzCv daif +PAN -UAO -TCO +DIT -SSBS BTYPE=--) pc : exception_target_el+0x88/0x8c lr : pend_serror_exception+0x18/0x13c sp : ffff800082f03a10 x29: ffff800082f03a10 x28: ffff0000cb132280 x27: 0000000000000000 x26: 0000000000000000 x25: ffff0000c2a99c20 x24: 0000000000000000 x23: 0000000000008000 x22: 0000000000000002 x21: 0000000000000004 x20: 0000000000008000 x19: ffff0000c2a99c20 x18: 0000000000000000 x17: 0000000000000000 x16: 0000000000000000 x15: 00000000200000c0 x14: 0000000000000000 x13: 0000000000000000 x12: 0000000000000000 x11: 0000000000000000 x10: 0000000000000000 x9 : 0000000000000000 x8 : ffff800082f03af8 x7 : 0000000000000000 x6 : 0000000000000000 x5 : ffff800080f621f0 x4 : 0000000000000000 x3 : 0000000000000000 x2 : 000000000040009b x1 : 0000000000000003 x0 : ffff0000c2a99c20 Call trace: exception_target_el+0x88/0x8c (P) kvm_inject_serror_esr+0x40/0x3b4 __kvm_arm_vcpu_set_events+0xf0/0x100 kvm_arch_vcpu_ioctl+0x180/0x9d4 kvm_vcpu_ioctl+0x60c/0x9f4 __arm64_sys_ioctl+0xac/0x104 invoke_syscall+0x48/0x110 el0_svc_common.constprop.0+0x40/0xe0 do_el0_svc+0x1c/0x28 el0_svc+0x34/0xf0 el0t_64_sync_handler+0xa0/0xe4 el0t_64_sync+0x198/0x19c Code: f946bc01 b4fffe61 9101e020 17fffff2 (d4210000) Reject the ioctls outright as no sane VMM would call these before KVM_ARM_VCPU_INIT anyway. Even if it did the exception would've been thrown away by the eventual reset of the vCPU's state. Cc: stable@vger.kernel.org # 6.17 Fixes: b7b27facc7b5 ("arm/arm64: KVM: Add KVM_GET/SET_VCPU_EVENTS") Signed-off-by: Oliver Upton Signed-off-by: Marc Zyngier --- arch/arm64/kvm/arm.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index f21d1b7f20f8..f01cacb669cf 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -1794,6 +1794,9 @@ long kvm_arch_vcpu_ioctl(struct file *filp, case KVM_GET_VCPU_EVENTS: { struct kvm_vcpu_events events; + if (!kvm_vcpu_initialized(vcpu)) + return -ENOEXEC; + if (kvm_arm_vcpu_get_events(vcpu, &events)) return -EINVAL; @@ -1805,6 +1808,9 @@ long kvm_arch_vcpu_ioctl(struct file *filp, case KVM_SET_VCPU_EVENTS: { struct kvm_vcpu_events events; + if (!kvm_vcpu_initialized(vcpu)) + return -ENOEXEC; + if (copy_from_user(&events, argp, sizeof(events))) return -EFAULT; From cc4309324dc695f62d25d56c0b29805e9724170c Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Tue, 30 Sep 2025 16:36:20 -0700 Subject: [PATCH 062/330] KVM: arm64: Document vCPU event ioctls as requiring init'ed vCPU KVM rejects calls to KVM_{GET,SET}_VCPU_EVENTS for an uninitialized vCPU as of commit cc96679f3c03 ("KVM: arm64: Prevent access to vCPU events before init"). Update the corresponding API documentation. Signed-off-by: Oliver Upton Signed-off-by: Marc Zyngier --- Documentation/virt/kvm/api.rst | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/Documentation/virt/kvm/api.rst b/Documentation/virt/kvm/api.rst index 6ae24c5ca559..4973c74db5c6 100644 --- a/Documentation/virt/kvm/api.rst +++ b/Documentation/virt/kvm/api.rst @@ -1229,6 +1229,9 @@ It is not possible to read back a pending external abort (injected via KVM_SET_VCPU_EVENTS or otherwise) because such an exception is always delivered directly to the virtual CPU). +Calling this ioctl on a vCPU that hasn't been initialized will return +-ENOEXEC. + :: struct kvm_vcpu_events { @@ -1309,6 +1312,8 @@ exceptions by manipulating individual registers using the KVM_SET_ONE_REG API. See KVM_GET_VCPU_EVENTS for the data structure. +Calling this ioctl on a vCPU that hasn't been initialized will return +-ENOEXEC. 4.33 KVM_GET_DEBUGREGS ---------------------- From a133052666bed0dc0b169952e9d3f9e6b2125f9a Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Tue, 30 Sep 2025 12:33:02 -0700 Subject: [PATCH 063/330] KVM: selftests: Fix irqfd_test for non-x86 architectures The KVM_IRQFD ioctl fails if no irqchip is present in-kernel, which isn't too surprising as there's not much KVM can do for an IRQ if it cannot resolve a destination. As written the irqfd_test assumes that a 'default' VM created in selftests has an in-kernel irqchip created implicitly. That may be the case on x86 but it isn't necessarily true on other architectures. Add an arch predicate indicating if 'default' VMs get an irqchip and make the irqfd_test depend on it. Work around arm64 VGIC initialization requirements by using vm_create_with_one_vcpu(), ignoring the created vCPU as it isn't used for the test. Reported-by: Sebastian Ott Reported-by: Naresh Kamboju Acked-by: Sean Christopherson Fixes: 7e9b231c402a ("KVM: selftests: Add a KVM_IRQFD test to verify uniqueness requirements") Signed-off-by: Oliver Upton Signed-off-by: Marc Zyngier --- tools/testing/selftests/kvm/include/kvm_util.h | 2 ++ tools/testing/selftests/kvm/irqfd_test.c | 14 +++++++++++--- tools/testing/selftests/kvm/lib/arm64/processor.c | 5 +++++ tools/testing/selftests/kvm/lib/kvm_util.c | 5 +++++ tools/testing/selftests/kvm/lib/s390/processor.c | 5 +++++ tools/testing/selftests/kvm/lib/x86/processor.c | 5 +++++ 6 files changed, 33 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/kvm/include/kvm_util.h b/tools/testing/selftests/kvm/include/kvm_util.h index 26cc30290e76..112d3f443a17 100644 --- a/tools/testing/selftests/kvm/include/kvm_util.h +++ b/tools/testing/selftests/kvm/include/kvm_util.h @@ -1273,4 +1273,6 @@ bool vm_is_gpa_protected(struct kvm_vm *vm, vm_paddr_t paddr); uint32_t guest_get_vcpuid(void); +bool kvm_arch_has_default_irqchip(void); + #endif /* SELFTEST_KVM_UTIL_H */ diff --git a/tools/testing/selftests/kvm/irqfd_test.c b/tools/testing/selftests/kvm/irqfd_test.c index 7c301b4c7005..5d7590d01868 100644 --- a/tools/testing/selftests/kvm/irqfd_test.c +++ b/tools/testing/selftests/kvm/irqfd_test.c @@ -89,11 +89,19 @@ static void juggle_eventfd_primary(struct kvm_vm *vm, int eventfd) int main(int argc, char *argv[]) { pthread_t racing_thread; + struct kvm_vcpu *unused; int r, i; - /* Create "full" VMs, as KVM_IRQFD requires an in-kernel IRQ chip. */ - vm1 = vm_create(1); - vm2 = vm_create(1); + TEST_REQUIRE(kvm_arch_has_default_irqchip()); + + /* + * Create "full" VMs, as KVM_IRQFD requires an in-kernel IRQ chip. Also + * create an unused vCPU as certain architectures (like arm64) need to + * complete IRQ chip initialization after all possible vCPUs for a VM + * have been created. + */ + vm1 = vm_create_with_one_vcpu(&unused, NULL); + vm2 = vm_create_with_one_vcpu(&unused, NULL); WRITE_ONCE(__eventfd, kvm_new_eventfd()); diff --git a/tools/testing/selftests/kvm/lib/arm64/processor.c b/tools/testing/selftests/kvm/lib/arm64/processor.c index 369a4c87dd8f..54f6d17c78f7 100644 --- a/tools/testing/selftests/kvm/lib/arm64/processor.c +++ b/tools/testing/selftests/kvm/lib/arm64/processor.c @@ -725,3 +725,8 @@ void kvm_arch_vm_release(struct kvm_vm *vm) if (vm->arch.has_gic) close(vm->arch.gic_fd); } + +bool kvm_arch_has_default_irqchip(void) +{ + return request_vgic && kvm_supports_vgic_v3(); +} diff --git a/tools/testing/selftests/kvm/lib/kvm_util.c b/tools/testing/selftests/kvm/lib/kvm_util.c index 6743fbd9bd67..a35adfebfa23 100644 --- a/tools/testing/selftests/kvm/lib/kvm_util.c +++ b/tools/testing/selftests/kvm/lib/kvm_util.c @@ -2344,3 +2344,8 @@ bool vm_is_gpa_protected(struct kvm_vm *vm, vm_paddr_t paddr) pg = paddr >> vm->page_shift; return sparsebit_is_set(region->protected_phy_pages, pg); } + +__weak bool kvm_arch_has_default_irqchip(void) +{ + return false; +} diff --git a/tools/testing/selftests/kvm/lib/s390/processor.c b/tools/testing/selftests/kvm/lib/s390/processor.c index 20cfe970e3e3..8ceeb17c819a 100644 --- a/tools/testing/selftests/kvm/lib/s390/processor.c +++ b/tools/testing/selftests/kvm/lib/s390/processor.c @@ -221,3 +221,8 @@ void vcpu_arch_dump(FILE *stream, struct kvm_vcpu *vcpu, uint8_t indent) void assert_on_unhandled_exception(struct kvm_vcpu *vcpu) { } + +bool kvm_arch_has_default_irqchip(void) +{ + return true; +} diff --git a/tools/testing/selftests/kvm/lib/x86/processor.c b/tools/testing/selftests/kvm/lib/x86/processor.c index c748cd9b2eef..b418502c5ecc 100644 --- a/tools/testing/selftests/kvm/lib/x86/processor.c +++ b/tools/testing/selftests/kvm/lib/x86/processor.c @@ -1318,3 +1318,8 @@ bool sys_clocksource_is_based_on_tsc(void) return ret; } + +bool kvm_arch_has_default_irqchip(void) +{ + return true; +} From 05a02490faeb952f1c8d2f5c38346fa0a717a483 Mon Sep 17 00:00:00 2001 From: Osama Abdelkader Date: Tue, 30 Sep 2025 16:56:21 +0300 Subject: [PATCH 064/330] KVM: arm64: Remove unreachable break after return Remove an unnecessary 'break' statement that follows a 'return' in arch/arm64/kvm/at.c. The break is unreachable. Signed-off-by: Osama Abdelkader Reviewed-by: Zenghui Yu Signed-off-by: Marc Zyngier --- arch/arm64/kvm/at.c | 1 - 1 file changed, 1 deletion(-) diff --git a/arch/arm64/kvm/at.c b/arch/arm64/kvm/at.c index e2e06ec8a67b..be26d5aa668c 100644 --- a/arch/arm64/kvm/at.c +++ b/arch/arm64/kvm/at.c @@ -91,7 +91,6 @@ static enum trans_regime compute_translation_regime(struct kvm_vcpu *vcpu, u32 o case OP_AT_S1E2W: case OP_AT_S1E2A: return vcpu_el2_e2h_is_set(vcpu) ? TR_EL20 : TR_EL2; - break; default: return (vcpu_el2_e2h_is_set(vcpu) && vcpu_el2_tge_is_set(vcpu)) ? TR_EL20 : TR_EL10; From 9a7f87eb587da49993f47f44c4c5535d8de76750 Mon Sep 17 00:00:00 2001 From: Zenghui Yu Date: Sun, 12 Oct 2025 23:43:52 +0800 Subject: [PATCH 065/330] KVM: arm64: selftests: Sync ID_AA64PFR1, MPIDR, CLIDR in guest We forgot to sync several registers (ID_AA64PFR1, MPIDR, CLIDR) in guest to make sure that the guest had seen the written value. Add them to the list. Signed-off-by: Zenghui Yu Reviewed-By: Ben Horgan Signed-off-by: Marc Zyngier --- tools/testing/selftests/kvm/arm64/set_id_regs.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/tools/testing/selftests/kvm/arm64/set_id_regs.c b/tools/testing/selftests/kvm/arm64/set_id_regs.c index 8ff1e853f7f8..5e24f77868b5 100644 --- a/tools/testing/selftests/kvm/arm64/set_id_regs.c +++ b/tools/testing/selftests/kvm/arm64/set_id_regs.c @@ -249,11 +249,14 @@ static void guest_code(void) GUEST_REG_SYNC(SYS_ID_AA64ISAR2_EL1); GUEST_REG_SYNC(SYS_ID_AA64ISAR3_EL1); GUEST_REG_SYNC(SYS_ID_AA64PFR0_EL1); + GUEST_REG_SYNC(SYS_ID_AA64PFR1_EL1); GUEST_REG_SYNC(SYS_ID_AA64MMFR0_EL1); GUEST_REG_SYNC(SYS_ID_AA64MMFR1_EL1); GUEST_REG_SYNC(SYS_ID_AA64MMFR2_EL1); GUEST_REG_SYNC(SYS_ID_AA64MMFR3_EL1); GUEST_REG_SYNC(SYS_ID_AA64ZFR0_EL1); + GUEST_REG_SYNC(SYS_MPIDR_EL1); + GUEST_REG_SYNC(SYS_CLIDR_EL1); GUEST_REG_SYNC(SYS_CTR_EL0); GUEST_REG_SYNC(SYS_MIDR_EL1); GUEST_REG_SYNC(SYS_REVIDR_EL1); From b9ce79887e270ed64ed499aa69f903cdca401c2f Mon Sep 17 00:00:00 2001 From: Markus Elfring Date: Fri, 10 Oct 2025 21:04:16 +0200 Subject: [PATCH 066/330] smb: client: Return a status code only as a constant in sid_to_id() Return a status code without storing it in an intermediate variable. This issue was detected by using the Coccinelle software. Signed-off-by: Markus Elfring Signed-off-by: Steve French --- fs/smb/client/cifsacl.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/fs/smb/client/cifsacl.c b/fs/smb/client/cifsacl.c index 63b3b1290bed..ce2ebc213a1d 100644 --- a/fs/smb/client/cifsacl.c +++ b/fs/smb/client/cifsacl.c @@ -339,7 +339,6 @@ int sid_to_id(struct cifs_sb_info *cifs_sb, struct smb_sid *psid, struct cifs_fattr *fattr, uint sidtype) { - int rc = 0; struct key *sidkey; char *sidstr; const struct cred *saved_cred; @@ -446,12 +445,12 @@ sid_to_id(struct cifs_sb_info *cifs_sb, struct smb_sid *psid, * fails then we just fall back to using the ctx->linux_uid/linux_gid. */ got_valid_id: - rc = 0; if (sidtype == SIDOWNER) fattr->cf_uid = fuid; else fattr->cf_gid = fgid; - return rc; + + return 0; } int From 911063b590ce77d473e92716f05f34712f97ef95 Mon Sep 17 00:00:00 2001 From: Markus Elfring Date: Fri, 10 Oct 2025 14:48:13 +0200 Subject: [PATCH 067/330] smb: client: Omit one redundant variable assignment in cifs_xattr_set() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The local variable “rc” is assigned a value in an if branch without using it before it is reassigned there. Thus delete this assignment statement. Signed-off-by: Markus Elfring Signed-off-by: Steve French --- fs/smb/client/xattr.c | 1 - 1 file changed, 1 deletion(-) diff --git a/fs/smb/client/xattr.c b/fs/smb/client/xattr.c index b88fa04f5792..029910d56c22 100644 --- a/fs/smb/client/xattr.c +++ b/fs/smb/client/xattr.c @@ -178,7 +178,6 @@ static int cifs_xattr_set(const struct xattr_handler *handler, memcpy(pacl, value, size); if (pTcon->ses->server->ops->set_acl) { int aclflags = 0; - rc = 0; switch (handler->flags) { case XATTR_CIFS_NTSD_FULL: From e487f13cc94fa80c71f217d7b176e48fbb5d6c46 Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:21 +0200 Subject: [PATCH 068/330] smb: smbdirect: introduce smbdirect_mr_io.{kref,mutex} and SMBDIRECT_MR_DISABLED This will be used in the next commits in order to improve the client code. A broken connection can just disable the smbdirect_mr_io while keeping the memory arround for the caller. Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/common/smbdirect/smbdirect_socket.h | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/fs/smb/common/smbdirect/smbdirect_socket.h b/fs/smb/common/smbdirect/smbdirect_socket.h index db22a1d0546b..361db7f9f623 100644 --- a/fs/smb/common/smbdirect/smbdirect_socket.h +++ b/fs/smb/common/smbdirect/smbdirect_socket.h @@ -437,13 +437,22 @@ enum smbdirect_mr_state { SMBDIRECT_MR_READY, SMBDIRECT_MR_REGISTERED, SMBDIRECT_MR_INVALIDATED, - SMBDIRECT_MR_ERROR + SMBDIRECT_MR_ERROR, + SMBDIRECT_MR_DISABLED }; struct smbdirect_mr_io { struct smbdirect_socket *socket; struct ib_cqe cqe; + /* + * We can have up to two references: + * 1. by the connection + * 2. by the registration + */ + struct kref kref; + struct mutex mutex; + struct list_head list; enum smbdirect_mr_state state; From abe5b71c391352127925bf951f3169d205c5caa7 Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:22 +0200 Subject: [PATCH 069/330] smb: client: change smbd_deregister_mr() to return void No callers checks the return value and this makes further changes easier. Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 4 +--- fs/smb/client/smbdirect.h | 2 +- 2 files changed, 2 insertions(+), 4 deletions(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index 316f398c70f4..a20aa2ddf57d 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -2612,7 +2612,7 @@ static void local_inv_done(struct ib_cq *cq, struct ib_wc *wc) * and we have to locally invalidate the buffer to prevent data is being * modified by remote peer after upper layer consumes it */ -int smbd_deregister_mr(struct smbdirect_mr_io *smbdirect_mr) +void smbd_deregister_mr(struct smbdirect_mr_io *smbdirect_mr) { struct ib_send_wr *wr; struct smbdirect_socket *sc = smbdirect_mr->socket; @@ -2662,8 +2662,6 @@ int smbd_deregister_mr(struct smbdirect_mr_io *smbdirect_mr) done: if (atomic_dec_and_test(&sc->mr_io.used.count)) wake_up(&sc->mr_io.cleanup.wait_queue); - - return rc; } static bool smb_set_sge(struct smb_extract_to_rdma *rdma, diff --git a/fs/smb/client/smbdirect.h b/fs/smb/client/smbdirect.h index d67ac5ddaff4..577d37dbeb8a 100644 --- a/fs/smb/client/smbdirect.h +++ b/fs/smb/client/smbdirect.h @@ -60,7 +60,7 @@ int smbd_send(struct TCP_Server_Info *server, struct smbdirect_mr_io *smbd_register_mr( struct smbd_connection *info, struct iov_iter *iter, bool writing, bool need_invalidate); -int smbd_deregister_mr(struct smbdirect_mr_io *mr); +void smbd_deregister_mr(struct smbdirect_mr_io *mr); #else #define cifs_rdma_enabled(server) 0 From 19421ec198981f60373080af67e67b6a6fcf191e Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:23 +0200 Subject: [PATCH 070/330] smb: client: let destroy_mr_list() call list_del(&mr->list) This makes the code clearer and will make further changes easier. Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index a20aa2ddf57d..b7be67dacd09 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -2363,6 +2363,7 @@ static void destroy_mr_list(struct smbdirect_socket *sc) mr->sgt.nents, mr->dir); ib_dereg_mr(mr->mr); kfree(mr->sgt.sgl); + list_del(&mr->list); kfree(mr); } } From a8e128b293e2b08d4ecca7c63ed1cb5b97f30af9 Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:24 +0200 Subject: [PATCH 071/330] smb: client: let destroy_mr_list() remove locked from the list This should make sure get_mr() can't see the removed entries. Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index b7be67dacd09..b974ca4e0b2e 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -2355,9 +2355,16 @@ static void smbd_mr_recovery_work(struct work_struct *work) static void destroy_mr_list(struct smbdirect_socket *sc) { struct smbdirect_mr_io *mr, *tmp; + LIST_HEAD(all_list); + unsigned long flags; disable_work_sync(&sc->mr_io.recovery_work); - list_for_each_entry_safe(mr, tmp, &sc->mr_io.all.list, list) { + + spin_lock_irqsave(&sc->mr_io.all.lock, flags); + list_splice_tail_init(&sc->mr_io.all.list, &all_list); + spin_unlock_irqrestore(&sc->mr_io.all.lock, flags); + + list_for_each_entry_safe(mr, tmp, &all_list, list) { if (mr->state == SMBDIRECT_MR_INVALIDATED) ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); From 9bebb8924b27f06e7072f0b18a5f78cef561c810 Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:25 +0200 Subject: [PATCH 072/330] smb: client: improve logic in allocate_mr_list() - use 'mr' as variable name - use goto lables for easier cleanup - use destroy_mr_list() - style fixes - INIT_WORK(&sc->mr_io.recovery_work, smbd_mr_recovery_work) on success This will make further changes easier. Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 69 +++++++++++++++++++++------------------ 1 file changed, 37 insertions(+), 32 deletions(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index b974ca4e0b2e..658ca11cb26c 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -2385,10 +2385,9 @@ static void destroy_mr_list(struct smbdirect_socket *sc) static int allocate_mr_list(struct smbdirect_socket *sc) { struct smbdirect_socket_parameters *sp = &sc->parameters; - int i; - struct smbdirect_mr_io *smbdirect_mr, *tmp; - - INIT_WORK(&sc->mr_io.recovery_work, smbd_mr_recovery_work); + struct smbdirect_mr_io *mr; + int ret; + u32 i; if (sp->responder_resources == 0) { log_rdma_mr(ERR, "responder_resources negotiated as 0\n"); @@ -2397,42 +2396,48 @@ static int allocate_mr_list(struct smbdirect_socket *sc) /* Allocate more MRs (2x) than hardware responder_resources */ for (i = 0; i < sp->responder_resources * 2; i++) { - smbdirect_mr = kzalloc(sizeof(*smbdirect_mr), GFP_KERNEL); - if (!smbdirect_mr) - goto cleanup_entries; - smbdirect_mr->mr = ib_alloc_mr(sc->ib.pd, sc->mr_io.type, - sp->max_frmr_depth); - if (IS_ERR(smbdirect_mr->mr)) { + mr = kzalloc(sizeof(*mr), GFP_KERNEL); + if (!mr) { + ret = -ENOMEM; + goto kzalloc_mr_failed; + } + + mr->mr = ib_alloc_mr(sc->ib.pd, + sc->mr_io.type, + sp->max_frmr_depth); + if (IS_ERR(mr->mr)) { + ret = PTR_ERR(mr->mr); log_rdma_mr(ERR, "ib_alloc_mr failed mr_type=%x max_frmr_depth=%x\n", sc->mr_io.type, sp->max_frmr_depth); - goto out; + goto ib_alloc_mr_failed; } - smbdirect_mr->sgt.sgl = kcalloc(sp->max_frmr_depth, - sizeof(struct scatterlist), - GFP_KERNEL); - if (!smbdirect_mr->sgt.sgl) { - log_rdma_mr(ERR, "failed to allocate sgl\n"); - ib_dereg_mr(smbdirect_mr->mr); - goto out; - } - smbdirect_mr->state = SMBDIRECT_MR_READY; - smbdirect_mr->socket = sc; - list_add_tail(&smbdirect_mr->list, &sc->mr_io.all.list); + mr->sgt.sgl = kcalloc(sp->max_frmr_depth, + sizeof(struct scatterlist), + GFP_KERNEL); + if (!mr->sgt.sgl) { + ret = -ENOMEM; + log_rdma_mr(ERR, "failed to allocate sgl\n"); + goto kcalloc_sgl_failed; + } + mr->state = SMBDIRECT_MR_READY; + mr->socket = sc; + + list_add_tail(&mr->list, &sc->mr_io.all.list); atomic_inc(&sc->mr_io.ready.count); } + + INIT_WORK(&sc->mr_io.recovery_work, smbd_mr_recovery_work); + return 0; -out: - kfree(smbdirect_mr); -cleanup_entries: - list_for_each_entry_safe(smbdirect_mr, tmp, &sc->mr_io.all.list, list) { - list_del(&smbdirect_mr->list); - ib_dereg_mr(smbdirect_mr->mr); - kfree(smbdirect_mr->sgt.sgl); - kfree(smbdirect_mr); - } - return -ENOMEM; +kcalloc_sgl_failed: + ib_dereg_mr(mr->mr); +ib_alloc_mr_failed: + kfree(mr); +kzalloc_mr_failed: + destroy_mr_list(sc); + return ret; } /* From c8478502960eb8fb9847b36a66380adf421cdc62 Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:26 +0200 Subject: [PATCH 073/330] smb: client: improve logic in smbd_register_mr() - use 'mr' as variable name - style fixes This will make further changes easier. Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 52 +++++++++++++++++---------------------- 1 file changed, 23 insertions(+), 29 deletions(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index 658ca11cb26c..a863b6fff87a 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -2517,9 +2517,8 @@ struct smbdirect_mr_io *smbd_register_mr(struct smbd_connection *info, { struct smbdirect_socket *sc = &info->socket; struct smbdirect_socket_parameters *sp = &sc->parameters; - struct smbdirect_mr_io *smbdirect_mr; + struct smbdirect_mr_io *mr; int rc, num_pages; - enum dma_data_direction dir; struct ib_reg_wr *reg_wr; num_pages = iov_iter_npages(iter, sp->max_frmr_depth + 1); @@ -2530,49 +2529,45 @@ struct smbdirect_mr_io *smbd_register_mr(struct smbd_connection *info, return NULL; } - smbdirect_mr = get_mr(sc); - if (!smbdirect_mr) { + mr = get_mr(sc); + if (!mr) { log_rdma_mr(ERR, "get_mr returning NULL\n"); return NULL; } - dir = writing ? DMA_FROM_DEVICE : DMA_TO_DEVICE; - smbdirect_mr->dir = dir; - smbdirect_mr->need_invalidate = need_invalidate; - smbdirect_mr->sgt.nents = 0; - smbdirect_mr->sgt.orig_nents = 0; + mr->dir = writing ? DMA_FROM_DEVICE : DMA_TO_DEVICE; + mr->need_invalidate = need_invalidate; + mr->sgt.nents = 0; + mr->sgt.orig_nents = 0; log_rdma_mr(INFO, "num_pages=0x%x count=0x%zx depth=%u\n", num_pages, iov_iter_count(iter), sp->max_frmr_depth); - smbd_iter_to_mr(iter, &smbdirect_mr->sgt, sp->max_frmr_depth); + smbd_iter_to_mr(iter, &mr->sgt, sp->max_frmr_depth); - rc = ib_dma_map_sg(sc->ib.dev, smbdirect_mr->sgt.sgl, - smbdirect_mr->sgt.nents, dir); + rc = ib_dma_map_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); if (!rc) { log_rdma_mr(ERR, "ib_dma_map_sg num_pages=%x dir=%x rc=%x\n", - num_pages, dir, rc); + num_pages, mr->dir, rc); goto dma_map_error; } - rc = ib_map_mr_sg(smbdirect_mr->mr, smbdirect_mr->sgt.sgl, - smbdirect_mr->sgt.nents, NULL, PAGE_SIZE); - if (rc != smbdirect_mr->sgt.nents) { + rc = ib_map_mr_sg(mr->mr, mr->sgt.sgl, mr->sgt.nents, NULL, PAGE_SIZE); + if (rc != mr->sgt.nents) { log_rdma_mr(ERR, - "ib_map_mr_sg failed rc = %d nents = %x\n", - rc, smbdirect_mr->sgt.nents); + "ib_map_mr_sg failed rc = %d nents = %x\n", + rc, mr->sgt.nents); goto map_mr_error; } - ib_update_fast_reg_key(smbdirect_mr->mr, - ib_inc_rkey(smbdirect_mr->mr->rkey)); - reg_wr = &smbdirect_mr->wr; + ib_update_fast_reg_key(mr->mr, ib_inc_rkey(mr->mr->rkey)); + reg_wr = &mr->wr; reg_wr->wr.opcode = IB_WR_REG_MR; - smbdirect_mr->cqe.done = register_mr_done; - reg_wr->wr.wr_cqe = &smbdirect_mr->cqe; + mr->cqe.done = register_mr_done; + reg_wr->wr.wr_cqe = &mr->cqe; reg_wr->wr.num_sge = 0; reg_wr->wr.send_flags = IB_SEND_SIGNALED; - reg_wr->mr = smbdirect_mr->mr; - reg_wr->key = smbdirect_mr->mr->rkey; + reg_wr->mr = mr->mr; + reg_wr->key = mr->mr->rkey; reg_wr->access = writing ? IB_ACCESS_REMOTE_WRITE | IB_ACCESS_LOCAL_WRITE : IB_ACCESS_REMOTE_READ; @@ -2584,18 +2579,17 @@ struct smbdirect_mr_io *smbd_register_mr(struct smbd_connection *info, */ rc = ib_post_send(sc->ib.qp, ®_wr->wr, NULL); if (!rc) - return smbdirect_mr; + return mr; log_rdma_mr(ERR, "ib_post_send failed rc=%x reg_wr->key=%x\n", rc, reg_wr->key); /* If all failed, attempt to recover this MR by setting it SMBDIRECT_MR_ERROR*/ map_mr_error: - ib_dma_unmap_sg(sc->ib.dev, smbdirect_mr->sgt.sgl, - smbdirect_mr->sgt.nents, smbdirect_mr->dir); + ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); dma_map_error: - smbdirect_mr->state = SMBDIRECT_MR_ERROR; + mr->state = SMBDIRECT_MR_ERROR; if (atomic_dec_and_test(&sc->mr_io.used.count)) wake_up(&sc->mr_io.cleanup.wait_queue); From 56c817e31acedc8a9041b181a15755e5a7b55f2b Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:27 +0200 Subject: [PATCH 074/330] smb: client: improve logic in smbd_deregister_mr() - use 'mr' as variable name - style fixes This will make further changes easier. Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 35 ++++++++++++++++------------------- 1 file changed, 16 insertions(+), 19 deletions(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index a863b6fff87a..af0642e94d7e 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -2619,44 +2619,41 @@ static void local_inv_done(struct ib_cq *cq, struct ib_wc *wc) * and we have to locally invalidate the buffer to prevent data is being * modified by remote peer after upper layer consumes it */ -void smbd_deregister_mr(struct smbdirect_mr_io *smbdirect_mr) +void smbd_deregister_mr(struct smbdirect_mr_io *mr) { - struct ib_send_wr *wr; - struct smbdirect_socket *sc = smbdirect_mr->socket; - int rc = 0; + struct smbdirect_socket *sc = mr->socket; + + if (mr->need_invalidate) { + struct ib_send_wr *wr = &mr->inv_wr; + int rc; - if (smbdirect_mr->need_invalidate) { /* Need to finish local invalidation before returning */ - wr = &smbdirect_mr->inv_wr; wr->opcode = IB_WR_LOCAL_INV; - smbdirect_mr->cqe.done = local_inv_done; - wr->wr_cqe = &smbdirect_mr->cqe; + mr->cqe.done = local_inv_done; + wr->wr_cqe = &mr->cqe; wr->num_sge = 0; - wr->ex.invalidate_rkey = smbdirect_mr->mr->rkey; + wr->ex.invalidate_rkey = mr->mr->rkey; wr->send_flags = IB_SEND_SIGNALED; - init_completion(&smbdirect_mr->invalidate_done); + init_completion(&mr->invalidate_done); rc = ib_post_send(sc->ib.qp, wr, NULL); if (rc) { log_rdma_mr(ERR, "ib_post_send failed rc=%x\n", rc); smbd_disconnect_rdma_connection(sc); goto done; } - wait_for_completion(&smbdirect_mr->invalidate_done); - smbdirect_mr->need_invalidate = false; + wait_for_completion(&mr->invalidate_done); + mr->need_invalidate = false; } else /* * For remote invalidation, just set it to SMBDIRECT_MR_INVALIDATED * and defer to mr_recovery_work to recover the MR for next use */ - smbdirect_mr->state = SMBDIRECT_MR_INVALIDATED; + mr->state = SMBDIRECT_MR_INVALIDATED; - if (smbdirect_mr->state == SMBDIRECT_MR_INVALIDATED) { - ib_dma_unmap_sg( - sc->ib.dev, smbdirect_mr->sgt.sgl, - smbdirect_mr->sgt.nents, - smbdirect_mr->dir); - smbdirect_mr->state = SMBDIRECT_MR_READY; + if (mr->state == SMBDIRECT_MR_INVALIDATED) { + ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); + mr->state = SMBDIRECT_MR_READY; if (atomic_inc_return(&sc->mr_io.ready.count) == 1) wake_up(&sc->mr_io.ready.wait_queue); } else From b9c0becc2fceb53e4a958575344e92c0e4f812bb Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:28 +0200 Subject: [PATCH 075/330] smb: client: call ib_dma_unmap_sg if mr->sgt.nents is not 0 This seems to be the more reliable way to check if we need to call ib_dma_unmap_sg(). Fixes: c7398583340a ("CIFS: SMBD: Implement RDMA memory registration") Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 12 ++++++++---- 1 file changed, 8 insertions(+), 4 deletions(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index af0642e94d7e..21dcd326af3d 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -2365,9 +2365,8 @@ static void destroy_mr_list(struct smbdirect_socket *sc) spin_unlock_irqrestore(&sc->mr_io.all.lock, flags); list_for_each_entry_safe(mr, tmp, &all_list, list) { - if (mr->state == SMBDIRECT_MR_INVALIDATED) - ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, - mr->sgt.nents, mr->dir); + if (mr->sgt.nents) + ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); ib_dereg_mr(mr->mr); kfree(mr->sgt.sgl); list_del(&mr->list); @@ -2589,6 +2588,7 @@ struct smbdirect_mr_io *smbd_register_mr(struct smbd_connection *info, ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); dma_map_error: + mr->sgt.nents = 0; mr->state = SMBDIRECT_MR_ERROR; if (atomic_dec_and_test(&sc->mr_io.used.count)) wake_up(&sc->mr_io.cleanup.wait_queue); @@ -2651,8 +2651,12 @@ void smbd_deregister_mr(struct smbdirect_mr_io *mr) */ mr->state = SMBDIRECT_MR_INVALIDATED; - if (mr->state == SMBDIRECT_MR_INVALIDATED) { + if (mr->sgt.nents) { ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); + mr->sgt.nents = 0; + } + + if (mr->state == SMBDIRECT_MR_INVALIDATED) { mr->state = SMBDIRECT_MR_READY; if (atomic_inc_return(&sc->mr_io.ready.count) == 1) wake_up(&sc->mr_io.ready.wait_queue); From 1ef0e16c3d7ca07432987840d8eef1a9ffb67dec Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:29 +0200 Subject: [PATCH 076/330] smb: client: let destroy_mr_list() call ib_dereg_mr() before ib_dma_unmap_sg() This is more consistent as we call ib_dma_unmap_sg() only when the memory is no longer registered. This is the same pattern as calling ib_dma_unmap_sg() after IB_WR_LOCAL_INV. Fixes: c7398583340a ("CIFS: SMBD: Implement RDMA memory registration") Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index 21dcd326af3d..c3330e43488f 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -2365,9 +2365,10 @@ static void destroy_mr_list(struct smbdirect_socket *sc) spin_unlock_irqrestore(&sc->mr_io.all.lock, flags); list_for_each_entry_safe(mr, tmp, &all_list, list) { + if (mr->mr) + ib_dereg_mr(mr->mr); if (mr->sgt.nents) ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); - ib_dereg_mr(mr->mr); kfree(mr->sgt.sgl); list_del(&mr->list); kfree(mr); From c35dd838666d47de2848639234ec32e3ba22b49f Mon Sep 17 00:00:00 2001 From: Mukesh Ojha Date: Fri, 10 Oct 2025 23:17:07 +0530 Subject: [PATCH 077/330] KVM: arm64: Guard PMSCR_EL1 initialization with SPE presence check Commit efad60e46057 ("KVM: arm64: Initialize PMSCR_EL1 when in VHE") does not perform sufficient check before initializing PMSCR_EL1 to 0 when running in VHE mode. On some platforms, this causes the system to hang during boot, as EL3 has not delegated access to the Profiling Buffer to the Non-secure world, nor does it reinject an UNDEF on sysreg trap. To avoid this issue, restrict the PMSCR_EL1 initialization to CPUs that support Statistical Profiling Extension (FEAT_SPE) and have the Profiling Buffer accessible in Non-secure EL1. This is determined via a new helper `cpu_has_spe()` which checks both PMSVer and PMBIDR_EL1.P. This ensures the initialization only affects CPUs where SPE is implemented and usable, preventing boot failures on platforms where SPE is not properly configured. Fixes: efad60e46057 ("KVM: arm64: Initialize PMSCR_EL1 when in VHE") Signed-off-by: Mukesh Ojha Signed-off-by: Marc Zyngier --- arch/arm64/kvm/debug.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) diff --git a/arch/arm64/kvm/debug.c b/arch/arm64/kvm/debug.c index 3515a273eaa2..3ad6b7c6e4ba 100644 --- a/arch/arm64/kvm/debug.c +++ b/arch/arm64/kvm/debug.c @@ -15,6 +15,12 @@ #include #include +static int cpu_has_spe(u64 dfr0) +{ + return cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_EL1_PMSVer_SHIFT) && + !(read_sysreg_s(SYS_PMBIDR_EL1) & PMBIDR_EL1_P); +} + /** * kvm_arm_setup_mdcr_el2 - configure vcpu mdcr_el2 value * @@ -77,13 +83,12 @@ void kvm_init_host_debug_data(void) *host_data_ptr(debug_brps) = SYS_FIELD_GET(ID_AA64DFR0_EL1, BRPs, dfr0); *host_data_ptr(debug_wrps) = SYS_FIELD_GET(ID_AA64DFR0_EL1, WRPs, dfr0); + if (cpu_has_spe(dfr0)) + host_data_set_flag(HAS_SPE); + if (has_vhe()) return; - if (cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_EL1_PMSVer_SHIFT) && - !(read_sysreg_s(SYS_PMBIDR_EL1) & PMBIDR_EL1_P)) - host_data_set_flag(HAS_SPE); - /* Check if we have BRBE implemented and available at the host */ if (cpuid_feature_extract_unsigned_field(dfr0, ID_AA64DFR0_EL1_BRBE_SHIFT)) host_data_set_flag(HAS_BRBE); @@ -102,7 +107,7 @@ void kvm_init_host_debug_data(void) void kvm_debug_init_vhe(void) { /* Clear PMSCR_EL1.E{0,1}SPE which reset to UNKNOWN values. */ - if (SYS_FIELD_GET(ID_AA64DFR0_EL1, PMSVer, read_sysreg(id_aa64dfr0_el1))) + if (host_data_test_flag(HAS_SPE)) write_sysreg_el1(0, SYS_PMSCR); } From 2192d348c0aa0cc2e7249dc3709f21bfe0a0170c Mon Sep 17 00:00:00 2001 From: Zenghui Yu Date: Wed, 8 Oct 2025 23:45:20 +0800 Subject: [PATCH 078/330] KVM: arm64: selftests: Allocate vcpus with correct size vcpus array contains pointers to struct kvm_vcpu {}. It is way overkill to allocate the array with (nr_cpus * sizeof(struct kvm_vcpu)). Fix the allocation by using the correct size. Signed-off-by: Zenghui Yu Signed-off-by: Marc Zyngier --- tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c b/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c index 87922a89b134..b134a304f0a6 100644 --- a/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c +++ b/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c @@ -331,7 +331,7 @@ static void setup_vm(void) { int i; - vcpus = malloc(test_data.nr_cpus * sizeof(struct kvm_vcpu)); + vcpus = malloc(test_data.nr_cpus * sizeof(struct kvm_vcpu *)); TEST_ASSERT(vcpus, "Failed to allocate vCPU array"); vm = vm_create_with_vcpus(test_data.nr_cpus, guest_code, vcpus); From d5e6310a0d996493b1af9f3eeec418350523388b Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Tue, 7 Oct 2025 12:52:55 -0700 Subject: [PATCH 079/330] KVM: arm64: selftests: Actually enable IRQs in vgic_lpi_stress vgic_lpi_stress rather hilariously leaves IRQs disabled for the duration of the test. While the ITS translation of MSIs happens regardless of this, for completeness the guest should actually handle the LPIs. Signed-off-by: Oliver Upton Reviewed-by: Zenghui Yu Signed-off-by: Marc Zyngier --- tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c | 1 + 1 file changed, 1 insertion(+) diff --git a/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c b/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c index b134a304f0a6..687d04463983 100644 --- a/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c +++ b/tools/testing/selftests/kvm/arm64/vgic_lpi_stress.c @@ -123,6 +123,7 @@ static void guest_setup_gic(void) static void guest_code(size_t nr_lpis) { guest_setup_gic(); + local_irq_enable(); GUEST_SYNC(0); From 3193287ddffbce29fd1a79d812f543c0fe4861d1 Mon Sep 17 00:00:00 2001 From: Sascha Bischoff Date: Tue, 7 Oct 2025 16:07:13 +0000 Subject: [PATCH 080/330] KVM: arm64: gic-v3: Only set ICH_HCR traps for v2-on-v3 or v3 guests The ICH_HCR_EL2 traps are used when running on GICv3 hardware, or when running a GICv3-based guest using FEAT_GCIE_LEGACY on GICv5 hardware. When running a GICv2 guest on GICv3 hardware the traps are used to ensure that the guest never sees any part of GICv3 (only GICv2 is visible to the guest), and when running a GICv3 guest they are used to trap in specific scenarios. They are not applicable for a GICv2-native guest, and won't be applicable for a(n upcoming) GICv5 guest. The traps themselves are configured in the vGIC CPU IF state, which is stored as a union. Updating the wrong aperture of the union risks corrupting state, and therefore needs to be avoided at all costs. Bail early if we're not running a compatible guest (GICv2 on GICv3 hardware, GICv3 native, GICv3 on GICv5 hardware). Trap everything unconditionally if we're running a GICv2 guest on GICv3 hardware. Otherwise, conditionally set up GICv3-native trapping. Signed-off-by: Sascha Bischoff Signed-off-by: Marc Zyngier --- arch/arm64/kvm/vgic/vgic-v3.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/arch/arm64/kvm/vgic/vgic-v3.c b/arch/arm64/kvm/vgic/vgic-v3.c index f1c153106c56..6fbb4b099855 100644 --- a/arch/arm64/kvm/vgic/vgic-v3.c +++ b/arch/arm64/kvm/vgic/vgic-v3.c @@ -297,8 +297,11 @@ void vcpu_set_ich_hcr(struct kvm_vcpu *vcpu) { struct vgic_v3_cpu_if *vgic_v3 = &vcpu->arch.vgic_cpu.vgic_v3; + if (!vgic_is_v3(vcpu->kvm)) + return; + /* Hide GICv3 sysreg if necessary */ - if (!kvm_has_gicv3(vcpu->kvm)) { + if (vcpu->kvm->arch.vgic.vgic_model == KVM_DEV_TYPE_ARM_VGIC_V2) { vgic_v3->vgic_hcr |= (ICH_HCR_EL2_TALL0 | ICH_HCR_EL2_TALL1 | ICH_HCR_EL2_TC); return; From 164ecbf73c3ea61455e07eefdad8050a7b569558 Mon Sep 17 00:00:00 2001 From: Sascha Bischoff Date: Tue, 7 Oct 2025 15:48:54 +0000 Subject: [PATCH 081/330] Documentation: KVM: Update GICv3 docs for GICv5 hosts GICv5 hosts optionally include FEAT_GCIE_LEGACY, which allows them to execute GICv3-based VMs on GICv5 hardware. Update the GICv3 documentation to reflect this now that GICv3 guests are supports on compatible GICv5 hosts. Signed-off-by: Sascha Bischoff Signed-off-by: Marc Zyngier --- Documentation/virt/kvm/devices/arm-vgic-v3.rst | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/Documentation/virt/kvm/devices/arm-vgic-v3.rst b/Documentation/virt/kvm/devices/arm-vgic-v3.rst index ff02102f7141..5395ee66fc32 100644 --- a/Documentation/virt/kvm/devices/arm-vgic-v3.rst +++ b/Documentation/virt/kvm/devices/arm-vgic-v3.rst @@ -13,7 +13,8 @@ will act as the VM interrupt controller, requiring emulated user-space devices to inject interrupts to the VGIC instead of directly to CPUs. It is not possible to create both a GICv3 and GICv2 on the same VM. -Creating a guest GICv3 device requires a host GICv3 as well. +Creating a guest GICv3 device requires a host GICv3 host, or a GICv5 host with +support for FEAT_GCIE_LEGACY. Groups: From 4cab5c857d1f92b4b322e30349fdc5e2e38e7a2f Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:45 +0100 Subject: [PATCH 082/330] KVM: arm64: Hide CNTHV_*_EL2 from userspace for nVHE guests Although we correctly UNDEF any CNTHV_*_EL2 access from the guest when E2H==0, we still expose these registers to userspace, which is a bad idea. Drop the ad-hoc UNDEF injection and switch to a .visibility() callback which will also hide the register from userspace. Fixes: 0e45981028550 ("KVM: arm64: timer: Don't adjust the EL2 virtual timer offset") Signed-off-by: Marc Zyngier --- arch/arm64/kvm/sys_regs.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index ee8a7033c85b..9f2f4e0b042e 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -1594,16 +1594,6 @@ static bool access_arch_timer(struct kvm_vcpu *vcpu, return true; } -static bool access_hv_timer(struct kvm_vcpu *vcpu, - struct sys_reg_params *p, - const struct sys_reg_desc *r) -{ - if (!vcpu_el2_e2h_is_set(vcpu)) - return undef_access(vcpu, p, r); - - return access_arch_timer(vcpu, p, r); -} - static s64 kvm_arm64_ftr_safe_value(u32 id, const struct arm64_ftr_bits *ftrp, s64 new, s64 cur) { @@ -2831,6 +2821,16 @@ static unsigned int s1pie_el2_visibility(const struct kvm_vcpu *vcpu, return __el2_visibility(vcpu, rd, s1pie_visibility); } +static unsigned int cnthv_visibility(const struct kvm_vcpu *vcpu, + const struct sys_reg_desc *rd) +{ + if (vcpu_has_nv(vcpu) && + !vcpu_has_feature(vcpu, KVM_ARM_VCPU_HAS_EL2_E2H0)) + return 0; + + return REG_HIDDEN; +} + static bool access_mdcr(struct kvm_vcpu *vcpu, struct sys_reg_params *p, const struct sys_reg_desc *r) @@ -3691,9 +3691,9 @@ static const struct sys_reg_desc sys_reg_descs[] = { EL2_REG(CNTHP_CTL_EL2, access_arch_timer, reset_val, 0), EL2_REG(CNTHP_CVAL_EL2, access_arch_timer, reset_val, 0), - { SYS_DESC(SYS_CNTHV_TVAL_EL2), access_hv_timer }, - EL2_REG(CNTHV_CTL_EL2, access_hv_timer, reset_val, 0), - EL2_REG(CNTHV_CVAL_EL2, access_hv_timer, reset_val, 0), + { SYS_DESC(SYS_CNTHV_TVAL_EL2), access_arch_timer, .visibility = cnthv_visibility }, + EL2_REG_FILTERED(CNTHV_CTL_EL2, access_arch_timer, reset_val, 0, cnthv_visibility), + EL2_REG_FILTERED(CNTHV_CVAL_EL2, access_arch_timer, reset_val, 0, cnthv_visibility), { SYS_DESC(SYS_CNTKCTL_EL12), access_cntkctl_el12 }, From aa68975c973ed3b0bd4ff513113495588afb855c Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:46 +0100 Subject: [PATCH 083/330] KVM: arm64: Introduce timer_context_to_vcpu() helper We currently have a vcpu pointer nested into each timer context. As we are about to remove this pointer, introduce a helper (aptly named timer_context_to_vcpu()) that returns this pointer, at least until we repaint the data structure. Signed-off-by: Marc Zyngier --- arch/arm64/kvm/arch_timer.c | 25 +++++++++++++------------ include/kvm/arm_arch_timer.h | 2 +- 2 files changed, 14 insertions(+), 13 deletions(-) diff --git a/arch/arm64/kvm/arch_timer.c b/arch/arm64/kvm/arch_timer.c index dbd74e4885e2..e5a25e743f5b 100644 --- a/arch/arm64/kvm/arch_timer.c +++ b/arch/arm64/kvm/arch_timer.c @@ -66,7 +66,7 @@ static int nr_timers(struct kvm_vcpu *vcpu) u32 timer_get_ctl(struct arch_timer_context *ctxt) { - struct kvm_vcpu *vcpu = ctxt->vcpu; + struct kvm_vcpu *vcpu = timer_context_to_vcpu(ctxt); switch(arch_timer_ctx_index(ctxt)) { case TIMER_VTIMER: @@ -85,7 +85,7 @@ u32 timer_get_ctl(struct arch_timer_context *ctxt) u64 timer_get_cval(struct arch_timer_context *ctxt) { - struct kvm_vcpu *vcpu = ctxt->vcpu; + struct kvm_vcpu *vcpu = timer_context_to_vcpu(ctxt); switch(arch_timer_ctx_index(ctxt)) { case TIMER_VTIMER: @@ -104,7 +104,7 @@ u64 timer_get_cval(struct arch_timer_context *ctxt) static void timer_set_ctl(struct arch_timer_context *ctxt, u32 ctl) { - struct kvm_vcpu *vcpu = ctxt->vcpu; + struct kvm_vcpu *vcpu = timer_context_to_vcpu(ctxt); switch(arch_timer_ctx_index(ctxt)) { case TIMER_VTIMER: @@ -126,7 +126,7 @@ static void timer_set_ctl(struct arch_timer_context *ctxt, u32 ctl) static void timer_set_cval(struct arch_timer_context *ctxt, u64 cval) { - struct kvm_vcpu *vcpu = ctxt->vcpu; + struct kvm_vcpu *vcpu = timer_context_to_vcpu(ctxt); switch(arch_timer_ctx_index(ctxt)) { case TIMER_VTIMER: @@ -343,7 +343,7 @@ static enum hrtimer_restart kvm_hrtimer_expire(struct hrtimer *hrt) u64 ns; ctx = container_of(hrt, struct arch_timer_context, hrtimer); - vcpu = ctx->vcpu; + vcpu = timer_context_to_vcpu(ctx); trace_kvm_timer_hrtimer_expire(ctx); @@ -436,8 +436,9 @@ static void kvm_timer_update_status(struct arch_timer_context *ctx, bool level) * * But hey, it's fast, right? */ - if (is_hyp_ctxt(ctx->vcpu) && - (ctx == vcpu_vtimer(ctx->vcpu) || ctx == vcpu_ptimer(ctx->vcpu))) { + struct kvm_vcpu *vcpu = timer_context_to_vcpu(ctx); + if (is_hyp_ctxt(vcpu) && + (ctx == vcpu_vtimer(vcpu) || ctx == vcpu_ptimer(vcpu))) { unsigned long val = timer_get_ctl(ctx); __assign_bit(__ffs(ARCH_TIMER_CTRL_IT_STAT), &val, level); timer_set_ctl(ctx, val); @@ -470,7 +471,7 @@ static void timer_emulate(struct arch_timer_context *ctx) trace_kvm_timer_emulate(ctx, should_fire); if (should_fire != ctx->irq.level) - kvm_timer_update_irq(ctx->vcpu, should_fire, ctx); + kvm_timer_update_irq(timer_context_to_vcpu(ctx), should_fire, ctx); kvm_timer_update_status(ctx, should_fire); @@ -498,7 +499,7 @@ static void set_cntpoff(u64 cntpoff) static void timer_save_state(struct arch_timer_context *ctx) { - struct arch_timer_cpu *timer = vcpu_timer(ctx->vcpu); + struct arch_timer_cpu *timer = vcpu_timer(timer_context_to_vcpu(ctx)); enum kvm_arch_timers index = arch_timer_ctx_index(ctx); unsigned long flags; @@ -609,7 +610,7 @@ static void kvm_timer_unblocking(struct kvm_vcpu *vcpu) static void timer_restore_state(struct arch_timer_context *ctx) { - struct arch_timer_cpu *timer = vcpu_timer(ctx->vcpu); + struct arch_timer_cpu *timer = vcpu_timer(timer_context_to_vcpu(ctx)); enum kvm_arch_timers index = arch_timer_ctx_index(ctx); unsigned long flags; @@ -668,7 +669,7 @@ static inline void set_timer_irq_phys_active(struct arch_timer_context *ctx, boo static void kvm_timer_vcpu_load_gic(struct arch_timer_context *ctx) { - struct kvm_vcpu *vcpu = ctx->vcpu; + struct kvm_vcpu *vcpu = timer_context_to_vcpu(ctx); bool phys_active = false; /* @@ -677,7 +678,7 @@ static void kvm_timer_vcpu_load_gic(struct arch_timer_context *ctx) * this point and the register restoration, we'll take the * interrupt anyway. */ - kvm_timer_update_irq(ctx->vcpu, kvm_timer_should_fire(ctx), ctx); + kvm_timer_update_irq(vcpu, kvm_timer_should_fire(ctx), ctx); if (irqchip_in_kernel(vcpu->kvm)) phys_active = kvm_vgic_map_is_active(vcpu, timer_irq(ctx)); diff --git a/include/kvm/arm_arch_timer.h b/include/kvm/arm_arch_timer.h index 681cf0c8b9df..d188c716d03c 100644 --- a/include/kvm/arm_arch_timer.h +++ b/include/kvm/arm_arch_timer.h @@ -128,7 +128,7 @@ void kvm_timer_init_vhe(void); #define vcpu_hptimer(v) (&(v)->arch.timer_cpu.timers[TIMER_HPTIMER]) #define arch_timer_ctx_index(ctx) ((ctx) - vcpu_timer((ctx)->vcpu)->timers) - +#define timer_context_to_vcpu(ctx) ((ctx)->vcpu) #define timer_vm_data(ctx) (&(ctx)->vcpu->kvm->arch.timer_data) #define timer_irq(ctx) (timer_vm_data(ctx)->ppi[arch_timer_ctx_index(ctx)]) From 8625a670afb05f1e1d69d50a74dbcc9d1b855efe Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:47 +0100 Subject: [PATCH 084/330] KVM: arm64: Replace timer context vcpu pointer with timer_id Having to follow a pointer to a vcpu is pretty dumb, when the timers are are a fixed offset in the vcpu structure itself. Trade the vcpu pointer for a timer_id, which can then be used to compute the vcpu address as needed. Reviewed-by: Joey Gouly Signed-off-by: Marc Zyngier --- arch/arm64/kvm/arch_timer.c | 4 ++-- include/kvm/arm_arch_timer.h | 11 ++++++----- 2 files changed, 8 insertions(+), 7 deletions(-) diff --git a/arch/arm64/kvm/arch_timer.c b/arch/arm64/kvm/arch_timer.c index e5a25e743f5b..c832c293676a 100644 --- a/arch/arm64/kvm/arch_timer.c +++ b/arch/arm64/kvm/arch_timer.c @@ -149,7 +149,7 @@ static void timer_set_cval(struct arch_timer_context *ctxt, u64 cval) static void timer_set_offset(struct arch_timer_context *ctxt, u64 offset) { if (!ctxt->offset.vm_offset) { - WARN(offset, "timer %ld\n", arch_timer_ctx_index(ctxt)); + WARN(offset, "timer %d\n", arch_timer_ctx_index(ctxt)); return; } @@ -1064,7 +1064,7 @@ static void timer_context_init(struct kvm_vcpu *vcpu, int timerid) struct arch_timer_context *ctxt = vcpu_get_timer(vcpu, timerid); struct kvm *kvm = vcpu->kvm; - ctxt->vcpu = vcpu; + ctxt->timer_id = timerid; if (timerid == TIMER_VTIMER) ctxt->offset.vm_offset = &kvm->arch.timer_data.voffset; diff --git a/include/kvm/arm_arch_timer.h b/include/kvm/arm_arch_timer.h index d188c716d03c..d8e400cb2bff 100644 --- a/include/kvm/arm_arch_timer.h +++ b/include/kvm/arm_arch_timer.h @@ -51,8 +51,6 @@ struct arch_timer_vm_data { }; struct arch_timer_context { - struct kvm_vcpu *vcpu; - /* Emulated Timer (may be unused) */ struct hrtimer hrtimer; u64 ns_frac; @@ -71,6 +69,9 @@ struct arch_timer_context { bool level; } irq; + /* Who am I? */ + enum kvm_arch_timers timer_id; + /* Duplicated state from arch_timer.c for convenience */ u32 host_timer_irq; }; @@ -127,9 +128,9 @@ void kvm_timer_init_vhe(void); #define vcpu_hvtimer(v) (&(v)->arch.timer_cpu.timers[TIMER_HVTIMER]) #define vcpu_hptimer(v) (&(v)->arch.timer_cpu.timers[TIMER_HPTIMER]) -#define arch_timer_ctx_index(ctx) ((ctx) - vcpu_timer((ctx)->vcpu)->timers) -#define timer_context_to_vcpu(ctx) ((ctx)->vcpu) -#define timer_vm_data(ctx) (&(ctx)->vcpu->kvm->arch.timer_data) +#define arch_timer_ctx_index(ctx) ((ctx)->timer_id) +#define timer_context_to_vcpu(ctx) container_of((ctx), struct kvm_vcpu, arch.timer_cpu.timers[(ctx)->timer_id]) +#define timer_vm_data(ctx) (&(timer_context_to_vcpu(ctx)->kvm->arch.timer_data)) #define timer_irq(ctx) (timer_vm_data(ctx)->ppi[arch_timer_ctx_index(ctx)]) u64 kvm_arm_timer_read_sysreg(struct kvm_vcpu *vcpu, From a92d552266890f83126fdef4f777a985cc1302bd Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:48 +0100 Subject: [PATCH 085/330] KVM: arm64: Make timer_set_offset() generally accessible Move the timer_set_offset() helper to arm_arch_timer.h, so that it is next to timer_get_offset(), and accessible by the rest of KVM. Signed-off-by: Marc Zyngier --- arch/arm64/kvm/arch_timer.c | 10 ---------- include/kvm/arm_arch_timer.h | 10 ++++++++++ 2 files changed, 10 insertions(+), 10 deletions(-) diff --git a/arch/arm64/kvm/arch_timer.c b/arch/arm64/kvm/arch_timer.c index c832c293676a..27662a3a3043 100644 --- a/arch/arm64/kvm/arch_timer.c +++ b/arch/arm64/kvm/arch_timer.c @@ -146,16 +146,6 @@ static void timer_set_cval(struct arch_timer_context *ctxt, u64 cval) } } -static void timer_set_offset(struct arch_timer_context *ctxt, u64 offset) -{ - if (!ctxt->offset.vm_offset) { - WARN(offset, "timer %d\n", arch_timer_ctx_index(ctxt)); - return; - } - - WRITE_ONCE(*ctxt->offset.vm_offset, offset); -} - u64 kvm_phys_timer_read(void) { return timecounter->cc->read(timecounter->cc); diff --git a/include/kvm/arm_arch_timer.h b/include/kvm/arm_arch_timer.h index d8e400cb2bff..5f7f2ed8817c 100644 --- a/include/kvm/arm_arch_timer.h +++ b/include/kvm/arm_arch_timer.h @@ -179,4 +179,14 @@ static inline u64 timer_get_offset(struct arch_timer_context *ctxt) return offset; } +static inline void timer_set_offset(struct arch_timer_context *ctxt, u64 offset) +{ + if (!ctxt->offset.vm_offset) { + WARN(offset, "timer %d\n", arch_timer_ctx_index(ctxt)); + return; + } + + WRITE_ONCE(*ctxt->offset.vm_offset, offset); +} + #endif From 77a0c42eaf03c66936429d190bb2ea1a214bd528 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:49 +0100 Subject: [PATCH 086/330] KVM: arm64: Add timer UAPI workaround to sysreg infrastructure Amongst the numerous bugs that plague the KVM/arm64 UAPI, one of the most annoying thing is that the userspace view of the virtual timer has its CVAL and CNT encodings swapped. In order to reduce the amount of code that has to know about this, start by adding handling for this bug in the sys_reg code. Nothing is making use of it yet, as the code responsible for userspace interaction is catching the accesses early. Signed-off-by: Marc Zyngier --- arch/arm64/kvm/sys_regs.c | 33 ++++++++++++++++++++++++++++++--- arch/arm64/kvm/sys_regs.h | 6 ++++++ 2 files changed, 36 insertions(+), 3 deletions(-) diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index 9f2f4e0b042e..8e6f50f54b4b 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -5231,15 +5231,28 @@ static int demux_c15_set(struct kvm_vcpu *vcpu, u64 id, void __user *uaddr) } } +static u64 kvm_one_reg_to_id(const struct kvm_one_reg *reg) +{ + switch(reg->id) { + case KVM_REG_ARM_TIMER_CVAL: + return TO_ARM64_SYS_REG(CNTV_CVAL_EL0); + case KVM_REG_ARM_TIMER_CNT: + return TO_ARM64_SYS_REG(CNTVCT_EL0); + default: + return reg->id; + } +} + int kvm_sys_reg_get_user(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg, const struct sys_reg_desc table[], unsigned int num) { u64 __user *uaddr = (u64 __user *)(unsigned long)reg->addr; const struct sys_reg_desc *r; + u64 id = kvm_one_reg_to_id(reg); u64 val; int ret; - r = id_to_sys_reg_desc(vcpu, reg->id, table, num); + r = id_to_sys_reg_desc(vcpu, id, table, num); if (!r || sysreg_hidden(vcpu, r)) return -ENOENT; @@ -5272,13 +5285,14 @@ int kvm_sys_reg_set_user(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg, { u64 __user *uaddr = (u64 __user *)(unsigned long)reg->addr; const struct sys_reg_desc *r; + u64 id = kvm_one_reg_to_id(reg); u64 val; int ret; if (get_user(val, uaddr)) return -EFAULT; - r = id_to_sys_reg_desc(vcpu, reg->id, table, num); + r = id_to_sys_reg_desc(vcpu, id, table, num); if (!r || sysreg_hidden(vcpu, r)) return -ENOENT; @@ -5338,10 +5352,23 @@ static u64 sys_reg_to_index(const struct sys_reg_desc *reg) static bool copy_reg_to_user(const struct sys_reg_desc *reg, u64 __user **uind) { + u64 idx; + if (!*uind) return true; - if (put_user(sys_reg_to_index(reg), *uind)) + switch (reg_to_encoding(reg)) { + case SYS_CNTV_CVAL_EL0: + idx = KVM_REG_ARM_TIMER_CVAL; + break; + case SYS_CNTVCT_EL0: + idx = KVM_REG_ARM_TIMER_CNT; + break; + default: + idx = sys_reg_to_index(reg); + } + + if (put_user(idx, *uind)) return false; (*uind)++; diff --git a/arch/arm64/kvm/sys_regs.h b/arch/arm64/kvm/sys_regs.h index 317abc490368..b3f904472fac 100644 --- a/arch/arm64/kvm/sys_regs.h +++ b/arch/arm64/kvm/sys_regs.h @@ -257,4 +257,10 @@ int kvm_finalize_sys_regs(struct kvm_vcpu *vcpu); (val); \ }) +#define TO_ARM64_SYS_REG(r) ARM64_SYS_REG(sys_reg_Op0(SYS_ ## r), \ + sys_reg_Op1(SYS_ ## r), \ + sys_reg_CRn(SYS_ ## r), \ + sys_reg_CRm(SYS_ ## r), \ + sys_reg_Op2(SYS_ ## r)) + #endif /* __ARM64_KVM_SYS_REGS_LOCAL_H__ */ From 09424d5d7d4e8b427ee4a737fb7765103789e08a Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:50 +0100 Subject: [PATCH 087/330] KVM: arm64: Move CNT*_CTL_EL0 userspace accessors to generic infrastructure Remove the handling of CNT*_CTL_EL0 from guest.c, and move it to sys_regs.c, using a new TIMER_REG() definition to encapsulate it. Signed-off-by: Marc Zyngier --- arch/arm64/kvm/guest.c | 4 ---- arch/arm64/kvm/sys_regs.c | 36 +++++++++++++++++++++++++++++++----- 2 files changed, 31 insertions(+), 9 deletions(-) diff --git a/arch/arm64/kvm/guest.c b/arch/arm64/kvm/guest.c index 16ba5e9ac86c..dea648706fd5 100644 --- a/arch/arm64/kvm/guest.c +++ b/arch/arm64/kvm/guest.c @@ -592,10 +592,8 @@ static unsigned long num_core_regs(const struct kvm_vcpu *vcpu) } static const u64 timer_reg_list[] = { - KVM_REG_ARM_TIMER_CTL, KVM_REG_ARM_TIMER_CNT, KVM_REG_ARM_TIMER_CVAL, - KVM_REG_ARM_PTIMER_CTL, KVM_REG_ARM_PTIMER_CNT, KVM_REG_ARM_PTIMER_CVAL, }; @@ -605,10 +603,8 @@ static const u64 timer_reg_list[] = { static bool is_timer_reg(u64 index) { switch (index) { - case KVM_REG_ARM_TIMER_CTL: case KVM_REG_ARM_TIMER_CNT: case KVM_REG_ARM_TIMER_CVAL: - case KVM_REG_ARM_PTIMER_CTL: case KVM_REG_ARM_PTIMER_CNT: case KVM_REG_ARM_PTIMER_CVAL: return true; diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index 8e6f50f54b4b..d97aacf4c1dc 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -1594,6 +1594,23 @@ static bool access_arch_timer(struct kvm_vcpu *vcpu, return true; } +static int arch_timer_set_user(struct kvm_vcpu *vcpu, + const struct sys_reg_desc *rd, + u64 val) +{ + switch (reg_to_encoding(rd)) { + case SYS_CNTV_CTL_EL0: + case SYS_CNTP_CTL_EL0: + case SYS_CNTHV_CTL_EL2: + case SYS_CNTHP_CTL_EL2: + val &= ~ARCH_TIMER_CTRL_IT_STAT; + break; + } + + __vcpu_assign_sys_reg(vcpu, rd->reg, val); + return 0; +} + static s64 kvm_arm64_ftr_safe_value(u32 id, const struct arm64_ftr_bits *ftrp, s64 new, s64 cur) { @@ -2496,15 +2513,20 @@ static bool bad_redir_trap(struct kvm_vcpu *vcpu, "trap of EL2 register redirected to EL1"); } -#define EL2_REG_FILTERED(name, acc, rst, v, filter) { \ +#define SYS_REG_USER_FILTER(name, acc, rst, v, gu, su, filter) { \ SYS_DESC(SYS_##name), \ .access = acc, \ .reset = rst, \ .reg = name, \ + .get_user = gu, \ + .set_user = su, \ .visibility = filter, \ .val = v, \ } +#define EL2_REG_FILTERED(name, acc, rst, v, filter) \ + SYS_REG_USER_FILTER(name, acc, rst, v, NULL, NULL, filter) + #define EL2_REG(name, acc, rst, v) \ EL2_REG_FILTERED(name, acc, rst, v, el2_visibility) @@ -2515,6 +2537,10 @@ static bool bad_redir_trap(struct kvm_vcpu *vcpu, EL2_REG_VNCR_FILT(name, hidden_visibility) #define EL2_REG_REDIR(name, rst, v) EL2_REG(name, bad_redir_trap, rst, v) +#define TIMER_REG(name, vis) \ + SYS_REG_USER_FILTER(name, access_arch_timer, reset_val, 0, \ + NULL, arch_timer_set_user, vis) + /* * Since reset() callback and field val are not used for idregs, they will be * used for specific purposes for idregs. @@ -3485,11 +3511,11 @@ static const struct sys_reg_desc sys_reg_descs[] = { { SYS_DESC(SYS_CNTPCTSS_EL0), access_arch_timer }, { SYS_DESC(SYS_CNTVCTSS_EL0), access_arch_timer }, { SYS_DESC(SYS_CNTP_TVAL_EL0), access_arch_timer }, - { SYS_DESC(SYS_CNTP_CTL_EL0), access_arch_timer }, + TIMER_REG(CNTP_CTL_EL0, NULL), { SYS_DESC(SYS_CNTP_CVAL_EL0), access_arch_timer }, { SYS_DESC(SYS_CNTV_TVAL_EL0), access_arch_timer }, - { SYS_DESC(SYS_CNTV_CTL_EL0), access_arch_timer }, + TIMER_REG(CNTV_CTL_EL0, NULL), { SYS_DESC(SYS_CNTV_CVAL_EL0), access_arch_timer }, /* PMEVCNTRn_EL0 */ @@ -3688,11 +3714,11 @@ static const struct sys_reg_desc sys_reg_descs[] = { EL2_REG_VNCR(CNTVOFF_EL2, reset_val, 0), EL2_REG(CNTHCTL_EL2, access_rw, reset_val, 0), { SYS_DESC(SYS_CNTHP_TVAL_EL2), access_arch_timer }, - EL2_REG(CNTHP_CTL_EL2, access_arch_timer, reset_val, 0), + TIMER_REG(CNTHP_CTL_EL2, el2_visibility), EL2_REG(CNTHP_CVAL_EL2, access_arch_timer, reset_val, 0), { SYS_DESC(SYS_CNTHV_TVAL_EL2), access_arch_timer, .visibility = cnthv_visibility }, - EL2_REG_FILTERED(CNTHV_CTL_EL2, access_arch_timer, reset_val, 0, cnthv_visibility), + TIMER_REG(CNTHV_CTL_EL2, cnthv_visibility), EL2_REG_FILTERED(CNTHV_CVAL_EL2, access_arch_timer, reset_val, 0, cnthv_visibility), { SYS_DESC(SYS_CNTKCTL_EL12), access_cntkctl_el12 }, From 8af198980eff2ed2a5df3d2ee39f8c9d61f40559 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:51 +0100 Subject: [PATCH 088/330] KVM: arm64: Move CNT*_CVAL_EL0 userspace accessors to generic infrastructure As for the control registers, move the comparator registers to the common infrastructure. Signed-off-by: Marc Zyngier --- arch/arm64/kvm/guest.c | 4 ---- arch/arm64/kvm/sys_regs.c | 8 ++++---- 2 files changed, 4 insertions(+), 8 deletions(-) diff --git a/arch/arm64/kvm/guest.c b/arch/arm64/kvm/guest.c index dea648706fd5..c23ec9be4ce2 100644 --- a/arch/arm64/kvm/guest.c +++ b/arch/arm64/kvm/guest.c @@ -593,9 +593,7 @@ static unsigned long num_core_regs(const struct kvm_vcpu *vcpu) static const u64 timer_reg_list[] = { KVM_REG_ARM_TIMER_CNT, - KVM_REG_ARM_TIMER_CVAL, KVM_REG_ARM_PTIMER_CNT, - KVM_REG_ARM_PTIMER_CVAL, }; #define NUM_TIMER_REGS ARRAY_SIZE(timer_reg_list) @@ -604,9 +602,7 @@ static bool is_timer_reg(u64 index) { switch (index) { case KVM_REG_ARM_TIMER_CNT: - case KVM_REG_ARM_TIMER_CVAL: case KVM_REG_ARM_PTIMER_CNT: - case KVM_REG_ARM_PTIMER_CVAL: return true; } return false; diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index d97aacf4c1dc..68e88d5c0dfb 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -3512,11 +3512,11 @@ static const struct sys_reg_desc sys_reg_descs[] = { { SYS_DESC(SYS_CNTVCTSS_EL0), access_arch_timer }, { SYS_DESC(SYS_CNTP_TVAL_EL0), access_arch_timer }, TIMER_REG(CNTP_CTL_EL0, NULL), - { SYS_DESC(SYS_CNTP_CVAL_EL0), access_arch_timer }, + TIMER_REG(CNTP_CVAL_EL0, NULL), { SYS_DESC(SYS_CNTV_TVAL_EL0), access_arch_timer }, TIMER_REG(CNTV_CTL_EL0, NULL), - { SYS_DESC(SYS_CNTV_CVAL_EL0), access_arch_timer }, + TIMER_REG(CNTV_CVAL_EL0, NULL), /* PMEVCNTRn_EL0 */ PMU_PMEVCNTR_EL0(0), @@ -3715,11 +3715,11 @@ static const struct sys_reg_desc sys_reg_descs[] = { EL2_REG(CNTHCTL_EL2, access_rw, reset_val, 0), { SYS_DESC(SYS_CNTHP_TVAL_EL2), access_arch_timer }, TIMER_REG(CNTHP_CTL_EL2, el2_visibility), - EL2_REG(CNTHP_CVAL_EL2, access_arch_timer, reset_val, 0), + TIMER_REG(CNTHP_CVAL_EL2, el2_visibility), { SYS_DESC(SYS_CNTHV_TVAL_EL2), access_arch_timer, .visibility = cnthv_visibility }, TIMER_REG(CNTHV_CTL_EL2, cnthv_visibility), - EL2_REG_FILTERED(CNTHV_CVAL_EL2, access_arch_timer, reset_val, 0, cnthv_visibility), + TIMER_REG(CNTHV_CVAL_EL2, cnthv_visibility), { SYS_DESC(SYS_CNTKCTL_EL12), access_cntkctl_el12 }, From c3be3a48fb18f9d243fac452e0be41469bb246b4 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:52 +0100 Subject: [PATCH 089/330] KVM: arm64: Move CNT*CT_EL0 userspace accessors to generic infrastructure Moving the counter registers is a bit more involved than for the control and comparator (there is no shadow data for the counter), but still pretty manageable. Reviewed-by: Joey Gouly Signed-off-by: Marc Zyngier --- arch/arm64/kvm/guest.c | 7 ------- arch/arm64/kvm/sys_regs.c | 34 +++++++++++++++++++++++++++++++--- 2 files changed, 31 insertions(+), 10 deletions(-) diff --git a/arch/arm64/kvm/guest.c b/arch/arm64/kvm/guest.c index c23ec9be4ce2..138e5e2dc10c 100644 --- a/arch/arm64/kvm/guest.c +++ b/arch/arm64/kvm/guest.c @@ -592,19 +592,12 @@ static unsigned long num_core_regs(const struct kvm_vcpu *vcpu) } static const u64 timer_reg_list[] = { - KVM_REG_ARM_TIMER_CNT, - KVM_REG_ARM_PTIMER_CNT, }; #define NUM_TIMER_REGS ARRAY_SIZE(timer_reg_list) static bool is_timer_reg(u64 index) { - switch (index) { - case KVM_REG_ARM_TIMER_CNT: - case KVM_REG_ARM_PTIMER_CNT: - return true; - } return false; } diff --git a/arch/arm64/kvm/sys_regs.c b/arch/arm64/kvm/sys_regs.c index 68e88d5c0dfb..e67eb39ddc11 100644 --- a/arch/arm64/kvm/sys_regs.c +++ b/arch/arm64/kvm/sys_regs.c @@ -1605,12 +1605,38 @@ static int arch_timer_set_user(struct kvm_vcpu *vcpu, case SYS_CNTHP_CTL_EL2: val &= ~ARCH_TIMER_CTRL_IT_STAT; break; + case SYS_CNTVCT_EL0: + if (!test_bit(KVM_ARCH_FLAG_VM_COUNTER_OFFSET, &vcpu->kvm->arch.flags)) + timer_set_offset(vcpu_vtimer(vcpu), kvm_phys_timer_read() - val); + return 0; + case SYS_CNTPCT_EL0: + if (!test_bit(KVM_ARCH_FLAG_VM_COUNTER_OFFSET, &vcpu->kvm->arch.flags)) + timer_set_offset(vcpu_ptimer(vcpu), kvm_phys_timer_read() - val); + return 0; } __vcpu_assign_sys_reg(vcpu, rd->reg, val); return 0; } +static int arch_timer_get_user(struct kvm_vcpu *vcpu, + const struct sys_reg_desc *rd, + u64 *val) +{ + switch (reg_to_encoding(rd)) { + case SYS_CNTVCT_EL0: + *val = kvm_phys_timer_read() - timer_get_offset(vcpu_vtimer(vcpu)); + break; + case SYS_CNTPCT_EL0: + *val = kvm_phys_timer_read() - timer_get_offset(vcpu_ptimer(vcpu)); + break; + default: + *val = __vcpu_sys_reg(vcpu, rd->reg); + } + + return 0; +} + static s64 kvm_arm64_ftr_safe_value(u32 id, const struct arm64_ftr_bits *ftrp, s64 new, s64 cur) { @@ -2539,7 +2565,7 @@ static bool bad_redir_trap(struct kvm_vcpu *vcpu, #define TIMER_REG(name, vis) \ SYS_REG_USER_FILTER(name, access_arch_timer, reset_val, 0, \ - NULL, arch_timer_set_user, vis) + arch_timer_get_user, arch_timer_set_user, vis) /* * Since reset() callback and field val are not used for idregs, they will be @@ -3506,8 +3532,10 @@ static const struct sys_reg_desc sys_reg_descs[] = { AMU_AMEVTYPER1_EL0(14), AMU_AMEVTYPER1_EL0(15), - { SYS_DESC(SYS_CNTPCT_EL0), access_arch_timer }, - { SYS_DESC(SYS_CNTVCT_EL0), access_arch_timer }, + { SYS_DESC(SYS_CNTPCT_EL0), .access = access_arch_timer, + .get_user = arch_timer_get_user, .set_user = arch_timer_set_user }, + { SYS_DESC(SYS_CNTVCT_EL0), .access = access_arch_timer, + .get_user = arch_timer_get_user, .set_user = arch_timer_set_user }, { SYS_DESC(SYS_CNTPCTSS_EL0), access_arch_timer }, { SYS_DESC(SYS_CNTVCTSS_EL0), access_arch_timer }, { SYS_DESC(SYS_CNTP_TVAL_EL0), access_arch_timer }, From 892f7c38ba3b7de19b3dffb8e148d5fbf1228f20 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:53 +0100 Subject: [PATCH 090/330] KVM: arm64: Fix WFxT handling of nested virt The spec for WFxT indicates that the parameter to the WFxT instruction is relative to the reading of CNTVCT_EL0. This means that the implementation needs to take the execution context into account, as CNTVOFF_EL2 does not always affect readings of CNTVCT_EL0 (such as when HCR_EL2.E2H is 1 and that we're in host context). This also rids us of the last instance of KVM_REG_ARM_TIMER_CNT outside of the userspace interaction code. Signed-off-by: Marc Zyngier --- arch/arm64/kvm/handle_exit.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/arch/arm64/kvm/handle_exit.c b/arch/arm64/kvm/handle_exit.c index bca8c80e11da..cc7d5d1709cb 100644 --- a/arch/arm64/kvm/handle_exit.c +++ b/arch/arm64/kvm/handle_exit.c @@ -147,7 +147,12 @@ static int kvm_handle_wfx(struct kvm_vcpu *vcpu) if (esr & ESR_ELx_WFx_ISS_RV) { u64 val, now; - now = kvm_arm_timer_get_reg(vcpu, KVM_REG_ARM_TIMER_CNT); + now = kvm_phys_timer_read(); + if (is_hyp_ctxt(vcpu) && vcpu_el2_e2h_is_set(vcpu)) + now -= timer_get_offset(vcpu_hvtimer(vcpu)); + else + now -= timer_get_offset(vcpu_vtimer(vcpu)); + val = vcpu_get_reg(vcpu, kvm_vcpu_sys_get_rt(vcpu)); if (now >= val) From 386aac77da112651a5cdadc4a6b29181592f5aa0 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:54 +0100 Subject: [PATCH 091/330] KVM: arm64: Kill leftovers of ad-hoc timer userspace access Now that the whole timer infrastructure is handled as system register accesses, get rid of the now unused ad-hoc infrastructure. Signed-off-by: Marc Zyngier --- arch/arm64/kvm/arch_timer.c | 68 ------------------------------------ arch/arm64/kvm/guest.c | 55 ----------------------------- include/kvm/arm_arch_timer.h | 3 -- 3 files changed, 126 deletions(-) diff --git a/arch/arm64/kvm/arch_timer.c b/arch/arm64/kvm/arch_timer.c index 27662a3a3043..3f675875abea 100644 --- a/arch/arm64/kvm/arch_timer.c +++ b/arch/arm64/kvm/arch_timer.c @@ -1112,49 +1112,6 @@ void kvm_timer_cpu_down(void) disable_percpu_irq(host_ptimer_irq); } -int kvm_arm_timer_set_reg(struct kvm_vcpu *vcpu, u64 regid, u64 value) -{ - struct arch_timer_context *timer; - - switch (regid) { - case KVM_REG_ARM_TIMER_CTL: - timer = vcpu_vtimer(vcpu); - kvm_arm_timer_write(vcpu, timer, TIMER_REG_CTL, value); - break; - case KVM_REG_ARM_TIMER_CNT: - if (!test_bit(KVM_ARCH_FLAG_VM_COUNTER_OFFSET, - &vcpu->kvm->arch.flags)) { - timer = vcpu_vtimer(vcpu); - timer_set_offset(timer, kvm_phys_timer_read() - value); - } - break; - case KVM_REG_ARM_TIMER_CVAL: - timer = vcpu_vtimer(vcpu); - kvm_arm_timer_write(vcpu, timer, TIMER_REG_CVAL, value); - break; - case KVM_REG_ARM_PTIMER_CTL: - timer = vcpu_ptimer(vcpu); - kvm_arm_timer_write(vcpu, timer, TIMER_REG_CTL, value); - break; - case KVM_REG_ARM_PTIMER_CNT: - if (!test_bit(KVM_ARCH_FLAG_VM_COUNTER_OFFSET, - &vcpu->kvm->arch.flags)) { - timer = vcpu_ptimer(vcpu); - timer_set_offset(timer, kvm_phys_timer_read() - value); - } - break; - case KVM_REG_ARM_PTIMER_CVAL: - timer = vcpu_ptimer(vcpu); - kvm_arm_timer_write(vcpu, timer, TIMER_REG_CVAL, value); - break; - - default: - return -1; - } - - return 0; -} - static u64 read_timer_ctl(struct arch_timer_context *timer) { /* @@ -1171,31 +1128,6 @@ static u64 read_timer_ctl(struct arch_timer_context *timer) return ctl; } -u64 kvm_arm_timer_get_reg(struct kvm_vcpu *vcpu, u64 regid) -{ - switch (regid) { - case KVM_REG_ARM_TIMER_CTL: - return kvm_arm_timer_read(vcpu, - vcpu_vtimer(vcpu), TIMER_REG_CTL); - case KVM_REG_ARM_TIMER_CNT: - return kvm_arm_timer_read(vcpu, - vcpu_vtimer(vcpu), TIMER_REG_CNT); - case KVM_REG_ARM_TIMER_CVAL: - return kvm_arm_timer_read(vcpu, - vcpu_vtimer(vcpu), TIMER_REG_CVAL); - case KVM_REG_ARM_PTIMER_CTL: - return kvm_arm_timer_read(vcpu, - vcpu_ptimer(vcpu), TIMER_REG_CTL); - case KVM_REG_ARM_PTIMER_CNT: - return kvm_arm_timer_read(vcpu, - vcpu_ptimer(vcpu), TIMER_REG_CNT); - case KVM_REG_ARM_PTIMER_CVAL: - return kvm_arm_timer_read(vcpu, - vcpu_ptimer(vcpu), TIMER_REG_CVAL); - } - return (u64)-1; -} - static u64 kvm_arm_timer_read(struct kvm_vcpu *vcpu, struct arch_timer_context *timer, enum kvm_arch_timer_regs treg) diff --git a/arch/arm64/kvm/guest.c b/arch/arm64/kvm/guest.c index 138e5e2dc10c..1c87699fd886 100644 --- a/arch/arm64/kvm/guest.c +++ b/arch/arm64/kvm/guest.c @@ -591,49 +591,6 @@ static unsigned long num_core_regs(const struct kvm_vcpu *vcpu) return copy_core_reg_indices(vcpu, NULL); } -static const u64 timer_reg_list[] = { -}; - -#define NUM_TIMER_REGS ARRAY_SIZE(timer_reg_list) - -static bool is_timer_reg(u64 index) -{ - return false; -} - -static int copy_timer_indices(struct kvm_vcpu *vcpu, u64 __user *uindices) -{ - for (int i = 0; i < NUM_TIMER_REGS; i++) { - if (put_user(timer_reg_list[i], uindices)) - return -EFAULT; - uindices++; - } - - return 0; -} - -static int set_timer_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg) -{ - void __user *uaddr = (void __user *)(long)reg->addr; - u64 val; - int ret; - - ret = copy_from_user(&val, uaddr, KVM_REG_SIZE(reg->id)); - if (ret != 0) - return -EFAULT; - - return kvm_arm_timer_set_reg(vcpu, reg->id, val); -} - -static int get_timer_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg) -{ - void __user *uaddr = (void __user *)(long)reg->addr; - u64 val; - - val = kvm_arm_timer_get_reg(vcpu, reg->id); - return copy_to_user(uaddr, &val, KVM_REG_SIZE(reg->id)) ? -EFAULT : 0; -} - static unsigned long num_sve_regs(const struct kvm_vcpu *vcpu) { const unsigned int slices = vcpu_sve_slices(vcpu); @@ -709,7 +666,6 @@ unsigned long kvm_arm_num_regs(struct kvm_vcpu *vcpu) res += num_sve_regs(vcpu); res += kvm_arm_num_sys_reg_descs(vcpu); res += kvm_arm_get_fw_num_regs(vcpu); - res += NUM_TIMER_REGS; return res; } @@ -740,11 +696,6 @@ int kvm_arm_copy_reg_indices(struct kvm_vcpu *vcpu, u64 __user *uindices) return ret; uindices += kvm_arm_get_fw_num_regs(vcpu); - ret = copy_timer_indices(vcpu, uindices); - if (ret < 0) - return ret; - uindices += NUM_TIMER_REGS; - return kvm_arm_copy_sys_reg_indices(vcpu, uindices); } @@ -762,9 +713,6 @@ int kvm_arm_get_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg) case KVM_REG_ARM64_SVE: return get_sve_reg(vcpu, reg); } - if (is_timer_reg(reg->id)) - return get_timer_reg(vcpu, reg); - return kvm_arm_sys_reg_get_reg(vcpu, reg); } @@ -782,9 +730,6 @@ int kvm_arm_set_reg(struct kvm_vcpu *vcpu, const struct kvm_one_reg *reg) case KVM_REG_ARM64_SVE: return set_sve_reg(vcpu, reg); } - if (is_timer_reg(reg->id)) - return set_timer_reg(vcpu, reg); - return kvm_arm_sys_reg_set_reg(vcpu, reg); } diff --git a/include/kvm/arm_arch_timer.h b/include/kvm/arm_arch_timer.h index 5f7f2ed8817c..7310841f4512 100644 --- a/include/kvm/arm_arch_timer.h +++ b/include/kvm/arm_arch_timer.h @@ -107,9 +107,6 @@ void kvm_timer_vcpu_terminate(struct kvm_vcpu *vcpu); void kvm_timer_init_vm(struct kvm *kvm); -u64 kvm_arm_timer_get_reg(struct kvm_vcpu *, u64 regid); -int kvm_arm_timer_set_reg(struct kvm_vcpu *, u64 regid, u64 value); - int kvm_arm_timer_set_attr(struct kvm_vcpu *vcpu, struct kvm_device_attr *attr); int kvm_arm_timer_get_attr(struct kvm_vcpu *vcpu, struct kvm_device_attr *attr); int kvm_arm_timer_has_attr(struct kvm_vcpu *vcpu, struct kvm_device_attr *attr); From 6418330c8478735f625398bc4e96d3ac6ce1e055 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:55 +0100 Subject: [PATCH 092/330] KVM: arm64: selftests: Make dependencies on VHE-specific registers explicit The hyp virtual timer registers only exist when VHE is present, Similarly, VNCR_EL2 only exists when NV2 is present. Make these dependencies explicit. Signed-off-by: Marc Zyngier --- tools/testing/selftests/kvm/arm64/get-reg-list.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/tools/testing/selftests/kvm/arm64/get-reg-list.c b/tools/testing/selftests/kvm/arm64/get-reg-list.c index 011fad95dd02..0a4cfb368512 100644 --- a/tools/testing/selftests/kvm/arm64/get-reg-list.c +++ b/tools/testing/selftests/kvm/arm64/get-reg-list.c @@ -65,6 +65,9 @@ static struct feature_id_reg feat_id_regs[] = { REG_FEAT(SCTLR2_EL1, ID_AA64MMFR3_EL1, SCTLRX, IMP), REG_FEAT(VDISR_EL2, ID_AA64PFR0_EL1, RAS, IMP), REG_FEAT(VSESR_EL2, ID_AA64PFR0_EL1, RAS, IMP), + REG_FEAT(VNCR_EL2, ID_AA64MMFR4_EL1, NV_frac, NV2_ONLY), + REG_FEAT(CNTHV_CTL_EL2, ID_AA64MMFR1_EL1, VH, IMP), + REG_FEAT(CNTHV_CVAL_EL2,ID_AA64MMFR1_EL1, VH, IMP), }; bool filter_reg(__u64 reg) From 4da5a9af78b74fb771a4d25dc794296d10e170b1 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:56 +0100 Subject: [PATCH 093/330] KVM: arm64: selftests: Add an E2H=0-specific configuration to get_reg_list Add yet another configuration, this time dealing E2H=0. Signed-off-by: Marc Zyngier --- .../selftests/kvm/arm64/get-reg-list.c | 79 +++++++++++++++++++ 1 file changed, 79 insertions(+) diff --git a/tools/testing/selftests/kvm/arm64/get-reg-list.c b/tools/testing/selftests/kvm/arm64/get-reg-list.c index 0a4cfb368512..7a238755f072 100644 --- a/tools/testing/selftests/kvm/arm64/get-reg-list.c +++ b/tools/testing/selftests/kvm/arm64/get-reg-list.c @@ -758,6 +758,10 @@ static __u64 el2_regs[] = { SYS_REG(VSESR_EL2), }; +static __u64 el2_e2h0_regs[] = { + /* Empty */ +}; + #define BASE_SUBLIST \ { "base", .regs = base_regs, .regs_n = ARRAY_SIZE(base_regs), } #define VREGS_SUBLIST \ @@ -792,6 +796,15 @@ static __u64 el2_regs[] = { .regs = el2_regs, \ .regs_n = ARRAY_SIZE(el2_regs), \ } +#define EL2_E2H0_SUBLIST \ + EL2_SUBLIST, \ + { \ + .name = "EL2 E2H0", \ + .capability = KVM_CAP_ARM_EL2_E2H0, \ + .feature = KVM_ARM_VCPU_HAS_EL2_E2H0, \ + .regs = el2_e2h0_regs, \ + .regs_n = ARRAY_SIZE(el2_e2h0_regs), \ + } static struct vcpu_reg_list vregs_config = { .sublists = { @@ -900,6 +913,65 @@ static struct vcpu_reg_list el2_pauth_pmu_config = { }, }; +static struct vcpu_reg_list el2_e2h0_vregs_config = { + .sublists = { + BASE_SUBLIST, + EL2_E2H0_SUBLIST, + VREGS_SUBLIST, + {0}, + }, +}; + +static struct vcpu_reg_list el2_e2h0_vregs_pmu_config = { + .sublists = { + BASE_SUBLIST, + EL2_E2H0_SUBLIST, + VREGS_SUBLIST, + PMU_SUBLIST, + {0}, + }, +}; + +static struct vcpu_reg_list el2_e2h0_sve_config = { + .sublists = { + BASE_SUBLIST, + EL2_E2H0_SUBLIST, + SVE_SUBLIST, + {0}, + }, +}; + +static struct vcpu_reg_list el2_e2h0_sve_pmu_config = { + .sublists = { + BASE_SUBLIST, + EL2_E2H0_SUBLIST, + SVE_SUBLIST, + PMU_SUBLIST, + {0}, + }, +}; + +static struct vcpu_reg_list el2_e2h0_pauth_config = { + .sublists = { + BASE_SUBLIST, + EL2_E2H0_SUBLIST, + VREGS_SUBLIST, + PAUTH_SUBLIST, + {0}, + }, +}; + +static struct vcpu_reg_list el2_e2h0_pauth_pmu_config = { + .sublists = { + BASE_SUBLIST, + EL2_E2H0_SUBLIST, + VREGS_SUBLIST, + PAUTH_SUBLIST, + PMU_SUBLIST, + {0}, + }, +}; + struct vcpu_reg_list *vcpu_configs[] = { &vregs_config, &vregs_pmu_config, @@ -914,5 +986,12 @@ struct vcpu_reg_list *vcpu_configs[] = { &el2_sve_pmu_config, &el2_pauth_config, &el2_pauth_pmu_config, + + &el2_e2h0_vregs_config, + &el2_e2h0_vregs_pmu_config, + &el2_e2h0_sve_config, + &el2_e2h0_sve_pmu_config, + &el2_e2h0_pauth_config, + &el2_e2h0_pauth_pmu_config, }; int vcpu_configs_n = ARRAY_SIZE(vcpu_configs); From 5c7cf1e44e94a5408b1b5277810502b0f82b77fe Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Mon, 29 Sep 2025 17:04:57 +0100 Subject: [PATCH 094/330] KVM: arm64: selftests: Fix misleading comment about virtual timer encoding The userspace-visible encoding for CNTV_CVAL_EL0 and CNTVCNT_EL0 have been swapped for as long as usersapce has had access to the registers. This is documented in arch/arm64/include/uapi/asm/kvm.h. Despite that, the get_reg_list test has unhelpful comments indicating the wrong register for the encoding. Replace this with definitions exposed in the include file, and a comment explaining again the brokenness. Signed-off-by: Marc Zyngier --- .../testing/selftests/kvm/arm64/get-reg-list.c | 17 ++++++++++++++--- 1 file changed, 14 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/kvm/arm64/get-reg-list.c b/tools/testing/selftests/kvm/arm64/get-reg-list.c index 7a238755f072..c9b84eeaab6b 100644 --- a/tools/testing/selftests/kvm/arm64/get-reg-list.c +++ b/tools/testing/selftests/kvm/arm64/get-reg-list.c @@ -348,9 +348,20 @@ static __u64 base_regs[] = { KVM_REG_ARM_FW_FEAT_BMAP_REG(1), /* KVM_REG_ARM_STD_HYP_BMAP */ KVM_REG_ARM_FW_FEAT_BMAP_REG(2), /* KVM_REG_ARM_VENDOR_HYP_BMAP */ KVM_REG_ARM_FW_FEAT_BMAP_REG(3), /* KVM_REG_ARM_VENDOR_HYP_BMAP_2 */ - ARM64_SYS_REG(3, 3, 14, 3, 1), /* CNTV_CTL_EL0 */ - ARM64_SYS_REG(3, 3, 14, 3, 2), /* CNTV_CVAL_EL0 */ - ARM64_SYS_REG(3, 3, 14, 0, 2), + + /* + * EL0 Virtual Timer Registers + * + * WARNING: + * KVM_REG_ARM_TIMER_CVAL and KVM_REG_ARM_TIMER_CNT are not defined + * with the appropriate register encodings. Their values have been + * accidentally swapped. As this is set API, the definitions here + * must be used, rather than ones derived from the encodings. + */ + KVM_ARM64_SYS_REG(SYS_CNTV_CTL_EL0), + KVM_REG_ARM_TIMER_CVAL, + KVM_REG_ARM_TIMER_CNT, + ARM64_SYS_REG(3, 0, 0, 0, 0), /* MIDR_EL1 */ ARM64_SYS_REG(3, 0, 0, 0, 6), /* REVIDR_EL1 */ ARM64_SYS_REG(3, 1, 0, 0, 1), /* CLIDR_EL1 */ From fb10ddf35c1cc3b2888a944c0a3b1aa3baea585e Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Wed, 24 Sep 2025 16:51:49 -0700 Subject: [PATCH 095/330] KVM: arm64: Compute per-vCPU FGTs at vcpu_load() To date KVM has used the fine-grained traps for the sake of UNDEF enforcement (so-called FGUs), meaning the constituent parts could be computed on a per-VM basis and folded into the effective value when programmed. Prepare for traps changing based on the vCPU context by computing the whole mess of them at vcpu_load(). Aggressively inline all the helpers to preserve the build-time checks that were there before. Signed-off-by: Oliver Upton Reviewed-by: Joey Gouly Signed-off-by: Marc Zyngier --- arch/arm64/include/asm/kvm_host.h | 50 ++++++++ arch/arm64/kvm/arm.c | 1 + arch/arm64/kvm/config.c | 82 +++++++++++++ arch/arm64/kvm/hyp/include/hyp/switch.h | 148 +++--------------------- arch/arm64/kvm/hyp/nvhe/pkvm.c | 1 + 5 files changed, 151 insertions(+), 131 deletions(-) diff --git a/arch/arm64/include/asm/kvm_host.h b/arch/arm64/include/asm/kvm_host.h index b763293281c8..64302c438355 100644 --- a/arch/arm64/include/asm/kvm_host.h +++ b/arch/arm64/include/asm/kvm_host.h @@ -816,6 +816,11 @@ struct kvm_vcpu_arch { u64 hcrx_el2; u64 mdcr_el2; + struct { + u64 r; + u64 w; + } fgt[__NR_FGT_GROUP_IDS__]; + /* Exception Information */ struct kvm_vcpu_fault_info fault; @@ -1600,6 +1605,51 @@ static inline bool kvm_arch_has_irq_bypass(void) void compute_fgu(struct kvm *kvm, enum fgt_group_id fgt); void get_reg_fixed_bits(struct kvm *kvm, enum vcpu_sysreg reg, u64 *res0, u64 *res1); void check_feature_map(void); +void kvm_vcpu_load_fgt(struct kvm_vcpu *vcpu); +static __always_inline enum fgt_group_id __fgt_reg_to_group_id(enum vcpu_sysreg reg) +{ + switch (reg) { + case HFGRTR_EL2: + case HFGWTR_EL2: + return HFGRTR_GROUP; + case HFGITR_EL2: + return HFGITR_GROUP; + case HDFGRTR_EL2: + case HDFGWTR_EL2: + return HDFGRTR_GROUP; + case HAFGRTR_EL2: + return HAFGRTR_GROUP; + case HFGRTR2_EL2: + case HFGWTR2_EL2: + return HFGRTR2_GROUP; + case HFGITR2_EL2: + return HFGITR2_GROUP; + case HDFGRTR2_EL2: + case HDFGWTR2_EL2: + return HDFGRTR2_GROUP; + default: + BUILD_BUG_ON(1); + } +} + +#define vcpu_fgt(vcpu, reg) \ + ({ \ + enum fgt_group_id id = __fgt_reg_to_group_id(reg); \ + u64 *p; \ + switch (reg) { \ + case HFGWTR_EL2: \ + case HDFGWTR_EL2: \ + case HFGWTR2_EL2: \ + case HDFGWTR2_EL2: \ + p = &(vcpu)->arch.fgt[id].w; \ + break; \ + default: \ + p = &(vcpu)->arch.fgt[id].r; \ + break; \ + } \ + \ + p; \ + }) #endif /* __ARM64_KVM_HOST_H__ */ diff --git a/arch/arm64/kvm/arm.c b/arch/arm64/kvm/arm.c index f01cacb669cf..870953b4a8a7 100644 --- a/arch/arm64/kvm/arm.c +++ b/arch/arm64/kvm/arm.c @@ -642,6 +642,7 @@ void kvm_arch_vcpu_load(struct kvm_vcpu *vcpu, int cpu) vcpu->arch.hcr_el2 |= HCR_TWI; vcpu_set_pauth_traps(vcpu); + kvm_vcpu_load_fgt(vcpu); if (is_protected_kvm_enabled()) { kvm_call_hyp_nvhe(__pkvm_vcpu_load, diff --git a/arch/arm64/kvm/config.c b/arch/arm64/kvm/config.c index fbd8944a3dea..b1cf7660efe1 100644 --- a/arch/arm64/kvm/config.c +++ b/arch/arm64/kvm/config.c @@ -5,6 +5,8 @@ */ #include +#include +#include #include /* @@ -1428,3 +1430,83 @@ void get_reg_fixed_bits(struct kvm *kvm, enum vcpu_sysreg reg, u64 *res0, u64 *r break; } } + +static __always_inline struct fgt_masks *__fgt_reg_to_masks(enum vcpu_sysreg reg) +{ + switch (reg) { + case HFGRTR_EL2: + return &hfgrtr_masks; + case HFGWTR_EL2: + return &hfgwtr_masks; + case HFGITR_EL2: + return &hfgitr_masks; + case HDFGRTR_EL2: + return &hdfgrtr_masks; + case HDFGWTR_EL2: + return &hdfgwtr_masks; + case HAFGRTR_EL2: + return &hafgrtr_masks; + case HFGRTR2_EL2: + return &hfgrtr2_masks; + case HFGWTR2_EL2: + return &hfgwtr2_masks; + case HFGITR2_EL2: + return &hfgitr2_masks; + case HDFGRTR2_EL2: + return &hdfgrtr2_masks; + case HDFGWTR2_EL2: + return &hdfgwtr2_masks; + default: + BUILD_BUG_ON(1); + } +} + +static __always_inline void __compute_fgt(struct kvm_vcpu *vcpu, enum vcpu_sysreg reg) +{ + u64 fgu = vcpu->kvm->arch.fgu[__fgt_reg_to_group_id(reg)]; + struct fgt_masks *m = __fgt_reg_to_masks(reg); + u64 clear = 0, set = 0, val = m->nmask; + + set |= fgu & m->mask; + clear |= fgu & m->nmask; + + if (is_nested_ctxt(vcpu)) { + u64 nested = __vcpu_sys_reg(vcpu, reg); + set |= nested & m->mask; + clear |= ~nested & m->nmask; + } + + val |= set; + val &= ~clear; + *vcpu_fgt(vcpu, reg) = val; +} + +static void __compute_hfgwtr(struct kvm_vcpu *vcpu) +{ + __compute_fgt(vcpu, HFGWTR_EL2); + + if (cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38)) + *vcpu_fgt(vcpu, HFGWTR_EL2) |= HFGWTR_EL2_TCR_EL1; +} + +void kvm_vcpu_load_fgt(struct kvm_vcpu *vcpu) +{ + if (!cpus_have_final_cap(ARM64_HAS_FGT)) + return; + + __compute_fgt(vcpu, HFGRTR_EL2); + __compute_hfgwtr(vcpu); + __compute_fgt(vcpu, HFGITR_EL2); + __compute_fgt(vcpu, HDFGRTR_EL2); + __compute_fgt(vcpu, HDFGWTR_EL2); + __compute_fgt(vcpu, HAFGRTR_EL2); + + if (!cpus_have_final_cap(ARM64_HAS_FGT2)) + return; + + __compute_fgt(vcpu, HFGRTR2_EL2); + __compute_fgt(vcpu, HFGWTR2_EL2); + __compute_fgt(vcpu, HFGITR2_EL2); + __compute_fgt(vcpu, HDFGRTR2_EL2); + __compute_fgt(vcpu, HDFGWTR2_EL2); +} diff --git a/arch/arm64/kvm/hyp/include/hyp/switch.h b/arch/arm64/kvm/hyp/include/hyp/switch.h index b6682202edf3..c5d5e5b86eaf 100644 --- a/arch/arm64/kvm/hyp/include/hyp/switch.h +++ b/arch/arm64/kvm/hyp/include/hyp/switch.h @@ -195,123 +195,6 @@ static inline void __deactivate_cptr_traps(struct kvm_vcpu *vcpu) __deactivate_cptr_traps_nvhe(vcpu); } -#define reg_to_fgt_masks(reg) \ - ({ \ - struct fgt_masks *m; \ - switch(reg) { \ - case HFGRTR_EL2: \ - m = &hfgrtr_masks; \ - break; \ - case HFGWTR_EL2: \ - m = &hfgwtr_masks; \ - break; \ - case HFGITR_EL2: \ - m = &hfgitr_masks; \ - break; \ - case HDFGRTR_EL2: \ - m = &hdfgrtr_masks; \ - break; \ - case HDFGWTR_EL2: \ - m = &hdfgwtr_masks; \ - break; \ - case HAFGRTR_EL2: \ - m = &hafgrtr_masks; \ - break; \ - case HFGRTR2_EL2: \ - m = &hfgrtr2_masks; \ - break; \ - case HFGWTR2_EL2: \ - m = &hfgwtr2_masks; \ - break; \ - case HFGITR2_EL2: \ - m = &hfgitr2_masks; \ - break; \ - case HDFGRTR2_EL2: \ - m = &hdfgrtr2_masks; \ - break; \ - case HDFGWTR2_EL2: \ - m = &hdfgwtr2_masks; \ - break; \ - default: \ - BUILD_BUG_ON(1); \ - } \ - \ - m; \ - }) - -#define compute_clr_set(vcpu, reg, clr, set) \ - do { \ - u64 hfg = __vcpu_sys_reg(vcpu, reg); \ - struct fgt_masks *m = reg_to_fgt_masks(reg); \ - set |= hfg & m->mask; \ - clr |= ~hfg & m->nmask; \ - } while(0) - -#define reg_to_fgt_group_id(reg) \ - ({ \ - enum fgt_group_id id; \ - switch(reg) { \ - case HFGRTR_EL2: \ - case HFGWTR_EL2: \ - id = HFGRTR_GROUP; \ - break; \ - case HFGITR_EL2: \ - id = HFGITR_GROUP; \ - break; \ - case HDFGRTR_EL2: \ - case HDFGWTR_EL2: \ - id = HDFGRTR_GROUP; \ - break; \ - case HAFGRTR_EL2: \ - id = HAFGRTR_GROUP; \ - break; \ - case HFGRTR2_EL2: \ - case HFGWTR2_EL2: \ - id = HFGRTR2_GROUP; \ - break; \ - case HFGITR2_EL2: \ - id = HFGITR2_GROUP; \ - break; \ - case HDFGRTR2_EL2: \ - case HDFGWTR2_EL2: \ - id = HDFGRTR2_GROUP; \ - break; \ - default: \ - BUILD_BUG_ON(1); \ - } \ - \ - id; \ - }) - -#define compute_undef_clr_set(vcpu, kvm, reg, clr, set) \ - do { \ - u64 hfg = kvm->arch.fgu[reg_to_fgt_group_id(reg)]; \ - struct fgt_masks *m = reg_to_fgt_masks(reg); \ - set |= hfg & m->mask; \ - clr |= hfg & m->nmask; \ - } while(0) - -#define update_fgt_traps_cs(hctxt, vcpu, kvm, reg, clr, set) \ - do { \ - struct fgt_masks *m = reg_to_fgt_masks(reg); \ - u64 c = clr, s = set; \ - u64 val; \ - \ - ctxt_sys_reg(hctxt, reg) = read_sysreg_s(SYS_ ## reg); \ - if (is_nested_ctxt(vcpu)) \ - compute_clr_set(vcpu, reg, c, s); \ - \ - compute_undef_clr_set(vcpu, kvm, reg, c, s); \ - \ - val = m->nmask; \ - val |= s; \ - val &= ~c; \ - write_sysreg_s(val, SYS_ ## reg); \ - } while(0) - -#define update_fgt_traps(hctxt, vcpu, kvm, reg) \ - update_fgt_traps_cs(hctxt, vcpu, kvm, reg, 0, 0) - static inline bool cpu_has_amu(void) { u64 pfr0 = read_sysreg_s(SYS_ID_AA64PFR0_EL1); @@ -320,33 +203,36 @@ static inline bool cpu_has_amu(void) ID_AA64PFR0_EL1_AMU_SHIFT); } +#define __activate_fgt(hctxt, vcpu, reg) \ + do { \ + ctxt_sys_reg(hctxt, reg) = read_sysreg_s(SYS_ ## reg); \ + write_sysreg_s(*vcpu_fgt(vcpu, reg), SYS_ ## reg); \ + } while (0) + static inline void __activate_traps_hfgxtr(struct kvm_vcpu *vcpu) { struct kvm_cpu_context *hctxt = host_data_ptr(host_ctxt); - struct kvm *kvm = kern_hyp_va(vcpu->kvm); if (!cpus_have_final_cap(ARM64_HAS_FGT)) return; - update_fgt_traps(hctxt, vcpu, kvm, HFGRTR_EL2); - update_fgt_traps_cs(hctxt, vcpu, kvm, HFGWTR_EL2, 0, - cpus_have_final_cap(ARM64_WORKAROUND_AMPERE_AC03_CPU_38) ? - HFGWTR_EL2_TCR_EL1_MASK : 0); - update_fgt_traps(hctxt, vcpu, kvm, HFGITR_EL2); - update_fgt_traps(hctxt, vcpu, kvm, HDFGRTR_EL2); - update_fgt_traps(hctxt, vcpu, kvm, HDFGWTR_EL2); + __activate_fgt(hctxt, vcpu, HFGRTR_EL2); + __activate_fgt(hctxt, vcpu, HFGWTR_EL2); + __activate_fgt(hctxt, vcpu, HFGITR_EL2); + __activate_fgt(hctxt, vcpu, HDFGRTR_EL2); + __activate_fgt(hctxt, vcpu, HDFGWTR_EL2); if (cpu_has_amu()) - update_fgt_traps(hctxt, vcpu, kvm, HAFGRTR_EL2); + __activate_fgt(hctxt, vcpu, HAFGRTR_EL2); if (!cpus_have_final_cap(ARM64_HAS_FGT2)) return; - update_fgt_traps(hctxt, vcpu, kvm, HFGRTR2_EL2); - update_fgt_traps(hctxt, vcpu, kvm, HFGWTR2_EL2); - update_fgt_traps(hctxt, vcpu, kvm, HFGITR2_EL2); - update_fgt_traps(hctxt, vcpu, kvm, HDFGRTR2_EL2); - update_fgt_traps(hctxt, vcpu, kvm, HDFGWTR2_EL2); + __activate_fgt(hctxt, vcpu, HFGRTR2_EL2); + __activate_fgt(hctxt, vcpu, HFGWTR2_EL2); + __activate_fgt(hctxt, vcpu, HFGITR2_EL2); + __activate_fgt(hctxt, vcpu, HDFGRTR2_EL2); + __activate_fgt(hctxt, vcpu, HDFGWTR2_EL2); } #define __deactivate_fgt(htcxt, vcpu, reg) \ diff --git a/arch/arm64/kvm/hyp/nvhe/pkvm.c b/arch/arm64/kvm/hyp/nvhe/pkvm.c index 05774aed09cb..43bde061b65d 100644 --- a/arch/arm64/kvm/hyp/nvhe/pkvm.c +++ b/arch/arm64/kvm/hyp/nvhe/pkvm.c @@ -172,6 +172,7 @@ static int pkvm_vcpu_init_traps(struct pkvm_hyp_vcpu *hyp_vcpu) /* Trust the host for non-protected vcpu features. */ vcpu->arch.hcrx_el2 = host_vcpu->arch.hcrx_el2; + memcpy(vcpu->arch.fgt, host_vcpu->arch.fgt, sizeof(vcpu->arch.fgt)); return 0; } From e0b5a7967dec05144bc98125f98c47f74fd1152b Mon Sep 17 00:00:00 2001 From: Oliver Upton Date: Wed, 24 Sep 2025 16:51:50 -0700 Subject: [PATCH 096/330] KVM: arm64: nv: Use FGT write trap of MDSCR_EL1 when available Marc reports that the performance of running an L3 guest has regressed by 60% as a result of setting MDCR_EL2.TDA to hide bad architecture. That's of course terrible for the single user of recursive NV ;-) While there's nothing to be done on non-FGT systems, take advantage of the precise write trap of MDSCR_EL1 and leave the rest of the debug registers untrapped. Reported-by: Marc Zyngier Signed-off-by: Oliver Upton Reviewed-by: Joey Gouly Signed-off-by: Marc Zyngier --- arch/arm64/kvm/config.c | 10 +++++++++- arch/arm64/kvm/nested.c | 9 ++++++--- 2 files changed, 15 insertions(+), 4 deletions(-) diff --git a/arch/arm64/kvm/config.c b/arch/arm64/kvm/config.c index b1cf7660efe1..24bb3f36e9d5 100644 --- a/arch/arm64/kvm/config.c +++ b/arch/arm64/kvm/config.c @@ -1489,6 +1489,14 @@ static void __compute_hfgwtr(struct kvm_vcpu *vcpu) *vcpu_fgt(vcpu, HFGWTR_EL2) |= HFGWTR_EL2_TCR_EL1; } +static void __compute_hdfgwtr(struct kvm_vcpu *vcpu) +{ + __compute_fgt(vcpu, HDFGWTR_EL2); + + if (is_hyp_ctxt(vcpu)) + *vcpu_fgt(vcpu, HDFGWTR_EL2) |= HDFGWTR_EL2_MDSCR_EL1; +} + void kvm_vcpu_load_fgt(struct kvm_vcpu *vcpu) { if (!cpus_have_final_cap(ARM64_HAS_FGT)) @@ -1498,7 +1506,7 @@ void kvm_vcpu_load_fgt(struct kvm_vcpu *vcpu) __compute_hfgwtr(vcpu); __compute_fgt(vcpu, HFGITR_EL2); __compute_fgt(vcpu, HDFGRTR_EL2); - __compute_fgt(vcpu, HDFGWTR_EL2); + __compute_hdfgwtr(vcpu); __compute_fgt(vcpu, HAFGRTR_EL2); if (!cpus_have_final_cap(ARM64_HAS_FGT2)) diff --git a/arch/arm64/kvm/nested.c b/arch/arm64/kvm/nested.c index 7a045cad6bdf..f04cda40545b 100644 --- a/arch/arm64/kvm/nested.c +++ b/arch/arm64/kvm/nested.c @@ -1859,13 +1859,16 @@ void kvm_nested_setup_mdcr_el2(struct kvm_vcpu *vcpu) { u64 guest_mdcr = __vcpu_sys_reg(vcpu, MDCR_EL2); + if (is_nested_ctxt(vcpu)) + vcpu->arch.mdcr_el2 |= (guest_mdcr & NV_MDCR_GUEST_INCLUDE); /* * In yet another example where FEAT_NV2 is fscking broken, accesses * to MDSCR_EL1 are redirected to the VNCR despite having an effect * at EL2. Use a big hammer to apply sanity. + * + * Unless of course we have FEAT_FGT, in which case we can precisely + * trap MDSCR_EL1. */ - if (is_hyp_ctxt(vcpu)) + else if (!cpus_have_final_cap(ARM64_HAS_FGT)) vcpu->arch.mdcr_el2 |= MDCR_EL2_TDA; - else - vcpu->arch.mdcr_el2 |= (guest_mdcr & NV_MDCR_GUEST_INCLUDE); } From 1696b0cfcf004a3af34ffe4c57a14e837ef18144 Mon Sep 17 00:00:00 2001 From: Zhanjun Dong Date: Mon, 29 Sep 2025 11:29:04 -0400 Subject: [PATCH 097/330] drm/i915/guc: Skip communication warning on reset in progress GuC IRQ and tasklet handler receive just single G2H message, and let other messages to be received from next tasklet. During this chained tasklet process, if reset process started, communication will be disabled. Skip warning for this condition. Fixes: 65dd4ed0f4e1 ("drm/i915/guc: Don't receive all G2H messages in irq handler") Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/15018 Signed-off-by: Zhanjun Dong Reviewed-by: Vinay Belgaumkar Signed-off-by: Daniele Ceraolo Spurio Link: https://lore.kernel.org/r/20250929152904.269776-1-zhanjun.dong@intel.com (cherry picked from commit 604b5ee4a653a70979ce689dbd6a5d942eb016bf) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c index 3e7e5badcc2b..2c651ec024ef 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c @@ -1325,9 +1325,16 @@ static int ct_receive(struct intel_guc_ct *ct) static void ct_try_receive_message(struct intel_guc_ct *ct) { + struct intel_guc *guc = ct_to_guc(ct); int ret; - if (GEM_WARN_ON(!ct->enabled)) + if (!ct->enabled) { + GEM_WARN_ON(!guc_to_gt(guc)->uc.reset_in_progress); + return; + } + + /* When interrupt disabled, message handling is not expected */ + if (!guc->interrupts.enabled) return; ret = ct_receive(ct); From 760039c95c78490c5c66ef584fcd536797ed6a2f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Fri, 3 Oct 2025 17:57:30 +0300 Subject: [PATCH 098/330] drm/i915/frontbuffer: Move bo refcounting intel_frontbuffer_{get,release}() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Currently xe's intel_frontbuffer implementation forgets to hold a reference on the bo. This makes the entire thing extremely fragile as the cleanup order now depends on bo references held by other things (namely intel_fb_bo_framebuffer_fini()). Move the bo refcounting to intel_frontbuffer_{get,release}() so that both i915 and xe do this the same way. I first tried to fix this by having xe do the refcounting from its intel_bo_set_frontbuffer() implementation (which is what i915 does currently), but turns out xe's drm_gem_object_free() can sleep and thus drm_gem_object_put() isn't safe to call while we hold fb_tracking.lock. Fixes: 10690b8a49bc ("drm/i915/display: Add intel_fb_bo_framebuffer_fini") Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20251003145734.7634-2-ville.syrjala@linux.intel.com Reviewed-by: Jani Nikula (cherry picked from commit eb4d490729a5fd8dc5a76d334f8d01fec7c14bbe) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_frontbuffer.c | 10 +++++++++- drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h | 2 -- 2 files changed, 9 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_frontbuffer.c b/drivers/gpu/drm/i915/display/intel_frontbuffer.c index 43be5377ddc1..73ed28ac9573 100644 --- a/drivers/gpu/drm/i915/display/intel_frontbuffer.c +++ b/drivers/gpu/drm/i915/display/intel_frontbuffer.c @@ -270,6 +270,8 @@ static void frontbuffer_release(struct kref *ref) spin_unlock(&display->fb_tracking.lock); i915_active_fini(&front->write); + + drm_gem_object_put(obj); kfree_rcu(front, rcu); } @@ -287,6 +289,8 @@ intel_frontbuffer_get(struct drm_gem_object *obj) if (!front) return NULL; + drm_gem_object_get(obj); + front->obj = obj; kref_init(&front->ref); atomic_set(&front->bits, 0); @@ -299,8 +303,12 @@ intel_frontbuffer_get(struct drm_gem_object *obj) spin_lock(&display->fb_tracking.lock); cur = intel_bo_set_frontbuffer(obj, front); spin_unlock(&display->fb_tracking.lock); - if (cur != front) + + if (cur != front) { + drm_gem_object_put(obj); kfree(front); + } + return cur; } diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h b/drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h index b6dc3d1b9bb1..b682969e3a29 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_frontbuffer.h @@ -89,12 +89,10 @@ i915_gem_object_set_frontbuffer(struct drm_i915_gem_object *obj, if (!front) { RCU_INIT_POINTER(obj->frontbuffer, NULL); - drm_gem_object_put(intel_bo_to_drm_bo(obj)); } else if (rcu_access_pointer(obj->frontbuffer)) { cur = rcu_dereference_protected(obj->frontbuffer, true); kref_get(&cur->ref); } else { - drm_gem_object_get(intel_bo_to_drm_bo(obj)); rcu_assign_pointer(obj->frontbuffer, front); } From 86af6b90e0556fcefbc6e98eb78bdce90327ee76 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Fri, 3 Oct 2025 17:57:31 +0300 Subject: [PATCH 099/330] drm/i915/fb: Fix the set_tiling vs. addfb race, again MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit intel_frontbuffer_get() is what locks out subsequent set_tiling changes to the bo. Thus the fence vs. modifier check must be done after intel_frontbuffer_get(), or else a concurrent set_tiling ioctl might sneak in and change the fence after the check has been done. Close the race again. See commit dd689287b977 ("drm/i915: Prevent concurrent tiling/framebuffer modifications") for the previous instance. v2: Reorder intel_user_framebuffer_destroy() to match the unwind (Jani) Cc: Jouni Högander Reviewed-by: Jani Nikula Fixes: 10690b8a49bc ("drm/i915/display: Add intel_fb_bo_framebuffer_fini") Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20251003145734.7634-3-ville.syrjala@linux.intel.com (cherry picked from commit 1d1e4ded216017f8febd91332ee337f0e0e79285) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_fb.c | 38 +++++++++++++------------ 1 file changed, 20 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_fb.c b/drivers/gpu/drm/i915/display/intel_fb.c index 22a4a1575d22..b817ff44c043 100644 --- a/drivers/gpu/drm/i915/display/intel_fb.c +++ b/drivers/gpu/drm/i915/display/intel_fb.c @@ -2113,10 +2113,10 @@ static void intel_user_framebuffer_destroy(struct drm_framebuffer *fb) if (intel_fb_uses_dpt(fb)) intel_dpt_destroy(intel_fb->dpt_vm); - intel_frontbuffer_put(intel_fb->frontbuffer); - intel_fb_bo_framebuffer_fini(intel_fb_bo(fb)); + intel_frontbuffer_put(intel_fb->frontbuffer); + kfree(intel_fb); } @@ -2218,15 +2218,17 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, int ret = -EINVAL; int i; + /* + * intel_frontbuffer_get() must be done before + * intel_fb_bo_framebuffer_init() to avoid set_tiling vs. addfb race. + */ + intel_fb->frontbuffer = intel_frontbuffer_get(obj); + if (!intel_fb->frontbuffer) + return -ENOMEM; + ret = intel_fb_bo_framebuffer_init(fb, obj, mode_cmd); if (ret) - return ret; - - intel_fb->frontbuffer = intel_frontbuffer_get(obj); - if (!intel_fb->frontbuffer) { - ret = -ENOMEM; - goto err; - } + goto err_frontbuffer_put; ret = -EINVAL; if (!drm_any_plane_has_format(display->drm, @@ -2235,7 +2237,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_dbg_kms(display->drm, "unsupported pixel format %p4cc / modifier 0x%llx\n", &mode_cmd->pixel_format, mode_cmd->modifier[0]); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } max_stride = intel_fb_max_stride(display, mode_cmd->pixel_format, @@ -2246,7 +2248,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, mode_cmd->modifier[0] != DRM_FORMAT_MOD_LINEAR ? "tiled" : "linear", mode_cmd->pitches[0], max_stride); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } /* FIXME need to adjust LINOFF/TILEOFF accordingly. */ @@ -2254,7 +2256,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_dbg_kms(display->drm, "plane 0 offset (0x%08x) must be 0\n", mode_cmd->offsets[0]); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } drm_helper_mode_fill_fb_struct(display->drm, fb, info, mode_cmd); @@ -2264,7 +2266,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, if (mode_cmd->handles[i] != mode_cmd->handles[0]) { drm_dbg_kms(display->drm, "bad plane %d handle\n", i); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } stride_alignment = intel_fb_stride_alignment(fb, i); @@ -2272,7 +2274,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_dbg_kms(display->drm, "plane %d pitch (%d) must be at least %u byte aligned\n", i, fb->pitches[i], stride_alignment); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } if (intel_fb_is_gen12_ccs_aux_plane(fb, i)) { @@ -2282,7 +2284,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, drm_dbg_kms(display->drm, "ccs aux plane %d pitch (%d) must be %d\n", i, fb->pitches[i], ccs_aux_stride); - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; } } @@ -2291,7 +2293,7 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, ret = intel_fill_fb_info(display, intel_fb); if (ret) - goto err_frontbuffer_put; + goto err_bo_framebuffer_fini; if (intel_fb_uses_dpt(fb)) { struct i915_address_space *vm; @@ -2317,10 +2319,10 @@ int intel_framebuffer_init(struct intel_framebuffer *intel_fb, err_free_dpt: if (intel_fb_uses_dpt(fb)) intel_dpt_destroy(intel_fb->dpt_vm); +err_bo_framebuffer_fini: + intel_fb_bo_framebuffer_fini(obj); err_frontbuffer_put: intel_frontbuffer_put(intel_fb->frontbuffer); -err: - intel_fb_bo_framebuffer_fini(obj); return ret; } From 9858ea4c29c283f0a8a3cdbb42108d464ece90a8 Mon Sep 17 00:00:00 2001 From: Matthew Schwartz Date: Thu, 9 Oct 2025 14:19:00 +0200 Subject: [PATCH 100/330] Revert "drm/amd/display: Only restore backlight after amdgpu_dm_init or dm_resume" This fix regressed the original issue that commit 7875afafba84 ("drm/amd/display: Fix brightness level not retained over reboot") solved, so revert it until a different approach to solve the regression that it caused with AMD_PRIVATE_COLOR is found. Fixes: a490c8d77d50 ("drm/amd/display: Only restore backlight after amdgpu_dm_init or dm_resume") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4620 Cc: stable@vger.kernel.org Signed-off-by: Matthew Schwartz Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 12 ++++-------- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h | 7 ------- 2 files changed, 4 insertions(+), 15 deletions(-) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 0d03e324d5b9..6597475e245d 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -2085,8 +2085,6 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) dc_hardware_init(adev->dm.dc); - adev->dm.restore_backlight = true; - adev->dm.hpd_rx_offload_wq = hpd_rx_irq_create_workqueue(adev); if (!adev->dm.hpd_rx_offload_wq) { drm_err(adev_to_drm(adev), "failed to create hpd rx offload workqueue.\n"); @@ -3442,7 +3440,6 @@ static int dm_resume(struct amdgpu_ip_block *ip_block) dc_set_power_state(dm->dc, DC_ACPI_CM_POWER_STATE_D0); dc_resume(dm->dc); - adev->dm.restore_backlight = true; amdgpu_dm_irq_resume_early(adev); @@ -9969,6 +9966,7 @@ static void amdgpu_dm_commit_streams(struct drm_atomic_state *state, bool mode_set_reset_required = false; u32 i; struct dc_commit_streams_params params = {dc_state->streams, dc_state->stream_count}; + bool set_backlight_level = false; /* Disable writeback */ for_each_old_connector_in_state(state, connector, old_con_state, i) { @@ -10088,6 +10086,7 @@ static void amdgpu_dm_commit_streams(struct drm_atomic_state *state, acrtc->hw_mode = new_crtc_state->mode; crtc->hwmode = new_crtc_state->mode; mode_set_reset_required = true; + set_backlight_level = true; } else if (modereset_required(new_crtc_state)) { drm_dbg_atomic(dev, "Atomic commit: RESET. crtc id %d:[%p]\n", @@ -10144,16 +10143,13 @@ static void amdgpu_dm_commit_streams(struct drm_atomic_state *state, * to fix a flicker issue. * It will cause the dm->actual_brightness is not the current panel brightness * level. (the dm->brightness is the correct panel level) - * So we set the backlight level with dm->brightness value after initial - * set mode. Use restore_backlight flag to avoid setting backlight level - * for every subsequent mode set. + * So we set the backlight level with dm->brightness value after set mode */ - if (dm->restore_backlight) { + if (set_backlight_level) { for (i = 0; i < dm->num_of_edps; i++) { if (dm->backlight_dev[i]) amdgpu_dm_backlight_set_level(dm, i, dm->brightness[i]); } - dm->restore_backlight = false; } } diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h index 009f206226f0..db75e991ac7b 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h @@ -630,13 +630,6 @@ struct amdgpu_display_manager { */ u32 actual_brightness[AMDGPU_DM_MAX_NUM_EDP]; - /** - * @restore_backlight: - * - * Flag to indicate whether to restore backlight after modeset. - */ - bool restore_backlight; - /** * @aux_hpd_discon_quirk: * From 5c05bcf6ae7732da1bd4dc1958d527b5f07f216a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Timur=20Krist=C3=B3f?= Date: Fri, 26 Sep 2025 20:26:12 +0200 Subject: [PATCH 101/330] drm/amd/pm: Disable MCLK switching on SI at high pixel clocks MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On various SI GPUs, a flickering can be observed near the bottom edge of the screen when using a single 4K 60Hz monitor over DP. Disabling MCLK switching works around this problem. Reviewed-by: Alex Deucher Signed-off-by: Timur Kristóf Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c b/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c index cf9932e68055..3a9522c17fee 100644 --- a/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c +++ b/drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c @@ -3500,6 +3500,11 @@ static void si_apply_state_adjust_rules(struct amdgpu_device *adev, * for these GPUs to calculate bandwidth requirements. */ if (high_pixelclock_count) { + /* Work around flickering lines at the bottom edge + * of the screen when using a single 4K 60Hz monitor. + */ + disable_mclk_switching = true; + /* On Oland, we observe some flickering when two 4K 60Hz * displays are connected, possibly because voltage is too low. * Raise the voltage by requiring a higher SCLK. From 7bdd91abf0cb3ea78160e2e78fb58b12f6a38d55 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Timur=20Krist=C3=B3f?= Date: Fri, 26 Sep 2025 20:26:13 +0200 Subject: [PATCH 102/330] drm/amd: Disable ASPM on SI MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Enabling ASPM causes randoms hangs on Tahiti and Oland on Zen4. It's unclear if this is a platform-specific or GPU-specific issue. Disable ASPM on SI for the time being. Reviewed-by: Alex Deucher Signed-off-by: Timur Kristóf Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 7a899fb4de29..3d032c4e2dce 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -1882,6 +1882,13 @@ static bool amdgpu_device_pcie_dynamic_switching_supported(struct amdgpu_device static bool amdgpu_device_aspm_support_quirk(struct amdgpu_device *adev) { + /* Enabling ASPM causes randoms hangs on Tahiti and Oland on Zen4. + * It's unclear if this is a platform-specific or GPU-specific issue. + * Disable ASPM on SI for the time being. + */ + if (adev->family == AMDGPU_FAMILY_SI) + return true; + #if IS_ENABLED(CONFIG_X86) struct cpuinfo_x86 *c = &cpu_data(0); From 5d55ed19d4190d2c210ac05ac7a53f800a8c6fe5 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Wed, 27 Aug 2025 14:47:23 +0200 Subject: [PATCH 103/330] drm/amdgpu: remove two invalid BUG_ON()s MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Those can be triggered trivially by userspace. Signed-off-by: Christian König Reviewed-by: Alex Deucher Acked-by: Timur Kristóf Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c | 2 -- drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c | 2 -- 2 files changed, 4 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 66c47c466532..d61eb9f187c6 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -5862,8 +5862,6 @@ static void gfx_v11_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 header, control = 0; - BUG_ON(ib->flags & AMDGPU_IB_FLAG_CE); - header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); control |= ib->length_dw | (vmid << 24); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c index 710ec9c34e43..93fde0f9af87 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c @@ -4419,8 +4419,6 @@ static void gfx_v12_0_ring_emit_ib_gfx(struct amdgpu_ring *ring, unsigned vmid = AMDGPU_JOB_GET_VMID(job); u32 header, control = 0; - BUG_ON(ib->flags & AMDGPU_IB_FLAG_CE); - header = PACKET3(PACKET3_INDIRECT_BUFFER, 2); control |= ib->length_dw | (vmid << 24); From 8f74c70be57527d7b79e2ecf6de1a154d148254d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Mon, 22 Sep 2025 14:18:16 +0200 Subject: [PATCH 104/330] drm/amdgpu: block CE CS if not explicitely allowed by module option MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The Constant Engine found on gfx6-gfx10 HW has been a notorious source of problems. RADV never used it in the first place, radeonsi only used it for a few releases around 2017 for gfx6-gfx9 before dropping support for it as well. While investigating another problem I just recently found that submitting to the CE seems to be completely broken on gfx9 for quite a while. Since nobody complained about that problem it most likely means that nobody is using any of the affected radeonsi versions on current Linux kernels any more. So to potentially phase out the support for the CE and eliminate another source of problems block submitting CE IBs unless it is enabled again using a debug flag. Signed-off-by: Christian König Reviewed-by: Alex Deucher Acked-by: Timur Kristóf Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 6 ++++++ drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 8 +++++++- 3 files changed, 14 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 2a0df4cabb99..6f5b4a0e0a34 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -1290,6 +1290,7 @@ struct amdgpu_device { bool debug_disable_gpu_ring_reset; bool debug_vm_userptr; bool debug_disable_ce_logs; + bool debug_enable_ce_cs; /* Protection for the following isolation structure */ struct mutex enforce_isolation_mutex; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 9cd7741d2254..ba9fb08db094 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -364,6 +364,12 @@ static int amdgpu_cs_p2_ib(struct amdgpu_cs_parser *p, if (p->uf_bo && ring->funcs->no_user_fence) return -EINVAL; + if (!p->adev->debug_enable_ce_cs && + chunk_ib->flags & AMDGPU_IB_FLAG_CE) { + dev_err_ratelimited(p->adev->dev, "CE CS is blocked, use debug=0x400 to override\n"); + return -EINVAL; + } + if (chunk_ib->ip_type == AMDGPU_HW_IP_GFX && chunk_ib->flags & AMDGPU_IB_FLAG_PREEMPT) { if (chunk_ib->flags & AMDGPU_IB_FLAG_CE) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index bff25ef3e2d0..61268aa82df4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -144,7 +144,8 @@ enum AMDGPU_DEBUG_MASK { AMDGPU_DEBUG_DISABLE_GPU_RING_RESET = BIT(6), AMDGPU_DEBUG_SMU_POOL = BIT(7), AMDGPU_DEBUG_VM_USERPTR = BIT(8), - AMDGPU_DEBUG_DISABLE_RAS_CE_LOG = BIT(9) + AMDGPU_DEBUG_DISABLE_RAS_CE_LOG = BIT(9), + AMDGPU_DEBUG_ENABLE_CE_CS = BIT(10) }; unsigned int amdgpu_vram_limit = UINT_MAX; @@ -2289,6 +2290,11 @@ static void amdgpu_init_debug_options(struct amdgpu_device *adev) pr_info("debug: disable kernel logs of correctable errors\n"); adev->debug_disable_ce_logs = true; } + + if (amdgpu_debug_mask & AMDGPU_DEBUG_ENABLE_CE_CS) { + pr_info("debug: allowing command submission to CE engine\n"); + adev->debug_enable_ce_cs = true; + } } static unsigned long amdgpu_fix_asic_type(struct pci_dev *pdev, unsigned long flags) From 357d90be2c7aaa526a840cddffd2b8d676fe75a6 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 26 Sep 2025 17:31:32 -0400 Subject: [PATCH 105/330] drm/amdgpu: fix handling of harvesting for ip_discovery firmware Chips which use the IP discovery firmware loaded by the driver reported incorrect harvesting information in the ip discovery table in sysfs because the driver only uses the ip discovery firmware for populating sysfs and not for direct parsing for the driver itself as such, the fields that are used to print the harvesting info in sysfs report incorrect data for some IPs. Populate the relevant fields for this case as well. Fixes: 514678da56da ("drm/amdgpu/discovery: fix fw based ip discovery") Acked-by: Tom St Denis Reviewed-by: Lijo Lazar Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c | 18 +++++++++++++++++- 1 file changed, 17 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c index 73401f0aeb34..dd7b2b796427 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c @@ -1033,7 +1033,9 @@ static uint8_t amdgpu_discovery_get_harvest_info(struct amdgpu_device *adev, /* Until a uniform way is figured, get mask based on hwid */ switch (hw_id) { case VCN_HWID: - harvest = ((1 << inst) & adev->vcn.inst_mask) == 0; + /* VCN vs UVD+VCE */ + if (!amdgpu_ip_version(adev, VCE_HWIP, 0)) + harvest = ((1 << inst) & adev->vcn.inst_mask) == 0; break; case DMU_HWID: if (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK) @@ -2565,7 +2567,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 0, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 0, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 0); @@ -2592,7 +2596,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 3, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 3, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 0, 1); @@ -2619,8 +2625,10 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); vega10_reg_base_init(adev); adev->sdma.num_instances = 1; + adev->sdma.sdma_mask = 1; adev->vcn.num_vcn_inst = 1; adev->gmc.num_umc = 2; + adev->gfx.xcc_mask = 1; if (adev->apu_flags & AMD_APU_IS_RAVEN2) { adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 2, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 2, 0); @@ -2665,7 +2673,9 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); vega20_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; adev->gmc.num_umc = 8; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 0); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 0); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 0); @@ -2693,8 +2703,10 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); arct_reg_base_init(adev); adev->sdma.num_instances = 8; + adev->sdma.sdma_mask = 0xff; adev->vcn.num_vcn_inst = 2; adev->gmc.num_umc = 8; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 1); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 1); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 2, 1); @@ -2726,8 +2738,10 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) amdgpu_discovery_init(adev); aldebaran_reg_base_init(adev); adev->sdma.num_instances = 5; + adev->sdma.sdma_mask = 0x1f; adev->vcn.num_vcn_inst = 2; adev->gmc.num_umc = 4; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(9, 4, 2); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(9, 4, 2); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(4, 4, 0); @@ -2762,6 +2776,8 @@ int amdgpu_discovery_set_ip_blocks(struct amdgpu_device *adev) } else { cyan_skillfish_reg_base_init(adev); adev->sdma.num_instances = 2; + adev->sdma.sdma_mask = 3; + adev->gfx.xcc_mask = 1; adev->ip_versions[MMHUB_HWIP][0] = IP_VERSION(2, 0, 3); adev->ip_versions[ATHUB_HWIP][0] = IP_VERSION(2, 0, 3); adev->ip_versions[OSSSYS_HWIP][0] = IP_VERSION(5, 0, 1); From 1f22fcb88bfef26a966e9eb242c692c6bf253d47 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 15 Sep 2025 12:37:32 -0400 Subject: [PATCH 106/330] drm/amdgpu: handle wrap around in reemit handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Compare the sequence numbers directly. Fixes: 77cc0da39c7c ("drm/amdgpu: track ring state associated with a fence") Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index fd8cca241da6..e270df30c279 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -790,14 +790,19 @@ void amdgpu_ring_backup_unprocessed_commands(struct amdgpu_ring *ring, struct dma_fence *unprocessed; struct dma_fence __rcu **ptr; struct amdgpu_fence *fence; - u64 wptr, i, seqno; + u64 wptr; + u32 seq, last_seq; - seqno = amdgpu_fence_read(ring); + last_seq = amdgpu_fence_read(ring) & ring->fence_drv.num_fences_mask; + seq = ring->fence_drv.sync_seq & ring->fence_drv.num_fences_mask; wptr = ring->fence_drv.signalled_wptr; ring->ring_backup_entries_to_copy = 0; - for (i = seqno + 1; i <= ring->fence_drv.sync_seq; ++i) { - ptr = &ring->fence_drv.fences[i & ring->fence_drv.num_fences_mask]; + do { + last_seq++; + last_seq &= ring->fence_drv.num_fences_mask; + + ptr = &ring->fence_drv.fences[last_seq]; rcu_read_lock(); unprocessed = rcu_dereference(*ptr); @@ -813,7 +818,7 @@ void amdgpu_ring_backup_unprocessed_commands(struct amdgpu_ring *ring, wptr = fence->wptr; } rcu_read_unlock(); - } + } while (last_seq != seq); } /* From ff780f4f80323148d43198f2052c14160c8428d3 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 3 Sep 2025 13:48:23 -0400 Subject: [PATCH 107/330] drm/amdgpu: set an error on all fences from a bad context MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When we backup ring contents to reemit after a queue reset, we don't backup ring contents from the bad context. When we signal the fences, we should set an error on those fences as well. v2: misc cleanups v3: add locking for fence error, fix comment (Christian) v4: fix wrap around, locking (Christian) Fixes: 77cc0da39c7c ("drm/amdgpu: track ring state associated with a fence") Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c | 39 ++++++++++++++++++++--- drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 2 +- 3 files changed, 37 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index e270df30c279..18a7829122d2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -758,11 +758,42 @@ void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring) * @fence: fence of the ring to signal * */ -void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *fence) +void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *af) { - dma_fence_set_error(&fence->base, -ETIME); - amdgpu_fence_write(fence->ring, fence->seq); - amdgpu_fence_process(fence->ring); + struct dma_fence *unprocessed; + struct dma_fence __rcu **ptr; + struct amdgpu_fence *fence; + struct amdgpu_ring *ring = af->ring; + unsigned long flags; + u32 seq, last_seq; + + last_seq = amdgpu_fence_read(ring) & ring->fence_drv.num_fences_mask; + seq = ring->fence_drv.sync_seq & ring->fence_drv.num_fences_mask; + + /* mark all fences from the guilty context with an error */ + spin_lock_irqsave(&ring->fence_drv.lock, flags); + do { + last_seq++; + last_seq &= ring->fence_drv.num_fences_mask; + + ptr = &ring->fence_drv.fences[last_seq]; + rcu_read_lock(); + unprocessed = rcu_dereference(*ptr); + + if (unprocessed && !dma_fence_is_signaled_locked(unprocessed)) { + fence = container_of(unprocessed, struct amdgpu_fence, base); + + if (fence == af) + dma_fence_set_error(&fence->base, -ETIME); + else if (fence->context == af->context) + dma_fence_set_error(&fence->base, -ECANCELED); + } + rcu_read_unlock(); + } while (last_seq != seq); + spin_unlock_irqrestore(&ring->fence_drv.lock, flags); + /* signal the guilty fence */ + amdgpu_fence_write(ring, af->seq); + amdgpu_fence_process(ring); } void amdgpu_fence_save_wptr(struct dma_fence *fence) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index 8f6ce948c684..5ec5c3ff22bb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -811,7 +811,7 @@ int amdgpu_ring_reset_helper_end(struct amdgpu_ring *ring, if (r) return r; - /* signal the fence of the bad job */ + /* signal the guilty fence and set an error on all fences from the context */ if (guilty_fence) amdgpu_fence_driver_guilty_force_completion(guilty_fence); /* Re-emit the non-guilty commands */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h index b6b649179776..4b46e3c26ff3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h @@ -155,7 +155,7 @@ extern const struct drm_sched_backend_ops amdgpu_sched_ops; void amdgpu_fence_driver_clear_job_fences(struct amdgpu_ring *ring); void amdgpu_fence_driver_set_error(struct amdgpu_ring *ring, int error); void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring); -void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *fence); +void amdgpu_fence_driver_guilty_force_completion(struct amdgpu_fence *af); void amdgpu_fence_save_wptr(struct dma_fence *fence); int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring); From 6df8e84aa6b5b1812cc2cacd6b3f5ccbb18cda2b Mon Sep 17 00:00:00 2001 From: Gui-Dong Han Date: Wed, 8 Oct 2025 03:43:27 +0000 Subject: [PATCH 108/330] drm/amdgpu: use atomic functions with memory barriers for vm fault info The atomic variable vm_fault_info_updated is used to synchronize access to adev->gmc.vm_fault_info between the interrupt handler and get_vm_fault_info(). The default atomic functions like atomic_set() and atomic_read() do not provide memory barriers. This allows for CPU instruction reordering, meaning the memory accesses to vm_fault_info and the vm_fault_info_updated flag are not guaranteed to occur in the intended order. This creates a race condition that can lead to inconsistent or stale data being used. The previous implementation, which used an explicit mb(), was incomplete and inefficient. It failed to account for all potential CPU reorderings, such as the access of vm_fault_info being reordered before the atomic_read of the flag. This approach is also more verbose and less performant than using the proper atomic functions with acquire/release semantics. Fix this by switching to atomic_set_release() and atomic_read_acquire(). These functions provide the necessary acquire and release semantics, which act as memory barriers to ensure the correct order of operations. It is also more efficient and idiomatic than using explicit full memory barriers. Fixes: b97dfa27ef3a ("drm/amdgpu: save vm fault information for amdkfd") Cc: stable@vger.kernel.org Signed-off-by: Gui-Dong Han Signed-off-by: Felix Kuehling Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 5 ++--- drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c | 7 +++---- drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c | 7 +++---- 3 files changed, 8 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 83020963dfde..a2ca9acf8c4e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -2329,10 +2329,9 @@ void amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(struct kgd_mem *mem) int amdgpu_amdkfd_gpuvm_get_vm_fault_info(struct amdgpu_device *adev, struct kfd_vm_fault_info *mem) { - if (atomic_read(&adev->gmc.vm_fault_info_updated) == 1) { + if (atomic_read_acquire(&adev->gmc.vm_fault_info_updated) == 1) { *mem = *adev->gmc.vm_fault_info; - mb(); /* make sure read happened */ - atomic_set(&adev->gmc.vm_fault_info_updated, 0); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 0); } return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index 93d7ccb7d013..0e5e54d0a9a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -1068,7 +1068,7 @@ static int gmc_v7_0_sw_init(struct amdgpu_ip_block *ip_block) GFP_KERNEL); if (!adev->gmc.vm_fault_info) return -ENOMEM; - atomic_set(&adev->gmc.vm_fault_info_updated, 0); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 0); return 0; } @@ -1290,7 +1290,7 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev, vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID); if (amdgpu_amdkfd_is_kfd_vmid(adev, vmid) - && !atomic_read(&adev->gmc.vm_fault_info_updated)) { + && !atomic_read_acquire(&adev->gmc.vm_fault_info_updated)) { struct kfd_vm_fault_info *info = adev->gmc.vm_fault_info; u32 protections = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, @@ -1306,8 +1306,7 @@ static int gmc_v7_0_process_interrupt(struct amdgpu_device *adev, info->prot_read = protections & 0x8 ? true : false; info->prot_write = protections & 0x10 ? true : false; info->prot_exec = protections & 0x20 ? true : false; - mb(); - atomic_set(&adev->gmc.vm_fault_info_updated, 1); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 1); } return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index c5e2a2c41e06..e1509480dfc2 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -1183,7 +1183,7 @@ static int gmc_v8_0_sw_init(struct amdgpu_ip_block *ip_block) GFP_KERNEL); if (!adev->gmc.vm_fault_info) return -ENOMEM; - atomic_set(&adev->gmc.vm_fault_info_updated, 0); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 0); return 0; } @@ -1478,7 +1478,7 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev, vmid = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, VMID); if (amdgpu_amdkfd_is_kfd_vmid(adev, vmid) - && !atomic_read(&adev->gmc.vm_fault_info_updated)) { + && !atomic_read_acquire(&adev->gmc.vm_fault_info_updated)) { struct kfd_vm_fault_info *info = adev->gmc.vm_fault_info; u32 protections = REG_GET_FIELD(status, VM_CONTEXT1_PROTECTION_FAULT_STATUS, @@ -1494,8 +1494,7 @@ static int gmc_v8_0_process_interrupt(struct amdgpu_device *adev, info->prot_read = protections & 0x8 ? true : false; info->prot_write = protections & 0x10 ? true : false; info->prot_exec = protections & 0x20 ? true : false; - mb(); - atomic_set(&adev->gmc.vm_fault_info_updated, 1); + atomic_set_release(&adev->gmc.vm_fault_info_updated, 1); } return 0; From ef38b4eab146715bc68d45029257f5e69ea3f2cd Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 10 Oct 2025 16:40:57 -0400 Subject: [PATCH 109/330] drm/amdgpu: drop unused structures in amdgpu_drm.h These were never used and are duplicated with the interface that is used. Maybe leftovers from a previous revision of the patch that added them. Fixes: 90c448fef312 ("drm/amdgpu: add new AMDGPU_INFO subquery for userq objects") Reviewed-by: Prike Liang Signed-off-by: Alex Deucher --- include/uapi/drm/amdgpu_drm.h | 21 --------------------- 1 file changed, 21 deletions(-) diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h index cd7402e36b6d..406a42be429b 100644 --- a/include/uapi/drm/amdgpu_drm.h +++ b/include/uapi/drm/amdgpu_drm.h @@ -1555,27 +1555,6 @@ struct drm_amdgpu_info_hw_ip { __u32 userq_num_slots; }; -/* GFX metadata BO sizes and alignment info (in bytes) */ -struct drm_amdgpu_info_uq_fw_areas_gfx { - /* shadow area size */ - __u32 shadow_size; - /* shadow area base virtual mem alignment */ - __u32 shadow_alignment; - /* context save area size */ - __u32 csa_size; - /* context save area base virtual mem alignment */ - __u32 csa_alignment; -}; - -/* IP specific fw related information used in the - * subquery AMDGPU_INFO_UQ_FW_AREAS - */ -struct drm_amdgpu_info_uq_fw_areas { - union { - struct drm_amdgpu_info_uq_fw_areas_gfx gfx; - }; -}; - struct drm_amdgpu_info_num_handles { /** Max handles as supported by firmware for UVD */ __u32 uvd_max_handles; From 6917112af2ba36c5f19075eb9f2933ffd07e55bf Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Timur=20Krist=C3=B3f?= Date: Mon, 13 Oct 2025 08:06:42 +0200 Subject: [PATCH 110/330] drm/amd/powerplay: Fix CIK shutdown temperature MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Remove extra multiplication. CIK GPUs such as Hawaii appear to use PP_TABLE_V0 in which case the shutdown temperature is hardcoded in smu7_init_dpm_defaults and is already multiplied by 1000. The value was mistakenly multiplied another time by smu7_get_thermal_temperature_range. Fixes: 4ba082572a42 ("drm/amd/powerplay: export the thermal ranges of VI asics (V2)") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/1676 Reviewed-by: Alex Deucher Signed-off-by: Timur Kristóf Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c index 8da882c51856..9b28c0728269 100644 --- a/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c +++ b/drivers/gpu/drm/amd/pm/powerplay/hwmgr/smu7_hwmgr.c @@ -5444,8 +5444,7 @@ static int smu7_get_thermal_temperature_range(struct pp_hwmgr *hwmgr, thermal_data->max = table_info->cac_dtp_table->usSoftwareShutdownTemp * PP_TEMPERATURE_UNITS_PER_CENTIGRADES; else if (hwmgr->pp_table_version == PP_TABLE_V0) - thermal_data->max = data->thermal_temp_setting.temperature_shutdown * - PP_TEMPERATURE_UNITS_PER_CENTIGRADES; + thermal_data->max = data->thermal_temp_setting.temperature_shutdown; thermal_data->sw_ctf_threshold = thermal_data->max; From 74de0eaa00eac2e0cbad1dda6dcf8f44ab27629e Mon Sep 17 00:00:00 2001 From: Sathishkumar S Date: Fri, 10 Oct 2025 23:32:40 +0530 Subject: [PATCH 111/330] drm/amdgpu: fix bit shift logic BIT_ULL(n) sets nth bit, remove explicit shift and set the position Fixes: a7a411e24626 ("drm/amdgpu: fix shift-out-of-bounds in amdgpu_debugfs_jpeg_sched_mask_set") Signed-off-by: Sathishkumar S Reviewed-by: Leo Liu Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c index 6b7d66b6d4cc..63ee6ba6a931 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_jpeg.c @@ -371,7 +371,7 @@ static int amdgpu_debugfs_jpeg_sched_mask_set(void *data, u64 val) for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { ring = &adev->jpeg.inst[i].ring_dec[j]; - if (val & (BIT_ULL(1) << ((i * adev->jpeg.num_jpeg_rings) + j))) + if (val & (BIT_ULL((i * adev->jpeg.num_jpeg_rings) + j))) ring->sched.ready = true; else ring->sched.ready = false; From 33cc891b56b93cad1a83263eaf2e417436f70c82 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Tue, 7 Oct 2025 10:10:52 +0200 Subject: [PATCH 112/330] drm/amdgpu: hide VRAM sysfs attributes on GPUs without VRAM MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Otherwise accessing them can cause a crash. Signed-off-by: Christian König Tested-by: Mangesh Gadre Acked-by: Alex Deucher Reviewed-by: Arunpravin Paneer Selvam Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c index a5adb2ed9b3c..9d934c07fa6b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vram_mgr.c @@ -234,6 +234,9 @@ static umode_t amdgpu_vram_attrs_is_visible(struct kobject *kobj, !adev->gmc.vram_vendor) return 0; + if (!ttm_resource_manager_used(&adev->mman.vram_mgr.manager)) + return 0; + return attr->mode; } From 883f309add55060233bf11c1ea6947140372920f Mon Sep 17 00:00:00 2001 From: "Jesse.Zhang" Date: Mon, 13 Oct 2025 13:46:12 +0800 Subject: [PATCH 113/330] drm/amdgpu: Fix NULL pointer dereference in VRAM logic for APU devices MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Previously, APU platforms (and other scenarios with uninitialized VRAM managers) triggered a NULL pointer dereference in `ttm_resource_manager_usage()`. The root cause is not that the `struct ttm_resource_manager *man` pointer itself is NULL, but that `man->bdev` (the backing device pointer within the manager) remains uninitialized (NULL) on APUs—since APUs lack dedicated VRAM and do not fully set up VRAM manager structures. When `ttm_resource_manager_usage()` attempts to acquire `man->bdev->lru_lock`, it dereferences the NULL `man->bdev`, leading to a kernel OOPS. 1. **amdgpu_cs.c**: Extend the existing bandwidth control check in `amdgpu_cs_get_threshold_for_moves()` to include a check for `ttm_resource_manager_used()`. If the manager is not used (uninitialized `bdev`), return 0 for migration thresholds immediately—skipping VRAM-specific logic that would trigger the NULL dereference. 2. **amdgpu_kms.c**: Update the `AMDGPU_INFO_VRAM_USAGE` ioctl and memory info reporting to use a conditional: if the manager is used, return the real VRAM usage; otherwise, return 0. This avoids accessing `man->bdev` when it is NULL. 3. **amdgpu_virt.c**: Modify the vf2pf (virtual function to physical function) data write path. Use `ttm_resource_manager_used()` to check validity: if the manager is usable, calculate `fb_usage` from VRAM usage; otherwise, set `fb_usage` to 0 (APUs have no discrete framebuffer to report). This approach is more robust than APU-specific checks because it: - Works for all scenarios where the VRAM manager is uninitialized (not just APUs), - Aligns with TTM's design by using its native helper function, - Preserves correct behavior for discrete GPUs (which have fully initialized `man->bdev` and pass the `ttm_resource_manager_used()` check). v4: use ttm_resource_manager_used(&adev->mman.vram_mgr.manager) instead of checking the adev->gmc.is_app_apu flag (Christian) Reviewed-by: Christian König Suggested-by: Lijo Lazar Signed-off-by: Jesse Zhang Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 7 ++++--- drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 4 ++-- 3 files changed, 7 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index ba9fb08db094..2f6a96af7fb1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -708,7 +708,7 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev, */ const s64 us_upper_bound = 200000; - if (!adev->mm_stats.log2_max_MBps) { + if ((!adev->mm_stats.log2_max_MBps) || !ttm_resource_manager_used(&adev->mman.vram_mgr.manager)) { *max_bytes = 0; *max_vis_bytes = 0; return; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index a9327472c651..b3e6b3fcdf2c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -758,7 +758,8 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) ui64 = atomic64_read(&adev->num_vram_cpu_page_faults); return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0; case AMDGPU_INFO_VRAM_USAGE: - ui64 = ttm_resource_manager_usage(&adev->mman.vram_mgr.manager); + ui64 = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ? + ttm_resource_manager_usage(&adev->mman.vram_mgr.manager) : 0; return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0; case AMDGPU_INFO_VIS_VRAM_USAGE: ui64 = amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr); @@ -804,8 +805,8 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) mem.vram.usable_heap_size = adev->gmc.real_vram_size - atomic64_read(&adev->vram_pin_size) - AMDGPU_VM_RESERVED_VRAM; - mem.vram.heap_usage = - ttm_resource_manager_usage(vram_man); + mem.vram.heap_usage = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ? + ttm_resource_manager_usage(vram_man) : 0; mem.vram.max_allocation = mem.vram.usable_heap_size * 3 / 4; mem.cpu_accessible_vram.total_heap_size = diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index 3328ab63376b..f96beb96c75c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -598,8 +598,8 @@ static int amdgpu_virt_write_vf2pf_data(struct amdgpu_device *adev) vf2pf_info->driver_cert = 0; vf2pf_info->os_info.all = 0; - vf2pf_info->fb_usage = - ttm_resource_manager_usage(&adev->mman.vram_mgr.manager) >> 20; + vf2pf_info->fb_usage = ttm_resource_manager_used(&adev->mman.vram_mgr.manager) ? + ttm_resource_manager_usage(&adev->mman.vram_mgr.manager) >> 20 : 0; vf2pf_info->fb_vis_usage = amdgpu_vram_mgr_vis_usage(&adev->mman.vram_mgr) >> 20; vf2pf_info->fb_size = adev->gmc.real_vram_size >> 20; From d0de79f66a80eeb849033fae34bd07a69ce72235 Mon Sep 17 00:00:00 2001 From: Jonathan Kim Date: Thu, 9 Oct 2025 10:45:42 -0400 Subject: [PATCH 114/330] drm/amdgpu: fix gfx12 mes packet status return check GFX12 MES uses low 32 bits of status return for success (1 or 0) and high bits for debug information if low bits are 0. GFX11 MES doesn't do this so checking full 64-bit status return for 1 or 0 is still valid. Signed-off-by: Jonathan Kim Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/mes_v12_0.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c index aff06f06aeee..e3149196143e 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c @@ -228,7 +228,12 @@ static int mes_v12_0_submit_pkt_and_poll_completion(struct amdgpu_mes *mes, pipe, x_pkt->header.opcode); r = amdgpu_fence_wait_polling(ring, seq, timeout); - if (r < 1 || !*status_ptr) { + + /* + * status_ptr[31:0] == 0 (fail) or status_ptr[63:0] == 1 (success). + * If status_ptr[31:0] == 0 then status_ptr[63:32] will have debug error information. + */ + if (r < 1 || !(lower_32_bits(*status_ptr))) { if (misc_op_str) dev_err(adev->dev, "MES(%d) failed to respond to msg=%s (%s)\n", From 8745ca5efb2aad0b6591d9b8cd48573ea49c929d Mon Sep 17 00:00:00 2001 From: Jonathan Kim Date: Thu, 9 Oct 2025 10:48:09 -0400 Subject: [PATCH 115/330] drm/amdgpu: fix initialization of doorbell array for detect and hang Initialized doorbells should be set to invalid rather than 0 to prevent driver from over counting hung doorbells since it checks against the invalid value to begin with. Signed-off-by: Jonathan Kim Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index 5bf9be073cdd..30e1fb510600 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -409,7 +409,7 @@ int amdgpu_mes_detect_and_reset_hung_queues(struct amdgpu_device *adev, return -EINVAL; /* Clear the doorbell array before detection */ - memset(adev->mes.hung_queue_db_array_cpu_addr, 0, + memset(adev->mes.hung_queue_db_array_cpu_addr, AMDGPU_MES_INVALID_DB_OFFSET, adev->mes.hung_queue_db_array_size * sizeof(u32)); input.queue_type = queue_type; input.detect_only = detect_only; From 0ef930e1faca6418316e5b9a3b4d1f6ae9e5b240 Mon Sep 17 00:00:00 2001 From: Jonathan Kim Date: Thu, 9 Oct 2025 11:28:19 -0400 Subject: [PATCH 116/330] drm/amdgpu: fix hung reset queue array memory allocation By design the MES will return an array result that is twice the number of hung doorbells it can report. i.e. if up k reported doorbells are supported, then the second half of the array, also of length k, holds the HQD information (type/queue/pipe) where queue 1 corresponds to index 0 and k, queue 2 corresponds to index 1 and k + 1 etc ... The driver will use the HDQ info to target queue/pipe reset for hardware scheduled user compute queues. Signed-off-by: Jonathan Kim Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c | 7 ++++++- drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h | 1 + drivers/gpu/drm/amd/amdgpu/mes_userqueue.c | 6 +++--- drivers/gpu/drm/amd/amdgpu/mes_v11_0.c | 8 +++++--- drivers/gpu/drm/amd/amdgpu/mes_v12_0.c | 8 +++++--- 5 files changed, 20 insertions(+), 10 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index 30e1fb510600..94973018f761 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -420,12 +420,17 @@ int amdgpu_mes_detect_and_reset_hung_queues(struct amdgpu_device *adev, dev_err(adev->dev, "failed to detect and reset\n"); } else { *hung_db_num = 0; - for (i = 0; i < adev->mes.hung_queue_db_array_size; i++) { + for (i = 0; i < adev->mes.hung_queue_hqd_info_offset; i++) { if (db_array[i] != AMDGPU_MES_INVALID_DB_OFFSET) { hung_db_array[i] = db_array[i]; *hung_db_num += 1; } } + + /* + * TODO: return HQD info for MES scheduled user compute queue reset cases + * stored in hung_db_array hqd info offset to full array size + */ } return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h index 6b506fc72f58..97c137c90f97 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.h @@ -149,6 +149,7 @@ struct amdgpu_mes { void *resource_1_addr[AMDGPU_MAX_MES_PIPES]; int hung_queue_db_array_size; + int hung_queue_hqd_info_offset; struct amdgpu_bo *hung_queue_db_array_gpu_obj; uint64_t hung_queue_db_array_gpu_addr; void *hung_queue_db_array_cpu_addr; diff --git a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c index 2db9b2c63693..1cd9eaeef38f 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_userqueue.c @@ -208,10 +208,10 @@ static int mes_userq_detect_and_reset(struct amdgpu_device *adev, struct amdgpu_userq_mgr *uqm, *tmp; unsigned int hung_db_num = 0; int queue_id, r, i; - u32 db_array[4]; + u32 db_array[8]; - if (db_array_size > 4) { - dev_err(adev->dev, "DB array size (%d vs 4) too small\n", + if (db_array_size > 8) { + dev_err(adev->dev, "DB array size (%d vs 8) too small\n", db_array_size); return -EINVAL; } diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index e82188431f79..da575bb1377f 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -66,7 +66,8 @@ static int mes_v11_0_kiq_hw_fini(struct amdgpu_device *adev); #define GFX_MES_DRAM_SIZE 0x80000 #define MES11_HW_RESOURCE_1_SIZE (128 * AMDGPU_GPU_PAGE_SIZE) -#define MES11_HUNG_DB_OFFSET_ARRAY_SIZE 4 +#define MES11_HUNG_DB_OFFSET_ARRAY_SIZE 8 /* [0:3] = db offset, [4:7] = hqd info */ +#define MES11_HUNG_HQD_INFO_OFFSET 4 static void mes_v11_0_ring_set_wptr(struct amdgpu_ring *ring) { @@ -1720,8 +1721,9 @@ static int mes_v11_0_early_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int pipe, r; - adev->mes.hung_queue_db_array_size = - MES11_HUNG_DB_OFFSET_ARRAY_SIZE; + adev->mes.hung_queue_db_array_size = MES11_HUNG_DB_OFFSET_ARRAY_SIZE; + adev->mes.hung_queue_hqd_info_offset = MES11_HUNG_HQD_INFO_OFFSET; + for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) { if (!adev->enable_mes_kiq && pipe == AMDGPU_MES_KIQ_PIPE) continue; diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c index e3149196143e..7f3512d9de07 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v12_0.c @@ -47,7 +47,8 @@ static int mes_v12_0_kiq_hw_fini(struct amdgpu_device *adev); #define MES_EOP_SIZE 2048 -#define MES12_HUNG_DB_OFFSET_ARRAY_SIZE 4 +#define MES12_HUNG_DB_OFFSET_ARRAY_SIZE 8 /* [0:3] = db offset [4:7] hqd info */ +#define MES12_HUNG_HQD_INFO_OFFSET 4 static void mes_v12_0_ring_set_wptr(struct amdgpu_ring *ring) { @@ -1904,8 +1905,9 @@ static int mes_v12_0_early_init(struct amdgpu_ip_block *ip_block) struct amdgpu_device *adev = ip_block->adev; int pipe, r; - adev->mes.hung_queue_db_array_size = - MES12_HUNG_DB_OFFSET_ARRAY_SIZE; + adev->mes.hung_queue_db_array_size = MES12_HUNG_DB_OFFSET_ARRAY_SIZE; + adev->mes.hung_queue_hqd_info_offset = MES12_HUNG_HQD_INFO_OFFSET; + for (pipe = 0; pipe < AMDGPU_MAX_MES_PIPES; pipe++) { r = amdgpu_mes_init_microcode(adev, pipe); if (r) From 277bb0f83e98261018ddd82b7ab8154bb9b93237 Mon Sep 17 00:00:00 2001 From: Jonathan Kim Date: Thu, 5 Jun 2025 10:18:37 -0400 Subject: [PATCH 117/330] drm/amdgpu: enable suspend/resume all for gfx 12 Suspend/resume all gangs has been available for GFX12 for a while now so enable it. Signed-off-by: Jonathan Kim Acked-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c | 11 ++++------- 1 file changed, 4 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c index 94973018f761..4883adcfbb4b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c @@ -691,14 +691,11 @@ int amdgpu_mes_init_microcode(struct amdgpu_device *adev, int pipe) bool amdgpu_mes_suspend_resume_all_supported(struct amdgpu_device *adev) { uint32_t mes_rev = adev->mes.sched_version & AMDGPU_MES_VERSION_MASK; - bool is_supported = false; - if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) && - amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(12, 0, 0) && - mes_rev >= 0x63) - is_supported = true; - - return is_supported; + return ((amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0) && + amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(12, 0, 0) && + mes_rev >= 0x63) || + amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(12, 0, 0)); } /* Fix me -- node_id is used to identify the correct MES instances in the future */ From 079ae5118e1f0dcf5b1ab68ffdb5760b06ed79a2 Mon Sep 17 00:00:00 2001 From: Jonathan Kim Date: Wed, 18 Jun 2025 10:31:15 -0400 Subject: [PATCH 118/330] drm/amdkfd: fix suspend/resume all calls in mes based eviction path Suspend/resume all gangs should be done with the device lock is held. Signed-off-by: Jonathan Kim Acked-by: Alex Deucher Reviewed-by: Harish Kasiviswanathan Signed-off-by: Alex Deucher --- .../drm/amd/amdkfd/kfd_device_queue_manager.c | 73 ++++++------------- 1 file changed, 21 insertions(+), 52 deletions(-) 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 6c5c7c1bf5ed..6e7bc983fc0b 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c @@ -1209,6 +1209,15 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm, pr_debug_ratelimited("Evicting process pid %d queues\n", pdd->process->lead_thread->pid); + if (dqm->dev->kfd->shared_resources.enable_mes) { + pdd->last_evict_timestamp = get_jiffies_64(); + retval = suspend_all_queues_mes(dqm); + if (retval) { + dev_err(dev, "Suspending all queues failed"); + goto out; + } + } + /* Mark all queues as evicted. Deactivate all active queues on * the qpd. */ @@ -1221,23 +1230,27 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm, decrement_queue_count(dqm, qpd, q); if (dqm->dev->kfd->shared_resources.enable_mes) { - int err; - - err = remove_queue_mes(dqm, q, qpd); - if (err) { + retval = remove_queue_mes(dqm, q, qpd); + if (retval) { dev_err(dev, "Failed to evict queue %d\n", q->properties.queue_id); - retval = err; + goto out; } } } - pdd->last_evict_timestamp = get_jiffies_64(); - if (!dqm->dev->kfd->shared_resources.enable_mes) + + if (!dqm->dev->kfd->shared_resources.enable_mes) { + pdd->last_evict_timestamp = get_jiffies_64(); retval = execute_queues_cpsch(dqm, qpd->is_debug ? KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES : KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0, USE_DEFAULT_GRACE_PERIOD); + } else { + retval = resume_all_queues_mes(dqm); + if (retval) + dev_err(dev, "Resuming all queues failed"); + } out: dqm_unlock(dqm); @@ -3098,61 +3111,17 @@ int kfd_dqm_suspend_bad_queue_mes(struct kfd_node *knode, u32 pasid, u32 doorbel return ret; } -static int kfd_dqm_evict_pasid_mes(struct device_queue_manager *dqm, - struct qcm_process_device *qpd) -{ - struct device *dev = dqm->dev->adev->dev; - int ret = 0; - - /* Check if process is already evicted */ - dqm_lock(dqm); - if (qpd->evicted) { - /* Increment the evicted count to make sure the - * process stays evicted before its terminated. - */ - qpd->evicted++; - dqm_unlock(dqm); - goto out; - } - dqm_unlock(dqm); - - ret = suspend_all_queues_mes(dqm); - if (ret) { - dev_err(dev, "Suspending all queues failed"); - goto out; - } - - ret = dqm->ops.evict_process_queues(dqm, qpd); - if (ret) { - dev_err(dev, "Evicting process queues failed"); - goto out; - } - - ret = resume_all_queues_mes(dqm); - if (ret) - dev_err(dev, "Resuming all queues failed"); - -out: - return ret; -} - int kfd_evict_process_device(struct kfd_process_device *pdd) { struct device_queue_manager *dqm; struct kfd_process *p; - int ret = 0; p = pdd->process; dqm = pdd->dev->dqm; WARN(debug_evictions, "Evicting pid %d", p->lead_thread->pid); - if (dqm->dev->kfd->shared_resources.enable_mes) - ret = kfd_dqm_evict_pasid_mes(dqm, &pdd->qpd); - else - ret = dqm->ops.evict_process_queues(dqm, &pdd->qpd); - - return ret; + return dqm->ops.evict_process_queues(dqm, &pdd->qpd); } int reserve_debug_trap_vmid(struct device_queue_manager *dqm, From 927069c4ac2cd1a37efa468596fb5b8f86db9df0 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Mon, 13 Oct 2025 12:05:31 -0600 Subject: [PATCH 119/330] Revert "io_uring/rw: drop -EOPNOTSUPP check in __io_complete_rw_common()" This reverts commit 90bfb28d5fa8127a113a140c9791ea0b40ab156a. Kevin reports that this commit causes an issue for him with LVM snapshots, most likely because of turning off NOWAIT support while a snapshot is being created. This makes -EOPNOTSUPP bubble back through the completion handler, where io_uring read/write handling should just retry it. Reinstate the previous check removed by the referenced commit. Cc: stable@vger.kernel.org Fixes: 90bfb28d5fa8 ("io_uring/rw: drop -EOPNOTSUPP check in __io_complete_rw_common()") Reported-by: Salvatore Bonaccorso Reported-by: Kevin Lumik Link: https://lore.kernel.org/io-uring/cceb723c-051b-4de2-9a4c-4aa82e1619ee@kernel.dk/ Signed-off-by: Jens Axboe --- io_uring/rw.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/io_uring/rw.c b/io_uring/rw.c index 08882648d569..a0f9d2021e3f 100644 --- a/io_uring/rw.c +++ b/io_uring/rw.c @@ -542,7 +542,7 @@ static void __io_complete_rw_common(struct io_kiocb *req, long res) { if (res == req->cqe.res) return; - if (res == -EAGAIN && io_rw_should_reissue(req)) { + if ((res == -EOPNOTSUPP || res == -EAGAIN) && io_rw_should_reissue(req)) { req->flags |= REQ_F_REISSUE | REQ_F_BL_NO_RECYCLE; } else { req_set_fail(req); From 8db4a1d146f83c6bdb0f5b98c50c509ae8549827 Mon Sep 17 00:00:00 2001 From: Mike Snitzer Date: Tue, 7 Oct 2025 13:39:05 -0400 Subject: [PATCH 120/330] NFSv4/flexfiles: fix to allocate mirror->dss before use Move mirror_array's dss_count initialization and dss allocation to ff_layout_alloc_mirror(), just before the loop that initializes each nfs4_ff_layout_ds_stripe's nfs_file_localio. Also handle NULL return from kcalloc() and remove one level of indent in ff_layout_alloc_mirror(). This commit fixes dangling nfsd_serv refcount issues seen when using NFS LOCALIO and then attempting to stop the NFSD service. Fixes: 20b1d75fb840 ("NFSv4/flexfiles: Add support for striped layouts") Signed-off-by: Mike Snitzer Signed-off-by: Anna Schumaker --- fs/nfs/flexfilelayout/flexfilelayout.c | 35 +++++++++++++++----------- 1 file changed, 21 insertions(+), 14 deletions(-) diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index df01d2876b68..9056f05a67dc 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -270,19 +270,31 @@ ff_layout_remove_mirror(struct nfs4_ff_layout_mirror *mirror) mirror->layout = NULL; } -static struct nfs4_ff_layout_mirror *ff_layout_alloc_mirror(gfp_t gfp_flags) +static struct nfs4_ff_layout_mirror *ff_layout_alloc_mirror(u32 dss_count, + gfp_t gfp_flags) { struct nfs4_ff_layout_mirror *mirror; - u32 dss_id; mirror = kzalloc(sizeof(*mirror), gfp_flags); - if (mirror != NULL) { - spin_lock_init(&mirror->lock); - refcount_set(&mirror->ref, 1); - INIT_LIST_HEAD(&mirror->mirrors); - for (dss_id = 0; dss_id < mirror->dss_count; dss_id++) - nfs_localio_file_init(&mirror->dss[dss_id].nfl); + if (mirror == NULL) + return NULL; + + spin_lock_init(&mirror->lock); + refcount_set(&mirror->ref, 1); + INIT_LIST_HEAD(&mirror->mirrors); + + mirror->dss_count = dss_count; + mirror->dss = + kcalloc(dss_count, sizeof(struct nfs4_ff_layout_ds_stripe), + gfp_flags); + if (mirror->dss == NULL) { + kfree(mirror); + return NULL; } + + for (u32 dss_id = 0; dss_id < mirror->dss_count; dss_id++) + nfs_localio_file_init(&mirror->dss[dss_id].nfl); + return mirror; } @@ -507,17 +519,12 @@ ff_layout_alloc_lseg(struct pnfs_layout_hdr *lh, if (dss_count > 1 && stripe_unit == 0) goto out_err_free; - fls->mirror_array[i] = ff_layout_alloc_mirror(gfp_flags); + fls->mirror_array[i] = ff_layout_alloc_mirror(dss_count, gfp_flags); if (fls->mirror_array[i] == NULL) { rc = -ENOMEM; goto out_err_free; } - fls->mirror_array[i]->dss_count = dss_count; - fls->mirror_array[i]->dss = - kcalloc(dss_count, sizeof(struct nfs4_ff_layout_ds_stripe), - gfp_flags); - for (dss_id = 0; dss_id < dss_count; dss_id++) { dss_info = &fls->mirror_array[i]->dss[dss_id]; dss_info->mirror = fls->mirror_array[i]; From 7a84394f02ab1985ebbe0a8d6f6d69bd040de4b3 Mon Sep 17 00:00:00 2001 From: Joshua Watt Date: Tue, 7 Oct 2025 15:22:58 -0600 Subject: [PATCH 121/330] NFS4: Apply delay_retrans to async operations The setting of delay_retrans is applied to synchronous RPC operations because the retransmit count is stored in same struct nfs4_exception that is passed each time an error is checked. However, for asynchronous operations (READ, WRITE, LOCKU, CLOSE, DELEGRETURN), a new struct nfs4_exception is made on the stack each time the task callback is invoked. This means that the retransmit count is always zero and thus delay_retrans never takes effect. Apply delay_retrans to these operations by tracking and updating their retransmit count. Change-Id: Ieb33e046c2b277cb979caa3faca7f52faf0568c9 Signed-off-by: Joshua Watt Reviewed-by: Benjamin Coddington Signed-off-by: Anna Schumaker --- fs/nfs/nfs4proc.c | 13 +++++++++++++ include/linux/nfs_xdr.h | 1 + 2 files changed, 14 insertions(+) diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index f58098417142..411776718494 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3636,6 +3636,7 @@ struct nfs4_closedata { } lr; struct nfs_fattr fattr; unsigned long timestamp; + unsigned short retrans; }; static void nfs4_free_closedata(void *data) @@ -3664,6 +3665,7 @@ static void nfs4_close_done(struct rpc_task *task, void *data) .state = state, .inode = calldata->inode, .stateid = &calldata->arg.stateid, + .retrans = calldata->retrans, }; if (!nfs4_sequence_done(task, &calldata->res.seq_res)) @@ -3711,6 +3713,7 @@ static void nfs4_close_done(struct rpc_task *task, void *data) default: task->tk_status = nfs4_async_handle_exception(task, server, task->tk_status, &exception); + calldata->retrans = exception.retrans; if (exception.retry) goto out_restart; } @@ -5593,9 +5596,11 @@ static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_pgio_header *hdr) .inode = hdr->inode, .state = hdr->args.context->state, .stateid = &hdr->args.stateid, + .retrans = hdr->retrans, }; task->tk_status = nfs4_async_handle_exception(task, server, task->tk_status, &exception); + hdr->retrans = exception.retrans; if (exception.retry) { rpc_restart_call_prepare(task); return -EAGAIN; @@ -5709,10 +5714,12 @@ static int nfs4_write_done_cb(struct rpc_task *task, .inode = hdr->inode, .state = hdr->args.context->state, .stateid = &hdr->args.stateid, + .retrans = hdr->retrans, }; task->tk_status = nfs4_async_handle_exception(task, NFS_SERVER(inode), task->tk_status, &exception); + hdr->retrans = exception.retrans; if (exception.retry) { rpc_restart_call_prepare(task); return -EAGAIN; @@ -6726,6 +6733,7 @@ struct nfs4_delegreturndata { struct nfs_fh fh; nfs4_stateid stateid; unsigned long timestamp; + unsigned short retrans; struct { struct nfs4_layoutreturn_args arg; struct nfs4_layoutreturn_res res; @@ -6746,6 +6754,7 @@ static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata) .inode = data->inode, .stateid = &data->stateid, .task_is_privileged = data->args.seq_args.sa_privileged, + .retrans = data->retrans, }; if (!nfs4_sequence_done(task, &data->res.seq_res)) @@ -6817,6 +6826,7 @@ static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata) task->tk_status = nfs4_async_handle_exception(task, data->res.server, task->tk_status, &exception); + data->retrans = exception.retrans; if (exception.retry) goto out_restart; } @@ -7093,6 +7103,7 @@ struct nfs4_unlockdata { struct file_lock fl; struct nfs_server *server; unsigned long timestamp; + unsigned short retrans; }; static struct nfs4_unlockdata *nfs4_alloc_unlockdata(struct file_lock *fl, @@ -7147,6 +7158,7 @@ static void nfs4_locku_done(struct rpc_task *task, void *data) struct nfs4_exception exception = { .inode = calldata->lsp->ls_state->inode, .stateid = &calldata->arg.stateid, + .retrans = calldata->retrans, }; if (!nfs4_sequence_done(task, &calldata->res.seq_res)) @@ -7180,6 +7192,7 @@ static void nfs4_locku_done(struct rpc_task *task, void *data) task->tk_status = nfs4_async_handle_exception(task, calldata->server, task->tk_status, &exception); + calldata->retrans = exception.retrans; if (exception.retry) rpc_restart_call_prepare(task); } diff --git a/include/linux/nfs_xdr.h b/include/linux/nfs_xdr.h index d56583572c98..31463286402f 100644 --- a/include/linux/nfs_xdr.h +++ b/include/linux/nfs_xdr.h @@ -1659,6 +1659,7 @@ struct nfs_pgio_header { void *netfs; #endif + unsigned short retrans; int pnfs_error; int error; /* merge with pnfs_error */ unsigned int good_bytes; /* boundary of good data */ From 9ff022f3820a31507cb93be6661bf5f3ca0609a4 Mon Sep 17 00:00:00 2001 From: Scott Mayhew Date: Thu, 9 Oct 2025 16:42:12 -0400 Subject: [PATCH 122/330] NFS: check if suid/sgid was cleared after a write as needed I noticed xfstests generic/193 and generic/355 started failing against knfsd after commit e7a8ebc305f2 ("NFSD: Offer write delegation for OPEN with OPEN4_SHARE_ACCESS_WRITE"). I ran those same tests against ONTAP (which has had write delegation support for a lot longer than knfsd) and they fail there too... so while it's a new failure against knfsd, it isn't an entirely new failure. Add the NFS_INO_REVAL_FORCED flag so that the presence of a delegation doesn't keep the inode from being revalidated to fetch the updated mode. Signed-off-by: Scott Mayhew Signed-off-by: Anna Schumaker --- fs/nfs/write.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 0fb6905736d5..336c510f3750 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1535,7 +1535,8 @@ static int nfs_writeback_done(struct rpc_task *task, /* Deal with the suid/sgid bit corner case */ if (nfs_should_remove_suid(inode)) { spin_lock(&inode->i_lock); - nfs_set_cache_invalid(inode, NFS_INO_INVALID_MODE); + nfs_set_cache_invalid(inode, NFS_INO_INVALID_MODE + | NFS_INO_REVAL_FORCED); spin_unlock(&inode->i_lock); } return 0; From 9bb3baa9d1604cd20f49ae7dac9306b4037a0e7a Mon Sep 17 00:00:00 2001 From: Joshua Watt Date: Thu, 9 Oct 2025 15:48:04 -0600 Subject: [PATCH 123/330] NFS4: Fix state renewals missing after boot Since the last renewal time was initialized to 0 and jiffies start counting at -5 minutes, any clients connected in the first 5 minutes after a reboot would have their renewal timer set to a very long interval. If the connection was idle, this would result in the client state timing out on the server and the next call to the server would return NFS4ERR_BADSESSION. Fix this by initializing the last renewal time to the current jiffies instead of 0. Signed-off-by: Joshua Watt Signed-off-by: Anna Schumaker --- fs/nfs/nfs4client.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c index 6fddf43d729c..5998d6bd8a4f 100644 --- a/fs/nfs/nfs4client.c +++ b/fs/nfs/nfs4client.c @@ -222,6 +222,7 @@ struct nfs_client *nfs4_alloc_client(const struct nfs_client_initdata *cl_init) clp->cl_state = 1 << NFS4CLNT_LEASE_EXPIRED; clp->cl_mvops = nfs_v4_minor_ops[cl_init->minorversion]; clp->cl_mig_gen = 1; + clp->cl_last_renewal = jiffies; #if IS_ENABLED(CONFIG_NFS_V4_1) init_waitqueue_head(&clp->cl_lock_waitq); #endif From 02e7567f5da023524476053a38c54f4f19130959 Mon Sep 17 00:00:00 2001 From: Li Ming Date: Wed, 1 Oct 2025 14:03:37 +0800 Subject: [PATCH 124/330] cxl/port: Avoid missing port component registers setup port->nr_dports is used to represent how many dports added to the cxl port, it will increase in add_dport() when a new dport is being added to the cxl port, but it will not be reduced when a dport is removed from the cxl port. Currently, when the first dport is added to a cxl port, it will trigger component registers setup on the cxl port, the implementation is using port->nr_dports to confirm if the dport is the first dport. A corner case here is that adding dport could fail after port->nr_dports updating and before checking port->nr_dports for component registers setup. If the failure happens during the first dport attaching, it will cause that CXL subsystem has not chance to execute component registers setup for the cxl port. the failure flow like below: port->nr_dports = 0 dport 1 adding to the port: add_dport() # port->nr_dports: 1 failed on devm_add_action_or_reset() or sysfs_create_link() return error # port->nr_dports: 1 dport 2 adding to the port: add_dport() # port->nr_dports: 2 no failure skip component registers setup because of port->nr_dports is 2 The solution here is that moving component registers setup closer to add_dport(), so if add_dport() is executed correctly for the first dport, component registers setup on the port will be executed immediately after that. Fixes: f6ee24913de2 ("cxl: Move port register setup to when first dport appear") Signed-off-by: Li Ming Reviewed-by: Dave Jiang Reviewed-by: Davidlohr Bueso Reviewed-by: Jonathan Cameron Signed-off-by: Dave Jiang --- drivers/cxl/core/port.c | 26 ++++++++++++++------------ 1 file changed, 14 insertions(+), 12 deletions(-) diff --git a/drivers/cxl/core/port.c b/drivers/cxl/core/port.c index d5f71eb1ade8..8128fd2b5b31 100644 --- a/drivers/cxl/core/port.c +++ b/drivers/cxl/core/port.c @@ -1182,6 +1182,20 @@ __devm_cxl_add_dport(struct cxl_port *port, struct device *dport_dev, if (rc) return ERR_PTR(rc); + /* + * Setup port register if this is the first dport showed up. Having + * a dport also means that there is at least 1 active link. + */ + if (port->nr_dports == 1 && + port->component_reg_phys != CXL_RESOURCE_NONE) { + rc = cxl_port_setup_regs(port, port->component_reg_phys); + if (rc) { + xa_erase(&port->dports, (unsigned long)dport->dport_dev); + return ERR_PTR(rc); + } + port->component_reg_phys = CXL_RESOURCE_NONE; + } + get_device(dport_dev); rc = devm_add_action_or_reset(host, cxl_dport_remove, dport); if (rc) @@ -1200,18 +1214,6 @@ __devm_cxl_add_dport(struct cxl_port *port, struct device *dport_dev, cxl_debugfs_create_dport_dir(dport); - /* - * Setup port register if this is the first dport showed up. Having - * a dport also means that there is at least 1 active link. - */ - if (port->nr_dports == 1 && - port->component_reg_phys != CXL_RESOURCE_NONE) { - rc = cxl_port_setup_regs(port, port->component_reg_phys); - if (rc) - return ERR_PTR(rc); - port->component_reg_phys = CXL_RESOURCE_NONE; - } - return dport; } From 2b929b6eec0c7c45eb554256d349c16c0ba1df3c Mon Sep 17 00:00:00 2001 From: Pauli Virtanen Date: Mon, 13 Oct 2025 20:33:06 +0300 Subject: [PATCH 125/330] ALSA: usb-audio: add mixer_playback_min_mute quirk for Logitech H390 ID 046d:0a8f Logitech, Inc. H390 headset with microphone is reported to have muted min playback volume. Apply quirk for that. Link: https://gitlab.freedesktop.org/pipewire/pipewire/-/issues/4929 Signed-off-by: Pauli Virtanen Signed-off-by: Takashi Iwai --- sound/usb/quirks.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/sound/usb/quirks.c b/sound/usb/quirks.c index 634cb4fb586f..43793a5c3f51 100644 --- a/sound/usb/quirks.c +++ b/sound/usb/quirks.c @@ -2180,6 +2180,8 @@ static const struct usb_audio_quirk_flags_table quirk_flags_table[] = { QUIRK_FLAG_CTL_MSG_DELAY_1M | QUIRK_FLAG_MIC_RES_384), DEVICE_FLG(0x046d, 0x09a4, /* Logitech QuickCam E 3500 */ QUIRK_FLAG_CTL_MSG_DELAY_1M | QUIRK_FLAG_IGNORE_CTL_ERROR), + DEVICE_FLG(0x046d, 0x0a8f, /* Logitech H390 headset */ + QUIRK_FLAG_MIXER_PLAYBACK_MIN_MUTE), DEVICE_FLG(0x0499, 0x1506, /* Yamaha THR5 */ QUIRK_FLAG_GENERIC_IMPLICIT_FB), DEVICE_FLG(0x0499, 0x1509, /* Steinberg UR22 */ From 15292f1b4c55a3a7c940dbcb6cb8793871ed3d92 Mon Sep 17 00:00:00 2001 From: Babu Moger Date: Fri, 10 Oct 2025 12:08:35 -0500 Subject: [PATCH 126/330] x86/resctrl: Fix miscount of bandwidth event when reactivating previously unavailable RMID Users can create as many monitoring groups as the number of RMIDs supported by the hardware. However, on AMD systems, only a limited number of RMIDs are guaranteed to be actively tracked by the hardware. RMIDs that exceed this limit are placed in an "Unavailable" state. When a bandwidth counter is read for such an RMID, the hardware sets MSR_IA32_QM_CTR.Unavailable (bit 62). When such an RMID starts being tracked again the hardware counter is reset to zero. MSR_IA32_QM_CTR.Unavailable remains set on first read after tracking re-starts and is clear on all subsequent reads as long as the RMID is tracked. resctrl miscounts the bandwidth events after an RMID transitions from the "Unavailable" state back to being tracked. This happens because when the hardware starts counting again after resetting the counter to zero, resctrl in turn compares the new count against the counter value stored from the previous time the RMID was tracked. This results in resctrl computing an event value that is either undercounting (when new counter is more than stored counter) or a mistaken overflow (when new counter is less than stored counter). Reset the stored value (arch_mbm_state::prev_msr) of MSR_IA32_QM_CTR to zero whenever the RMID is in the "Unavailable" state to ensure accurate counting after the RMID resets to zero when it starts to be tracked again. Example scenario that results in mistaken overflow ================================================== 1. The resctrl filesystem is mounted, and a task is assigned to a monitoring group. $mount -t resctrl resctrl /sys/fs/resctrl $mkdir /sys/fs/resctrl/mon_groups/test1/ $echo 1234 > /sys/fs/resctrl/mon_groups/test1/tasks $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 21323 <- Total bytes on domain 0 "Unavailable" <- Total bytes on domain 1 Task is running on domain 0. Counter on domain 1 is "Unavailable". 2. The task runs on domain 0 for a while and then moves to domain 1. The counter starts incrementing on domain 1. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 7345357 <- Total bytes on domain 0 4545 <- Total bytes on domain 1 3. At some point, the RMID in domain 0 transitions to the "Unavailable" state because the task is no longer executing in that domain. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes "Unavailable" <- Total bytes on domain 0 434341 <- Total bytes on domain 1 4. Since the task continues to migrate between domains, it may eventually return to domain 0. $cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 17592178699059 <- Overflow on domain 0 3232332 <- Total bytes on domain 1 In this case, the RMID on domain 0 transitions from "Unavailable" state to active state. The hardware sets MSR_IA32_QM_CTR.Unavailable (bit 62) when the counter is read and begins tracking the RMID counting from 0. Subsequent reads succeed but return a value smaller than the previously saved MSR value (7345357). Consequently, the resctrl's overflow logic is triggered, it compares the previous value (7345357) with the new, smaller value and incorrectly interprets this as a counter overflow, adding a large delta. In reality, this is a false positive: the counter did not overflow but was simply reset when the RMID transitioned from "Unavailable" back to active state. Here is the text from APM [1] available from [2]. "In PQOS Version 2.0 or higher, the MBM hardware will set the U bit on the first QM_CTR read when it begins tracking an RMID that it was not previously tracking. The U bit will be zero for all subsequent reads from that RMID while it is still tracked by the hardware. Therefore, a QM_CTR read with the U bit set when that RMID is in use by a processor can be considered 0 when calculating the difference with a subsequent read." [1] AMD64 Architecture Programmer's Manual Volume 2: System Programming Publication # 24593 Revision 3.41 section 19.3.3 Monitoring L3 Memory Bandwidth (MBM). [ bp: Split commit message into smaller paragraph chunks for better consumption. ] Fixes: 4d05bf71f157d ("x86/resctrl: Introduce AMD QOS feature") Signed-off-by: Babu Moger Signed-off-by: Borislav Petkov (AMD) Reviewed-by: Reinette Chatre Tested-by: Reinette Chatre Cc: stable@vger.kernel.org # needs adjustments for <= v6.17 Link: https://bugzilla.kernel.org/show_bug.cgi?id=206537 # [2] --- arch/x86/kernel/cpu/resctrl/monitor.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/arch/x86/kernel/cpu/resctrl/monitor.c b/arch/x86/kernel/cpu/resctrl/monitor.c index c8945610d455..2cd25a0d4637 100644 --- a/arch/x86/kernel/cpu/resctrl/monitor.c +++ b/arch/x86/kernel/cpu/resctrl/monitor.c @@ -242,7 +242,9 @@ int resctrl_arch_rmid_read(struct rdt_resource *r, struct rdt_mon_domain *d, u32 unused, u32 rmid, enum resctrl_event_id eventid, u64 *val, void *ignored) { + struct rdt_hw_mon_domain *hw_dom = resctrl_to_arch_mon_dom(d); int cpu = cpumask_any(&d->hdr.cpu_mask); + struct arch_mbm_state *am; u64 msr_val; u32 prmid; int ret; @@ -251,12 +253,16 @@ int resctrl_arch_rmid_read(struct rdt_resource *r, struct rdt_mon_domain *d, prmid = logical_rmid_to_physical_rmid(cpu, rmid); ret = __rmid_read_phys(prmid, eventid, &msr_val); - if (ret) - return ret; - *val = get_corrected_val(r, d, rmid, eventid, msr_val); + if (!ret) { + *val = get_corrected_val(r, d, rmid, eventid, msr_val); + } else if (ret == -EINVAL) { + am = get_arch_mbm_state(hw_dom, rmid, eventid); + if (am) + am->prev_msr = 0; + } - return 0; + return ret; } static int __cntr_id_read(u32 cntr_id, u64 *val) From e5ae8d1eb08a3e27fff4ae264af4c8056d908639 Mon Sep 17 00:00:00 2001 From: Kenneth Graunke Date: Fri, 12 Sep 2025 15:31:45 -0700 Subject: [PATCH 127/330] drm/xe: Increase global invalidation timeout to 1000us The previous timeout of 500us seems to be too small; panning the map in the Roll20 VTT in Firefox on a KDE/Wayland desktop reliably triggered timeouts within a few seconds of usage, causing the monitor to freeze and the following to be printed to dmesg: [Jul30 13:44] xe 0000:03:00.0: [drm] *ERROR* GT0: Global invalidation timeout [Jul30 13:48] xe 0000:03:00.0: [drm] *ERROR* [CRTC:82:pipe A] flip_done timed out I haven't hit a single timeout since increasing it to 1000us even after several multi-hour testing sessions. Fixes: 0dd2dd0182bc ("drm/xe: Move DSB l2 flush to a more sensible place") Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/5710 Signed-off-by: Kenneth Graunke Cc: stable@vger.kernel.org Cc: Maarten Lankhorst Reviewed-by: Shuicheng Lin Link: https://lore.kernel.org/r/20250912223254.147940-1-kenneth@whitecape.org Signed-off-by: Lucas De Marchi (cherry picked from commit 146046907b56578263434107f5a7d5051847c459) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_device.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c index 2883b39c9b37..34d33965eac2 100644 --- a/drivers/gpu/drm/xe/xe_device.c +++ b/drivers/gpu/drm/xe/xe_device.c @@ -1070,7 +1070,7 @@ void xe_device_l2_flush(struct xe_device *xe) spin_lock(>->global_invl_lock); xe_mmio_write32(>->mmio, XE2_GLOBAL_INVAL, 0x1); - if (xe_mmio_wait32(>->mmio, XE2_GLOBAL_INVAL, 0x1, 0x0, 500, NULL, true)) + if (xe_mmio_wait32(>->mmio, XE2_GLOBAL_INVAL, 0x1, 0x0, 1000, NULL, true)) xe_gt_err_once(gt, "Global invalidation timeout\n"); spin_unlock(>->global_invl_lock); From 7ac74613e5f2ef3450f44fd2127198662c2563a9 Mon Sep 17 00:00:00 2001 From: Matthew Brost Date: Thu, 9 Oct 2025 04:06:18 -0700 Subject: [PATCH 128/330] drm/xe: Don't allow evicting of BOs in same VM in array of VM binds MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit An array of VM binds can potentially evict other buffer objects (BOs) within the same VM under certain conditions, which may lead to NULL pointer dereferences later in the bind pipeline. To prevent this, clear the allow_res_evict flag in the xe_bo_validate call. v2: - Invert polarity of no_res_evict (Thomas) - Add comment in code explaining issue (Thomas) Cc: stable@vger.kernel.org Reported-by: Paulo Zanoni Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/6268 Fixes: 774b5fa509a9 ("drm/xe: Avoid evicting object of the same vm in none fault mode") Fixes: 77f2ef3f16f5 ("drm/xe: Lock all gpuva ops during VM bind IOCTL") Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs") Signed-off-by: Matthew Brost Tested-by: Paulo Zanoni Reviewed-by: Thomas Hellström Link: https://lore.kernel.org/r/20251009110618.3481870-1-matthew.brost@intel.com (cherry picked from commit 8b9ba8d6d95fe75fed6b0480bb03da4b321bea08) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_vm.c | 32 +++++++++++++++++++++++--------- drivers/gpu/drm/xe/xe_vm_types.h | 1 + 2 files changed, 24 insertions(+), 9 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c index 027e6ce648c5..f602b874e054 100644 --- a/drivers/gpu/drm/xe/xe_vm.c +++ b/drivers/gpu/drm/xe/xe_vm.c @@ -2832,7 +2832,7 @@ static void vm_bind_ioctl_ops_unwind(struct xe_vm *vm, } static int vma_lock_and_validate(struct drm_exec *exec, struct xe_vma *vma, - bool validate) + bool res_evict, bool validate) { struct xe_bo *bo = xe_vma_bo(vma); struct xe_vm *vm = xe_vma_vm(vma); @@ -2843,7 +2843,8 @@ static int vma_lock_and_validate(struct drm_exec *exec, struct xe_vma *vma, err = drm_exec_lock_obj(exec, &bo->ttm.base); if (!err && validate) err = xe_bo_validate(bo, vm, - !xe_vm_in_preempt_fence_mode(vm), exec); + !xe_vm_in_preempt_fence_mode(vm) && + res_evict, exec); } return err; @@ -2913,14 +2914,23 @@ static int prefetch_ranges(struct xe_vm *vm, struct xe_vma_op *op) } static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, - struct xe_vma_op *op) + struct xe_vma_ops *vops, struct xe_vma_op *op) { int err = 0; + bool res_evict; + + /* + * We only allow evicting a BO within the VM if it is not part of an + * array of binds, as an array of binds can evict another BO within the + * bind. + */ + res_evict = !(vops->flags & XE_VMA_OPS_ARRAY_OF_BINDS); switch (op->base.op) { case DRM_GPUVA_OP_MAP: if (!op->map.invalidate_on_bind) err = vma_lock_and_validate(exec, op->map.vma, + res_evict, !xe_vm_in_fault_mode(vm) || op->map.immediate); break; @@ -2931,11 +2941,13 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.remap.unmap->va), - false); + res_evict, false); if (!err && op->remap.prev) - err = vma_lock_and_validate(exec, op->remap.prev, true); + err = vma_lock_and_validate(exec, op->remap.prev, + res_evict, true); if (!err && op->remap.next) - err = vma_lock_and_validate(exec, op->remap.next, true); + err = vma_lock_and_validate(exec, op->remap.next, + res_evict, true); break; case DRM_GPUVA_OP_UNMAP: err = check_ufence(gpuva_to_vma(op->base.unmap.va)); @@ -2944,7 +2956,7 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.unmap.va), - false); + res_evict, false); break; case DRM_GPUVA_OP_PREFETCH: { @@ -2959,7 +2971,7 @@ static int op_lock_and_prep(struct drm_exec *exec, struct xe_vm *vm, err = vma_lock_and_validate(exec, gpuva_to_vma(op->base.prefetch.va), - false); + res_evict, false); if (!err && !xe_vma_has_no_bo(vma)) err = xe_bo_migrate(xe_vma_bo(vma), region_to_mem_type[region], @@ -3005,7 +3017,7 @@ static int vm_bind_ioctl_ops_lock_and_prep(struct drm_exec *exec, return err; list_for_each_entry(op, &vops->list, link) { - err = op_lock_and_prep(exec, vm, op); + err = op_lock_and_prep(exec, vm, vops, op); if (err) return err; } @@ -3638,6 +3650,8 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file) } xe_vma_ops_init(&vops, vm, q, syncs, num_syncs); + if (args->num_binds > 1) + vops.flags |= XE_VMA_OPS_ARRAY_OF_BINDS; for (i = 0; i < args->num_binds; ++i) { u64 range = bind_ops[i].range; u64 addr = bind_ops[i].addr; diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h index da39940501d8..413353e1c225 100644 --- a/drivers/gpu/drm/xe/xe_vm_types.h +++ b/drivers/gpu/drm/xe/xe_vm_types.h @@ -476,6 +476,7 @@ struct xe_vma_ops { /** @flag: signify the properties within xe_vma_ops*/ #define XE_VMA_OPS_FLAG_HAS_SVM_PREFETCH BIT(0) #define XE_VMA_OPS_FLAG_MADVISE BIT(1) +#define XE_VMA_OPS_ARRAY_OF_BINDS BIT(2) u32 flags; #ifdef TEST_VM_OPS_ERROR /** @inject_error: inject error to test error handling */ From d30203739be798d3de5c84db3060e96f00c54e82 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Thu, 18 Sep 2025 13:58:57 -0700 Subject: [PATCH 129/330] drm/xe: Move rebar to be done earlier MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There may be cases in which the BAR0 also needs to move to accommodate the bigger BAR2. However if it's not released, the BAR2 resize fails. During the vram probe it can't be released as it's already in use by xe_mmio for early register access. Add a new function in xe_vram and let xe_pci call it directly before even early device probe. This allows the BAR2 to resize in cases BAR0 also needs to move, assuming there aren't other reasons to hold that move: [] xe 0000:03:00.0: vgaarb: deactivate vga console [] xe 0000:03:00.0: [drm] Attempting to resize bar from 8192MiB -> 16384MiB [] xe 0000:03:00.0: BAR 0 [mem 0x83000000-0x83ffffff 64bit]: releasing [] xe 0000:03:00.0: BAR 2 [mem 0x4000000000-0x41ffffffff 64bit pref]: releasing [] pcieport 0000:02:01.0: bridge window [mem 0x4000000000-0x41ffffffff 64bit pref]: releasing [] pcieport 0000:01:00.0: bridge window [mem 0x4000000000-0x41ffffffff 64bit pref]: releasing [] pcieport 0000:01:00.0: bridge window [mem 0x4000000000-0x43ffffffff 64bit pref]: assigned [] pcieport 0000:02:01.0: bridge window [mem 0x4000000000-0x43ffffffff 64bit pref]: assigned [] xe 0000:03:00.0: BAR 2 [mem 0x4000000000-0x43ffffffff 64bit pref]: assigned [] xe 0000:03:00.0: BAR 0 [mem 0x83000000-0x83ffffff 64bit]: assigned [] pcieport 0000:00:01.0: PCI bridge to [bus 01-04] [] pcieport 0000:00:01.0: bridge window [mem 0x83000000-0x840fffff] [] pcieport 0000:00:01.0: bridge window [mem 0x4000000000-0x44007fffff 64bit pref] [] pcieport 0000:01:00.0: PCI bridge to [bus 02-04] [] pcieport 0000:01:00.0: bridge window [mem 0x83000000-0x840fffff] [] pcieport 0000:01:00.0: bridge window [mem 0x4000000000-0x43ffffffff 64bit pref] [] pcieport 0000:02:01.0: PCI bridge to [bus 03] [] pcieport 0000:02:01.0: bridge window [mem 0x83000000-0x83ffffff] [] pcieport 0000:02:01.0: bridge window [mem 0x4000000000-0x43ffffffff 64bit pref] [] xe 0000:03:00.0: [drm] BAR2 resized to 16384M [] xe 0000:03:00.0: [drm:xe_pci_probe [xe]] BATTLEMAGE e221:0000 dgfx:1 gfx:Xe2_HPG (20.02) ... For BMG there are additional fix needed in the PCI side, but this helps getting it to a working resize. All the rebar logic is more pci-specific than xe-specific and can be done very early in the probe sequence. In future it would be good to move it out of xe_vram.c, but this refactor is left for later. Cc: Ilpo Järvinen Cc: stable@vger.kernel.org # 6.12+ Link: https://lore.kernel.org/intel-xe/fafda2a3-fc63-ce97-d22b-803f771a4d19@linux.intel.com Reviewed-by: Ilpo Järvinen Link: https://lore.kernel.org/r/20250918-xe-pci-rebar-2-v1-2-6c094702a074@intel.com Signed-off-by: Lucas De Marchi (cherry picked from commit 45e33f220fd625492c11e15733d8e9b4f9db82a4) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_pci.c | 2 ++ drivers/gpu/drm/xe/xe_vram.c | 34 ++++++++++++++++++++++++++-------- drivers/gpu/drm/xe/xe_vram.h | 1 + 3 files changed, 29 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c index be91343829dd..9a6df79fc5b6 100644 --- a/drivers/gpu/drm/xe/xe_pci.c +++ b/drivers/gpu/drm/xe/xe_pci.c @@ -867,6 +867,8 @@ static int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (err) return err; + xe_vram_resize_bar(xe); + err = xe_device_probe_early(xe); /* * In Boot Survivability mode, no drm card is exposed and driver diff --git a/drivers/gpu/drm/xe/xe_vram.c b/drivers/gpu/drm/xe/xe_vram.c index b44ebf50fedb..652df7a5f4f6 100644 --- a/drivers/gpu/drm/xe/xe_vram.c +++ b/drivers/gpu/drm/xe/xe_vram.c @@ -26,15 +26,35 @@ #define BAR_SIZE_SHIFT 20 -static void -_resize_bar(struct xe_device *xe, int resno, resource_size_t size) +/* + * Release all the BARs that could influence/block LMEMBAR resizing, i.e. + * assigned IORESOURCE_MEM_64 BARs + */ +static void release_bars(struct pci_dev *pdev) +{ + struct resource *res; + int i; + + pci_dev_for_each_resource(pdev, res, i) { + /* Resource already un-assigned, do not reset it */ + if (!res->parent) + continue; + + /* No need to release unrelated BARs */ + if (!(res->flags & IORESOURCE_MEM_64)) + continue; + + pci_release_resource(pdev, i); + } +} + +static void resize_bar(struct xe_device *xe, int resno, resource_size_t size) { struct pci_dev *pdev = to_pci_dev(xe->drm.dev); int bar_size = pci_rebar_bytes_to_size(size); int ret; - if (pci_resource_len(pdev, resno)) - pci_release_resource(pdev, resno); + release_bars(pdev); ret = pci_resize_resource(pdev, resno, bar_size); if (ret) { @@ -50,7 +70,7 @@ _resize_bar(struct xe_device *xe, int resno, resource_size_t size) * if force_vram_bar_size is set, attempt to set to the requested size * else set to maximum possible size */ -static void resize_vram_bar(struct xe_device *xe) +void xe_vram_resize_bar(struct xe_device *xe) { int force_vram_bar_size = xe_modparam.force_vram_bar_size; struct pci_dev *pdev = to_pci_dev(xe->drm.dev); @@ -119,7 +139,7 @@ static void resize_vram_bar(struct xe_device *xe) pci_read_config_dword(pdev, PCI_COMMAND, &pci_cmd); pci_write_config_dword(pdev, PCI_COMMAND, pci_cmd & ~PCI_COMMAND_MEMORY); - _resize_bar(xe, LMEM_BAR, rebar_size); + resize_bar(xe, LMEM_BAR, rebar_size); pci_assign_unassigned_bus_resources(pdev->bus); pci_write_config_dword(pdev, PCI_COMMAND, pci_cmd); @@ -148,8 +168,6 @@ static int determine_lmem_bar_size(struct xe_device *xe, struct xe_vram_region * return -ENXIO; } - resize_vram_bar(xe); - lmem_bar->io_start = pci_resource_start(pdev, LMEM_BAR); lmem_bar->io_size = pci_resource_len(pdev, LMEM_BAR); if (!lmem_bar->io_size) diff --git a/drivers/gpu/drm/xe/xe_vram.h b/drivers/gpu/drm/xe/xe_vram.h index 72860f714fc6..13505cfb184d 100644 --- a/drivers/gpu/drm/xe/xe_vram.h +++ b/drivers/gpu/drm/xe/xe_vram.h @@ -11,6 +11,7 @@ struct xe_device; struct xe_vram_region; +void xe_vram_resize_bar(struct xe_device *xe); int xe_vram_probe(struct xe_device *xe); struct xe_vram_region *xe_vram_region_alloc(struct xe_device *xe, u8 id, u32 placement); From 1117e7d1e8e66bf7e40291178b829a8513f83a7a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Thomas=20Hellstr=C3=B6m?= Date: Wed, 10 Sep 2025 18:09:39 +0200 Subject: [PATCH 130/330] drm/xe/migrate: Fix an error path MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The exhaustive eviction accidently changed an error path goto to a return. Fix this. Fixes: 59eabff2a352 ("drm/xe: Convert xe_bo_create_pin_map() for exhaustive eviction") Cc: Matthew Brost Signed-off-by: Thomas Hellström Reviewed-by: Francois Dugast Link: https://lore.kernel.org/r/20250910160939.103473-1-thomas.hellstrom@linux.intel.com (cherry picked from commit 381f1ed15159c4b3f00dd37cc70924dedebeb111) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_migrate.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c index 1d667fa36cf3..569869a2b339 100644 --- a/drivers/gpu/drm/xe/xe_migrate.c +++ b/drivers/gpu/drm/xe/xe_migrate.c @@ -434,7 +434,7 @@ int xe_migrate_init(struct xe_migrate *m) err = xe_migrate_lock_prepare_vm(tile, m, vm); if (err) - return err; + goto err_out; if (xe->info.has_usm) { struct xe_hw_engine *hwe = xe_gt_hw_engine(primary_gt, From 7413e9f2be6b2b0caff9c517efa123d988914bba Mon Sep 17 00:00:00 2001 From: Matthew Brost Date: Thu, 9 Oct 2025 06:06:29 -0700 Subject: [PATCH 131/330] drm/xe: Handle mixed mappings and existing VRAM on atomic faults Moving to VRAM will fail if mixed mappings are present or if the page is already located in VRAM. Atomic faults that require a move to VRAM currently retry without attempting to evict mixed mappings or locate existing VRAM mappings. This patch fixes the issue by attempting to evict mixed mappings or find existing VRAM pages when a move to VRAM fails during atomic fault handling. Fixes: a9ac0fa455b0 ("drm/xe: Strict migration policy for atomic SVM faults") Signed-off-by: Matthew Brost Reviewed-by: Himal Prasad Ghimiray Link: https://lore.kernel.org/r/20251009130629.3531962-1-matthew.brost@intel.com (cherry picked from commit 75188605c56d10c1bd3b1cd94f4872f349c3a9c8) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_svm.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c index 7e2db71ff34e..b268ee0d2271 100644 --- a/drivers/gpu/drm/xe/xe_svm.c +++ b/drivers/gpu/drm/xe/xe_svm.c @@ -1073,7 +1073,17 @@ static int __xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma, drm_dbg(&vm->xe->drm, "VRAM allocation failed, falling back to retrying fault, asid=%u, errno=%pe\n", vm->usm.asid, ERR_PTR(err)); - goto retry; + + /* + * In the devmem-only case, mixed mappings may + * be found. The get_pages function will fix + * these up to a single location, allowing the + * page fault handler to make forward progress. + */ + if (ctx.devmem_only) + goto get_pages; + else + goto retry; } else { drm_err(&vm->xe->drm, "VRAM allocation failed, retry count exceeded, asid=%u, errno=%pe\n", @@ -1083,6 +1093,7 @@ static int __xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma, } } +get_pages: get_pages_start = xe_svm_stats_ktime_get(); range_debug(range, "GET PAGES"); From 1852d27aa998272696680607b65a2ceac966104e Mon Sep 17 00:00:00 2001 From: Vinay Belgaumkar Date: Thu, 9 Oct 2025 18:10:47 -0700 Subject: [PATCH 132/330] drm/xe: Enable media sampler power gating Where applicable, enable media sampler power gating. Also, add it to the powergate_info debugfs. v2: Remove the sampler powergate status since it is cleared quickly anyway. v3: Use vcs mask (Rodrigo) and fix the version check for media v4: Remove extra spaces v5: Media samplers are independent of vcs mask, use Media version 1255 (Matt Roper) Fixes: 38e8c4184ea0 ("drm/xe: Enable Coarse Power Gating") Cc: Rodrigo Vivi Cc: Matt Roper Reviewed-by: Rodrigo Vivi Signed-off-by: Vinay Belgaumkar Link: https://lore.kernel.org/r/20251010011047.2047584-1-vinay.belgaumkar@intel.com Signed-off-by: Rodrigo Vivi (cherry picked from commit 4cbc08649a54c3d533df9832342d52d409dfbbf0) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/regs/xe_gt_regs.h | 1 + drivers/gpu/drm/xe/xe_gt_idle.c | 8 ++++++++ 2 files changed, 9 insertions(+) diff --git a/drivers/gpu/drm/xe/regs/xe_gt_regs.h b/drivers/gpu/drm/xe/regs/xe_gt_regs.h index 06cb6b02ec64..51f2a03847f9 100644 --- a/drivers/gpu/drm/xe/regs/xe_gt_regs.h +++ b/drivers/gpu/drm/xe/regs/xe_gt_regs.h @@ -342,6 +342,7 @@ #define POWERGATE_ENABLE XE_REG(0xa210) #define RENDER_POWERGATE_ENABLE REG_BIT(0) #define MEDIA_POWERGATE_ENABLE REG_BIT(1) +#define MEDIA_SAMPLERS_POWERGATE_ENABLE REG_BIT(2) #define VDN_HCP_POWERGATE_ENABLE(n) REG_BIT(3 + 2 * (n)) #define VDN_MFXVDENC_POWERGATE_ENABLE(n) REG_BIT(4 + 2 * (n)) diff --git a/drivers/gpu/drm/xe/xe_gt_idle.c b/drivers/gpu/drm/xe/xe_gt_idle.c index f8950a52d0a4..bdc9d9877ec4 100644 --- a/drivers/gpu/drm/xe/xe_gt_idle.c +++ b/drivers/gpu/drm/xe/xe_gt_idle.c @@ -124,6 +124,9 @@ void xe_gt_idle_enable_pg(struct xe_gt *gt) if (xe_gt_is_main_type(gt)) gtidle->powergate_enable |= RENDER_POWERGATE_ENABLE; + if (MEDIA_VERx100(xe) >= 1100 && MEDIA_VERx100(xe) < 1255) + gtidle->powergate_enable |= MEDIA_SAMPLERS_POWERGATE_ENABLE; + if (xe->info.platform != XE_DG1) { for (i = XE_HW_ENGINE_VCS0, j = 0; i <= XE_HW_ENGINE_VCS7; ++i, ++j) { if ((gt->info.engine_mask & BIT(i))) @@ -246,6 +249,11 @@ int xe_gt_idle_pg_print(struct xe_gt *gt, struct drm_printer *p) drm_printf(p, "Media Slice%d Power Gate Status: %s\n", n, str_up_down(pg_status & media_slices[n].status_bit)); } + + if (MEDIA_VERx100(xe) >= 1100 && MEDIA_VERx100(xe) < 1255) + drm_printf(p, "Media Samplers Power Gating Enabled: %s\n", + str_yes_no(pg_enabled & MEDIA_SAMPLERS_POWERGATE_ENABLE)); + return 0; } From 9f64b3cd051b825de0a2a9f145c8e003200cedd5 Mon Sep 17 00:00:00 2001 From: Shuicheng Lin Date: Fri, 10 Oct 2025 17:25:29 +0000 Subject: [PATCH 133/330] drm/xe/guc: Check GuC running state before deregistering exec queue In normal operation, a registered exec queue is disabled and deregistered through the GuC, and freed only after the GuC confirms completion. However, if the driver is forced to unbind while the exec queue is still running, the user may call exec_destroy() after the GuC has already been stopped and CT communication disabled. In this case, the driver cannot receive a response from the GuC, preventing proper cleanup of exec queue resources. Fix this by directly releasing the resources when GuC is not running. Here is the failure dmesg log: " [ 468.089581] ---[ end trace 0000000000000000 ]--- [ 468.089608] pci 0000:03:00.0: [drm] *ERROR* GT0: GUC ID manager unclean (1/65535) [ 468.090558] pci 0000:03:00.0: [drm] GT0: total 65535 [ 468.090562] pci 0000:03:00.0: [drm] GT0: used 1 [ 468.090564] pci 0000:03:00.0: [drm] GT0: range 1..1 (1) [ 468.092716] ------------[ cut here ]------------ [ 468.092719] WARNING: CPU: 14 PID: 4775 at drivers/gpu/drm/xe/xe_ttm_vram_mgr.c:298 ttm_vram_mgr_fini+0xf8/0x130 [xe] " v2: use xe_uc_fw_is_running() instead of xe_guc_ct_enabled(). As CT may go down and come back during VF migration. Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs") Cc: stable@vger.kernel.org Cc: Matthew Brost Signed-off-by: Shuicheng Lin Reviewed-by: Matthew Brost Signed-off-by: Matthew Brost Link: https://lore.kernel.org/r/20251010172529.2967639-2-shuicheng.lin@intel.com (cherry picked from commit 9b42321a02c50a12b2beb6ae9469606257fbecea) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_guc_submit.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c index 53024eb5670b..94ed8159496f 100644 --- a/drivers/gpu/drm/xe/xe_guc_submit.c +++ b/drivers/gpu/drm/xe/xe_guc_submit.c @@ -44,6 +44,7 @@ #include "xe_ring_ops_types.h" #include "xe_sched_job.h" #include "xe_trace.h" +#include "xe_uc_fw.h" #include "xe_vm.h" static struct xe_guc * @@ -1489,7 +1490,17 @@ static void __guc_exec_queue_process_msg_cleanup(struct xe_sched_msg *msg) xe_gt_assert(guc_to_gt(guc), !(q->flags & EXEC_QUEUE_FLAG_PERMANENT)); trace_xe_exec_queue_cleanup_entity(q); - if (exec_queue_registered(q)) + /* + * Expected state transitions for cleanup: + * - If the exec queue is registered and GuC firmware is running, we must first + * disable scheduling and deregister the queue to ensure proper teardown and + * resource release in the GuC, then destroy the exec queue on driver side. + * - If the GuC is already stopped (e.g., during driver unload or GPU reset), + * we cannot expect a response for the deregister request. In this case, + * it is safe to directly destroy the exec queue on driver side, as the GuC + * will not process further requests and all resources must be cleaned up locally. + */ + if (exec_queue_registered(q) && xe_uc_fw_is_running(&guc->fw)) disable_scheduling_deregister(guc, q); else __guc_exec_queue_destroy(guc, q); From 7e5a5983edda664e8e4bb20af17b80f5135c655c Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Wed, 24 Sep 2025 16:10:38 +0100 Subject: [PATCH 134/330] btrfs: fix clearing of BTRFS_FS_RELOC_RUNNING if relocation already running When starting relocation, at reloc_chunk_start(), if we happen to find the flag BTRFS_FS_RELOC_RUNNING is already set we return an error (-EINPROGRESS) to the callers, however the callers call reloc_chunk_end() which will clear the flag BTRFS_FS_RELOC_RUNNING, which is wrong since relocation was started by another task and still running. Finding the BTRFS_FS_RELOC_RUNNING flag already set is an unexpected scenario, but still our current behaviour is not correct. Fix this by never calling reloc_chunk_end() if reloc_chunk_start() has returned an error, which is what logically makes sense, since the general widespread pattern is to have end functions called only if the counterpart start functions succeeded. This requires changing reloc_chunk_start() to clear BTRFS_FS_RELOC_RUNNING if there's a pending cancel request. Fixes: 907d2710d727 ("btrfs: add cancellable chunk relocation support") CC: stable@vger.kernel.org # 5.15+ Reviewed-by: Boris Burkov Reviewed-by: Johannes Thumshirn Reviewed-by: Qu Wenruo Signed-off-by: Filipe Manana Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/relocation.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c index 8dd8de6b9fb8..0765e06d00b8 100644 --- a/fs/btrfs/relocation.c +++ b/fs/btrfs/relocation.c @@ -3780,6 +3780,7 @@ static noinline_for_stack struct inode *create_reloc_inode( /* * Mark start of chunk relocation that is cancellable. Check if the cancellation * has been requested meanwhile and don't start in that case. + * NOTE: if this returns an error, reloc_chunk_end() must not be called. * * Return: * 0 success @@ -3796,10 +3797,8 @@ static int reloc_chunk_start(struct btrfs_fs_info *fs_info) if (atomic_read(&fs_info->reloc_cancel_req) > 0) { btrfs_info(fs_info, "chunk relocation canceled on start"); - /* - * On cancel, clear all requests but let the caller mark - * the end after cleanup operations. - */ + /* On cancel, clear all requests. */ + clear_and_wake_up_bit(BTRFS_FS_RELOC_RUNNING, &fs_info->flags); atomic_set(&fs_info->reloc_cancel_req, 0); return -ECANCELED; } @@ -3808,9 +3807,11 @@ static int reloc_chunk_start(struct btrfs_fs_info *fs_info) /* * Mark end of chunk relocation that is cancellable and wake any waiters. + * NOTE: call only if a previous call to reloc_chunk_start() succeeded. */ static void reloc_chunk_end(struct btrfs_fs_info *fs_info) { + ASSERT(test_bit(BTRFS_FS_RELOC_RUNNING, &fs_info->flags)); /* Requested after start, clear bit first so any waiters can continue */ if (atomic_read(&fs_info->reloc_cancel_req) > 0) btrfs_info(fs_info, "chunk relocation canceled during operation"); @@ -4023,9 +4024,9 @@ int btrfs_relocate_block_group(struct btrfs_fs_info *fs_info, u64 group_start, if (err && rw) btrfs_dec_block_group_ro(rc->block_group); iput(rc->data_inode); + reloc_chunk_end(fs_info); out_put_bg: btrfs_put_block_group(bg); - reloc_chunk_end(fs_info); free_reloc_control(rc); return err; } @@ -4208,8 +4209,8 @@ int btrfs_recover_relocation(struct btrfs_fs_info *fs_info) ret = ret2; out_unset: unset_reloc_control(rc); -out_end: reloc_chunk_end(fs_info); +out_end: free_reloc_control(rc); out: free_reloc_roots(&reloc_roots); From 53a4acbfc1de85fa637521ffab4f4e2ee03cbeeb Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Miquel=20Sabat=C3=A9=20Sol=C3=A0?= Date: Thu, 25 Sep 2025 20:41:39 +0200 Subject: [PATCH 135/330] btrfs: fix memory leak on duplicated memory in the qgroup assign ioctl MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On 'btrfs_ioctl_qgroup_assign' we first duplicate the argument as provided by the user, which is kfree'd in the end. But this was not the case when allocating memory for 'prealloc'. In this case, if it somehow failed, then the previous code would go directly into calling 'mnt_drop_write_file', without freeing the string duplicated from the user space. Fixes: 4addc1ffd67a ("btrfs: qgroup: preallocate memory before adding a relation") CC: stable@vger.kernel.org # 6.12+ Reviewed-by: Boris Burkov Reviewed-by: Filipe Manana Signed-off-by: Miquel Sabaté Solà Signed-off-by: Filipe Manana Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ioctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c index a454b5ba2097..938286bee6a8 100644 --- a/fs/btrfs/ioctl.c +++ b/fs/btrfs/ioctl.c @@ -3740,7 +3740,7 @@ static long btrfs_ioctl_qgroup_assign(struct file *file, void __user *arg) prealloc = kzalloc(sizeof(*prealloc), GFP_KERNEL); if (!prealloc) { ret = -ENOMEM; - goto drop_write; + goto out; } } From b7fdfd29a136a17c5c8ad9e9bbf89c48919c3d19 Mon Sep 17 00:00:00 2001 From: Qu Wenruo Date: Fri, 26 Sep 2025 14:20:11 +0930 Subject: [PATCH 136/330] btrfs: only set the device specific options after devices are opened [BUG] With v6.17-rc kernels, btrfs will always set 'ssd' mount option even if the block device is not a rotating one: # cat /sys/block/sdd/queue/rotational 1 # cat /etc/fstab: LABEL=DATA2 /data2 btrfs rw,relatime,space_cache=v2,subvolid=5,subvol=/,nofail,nosuid,nodev 0 0 # mount [...] /dev/sdd on /data2 type btrfs (rw,nosuid,nodev,relatime,ssd,space_cache=v2,subvolid=5,subvol=/) [CAUSE] The 'ssd' mount option is set by set_device_specific_options(), and it expects that if there is any rotating device in the btrfs, it will set fs_devices::rotating. However after commit bddf57a70781 ("btrfs: delay btrfs_open_devices() until super block is created"), the device opening is delayed until the super block is created. But the timing of set_device_specific_options() is still left as is, this makes the function be called without any device opened. Since no device is opened, thus fs_devices::rotating will never be set, making btrfs incorrectly set 'ssd' mount option. [FIX] Only call set_device_specific_options() after btrfs_open_devices(). Also only call set_device_specific_options() after a new mount, if we're mounting a mounted btrfs, there is no need to set the device specific mount options again. Reported-by: HAN Yuwei Link: https://lore.kernel.org/linux-btrfs/C8FF75669DFFC3C5+5f93bf8a-80a0-48a6-81bf-4ec890abc99a@bupt.moe/ Fixes: bddf57a70781 ("btrfs: delay btrfs_open_devices() until super block is created") CC: stable@vger.kernel.org # 6.17 Signed-off-by: Qu Wenruo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/super.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index d6e496436539..aadc02374b2a 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -1900,8 +1900,6 @@ static int btrfs_get_tree_super(struct fs_context *fc) return PTR_ERR(sb); } - set_device_specific_options(fs_info); - if (sb->s_root) { /* * Not the first mount of the fs thus got an existing super block. @@ -1946,6 +1944,7 @@ static int btrfs_get_tree_super(struct fs_context *fc) deactivate_locked_super(sb); return -EACCES; } + set_device_specific_options(fs_info); bdev = fs_devices->latest_dev->bdev; snprintf(sb->s_id, sizeof(sb->s_id), "%pg", bdev); shrinker_debugfs_rename(sb->s_shrink, "sb-btrfs:%s", sb->s_id); From 42d3a055d946878a327ee030f0e0c7df0f0f15c8 Mon Sep 17 00:00:00 2001 From: Qu Wenruo Date: Tue, 30 Sep 2025 07:54:30 +0930 Subject: [PATCH 137/330] btrfs: do not use folio_test_partial_kmap() in ASSERT()s [BUG] Syzbot reported an ASSERT() triggered inside scrub: BTRFS info (device loop0): scrub: started on devid 1 assertion failed: !folio_test_partial_kmap(folio) :: 0, in fs/btrfs/scrub.c:697 ------------[ cut here ]------------ kernel BUG at fs/btrfs/scrub.c:697! Oops: invalid opcode: 0000 [#1] SMP KASAN PTI CPU: 0 UID: 0 PID: 6077 Comm: syz.0.17 Not tainted syzkaller #0 PREEMPT(full) Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 08/18/2025 RIP: 0010:scrub_stripe_get_kaddr+0x1bb/0x1c0 fs/btrfs/scrub.c:697 Call Trace: scrub_bio_add_sector fs/btrfs/scrub.c:932 [inline] scrub_submit_initial_read+0xf21/0x1120 fs/btrfs/scrub.c:1897 submit_initial_group_read+0x423/0x5b0 fs/btrfs/scrub.c:1952 flush_scrub_stripes+0x18f/0x1150 fs/btrfs/scrub.c:1973 scrub_stripe+0xbea/0x2a30 fs/btrfs/scrub.c:2516 scrub_chunk+0x2a3/0x430 fs/btrfs/scrub.c:2575 scrub_enumerate_chunks+0xa70/0x1350 fs/btrfs/scrub.c:2839 btrfs_scrub_dev+0x6e7/0x10e0 fs/btrfs/scrub.c:3153 btrfs_ioctl_scrub+0x249/0x4b0 fs/btrfs/ioctl.c:3163 vfs_ioctl fs/ioctl.c:51 [inline] __do_sys_ioctl fs/ioctl.c:597 [inline] __se_sys_ioctl+0xfc/0x170 fs/ioctl.c:583 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0xfa/0xfa0 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x77/0x7f ---[ end trace 0000000000000000 ]--- Which doesn't make much sense, as all the folios we allocated for scrub should not be highmem. [CAUSE] Thankfully syzbot has a detailed kernel config file, showing that CONFIG_DEBUG_KMAP_LOCAL_FORCE_MAP is set to y. And that debug option will force all folio_test_partial_kmap() to return true, to improve coverage on highmem tests. But in our case we really just want to make sure the folios we allocated are not highmem (and they are indeed not). Such incorrect result from folio_test_partial_kmap() is just screwing up everything. [FIX] Replace folio_test_partial_kmap() to folio_test_highmem() so that we won't bother those highmem specific debuging options. Fixes: 5fbaae4b8567 ("btrfs: prepare scrub to support bs > ps cases") Reported-by: syzbot+bde59221318c592e6346@syzkaller.appspotmail.com Signed-off-by: Qu Wenruo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/scrub.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index 4691d0bdb2e8..651b11884f82 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -694,7 +694,7 @@ static void *scrub_stripe_get_kaddr(struct scrub_stripe *stripe, int sector_nr) /* stripe->folios[] is allocated by us and no highmem is allowed. */ ASSERT(folio); - ASSERT(!folio_test_partial_kmap(folio)); + ASSERT(!folio_test_highmem(folio)); return folio_address(folio) + offset_in_folio(folio, offset); } @@ -707,7 +707,7 @@ static phys_addr_t scrub_stripe_get_paddr(struct scrub_stripe *stripe, int secto /* stripe->folios[] is allocated by us and no highmem is allowed. */ ASSERT(folio); - ASSERT(!folio_test_partial_kmap(folio)); + ASSERT(!folio_test_highmem(folio)); /* And the range must be contained inside the folio. */ ASSERT(offset_in_folio(folio, offset) + fs_info->sectorsize <= folio_size(folio)); return page_to_phys(folio_page(folio, 0)) + offset_in_folio(folio, offset); From a5a51bf4e9b7354ce7cd697e610d72c1b33fd949 Mon Sep 17 00:00:00 2001 From: Filipe Manana Date: Wed, 1 Oct 2025 11:08:13 +0100 Subject: [PATCH 138/330] btrfs: do not assert we found block group item when creating free space tree Currently, when building a free space tree at populate_free_space_tree(), if we are not using the block group tree feature, we always expect to find block group items (either extent items or a block group item with key type BTRFS_BLOCK_GROUP_ITEM_KEY) when we search the extent tree with btrfs_search_slot_for_read(), so we assert that we found an item. However this expectation is wrong since we can have a new block group created in the current transaction which is still empty and for which we still have not added the block group's item to the extent tree, in which case we do not have any items in the extent tree associated to the block group. The insertion of a new block group's block group item in the extent tree happens at btrfs_create_pending_block_groups() when it calls the helper insert_block_group_item(). This typically is done when a transaction handle is released, committed or when running delayed refs (either as part of a transaction commit or when serving tickets for space reservation if we are low on free space). So remove the assertion at populate_free_space_tree() even when the block group tree feature is not enabled and update the comment to mention this case. Syzbot reported this with the following stack trace: BTRFS info (device loop3 state M): rebuilding free space tree assertion failed: ret == 0 :: 0, in fs/btrfs/free-space-tree.c:1115 ------------[ cut here ]------------ kernel BUG at fs/btrfs/free-space-tree.c:1115! Oops: invalid opcode: 0000 [#1] SMP KASAN PTI CPU: 1 UID: 0 PID: 6352 Comm: syz.3.25 Not tainted syzkaller #0 PREEMPT(full) Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 08/18/2025 RIP: 0010:populate_free_space_tree+0x700/0x710 fs/btrfs/free-space-tree.c:1115 Code: ff ff e8 d3 (...) RSP: 0018:ffffc9000430f780 EFLAGS: 00010246 RAX: 0000000000000043 RBX: ffff88805b709630 RCX: fea61d0e2e79d000 RDX: 0000000000000000 RSI: 0000000080000000 RDI: 0000000000000000 RBP: ffffc9000430f8b0 R08: ffffc9000430f4a7 R09: 1ffff92000861e94 R10: dffffc0000000000 R11: fffff52000861e95 R12: 0000000000000001 R13: 1ffff92000861f00 R14: dffffc0000000000 R15: 0000000000000000 FS: 00007f424d9fe6c0(0000) GS:ffff888125afc000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fd78ad212c0 CR3: 0000000076d68000 CR4: 00000000003526f0 Call Trace: btrfs_rebuild_free_space_tree+0x1ba/0x6d0 fs/btrfs/free-space-tree.c:1364 btrfs_start_pre_rw_mount+0x128f/0x1bf0 fs/btrfs/disk-io.c:3062 btrfs_remount_rw fs/btrfs/super.c:1334 [inline] btrfs_reconfigure+0xaed/0x2160 fs/btrfs/super.c:1559 reconfigure_super+0x227/0x890 fs/super.c:1076 do_remount fs/namespace.c:3279 [inline] path_mount+0xd1a/0xfe0 fs/namespace.c:4027 do_mount fs/namespace.c:4048 [inline] __do_sys_mount fs/namespace.c:4236 [inline] __se_sys_mount+0x313/0x410 fs/namespace.c:4213 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0xfa/0xfa0 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0033:0x7f424e39066a Code: d8 64 89 02 (...) RSP: 002b:00007f424d9fde68 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5 RAX: ffffffffffffffda RBX: 00007f424d9fdef0 RCX: 00007f424e39066a RDX: 0000200000000180 RSI: 0000200000000380 RDI: 0000000000000000 RBP: 0000200000000180 R08: 00007f424d9fdef0 R09: 0000000000000020 R10: 0000000000000020 R11: 0000000000000246 R12: 0000200000000380 R13: 00007f424d9fdeb0 R14: 0000000000000000 R15: 00002000000002c0 Modules linked in: ---[ end trace 0000000000000000 ]--- Reported-by: syzbot+884dc4621377ba579a6f@syzkaller.appspotmail.com Link: https://lore.kernel.org/linux-btrfs/68dc3dab.a00a0220.102ee.004e.GAE@google.com/ Fixes: a5ed91828518 ("Btrfs: implement the free space B-tree") CC: # 6.1.x: 1961d20f6fa8: btrfs: fix assertion when building free space tree CC: # 6.1.x Reviewed-by: Qu Wenruo Signed-off-by: Filipe Manana Signed-off-by: David Sterba --- fs/btrfs/free-space-tree.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) diff --git a/fs/btrfs/free-space-tree.c b/fs/btrfs/free-space-tree.c index dad0b492a663..d86541073d42 100644 --- a/fs/btrfs/free-space-tree.c +++ b/fs/btrfs/free-space-tree.c @@ -1106,14 +1106,15 @@ static int populate_free_space_tree(struct btrfs_trans_handle *trans, * If ret is 1 (no key found), it means this is an empty block group, * without any extents allocated from it and there's no block group * item (key BTRFS_BLOCK_GROUP_ITEM_KEY) located in the extent tree - * because we are using the block group tree feature, so block group - * items are stored in the block group tree. It also means there are no - * extents allocated for block groups with a start offset beyond this - * block group's end offset (this is the last, highest, block group). + * because we are using the block group tree feature (so block group + * items are stored in the block group tree) or this is a new block + * group created in the current transaction and its block group item + * was not yet inserted in the extent tree (that happens in + * btrfs_create_pending_block_groups() -> insert_block_group_item()). + * It also means there are no extents allocated for block groups with a + * start offset beyond this block group's end offset (this is the last, + * highest, block group). */ - if (!btrfs_fs_compat_ro(trans->fs_info, BLOCK_GROUP_TREE)) - ASSERT(ret == 0); - start = block_group->start; end = block_group->start + block_group->length; while (ret == 0) { From 8ab2fa69691b2913a67f3c54fbb991247b3755be Mon Sep 17 00:00:00 2001 From: Boris Burkov Date: Tue, 30 Sep 2025 21:05:17 -0700 Subject: [PATCH 139/330] btrfs: fix incorrect readahead expansion length The intent of btrfs_readahead_expand() was to expand to the length of the current compressed extent being read. However, "ram_bytes" is *not* that, in the case where a single physical compressed extent is used for multiple file extents. Consider this case with a large compressed extent C and then later two non-compressed extents N1 and N2 written over C, leaving C1 and C2 pointing to offset/len pairs of C: [ C ] [ N1 ][ C1 ][ N2 ][ C2 ] In such a case, ram_bytes for both C1 and C2 is the full uncompressed length of C. So starting readahead in C1 will expand the readahead past the end of C1, past N2, and into C2. This will then expand readahead again, to C2_start + ram_bytes, way past EOF. First of all, this is totally undesirable, we don't want to read the whole file in arbitrary chunks of the large underlying extent if it happens to exist. Secondly, it results in zeroing the range past the end of C2 up to ram_bytes. This is particularly unpleasant with fs-verity as it can zero and set uptodate pages in the verity virtual space past EOF. This incorrect readahead behavior can lead to verity verification errors, if we iterate in a way that happens to do the wrong readahead. Fix this by using em->len for readahead expansion, not em->ram_bytes, resulting in the expected behavior of stopping readahead at the extent boundary. Reported-by: Max Chernoff Link: https://bugzilla.redhat.com/show_bug.cgi?id=2399898 Fixes: 9e9ff875e417 ("btrfs: use readahead_expand() on compressed extents") CC: stable@vger.kernel.org # 6.17 Reviewed-by: Filipe Manana Signed-off-by: Boris Burkov Signed-off-by: David Sterba --- fs/btrfs/extent_io.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index c123a3ef154a..755ec6dfd51c 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -973,7 +973,7 @@ static void btrfs_readahead_expand(struct readahead_control *ractl, { const u64 ra_pos = readahead_pos(ractl); const u64 ra_end = ra_pos + readahead_length(ractl); - const u64 em_end = em->start + em->ram_bytes; + const u64 em_end = em->start + em->len; /* No expansion for holes and inline extents. */ if (em->disk_bytenr > EXTENT_MAP_LAST_BYTE) From fec9b9d3ced39f16be8d7afdf81f4dd2653da319 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Miquel=20Sabat=C3=A9=20Sol=C3=A0?= Date: Wed, 8 Oct 2025 14:18:59 +0200 Subject: [PATCH 140/330] btrfs: fix memory leaks when rejecting a non SINGLE data profile without an RST MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit At the end of btrfs_load_block_group_zone_info() the first thing we do is to ensure that if the mapping type is not a SINGLE one and there is no RAID stripe tree, then we return early with an error. Doing that, though, prevents the code from running the last calls from this function which are about freeing memory allocated during its run. Hence, in this case, instead of returning early, we set the ret value and fall through the rest of the cleanup code. Fixes: 5906333cc4af ("btrfs: zoned: don't skip block group profile checks on conventional zones") CC: stable@vger.kernel.org # 6.8+ Reviewed-by: Johannes Thumshirn Signed-off-by: Miquel Sabaté Solà Signed-off-by: David Sterba --- fs/btrfs/zoned.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c index e3341a84f4ab..838149fa60ce 100644 --- a/fs/btrfs/zoned.c +++ b/fs/btrfs/zoned.c @@ -1753,7 +1753,7 @@ int btrfs_load_block_group_zone_info(struct btrfs_block_group *cache, bool new) !fs_info->stripe_root) { btrfs_err(fs_info, "zoned: data %s needs raid-stripe-tree", btrfs_bg_type_to_raid_name(map->type)); - return -EINVAL; + ret = -EINVAL; } if (unlikely(cache->alloc_offset > cache->zone_capacity)) { From e92c2941204de7b62e9c2deecfeb9eaefe54a22a Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Wed, 8 Oct 2025 18:08:58 +0300 Subject: [PATCH 141/330] btrfs: tree-checker: fix bounds check in check_inode_extref() The parentheses for the unlikely() annotation were put in the wrong place so it means that the condition is basically never true and the bounds checking is skipped. Fixes: aab9458b9f00 ("btrfs: tree-checker: add inode extref checks") Signed-off-by: Dan Carpenter Reviewed-by: Qu Wenruo Signed-off-by: David Sterba --- fs/btrfs/tree-checker.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/btrfs/tree-checker.c b/fs/btrfs/tree-checker.c index ca30b15ea452..c10b4c242acf 100644 --- a/fs/btrfs/tree-checker.c +++ b/fs/btrfs/tree-checker.c @@ -1797,7 +1797,7 @@ static int check_inode_extref(struct extent_buffer *leaf, struct btrfs_inode_extref *extref = (struct btrfs_inode_extref *)ptr; u16 namelen; - if (unlikely(ptr + sizeof(*extref)) > end) { + if (unlikely(ptr + sizeof(*extref) > end)) { inode_ref_err(leaf, slot, "inode extref overflow, ptr %lu end %lu inode_extref size %zu", ptr, end, sizeof(*extref)); From 8aec9dbf2db2e958de5bd20e23b8fbb8f2aa1fa6 Mon Sep 17 00:00:00 2001 From: "Gustavo A. R. Silva" Date: Fri, 3 Oct 2025 15:11:06 +0100 Subject: [PATCH 142/330] btrfs: send: fix -Wflex-array-member-not-at-end warning in struct send_ctx The warning -Wflex-array-member-not-at-end was introduced in GCC-14, and we are getting ready to enable it, globally. Fix the following warning: fs/btrfs/send.c:181:24: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end] and move the declaration of send_ctx::cur_inode_path to the end. Notice that struct fs_path contains a flexible array member inline_buf, but also a padding array and a limit calculated for the usable space of inline_buf (FS_PATH_INLINE_SIZE). It is not the pattern where flexible array is in the middle of a structure and could potentially overwrite other members. Signed-off-by: Gustavo A. R. Silva Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/send.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index 9230e5066fc6..6144e66661f5 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -178,7 +178,6 @@ struct send_ctx { u64 cur_inode_rdev; u64 cur_inode_last_extent; u64 cur_inode_next_write_offset; - struct fs_path cur_inode_path; bool cur_inode_new; bool cur_inode_new_gen; bool cur_inode_deleted; @@ -305,6 +304,9 @@ struct send_ctx { struct btrfs_lru_cache dir_created_cache; struct btrfs_lru_cache dir_utimes_cache; + + /* Must be last as it ends in a flexible-array member. */ + struct fs_path cur_inode_path; }; struct pending_dir_move { From a375246fcf2bbdaeb1df7fa7ee5a8b884a89085e Mon Sep 17 00:00:00 2001 From: Dave Jiang Date: Thu, 9 Oct 2025 08:40:01 -0700 Subject: [PATCH 143/330] cxl/features: Add check for no entries in cxl_feature_info cxl EDAC calls cxl_feature_info() to get the feature information and if the hardware has no Features support, cxlfs may be passed in as NULL. [ 51.957498] BUG: kernel NULL pointer dereference, address: 0000000000000008 [ 51.965571] #PF: supervisor read access in kernel mode [ 51.971559] #PF: error_code(0x0000) - not-present page [ 51.977542] PGD 17e4f6067 P4D 0 [ 51.981384] Oops: Oops: 0000 [#1] SMP NOPTI [ 51.986300] CPU: 49 UID: 0 PID: 3782 Comm: systemd-udevd Not tainted 6.17.0dj test+ #64 PREEMPT(voluntary) [ 51.997355] Hardware name: [ 52.009790] RIP: 0010:cxl_feature_info+0xa/0x80 [cxl_core] Add a check for cxlfs before dereferencing it and return -EOPNOTSUPP if there is no cxlfs created due to no hardware support. Fixes: eb5dfcb9e36d ("cxl: Add support to handle user feature commands for set feature") Reviewed-by: Davidlohr Bueso Reviewed-by: Alison Schofield Signed-off-by: Dave Jiang --- drivers/cxl/core/features.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/cxl/core/features.c b/drivers/cxl/core/features.c index 7c750599ea69..4bc484b46f43 100644 --- a/drivers/cxl/core/features.c +++ b/drivers/cxl/core/features.c @@ -371,6 +371,9 @@ cxl_feature_info(struct cxl_features_state *cxlfs, { struct cxl_feat_entry *feat; + if (!cxlfs || !cxlfs->entries) + return ERR_PTR(-EOPNOTSUPP); + for (int i = 0; i < cxlfs->entries->num_features; i++) { feat = &cxlfs->entries->ent[i]; if (uuid_equal(uuid, &feat->uuid)) From f25785f9b088ed65089dd0d0034da52858417839 Mon Sep 17 00:00:00 2001 From: Rik van Riel Date: Sun, 5 Oct 2025 23:48:05 -0400 Subject: [PATCH 144/330] x86/mm: Fix overflow in __cpa_addr() The change to have cpa_flush() call flush_kernel_pages() introduced a bug where __cpa_addr() can access an address one larger than the largest one in the cpa->pages array. KASAN reports the issue like this: BUG: KASAN: slab-out-of-bounds in __cpa_addr arch/x86/mm/pat/set_memory.c:309 [inline] BUG: KASAN: slab-out-of-bounds in __cpa_addr+0x1d3/0x220 arch/x86/mm/pat/set_memory.c:306 Read of size 8 at addr ffff88801f75e8f8 by task syz.0.17/5978 This bug could cause cpa_flush() to not properly flush memory, which somehow never showed any symptoms in my tests, possibly because cpa_flush() is called so rarely, but could potentially cause issues for other people. Fix the issue by directly calculating the flush end address from the start address. Fixes: 86e6815b316e ("x86/mm: Change cpa_flush() to call flush_kernel_range() directly") Reported-by: syzbot+afec6555eef563c66c97@syzkaller.appspotmail.com Signed-off-by: Rik van Riel Signed-off-by: Dave Hansen Reviewed-by: Kiryl Shutsemau Link: https://lore.kernel.org/all/68e2ff90.050a0220.2c17c1.0038.GAE@google.com/ --- arch/x86/mm/pat/set_memory.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/x86/mm/pat/set_memory.c b/arch/x86/mm/pat/set_memory.c index d2d54b8c4dbb..970981893c9b 100644 --- a/arch/x86/mm/pat/set_memory.c +++ b/arch/x86/mm/pat/set_memory.c @@ -446,7 +446,7 @@ static void cpa_flush(struct cpa_data *cpa, int cache) } start = fix_addr(__cpa_addr(cpa, 0)); - end = fix_addr(__cpa_addr(cpa, cpa->numpages)); + end = start + cpa->numpages * PAGE_SIZE; if (cpa->force_flush_all) end = TLB_FLUSH_ALL; From 83b0177a6c4889b3a6e865da5e21b2c9d97d0551 Mon Sep 17 00:00:00 2001 From: Ingo Molnar Date: Fri, 16 May 2025 15:43:04 +0200 Subject: [PATCH 145/330] x86/mm: Fix SMP ordering in switch_mm_irqs_off() Stephen noted that it is possible to not have an smp_mb() between the loaded_mm store and the tlb_gen load in switch_mm(), meaning the ordering against flush_tlb_mm_range() goes out the window, and it becomes possible for switch_mm() to not observe a recent tlb_gen update and fail to flush the TLBs. [ dhansen: merge conflict fixed by Ingo ] Fixes: 209954cbc7d0 ("x86/mm/tlb: Update mm_cpumask lazily") Reported-by: Stephen Dolan Closes: https://lore.kernel.org/all/CAHDw0oGd0B4=uuv8NGqbUQ_ZVmSheU2bN70e4QhFXWvuAZdt2w@mail.gmail.com/ Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Ingo Molnar Signed-off-by: Dave Hansen --- arch/x86/mm/tlb.c | 24 ++++++++++++++++++++++-- 1 file changed, 22 insertions(+), 2 deletions(-) diff --git a/arch/x86/mm/tlb.c b/arch/x86/mm/tlb.c index 39f80111e6f1..5d221709353e 100644 --- a/arch/x86/mm/tlb.c +++ b/arch/x86/mm/tlb.c @@ -911,11 +911,31 @@ void switch_mm_irqs_off(struct mm_struct *unused, struct mm_struct *next, * CR3 and cpu_tlbstate.loaded_mm are not all in sync. */ this_cpu_write(cpu_tlbstate.loaded_mm, LOADED_MM_SWITCHING); - barrier(); - /* Start receiving IPIs and then read tlb_gen (and LAM below) */ + /* + * Make sure this CPU is set in mm_cpumask() such that we'll + * receive invalidation IPIs. + * + * Rely on the smp_mb() implied by cpumask_set_cpu()'s atomic + * operation, or explicitly provide one. Such that: + * + * switch_mm_irqs_off() flush_tlb_mm_range() + * smp_store_release(loaded_mm, SWITCHING); atomic64_inc_return(tlb_gen) + * smp_mb(); // here // smp_mb() implied + * atomic64_read(tlb_gen); this_cpu_read(loaded_mm); + * + * we properly order against flush_tlb_mm_range(), where the + * loaded_mm load can happen in mative_flush_tlb_multi() -> + * should_flush_tlb(). + * + * This way switch_mm() must see the new tlb_gen or + * flush_tlb_mm_range() must see the new loaded_mm, or both. + */ if (next != &init_mm && !cpumask_test_cpu(cpu, mm_cpumask(next))) cpumask_set_cpu(cpu, mm_cpumask(next)); + else + smp_mb(); + next_tlb_gen = atomic64_read(&next->context.tlb_gen); ns = choose_new_asid(next, next_tlb_gen); From d6fc45100aa8c02be3ddd16fae569b84086c15a9 Mon Sep 17 00:00:00 2001 From: Hans Zhang <18255117159@163.com> Date: Fri, 10 Oct 2025 22:43:07 +0800 Subject: [PATCH 146/330] PCI: cadence: Search for MSI Capability with correct ID 907912c1daa7 ("PCI: cadence: Use cdns_pcie_find_*capability() to avoid hardcoding offsets") incorrectly searched for the MSI-X Capability ID instead of the MSI Capability ID in cdns_pcie_ep_get_msi(). Search for PCI_CAP_ID_MSI, not PCI_CAP_ID_MSIX, to fix this problem. Fixes: 907912c1daa7 ("PCI: cadence: Use cdns_pcie_find_*capability() to avoid hardcoding offsets") Reported-by: Sasha Levin Closes: https://lore.kernel.org/r/aOfMk9BW8BH2P30V@laps/ Signed-off-by: Hans Zhang <18255117159@163.com> Signed-off-by: Bjorn Helgaas Link: https://patch.msgid.link/20251010144307.12979-1-18255117159@163.com --- drivers/pci/controller/cadence/pcie-cadence-ep.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/controller/cadence/pcie-cadence-ep.c b/drivers/pci/controller/cadence/pcie-cadence-ep.c index 1eac012a8226..c0e1194a936b 100644 --- a/drivers/pci/controller/cadence/pcie-cadence-ep.c +++ b/drivers/pci/controller/cadence/pcie-cadence-ep.c @@ -255,7 +255,7 @@ static int cdns_pcie_ep_get_msi(struct pci_epc *epc, u8 fn, u8 vfn) u16 flags, mme; u8 cap; - cap = cdns_pcie_find_capability(pcie, PCI_CAP_ID_MSIX); + cap = cdns_pcie_find_capability(pcie, PCI_CAP_ID_MSI); fn = cdns_pcie_get_fn_from_vfn(pcie, fn, vfn); /* Validate that the MSI feature is actually enabled. */ From 1ee889fdf409ce68c1e3b62912333a5cc69acaa0 Mon Sep 17 00:00:00 2001 From: Mateusz Guzik Date: Wed, 1 Oct 2025 01:29:57 +0200 Subject: [PATCH 147/330] f2fs: don't call iput() from f2fs_drop_inode() iput() calls the problematic routine, which does a ->i_count inc/dec cycle. Undoing it with iput() recurses into the problem. Note f2fs should not be playing games with the refcount to begin with, but that will be handled later. Right now solve the immediate regression. Fixes: bc986b1d756482a ("fs: stop accessing ->i_count directly in f2fs and gfs2") Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-lkp/202509301450.138b448f-lkp@intel.com Signed-off-by: Mateusz Guzik Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index fd8e7b0b2166..db7afb806411 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -1820,7 +1820,7 @@ static int f2fs_drop_inode(struct inode *inode) sb_end_intwrite(inode->i_sb); spin_lock(&inode->i_lock); - iput(inode); + atomic_dec(&inode->i_count); } trace_f2fs_drop_inode(inode, 0); return 0; From 9d5c4f5c7a2c7677e1b3942772122b032c265aae Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Tue, 7 Oct 2025 03:32:30 +0000 Subject: [PATCH 148/330] f2fs: fix wrong block mapping for multi-devices Assuming the disk layout as below, disk0: 0 --- 0x00035abfff disk1: 0x00035ac000 --- 0x00037abfff disk2: 0x00037ac000 --- 0x00037ebfff and we want to read data from offset=13568 having len=128 across the block devices, we can illustrate the block addresses like below. 0 .. 0x00037ac000 ------------------- 0x00037ebfff, 0x00037ec000 ------- | ^ ^ ^ | fofs 0 13568 13568+128 | ------------------------------------------------------ | LBA 0x37e8aa9 0x37ebfa9 0x37ec029 --- map 0x3caa9 0x3ffa9 In this example, we should give the relative map of the target block device ranging from 0x3caa9 to 0x3ffa9 where the length should be calculated by 0x37ebfff + 1 - 0x37ebfa9. In the below equation, however, map->m_pblk was supposed to be the original address instead of the one from the target block address. - map->m_len = min(map->m_len, dev->end_blk + 1 - map->m_pblk); Cc: stable@vger.kernel.org Fixes: 71f2c8206202 ("f2fs: multidevice: support direct IO") Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index ef38e62cda8f..775aa4f63aa3 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1497,8 +1497,8 @@ static bool f2fs_map_blocks_cached(struct inode *inode, struct f2fs_dev_info *dev = &sbi->devs[bidx]; map->m_bdev = dev->bdev; - map->m_pblk -= dev->start_blk; map->m_len = min(map->m_len, dev->end_blk + 1 - map->m_pblk); + map->m_pblk -= dev->start_blk; } else { map->m_bdev = inode->i_sb->s_bdev; } From ca88ecdce5f51874a7c151809bd2c936ee0d3805 Mon Sep 17 00:00:00 2001 From: Marc Zyngier Date: Wed, 8 Oct 2025 22:35:15 +0100 Subject: [PATCH 149/330] arm64: Revamp HCR_EL2.E2H RES1 detection We currently have two ways to identify CPUs that only implement FEAT_VHE and not FEAT_E2H0: - either they advertise it via ID_AA64MMFR4_EL1.E2H0, - or the HCR_EL2.E2H bit is RAO/WI However, there is a third category of "cpus" that fall between these two cases: on CPUs that do not implement FEAT_FGT, it is IMPDEF whether an access to ID_AA64MMFR4_EL1 can trap to EL2 when the register value is zero. A consequence of this is that on systems such as Neoverse V2, a NV guest cannot reliably detect that it is in a VHE-only configuration (E2H is writable, and ID_AA64MMFR0_EL1 is 0), despite the hypervisor's best effort to repaint the id register. Replace the RAO/WI test by a sequence that makes use of the VHE register remnapping between EL1 and EL2 to detect this situation, and work out whether we get the VHE behaviour even after having set HCR_EL2.E2H to 0. This solves the NV problem, and provides a more reliable acid test for CPUs that do not completely follow the letter of the architecture while providing a RES1 behaviour for HCR_EL2.E2H. Suggested-by: Mark Rutland Acked-by: Mark Rutland Acked-by: Catalin Marinas Reviewed-by: Oliver Upton Tested-by: Jan Kotas Signed-off-by: Marc Zyngier Link: https://lore.kernel.org/r/15A85F2B-1A0C-4FA7-9FE4-EEC2203CC09E@global.cadence.com --- arch/arm64/include/asm/el2_setup.h | 38 +++++++++++++++++++++++++----- 1 file changed, 32 insertions(+), 6 deletions(-) diff --git a/arch/arm64/include/asm/el2_setup.h b/arch/arm64/include/asm/el2_setup.h index b37da3ee8529..99a7c0235e6d 100644 --- a/arch/arm64/include/asm/el2_setup.h +++ b/arch/arm64/include/asm/el2_setup.h @@ -24,22 +24,48 @@ * ID_AA64MMFR4_EL1.E2H0 < 0. On such CPUs HCR_EL2.E2H is RES1, but it * can reset into an UNKNOWN state and might not read as 1 until it has * been initialized explicitly. - * - * Fruity CPUs seem to have HCR_EL2.E2H set to RAO/WI, but - * don't advertise it (they predate this relaxation). - * * Initalize HCR_EL2.E2H so that later code can rely upon HCR_EL2.E2H * indicating whether the CPU is running in E2H mode. */ mrs_s x1, SYS_ID_AA64MMFR4_EL1 sbfx x1, x1, #ID_AA64MMFR4_EL1_E2H0_SHIFT, #ID_AA64MMFR4_EL1_E2H0_WIDTH cmp x1, #0 - b.ge .LnVHE_\@ + b.lt .LnE2H0_\@ + /* + * Unfortunately, HCR_EL2.E2H can be RES1 even if not advertised + * as such via ID_AA64MMFR4_EL1.E2H0: + * + * - Fruity CPUs predate the !FEAT_E2H0 relaxation, and seem to + * have HCR_EL2.E2H implemented as RAO/WI. + * + * - On CPUs that lack FEAT_FGT, a hypervisor can't trap guest + * reads of ID_AA64MMFR4_EL1 to advertise !FEAT_E2H0. NV + * guests on these hosts can write to HCR_EL2.E2H without + * trapping to the hypervisor, but these writes have no + * functional effect. + * + * Handle both cases by checking for an essential VHE property + * (system register remapping) to decide whether we're + * effectively VHE-only or not. + */ + msr_hcr_el2 x0 // Setup HCR_EL2 as nVHE + isb + mov x1, #1 // Write something to FAR_EL1 + msr far_el1, x1 + isb + mov x1, #2 // Try to overwrite it via FAR_EL2 + msr far_el2, x1 + isb + mrs x1, far_el1 // If we see the latest write in FAR_EL1, + cmp x1, #2 // we can safely assume we are VHE only. + b.ne .LnVHE_\@ // Otherwise, we know that nVHE works. + +.LnE2H0_\@: orr x0, x0, #HCR_E2H -.LnVHE_\@: msr_hcr_el2 x0 isb +.LnVHE_\@: .endm .macro __init_el2_sctlr From 095232711f23179053ca26bcf046ca121a91a465 Mon Sep 17 00:00:00 2001 From: Francesco Valla Date: Fri, 3 Oct 2025 12:33:03 +0200 Subject: [PATCH 150/330] drm/draw: fix color truncation in drm_draw_fill24 The color parameter passed to drm_draw_fill24() was truncated to 16 bits, leading to an incorrect color drawn to the target iosys_map. Fix this behavior, widening the parameter to 32 bits. Fixes: 31fa2c1ca0b2 ("drm/panic: Move drawing functions to drm_draw") Signed-off-by: Francesco Valla Reviewed-by: Jocelyn Falempe Link: https://lore.kernel.org/r/20251003-drm_draw_fill24_fix-v1-1-8fb7c1c2a893@valla.it Signed-off-by: Jocelyn Falempe --- drivers/gpu/drm/drm_draw.c | 2 +- drivers/gpu/drm/drm_draw_internal.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/drm_draw.c b/drivers/gpu/drm/drm_draw.c index 9dc0408fbbea..5b956229c82f 100644 --- a/drivers/gpu/drm/drm_draw.c +++ b/drivers/gpu/drm/drm_draw.c @@ -127,7 +127,7 @@ EXPORT_SYMBOL(drm_draw_fill16); void drm_draw_fill24(struct iosys_map *dmap, unsigned int dpitch, unsigned int height, unsigned int width, - u16 color) + u32 color) { unsigned int y, x; diff --git a/drivers/gpu/drm/drm_draw_internal.h b/drivers/gpu/drm/drm_draw_internal.h index f121ee7339dc..20cb404e23ea 100644 --- a/drivers/gpu/drm/drm_draw_internal.h +++ b/drivers/gpu/drm/drm_draw_internal.h @@ -47,7 +47,7 @@ void drm_draw_fill16(struct iosys_map *dmap, unsigned int dpitch, void drm_draw_fill24(struct iosys_map *dmap, unsigned int dpitch, unsigned int height, unsigned int width, - u16 color); + u32 color); void drm_draw_fill32(struct iosys_map *dmap, unsigned int dpitch, unsigned int height, unsigned int width, From 7f38a1487555604bc4e210fa7cc9b1bce981c40e Mon Sep 17 00:00:00 2001 From: Alok Tiwari Date: Sun, 12 Oct 2025 07:20:01 -0700 Subject: [PATCH 151/330] drm/rockchip: vop2: use correct destination rectangle height check The vop2_plane_atomic_check() function incorrectly checks drm_rect_width(dest) twice instead of verifying both width and height. Fix the second condition to use drm_rect_height(dest) so that invalid destination rectangles with height < 4 are correctly rejected. Fixes: 604be85547ce ("drm/rockchip: Add VOP2 driver") Signed-off-by: Alok Tiwari Reviewed-by: Andy Yan Signed-off-by: Heiko Stuebner Link: https://lore.kernel.org/r/20251012142005.660727-1-alok.a.tiwari@oracle.com --- drivers/gpu/drm/rockchip/rockchip_drm_vop2.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c index b50927a824b4..7ec7bea5e38e 100644 --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.c @@ -1031,7 +1031,7 @@ static int vop2_plane_atomic_check(struct drm_plane *plane, return format; if (drm_rect_width(src) >> 16 < 4 || drm_rect_height(src) >> 16 < 4 || - drm_rect_width(dest) < 4 || drm_rect_width(dest) < 4) { + drm_rect_width(dest) < 4 || drm_rect_height(dest) < 4) { drm_err(vop2->drm, "Invalid size: %dx%d->%dx%d, min size is 4x4\n", drm_rect_width(src) >> 16, drm_rect_height(src) >> 16, drm_rect_width(dest), drm_rect_height(dest)); From 62685ab071de7c39499212bff19f1b5bc0148bc7 Mon Sep 17 00:00:00 2001 From: Jiri Olsa Date: Wed, 1 Oct 2025 15:24:49 +0200 Subject: [PATCH 152/330] uprobe: Move arch_uprobe_optimize right after handlers execution It's less confusing to optimize uprobe right after handlers execution and before we do the check for changed ip register to avoid situations where changed ip register would skip uprobe optimization. Suggested-by: Linus Torvalds Signed-off-by: Jiri Olsa Signed-off-by: Peter Zijlstra (Intel) Reviewed-by: Masami Hiramatsu (Google) Acked-by: Andrii Nakryiko Acked-by: Oleg Nesterov --- kernel/events/uprobes.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c index 8709c69118b5..f11ceb8be8c4 100644 --- a/kernel/events/uprobes.c +++ b/kernel/events/uprobes.c @@ -2765,6 +2765,9 @@ static void handle_swbp(struct pt_regs *regs) handler_chain(uprobe, regs); + /* Try to optimize after first hit. */ + arch_uprobe_optimize(&uprobe->arch, bp_vaddr); + /* * If user decided to take execution elsewhere, it makes little sense * to execute the original instruction, so let's skip it. @@ -2772,9 +2775,6 @@ static void handle_swbp(struct pt_regs *regs) if (instruction_pointer(regs) != bp_vaddr) goto out; - /* Try to optimize after first hit. */ - arch_uprobe_optimize(&uprobe->arch, bp_vaddr); - if (arch_uprobe_skip_sstep(&uprobe->arch, regs)) goto out; From ebfc8542ad62d066771e46c8aa30f5624b89cad8 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 13 Oct 2025 10:22:42 +0300 Subject: [PATCH 153/330] perf/core: Fix address filter match with backing files It was reported that Intel PT address filters do not work in Docker containers. That relates to the use of overlayfs. overlayfs records the backing file in struct vm_area_struct vm_file, instead of the user file that the user mmapped. In order for an address filter to match, it must compare to the user file inode. There is an existing helper file_user_inode() for that situation. Use file_user_inode() instead of file_inode() to get the inode for address filter matching. Example: Setup: # cd /root # mkdir test ; cd test ; mkdir lower upper work merged # cp `which cat` lower # mount -t overlay overlay -olowerdir=lower,upperdir=upper,workdir=work merged # perf record --buildid-mmap -e intel_pt//u --filter 'filter * @ /root/test/merged/cat' -- /root/test/merged/cat /proc/self/maps ... 55d61d246000-55d61d2e1000 r-xp 00018000 00:1a 3418 /root/test/merged/cat ... [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.015 MB perf.data ] # perf buildid-cache --add /root/test/merged/cat Before: Address filter does not match so there are no control flow packets # perf script --itrace=e # perf script --itrace=b | wc -l 0 # perf script -D | grep 'TIP.PGE' | wc -l 0 # After: Address filter does match so there are control flow packets # perf script --itrace=e # perf script --itrace=b | wc -l 235 # perf script -D | grep 'TIP.PGE' | wc -l 57 # With respect to stable kernels, overlayfs mmap function ovl_mmap() was added in v4.19 but file_user_inode() was not added until v6.8 and never back-ported to stable kernels. FMODE_BACKING that it depends on was added in v6.5. This issue has gone largely unnoticed, so back-porting before v6.8 is probably not worth it, so put 6.8 as the stable kernel prerequisite version, although in practice the next long term kernel is 6.12. Closes: https://lore.kernel.org/linux-perf-users/aBCwoq7w8ohBRQCh@fremen.lan Reported-by: Edd Barrett Signed-off-by: Adrian Hunter Signed-off-by: Peter Zijlstra (Intel) Acked-by: Amir Goldstein Cc: stable@vger.kernel.org # 6.8 --- kernel/events/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/events/core.c b/kernel/events/core.c index 7541f6f85fcb..cd63ec84e386 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -9492,7 +9492,7 @@ static bool perf_addr_filter_match(struct perf_addr_filter *filter, if (!filter->path.dentry) return false; - if (d_inode(filter->path.dentry) != file_inode(file)) + if (d_inode(filter->path.dentry) != file_user_inode(file)) return false; if (filter->offset > offset + size) From 8818f507a9391019a3ec7c57b1a32e4b386e48a5 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 13 Oct 2025 10:22:43 +0300 Subject: [PATCH 154/330] perf/core: Fix MMAP event path names with backing files Some file systems like FUSE-based ones or overlayfs may record the backing file in struct vm_area_struct vm_file, instead of the user file that the user mmapped. Since commit def3ae83da02f ("fs: store real path instead of fake path in backing file f_path"), file_path() no longer returns the user file path when applied to a backing file. There is an existing helper file_user_path() for that situation. Use file_user_path() instead of file_path() to get the path for MMAP and MMAP2 events. Example: Setup: # cd /root # mkdir test ; cd test ; mkdir lower upper work merged # cp `which cat` lower # mount -t overlay overlay -olowerdir=lower,upperdir=upper,workdir=work merged # perf record -e intel_pt//u -- /root/test/merged/cat /proc/self/maps ... 55b0ba399000-55b0ba434000 r-xp 00018000 00:1a 3419 /root/test/merged/cat ... [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.060 MB perf.data ] # Before: File name is wrong (/cat), so decoding fails: # perf script --no-itrace --show-mmap-events cat 367 [016] 100.491492: PERF_RECORD_MMAP2 367/367: [0x55b0ba399000(0x9b000) @ 0x18000 00:02 3419 489959280]: r-xp /cat ... # perf script --itrace=e | wc -l Warning: 19 instruction trace errors 19 # After: File name is correct (/root/test/merged/cat), so decoding is ok: # perf script --no-itrace --show-mmap-events cat 364 [016] 72.153006: PERF_RECORD_MMAP2 364/364: [0x55ce4003d000(0x9b000) @ 0x18000 00:02 3419 3132534314]: r-xp /root/test/merged/cat # perf script --itrace=e # perf script --itrace=e | wc -l 0 # Fixes: def3ae83da02f ("fs: store real path instead of fake path in backing file f_path") Signed-off-by: Adrian Hunter Signed-off-by: Peter Zijlstra (Intel) Acked-by: Amir Goldstein Cc: stable@vger.kernel.org --- kernel/events/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/events/core.c b/kernel/events/core.c index cd63ec84e386..7b5c2373a8d7 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -9416,7 +9416,7 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) * need to add enough zero bytes after the string to handle * the 64bit alignment we do later. */ - name = file_path(file, buf, PATH_MAX - sizeof(u64)); + name = d_path(file_user_path(file), buf, PATH_MAX - sizeof(u64)); if (IS_ERR(name)) { name = "//toolong"; goto cpy_name; From fa4f4bae893fbce8a3edfff1ab7ece0c01dc1328 Mon Sep 17 00:00:00 2001 From: Adrian Hunter Date: Mon, 13 Oct 2025 10:22:44 +0300 Subject: [PATCH 155/330] perf/core: Fix MMAP2 event device with backing files Some file systems like FUSE-based ones or overlayfs may record the backing file in struct vm_area_struct vm_file, instead of the user file that the user mmapped. That causes perf to misreport the device major/minor numbers of the file system of the file, and the generation of the file, and potentially other inode details. There is an existing helper file_user_inode() for that situation. Use file_user_inode() instead of file_inode() to get the inode for MMAP2 events. Example: Setup: # cd /root # mkdir test ; cd test ; mkdir lower upper work merged # cp `which cat` lower # mount -t overlay overlay -olowerdir=lower,upperdir=upper,workdir=work merged # perf record -e cycles:u -- /root/test/merged/cat /proc/self/maps ... 55b2c91d0000-55b2c926b000 r-xp 00018000 00:1a 3419 /root/test/merged/cat ... [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.004 MB perf.data (5 samples) ] # # stat /root/test/merged/cat File: /root/test/merged/cat Size: 1127792 Blocks: 2208 IO Block: 4096 regular file Device: 0,26 Inode: 3419 Links: 1 Access: (0755/-rwxr-xr-x) Uid: ( 0/ root) Gid: ( 0/ root) Access: 2025-09-08 12:23:59.453309624 +0000 Modify: 2025-09-08 12:23:59.454309624 +0000 Change: 2025-09-08 12:23:59.454309624 +0000 Birth: 2025-09-08 12:23:59.453309624 +0000 Before: Device reported 00:02 differs from stat output and /proc/self/maps # perf script --show-mmap-events | grep /root/test/merged/cat cat 377 [-01] 243.078558: PERF_RECORD_MMAP2 377/377: [0x55b2c91d0000(0x9b000) @ 0x18000 00:02 3419 2068525940]: r-xp /root/test/merged/cat After: Device reported 00:1a is the same as stat output and /proc/self/maps # perf script --show-mmap-events | grep /root/test/merged/cat cat 362 [-01] 127.755167: PERF_RECORD_MMAP2 362/362: [0x55ba6e781000(0x9b000) @ 0x18000 00:1a 3419 0]: r-xp /root/test/merged/cat With respect to stable kernels, overlayfs mmap function ovl_mmap() was added in v4.19 but file_user_inode() was not added until v6.8 and never back-ported to stable kernels. FMODE_BACKING that it depends on was added in v6.5. This issue has gone largely unnoticed, so back-porting before v6.8 is probably not worth it, so put 6.8 as the stable kernel prerequisite version, although in practice the next long term kernel is 6.12. Signed-off-by: Adrian Hunter Signed-off-by: Peter Zijlstra (Intel) Acked-by: Amir Goldstein Cc: stable@vger.kernel.org # 6.8 --- kernel/events/core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/events/core.c b/kernel/events/core.c index 7b5c2373a8d7..177e57c1a362 100644 --- a/kernel/events/core.c +++ b/kernel/events/core.c @@ -9403,7 +9403,7 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) flags |= MAP_HUGETLB; if (file) { - struct inode *inode; + const struct inode *inode; dev_t dev; buf = kmalloc(PATH_MAX, GFP_KERNEL); @@ -9421,7 +9421,7 @@ static void perf_event_mmap_event(struct perf_mmap_event *mmap_event) name = "//toolong"; goto cpy_name; } - inode = file_inode(vma->vm_file); + inode = file_user_inode(vma->vm_file); dev = inode->i_sb->s_dev; ino = inode->i_ino; gen = inode->i_generation; From 6c26c055523d915afb8d18e7277848eff66a3085 Mon Sep 17 00:00:00 2001 From: Xinpeng Sun Date: Thu, 9 Oct 2025 11:31:08 +0800 Subject: [PATCH 156/330] HID: intel-thc-hid: intel-quicki2c: Fix wrong type casting The type definition of qcdev->i2c_max_frame_size is already u32, so remove the unnecessary type casting le16_to_cpu. Signed-off-by: Xinpeng Sun Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202509280841.pxmgBzKW-lkp@intel.com/ Signed-off-by: Jiri Kosina --- drivers/hid/intel-thc-hid/intel-quicki2c/pci-quicki2c.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/hid/intel-thc-hid/intel-quicki2c/pci-quicki2c.c b/drivers/hid/intel-thc-hid/intel-quicki2c/pci-quicki2c.c index 8433a991e7f4..0156ab391778 100644 --- a/drivers/hid/intel-thc-hid/intel-quicki2c/pci-quicki2c.c +++ b/drivers/hid/intel-thc-hid/intel-quicki2c/pci-quicki2c.c @@ -466,7 +466,7 @@ static void quicki2c_dma_adv_enable(struct quicki2c_device *qcdev) dev_warn(qcdev->dev, "Max frame size is smaller than hid max input length!"); thc_i2c_set_rx_max_size(qcdev->thc_hw, - le16_to_cpu(qcdev->i2c_max_frame_size)); + qcdev->i2c_max_frame_size); } thc_i2c_rx_max_size_enable(qcdev->thc_hw, true); } From 8fe2cd8ec84b3592b57f40b080f9d5aeebd553af Mon Sep 17 00:00:00 2001 From: Even Xu Date: Fri, 19 Sep 2025 15:09:39 +0800 Subject: [PATCH 157/330] HID: intel-thc-hid: Intel-quickspi: switch first interrupt from level to edge detection The original implementation used level detection for the first interrupt after device reset to avoid potential interrupt line noise and missed interrupts during the initialization phase. However, this approach introduced unintended side effects when tested with certain touch panels, including: - Delayed hardware interrupt response - Multiple spurious interrupt triggers Switching back to edge detection for the first interrupt resolves these issues while maintaining reliable interrupt handling. Extensive testing across multiple platforms with touch panels from various vendors confirms this change introduces no regressions. [jkosina@suse.com: properly capitalize shortlog] Fixes: 9d8d51735a3a ("HID: intel-thc-hid: intel-quickspi: Add HIDSPI protocol implementation") Tested-by: Rui Zhang Signed-off-by: Even Xu Signed-off-by: Jiri Kosina --- drivers/hid/intel-thc-hid/intel-quickspi/quickspi-protocol.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/hid/intel-thc-hid/intel-quickspi/quickspi-protocol.c b/drivers/hid/intel-thc-hid/intel-quickspi/quickspi-protocol.c index e6ba2ddcc9cb..16f780bc879b 100644 --- a/drivers/hid/intel-thc-hid/intel-quickspi/quickspi-protocol.c +++ b/drivers/hid/intel-thc-hid/intel-quickspi/quickspi-protocol.c @@ -280,8 +280,7 @@ int reset_tic(struct quickspi_device *qsdev) qsdev->reset_ack = false; - /* First interrupt uses level trigger to avoid missing interrupt */ - thc_int_trigger_type_select(qsdev->thc_hw, false); + thc_int_trigger_type_select(qsdev->thc_hw, true); ret = acpi_tic_reset(qsdev); if (ret) From 50f1f782f8d621a90108340c632bcb6ab4307d2e Mon Sep 17 00:00:00 2001 From: Abhishek Tamboli Date: Wed, 24 Sep 2025 10:07:20 +0530 Subject: [PATCH 158/330] HID: intel-thc-hid: intel-quickspi: Add ARL PCI Device Id's Add the missing PCI ID for the quickspi device used on the Lenovo Yoga Pro 9i 16IAH10. Buglink: https://bugzilla.kernel.org/show_bug.cgi?id=220567 Signed-off-by: Abhishek Tamboli Reviewed-by: Even Xu Signed-off-by: Jiri Kosina --- drivers/hid/intel-thc-hid/intel-quickspi/pci-quickspi.c | 6 ++++++ drivers/hid/intel-thc-hid/intel-quickspi/quickspi-dev.h | 2 ++ 2 files changed, 8 insertions(+) diff --git a/drivers/hid/intel-thc-hid/intel-quickspi/pci-quickspi.c b/drivers/hid/intel-thc-hid/intel-quickspi/pci-quickspi.c index 84314989dc53..14cabd5dc6dd 100644 --- a/drivers/hid/intel-thc-hid/intel-quickspi/pci-quickspi.c +++ b/drivers/hid/intel-thc-hid/intel-quickspi/pci-quickspi.c @@ -33,6 +33,10 @@ struct quickspi_driver_data ptl = { .max_packet_size_value = MAX_PACKET_SIZE_VALUE_LNL, }; +struct quickspi_driver_data arl = { + .max_packet_size_value = MAX_PACKET_SIZE_VALUE_MTL, +}; + /* THC QuickSPI ACPI method to get device properties */ /* HIDSPI Method: {6e2ac436-0fcf-41af-a265-b32a220dcfab} */ static guid_t hidspi_guid = @@ -978,6 +982,8 @@ static const struct pci_device_id quickspi_pci_tbl[] = { {PCI_DEVICE_DATA(INTEL, THC_PTL_U_DEVICE_ID_SPI_PORT2, &ptl), }, {PCI_DEVICE_DATA(INTEL, THC_WCL_DEVICE_ID_SPI_PORT1, &ptl), }, {PCI_DEVICE_DATA(INTEL, THC_WCL_DEVICE_ID_SPI_PORT2, &ptl), }, + {PCI_DEVICE_DATA(INTEL, THC_ARL_DEVICE_ID_SPI_PORT1, &arl), }, + {PCI_DEVICE_DATA(INTEL, THC_ARL_DEVICE_ID_SPI_PORT2, &arl), }, {} }; MODULE_DEVICE_TABLE(pci, quickspi_pci_tbl); diff --git a/drivers/hid/intel-thc-hid/intel-quickspi/quickspi-dev.h b/drivers/hid/intel-thc-hid/intel-quickspi/quickspi-dev.h index f3532d866749..c30e1a42eb09 100644 --- a/drivers/hid/intel-thc-hid/intel-quickspi/quickspi-dev.h +++ b/drivers/hid/intel-thc-hid/intel-quickspi/quickspi-dev.h @@ -21,6 +21,8 @@ #define PCI_DEVICE_ID_INTEL_THC_PTL_U_DEVICE_ID_SPI_PORT2 0xE44B #define PCI_DEVICE_ID_INTEL_THC_WCL_DEVICE_ID_SPI_PORT1 0x4D49 #define PCI_DEVICE_ID_INTEL_THC_WCL_DEVICE_ID_SPI_PORT2 0x4D4B +#define PCI_DEVICE_ID_INTEL_THC_ARL_DEVICE_ID_SPI_PORT1 0x7749 +#define PCI_DEVICE_ID_INTEL_THC_ARL_DEVICE_ID_SPI_PORT2 0x774B /* HIDSPI special ACPI parameters DSM methods */ #define ACPI_QUICKSPI_REVISION_NUM 2 From 362f21536966d7039da1de762f28f4ad44565acc Mon Sep 17 00:00:00 2001 From: Deepak Sharma Date: Fri, 26 Sep 2025 20:28:11 +0530 Subject: [PATCH 159/330] HID: cp2112: Add parameter validation to data length Syzkaller reported a stack OOB access in cp2112_write_req caused by lack of parameter validation for the user input in I2C SMBUS ioctl in cp2112 driver Add the parameter validation for the data->block[0] to be bounded by I2C_SMBUS_BLOCK_MAX + the additional compatibility padding [jkosina@suse.com: fix whitespace damage] Reported-by: syzbot+7617e19c8a59edfbd879@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=7617e19c8a59edfbd879 Tested-by: syzbot+7617e19c8a59edfbd879@syzkaller.appspotmail.com Signed-off-by: Deepak Sharma Signed-off-by: Jiri Kosina --- drivers/hid/hid-cp2112.c | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/drivers/hid/hid-cp2112.c b/drivers/hid/hid-cp2112.c index 5a95ea3bec98..803b883ae875 100644 --- a/drivers/hid/hid-cp2112.c +++ b/drivers/hid/hid-cp2112.c @@ -689,7 +689,14 @@ static int cp2112_xfer(struct i2c_adapter *adap, u16 addr, count = cp2112_write_read_req(buf, addr, read_length, command, NULL, 0); } else { - count = cp2112_write_req(buf, addr, command, + /* Copy starts from data->block[1] so the length can + * be at max I2C_SMBUS_CLOCK_MAX + 1 + */ + + if (data->block[0] > I2C_SMBUS_BLOCK_MAX + 1) + count = -EINVAL; + else + count = cp2112_write_req(buf, addr, command, data->block + 1, data->block[0]); } @@ -700,7 +707,14 @@ static int cp2112_xfer(struct i2c_adapter *adap, u16 addr, I2C_SMBUS_BLOCK_MAX, command, NULL, 0); } else { - count = cp2112_write_req(buf, addr, command, + /* data_length here is data->block[0] + 1 + * so make sure that the data->block[0] is + * less than or equals I2C_SMBUS_BLOCK_MAX + 1 + */ + if (data->block[0] > I2C_SMBUS_BLOCK_MAX + 1) + count = -EINVAL; + else + count = cp2112_write_req(buf, addr, command, data->block, data->block[0] + 1); } @@ -709,7 +723,14 @@ static int cp2112_xfer(struct i2c_adapter *adap, u16 addr, size = I2C_SMBUS_BLOCK_DATA; read_write = I2C_SMBUS_READ; - count = cp2112_write_read_req(buf, addr, I2C_SMBUS_BLOCK_MAX, + /* data_length is data->block[0] + 1, so + * so data->block[0] should be less than or + * equal to the I2C_SMBUS_BLOCK_MAX + 1 + */ + if (data->block[0] > I2C_SMBUS_BLOCK_MAX + 1) + count = -EINVAL; + else + count = cp2112_write_read_req(buf, addr, I2C_SMBUS_BLOCK_MAX, command, data->block, data->block[0] + 1); break; From 0be4253bf878d9aaa2b96031ac8683fceeb81480 Mon Sep 17 00:00:00 2001 From: Tristan Lobb Date: Sun, 28 Sep 2025 18:25:43 +0200 Subject: [PATCH 160/330] HID: quirks: avoid Cooler Master MM712 dongle wakeup bug The Cooler Master Mice Dongle includes a vendor defined HID interface alongside its mouse interface. Not polling it will cause the mouse to stop responding to polls on any interface once woken up again after going into power saving mode. Add the HID_QUIRK_ALWAYS_POLL quirk alongside the Cooler Master VID and the Dongle's PID. Signed-off-by: Tristan Lobb Signed-off-by: Jiri Kosina --- drivers/hid/hid-ids.h | 3 +++ drivers/hid/hid-quirks.c | 1 + 2 files changed, 4 insertions(+) diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index 5721b8414bbd..d05a62bbafff 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -342,6 +342,9 @@ #define USB_DEVICE_ID_CODEMERCS_IOW_FIRST 0x1500 #define USB_DEVICE_ID_CODEMERCS_IOW_LAST 0x15ff +#define USB_VENDOR_ID_COOLER_MASTER 0x2516 +#define USB_DEVICE_ID_COOLER_MASTER_MICE_DONGLE 0x01b7 + #define USB_VENDOR_ID_CORSAIR 0x1b1c #define USB_DEVICE_ID_CORSAIR_K90 0x1b02 #define USB_DEVICE_ID_CORSAIR_K70R 0x1b09 diff --git a/drivers/hid/hid-quirks.c b/drivers/hid/hid-quirks.c index ffd034566e2e..d7105a839598 100644 --- a/drivers/hid/hid-quirks.c +++ b/drivers/hid/hid-quirks.c @@ -57,6 +57,7 @@ static const struct hid_device_id hid_quirks[] = { { HID_USB_DEVICE(USB_VENDOR_ID_CH, USB_DEVICE_ID_CH_FLIGHT_SIM_YOKE), HID_QUIRK_NOGET }, { HID_USB_DEVICE(USB_VENDOR_ID_CH, USB_DEVICE_ID_CH_PRO_PEDALS), HID_QUIRK_NOGET }, { HID_USB_DEVICE(USB_VENDOR_ID_CH, USB_DEVICE_ID_CH_PRO_THROTTLE), HID_QUIRK_NOGET }, + { HID_USB_DEVICE(USB_VENDOR_ID_COOLER_MASTER, USB_DEVICE_ID_COOLER_MASTER_MICE_DONGLE), HID_QUIRK_ALWAYS_POLL }, { HID_USB_DEVICE(USB_VENDOR_ID_CORSAIR, USB_DEVICE_ID_CORSAIR_K65RGB), HID_QUIRK_NO_INIT_REPORTS }, { HID_USB_DEVICE(USB_VENDOR_ID_CORSAIR, USB_DEVICE_ID_CORSAIR_K65RGB_RAPIDFIRE), HID_QUIRK_NO_INIT_REPORTS | HID_QUIRK_ALWAYS_POLL }, { HID_USB_DEVICE(USB_VENDOR_ID_CORSAIR, USB_DEVICE_ID_CORSAIR_K70RGB), HID_QUIRK_NO_INIT_REPORTS }, From 1141ed52348d3df82d3fd2316128b3fc6203a68c Mon Sep 17 00:00:00 2001 From: Oleg Makarenko Date: Mon, 29 Sep 2025 18:46:11 +0300 Subject: [PATCH 161/330] HID: quirks: Add ALWAYS_POLL quirk for VRS R295 steering wheel This patch adds ALWAYS_POLL quirk for the VRS R295 steering wheel joystick. This device reboots itself every 8-10 seconds if it is not polled. Signed-off-by: Oleg Makarenko Signed-off-by: Jiri Kosina --- drivers/hid/hid-ids.h | 1 + drivers/hid/hid-quirks.c | 1 + 2 files changed, 2 insertions(+) diff --git a/drivers/hid/hid-ids.h b/drivers/hid/hid-ids.h index d05a62bbafff..0723b4b1c9ec 100644 --- a/drivers/hid/hid-ids.h +++ b/drivers/hid/hid-ids.h @@ -1435,6 +1435,7 @@ #define USB_VENDOR_ID_VRS 0x0483 #define USB_DEVICE_ID_VRS_DFP 0xa355 +#define USB_DEVICE_ID_VRS_R295 0xa44c #define USB_VENDOR_ID_VTL 0x0306 #define USB_DEVICE_ID_VTL_MULTITOUCH_FF3F 0xff3f diff --git a/drivers/hid/hid-quirks.c b/drivers/hid/hid-quirks.c index d7105a839598..bcd4bccf1a7c 100644 --- a/drivers/hid/hid-quirks.c +++ b/drivers/hid/hid-quirks.c @@ -207,6 +207,7 @@ static const struct hid_device_id hid_quirks[] = { { HID_USB_DEVICE(USB_VENDOR_ID_UCLOGIC, USB_DEVICE_ID_UCLOGIC_TABLET_KNA5), HID_QUIRK_MULTI_INPUT }, { HID_USB_DEVICE(USB_VENDOR_ID_UCLOGIC, USB_DEVICE_ID_UCLOGIC_TABLET_TWA60), HID_QUIRK_MULTI_INPUT }, { HID_USB_DEVICE(USB_VENDOR_ID_UGTIZER, USB_DEVICE_ID_UGTIZER_TABLET_WP5540), HID_QUIRK_MULTI_INPUT }, + { HID_USB_DEVICE(USB_VENDOR_ID_VRS, USB_DEVICE_ID_VRS_R295), HID_QUIRK_ALWAYS_POLL }, { HID_USB_DEVICE(USB_VENDOR_ID_WALTOP, USB_DEVICE_ID_WALTOP_MEDIA_TABLET_10_6_INCH), HID_QUIRK_MULTI_INPUT }, { HID_USB_DEVICE(USB_VENDOR_ID_WALTOP, USB_DEVICE_ID_WALTOP_MEDIA_TABLET_14_1_INCH), HID_QUIRK_MULTI_INPUT }, { HID_USB_DEVICE(USB_VENDOR_ID_WALTOP, USB_DEVICE_ID_WALTOP_SIRIUS_BATTERY_FREE_TABLET), HID_QUIRK_MULTI_INPUT }, From 1d64624243af8329b4b219d8c39e28ea448f9929 Mon Sep 17 00:00:00 2001 From: Vicki Pfau Date: Mon, 6 Oct 2025 18:05:31 -0700 Subject: [PATCH 162/330] HID: core: Add printk_ratelimited variants to hid_warn() etc hid_warn_ratelimited() is needed. Add the others as part of the block. Signed-off-by: Vicki Pfau Signed-off-by: Jiri Kosina --- include/linux/hid.h | 11 +++++++++++ 1 file changed, 11 insertions(+) diff --git a/include/linux/hid.h b/include/linux/hid.h index e1b673ad7457..a4ddb94e3ee5 100644 --- a/include/linux/hid.h +++ b/include/linux/hid.h @@ -1292,4 +1292,15 @@ void hid_quirks_exit(__u16 bus); #define hid_dbg_once(hid, fmt, ...) \ dev_dbg_once(&(hid)->dev, fmt, ##__VA_ARGS__) +#define hid_err_ratelimited(hid, fmt, ...) \ + dev_err_ratelimited(&(hid)->dev, fmt, ##__VA_ARGS__) +#define hid_notice_ratelimited(hid, fmt, ...) \ + dev_notice_ratelimited(&(hid)->dev, fmt, ##__VA_ARGS__) +#define hid_warn_ratelimited(hid, fmt, ...) \ + dev_warn_ratelimited(&(hid)->dev, fmt, ##__VA_ARGS__) +#define hid_info_ratelimited(hid, fmt, ...) \ + dev_info_ratelimited(&(hid)->dev, fmt, ##__VA_ARGS__) +#define hid_dbg_ratelimited(hid, fmt, ...) \ + dev_dbg_ratelimited(&(hid)->dev, fmt, ##__VA_ARGS__) + #endif From b73bc6a51f0c0066912c7e181acee41091c70fe6 Mon Sep 17 00:00:00 2001 From: Vicki Pfau Date: Mon, 6 Oct 2025 18:05:32 -0700 Subject: [PATCH 163/330] HID: nintendo: Wait longer for initial probe Some third-party controllers, such as the PB Tails CHOC, won't always respond quickly on startup. Since this packet is needed for probe, and only once during probe, let's just wait an extra second, which makes connecting consistent. Signed-off-by: Vicki Pfau Signed-off-by: Jiri Kosina --- drivers/hid/hid-nintendo.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/hid/hid-nintendo.c b/drivers/hid/hid-nintendo.c index fb4985988615..e3e54f1df44f 100644 --- a/drivers/hid/hid-nintendo.c +++ b/drivers/hid/hid-nintendo.c @@ -2420,7 +2420,7 @@ static int joycon_read_info(struct joycon_ctlr *ctlr) struct joycon_input_report *report; req.subcmd_id = JC_SUBCMD_REQ_DEV_INFO; - ret = joycon_send_subcmd(ctlr, &req, 0, HZ); + ret = joycon_send_subcmd(ctlr, &req, 0, 2 * HZ); if (ret) { hid_err(ctlr->hdev, "Failed to get joycon info; ret=%d\n", ret); return ret; From b8874720b2f33a06ff1d4cf3827e7ec1195cb360 Mon Sep 17 00:00:00 2001 From: Vicki Pfau Date: Mon, 6 Oct 2025 18:05:33 -0700 Subject: [PATCH 164/330] HID: nintendo: Rate limit IMU compensation message Some controllers are very bad at updating the IMU, leading to these messages spamming the syslog. Rate-limiting them helps with this a bit. Signed-off-by: Vicki Pfau Signed-off-by: Jiri Kosina --- drivers/hid/hid-nintendo.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/hid/hid-nintendo.c b/drivers/hid/hid-nintendo.c index e3e54f1df44f..c2849a541f65 100644 --- a/drivers/hid/hid-nintendo.c +++ b/drivers/hid/hid-nintendo.c @@ -1455,10 +1455,10 @@ static void joycon_parse_imu_report(struct joycon_ctlr *ctlr, ctlr->imu_avg_delta_ms; ctlr->imu_timestamp_us += 1000 * ctlr->imu_avg_delta_ms; if (dropped_pkts > JC_IMU_DROPPED_PKT_WARNING) { - hid_warn(ctlr->hdev, + hid_warn_ratelimited(ctlr->hdev, "compensating for %u dropped IMU reports\n", dropped_pkts); - hid_warn(ctlr->hdev, + hid_warn_ratelimited(ctlr->hdev, "delta=%u avg_delta=%u\n", delta, ctlr->imu_avg_delta_ms); } From 083a4f3f3cc7d107728c8f297e4f6276f0876b2d Mon Sep 17 00:00:00 2001 From: Jonathan Denose Date: Mon, 13 Oct 2025 20:54:57 +0000 Subject: [PATCH 165/330] HID: Kconfig: Fix build error from CONFIG_HID_HAPTIC Temporarily change CONFIG_HID_HAPTIC to be bool instead of tristate, until we implement a permanent solution. Recently the CONFIG_HID_HAPTIC Kconfig option was reported as causing the following build errors: MODPOST Module.symvers ERROR: modpost: "hid_haptic_init" [drivers/hid/hid-multitouch.ko] undefined! ERROR: modpost: "hid_haptic_pressure_increase" [drivers/hid/hid-multitouch.ko] undefined! ERROR: modpost: "hid_haptic_check_pressure_unit" [drivers/hid/hid-multitouch.ko] undefined! ERROR: modpost: "hid_haptic_input_configured" [drivers/hid/hid-multitouch.ko] undefined! ERROR: modpost: "hid_haptic_input_mapping" [drivers/hid/hid-multitouch.ko] undefined! ERROR: modpost: "hid_haptic_feature_mapping" [drivers/hid/hid-multitouch.ko] undefined! ERROR: modpost: "hid_haptic_pressure_reset" [drivers/hid/hid-multitouch.ko] undefined! make[3]: *** [/home/thl/var/linux.dev/scripts/Makefile.modpost:147: Module.symvers] Error 1 when the kernel is compiled with the following configuration: CONFIG_HID=y CONFIG_HID_MULTITOUCH=m CONFIG_HID_HAPTIC=m To resolve this, temporarily change the CONFIG_HID_HAPTIC option to be bool, until we arrive at a permanent solution to enable CONFIG_HID_HAPTIC to be tristate. For a more detailed discussion, see [1]. [1]: https://lore.kernel.org/linux-input/auypydfkhx2eg7vp764way4batdilzc35inqda3exwzs3tk3ff@oagat6g46zto/ Signed-off-by: Jonathan Denose Signed-off-by: Jiri Kosina --- drivers/hid/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/hid/Kconfig b/drivers/hid/Kconfig index 7ff85c7200e5..986de05a9787 100644 --- a/drivers/hid/Kconfig +++ b/drivers/hid/Kconfig @@ -93,7 +93,7 @@ config HID_GENERIC If unsure, say Y. config HID_HAPTIC - tristate "Haptic touchpad support" + bool "Haptic touchpad support" default n help Support for touchpads with force sensors and haptic actuators instead of a From a7cdc2086c19e435d4cec3f9393b5f46899c0468 Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Sun, 28 Sep 2025 22:01:18 +0100 Subject: [PATCH 166/330] HID: hid-debug: Fix spelling mistake "Rechargable" -> "Rechargeable" There is a spelling mistake in HID description. Fix it. Signed-off-by: Colin Ian King Signed-off-by: Jiri Kosina --- drivers/hid/hid-debug.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/hid/hid-debug.c b/drivers/hid/hid-debug.c index 7107071c7c51..337d2dc81b4c 100644 --- a/drivers/hid/hid-debug.c +++ b/drivers/hid/hid-debug.c @@ -2523,7 +2523,7 @@ static const struct hid_usage_entry hid_usage_table[] = { { 0x85, 0x0088, "iDeviceName" }, { 0x85, 0x0089, "iDeviceChemistry" }, { 0x85, 0x008a, "ManufacturerData" }, - { 0x85, 0x008b, "Rechargable" }, + { 0x85, 0x008b, "Rechargeable" }, { 0x85, 0x008c, "WarningCapacityLimit" }, { 0x85, 0x008d, "CapacityGranularity1" }, { 0x85, 0x008e, "CapacityGranularity2" }, From ee6e44dfe6e50b4a5df853d933a96bdff5309e6e Mon Sep 17 00:00:00 2001 From: "Peter Zijlstra (Intel)" Date: Fri, 10 Oct 2025 00:17:27 +0530 Subject: [PATCH 167/330] sched/deadline: Stop dl_server before CPU goes offline IBM CI tool reported kernel warning[1] when running a CPU removal operation through drmgr[2]. i.e "drmgr -c cpu -r -q 1" WARNING: CPU: 0 PID: 0 at kernel/sched/cpudeadline.c:219 cpudl_set+0x58/0x170 NIP [c0000000002b6ed8] cpudl_set+0x58/0x170 LR [c0000000002b7cb8] dl_server_timer+0x168/0x2a0 Call Trace: [c000000002c2f8c0] init_stack+0x78c0/0x8000 (unreliable) [c0000000002b7cb8] dl_server_timer+0x168/0x2a0 [c00000000034df84] __hrtimer_run_queues+0x1a4/0x390 [c00000000034f624] hrtimer_interrupt+0x124/0x300 [c00000000002a230] timer_interrupt+0x140/0x320 Git bisects to: commit 4ae8d9aa9f9d ("sched/deadline: Fix dl_server getting stuck") This happens since: - dl_server hrtimer gets enqueued close to cpu offline, when kthread_park enqueues a fair task. - CPU goes offline and drmgr removes it from cpu_present_mask. - hrtimer fires and warning is hit. Fix it by stopping the dl_server before CPU is marked dead. [1]: https://lore.kernel.org/all/8218e149-7718-4432-9312-f97297c352b9@linux.ibm.com/ [2]: https://github.com/ibm-power-utilities/powerpc-utils/tree/next/src/drmgr [sshegde: wrote the changelog and tested it] Fixes: 4ae8d9aa9f9d ("sched/deadline: Fix dl_server getting stuck") Closes: https://lore.kernel.org/all/8218e149-7718-4432-9312-f97297c352b9@linux.ibm.com Signed-off-by: Peter Zijlstra (Intel) Reported-by: Venkat Rao Bagalkote Signed-off-by: Shrikanth Hegde Signed-off-by: Peter Zijlstra (Intel) Tested-by: Marek Szyprowski Tested-by: Shrikanth Hegde --- kernel/sched/core.c | 2 ++ kernel/sched/deadline.c | 3 +++ 2 files changed, 5 insertions(+) diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 198d2dd45f59..f1ebf67b48e2 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -8571,10 +8571,12 @@ int sched_cpu_dying(unsigned int cpu) sched_tick_stop(cpu); rq_lock_irqsave(rq, &rf); + update_rq_clock(rq); if (rq->nr_running != 1 || rq_has_pinned_tasks(rq)) { WARN(true, "Dying CPU not properly vacated!"); dump_rq_tasks(rq, KERN_WARNING); } + dl_server_stop(&rq->fair_server); rq_unlock_irqrestore(rq, &rf); calc_load_migrate(rq); diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c index 615411a0a881..7b7671060bf9 100644 --- a/kernel/sched/deadline.c +++ b/kernel/sched/deadline.c @@ -1582,6 +1582,9 @@ void dl_server_start(struct sched_dl_entity *dl_se) if (!dl_server(dl_se) || dl_se->dl_server_active) return; + if (WARN_ON_ONCE(!cpu_online(cpu_of(rq)))) + return; + dl_se->dl_server_active = 1; enqueue_dl_entity(dl_se, ENQUEUE_WAKEUP); if (!dl_task(dl_se->rq->curr) || dl_entity_preempt(dl_se, &rq->curr->dl)) From 17e3e88ed0b6318fde0d1c14df1a804711cab1b5 Mon Sep 17 00:00:00 2001 From: Vincent Guittot Date: Wed, 8 Oct 2025 15:12:14 +0200 Subject: [PATCH 168/330] sched/fair: Fix pelt lost idle time detection The check for some lost idle pelt time should be always done when pick_next_task_fair() fails to pick a task and not only when we call it from the fair fast-path. The case happens when the last running task on rq is a RT or DL task. When the latter goes to sleep and the /Sum of util_sum of the rq is at the max value, we don't account the lost of idle time whereas we should. Fixes: 67692435c411 ("sched: Rework pick_next_task() slow-path") Signed-off-by: Vincent Guittot Signed-off-by: Peter Zijlstra (Intel) --- kernel/sched/fair.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index bc0b7ce8a65d..cee1793e8277 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -8920,21 +8920,21 @@ pick_next_task_fair(struct rq *rq, struct task_struct *prev, struct rq_flags *rf return p; idle: - if (!rf) - return NULL; + if (rf) { + new_tasks = sched_balance_newidle(rq, rf); - new_tasks = sched_balance_newidle(rq, rf); + /* + * Because sched_balance_newidle() releases (and re-acquires) + * rq->lock, it is possible for any higher priority task to + * appear. In that case we must re-start the pick_next_entity() + * loop. + */ + if (new_tasks < 0) + return RETRY_TASK; - /* - * Because sched_balance_newidle() releases (and re-acquires) rq->lock, it is - * possible for any higher priority task to appear. In that case we - * must re-start the pick_next_entity() loop. - */ - if (new_tasks < 0) - return RETRY_TASK; - - if (new_tasks > 0) - goto again; + if (new_tasks > 0) + goto again; + } /* * rq is about to be idle, check if we need to update the From ae11e08c3d0c78d08dac4cea30bf39ede2130b03 Mon Sep 17 00:00:00 2001 From: Sakari Ailus Date: Fri, 4 Jul 2025 10:54:15 +0300 Subject: [PATCH 169/330] i2c: Remove redundant pm_runtime_mark_last_busy() calls pm_runtime_put_autosuspend(), pm_runtime_put_sync_autosuspend(), pm_runtime_autosuspend() and pm_request_autosuspend() now include a call to pm_runtime_mark_last_busy(). Remove the now-reduntant explicit call to pm_runtime_mark_last_busy(). Signed-off-by: Sakari Ailus Acked-by: Andi Shyti Signed-off-by: Wolfram Sang --- drivers/i2c/busses/i2c-amd-mp2.h | 1 - drivers/i2c/busses/i2c-at91-core.c | 1 - drivers/i2c/busses/i2c-at91-master.c | 1 - drivers/i2c/busses/i2c-cadence.c | 1 - drivers/i2c/busses/i2c-davinci.c | 2 -- drivers/i2c/busses/i2c-designware-master.c | 1 - drivers/i2c/busses/i2c-hix5hd2.c | 1 - drivers/i2c/busses/i2c-i801.c | 1 - drivers/i2c/busses/i2c-img-scb.c | 3 --- drivers/i2c/busses/i2c-imx-lpi2c.c | 4 ---- drivers/i2c/busses/i2c-imx.c | 3 --- drivers/i2c/busses/i2c-mv64xxx.c | 1 - drivers/i2c/busses/i2c-nvidia-gpu.c | 1 - drivers/i2c/busses/i2c-omap.c | 3 --- drivers/i2c/busses/i2c-qcom-cci.c | 2 -- drivers/i2c/busses/i2c-qcom-geni.c | 1 - drivers/i2c/busses/i2c-qup.c | 3 --- drivers/i2c/busses/i2c-riic.c | 2 -- drivers/i2c/busses/i2c-rzv2m.c | 1 - drivers/i2c/busses/i2c-sprd.c | 2 -- drivers/i2c/busses/i2c-stm32f7.c | 5 ----- drivers/i2c/busses/i2c-xiic.c | 1 - 22 files changed, 41 deletions(-) diff --git a/drivers/i2c/busses/i2c-amd-mp2.h b/drivers/i2c/busses/i2c-amd-mp2.h index 018a42de8b1e..9b7e9494dd12 100644 --- a/drivers/i2c/busses/i2c-amd-mp2.h +++ b/drivers/i2c/busses/i2c-amd-mp2.h @@ -207,7 +207,6 @@ static inline void amd_mp2_pm_runtime_get(struct amd_mp2_dev *mp2_dev) static inline void amd_mp2_pm_runtime_put(struct amd_mp2_dev *mp2_dev) { - pm_runtime_mark_last_busy(&mp2_dev->pci_dev->dev); pm_runtime_put_autosuspend(&mp2_dev->pci_dev->dev); } diff --git a/drivers/i2c/busses/i2c-at91-core.c b/drivers/i2c/busses/i2c-at91-core.c index edc047e3e535..b64adef778d4 100644 --- a/drivers/i2c/busses/i2c-at91-core.c +++ b/drivers/i2c/busses/i2c-at91-core.c @@ -313,7 +313,6 @@ static int __maybe_unused at91_twi_resume_noirq(struct device *dev) return ret; } - pm_runtime_mark_last_busy(dev); pm_request_autosuspend(dev); at91_init_twi_bus(twi_dev); diff --git a/drivers/i2c/busses/i2c-at91-master.c b/drivers/i2c/busses/i2c-at91-master.c index 59795c1c24ff..894cedbca99f 100644 --- a/drivers/i2c/busses/i2c-at91-master.c +++ b/drivers/i2c/busses/i2c-at91-master.c @@ -717,7 +717,6 @@ static int at91_twi_xfer(struct i2c_adapter *adap, struct i2c_msg *msg, int num) ret = (ret < 0) ? ret : num; out: - pm_runtime_mark_last_busy(dev->dev); pm_runtime_put_autosuspend(dev->dev); return ret; diff --git a/drivers/i2c/busses/i2c-cadence.c b/drivers/i2c/busses/i2c-cadence.c index 697d095afbe4..0fb728ade92e 100644 --- a/drivers/i2c/busses/i2c-cadence.c +++ b/drivers/i2c/busses/i2c-cadence.c @@ -1128,7 +1128,6 @@ static int cdns_i2c_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, cdns_i2c_set_mode(CDNS_I2C_MODE_SLAVE, id); #endif - pm_runtime_mark_last_busy(id->dev); pm_runtime_put_autosuspend(id->dev); return ret; } diff --git a/drivers/i2c/busses/i2c-davinci.c b/drivers/i2c/busses/i2c-davinci.c index 6a3d4e9e07f4..a773ba082321 100644 --- a/drivers/i2c/busses/i2c-davinci.c +++ b/drivers/i2c/busses/i2c-davinci.c @@ -543,7 +543,6 @@ i2c_davinci_xfer(struct i2c_adapter *adap, struct i2c_msg msgs[], int num) ret = num; out: - pm_runtime_mark_last_busy(dev->dev); pm_runtime_put_autosuspend(dev->dev); return ret; @@ -821,7 +820,6 @@ static int davinci_i2c_probe(struct platform_device *pdev) if (r) goto err_unuse_clocks; - pm_runtime_mark_last_busy(dev->dev); pm_runtime_put_autosuspend(dev->dev); return 0; diff --git a/drivers/i2c/busses/i2c-designware-master.c b/drivers/i2c/busses/i2c-designware-master.c index c7a72c28786c..41e9b5ecad20 100644 --- a/drivers/i2c/busses/i2c-designware-master.c +++ b/drivers/i2c/busses/i2c-designware-master.c @@ -901,7 +901,6 @@ i2c_dw_xfer(struct i2c_adapter *adap, struct i2c_msg msgs[], int num) i2c_dw_release_lock(dev); done_nolock: - pm_runtime_mark_last_busy(dev->dev); pm_runtime_put_autosuspend(dev->dev); return ret; diff --git a/drivers/i2c/busses/i2c-hix5hd2.c b/drivers/i2c/busses/i2c-hix5hd2.c index 5358f5ddf924..95ab910b80c0 100644 --- a/drivers/i2c/busses/i2c-hix5hd2.c +++ b/drivers/i2c/busses/i2c-hix5hd2.c @@ -373,7 +373,6 @@ static int hix5hd2_i2c_xfer(struct i2c_adapter *adap, ret = num; out: - pm_runtime_mark_last_busy(priv->dev); pm_runtime_put_autosuspend(priv->dev); return ret; } diff --git a/drivers/i2c/busses/i2c-i801.c b/drivers/i2c/busses/i2c-i801.c index cba992fa6557..57fbec1259be 100644 --- a/drivers/i2c/busses/i2c-i801.c +++ b/drivers/i2c/busses/i2c-i801.c @@ -930,7 +930,6 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr, */ iowrite8(SMBHSTSTS_INUSE_STS | STATUS_FLAGS, SMBHSTSTS(priv)); - pm_runtime_mark_last_busy(&priv->pci_dev->dev); pm_runtime_put_autosuspend(&priv->pci_dev->dev); return ret; } diff --git a/drivers/i2c/busses/i2c-img-scb.c b/drivers/i2c/busses/i2c-img-scb.c index a454f9f25146..88192c25c44c 100644 --- a/drivers/i2c/busses/i2c-img-scb.c +++ b/drivers/i2c/busses/i2c-img-scb.c @@ -1131,7 +1131,6 @@ static int img_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, break; } - pm_runtime_mark_last_busy(adap->dev.parent); pm_runtime_put_autosuspend(adap->dev.parent); return i2c->msg_status ? i2c->msg_status : num; @@ -1165,7 +1164,6 @@ static int img_i2c_init(struct img_i2c *i2c) "Unknown hardware revision (%d.%d.%d.%d)\n", (rev >> 24) & 0xff, (rev >> 16) & 0xff, (rev >> 8) & 0xff, rev & 0xff); - pm_runtime_mark_last_busy(i2c->adap.dev.parent); pm_runtime_put_autosuspend(i2c->adap.dev.parent); return -EINVAL; } @@ -1317,7 +1315,6 @@ static int img_i2c_init(struct img_i2c *i2c) /* Perform a synchronous sequence to reset the bus */ ret = img_i2c_reset_bus(i2c); - pm_runtime_mark_last_busy(i2c->adap.dev.parent); pm_runtime_put_autosuspend(i2c->adap.dev.parent); return ret; diff --git a/drivers/i2c/busses/i2c-imx-lpi2c.c b/drivers/i2c/busses/i2c-imx-lpi2c.c index 03b5a7e8c361..2a0962a0b441 100644 --- a/drivers/i2c/busses/i2c-imx-lpi2c.c +++ b/drivers/i2c/busses/i2c-imx-lpi2c.c @@ -363,7 +363,6 @@ static int lpi2c_imx_master_enable(struct lpi2c_imx_struct *lpi2c_imx) return 0; rpm_put: - pm_runtime_mark_last_busy(lpi2c_imx->adapter.dev.parent); pm_runtime_put_autosuspend(lpi2c_imx->adapter.dev.parent); return ret; @@ -377,7 +376,6 @@ static int lpi2c_imx_master_disable(struct lpi2c_imx_struct *lpi2c_imx) temp &= ~MCR_MEN; writel(temp, lpi2c_imx->base + LPI2C_MCR); - pm_runtime_mark_last_busy(lpi2c_imx->adapter.dev.parent); pm_runtime_put_autosuspend(lpi2c_imx->adapter.dev.parent); return 0; @@ -1462,7 +1460,6 @@ static int lpi2c_imx_probe(struct platform_device *pdev) if (ret) goto rpm_disable; - pm_runtime_mark_last_busy(&pdev->dev); pm_runtime_put_autosuspend(&pdev->dev); dev_info(&lpi2c_imx->adapter.dev, "LPI2C adapter registered\n"); @@ -1564,7 +1561,6 @@ static int lpi2c_suspend(struct device *dev) static int lpi2c_resume(struct device *dev) { - pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); return 0; diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c index 60f5c790ad7c..dcce882f3eba 100644 --- a/drivers/i2c/busses/i2c-imx.c +++ b/drivers/i2c/busses/i2c-imx.c @@ -1637,7 +1637,6 @@ static int i2c_imx_xfer(struct i2c_adapter *adapter, result = i2c_imx_xfer_common(adapter, msgs, num, false); - pm_runtime_mark_last_busy(i2c_imx->adapter.dev.parent); pm_runtime_put_autosuspend(i2c_imx->adapter.dev.parent); return result; @@ -1822,7 +1821,6 @@ static int i2c_imx_probe(struct platform_device *pdev) if (ret < 0) goto clk_notifier_unregister; - pm_runtime_mark_last_busy(&pdev->dev); pm_runtime_put_autosuspend(&pdev->dev); dev_dbg(&i2c_imx->adapter.dev, "claimed irq %d\n", irq); @@ -1928,7 +1926,6 @@ static int i2c_imx_suspend(struct device *dev) static int i2c_imx_resume(struct device *dev) { - pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); return 0; diff --git a/drivers/i2c/busses/i2c-mv64xxx.c b/drivers/i2c/busses/i2c-mv64xxx.c index 8fc26a511320..1acba628e16c 100644 --- a/drivers/i2c/busses/i2c-mv64xxx.c +++ b/drivers/i2c/busses/i2c-mv64xxx.c @@ -766,7 +766,6 @@ mv64xxx_i2c_xfer_core(struct i2c_adapter *adap, struct i2c_msg msgs[], int num) drv_data->num_msgs = 0; drv_data->msgs = NULL; - pm_runtime_mark_last_busy(&adap->dev); pm_runtime_put_autosuspend(&adap->dev); return ret; diff --git a/drivers/i2c/busses/i2c-nvidia-gpu.c b/drivers/i2c/busses/i2c-nvidia-gpu.c index 541d808d62d0..14c059b03945 100644 --- a/drivers/i2c/busses/i2c-nvidia-gpu.c +++ b/drivers/i2c/busses/i2c-nvidia-gpu.c @@ -216,7 +216,6 @@ static int gpu_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num) if (status2 < 0) dev_err(i2cd->dev, "i2c stop failed %d\n", status2); } - pm_runtime_mark_last_busy(i2cd->dev); pm_runtime_put_autosuspend(i2cd->dev); return status; } diff --git a/drivers/i2c/busses/i2c-omap.c b/drivers/i2c/busses/i2c-omap.c index 5fcc9f6c33e5..d9f590f0c384 100644 --- a/drivers/i2c/busses/i2c-omap.c +++ b/drivers/i2c/busses/i2c-omap.c @@ -828,7 +828,6 @@ omap_i2c_xfer_common(struct i2c_adapter *adap, struct i2c_msg msgs[], int num, omap->set_mpu_wkup_lat(omap->dev, -1); out: - pm_runtime_mark_last_busy(omap->dev); pm_runtime_put_autosuspend(omap->dev); return r; } @@ -1510,7 +1509,6 @@ omap_i2c_probe(struct platform_device *pdev) dev_info(omap->dev, "bus %d rev%d.%d at %d kHz\n", adap->nr, major, minor, omap->speed); - pm_runtime_mark_last_busy(omap->dev); pm_runtime_put_autosuspend(omap->dev); return 0; @@ -1605,7 +1603,6 @@ static int omap_i2c_suspend(struct device *dev) static int omap_i2c_resume(struct device *dev) { - pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); return 0; diff --git a/drivers/i2c/busses/i2c-qcom-cci.c b/drivers/i2c/busses/i2c-qcom-cci.c index a3afa11a71a1..e631d79baf14 100644 --- a/drivers/i2c/busses/i2c-qcom-cci.c +++ b/drivers/i2c/busses/i2c-qcom-cci.c @@ -450,7 +450,6 @@ static int cci_xfer(struct i2c_adapter *adap, struct i2c_msg msgs[], int num) ret = num; err: - pm_runtime_mark_last_busy(cci->dev); pm_runtime_put_autosuspend(cci->dev); return ret; @@ -508,7 +507,6 @@ static int __maybe_unused cci_suspend(struct device *dev) static int __maybe_unused cci_resume(struct device *dev) { cci_resume_runtime(dev); - pm_runtime_mark_last_busy(dev); pm_request_autosuspend(dev); return 0; diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c index 95a577764d5c..43fdd89b8beb 100644 --- a/drivers/i2c/busses/i2c-qcom-geni.c +++ b/drivers/i2c/busses/i2c-qcom-geni.c @@ -714,7 +714,6 @@ static int geni_i2c_xfer(struct i2c_adapter *adap, else ret = geni_i2c_fifo_xfer(gi2c, msgs, num); - pm_runtime_mark_last_busy(gi2c->se.dev); pm_runtime_put_autosuspend(gi2c->se.dev); gi2c->cur = NULL; gi2c->err = 0; diff --git a/drivers/i2c/busses/i2c-qup.c b/drivers/i2c/busses/i2c-qup.c index fc348924d522..a0e076fc5f36 100644 --- a/drivers/i2c/busses/i2c-qup.c +++ b/drivers/i2c/busses/i2c-qup.c @@ -1139,7 +1139,6 @@ static int qup_i2c_xfer(struct i2c_adapter *adap, ret = num; out: - pm_runtime_mark_last_busy(qup->dev); pm_runtime_put_autosuspend(qup->dev); return ret; @@ -1624,7 +1623,6 @@ static int qup_i2c_xfer_v2(struct i2c_adapter *adap, if (ret == 0) ret = num; out: - pm_runtime_mark_last_busy(qup->dev); pm_runtime_put_autosuspend(qup->dev); return ret; @@ -1991,7 +1989,6 @@ static int qup_i2c_suspend(struct device *device) static int qup_i2c_resume(struct device *device) { qup_i2c_pm_resume_runtime(device); - pm_runtime_mark_last_busy(device); pm_request_autosuspend(device); return 0; } diff --git a/drivers/i2c/busses/i2c-riic.c b/drivers/i2c/busses/i2c-riic.c index b0ee9ac45a97..3e8f126cb7f7 100644 --- a/drivers/i2c/busses/i2c-riic.c +++ b/drivers/i2c/busses/i2c-riic.c @@ -206,7 +206,6 @@ static int riic_xfer(struct i2c_adapter *adap, struct i2c_msg msgs[], int num) } out: - pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); return riic->err ?: num; @@ -452,7 +451,6 @@ static int riic_init_hw(struct riic_dev *riic) riic_clear_set_bit(riic, ICCR1_IICRST, 0, RIIC_ICCR1); - pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); return 0; } diff --git a/drivers/i2c/busses/i2c-rzv2m.c b/drivers/i2c/busses/i2c-rzv2m.c index b0e9c0b62429..238714850673 100644 --- a/drivers/i2c/busses/i2c-rzv2m.c +++ b/drivers/i2c/busses/i2c-rzv2m.c @@ -372,7 +372,6 @@ static int rzv2m_i2c_xfer(struct i2c_adapter *adap, ret = num; out: - pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); return ret; diff --git a/drivers/i2c/busses/i2c-sprd.c b/drivers/i2c/busses/i2c-sprd.c index 26ec34b19ad5..1b490525d8dd 100644 --- a/drivers/i2c/busses/i2c-sprd.c +++ b/drivers/i2c/busses/i2c-sprd.c @@ -302,7 +302,6 @@ static int sprd_i2c_xfer(struct i2c_adapter *i2c_adap, ret = sprd_i2c_handle_msg(i2c_adap, &msgs[im++], 1); err_msg: - pm_runtime_mark_last_busy(i2c_dev->dev); pm_runtime_put_autosuspend(i2c_dev->dev); return ret < 0 ? ret : im; @@ -559,7 +558,6 @@ static int sprd_i2c_probe(struct platform_device *pdev) goto err_rpm_put; } - pm_runtime_mark_last_busy(i2c_dev->dev); pm_runtime_put_autosuspend(i2c_dev->dev); return 0; diff --git a/drivers/i2c/busses/i2c-stm32f7.c b/drivers/i2c/busses/i2c-stm32f7.c index e6815f6cae78..dc69ed934ec8 100644 --- a/drivers/i2c/busses/i2c-stm32f7.c +++ b/drivers/i2c/busses/i2c-stm32f7.c @@ -1761,7 +1761,6 @@ static int stm32f7_i2c_xfer_core(struct i2c_adapter *i2c_adap, } pm_free: - pm_runtime_mark_last_busy(i2c_dev->dev); pm_runtime_put_autosuspend(i2c_dev->dev); return (ret < 0) ? ret : num; @@ -1870,7 +1869,6 @@ static int stm32f7_i2c_smbus_xfer(struct i2c_adapter *adapter, u16 addr, } pm_free: - pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); return ret; } @@ -1977,7 +1975,6 @@ static int stm32f7_i2c_reg_slave(struct i2c_client *slave) if (!stm32f7_i2c_is_slave_registered(i2c_dev)) stm32f7_i2c_enable_wakeup(i2c_dev, false); - pm_runtime_mark_last_busy(dev); pm_runtime_put_autosuspend(dev); return ret; @@ -2015,7 +2012,6 @@ static int stm32f7_i2c_unreg_slave(struct i2c_client *slave) stm32f7_i2c_enable_wakeup(i2c_dev, false); } - pm_runtime_mark_last_busy(i2c_dev->dev); pm_runtime_put_autosuspend(i2c_dev->dev); return 0; @@ -2328,7 +2324,6 @@ static int stm32f7_i2c_probe(struct platform_device *pdev) dev_info(i2c_dev->dev, "STM32F7 I2C-%d bus adapter\n", adap->nr); - pm_runtime_mark_last_busy(i2c_dev->dev); pm_runtime_put_autosuspend(i2c_dev->dev); return 0; diff --git a/drivers/i2c/busses/i2c-xiic.c b/drivers/i2c/busses/i2c-xiic.c index 607026c921d6..28015d77599d 100644 --- a/drivers/i2c/busses/i2c-xiic.c +++ b/drivers/i2c/busses/i2c-xiic.c @@ -1349,7 +1349,6 @@ static int xiic_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num) mutex_unlock(&i2c->lock); out: - pm_runtime_mark_last_busy(i2c->dev); pm_runtime_put_autosuspend(i2c->dev); return err; } From 72f437e674e54f1c143dccc67e5556d8d5acb241 Mon Sep 17 00:00:00 2001 From: Hans de Goede Date: Wed, 1 Oct 2025 23:23:10 +0200 Subject: [PATCH 170/330] i2c: usbio: Add ACPI device-id for MTL-CVF devices Add "INTC10D2" ACPI device-id for MTL-CVF devices, like the Dell Latitude 7450. Closes: https://bugzilla.redhat.com/show_bug.cgi?id=2368506 Signed-off-by: Hans de Goede Acked-by: Sakari Ailus Acked-by: Israel Cepeda Signed-off-by: Wolfram Sang --- drivers/i2c/busses/i2c-usbio.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/i2c/busses/i2c-usbio.c b/drivers/i2c/busses/i2c-usbio.c index d42f9ab6e9a5..e7799abf6787 100644 --- a/drivers/i2c/busses/i2c-usbio.c +++ b/drivers/i2c/busses/i2c-usbio.c @@ -27,6 +27,7 @@ static const struct acpi_device_id usbio_i2c_acpi_hids[] = { { "INTC1008" }, /* MTL */ { "INTC10B3" }, /* ARL */ { "INTC10B6" }, /* LNL */ + { "INTC10D2" }, /* MTL-CVF */ { "INTC10E3" }, /* PTL */ { } }; From 867537094124b0736ca2a40193de94fc5dc0b8d3 Mon Sep 17 00:00:00 2001 From: "Rob Herring (Arm)" Date: Mon, 13 Oct 2025 16:31:18 -0500 Subject: [PATCH 171/330] dt-bindings: i2c: Convert apm,xgene-slimpro-i2c to DT schema Convert APM X-Gene slimpro-i2c binding to DT schema format. It's a straight-forward conversion. Signed-off-by: Rob Herring (Arm) Signed-off-by: Wolfram Sang --- .../bindings/i2c/apm,xgene-slimpro-i2c.yaml | 36 +++++++++++++++++++ .../bindings/i2c/i2c-xgene-slimpro.txt | 15 -------- 2 files changed, 36 insertions(+), 15 deletions(-) create mode 100644 Documentation/devicetree/bindings/i2c/apm,xgene-slimpro-i2c.yaml delete mode 100644 Documentation/devicetree/bindings/i2c/i2c-xgene-slimpro.txt diff --git a/Documentation/devicetree/bindings/i2c/apm,xgene-slimpro-i2c.yaml b/Documentation/devicetree/bindings/i2c/apm,xgene-slimpro-i2c.yaml new file mode 100644 index 000000000000..9460c64071f2 --- /dev/null +++ b/Documentation/devicetree/bindings/i2c/apm,xgene-slimpro-i2c.yaml @@ -0,0 +1,36 @@ +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +%YAML 1.2 +--- +$id: http://devicetree.org/schemas/i2c/apm,xgene-slimpro-i2c.yaml# +$schema: http://devicetree.org/meta-schemas/core.yaml# + +title: APM X-Gene SLIMpro Mailbox I2C + +maintainers: + - Khuong Dinh + +description: + An I2C controller accessed over the "SLIMpro" mailbox. + +allOf: + - $ref: /schemas/i2c/i2c-controller.yaml# + +properties: + compatible: + const: apm,xgene-slimpro-i2c + + mboxes: + maxItems: 1 + +required: + - compatible + - mboxes + +unevaluatedProperties: false + +examples: + - | + i2c { + compatible = "apm,xgene-slimpro-i2c"; + mboxes = <&mailbox 0>; + }; diff --git a/Documentation/devicetree/bindings/i2c/i2c-xgene-slimpro.txt b/Documentation/devicetree/bindings/i2c/i2c-xgene-slimpro.txt deleted file mode 100644 index f6b2c20cfbf6..000000000000 --- a/Documentation/devicetree/bindings/i2c/i2c-xgene-slimpro.txt +++ /dev/null @@ -1,15 +0,0 @@ -APM X-Gene SLIMpro Mailbox I2C Driver - -An I2C controller accessed over the "SLIMpro" mailbox. - -Required properties : - - - compatible : should be "apm,xgene-slimpro-i2c" - - mboxes : use the label reference for the mailbox as the first parameter. - The second parameter is the channel number. - -Example : - i2cslimpro { - compatible = "apm,xgene-slimpro-i2c"; - mboxes = <&mailbox 0>; - }; From d41f68dff783d181a8fd462e612bda0fbab7f735 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Mon, 13 Oct 2025 20:05:52 -0700 Subject: [PATCH 172/330] ALSA: firewire: amdtp-stream: fix enum kernel-doc warnings Fix spelling of CIP_NO_HEADER to prevent a kernel-doc warning. Warning: amdtp-stream.h:57 Enum value 'CIP_NO_HEADER' not described in enum 'cip_flags' Warning: amdtp-stream.h:57 Excess enum value '%CIP_NO_HEADERS' description in 'cip_flags' Fixes: 3b196c394dd9f ("ALSA: firewire-lib: add no-header packet processing") Signed-off-by: Randy Dunlap Reviewed-by: Takashi Sakamoto Signed-off-by: Takashi Iwai --- sound/firewire/amdtp-stream.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/firewire/amdtp-stream.h b/sound/firewire/amdtp-stream.h index 775db3fc4959..ec10270c2cce 100644 --- a/sound/firewire/amdtp-stream.h +++ b/sound/firewire/amdtp-stream.h @@ -32,7 +32,7 @@ * allows 5 times as large as IEC 61883-6 defines. * @CIP_HEADER_WITHOUT_EOH: Only for in-stream. CIP Header doesn't include * valid EOH. - * @CIP_NO_HEADERS: a lack of headers in packets + * @CIP_NO_HEADER: a lack of headers in packets * @CIP_UNALIGHED_DBC: Only for in-stream. The value of dbc is not alighed to * the value of current SYT_INTERVAL; e.g. initial value is not zero. * @CIP_UNAWARE_SYT: For outgoing packet, the value in SYT field of CIP is 0xffff. From fd6e385528d8f85993b7bfc6430576136bb14c65 Mon Sep 17 00:00:00 2001 From: Jeffrey Hugo Date: Tue, 7 Oct 2025 13:57:50 +0200 Subject: [PATCH 173/330] accel/qaic: Fix bootlog initialization ordering As soon as we queue MHI buffers to receive the bootlog from the device, we could be receiving data. Therefore all the resources needed to process that data need to be setup prior to queuing the buffers. We currently initialize some of the resources after queuing the buffers which creates a race between the probe() and any data that comes back from the device. If the uninitialized resources are accessed, we could see page faults. Fix the init ordering to close the race. Fixes: 5f8df5c6def6 ("accel/qaic: Add bootlog debugfs") Signed-off-by: Jeffrey Hugo Signed-off-by: Youssef Samir Reviewed-by: Jeff Hugo Reviewed-by: Carl Vanderlip Signed-off-by: Jeff Hugo Link: https://lore.kernel.org/r/20251007115750.332169-1-youssef.abdulrahman@oss.qualcomm.com --- drivers/accel/qaic/qaic_debugfs.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/accel/qaic/qaic_debugfs.c b/drivers/accel/qaic/qaic_debugfs.c index a991b8198dc4..8dc4fe5bb560 100644 --- a/drivers/accel/qaic/qaic_debugfs.c +++ b/drivers/accel/qaic/qaic_debugfs.c @@ -218,6 +218,9 @@ static int qaic_bootlog_mhi_probe(struct mhi_device *mhi_dev, const struct mhi_d if (ret) goto destroy_workqueue; + dev_set_drvdata(&mhi_dev->dev, qdev); + qdev->bootlog_ch = mhi_dev; + for (i = 0; i < BOOTLOG_POOL_SIZE; i++) { msg = devm_kzalloc(&qdev->pdev->dev, sizeof(*msg), GFP_KERNEL); if (!msg) { @@ -233,8 +236,6 @@ static int qaic_bootlog_mhi_probe(struct mhi_device *mhi_dev, const struct mhi_d goto mhi_unprepare; } - dev_set_drvdata(&mhi_dev->dev, qdev); - qdev->bootlog_ch = mhi_dev; return 0; mhi_unprepare: From 11f08c30a3e4157305ba692f1d44cca5fc9a8fca Mon Sep 17 00:00:00 2001 From: Youssef Samir Date: Tue, 7 Oct 2025 14:23:20 +0200 Subject: [PATCH 174/330] accel/qaic: Treat remaining == 0 as error in find_and_map_user_pages() Currently, if find_and_map_user_pages() takes a DMA xfer request from the user with a length field set to 0, or in a rare case, the host receives QAIC_TRANS_DMA_XFER_CONT from the device where resources->xferred_dma_size is equal to the requested transaction size, the function will return 0 before allocating an sgt or setting the fields of the dma_xfer struct. In that case, encode_addr_size_pairs() will try to access the sgt which will lead to a general protection fault. Return an EINVAL in case the user provides a zero-sized ALP, or the device requests continuation after all of the bytes have been transferred. Fixes: 96d3c1cadedb ("accel/qaic: Clean up integer overflow checking in map_user_pages()") Signed-off-by: Youssef Samir Signed-off-by: Youssef Samir Reviewed-by: Jeff Hugo Reviewed-by: Carl Vanderlip Signed-off-by: Jeff Hugo Link: https://lore.kernel.org/r/20251007122320.339654-1-youssef.abdulrahman@oss.qualcomm.com --- drivers/accel/qaic/qaic_control.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/accel/qaic/qaic_control.c b/drivers/accel/qaic/qaic_control.c index d8bdab69f800..b86a8e48e731 100644 --- a/drivers/accel/qaic/qaic_control.c +++ b/drivers/accel/qaic/qaic_control.c @@ -407,7 +407,7 @@ static int find_and_map_user_pages(struct qaic_device *qdev, return -EINVAL; remaining = in_trans->size - resources->xferred_dma_size; if (remaining == 0) - return 0; + return -EINVAL; if (check_add_overflow(xfer_start_addr, remaining, &end)) return -EINVAL; From 52e59f7740ba23bbb664914967df9a00208ca10c Mon Sep 17 00:00:00 2001 From: Pranjal Ramajor Asha Kanojiya Date: Tue, 7 Oct 2025 08:18:37 +0200 Subject: [PATCH 175/330] accel/qaic: Synchronize access to DBC request queue head & tail pointer Two threads of the same process can potential read and write parallelly to head and tail pointers of the same DBC request queue. This could lead to a race condition and corrupt the DBC request queue. Fixes: ff13be830333 ("accel/qaic: Add datapath") Signed-off-by: Pranjal Ramajor Asha Kanojiya Signed-off-by: Youssef Samir Reviewed-by: Jeff Hugo Reviewed-by: Carl Vanderlip [jhugo: Add fixes tag] Signed-off-by: Jeff Hugo Link: https://lore.kernel.org/r/20251007061837.206132-1-youssef.abdulrahman@oss.qualcomm.com --- drivers/accel/qaic/qaic.h | 2 ++ drivers/accel/qaic/qaic_data.c | 12 ++++++++++-- drivers/accel/qaic/qaic_drv.c | 3 +++ 3 files changed, 15 insertions(+), 2 deletions(-) diff --git a/drivers/accel/qaic/qaic.h b/drivers/accel/qaic/qaic.h index c31081e42cee..820d133236dd 100644 --- a/drivers/accel/qaic/qaic.h +++ b/drivers/accel/qaic/qaic.h @@ -97,6 +97,8 @@ struct dma_bridge_chan { * response queue's head and tail pointer of this DBC. */ void __iomem *dbc_base; + /* Synchronizes access to Request queue's head and tail pointer */ + struct mutex req_lock; /* Head of list where each node is a memory handle queued in request queue */ struct list_head xfer_list; /* Synchronizes DBC readers during cleanup */ diff --git a/drivers/accel/qaic/qaic_data.c b/drivers/accel/qaic/qaic_data.c index 797289e9d780..c4f117edb266 100644 --- a/drivers/accel/qaic/qaic_data.c +++ b/drivers/accel/qaic/qaic_data.c @@ -1356,13 +1356,17 @@ static int __qaic_execute_bo_ioctl(struct drm_device *dev, void *data, struct dr goto release_ch_rcu; } + ret = mutex_lock_interruptible(&dbc->req_lock); + if (ret) + goto release_ch_rcu; + head = readl(dbc->dbc_base + REQHP_OFF); tail = readl(dbc->dbc_base + REQTP_OFF); if (head == U32_MAX || tail == U32_MAX) { /* PCI link error */ ret = -ENODEV; - goto release_ch_rcu; + goto unlock_req_lock; } queue_level = head <= tail ? tail - head : dbc->nelem - (head - tail); @@ -1370,11 +1374,12 @@ static int __qaic_execute_bo_ioctl(struct drm_device *dev, void *data, struct dr ret = send_bo_list_to_device(qdev, file_priv, exec, args->hdr.count, is_partial, dbc, head, &tail); if (ret) - goto release_ch_rcu; + goto unlock_req_lock; /* Finalize commit to hardware */ submit_ts = ktime_get_ns(); writel(tail, dbc->dbc_base + REQTP_OFF); + mutex_unlock(&dbc->req_lock); update_profiling_data(file_priv, exec, args->hdr.count, is_partial, received_ts, submit_ts, queue_level); @@ -1382,6 +1387,9 @@ static int __qaic_execute_bo_ioctl(struct drm_device *dev, void *data, struct dr if (datapath_polling) schedule_work(&dbc->poll_work); +unlock_req_lock: + if (ret) + mutex_unlock(&dbc->req_lock); release_ch_rcu: srcu_read_unlock(&dbc->ch_lock, rcu_id); unlock_dev_srcu: diff --git a/drivers/accel/qaic/qaic_drv.c b/drivers/accel/qaic/qaic_drv.c index e31bcb0ecfc9..e162f4b8a262 100644 --- a/drivers/accel/qaic/qaic_drv.c +++ b/drivers/accel/qaic/qaic_drv.c @@ -454,6 +454,9 @@ static struct qaic_device *create_qdev(struct pci_dev *pdev, return NULL; init_waitqueue_head(&qdev->dbc[i].dbc_release); INIT_LIST_HEAD(&qdev->dbc[i].bo_lists); + ret = drmm_mutex_init(drm, &qdev->dbc[i].req_lock); + if (ret) + return NULL; } return qdev; From 7e091add9c433bab6912228799bf508e2414acc3 Mon Sep 17 00:00:00 2001 From: Martin George Date: Mon, 8 Sep 2025 22:54:57 +0530 Subject: [PATCH 176/330] nvme-auth: update sc_c in host response The sc_c field is currently not updated in the host response to the controller challenge leading to failures while attempting secure channel concatenation. Fix this by adding a new sc_c variable to the dhchap queue context structure which is appropriately set during negotiate and then used in the host response. Fixes: e88a7595b57f ("nvme-tcp: request secure channel concatenation") Signed-off-by: Martin George Signed-off-by: Prashanth Adurthi Reviewed-by: Hannes Reinecke Signed-off-by: Keith Busch --- drivers/nvme/host/auth.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/nvme/host/auth.c b/drivers/nvme/host/auth.c index 012fcfc79a73..a01178caf15b 100644 --- a/drivers/nvme/host/auth.c +++ b/drivers/nvme/host/auth.c @@ -36,6 +36,7 @@ struct nvme_dhchap_queue_context { u8 status; u8 dhgroup_id; u8 hash_id; + u8 sc_c; size_t hash_len; u8 c1[64]; u8 c2[64]; @@ -154,6 +155,8 @@ static int nvme_auth_set_dhchap_negotiate_data(struct nvme_ctrl *ctrl, data->auth_protocol[0].dhchap.idlist[34] = NVME_AUTH_DHGROUP_6144; data->auth_protocol[0].dhchap.idlist[35] = NVME_AUTH_DHGROUP_8192; + chap->sc_c = data->sc_c; + return size; } @@ -489,7 +492,7 @@ static int nvme_auth_dhchap_setup_host_response(struct nvme_ctrl *ctrl, ret = crypto_shash_update(shash, buf, 2); if (ret) goto out; - memset(buf, 0, sizeof(buf)); + *buf = chap->sc_c; ret = crypto_shash_update(shash, buf, 1); if (ret) goto out; @@ -500,6 +503,7 @@ static int nvme_auth_dhchap_setup_host_response(struct nvme_ctrl *ctrl, strlen(ctrl->opts->host->nqn)); if (ret) goto out; + memset(buf, 0, sizeof(buf)); ret = crypto_shash_update(shash, buf, 1); if (ret) goto out; From df90f6cd29d8c77be6de4f9adf9cbe42ce2f0016 Mon Sep 17 00:00:00 2001 From: Vlastimil Babka Date: Tue, 14 Oct 2025 10:40:57 +0200 Subject: [PATCH 177/330] slab: fix clearing freelist in free_deferred_objects() defer_free() links pending objects using the slab's freelist offset which is fine as they are not free yet. free_deferred_objects() then clears this pointer to avoid confusing the debugging consistency checks that may be enabled for the cache. However, with CONFIG_SLAB_FREELIST_HARDENED, even the NULL pointer needs to be encoded appropriately using set_freepointer(), otherwise it's decoded as something else and triggers the consistency checks, as found by the kernel test robot. Use set_freepointer() to prevent the issue. Fixes: af92793e52c3 ("slab: Introduce kmalloc_nolock() and kfree_nolock().") Reported-and-tested-by: kernel test robot Closes: https://lore.kernel.org/oe-lkp/202510101652.7921fdc6-lkp@intel.com Acked-by: Alexei Starovoitov Reviewed-by: Harry Yoo Signed-off-by: Vlastimil Babka --- mm/slub.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/mm/slub.c b/mm/slub.c index b1f15598fbfd..13ae4491136a 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -6443,15 +6443,16 @@ static void free_deferred_objects(struct irq_work *work) slab = virt_to_slab(x); s = slab->slab_cache; + /* Point 'x' back to the beginning of allocated object */ + x -= s->offset; + /* * We used freepointer in 'x' to link 'x' into df->objects. * Clear it to NULL to avoid false positive detection * of "Freepointer corruption". */ - *(void **)x = NULL; + set_freepointer(s, x, NULL); - /* Point 'x' back to the beginning of allocated object */ - x -= s->offset; __slab_free(s, slab, x, x, 1, _THIS_IP_); } From 7f9ee5fc97e14682e36fe22ae2654c07e4998b82 Mon Sep 17 00:00:00 2001 From: Shardul Bankar Date: Tue, 14 Oct 2025 17:30:37 +0530 Subject: [PATCH 178/330] bpf: test_run: Fix ctx leak in bpf_prog_test_run_xdp error path Fix a memory leak in bpf_prog_test_run_xdp() where the context buffer allocated by bpf_ctx_init() is not freed when the function returns early due to a data size check. On the failing path: ctx = bpf_ctx_init(...); if (kattr->test.data_size_in - meta_sz < ETH_HLEN) return -EINVAL; The early return bypasses the cleanup label that kfree()s ctx, leading to a leak detectable by kmemleak under fuzzing. Change the return to jump to the existing free_ctx label. Fixes: fe9544ed1a2e ("bpf: Support specifying linear xdp packet data size for BPF_PROG_TEST_RUN") Reported-by: BPF Runtime Fuzzer (BRF) Signed-off-by: Shardul Bankar Signed-off-by: Martin KaFai Lau Acked-by: Jiri Olsa Acked-by: Daniel Borkmann Link: https://patch.msgid.link/20251014120037.1981316-1-shardulsb08@gmail.com --- net/bpf/test_run.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c index dfb03ee0bb62..1782e83de2cb 100644 --- a/net/bpf/test_run.c +++ b/net/bpf/test_run.c @@ -1269,7 +1269,7 @@ int bpf_prog_test_run_xdp(struct bpf_prog *prog, const union bpf_attr *kattr, goto free_ctx; if (kattr->test.data_size_in - meta_sz < ETH_HLEN) - return -EINVAL; + goto free_ctx; data = bpf_test_init(kattr, linear_sz, max_linear_sz, headroom, tailroom); if (IS_ERR(data)) { From 2e41e5a91a37202ff6743c3ae5329e106aeb1c6c Mon Sep 17 00:00:00 2001 From: Dave Jiang Date: Fri, 10 Oct 2025 13:57:53 -0700 Subject: [PATCH 179/330] cxl/acpi: Fix setup of memory resource in cxl_acpi_set_cache_size() In order to compare the resource against the HMAT memory target, the resource needs to be memory type. Change the DEFINE_RES() macro to DEFINE_RES_MEM() in order to set the correct resource type. hmat_get_extended_linear_cache_size() uses resource_contains() internally. This causes a regression for platforms with the extended linear cache enabled as the comparison always fails and the cache size is not set. User visible impact is that when 'cxl list' is issued, a CXL region with extended linear cache support will only report half the size of the actual size. And this also breaks MCE reporting of the memory region due to incorrect offset calculation for the memory. [dj: Fixup commit log suggested by djbw] [dj: Fixup stable address for cc] Fixes: 12b3d697c812 ("cxl: Remove core/acpi.c and cxl core dependency on ACPI") Cc: stable@vger.kernel.org Reviewed-by: Gregory Price Reviewed-by: Alison Schofield Reviewed-by: Dan Williams Signed-off-by: Dave Jiang --- drivers/cxl/acpi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/cxl/acpi.c b/drivers/cxl/acpi.c index d7a5539d07d4..bd2e282ca93a 100644 --- a/drivers/cxl/acpi.c +++ b/drivers/cxl/acpi.c @@ -348,7 +348,7 @@ static int cxl_acpi_set_cache_size(struct cxl_root_decoder *cxlrd) struct resource res; int nid, rc; - res = DEFINE_RES(start, size, 0); + res = DEFINE_RES_MEM(start, size); nid = phys_to_target_node(start); rc = hmat_get_extended_linear_cache_size(&res, nid, &cache_size); From 0f6f1982cb28abf1b8a3a8ba906e2c6ade6a70e8 Mon Sep 17 00:00:00 2001 From: Dave Jiang Date: Fri, 10 Oct 2025 13:57:54 -0700 Subject: [PATCH 180/330] cxl: Set range param for region_res_match_cxl_range() as const The function takes two parameters and compares them. The second parameter should be const since no modification should be done to it. Reviewed-by: Gregory Price Reviewed-by: Alison Schofield Signed-off-by: Dave Jiang --- drivers/cxl/core/region.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c index e14c1d305b22..858d4678628d 100644 --- a/drivers/cxl/core/region.c +++ b/drivers/cxl/core/region.c @@ -839,7 +839,7 @@ static int match_free_decoder(struct device *dev, const void *data) } static bool region_res_match_cxl_range(const struct cxl_region_params *p, - struct range *range) + const struct range *range) { if (!p->res) return false; From f4d027921c811ff7fc16e4d03c6bbbf4347cf37a Mon Sep 17 00:00:00 2001 From: Dave Jiang Date: Fri, 10 Oct 2025 13:57:55 -0700 Subject: [PATCH 181/330] cxl: Fix match_region_by_range() to use region_res_match_cxl_range() match_region_by_range() is not using the helper function that also takes extended linear cache size into account when comparing regions. This causes a x2 region to show up as 2 partial incomplete regions rather than a single CXL region with extended linear cache support. Replace the open coded compare logic with the proper helper function for comparison. User visible impact is that when 'cxl list' is issued, no activa CXL region(s) are shown. There may be multiple idle regions present. No actual active CXL region is present in the kernel. [dj: Fix stable address] Fixes: 0ec9849b6333 ("acpi/hmat / cxl: Add extended linear cache support for CXL") Cc: stable@vger.kernel.org Reviewed-by: Gregory Price Reviewed-by: Alison Schofield Reviewed-by: Dan Williams Signed-off-by: Dave Jiang --- drivers/cxl/core/region.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c index 858d4678628d..57ed85e332d3 100644 --- a/drivers/cxl/core/region.c +++ b/drivers/cxl/core/region.c @@ -3398,10 +3398,7 @@ static int match_region_by_range(struct device *dev, const void *data) p = &cxlr->params; guard(rwsem_read)(&cxl_rwsem.region); - if (p->res && p->res->start == r->start && p->res->end == r->end) - return 1; - - return 0; + return region_res_match_cxl_range(p, r); } static int cxl_extended_linear_cache_resize(struct cxl_region *cxlr, From 257c4b03a2f7d8c15f79c79b09a561af9734f6c4 Mon Sep 17 00:00:00 2001 From: Alison Schofield Date: Tue, 14 Oct 2025 00:31:04 -0700 Subject: [PATCH 182/330] cxl/region: Use %pa printk format to emit resource_size_t KASAN reports a stack-out-of-bounds access in validate_region_offset() while running the cxl-poison.sh unit test because the printk format specifier, %pr format, is not a match for the resource_size_t type of the variables. %pr expects struct resource pointers and attempts to dereference the structure fields, reading beyond the bounds of the stack variables. Since these messages emit an 'A exceeds B' type of message, keep the resource_size_t's and use the %pa specifier to be architecture safe. BUG: KASAN: stack-out-of-bounds in resource_string.isra.0+0xe9a/0x1690 [] Read of size 8 at addr ffff88800a7afb40 by task bash/1397 ... [] The buggy address belongs to stack of task bash/1397 [] and is located at offset 56 in frame: [] validate_region_offset+0x0/0x1c0 [cxl_core] Fixes: c3dd67681c70 ("cxl/region: Add inject and clear poison by region offset") Signed-off-by: Alison Schofield Reviewed-by: Dave Jiang Signed-off-by: Dave Jiang --- drivers/cxl/core/region.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c index 57ed85e332d3..b06fee1978ba 100644 --- a/drivers/cxl/core/region.c +++ b/drivers/cxl/core/region.c @@ -3663,14 +3663,14 @@ static int validate_region_offset(struct cxl_region *cxlr, u64 offset) if (offset < p->cache_size) { dev_err(&cxlr->dev, - "Offset %#llx is within extended linear cache %pr\n", + "Offset %#llx is within extended linear cache %pa\n", offset, &p->cache_size); return -EINVAL; } region_size = resource_size(p->res); if (offset >= region_size) { - dev_err(&cxlr->dev, "Offset %#llx exceeds region size %pr\n", + dev_err(&cxlr->dev, "Offset %#llx exceeds region size %pa\n", offset, ®ion_size); return -EINVAL; } From 469276c06affdfd2d9e88c9f228bb81119ec1a20 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= Date: Tue, 14 Oct 2025 19:36:02 +0300 Subject: [PATCH 183/330] PCI: Revert early bridge resource set up MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The commit a43ac325c7cb ("PCI: Set up bridge resources earlier") moved bridge window resources set up earlier than before. The change was necessary to support another change that got pulled on the last minute due to breaking s390 and other systems. The presence of valid bridge window resources earlier than before allows pci_assign_unassigned_root_bus_resources() call from pci_host_probe() assign the bridge windows. Some host bridges, however, have to wait first for the link up event before they can enumerate successfully (see e.g. qcom_pcie_global_irq_thread()) and thus the bus has not been enumerated yet while calling pci_host_probe(). Calling pci_assign_unassigned_root_bus_resources() without results from enumeration can result in sizing bridge windows with too small sizes which cannot be later corrected after the enumeration has completed because bridge windows have become pinned in place by the other resources. Interestingly, it seems pci_read_bridge_bases() is not called at all in the problematic case and the bridge window resource type setup is done by pci_bridge_check_ranges() and sizing by the usual resource fitting logic. The root problem behind all this looks pretty generic. If resource fitting is called too early, the hotplug reservation and old size lower bounding cause the bridge windows to be assigned without children but with non-zero size, which leads to these pinning problems. As such, this can likely be solved on the general level but the solution does not look trivial. As the commit a43ac325c7cb ("PCI: Set up bridge resources earlier") was prequisite for other change that did not end up into kernel yet, revert it to resolve the resource assignment failures and give time to code and test a generic solution. Fixes: a43ac325c7cb ("PCI: Set up bridge resources earlier") Reported-by: Val Packett Link: https://lore.kernel.org/r/017ff8df-511c-4da8-b3cf-edf2cb7f1a67@packett.cool Reported-by: Guenter Roeck Link: https://lore.kernel.org/r/df266709-a9b3-4fd8-af3a-c22eb3c9523a@roeck-us.net Signed-off-by: Ilpo Järvinen Signed-off-by: Bjorn Helgaas Link: https://patch.msgid.link/20251014163602.17138-1-ilpo.jarvinen@linux.intel.com --- drivers/pci/probe.c | 13 +++---------- 1 file changed, 3 insertions(+), 10 deletions(-) diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c index c83e75a0ec12..0ce98e18b5a8 100644 --- a/drivers/pci/probe.c +++ b/drivers/pci/probe.c @@ -538,14 +538,10 @@ static void pci_read_bridge_windows(struct pci_dev *bridge) } if (io) { bridge->io_window = 1; - pci_read_bridge_io(bridge, - pci_resource_n(bridge, PCI_BRIDGE_IO_WINDOW), - true); + pci_read_bridge_io(bridge, &res, true); } - pci_read_bridge_mmio(bridge, - pci_resource_n(bridge, PCI_BRIDGE_MEM_WINDOW), - true); + pci_read_bridge_mmio(bridge, &res, true); /* * DECchip 21050 pass 2 errata: the bridge may miss an address @@ -583,10 +579,7 @@ static void pci_read_bridge_windows(struct pci_dev *bridge) bridge->pref_64_window = 1; } - pci_read_bridge_mmio_pref(bridge, - pci_resource_n(bridge, - PCI_BRIDGE_PREF_MEM_WINDOW), - true); + pci_read_bridge_mmio_pref(bridge, &res, true); } void pci_read_bridge_bases(struct pci_bus *child) From a4bbb493a3247ef32f6191fd8b2a0657139f8e08 Mon Sep 17 00:00:00 2001 From: Alison Schofield Date: Tue, 14 Oct 2025 14:38:49 -0700 Subject: [PATCH 184/330] cxl/trace: Subtract to find an hpa_alias0 in cxl_poison events Traces of cxl_poison events include an hpa_alias0 field if the poison address is in a region configured with an ELC, Extended Linear Cache. Since the ELC always comes first in the region, the calculation needs to subtract the ELC size from the calculated HPA address. Fixes: 8c520c5f1e76 ("cxl: Add extended linear cache address alias emission for cxl events") Signed-off-by: Alison Schofield Reviewed-by: Dave Jiang Signed-off-by: Dave Jiang --- drivers/cxl/core/trace.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/cxl/core/trace.h b/drivers/cxl/core/trace.h index a53ec4798b12..a972e4ef1936 100644 --- a/drivers/cxl/core/trace.h +++ b/drivers/cxl/core/trace.h @@ -1068,7 +1068,7 @@ TRACE_EVENT(cxl_poison, __entry->hpa = cxl_dpa_to_hpa(cxlr, cxlmd, __entry->dpa); if (__entry->hpa != ULLONG_MAX && cxlr->params.cache_size) - __entry->hpa_alias0 = __entry->hpa + + __entry->hpa_alias0 = __entry->hpa - cxlr->params.cache_size; else __entry->hpa_alias0 = ULLONG_MAX; From e603a342cf7ecd64ef8f36207dfe1caacb9e2583 Mon Sep 17 00:00:00 2001 From: Andrii Nakryiko Date: Tue, 14 Oct 2025 13:20:37 -0700 Subject: [PATCH 185/330] selftests/bpf: make arg_parsing.c more robust to crashes We started getting a crash in BPF CI, which seems to originate from test_parse_test_list_file() test and is happening at this line: ASSERT_OK(strcmp("test_with_spaces", set.tests[0].name), "test 0 name"); One way we can crash there is if set.cnt zero, which is checked for with ASSERT_EQ() above, but we proceed after this regardless of the outcome. Instead of crashing, we should bail out with test failure early. Similarly, if parse_test_list_file() fails, we shouldn't be even looking at set, so bail even earlier if ASSERT_OK() fails. Fixes: 64276f01dce8 ("selftests/bpf: Test_progs can read test lists from file") Signed-off-by: Andrii Nakryiko Tested-by: Ihor Solodrai Link: https://lore.kernel.org/r/20251014202037.72922-1-andrii@kernel.org Signed-off-by: Alexei Starovoitov --- tools/testing/selftests/bpf/prog_tests/arg_parsing.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/bpf/prog_tests/arg_parsing.c b/tools/testing/selftests/bpf/prog_tests/arg_parsing.c index bb143de68875..fbf0d9c2f58b 100644 --- a/tools/testing/selftests/bpf/prog_tests/arg_parsing.c +++ b/tools/testing/selftests/bpf/prog_tests/arg_parsing.c @@ -146,9 +146,12 @@ static void test_parse_test_list_file(void) init_test_filter_set(&set); - ASSERT_OK(parse_test_list_file(tmpfile, &set, true), "parse file"); + if (!ASSERT_OK(parse_test_list_file(tmpfile, &set, true), "parse file")) + goto out_fclose; + + if (!ASSERT_EQ(set.cnt, 4, "test count")) + goto out_free_set; - ASSERT_EQ(set.cnt, 4, "test count"); ASSERT_OK(strcmp("test_with_spaces", set.tests[0].name), "test 0 name"); ASSERT_EQ(set.tests[0].subtest_cnt, 0, "test 0 subtest count"); ASSERT_OK(strcmp("testA", set.tests[1].name), "test 1 name"); @@ -158,8 +161,8 @@ static void test_parse_test_list_file(void) ASSERT_OK(strcmp("testB", set.tests[2].name), "test 2 name"); ASSERT_OK(strcmp("testC_no_eof_newline", set.tests[3].name), "test 3 name"); +out_free_set: free_test_filter_set(&set); - out_fclose: fclose(fp); out_remove: From 82ebecdc74ff555daf70b811d854b1f32a296bea Mon Sep 17 00:00:00 2001 From: Jaehun Gou Date: Tue, 14 Oct 2025 22:01:46 +0900 Subject: [PATCH 186/330] exfat: fix improper check of dentry.stream.valid_size MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We found an infinite loop bug in the exFAT file system that can lead to a Denial-of-Service (DoS) condition. When a dentry in an exFAT filesystem is malformed, the following system calls — SYS_openat, SYS_ftruncate, and SYS_pwrite64 — can cause the kernel to hang. Root cause analysis shows that the size validation code in exfat_find() does not check whether dentry.stream.valid_size is negative. As a result, the system calls mentioned above can succeed and eventually trigger the DoS issue. This patch adds a check for negative dentry.stream.valid_size to prevent this vulnerability. Co-developed-by: Seunghun Han Signed-off-by: Seunghun Han Co-developed-by: Jihoon Kwon Signed-off-by: Jihoon Kwon Signed-off-by: Jaehun Gou Signed-off-by: Namjae Jeon --- fs/exfat/namei.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c index 7eb9c67fd35f..2364b49f050a 100644 --- a/fs/exfat/namei.c +++ b/fs/exfat/namei.c @@ -642,10 +642,14 @@ static int exfat_find(struct inode *dir, const struct qstr *qname, info->type = exfat_get_entry_type(ep); info->attr = le16_to_cpu(ep->dentry.file.attr); - info->size = le64_to_cpu(ep2->dentry.stream.valid_size); info->valid_size = le64_to_cpu(ep2->dentry.stream.valid_size); info->size = le64_to_cpu(ep2->dentry.stream.size); + if (info->valid_size < 0) { + exfat_fs_error(sb, "data valid size is invalid(%lld)", info->valid_size); + return -EIO; + } + if (unlikely(EXFAT_B_TO_CLU_ROUND_UP(info->size, sbi) > sbi->used_clusters)) { exfat_fs_error(sb, "data size is invalid(%lld)", info->size); return -EIO; From 6f719373b943a955fee6fc2012aed207b65e2854 Mon Sep 17 00:00:00 2001 From: Thomas Zimmermann Date: Tue, 14 Oct 2025 10:46:34 +0200 Subject: [PATCH 187/330] drm/ast: Blank with VGACR17 sync enable, always clear VGACRB6 sync off Blank the display by disabling sync pulses with VGACR17<7>. Unblank by reenabling them. This VGA setting should be supported by all Aspeed hardware. Ast currently blanks via sync-off bits in VGACRB6. Not all BMCs handle VGACRB6 correctly. After disabling sync during a reboot, some BMCs do not reenable it after the soft reset. The display output remains dark. When the display is off during boot, some BMCs set the sync-off bits in VGACRB6, so the display remains dark. Observed with Blackbird AST2500 BMCs. Clearing the sync-off bits unconditionally fixes these issues. Also do not modify VGASR1's SD bit for blanking, as it only disables GPU access to video memory. v2: - init vgacrb6 correctly (Jocelyn) Signed-off-by: Thomas Zimmermann Fixes: ce3d99c83495 ("drm: Call drm_atomic_helper_shutdown() at shutdown time for misc drivers") Tested-by: Nick Bowler Reported-by: Nick Bowler Closes: https://lore.kernel.org/dri-devel/wpwd7rit6t4mnu6kdqbtsnk5bhftgslio6e2jgkz6kgw6cuvvr@xbfswsczfqsi/ Cc: Douglas Anderson Cc: Dave Airlie Cc: Thomas Zimmermann Cc: Jocelyn Falempe Cc: dri-devel@lists.freedesktop.org Cc: # v6.7+ Reviewed-by: Jocelyn Falempe Link: https://lore.kernel.org/r/20251014084743.18242-1-tzimmermann@suse.de --- drivers/gpu/drm/ast/ast_mode.c | 18 ++++++++++-------- drivers/gpu/drm/ast/ast_reg.h | 1 + 2 files changed, 11 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c index b4e8edc7c767..30b011ed0a05 100644 --- a/drivers/gpu/drm/ast/ast_mode.c +++ b/drivers/gpu/drm/ast/ast_mode.c @@ -836,22 +836,24 @@ ast_crtc_helper_atomic_flush(struct drm_crtc *crtc, static void ast_crtc_helper_atomic_enable(struct drm_crtc *crtc, struct drm_atomic_state *state) { struct ast_device *ast = to_ast_device(crtc->dev); + u8 vgacr17 = 0x00; + u8 vgacrb6 = 0xff; - ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0xb6, 0xfc, 0x00); - ast_set_index_reg_mask(ast, AST_IO_VGASRI, 0x01, 0xdf, 0x00); + vgacr17 |= AST_IO_VGACR17_SYNC_ENABLE; + vgacrb6 &= ~(AST_IO_VGACRB6_VSYNC_OFF | AST_IO_VGACRB6_HSYNC_OFF); + + ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0x17, 0x7f, vgacr17); + ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0xb6, 0xfc, vgacrb6); } static void ast_crtc_helper_atomic_disable(struct drm_crtc *crtc, struct drm_atomic_state *state) { struct drm_crtc_state *old_crtc_state = drm_atomic_get_old_crtc_state(state, crtc); struct ast_device *ast = to_ast_device(crtc->dev); - u8 vgacrb6; + u8 vgacr17 = 0xff; - ast_set_index_reg_mask(ast, AST_IO_VGASRI, 0x01, 0xdf, AST_IO_VGASR1_SD); - - vgacrb6 = AST_IO_VGACRB6_VSYNC_OFF | - AST_IO_VGACRB6_HSYNC_OFF; - ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0xb6, 0xfc, vgacrb6); + vgacr17 &= ~AST_IO_VGACR17_SYNC_ENABLE; + ast_set_index_reg_mask(ast, AST_IO_VGACRI, 0x17, 0x7f, vgacr17); /* * HW cursors require the underlying primary plane and CRTC to diff --git a/drivers/gpu/drm/ast/ast_reg.h b/drivers/gpu/drm/ast/ast_reg.h index e15adaf3a80e..30578e3b07e4 100644 --- a/drivers/gpu/drm/ast/ast_reg.h +++ b/drivers/gpu/drm/ast/ast_reg.h @@ -29,6 +29,7 @@ #define AST_IO_VGAGRI (0x4E) #define AST_IO_VGACRI (0x54) +#define AST_IO_VGACR17_SYNC_ENABLE BIT(7) /* called "Hardware reset" in docs */ #define AST_IO_VGACR80_PASSWORD (0xa8) #define AST_IO_VGACR99_VGAMEM_RSRV_MASK GENMASK(1, 0) #define AST_IO_VGACRA1_VGAIO_DISABLED BIT(1) From a49c4d48c3b60926e6a8cec217bf95aa65388ecc Mon Sep 17 00:00:00 2001 From: Kurt Borja Date: Tue, 14 Oct 2025 05:07:27 -0500 Subject: [PATCH 188/330] platform/x86: alienware-wmi-wmax: Fix NULL pointer dereference in sleep handlers MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Devices without the AWCC interface don't initialize `awcc`. Add a check before dereferencing it in sleep handlers. Cc: stable@vger.kernel.org Reported-by: Gal Hammer Tested-by: Gal Hammer Fixes: 07ac275981b1 ("platform/x86: alienware-wmi-wmax: Add support for manual fan control") Signed-off-by: Kurt Borja Link: https://patch.msgid.link/20251014-sleep-fix-v3-1-b5cb58da4638@gmail.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen --- drivers/platform/x86/dell/alienware-wmi-wmax.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/platform/x86/dell/alienware-wmi-wmax.c b/drivers/platform/x86/dell/alienware-wmi-wmax.c index 31f9643a6a3b..b106e8e407b3 100644 --- a/drivers/platform/x86/dell/alienware-wmi-wmax.c +++ b/drivers/platform/x86/dell/alienware-wmi-wmax.c @@ -1639,7 +1639,7 @@ static int wmax_wmi_probe(struct wmi_device *wdev, const void *context) static int wmax_wmi_suspend(struct device *dev) { - if (awcc->hwmon) + if (awcc && awcc->hwmon) awcc_hwmon_suspend(dev); return 0; @@ -1647,7 +1647,7 @@ static int wmax_wmi_suspend(struct device *dev) static int wmax_wmi_resume(struct device *dev) { - if (awcc->hwmon) + if (awcc && awcc->hwmon) awcc_hwmon_resume(dev); return 0; From a7b4747d8e0e7871c3d4971cded1dcc9af6af9e9 Mon Sep 17 00:00:00 2001 From: David Thompson Date: Mon, 13 Oct 2025 15:56:05 +0000 Subject: [PATCH 189/330] platform/mellanox: mlxbf-pmc: add sysfs_attr_init() to count_clock init MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The lock-related debug logic (CONFIG_LOCK_STAT) in the kernel is noting the following warning when the BlueField-3 SOC is booted: BUG: key ffff00008a3402a8 has not been registered! ------------[ cut here ]------------ DEBUG_LOCKS_WARN_ON(1) WARNING: CPU: 4 PID: 592 at kernel/locking/lockdep.c:4801 lockdep_init_map_type+0x1d4/0x2a0 Call trace: lockdep_init_map_type+0x1d4/0x2a0 __kernfs_create_file+0x84/0x140 sysfs_add_file_mode_ns+0xcc/0x1cc internal_create_group+0x110/0x3d4 internal_create_groups.part.0+0x54/0xcc sysfs_create_groups+0x24/0x40 device_add+0x6e8/0x93c device_register+0x28/0x40 __hwmon_device_register+0x4b0/0x8a0 devm_hwmon_device_register_with_groups+0x7c/0xe0 mlxbf_pmc_probe+0x1e8/0x3e0 [mlxbf_pmc] platform_probe+0x70/0x110 The mlxbf_pmc driver must call sysfs_attr_init() during the initialization of the "count_clock" data structure to avoid this warning. Fixes: 5efc800975d9 ("platform/mellanox: mlxbf-pmc: Add support for monitoring cycle count") Reviewed-by: Shravan Kumar Ramani Signed-off-by: David Thompson Link: https://patch.msgid.link/20251013155605.3589770-1-davthompson@nvidia.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen --- drivers/platform/mellanox/mlxbf-pmc.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/platform/mellanox/mlxbf-pmc.c b/drivers/platform/mellanox/mlxbf-pmc.c index 4776013e0764..16a2fd9fdd9b 100644 --- a/drivers/platform/mellanox/mlxbf-pmc.c +++ b/drivers/platform/mellanox/mlxbf-pmc.c @@ -2015,6 +2015,7 @@ static int mlxbf_pmc_init_perftype_counter(struct device *dev, unsigned int blk_ if (pmc->block[blk_num].type == MLXBF_PMC_TYPE_CRSPACE) { /* Program crspace counters to count clock cycles using "count_clock" sysfs */ attr = &pmc->block[blk_num].attr_count_clock; + sysfs_attr_init(&attr->dev_attr.attr); attr->dev_attr.attr.mode = 0644; attr->dev_attr.show = mlxbf_pmc_count_clock_show; attr->dev_attr.store = mlxbf_pmc_count_clock_store; From f4f868baf292550acbfc35839213de9da0d9e8ab Mon Sep 17 00:00:00 2001 From: Denis Benato Date: Fri, 3 Oct 2025 20:49:49 +0200 Subject: [PATCH 190/330] MAINTAINERS: add Denis Benato as maintainer for asus notebooks MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Add myself as maintainer for "ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS" as suggested by Hans de Goede and Armin Wolf. Signed-off-by: Denis Benato Link: https://lore.kernel.org/all/8128cd6b-50e3-464c-90c2-781f61c3963e@gmail.com Reviewed-by: Hans de Goede Acked-by: Luke Jones Reviewed-by: Mario Limonciello (AMD) Link: https://patch.msgid.link/20251003184949.1083030-1-benato.denis96@gmail.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 46126ce2f968..832f3279ea83 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3841,6 +3841,7 @@ F: drivers/hwmon/asus-ec-sensors.c ASUS NOTEBOOKS AND EEEPC ACPI/WMI EXTRAS DRIVERS M: Corentin Chary M: Luke D. Jones +M: Denis Benato L: platform-driver-x86@vger.kernel.org S: Maintained W: https://asus-linux.org/ From 28412b489b088fb88dff488305fd4e56bd47f6e4 Mon Sep 17 00:00:00 2001 From: Jiaming Zhang Date: Wed, 15 Oct 2025 13:16:45 +0800 Subject: [PATCH 191/330] ALSA: usb-audio: Fix NULL pointer deference in try_to_register_card In try_to_register_card(), the return value of usb_ifnum_to_if() is passed directly to usb_interface_claimed() without a NULL check, which will lead to a NULL pointer dereference when creating an invalid USB audio device. Fix this by adding a check to ensure the interface pointer is valid before passing it to usb_interface_claimed(). Fixes: 39efc9c8a973 ("ALSA: usb-audio: Fix last interface check for registration") Closes: https://lore.kernel.org/all/CANypQFYtQxHL5ghREs-BujZG413RPJGnO5TH=xjFBKpPts33tA@mail.gmail.com/ Signed-off-by: Jiaming Zhang Signed-off-by: Takashi Iwai --- sound/usb/card.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/sound/usb/card.c b/sound/usb/card.c index 1d5a65eac933..270dad84d825 100644 --- a/sound/usb/card.c +++ b/sound/usb/card.c @@ -891,10 +891,16 @@ get_alias_quirk(struct usb_device *dev, unsigned int id) */ static int try_to_register_card(struct snd_usb_audio *chip, int ifnum) { + struct usb_interface *iface; + if (check_delayed_register_option(chip) == ifnum || - chip->last_iface == ifnum || - usb_interface_claimed(usb_ifnum_to_if(chip->dev, chip->last_iface))) + chip->last_iface == ifnum) return snd_card_register(chip->card); + + iface = usb_ifnum_to_if(chip->dev, chip->last_iface); + if (iface && usb_interface_claimed(iface)) + return snd_card_register(chip->card); + return 0; } From 34cbd6e07fddf36e186c8bf26a456fb7f50af44e Mon Sep 17 00:00:00 2001 From: tr1x_em Date: Thu, 25 Sep 2025 09:10:03 +0530 Subject: [PATCH 192/330] platform/x86: alienware-wmi-wmax: Add AWCC support to Dell G15 5530 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Makes alienware-wmi load on G15 5530 by default Cc: stable@vger.kernel.org Signed-off-by: Saumya Reviewed-by: Kurt Borja Link: https://patch.msgid.link/20250925034010.31414-1-admin@trix.is-a.dev Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen --- drivers/platform/x86/dell/alienware-wmi-wmax.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/platform/x86/dell/alienware-wmi-wmax.c b/drivers/platform/x86/dell/alienware-wmi-wmax.c index b106e8e407b3..f417dcc9af35 100644 --- a/drivers/platform/x86/dell/alienware-wmi-wmax.c +++ b/drivers/platform/x86/dell/alienware-wmi-wmax.c @@ -209,6 +209,14 @@ static const struct dmi_system_id awcc_dmi_table[] __initconst = { }, .driver_data = &g_series_quirks, }, + { + .ident = "Dell Inc. G15 5530", + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), + DMI_MATCH(DMI_PRODUCT_NAME, "Dell G15 5530"), + }, + .driver_data = &g_series_quirks, + }, { .ident = "Dell Inc. G16 7630", .matches = { From 2d8636119b92970ba135c3c4da87d24dbfdeb8ca Mon Sep 17 00:00:00 2001 From: Jeongjun Park Date: Wed, 15 Oct 2025 16:34:54 +0900 Subject: [PATCH 193/330] exfat: fix out-of-bounds in exfat_nls_to_ucs2() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since the len argument value passed to exfat_ioctl_set_volume_label() from exfat_nls_to_utf16() is passed 1 too large, an out-of-bounds read occurs when dereferencing p_cstring in exfat_nls_to_ucs2() later. And because of the NLS_NAME_OVERLEN macro, another error occurs when creating a file with a period at the end using utf8 and other iocharsets. So to avoid this, you should remove the code that uses NLS_NAME_OVERLEN macro and make the len argument value be the length of the label string, but with a maximum length of FSLABEL_MAX - 1. Reported-by: syzbot+98cc76a76de46b3714d4@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=98cc76a76de46b3714d4 Fixes: d01579d590f7 ("exfat: Add support for FS_IOC_{GET,SET}FSLABEL") Suggested-by: Pali Rohár Signed-off-by: Jeongjun Park Signed-off-by: Namjae Jeon --- fs/exfat/exfat_fs.h | 1 - fs/exfat/file.c | 7 ++++--- fs/exfat/namei.c | 2 +- fs/exfat/nls.c | 3 --- 4 files changed, 5 insertions(+), 8 deletions(-) diff --git a/fs/exfat/exfat_fs.h b/fs/exfat/exfat_fs.h index 329697c89d09..38210fb6901c 100644 --- a/fs/exfat/exfat_fs.h +++ b/fs/exfat/exfat_fs.h @@ -29,7 +29,6 @@ enum exfat_error_mode { enum { NLS_NAME_NO_LOSSY = 0, /* no lossy */ NLS_NAME_LOSSY = 1 << 0, /* just detected incorrect filename(s) */ - NLS_NAME_OVERLEN = 1 << 1, /* the length is over than its limit */ }; #define EXFAT_HASH_BITS 8 diff --git a/fs/exfat/file.c b/fs/exfat/file.c index f246cf439588..adc37b4d7fc2 100644 --- a/fs/exfat/file.c +++ b/fs/exfat/file.c @@ -509,8 +509,8 @@ static int exfat_ioctl_get_volume_label(struct super_block *sb, unsigned long ar static int exfat_ioctl_set_volume_label(struct super_block *sb, unsigned long arg) { - int ret = 0, lossy; - char label[FSLABEL_MAX]; + int ret = 0, lossy, label_len; + char label[FSLABEL_MAX] = {0}; struct exfat_uni_name uniname; if (!capable(CAP_SYS_ADMIN)) @@ -520,8 +520,9 @@ static int exfat_ioctl_set_volume_label(struct super_block *sb, return -EFAULT; memset(&uniname, 0, sizeof(uniname)); + label_len = strnlen(label, FSLABEL_MAX - 1); if (label[0]) { - ret = exfat_nls_to_utf16(sb, label, FSLABEL_MAX, + ret = exfat_nls_to_utf16(sb, label, label_len, &uniname, &lossy); if (ret < 0) return ret; diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c index 2364b49f050a..745dce29ddb5 100644 --- a/fs/exfat/namei.c +++ b/fs/exfat/namei.c @@ -442,7 +442,7 @@ static int __exfat_resolve_path(struct inode *inode, const unsigned char *path, return namelen; /* return error value */ if ((lossy && !lookup) || !namelen) - return (lossy & NLS_NAME_OVERLEN) ? -ENAMETOOLONG : -EINVAL; + return -EINVAL; return 0; } diff --git a/fs/exfat/nls.c b/fs/exfat/nls.c index 8243d94ceaf4..57db08a5271c 100644 --- a/fs/exfat/nls.c +++ b/fs/exfat/nls.c @@ -616,9 +616,6 @@ static int exfat_nls_to_ucs2(struct super_block *sb, unilen++; } - if (p_cstring[i] != '\0') - lossy |= NLS_NAME_OVERLEN; - *uniname = '\0'; p_uniname->name_len = unilen; p_uniname->name_hash = exfat_calc_chksum16(upname, unilen << 1, 0, From 57b00ab3d33d2b177bf45d568fa1e6203cd099b6 Mon Sep 17 00:00:00 2001 From: Shuming Fan Date: Tue, 14 Oct 2025 17:21:57 +0800 Subject: [PATCH 194/330] ASoC: sdw_utils: add rt1321 part id to codec_info_list The SdW machine driver supports the part id rt1321 amplifier. Signed-off-by: Shuming Fan Link: https://patch.msgid.link/20251014092157.2789054-1-shumingf@realtek.com Signed-off-by: Mark Brown --- sound/soc/sdw_utils/soc_sdw_utils.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/sound/soc/sdw_utils/soc_sdw_utils.c b/sound/soc/sdw_utils/soc_sdw_utils.c index 56c72ef27e7b..270c66b90228 100644 --- a/sound/soc/sdw_utils/soc_sdw_utils.c +++ b/sound/soc/sdw_utils/soc_sdw_utils.c @@ -311,6 +311,26 @@ struct asoc_sdw_codec_info codec_info_list[] = { }, .dai_num = 1, }, + { + .part_id = 0x1321, + .dais = { + { + .direction = {true, false}, + .dai_name = "rt1320-aif1", + .component_name = "rt1320", + .dai_type = SOC_SDW_DAI_TYPE_AMP, + .dailink = {SOC_SDW_AMP_OUT_DAI_ID, SOC_SDW_UNUSED_DAI_ID}, + .init = asoc_sdw_rt_amp_init, + .exit = asoc_sdw_rt_amp_exit, + .rtd_init = asoc_sdw_rt_amp_spk_rtd_init, + .controls = generic_spk_controls, + .num_controls = ARRAY_SIZE(generic_spk_controls), + .widgets = generic_spk_widgets, + .num_widgets = ARRAY_SIZE(generic_spk_widgets), + }, + }, + .dai_num = 1, + }, { .part_id = 0x714, .version_id = 3, From d5cda96d0130effd4255f7c5e720a58760a032a4 Mon Sep 17 00:00:00 2001 From: Srinivas Kandagatla Date: Wed, 8 Oct 2025 15:58:01 +0100 Subject: [PATCH 195/330] ASoC: codecs: wcd938x-sdw: remove redundant runtime pm calls Component bind callbacks already does runtime pm calls, soundwire codec also tries to do the exactly same thing resulting in Unbalanced pm_runtime_enable and disable calls. Remove the redundant calls from wcd938x-sdw driver. Reported-by: Konrad Dybcio Fixes: ebaf88c0546d ("ASoC: codecs: wcd-common: move component ops to common") Signed-off-by: Srinivas Kandagatla Link: https://patch.msgid.link/20251008145801.3479-1-srinivas.kandagatla@oss.qualcomm.com Signed-off-by: Mark Brown --- sound/soc/codecs/wcd938x-sdw.c | 22 ++++------------------ 1 file changed, 4 insertions(+), 18 deletions(-) diff --git a/sound/soc/codecs/wcd938x-sdw.c b/sound/soc/codecs/wcd938x-sdw.c index add907cb2706..8c8f39d04972 100644 --- a/sound/soc/codecs/wcd938x-sdw.c +++ b/sound/soc/codecs/wcd938x-sdw.c @@ -1207,24 +1207,14 @@ static int wcd9380_probe(struct sdw_slave *pdev, regcache_cache_only(wcd->regmap, true); } - pm_runtime_set_autosuspend_delay(dev, 3000); - pm_runtime_use_autosuspend(dev); - pm_runtime_mark_last_busy(dev); - pm_runtime_set_active(dev); - pm_runtime_enable(dev); - ret = component_add(dev, &wcd_sdw_component_ops); if (ret) - goto err_disable_rpm; + return ret; + + /* Set suspended until aggregate device is bind */ + pm_runtime_set_suspended(dev); return 0; - -err_disable_rpm: - pm_runtime_disable(dev); - pm_runtime_set_suspended(dev); - pm_runtime_dont_use_autosuspend(dev); - - return ret; } static int wcd9380_remove(struct sdw_slave *pdev) @@ -1233,10 +1223,6 @@ static int wcd9380_remove(struct sdw_slave *pdev) component_del(dev, &wcd_sdw_component_ops); - pm_runtime_disable(dev); - pm_runtime_set_suspended(dev); - pm_runtime_dont_use_autosuspend(dev); - return 0; } From 5fb750e8a9ae123b2034771b864b8a21dbef65cd Mon Sep 17 00:00:00 2001 From: Alexei Starovoitov Date: Tue, 14 Oct 2025 17:07:00 -0700 Subject: [PATCH 196/330] bpf: Replace bpf_map_kmalloc_node() with kmalloc_nolock() to allocate bpf_async_cb structures. The following kmemleak splat: [ 8.105530] kmemleak: Trying to color unknown object at 0xff11000100e918c0 as Black [ 8.106521] Call Trace: [ 8.106521] [ 8.106521] dump_stack_lvl+0x4b/0x70 [ 8.106521] kvfree_call_rcu+0xcb/0x3b0 [ 8.106521] ? hrtimer_cancel+0x21/0x40 [ 8.106521] bpf_obj_free_fields+0x193/0x200 [ 8.106521] htab_map_update_elem+0x29c/0x410 [ 8.106521] bpf_prog_cfc8cd0f42c04044_overwrite_cb+0x47/0x4b [ 8.106521] bpf_prog_8c30cd7c4db2e963_overwrite_timer+0x65/0x86 [ 8.106521] bpf_prog_test_run_syscall+0xe1/0x2a0 happens due to the combination of features and fixes, but mainly due to commit 6d78b4473cdb ("bpf: Tell memcg to use allow_spinning=false path in bpf_timer_init()") It's using __GFP_HIGH, which instructs slub/kmemleak internals to skip kmemleak_alloc_recursive() on allocation, so subsequent kfree_rcu()-> kvfree_call_rcu()->kmemleak_ignore() complains with the above splat. To fix this imbalance, replace bpf_map_kmalloc_node() with kmalloc_nolock() and kfree_rcu() with call_rcu() + kfree_nolock() to make sure that the objects allocated with kmalloc_nolock() are freed with kfree_nolock() rather than the implicit kfree() that kfree_rcu() uses internally. Note, the kmalloc_nolock() happens under bpf_spin_lock_irqsave(), so it will always fail in PREEMPT_RT. This is not an issue at the moment, since bpf_timers are disabled in PREEMPT_RT. In the future bpf_spin_lock will be replaced with state machine similar to bpf_task_work. Fixes: 6d78b4473cdb ("bpf: Tell memcg to use allow_spinning=false path in bpf_timer_init()") Signed-off-by: Alexei Starovoitov Signed-off-by: Daniel Borkmann Reviewed-by: Shakeel Butt Acked-by: Harry Yoo Acked-by: Vlastimil Babka Cc: linux-mm@kvack.org Link: https://lore.kernel.org/bpf/20251015000700.28988-1-alexei.starovoitov@gmail.com --- include/linux/bpf.h | 4 ++++ kernel/bpf/helpers.c | 25 ++++++++++++++----------- kernel/bpf/syscall.c | 15 +++++++++++++++ 3 files changed, 33 insertions(+), 11 deletions(-) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index a98c83346134..d808253f2e94 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -2499,6 +2499,8 @@ int bpf_map_alloc_pages(const struct bpf_map *map, int nid, #ifdef CONFIG_MEMCG void *bpf_map_kmalloc_node(const struct bpf_map *map, size_t size, gfp_t flags, int node); +void *bpf_map_kmalloc_nolock(const struct bpf_map *map, size_t size, gfp_t flags, + int node); void *bpf_map_kzalloc(const struct bpf_map *map, size_t size, gfp_t flags); void *bpf_map_kvcalloc(struct bpf_map *map, size_t n, size_t size, gfp_t flags); @@ -2511,6 +2513,8 @@ void __percpu *bpf_map_alloc_percpu(const struct bpf_map *map, size_t size, */ #define bpf_map_kmalloc_node(_map, _size, _flags, _node) \ kmalloc_node(_size, _flags, _node) +#define bpf_map_kmalloc_nolock(_map, _size, _flags, _node) \ + kmalloc_nolock(_size, _flags, _node) #define bpf_map_kzalloc(_map, _size, _flags) \ kzalloc(_size, _flags) #define bpf_map_kvcalloc(_map, _n, _size, _flags) \ diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c index c9fab9a356df..8eb117c52817 100644 --- a/kernel/bpf/helpers.c +++ b/kernel/bpf/helpers.c @@ -1215,13 +1215,20 @@ static void bpf_wq_work(struct work_struct *work) rcu_read_unlock_trace(); } +static void bpf_async_cb_rcu_free(struct rcu_head *rcu) +{ + struct bpf_async_cb *cb = container_of(rcu, struct bpf_async_cb, rcu); + + kfree_nolock(cb); +} + static void bpf_wq_delete_work(struct work_struct *work) { struct bpf_work *w = container_of(work, struct bpf_work, delete_work); cancel_work_sync(&w->work); - kfree_rcu(w, cb.rcu); + call_rcu(&w->cb.rcu, bpf_async_cb_rcu_free); } static void bpf_timer_delete_work(struct work_struct *work) @@ -1230,13 +1237,13 @@ static void bpf_timer_delete_work(struct work_struct *work) /* Cancel the timer and wait for callback to complete if it was running. * If hrtimer_cancel() can be safely called it's safe to call - * kfree_rcu(t) right after for both preallocated and non-preallocated + * call_rcu() right after for both preallocated and non-preallocated * maps. The async->cb = NULL was already done and no code path can see * address 't' anymore. Timer if armed for existing bpf_hrtimer before * bpf_timer_cancel_and_free will have been cancelled. */ hrtimer_cancel(&t->timer); - kfree_rcu(t, cb.rcu); + call_rcu(&t->cb.rcu, bpf_async_cb_rcu_free); } static int __bpf_async_init(struct bpf_async_kern *async, struct bpf_map *map, u64 flags, @@ -1270,11 +1277,7 @@ static int __bpf_async_init(struct bpf_async_kern *async, struct bpf_map *map, u goto out; } - /* Allocate via bpf_map_kmalloc_node() for memcg accounting. Until - * kmalloc_nolock() is available, avoid locking issues by using - * __GFP_HIGH (GFP_ATOMIC & ~__GFP_RECLAIM). - */ - cb = bpf_map_kmalloc_node(map, size, __GFP_HIGH, map->numa_node); + cb = bpf_map_kmalloc_nolock(map, size, 0, map->numa_node); if (!cb) { ret = -ENOMEM; goto out; @@ -1315,7 +1318,7 @@ static int __bpf_async_init(struct bpf_async_kern *async, struct bpf_map *map, u * or pinned in bpffs. */ WRITE_ONCE(async->cb, NULL); - kfree(cb); + kfree_nolock(cb); ret = -EPERM; } out: @@ -1580,7 +1583,7 @@ void bpf_timer_cancel_and_free(void *val) * timer _before_ calling us, such that failing to cancel it here will * cause it to possibly use struct hrtimer after freeing bpf_hrtimer. * Therefore, we _need_ to cancel any outstanding timers before we do - * kfree_rcu, even though no more timers can be armed. + * call_rcu, even though no more timers can be armed. * * Moreover, we need to schedule work even if timer does not belong to * the calling callback_fn, as on two different CPUs, we can end up in a @@ -1607,7 +1610,7 @@ void bpf_timer_cancel_and_free(void *val) * completion. */ if (hrtimer_try_to_cancel(&t->timer) >= 0) - kfree_rcu(t, cb.rcu); + call_rcu(&t->cb.rcu, bpf_async_cb_rcu_free); else queue_work(system_dfl_wq, &t->cb.delete_work); } else { diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index 2a9456a3e730..8a129746bd6c 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -520,6 +520,21 @@ void *bpf_map_kmalloc_node(const struct bpf_map *map, size_t size, gfp_t flags, return ptr; } +void *bpf_map_kmalloc_nolock(const struct bpf_map *map, size_t size, gfp_t flags, + int node) +{ + struct mem_cgroup *memcg, *old_memcg; + void *ptr; + + memcg = bpf_map_get_memcg(map); + old_memcg = set_active_memcg(memcg); + ptr = kmalloc_nolock(size, flags | __GFP_ACCOUNT, node); + set_active_memcg(old_memcg); + mem_cgroup_put(memcg); + + return ptr; +} + void *bpf_map_kzalloc(const struct bpf_map *map, size_t size, gfp_t flags) { struct mem_cgroup *memcg, *old_memcg; From 6fced056d2cc8d01b326e6fcfabaacb9850b71a4 Mon Sep 17 00:00:00 2001 From: ZhangGuoDong Date: Sun, 12 Oct 2025 00:47:59 +0800 Subject: [PATCH 197/330] smb/server: fix possible memory leak in smb2_read() Memory leak occurs when ksmbd_vfs_read() fails. Fix this by adding the missing kvfree(). Co-developed-by: ChenXiaoSong Signed-off-by: ChenXiaoSong Signed-off-by: ZhangGuoDong Acked-by: Namjae Jeon Signed-off-by: Steve French --- fs/smb/server/smb2pdu.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/smb/server/smb2pdu.c b/fs/smb/server/smb2pdu.c index ab1d45fcebde..e81e615f322a 100644 --- a/fs/smb/server/smb2pdu.c +++ b/fs/smb/server/smb2pdu.c @@ -6824,6 +6824,7 @@ int smb2_read(struct ksmbd_work *work) nbytes = ksmbd_vfs_read(work, fp, length, &offset, aux_payload_buf); if (nbytes < 0) { + kvfree(aux_payload_buf); err = nbytes; goto out; } From 379510a815cb2e64eb0a379cb62295d6ade65df0 Mon Sep 17 00:00:00 2001 From: ZhangGuoDong Date: Sun, 12 Oct 2025 00:51:36 +0800 Subject: [PATCH 198/330] smb/server: fix possible refcount leak in smb2_sess_setup() Reference count of ksmbd_session will leak when session need reconnect. Fix this by adding the missing ksmbd_user_session_put(). Co-developed-by: ChenXiaoSong Signed-off-by: ChenXiaoSong Signed-off-by: ZhangGuoDong Acked-by: Namjae Jeon Signed-off-by: Steve French --- fs/smb/server/smb2pdu.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/smb/server/smb2pdu.c b/fs/smb/server/smb2pdu.c index e81e615f322a..b731d9b09408 100644 --- a/fs/smb/server/smb2pdu.c +++ b/fs/smb/server/smb2pdu.c @@ -1806,6 +1806,7 @@ int smb2_sess_setup(struct ksmbd_work *work) if (ksmbd_conn_need_reconnect(conn)) { rc = -EFAULT; + ksmbd_user_session_put(sess); sess = NULL; goto out_err; } From 88f170814fea74911ceab798a43cbd7c5599bed4 Mon Sep 17 00:00:00 2001 From: Marios Makassikis Date: Wed, 15 Oct 2025 09:25:46 +0200 Subject: [PATCH 199/330] ksmbd: fix recursive locking in RPC handle list access Since commit 305853cce3794 ("ksmbd: Fix race condition in RPC handle list access"), ksmbd_session_rpc_method() attempts to lock sess->rpc_lock. This causes hung connections / tasks when a client attempts to open a named pipe. Using Samba's rpcclient tool: $ rpcclient //192.168.1.254 -U user%password $ rpcclient $> srvinfo Kernel side: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message. task:kworker/0:0 state:D stack:0 pid:5021 tgid:5021 ppid:2 flags:0x00200000 Workqueue: ksmbd-io handle_ksmbd_work Call trace: __schedule from schedule+0x3c/0x58 schedule from schedule_preempt_disabled+0xc/0x10 schedule_preempt_disabled from rwsem_down_read_slowpath+0x1b0/0x1d8 rwsem_down_read_slowpath from down_read+0x28/0x30 down_read from ksmbd_session_rpc_method+0x18/0x3c ksmbd_session_rpc_method from ksmbd_rpc_open+0x34/0x68 ksmbd_rpc_open from ksmbd_session_rpc_open+0x194/0x228 ksmbd_session_rpc_open from create_smb2_pipe+0x8c/0x2c8 create_smb2_pipe from smb2_open+0x10c/0x27ac smb2_open from handle_ksmbd_work+0x238/0x3dc handle_ksmbd_work from process_scheduled_works+0x160/0x25c process_scheduled_works from worker_thread+0x16c/0x1e8 worker_thread from kthread+0xa8/0xb8 kthread from ret_from_fork+0x14/0x38 Exception stack(0x8529ffb0 to 0x8529fff8) The task deadlocks because the lock is already held: ksmbd_session_rpc_open down_write(&sess->rpc_lock) ksmbd_rpc_open ksmbd_session_rpc_method down_read(&sess->rpc_lock) <-- deadlock Adjust ksmbd_session_rpc_method() callers to take the lock when necessary. Fixes: 305853cce3794 ("ksmbd: Fix race condition in RPC handle list access") Signed-off-by: Marios Makassikis Acked-by: Namjae Jeon Signed-off-by: Steve French --- fs/smb/server/mgmt/user_session.c | 7 ++----- fs/smb/server/smb2pdu.c | 9 ++++++++- fs/smb/server/transport_ipc.c | 12 ++++++++++++ 3 files changed, 22 insertions(+), 6 deletions(-) diff --git a/fs/smb/server/mgmt/user_session.c b/fs/smb/server/mgmt/user_session.c index 6fa025374f2f..1c181ef99929 100644 --- a/fs/smb/server/mgmt/user_session.c +++ b/fs/smb/server/mgmt/user_session.c @@ -147,14 +147,11 @@ void ksmbd_session_rpc_close(struct ksmbd_session *sess, int id) int ksmbd_session_rpc_method(struct ksmbd_session *sess, int id) { struct ksmbd_session_rpc *entry; - int method; - down_read(&sess->rpc_lock); + lockdep_assert_held(&sess->rpc_lock); entry = xa_load(&sess->rpc_handle_list, id); - method = entry ? entry->method : 0; - up_read(&sess->rpc_lock); - return method; + return entry ? entry->method : 0; } void ksmbd_session_destroy(struct ksmbd_session *sess) diff --git a/fs/smb/server/smb2pdu.c b/fs/smb/server/smb2pdu.c index b731d9b09408..f901ae18e68a 100644 --- a/fs/smb/server/smb2pdu.c +++ b/fs/smb/server/smb2pdu.c @@ -4626,8 +4626,15 @@ static int smb2_get_info_file_pipe(struct ksmbd_session *sess, * pipe without opening it, checking error condition here */ id = req->VolatileFileId; - if (!ksmbd_session_rpc_method(sess, id)) + + lockdep_assert_not_held(&sess->rpc_lock); + + down_read(&sess->rpc_lock); + if (!ksmbd_session_rpc_method(sess, id)) { + up_read(&sess->rpc_lock); return -ENOENT; + } + up_read(&sess->rpc_lock); ksmbd_debug(SMB, "FileInfoClass %u, FileId 0x%llx\n", req->FileInfoClass, req->VolatileFileId); diff --git a/fs/smb/server/transport_ipc.c b/fs/smb/server/transport_ipc.c index 2aa1b29bea08..46f87fd1ce1c 100644 --- a/fs/smb/server/transport_ipc.c +++ b/fs/smb/server/transport_ipc.c @@ -825,6 +825,9 @@ struct ksmbd_rpc_command *ksmbd_rpc_write(struct ksmbd_session *sess, int handle if (!msg) return NULL; + lockdep_assert_not_held(&sess->rpc_lock); + + down_read(&sess->rpc_lock); msg->type = KSMBD_EVENT_RPC_REQUEST; req = (struct ksmbd_rpc_command *)msg->payload; req->handle = handle; @@ -833,6 +836,7 @@ struct ksmbd_rpc_command *ksmbd_rpc_write(struct ksmbd_session *sess, int handle req->flags |= KSMBD_RPC_WRITE_METHOD; req->payload_sz = payload_sz; memcpy(req->payload, payload, payload_sz); + up_read(&sess->rpc_lock); resp = ipc_msg_send_request(msg, req->handle); ipc_msg_free(msg); @@ -849,6 +853,9 @@ struct ksmbd_rpc_command *ksmbd_rpc_read(struct ksmbd_session *sess, int handle) if (!msg) return NULL; + lockdep_assert_not_held(&sess->rpc_lock); + + down_read(&sess->rpc_lock); msg->type = KSMBD_EVENT_RPC_REQUEST; req = (struct ksmbd_rpc_command *)msg->payload; req->handle = handle; @@ -856,6 +863,7 @@ struct ksmbd_rpc_command *ksmbd_rpc_read(struct ksmbd_session *sess, int handle) req->flags |= rpc_context_flags(sess); req->flags |= KSMBD_RPC_READ_METHOD; req->payload_sz = 0; + up_read(&sess->rpc_lock); resp = ipc_msg_send_request(msg, req->handle); ipc_msg_free(msg); @@ -876,6 +884,9 @@ struct ksmbd_rpc_command *ksmbd_rpc_ioctl(struct ksmbd_session *sess, int handle if (!msg) return NULL; + lockdep_assert_not_held(&sess->rpc_lock); + + down_read(&sess->rpc_lock); msg->type = KSMBD_EVENT_RPC_REQUEST; req = (struct ksmbd_rpc_command *)msg->payload; req->handle = handle; @@ -884,6 +895,7 @@ struct ksmbd_rpc_command *ksmbd_rpc_ioctl(struct ksmbd_session *sess, int handle req->flags |= KSMBD_RPC_IOCTL_METHOD; req->payload_sz = payload_sz; memcpy(req->payload, payload, payload_sz); + up_read(&sess->rpc_lock); resp = ipc_msg_send_request(msg, req->handle); ipc_msg_free(msg); From b0432201a11b3caaeca6c03f2b3e399275b2e489 Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Sun, 12 Oct 2025 21:10:30 +0200 Subject: [PATCH 200/330] smb: client: let destroy_mr_list() keep smbdirect_mr_io memory if registered If a smbdirect_mr_io structure if still visible to callers of smbd_register_mr() we can't free the related memory when the connection is disconnected! Otherwise smbd_deregister_mr() will crash. Now we use a mutex and refcounting in order to keep the memory around if the connection is disconnected. It means smbd_deregister_mr() can be called at any later time to free the memory, which is no longer referenced by nor referencing the connection. It also means smbd_destroy() no longer needs to wait for mr_io.used.count to become 0. Fixes: 050b8c374019 ("smbd: Make upper layer decide when to destroy the transport") Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 146 +++++++++++++++++++++++++++++++++----- 1 file changed, 127 insertions(+), 19 deletions(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index c3330e43488f..77de85d7cdc3 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -1624,19 +1624,7 @@ void smbd_destroy(struct TCP_Server_Info *server) log_rdma_event(INFO, "free receive buffers\n"); destroy_receive_buffers(sc); - /* - * For performance reasons, memory registration and deregistration - * are not locked by srv_mutex. It is possible some processes are - * blocked on transport srv_mutex while holding memory registration. - * Release the transport srv_mutex to allow them to hit the failure - * path when sending data, and then release memory registrations. - */ log_rdma_event(INFO, "freeing mr list\n"); - while (atomic_read(&sc->mr_io.used.count)) { - cifs_server_unlock(server); - msleep(1000); - cifs_server_lock(server); - } destroy_mr_list(sc); ib_free_cq(sc->ib.send_cq); @@ -2352,6 +2340,46 @@ static void smbd_mr_recovery_work(struct work_struct *work) } } +static void smbd_mr_disable_locked(struct smbdirect_mr_io *mr) +{ + struct smbdirect_socket *sc = mr->socket; + + lockdep_assert_held(&mr->mutex); + + if (mr->state == SMBDIRECT_MR_DISABLED) + return; + + if (mr->mr) + ib_dereg_mr(mr->mr); + if (mr->sgt.nents) + ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); + kfree(mr->sgt.sgl); + + mr->mr = NULL; + mr->sgt.sgl = NULL; + mr->sgt.nents = 0; + + mr->state = SMBDIRECT_MR_DISABLED; +} + +static void smbd_mr_free_locked(struct kref *kref) +{ + struct smbdirect_mr_io *mr = + container_of(kref, struct smbdirect_mr_io, kref); + + lockdep_assert_held(&mr->mutex); + + /* + * smbd_mr_disable_locked() should already be called! + */ + if (WARN_ON_ONCE(mr->state != SMBDIRECT_MR_DISABLED)) + smbd_mr_disable_locked(mr); + + mutex_unlock(&mr->mutex); + mutex_destroy(&mr->mutex); + kfree(mr); +} + static void destroy_mr_list(struct smbdirect_socket *sc) { struct smbdirect_mr_io *mr, *tmp; @@ -2365,13 +2393,31 @@ static void destroy_mr_list(struct smbdirect_socket *sc) spin_unlock_irqrestore(&sc->mr_io.all.lock, flags); list_for_each_entry_safe(mr, tmp, &all_list, list) { - if (mr->mr) - ib_dereg_mr(mr->mr); - if (mr->sgt.nents) - ib_dma_unmap_sg(sc->ib.dev, mr->sgt.sgl, mr->sgt.nents, mr->dir); - kfree(mr->sgt.sgl); + mutex_lock(&mr->mutex); + + smbd_mr_disable_locked(mr); list_del(&mr->list); - kfree(mr); + mr->socket = NULL; + + /* + * No kref_put_mutex() as it's already locked. + * + * If smbd_mr_free_locked() is called + * and the mutex is unlocked and mr is gone, + * in that case kref_put() returned 1. + * + * If kref_put() returned 0 we know that + * smbd_mr_free_locked() didn't + * run. Not by us nor by anyone else, as we + * still hold the mutex, so we need to unlock. + * + * If the mr is still registered it will + * be dangling (detached from the connection + * waiting for smbd_deregister_mr() to be + * called in order to free the memory. + */ + if (!kref_put(&mr->kref, smbd_mr_free_locked)) + mutex_unlock(&mr->mutex); } } @@ -2402,6 +2448,9 @@ static int allocate_mr_list(struct smbdirect_socket *sc) goto kzalloc_mr_failed; } + kref_init(&mr->kref); + mutex_init(&mr->mutex); + mr->mr = ib_alloc_mr(sc->ib.pd, sc->mr_io.type, sp->max_frmr_depth); @@ -2434,6 +2483,7 @@ static int allocate_mr_list(struct smbdirect_socket *sc) kcalloc_sgl_failed: ib_dereg_mr(mr->mr); ib_alloc_mr_failed: + mutex_destroy(&mr->mutex); kfree(mr); kzalloc_mr_failed: destroy_mr_list(sc); @@ -2471,6 +2521,7 @@ static struct smbdirect_mr_io *get_mr(struct smbdirect_socket *sc) list_for_each_entry(ret, &sc->mr_io.all.list, list) { if (ret->state == SMBDIRECT_MR_READY) { ret->state = SMBDIRECT_MR_REGISTERED; + kref_get(&ret->kref); spin_unlock_irqrestore(&sc->mr_io.all.lock, flags); atomic_dec(&sc->mr_io.ready.count); atomic_inc(&sc->mr_io.used.count); @@ -2535,6 +2586,8 @@ struct smbdirect_mr_io *smbd_register_mr(struct smbd_connection *info, return NULL; } + mutex_lock(&mr->mutex); + mr->dir = writing ? DMA_FROM_DEVICE : DMA_TO_DEVICE; mr->need_invalidate = need_invalidate; mr->sgt.nents = 0; @@ -2578,8 +2631,16 @@ struct smbdirect_mr_io *smbd_register_mr(struct smbd_connection *info, * on the next ib_post_send when we actually send I/O to remote peer */ rc = ib_post_send(sc->ib.qp, ®_wr->wr, NULL); - if (!rc) + if (!rc) { + /* + * get_mr() gave us a reference + * via kref_get(&mr->kref), we keep that and let + * the caller use smbd_deregister_mr() + * to remove it again. + */ + mutex_unlock(&mr->mutex); return mr; + } log_rdma_mr(ERR, "ib_post_send failed rc=%x reg_wr->key=%x\n", rc, reg_wr->key); @@ -2596,6 +2657,25 @@ struct smbdirect_mr_io *smbd_register_mr(struct smbd_connection *info, smbd_disconnect_rdma_connection(sc); + /* + * get_mr() gave us a reference + * via kref_get(&mr->kref), we need to remove it again + * on error. + * + * No kref_put_mutex() as it's already locked. + * + * If smbd_mr_free_locked() is called + * and the mutex is unlocked and mr is gone, + * in that case kref_put() returned 1. + * + * If kref_put() returned 0 we know that + * smbd_mr_free_locked() didn't + * run. Not by us nor by anyone else, as we + * still hold the mutex, so we need to unlock. + */ + if (!kref_put(&mr->kref, smbd_mr_free_locked)) + mutex_unlock(&mr->mutex); + return NULL; } @@ -2624,6 +2704,15 @@ void smbd_deregister_mr(struct smbdirect_mr_io *mr) { struct smbdirect_socket *sc = mr->socket; + mutex_lock(&mr->mutex); + if (mr->state == SMBDIRECT_MR_DISABLED) + goto put_kref; + + if (sc->status != SMBDIRECT_SOCKET_CONNECTED) { + smbd_mr_disable_locked(mr); + goto put_kref; + } + if (mr->need_invalidate) { struct ib_send_wr *wr = &mr->inv_wr; int rc; @@ -2640,6 +2729,7 @@ void smbd_deregister_mr(struct smbdirect_mr_io *mr) rc = ib_post_send(sc->ib.qp, wr, NULL); if (rc) { log_rdma_mr(ERR, "ib_post_send failed rc=%x\n", rc); + smbd_mr_disable_locked(mr); smbd_disconnect_rdma_connection(sc); goto done; } @@ -2671,6 +2761,24 @@ void smbd_deregister_mr(struct smbdirect_mr_io *mr) done: if (atomic_dec_and_test(&sc->mr_io.used.count)) wake_up(&sc->mr_io.cleanup.wait_queue); + +put_kref: + /* + * No kref_put_mutex() as it's already locked. + * + * If smbd_mr_free_locked() is called + * and the mutex is unlocked and mr is gone, + * in that case kref_put() returned 1. + * + * If kref_put() returned 0 we know that + * smbd_mr_free_locked() didn't + * run. Not by us nor by anyone else, as we + * still hold the mutex, so we need to unlock + * and keep the mr in SMBDIRECT_MR_READY or + * SMBDIRECT_MR_ERROR state. + */ + if (!kref_put(&mr->kref, smbd_mr_free_locked)) + mutex_unlock(&mr->mutex); } static bool smb_set_sge(struct smb_extract_to_rdma *rdma, From d877470b59910b5c50383d634dda3782386bba51 Mon Sep 17 00:00:00 2001 From: ZhangGuoDong Date: Mon, 13 Oct 2025 00:17:30 +0800 Subject: [PATCH 201/330] smb: move some duplicate definitions to common/cifsglob.h In order to maintain the code more easily, move duplicate definitions to new common header file. Co-developed-by: ChenXiaoSong Signed-off-by: ChenXiaoSong Signed-off-by: ZhangGuoDong Signed-off-by: Steve French --- fs/smb/client/cifsglob.h | 19 +------------------ fs/smb/common/cifsglob.h | 30 ++++++++++++++++++++++++++++++ fs/smb/server/smb_common.h | 14 +------------- 3 files changed, 32 insertions(+), 31 deletions(-) create mode 100644 fs/smb/common/cifsglob.h diff --git a/fs/smb/client/cifsglob.h b/fs/smb/client/cifsglob.h index 8f6f567d7474..c5034cf9ac9e 100644 --- a/fs/smb/client/cifsglob.h +++ b/fs/smb/client/cifsglob.h @@ -24,6 +24,7 @@ #include "cifsacl.h" #include #include +#include "../common/cifsglob.h" #include "../common/smb2pdu.h" #include "smb2pdu.h" #include @@ -702,12 +703,6 @@ get_rfc1002_length(void *buf) return be32_to_cpu(*((__be32 *)buf)) & 0xffffff; } -static inline void -inc_rfc1001_len(void *buf, int count) -{ - be32_add_cpu((__be32 *)buf, count); -} - struct TCP_Server_Info { struct list_head tcp_ses_list; struct list_head smb_ses_list; @@ -1021,8 +1016,6 @@ compare_mid(__u16 mid, const struct smb_hdr *smb) #define CIFS_MAX_RFC1002_WSIZE ((1<<17) - 1 - sizeof(WRITE_REQ) + 4) #define CIFS_MAX_RFC1002_RSIZE ((1<<17) - 1 - sizeof(READ_RSP) + 4) -#define CIFS_DEFAULT_IOSIZE (1024 * 1024) - /* * Windows only supports a max of 60kb reads and 65535 byte writes. Default to * those values when posix extensions aren't in force. In actuality here, we @@ -2148,30 +2141,20 @@ extern mempool_t cifs_io_request_pool; extern mempool_t cifs_io_subrequest_pool; /* Operations for different SMB versions */ -#define SMB1_VERSION_STRING "1.0" -#define SMB20_VERSION_STRING "2.0" #ifdef CONFIG_CIFS_ALLOW_INSECURE_LEGACY extern struct smb_version_operations smb1_operations; extern struct smb_version_values smb1_values; extern struct smb_version_operations smb20_operations; extern struct smb_version_values smb20_values; #endif /* CIFS_ALLOW_INSECURE_LEGACY */ -#define SMB21_VERSION_STRING "2.1" extern struct smb_version_operations smb21_operations; extern struct smb_version_values smb21_values; -#define SMBDEFAULT_VERSION_STRING "default" extern struct smb_version_values smbdefault_values; -#define SMB3ANY_VERSION_STRING "3" extern struct smb_version_values smb3any_values; -#define SMB30_VERSION_STRING "3.0" extern struct smb_version_operations smb30_operations; extern struct smb_version_values smb30_values; -#define SMB302_VERSION_STRING "3.02" -#define ALT_SMB302_VERSION_STRING "3.0.2" /*extern struct smb_version_operations smb302_operations;*/ /* not needed yet */ extern struct smb_version_values smb302_values; -#define SMB311_VERSION_STRING "3.1.1" -#define ALT_SMB311_VERSION_STRING "3.11" extern struct smb_version_operations smb311_operations; extern struct smb_version_values smb311_values; diff --git a/fs/smb/common/cifsglob.h b/fs/smb/common/cifsglob.h new file mode 100644 index 000000000000..00fd215e3eb5 --- /dev/null +++ b/fs/smb/common/cifsglob.h @@ -0,0 +1,30 @@ +/* SPDX-License-Identifier: LGPL-2.1 */ +/* + * + * Copyright (C) International Business Machines Corp., 2002,2008 + * Author(s): Steve French (sfrench@us.ibm.com) + * Jeremy Allison (jra@samba.org) + * + */ +#ifndef _COMMON_CIFS_GLOB_H +#define _COMMON_CIFS_GLOB_H + +static inline void inc_rfc1001_len(void *buf, int count) +{ + be32_add_cpu((__be32 *)buf, count); +} + +#define SMB1_VERSION_STRING "1.0" +#define SMB20_VERSION_STRING "2.0" +#define SMB21_VERSION_STRING "2.1" +#define SMBDEFAULT_VERSION_STRING "default" +#define SMB3ANY_VERSION_STRING "3" +#define SMB30_VERSION_STRING "3.0" +#define SMB302_VERSION_STRING "3.02" +#define ALT_SMB302_VERSION_STRING "3.0.2" +#define SMB311_VERSION_STRING "3.1.1" +#define ALT_SMB311_VERSION_STRING "3.11" + +#define CIFS_DEFAULT_IOSIZE (1024 * 1024) + +#endif /* _COMMON_CIFS_GLOB_H */ diff --git a/fs/smb/server/smb_common.h b/fs/smb/server/smb_common.h index d742ba754348..863716207a0d 100644 --- a/fs/smb/server/smb_common.h +++ b/fs/smb/server/smb_common.h @@ -10,6 +10,7 @@ #include "glob.h" #include "nterr.h" +#include "../common/cifsglob.h" #include "../common/smb2pdu.h" #include "smb2pdu.h" @@ -26,16 +27,8 @@ #define SMB311_PROT 6 #define BAD_PROT 0xFFFF -#define SMB1_VERSION_STRING "1.0" -#define SMB20_VERSION_STRING "2.0" -#define SMB21_VERSION_STRING "2.1" -#define SMB30_VERSION_STRING "3.0" -#define SMB302_VERSION_STRING "3.02" -#define SMB311_VERSION_STRING "3.1.1" - #define SMB_ECHO_INTERVAL (60 * HZ) -#define CIFS_DEFAULT_IOSIZE (64 * 1024) #define MAX_CIFS_SMALL_BUFFER_SIZE 448 /* big enough for most */ #define MAX_STREAM_PROT_LEN 0x00FFFFFF @@ -464,9 +457,4 @@ static inline unsigned int get_rfc1002_len(void *buf) { return be32_to_cpu(*((__be32 *)buf)) & 0xffffff; } - -static inline void inc_rfc1001_len(void *buf, int count) -{ - be32_add_cpu((__be32 *)buf, count); -} #endif /* __SMB_COMMON_H__ */ From 85d7dda5a9f665ea579741ec873a8841f37e8943 Mon Sep 17 00:00:00 2001 From: "Mario Limonciello (AMD)" Date: Tue, 23 Sep 2025 10:29:29 -0500 Subject: [PATCH 202/330] cpufreq/amd-pstate: Fix a regression leading to EPP 0 after hibernate After resuming from S4, all CPUs except the boot CPU have the wrong EPP hint programmed. This is because when the CPUs were offlined the EPP value was reset to 0. This is a similar problem as fixed by commit ba3319e590571 ("cpufreq/amd-pstate: Fix a regression leading to EPP 0 after resume") and the solution is also similar. When offlining rather than reset the values to zero, reset them to match those chosen by the policy. When the CPUs are onlined again these values will be restored. Closes: https://community.frame.work/t/increased-power-usage-after-resuming-from-suspend-on-ryzen-7040-kernel-6-15-regression/74531/20?u=mario_limonciello Fixes: 608a76b65288 ("cpufreq/amd-pstate: Add support for the "Requested CPU Min frequency" BIOS option") Reviewed-by: Gautham R. Shenoy Signed-off-by: Mario Limonciello (AMD) --- drivers/cpufreq/amd-pstate.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c index 298e92d8cc03..b44f0f7a5ba1 100644 --- a/drivers/cpufreq/amd-pstate.c +++ b/drivers/cpufreq/amd-pstate.c @@ -1614,7 +1614,11 @@ static int amd_pstate_cpu_offline(struct cpufreq_policy *policy) * min_perf value across kexec reboots. If this CPU is just onlined normally after this, the * limits, epp and desired perf will get reset to the cached values in cpudata struct */ - return amd_pstate_update_perf(policy, perf.bios_min_perf, 0U, 0U, 0U, false); + return amd_pstate_update_perf(policy, perf.bios_min_perf, + FIELD_GET(AMD_CPPC_DES_PERF_MASK, cpudata->cppc_req_cached), + FIELD_GET(AMD_CPPC_MAX_PERF_MASK, cpudata->cppc_req_cached), + FIELD_GET(AMD_CPPC_EPP_PERF_MASK, cpudata->cppc_req_cached), + false); } static int amd_pstate_suspend(struct cpufreq_policy *policy) From dc96cefef0d3032c69e46a21b345c60e56b18934 Mon Sep 17 00:00:00 2001 From: Yu Kuai Date: Wed, 15 Oct 2025 09:48:27 +0800 Subject: [PATCH 203/330] blk-mq: fix stale tag depth for shared sched tags in blk_mq_update_nr_requests() Commit 7f2799c546db ("blk-mq: cleanup shared tags case in blk_mq_update_nr_requests()") moves blk_mq_tag_update_sched_shared_tags() before q->nr_requests is updated, however, it's still using the old q->nr_requests to resize tag depth. Fix this problem by passing in expected new tag depth. Fixes: 7f2799c546db ("blk-mq: cleanup shared tags case in blk_mq_update_nr_requests()") Signed-off-by: Yu Kuai Reviewed-by: Ming Lei Reviewed-by: Nilay Shroff Reported-by: Chris Mason Link: https://lore.kernel.org/linux-block/20251014130507.4187235-2-clm@meta.com/ Signed-off-by: Jens Axboe --- block/blk-mq-sched.c | 2 +- block/blk-mq-tag.c | 5 +++-- block/blk-mq.c | 2 +- block/blk-mq.h | 3 ++- 4 files changed, 7 insertions(+), 5 deletions(-) diff --git a/block/blk-mq-sched.c b/block/blk-mq-sched.c index d06bb137a743..e0bed16485c3 100644 --- a/block/blk-mq-sched.c +++ b/block/blk-mq-sched.c @@ -557,7 +557,7 @@ int blk_mq_init_sched(struct request_queue *q, struct elevator_type *e, if (blk_mq_is_shared_tags(flags)) { /* Shared tags are stored at index 0 in @et->tags. */ q->sched_shared_tags = et->tags[0]; - blk_mq_tag_update_sched_shared_tags(q); + blk_mq_tag_update_sched_shared_tags(q, et->nr_requests); } queue_for_each_hw_ctx(q, hctx, i) { diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c index c7a4d4b9cc87..5b664dbdf655 100644 --- a/block/blk-mq-tag.c +++ b/block/blk-mq-tag.c @@ -622,10 +622,11 @@ void blk_mq_tag_resize_shared_tags(struct blk_mq_tag_set *set, unsigned int size sbitmap_queue_resize(&tags->bitmap_tags, size - set->reserved_tags); } -void blk_mq_tag_update_sched_shared_tags(struct request_queue *q) +void blk_mq_tag_update_sched_shared_tags(struct request_queue *q, + unsigned int nr) { sbitmap_queue_resize(&q->sched_shared_tags->bitmap_tags, - q->nr_requests - q->tag_set->reserved_tags); + nr - q->tag_set->reserved_tags); } /** diff --git a/block/blk-mq.c b/block/blk-mq.c index 09f579414161..d626d32f6e57 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -4941,7 +4941,7 @@ struct elevator_tags *blk_mq_update_nr_requests(struct request_queue *q, * tags can't grow, see blk_mq_alloc_sched_tags(). */ if (q->elevator) - blk_mq_tag_update_sched_shared_tags(q); + blk_mq_tag_update_sched_shared_tags(q, nr); else blk_mq_tag_resize_shared_tags(set, nr); } else if (!q->elevator) { diff --git a/block/blk-mq.h b/block/blk-mq.h index af42dc018808..c4fccdeb5441 100644 --- a/block/blk-mq.h +++ b/block/blk-mq.h @@ -186,7 +186,8 @@ void blk_mq_put_tag(struct blk_mq_tags *tags, struct blk_mq_ctx *ctx, void blk_mq_put_tags(struct blk_mq_tags *tags, int *tag_array, int nr_tags); void blk_mq_tag_resize_shared_tags(struct blk_mq_tag_set *set, unsigned int size); -void blk_mq_tag_update_sched_shared_tags(struct request_queue *q); +void blk_mq_tag_update_sched_shared_tags(struct request_queue *q, + unsigned int nr); void blk_mq_tag_wakeup_all(struct blk_mq_tags *tags, bool); void blk_mq_queue_tag_busy_iter(struct request_queue *q, busy_tag_iter_fn *fn, From 08823e89e3e269bf4c4a20b4c24a8119920cc7a4 Mon Sep 17 00:00:00 2001 From: Ming Lei Date: Wed, 15 Oct 2025 18:30:39 +0800 Subject: [PATCH 204/330] block: Remove elevator_lock usage from blkg_conf frozen operations Remove the acquisition and release of q->elevator_lock in the blkg_conf_open_bdev_frozen() and blkg_conf_exit_frozen() functions. The elevator lock is no longer needed in these code paths since commit 78c271344b6f ("block: move wbt_enable_default() out of queue freezing from sched ->exit()") which introduces `disk->rqos_state_mutex` for protecting wbt state change, and not necessary to abuse elevator_lock for this purpose. This change helps to solve the lockdep warning reported from Yu Kuai[1]. Pass blktests/throtl with lockdep enabled. Links: https://lore.kernel.org/linux-block/e5e7ac3f-2063-473a-aafb-4d8d43e5576e@yukuai.org.cn/ [1] Fixes: commit 78c271344b6f ("block: move wbt_enable_default() out of queue freezing from sched ->exit()") Signed-off-by: Ming Lei Reviewed-by: Nilay Shroff Signed-off-by: Jens Axboe --- block/blk-cgroup.c | 13 ++++--------- 1 file changed, 4 insertions(+), 9 deletions(-) diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c index f93de34fe87d..3cffb68ba5d8 100644 --- a/block/blk-cgroup.c +++ b/block/blk-cgroup.c @@ -812,8 +812,7 @@ int blkg_conf_open_bdev(struct blkg_conf_ctx *ctx) } /* * Similar to blkg_conf_open_bdev, but additionally freezes the queue, - * acquires q->elevator_lock, and ensures the correct locking order - * between q->elevator_lock and q->rq_qos_mutex. + * ensures the correct locking order between freeze queue and q->rq_qos_mutex. * * This function returns negative error on failure. On success it returns * memflags which must be saved and later passed to blkg_conf_exit_frozen @@ -834,13 +833,11 @@ unsigned long __must_check blkg_conf_open_bdev_frozen(struct blkg_conf_ctx *ctx) * At this point, we haven’t started protecting anything related to QoS, * so we release q->rq_qos_mutex here, which was first acquired in blkg_ * conf_open_bdev. Later, we re-acquire q->rq_qos_mutex after freezing - * the queue and acquiring q->elevator_lock to maintain the correct - * locking order. + * the queue to maintain the correct locking order. */ mutex_unlock(&ctx->bdev->bd_queue->rq_qos_mutex); memflags = blk_mq_freeze_queue(ctx->bdev->bd_queue); - mutex_lock(&ctx->bdev->bd_queue->elevator_lock); mutex_lock(&ctx->bdev->bd_queue->rq_qos_mutex); return memflags; @@ -995,9 +992,8 @@ void blkg_conf_exit(struct blkg_conf_ctx *ctx) EXPORT_SYMBOL_GPL(blkg_conf_exit); /* - * Similar to blkg_conf_exit, but also unfreezes the queue and releases - * q->elevator_lock. Should be used when blkg_conf_open_bdev_frozen - * is used to open the bdev. + * Similar to blkg_conf_exit, but also unfreezes the queue. Should be used + * when blkg_conf_open_bdev_frozen is used to open the bdev. */ void blkg_conf_exit_frozen(struct blkg_conf_ctx *ctx, unsigned long memflags) { @@ -1005,7 +1001,6 @@ void blkg_conf_exit_frozen(struct blkg_conf_ctx *ctx, unsigned long memflags) struct request_queue *q = ctx->bdev->bd_queue; blkg_conf_exit(ctx); - mutex_unlock(&q->elevator_lock); blk_mq_unfreeze_queue(q, memflags); } } From be7cab44ed099566c605a8dac686c3254db01b35 Mon Sep 17 00:00:00 2001 From: Pavel Begunkov Date: Wed, 15 Oct 2025 13:07:23 +0100 Subject: [PATCH 205/330] io_uring: protect mem region deregistration io_create_region_mmap_safe() protects publishing of a region against concurrent mmap calls, however we should also protect against it when removing a region. There is a gap io_register_mem_region() where it safely publishes a region, but then copy_to_user goes wrong and it unsafely frees the region. Cc: stable@vger.kernel.org Fixes: 087f997870a94 ("io_uring/memmap: implement mmap for regions") Signed-off-by: Pavel Begunkov Signed-off-by: Jens Axboe --- io_uring/register.c | 1 + 1 file changed, 1 insertion(+) diff --git a/io_uring/register.c b/io_uring/register.c index 43f04c47522c..58d43d624856 100644 --- a/io_uring/register.c +++ b/io_uring/register.c @@ -613,6 +613,7 @@ static int io_register_mem_region(struct io_ring_ctx *ctx, void __user *uarg) if (ret) return ret; if (copy_to_user(rd_uptr, &rd, sizeof(rd))) { + guard(mutex)(&ctx->mmap_lock); io_free_region(ctx, &ctx->param_region); return -EFAULT; } From 437c23357d897f5b5b7d297c477da44b56654d46 Mon Sep 17 00:00:00 2001 From: Pavel Begunkov Date: Wed, 15 Oct 2025 13:10:31 +0100 Subject: [PATCH 206/330] io_uring: fix unexpected placement on same size resizing There might be many reasons why a user is resizing a ring, e.g. moving to huge pages or for some memory compaction using IORING_SETUP_NO_MMAP. Don't bypass resizing, the user will definitely be surprised seeing 0 while the rings weren't actually moved to a new place. Signed-off-by: Pavel Begunkov Signed-off-by: Jens Axboe --- io_uring/register.c | 7 ------- 1 file changed, 7 deletions(-) diff --git a/io_uring/register.c b/io_uring/register.c index 58d43d624856..2e4717f1357c 100644 --- a/io_uring/register.c +++ b/io_uring/register.c @@ -421,13 +421,6 @@ static int io_register_resize_rings(struct io_ring_ctx *ctx, void __user *arg) if (unlikely(ret)) return ret; - /* nothing to do, but copy params back */ - if (p.sq_entries == ctx->sq_entries && p.cq_entries == ctx->cq_entries) { - if (copy_to_user(arg, &p, sizeof(p))) - return -EFAULT; - return 0; - } - size = rings_size(p.flags, p.sq_entries, p.cq_entries, &sq_array_offset); if (size == SIZE_MAX) From 95355766e5871e9cdc574be5a3b115392ad33aea Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jouni=20H=C3=B6gander?= Date: Mon, 22 Sep 2025 13:27:25 +0300 Subject: [PATCH 207/330] drm/i915/psr: Deactivate PSR only on LNL and when selective fetch enabled MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Using intel_psr_exit in frontbuffer flush on older platforms seems to be causing problems. Sending single full frame update using intel_psr_force_update is anyways more optimal compared to psr deactivate/activate -> move back to this approach on PSR1, PSR HW tracking and Panel Replay full frame update and use deactivate/activate only on LunarLake and only when selective fetch is enabled. Tested-by: Lemen Tested-by: Koos Vriezen Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/14946 Signed-off-by: Jouni Högander Reviewed-by: Mika Kahola Link: https://lore.kernel.org/r/20250922102725.2752742-1-jouni.hogander@intel.com (cherry picked from commit 924adb0bbdd8fef25fd229c76e3f602c3e8752ee) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_psr.c | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c index 01bf304c705f..10eb93a34cf2 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.c +++ b/drivers/gpu/drm/i915/display/intel_psr.c @@ -3402,6 +3402,7 @@ static void _psr_flush_handle(struct intel_dp *intel_dp) struct intel_display *display = to_intel_display(intel_dp); if (DISPLAY_VER(display) < 20 && intel_dp->psr.psr2_sel_fetch_enabled) { + /* Selective fetch prior LNL */ if (intel_dp->psr.psr2_sel_fetch_cff_enabled) { /* can we turn CFF off? */ if (intel_dp->psr.busy_frontbuffer_bits == 0) @@ -3420,12 +3421,19 @@ static void _psr_flush_handle(struct intel_dp *intel_dp) intel_psr_configure_full_frame_update(intel_dp); intel_psr_force_update(intel_dp); + } else if (!intel_dp->psr.psr2_sel_fetch_enabled) { + /* + * PSR1 on all platforms + * PSR2 HW tracking + * Panel Replay Full frame update + */ + intel_psr_force_update(intel_dp); } else { + /* Selective update LNL onwards */ intel_psr_exit(intel_dp); } - if ((!intel_dp->psr.psr2_sel_fetch_enabled || DISPLAY_VER(display) >= 20) && - !intel_dp->psr.busy_frontbuffer_bits) + if (!intel_dp->psr.active && !intel_dp->psr.busy_frontbuffer_bits) queue_work(display->wq.unordered, &intel_dp->psr.work); } From 3f9c60f4d3cc3e7f4dd7cac05011ea06d512050f Mon Sep 17 00:00:00 2001 From: Sebastian Chlad Date: Wed, 15 Oct 2025 12:33:56 +0200 Subject: [PATCH 208/330] selftests: cgroup: add values_close_report helper Some cgroup selftests, such as test_cpu, occasionally fail by a very small margin and if run in the CI context, it is useful to have detailed diagnostic output to understand the deviation. Introduce a values_close_report() helper which performs the same comparison as values_close(), but prints detailed information when the values differ beyond the allowed tolerance. Signed-off-by: Sebastian Chlad Signed-off-by: Tejun Heo --- .../cgroup/lib/include/cgroup_util.h | 20 +++++++++++++++++++ 1 file changed, 20 insertions(+) diff --git a/tools/testing/selftests/cgroup/lib/include/cgroup_util.h b/tools/testing/selftests/cgroup/lib/include/cgroup_util.h index 9dc90a1b386d..7ab2824ed7b5 100644 --- a/tools/testing/selftests/cgroup/lib/include/cgroup_util.h +++ b/tools/testing/selftests/cgroup/lib/include/cgroup_util.h @@ -25,6 +25,26 @@ static inline int values_close(long a, long b, int err) return labs(a - b) <= (a + b) / 100 * err; } +/* + * Checks if two given values differ by less than err% of their sum and assert + * with detailed debug info if not. + */ +static inline int values_close_report(long a, long b, int err) +{ + long diff = labs(a - b); + long limit = (a + b) / 100 * err; + double actual_err = (a + b) ? (100.0 * diff / (a + b)) : 0.0; + int close = diff <= limit; + + if (!close) + fprintf(stderr, + "[FAIL] actual=%ld expected=%ld | diff=%ld | limit=%ld | " + "tolerance=%d%% | actual_error=%.2f%%\n", + a, b, diff, limit, err, actual_err); + + return close; +} + extern ssize_t read_text(const char *path, char *buf, size_t max_len); extern ssize_t write_text(const char *path, char *buf, ssize_t len); From 4cdde87d723a0552f475c8c6b0db472a6945125f Mon Sep 17 00:00:00 2001 From: Sebastian Chlad Date: Wed, 15 Oct 2025 12:33:57 +0200 Subject: [PATCH 209/330] selftests: cgroup: Use values_close_report in test_cpu Convert test_cpu to use the newly added values_close_report() helper to print detailed diagnostics when a tolerance check fails. This provides clearer insight into deviations while run in the CI. Signed-off-by: Sebastian Chlad Signed-off-by: Tejun Heo --- tools/testing/selftests/cgroup/test_cpu.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/tools/testing/selftests/cgroup/test_cpu.c b/tools/testing/selftests/cgroup/test_cpu.c index 2a60e6c41940..d54e2317efff 100644 --- a/tools/testing/selftests/cgroup/test_cpu.c +++ b/tools/testing/selftests/cgroup/test_cpu.c @@ -219,7 +219,7 @@ static int test_cpucg_stats(const char *root) if (user_usec <= 0) goto cleanup; - if (!values_close(usage_usec, expected_usage_usec, 1)) + if (!values_close_report(usage_usec, expected_usage_usec, 1)) goto cleanup; ret = KSFT_PASS; @@ -291,7 +291,7 @@ static int test_cpucg_nice(const char *root) user_usec = cg_read_key_long(cpucg, "cpu.stat", "user_usec"); nice_usec = cg_read_key_long(cpucg, "cpu.stat", "nice_usec"); - if (!values_close(nice_usec, expected_nice_usec, 1)) + if (!values_close_report(nice_usec, expected_nice_usec, 1)) goto cleanup; ret = KSFT_PASS; @@ -404,7 +404,7 @@ overprovision_validate(const struct cpu_hogger *children, int num_children) goto cleanup; delta = children[i + 1].usage - children[i].usage; - if (!values_close(delta, children[0].usage, 35)) + if (!values_close_report(delta, children[0].usage, 35)) goto cleanup; } @@ -444,7 +444,7 @@ underprovision_validate(const struct cpu_hogger *children, int num_children) int ret = KSFT_FAIL, i; for (i = 0; i < num_children - 1; i++) { - if (!values_close(children[i + 1].usage, children[0].usage, 15)) + if (!values_close_report(children[i + 1].usage, children[0].usage, 15)) goto cleanup; } @@ -573,16 +573,16 @@ run_cpucg_nested_weight_test(const char *root, bool overprovisioned) nested_leaf_usage = leaf[1].usage + leaf[2].usage; if (overprovisioned) { - if (!values_close(leaf[0].usage, nested_leaf_usage, 15)) + if (!values_close_report(leaf[0].usage, nested_leaf_usage, 15)) goto cleanup; - } else if (!values_close(leaf[0].usage * 2, nested_leaf_usage, 15)) + } else if (!values_close_report(leaf[0].usage * 2, nested_leaf_usage, 15)) goto cleanup; child_usage = cg_read_key_long(child, "cpu.stat", "usage_usec"); if (child_usage <= 0) goto cleanup; - if (!values_close(child_usage, nested_leaf_usage, 1)) + if (!values_close_report(child_usage, nested_leaf_usage, 1)) goto cleanup; ret = KSFT_PASS; @@ -691,7 +691,7 @@ static int test_cpucg_max(const char *root) expected_usage_usec = n_periods * quota_usec + MIN(remainder_usec, quota_usec); - if (!values_close(usage_usec, expected_usage_usec, 10)) + if (!values_close_report(usage_usec, expected_usage_usec, 10)) goto cleanup; ret = KSFT_PASS; @@ -762,7 +762,7 @@ static int test_cpucg_max_nested(const char *root) expected_usage_usec = n_periods * quota_usec + MIN(remainder_usec, quota_usec); - if (!values_close(usage_usec, expected_usage_usec, 10)) + if (!values_close_report(usage_usec, expected_usage_usec, 10)) goto cleanup; ret = KSFT_PASS; From 0187c08058da3e7f11b356ac27e0c427d36f33f2 Mon Sep 17 00:00:00 2001 From: Dmitry Torokhov Date: Tue, 14 Oct 2025 21:28:44 -0700 Subject: [PATCH 210/330] HID: hid-input: only ignore 0 battery events for digitizers MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Commit 581c4484769e ("HID: input: map digitizer battery usage") added handling of battery events for digitizers (typically for batteries presented in stylii). Digitizers typically report correct battery levels only when stylus is actively touching the surface, and in other cases they may report battery level of 0. To avoid confusing consumers of the battery information the code was added to filer out reports with 0 battery levels. However there exist other kinds of devices that may legitimately report 0 battery levels. Fix this by filtering out 0-level reports only for digitizer usages, and continue reporting them for other kinds of devices (Smart Batteries, etc). Reported-by: 卢国宏 Fixes: 581c4484769e ("HID: input: map digitizer battery usage") Signed-off-by: Dmitry Torokhov Signed-off-by: Jiri Kosina --- drivers/hid/hid-input.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/hid/hid-input.c b/drivers/hid/hid-input.c index 5d7532d79d21..e56e7de53279 100644 --- a/drivers/hid/hid-input.c +++ b/drivers/hid/hid-input.c @@ -635,7 +635,10 @@ static void hidinput_update_battery(struct hid_device *dev, unsigned int usage, return; } - if (value == 0 || value < dev->battery_min || value > dev->battery_max) + if ((usage & HID_USAGE_PAGE) == HID_UP_DIGITIZER && value == 0) + return; + + if (value < dev->battery_min || value > dev->battery_max) return; capacity = hidinput_scale_battery_capacity(dev, value); From aa4daea418ee4215dca5c8636090660c545cb233 Mon Sep 17 00:00:00 2001 From: Thadeu Lima de Souza Cascardo Date: Wed, 8 Oct 2025 09:40:33 -0300 Subject: [PATCH 211/330] HID: multitouch: fix name of Stylus input devices HID_DG_PEN devices should have a suffix of "Stylus", as pointed out by commit c0ee1d571626 ("HID: hid-input: Add suffix also for HID_DG_PEN"). However, on multitouch devices, these suffixes may be overridden. Before that commit, HID_DG_PEN devices would get the "Stylus" suffix, but after that, multitouch would override them to have an "UNKNOWN" suffix. Just add HID_DG_PEN to the list of non-overriden suffixes in multitouch. Before this fix: [ 0.470981] input: ELAN9008:00 04F3:2E14 UNKNOWN as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-16/i2c-ELAN9008:00/0018:04F3:2E14.0001/input/input8 ELAN9008:00 04F3:2E14 UNKNOWN After this fix: [ 0.474332] input: ELAN9008:00 04F3:2E14 Stylus as /devices/pci0000:00/0000:00:15.1/i2c_designware.1/i2c-16/i2c-ELAN9008:00/0018:04F3:2E14.0001/input/input8 ELAN9008:00 04F3:2E14 Stylus Fixes: c0ee1d571626 ("HID: hid-input: Add suffix also for HID_DG_PEN") Signed-off-by: Thadeu Lima de Souza Cascardo Reviewed-by: Mika Westerberg Signed-off-by: Jiri Kosina --- drivers/hid/hid-multitouch.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/hid/hid-multitouch.c b/drivers/hid/hid-multitouch.c index 2879e65cf303..513b8673ad8d 100644 --- a/drivers/hid/hid-multitouch.c +++ b/drivers/hid/hid-multitouch.c @@ -1742,6 +1742,7 @@ static int mt_input_configured(struct hid_device *hdev, struct hid_input *hi) case HID_CP_CONSUMER_CONTROL: case HID_GD_WIRELESS_RADIO_CTLS: case HID_GD_SYSTEM_MULTIAXIS: + case HID_DG_PEN: /* already handled by hid core */ break; case HID_DG_TOUCHSCREEN: From 46f781e0d151844589dc2125c8cce3300546f92a Mon Sep 17 00:00:00 2001 From: Benjamin Tissoires Date: Wed, 8 Oct 2025 16:06:58 +0200 Subject: [PATCH 212/330] HID: multitouch: fix sticky fingers The sticky fingers quirk (MT_QUIRK_STICKY_FINGERS) was only considering the case when slots were not released during the last report. This can be problematic if the firmware forgets to release a finger while others are still present. This was observed on the Synaptics DLL0945 touchpad found on the Dell XPS 9310 and the Dell Inspiron 5406. Fixes: 4f4001bc76fd ("HID: multitouch: fix rare Win 8 cases when the touch up event gets missing") Cc: stable@vger.kernel.org Signed-off-by: Benjamin Tissoires Signed-off-by: Jiri Kosina --- drivers/hid/hid-multitouch.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) diff --git a/drivers/hid/hid-multitouch.c b/drivers/hid/hid-multitouch.c index 513b8673ad8d..179dc316b4b5 100644 --- a/drivers/hid/hid-multitouch.c +++ b/drivers/hid/hid-multitouch.c @@ -94,9 +94,8 @@ enum report_mode { TOUCHPAD_REPORT_ALL = TOUCHPAD_REPORT_BUTTONS | TOUCHPAD_REPORT_CONTACTS, }; -#define MT_IO_FLAGS_RUNNING 0 -#define MT_IO_FLAGS_ACTIVE_SLOTS 1 -#define MT_IO_FLAGS_PENDING_SLOTS 2 +#define MT_IO_SLOTS_MASK GENMASK(7, 0) /* reserve first 8 bits for slot tracking */ +#define MT_IO_FLAGS_RUNNING 32 static const bool mtrue = true; /* default for true */ static const bool mfalse; /* default for false */ @@ -172,7 +171,11 @@ struct mt_device { struct timer_list release_timer; /* to release sticky fingers */ struct hid_haptic_device *haptic; /* haptic related configuration */ struct hid_device *hdev; /* hid_device we're attached to */ - unsigned long mt_io_flags; /* mt flags (MT_IO_FLAGS_*) */ + unsigned long mt_io_flags; /* mt flags (MT_IO_FLAGS_RUNNING) + * first 8 bits are reserved for keeping the slot + * states, this is fine because we only support up + * to 250 slots (MT_MAX_MAXCONTACT) + */ __u8 inputmode_value; /* InputMode HID feature value */ __u8 maxcontacts; bool is_buttonpad; /* is this device a button pad? */ @@ -986,6 +989,7 @@ static void mt_release_pending_palms(struct mt_device *td, for_each_set_bit(slotnum, app->pending_palm_slots, td->maxcontacts) { clear_bit(slotnum, app->pending_palm_slots); + clear_bit(slotnum, &td->mt_io_flags); input_mt_slot(input, slotnum); input_mt_report_slot_inactive(input); @@ -1019,12 +1023,6 @@ static void mt_sync_frame(struct mt_device *td, struct mt_application *app, app->left_button_state = 0; if (td->is_haptic_touchpad) hid_haptic_pressure_reset(td->haptic); - - if (test_bit(MT_IO_FLAGS_ACTIVE_SLOTS, &td->mt_io_flags)) - set_bit(MT_IO_FLAGS_PENDING_SLOTS, &td->mt_io_flags); - else - clear_bit(MT_IO_FLAGS_PENDING_SLOTS, &td->mt_io_flags); - clear_bit(MT_IO_FLAGS_ACTIVE_SLOTS, &td->mt_io_flags); } static int mt_compute_timestamp(struct mt_application *app, __s32 value) @@ -1202,7 +1200,9 @@ static int mt_process_slot(struct mt_device *td, struct input_dev *input, input_event(input, EV_ABS, ABS_MT_TOUCH_MAJOR, major); input_event(input, EV_ABS, ABS_MT_TOUCH_MINOR, minor); - set_bit(MT_IO_FLAGS_ACTIVE_SLOTS, &td->mt_io_flags); + set_bit(slotnum, &td->mt_io_flags); + } else { + clear_bit(slotnum, &td->mt_io_flags); } return 0; @@ -1337,7 +1337,7 @@ static void mt_touch_report(struct hid_device *hid, * defect. */ if (app->quirks & MT_QUIRK_STICKY_FINGERS) { - if (test_bit(MT_IO_FLAGS_PENDING_SLOTS, &td->mt_io_flags)) + if (td->mt_io_flags & MT_IO_SLOTS_MASK) mod_timer(&td->release_timer, jiffies + msecs_to_jiffies(100)); else @@ -1814,6 +1814,7 @@ static void mt_release_contacts(struct hid_device *hid) for (i = 0; i < mt->num_slots; i++) { input_mt_slot(input_dev, i); input_mt_report_slot_inactive(input_dev); + clear_bit(i, &td->mt_io_flags); } input_mt_sync_frame(input_dev); input_sync(input_dev); @@ -1836,7 +1837,7 @@ static void mt_expired_timeout(struct timer_list *t) */ if (test_and_set_bit_lock(MT_IO_FLAGS_RUNNING, &td->mt_io_flags)) return; - if (test_bit(MT_IO_FLAGS_PENDING_SLOTS, &td->mt_io_flags)) + if (td->mt_io_flags & MT_IO_SLOTS_MASK) mt_release_contacts(hdev); clear_bit_unlock(MT_IO_FLAGS_RUNNING, &td->mt_io_flags); } From d9b3014a7f1425011909ad358dc0c8f187853a12 Mon Sep 17 00:00:00 2001 From: Benjamin Tissoires Date: Wed, 8 Oct 2025 16:06:59 +0200 Subject: [PATCH 213/330] selftests/hid: add tests for missing release on the Dell Synaptics Add a simple test for the corner case not currently covered by the sticky fingers quirk. Because it's a corner case test, we only test this on a couple of devices, not on all of them because the value of adding the same test over and over is rather moot. Signed-off-by: Benjamin Tissoires Signed-off-by: Jiri Kosina --- .../selftests/hid/tests/test_multitouch.py | 55 +++++++++++++++++++ 1 file changed, 55 insertions(+) diff --git a/tools/testing/selftests/hid/tests/test_multitouch.py b/tools/testing/selftests/hid/tests/test_multitouch.py index 5d2ffa3d5977..ece0ba8e7d34 100644 --- a/tools/testing/selftests/hid/tests/test_multitouch.py +++ b/tools/testing/selftests/hid/tests/test_multitouch.py @@ -1752,6 +1752,52 @@ class TestWin8TSConfidence(BaseTest.TestWin8Multitouch): assert evdev.slots[0][libevdev.EV_ABS.ABS_MT_TRACKING_ID] == -1 + @pytest.mark.skip_if_uhdev( + lambda uhdev: "Confidence" not in uhdev.fields, + "Device not compatible, missing Confidence usage", + ) + def test_mt_confidence_bad_multi_release(self): + """Check for the sticky finger being properly detected. + + We first inject 3 fingers, then release only the second. + After 100 ms, we should receive a generated event about the + 2 missing fingers being released. + """ + uhdev = self.uhdev + evdev = uhdev.get_evdev() + + # send 3 touches + t0 = Touch(1, 50, 10) + t1 = Touch(2, 150, 100) + t2 = Touch(3, 250, 200) + r = uhdev.event([t0, t1, t2]) + events = uhdev.next_sync_events() + self.debug_reports(r, uhdev, events) + + # release the second + t1.tipswitch = False + r = uhdev.event([t1]) + events = uhdev.next_sync_events() + self.debug_reports(r, uhdev, events) + + # only the second is released + assert evdev.slots[0][libevdev.EV_ABS.ABS_MT_TRACKING_ID] != -1 + assert evdev.slots[1][libevdev.EV_ABS.ABS_MT_TRACKING_ID] == -1 + assert evdev.slots[2][libevdev.EV_ABS.ABS_MT_TRACKING_ID] != -1 + + # wait for the timer to kick in + time.sleep(0.2) + + events = uhdev.next_sync_events() + self.debug_reports([], uhdev, events) + + # now all 3 fingers are released + assert libevdev.InputEvent(libevdev.EV_KEY.BTN_TOUCH, 0) in events + assert evdev.slots[0][libevdev.EV_ABS.ABS_MT_TRACKING_ID] == -1 + assert evdev.slots[1][libevdev.EV_ABS.ABS_MT_TRACKING_ID] == -1 + assert evdev.slots[2][libevdev.EV_ABS.ABS_MT_TRACKING_ID] == -1 + + class TestElanXPS9360(BaseTest.TestWin8Multitouch): def create_device(self): return Digitizer( @@ -2086,3 +2132,12 @@ class Testsynaptics_06cb_ce08(BaseTest.TestPTP): input_info=(BusType.I2C, 0x06CB, 0xCE08), rdesc="05 01 09 02 a1 01 85 02 09 01 a1 00 05 09 19 01 29 02 15 00 25 01 75 01 95 02 81 02 95 06 81 01 05 01 09 30 09 31 15 81 25 7f 75 08 95 02 81 06 c0 c0 05 01 09 02 a1 01 85 18 09 01 a1 00 05 09 19 01 29 03 46 00 00 15 00 25 01 75 01 95 03 81 02 95 05 81 01 05 01 09 30 09 31 15 81 25 7f 75 08 95 02 81 06 c0 c0 06 00 ff 09 02 a1 01 85 20 09 01 a1 00 09 03 15 00 26 ff 00 35 00 46 ff 00 75 08 95 05 81 02 c0 c0 05 0d 09 05 a1 01 85 03 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 f8 04 75 10 55 0e 65 11 09 30 35 00 46 24 04 95 01 81 02 46 30 02 26 a0 02 09 31 81 02 c0 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 f8 04 75 10 55 0e 65 11 09 30 35 00 46 24 04 95 01 81 02 46 30 02 26 a0 02 09 31 81 02 c0 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 f8 04 75 10 55 0e 65 11 09 30 35 00 46 24 04 95 01 81 02 46 30 02 26 a0 02 09 31 81 02 c0 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 f8 04 75 10 55 0e 65 11 09 30 35 00 46 24 04 95 01 81 02 46 30 02 26 a0 02 09 31 81 02 c0 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 f8 04 75 10 55 0e 65 11 09 30 35 00 46 24 04 95 01 81 02 46 30 02 26 a0 02 09 31 81 02 c0 05 0d 55 0c 66 01 10 47 ff ff 00 00 27 ff ff 00 00 75 10 95 01 09 56 81 02 09 54 25 7f 95 01 75 08 81 02 05 09 09 01 25 01 75 01 95 01 81 02 95 07 81 03 05 0d 85 08 09 55 09 59 75 04 95 02 25 0f b1 02 85 0d 09 60 75 01 95 01 15 00 25 01 b1 02 95 07 b1 03 85 07 06 00 ff 09 c5 15 00 26 ff 00 75 08 96 00 01 b1 02 c0 05 0d 09 0e a1 01 85 04 09 22 a1 02 09 52 15 00 25 0a 75 08 95 01 b1 02 c0 09 22 a1 00 85 06 09 57 09 58 75 01 95 02 25 01 b1 02 95 06 b1 03 c0 c0 06 00 ff 09 01 a1 01 85 09 09 02 15 00 26 ff 00 75 08 95 14 91 02 85 0a 09 03 15 00 26 ff 00 75 08 95 14 91 02 85 0b 09 04 15 00 26 ff 00 75 08 95 45 81 02 85 0c 09 05 15 00 26 ff 00 75 08 95 45 81 02 85 0f 09 06 15 00 26 ff 00 75 08 95 03 b1 02 85 0e 09 07 15 00 26 ff 00 75 08 95 01 b1 02 c0", ) + +class Testsynaptics_06cb_ce26(TestWin8TSConfidence): + def create_device(self): + return PTP( + "uhid test synaptics_06cb_ce26", + max_contacts=5, + input_info=(BusType.I2C, 0x06CB, 0xCE26), + rdesc="05 01 09 02 a1 01 85 02 09 01 a1 00 05 09 19 01 29 02 15 00 25 01 75 01 95 02 81 02 95 06 81 01 05 01 09 30 09 31 15 81 25 7f 75 08 95 02 81 06 c0 c0 05 0d 09 05 a1 01 85 03 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 45 05 75 10 55 0e 65 11 09 30 35 00 46 64 04 95 01 81 02 46 a2 02 26 29 03 09 31 81 02 c0 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 45 05 75 10 55 0e 65 11 09 30 35 00 46 64 04 95 01 81 02 46 a2 02 26 29 03 09 31 81 02 c0 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 45 05 75 10 55 0e 65 11 09 30 35 00 46 64 04 95 01 81 02 46 a2 02 26 29 03 09 31 81 02 c0 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 45 05 75 10 55 0e 65 11 09 30 35 00 46 64 04 95 01 81 02 46 a2 02 26 29 03 09 31 81 02 c0 05 0d 09 22 a1 02 15 00 25 01 09 47 09 42 95 02 75 01 81 02 95 01 75 03 25 05 09 51 81 02 75 01 95 03 81 03 05 01 15 00 26 45 05 75 10 55 0e 65 11 09 30 35 00 46 64 04 95 01 81 02 46 a2 02 26 29 03 09 31 81 02 c0 05 0d 55 0c 66 01 10 47 ff ff 00 00 27 ff ff 00 00 75 10 95 01 09 56 81 02 09 54 25 7f 95 01 75 08 81 02 05 09 09 01 25 01 75 01 95 01 81 02 95 07 81 03 05 0d 85 08 09 55 09 59 75 04 95 02 25 0f b1 02 85 0d 09 60 75 01 95 01 15 00 25 01 b1 02 95 07 b1 03 85 07 06 00 ff 09 c5 15 00 26 ff 00 75 08 96 00 01 b1 02 c0 05 0d 09 0e a1 01 85 04 09 22 a1 02 09 52 15 00 25 0a 75 08 95 01 b1 02 c0 09 22 a1 00 85 06 09 57 09 58 75 01 95 02 25 01 b1 02 95 06 b1 03 c0 c0 06 00 ff 09 01 a1 01 85 09 09 02 15 00 26 ff 00 75 08 95 14 91 02 85 0a 09 03 15 00 26 ff 00 75 08 95 14 91 02 85 0b 09 04 15 00 26 ff 00 75 08 95 3d 81 02 85 0c 09 05 15 00 26 ff 00 75 08 95 3d 81 02 85 0f 09 06 15 00 26 ff 00 75 08 95 03 b1 02 85 0e 09 07 15 00 26 ff 00 75 08 95 01 b1 02 c0", + ) From d451a0e88e9fa710df33f8dd5dc7ca63e22ef211 Mon Sep 17 00:00:00 2001 From: Stefan Metzmacher Date: Wed, 15 Oct 2025 17:05:04 +0200 Subject: [PATCH 214/330] smb: client: let smbd_destroy() wait for SMBDIRECT_SOCKET_DISCONNECTED We should wait for the rdma_cm to become SMBDIRECT_SOCKET_DISCONNECTED, it turns out that (at least running some xfstests e.g. cifs/001) often triggers the case where wait_event_interruptible() returns with -ERESTARTSYS instead of waiting for SMBDIRECT_SOCKET_DISCONNECTED to be reached. Or we are already in SMBDIRECT_SOCKET_DISCONNECTING and never wait for SMBDIRECT_SOCKET_DISCONNECTED. Fixes: 050b8c374019 ("smbd: Make upper layer decide when to destroy the transport") Fixes: e8b3bfe9bc65 ("cifs: smbd: Don't destroy transport on RDMA disconnect") Fixes: b0aa92a229ab ("smb: client: make sure smbd_disconnect_rdma_work() doesn't run after smbd_destroy() took over") Cc: Steve French Cc: Tom Talpey Cc: Long Li Cc: Namjae Jeon Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher Signed-off-by: Steve French --- fs/smb/client/smbdirect.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/fs/smb/client/smbdirect.c b/fs/smb/client/smbdirect.c index 77de85d7cdc3..49e2df3ad1f0 100644 --- a/fs/smb/client/smbdirect.c +++ b/fs/smb/client/smbdirect.c @@ -1575,12 +1575,12 @@ void smbd_destroy(struct TCP_Server_Info *server) disable_work_sync(&sc->disconnect_work); log_rdma_event(INFO, "destroying rdma session\n"); - if (sc->status < SMBDIRECT_SOCKET_DISCONNECTING) { + if (sc->status < SMBDIRECT_SOCKET_DISCONNECTING) smbd_disconnect_rdma_work(&sc->disconnect_work); + if (sc->status < SMBDIRECT_SOCKET_DISCONNECTED) { log_rdma_event(INFO, "wait for transport being disconnected\n"); - wait_event_interruptible( - sc->status_wait, - sc->status == SMBDIRECT_SOCKET_DISCONNECTED); + wait_event(sc->status_wait, sc->status == SMBDIRECT_SOCKET_DISCONNECTED); + log_rdma_event(INFO, "waited for transport being disconnected\n"); } /* From 2aab1f993c8cb753ccb3d5b848cd758e2e87d965 Mon Sep 17 00:00:00 2001 From: Ankan Biswas Date: Wed, 15 Oct 2025 20:50:57 +0530 Subject: [PATCH 215/330] drm/gpuvm: Fix kernel-doc warning for drm_gpuvm_map_req.map The kernel-doc for struct drm_gpuvm_map_req.map was added as '@op_map' instead of '@map', leading to this warning during htmldocs build: WARNING: include/drm/drm_gpuvm.h:1083 struct member 'map' not described in 'drm_gpuvm_map_req' Fixes: 000a45dce7ad ("drm/gpuvm: Pass map arguments through a struct") Reported-by: Stephen Rothwell Closes: https://lore.kernel.org/all/20250821133539.03aa298e@canb.auug.org.au/ Signed-off-by: Ankan Biswas Signed-off-by: Danilo Krummrich --- include/drm/drm_gpuvm.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/drm/drm_gpuvm.h b/include/drm/drm_gpuvm.h index 8890ded1d907..476990e761f8 100644 --- a/include/drm/drm_gpuvm.h +++ b/include/drm/drm_gpuvm.h @@ -1078,7 +1078,7 @@ struct drm_gpuva_ops { */ struct drm_gpuvm_map_req { /** - * @op_map: struct drm_gpuva_op_map + * @map: struct drm_gpuva_op_map */ struct drm_gpuva_op_map map; }; From fcd298fdc2a32f1d90cdf9a452c5c5fdc6e8d137 Mon Sep 17 00:00:00 2001 From: Frank Li Date: Fri, 10 Oct 2025 12:03:20 -0400 Subject: [PATCH 216/330] ASoC: dt-bindings: Add compatible string fsl,imx-audio-tlv320 Add compatible string fsl,imx-audio-tlv320 to fix below CHECK_DTBS warning: arch/arm/boot/dts/nxp/imx/imx6dl-gw5903.dtb: /sound: failed to match any schema with compatible: ['fsl,imx-audio-tlv320'] Signed-off-by: Frank Li Acked-by: Rob Herring (Arm) Link: https://patch.msgid.link/20251010160321.2130093-1-Frank.Li@nxp.com Signed-off-by: Mark Brown --- Documentation/devicetree/bindings/sound/fsl-asoc-card.yaml | 1 + 1 file changed, 1 insertion(+) diff --git a/Documentation/devicetree/bindings/sound/fsl-asoc-card.yaml b/Documentation/devicetree/bindings/sound/fsl-asoc-card.yaml index 92aa47ec72c7..88eb20bb008f 100644 --- a/Documentation/devicetree/bindings/sound/fsl-asoc-card.yaml +++ b/Documentation/devicetree/bindings/sound/fsl-asoc-card.yaml @@ -79,6 +79,7 @@ properties: - fsl,imx-audio-nau8822 - fsl,imx-audio-sgtl5000 - fsl,imx-audio-si476x + - fsl,imx-audio-tlv320 - fsl,imx-audio-tlv320aic31xx - fsl,imx-audio-tlv320aic32x4 - fsl,imx-audio-wm8524 From 7a37291ed40a33a5f6c3d370fdde5ee0d8f7d0e4 Mon Sep 17 00:00:00 2001 From: Sharique Mohammad Date: Wed, 15 Oct 2025 15:42:15 +0200 Subject: [PATCH 217/330] ASoC: max98090/91: fixed max98091 ALSA widget powering up/down The widgets DMIC3_ENA and DMIC4_ENA must be defined in the DAPM suppy widget, just like DMICL_ENA and DMICR_ENA. Whenever they are turned on or off, the required startup or shutdown sequences must be taken care by the max98090_shdn_event. Signed-off-by: Sharique Mohammad Link: https://patch.msgid.link/20251015134215.750001-1-sharq0406@gmail.com Signed-off-by: Mark Brown --- sound/soc/codecs/max98090.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/sound/soc/codecs/max98090.c b/sound/soc/codecs/max98090.c index 22177c1ce160..cb1508fc99f8 100644 --- a/sound/soc/codecs/max98090.c +++ b/sound/soc/codecs/max98090.c @@ -1234,9 +1234,11 @@ static const struct snd_soc_dapm_widget max98091_dapm_widgets[] = { SND_SOC_DAPM_INPUT("DMIC4"), SND_SOC_DAPM_SUPPLY("DMIC3_ENA", M98090_REG_DIGITAL_MIC_ENABLE, - M98090_DIGMIC3_SHIFT, 0, NULL, 0), + M98090_DIGMIC3_SHIFT, 0, max98090_shdn_event, + SND_SOC_DAPM_POST_PMU), SND_SOC_DAPM_SUPPLY("DMIC4_ENA", M98090_REG_DIGITAL_MIC_ENABLE, - M98090_DIGMIC4_SHIFT, 0, NULL, 0), + M98090_DIGMIC4_SHIFT, 0, max98090_shdn_event, + SND_SOC_DAPM_POST_PMU), }; static const struct snd_soc_dapm_route max98090_dapm_routes[] = { From 5726b68473f7153a7f6294185e5998b7e2a230a2 Mon Sep 17 00:00:00 2001 From: Li Qiang Date: Wed, 15 Oct 2025 15:55:30 +0800 Subject: [PATCH 218/330] ASoC: amd/sdw_utils: avoid NULL deref when devm_kasprintf() fails devm_kasprintf() may return NULL on memory allocation failure, but the debug message prints cpus->dai_name before checking it. Move the dev_dbg() call after the NULL check to prevent potential NULL pointer dereference. Fixes: cb8ea62e64020 ("ASoC: amd/sdw_utils: add sof based soundwire generic machine driver") Signed-off-by: Li Qiang Link: https://patch.msgid.link/20251015075530.146851-1-liqiang01@kylinos.cn Signed-off-by: Mark Brown --- sound/soc/amd/acp/acp-sdw-sof-mach.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/amd/acp/acp-sdw-sof-mach.c b/sound/soc/amd/acp/acp-sdw-sof-mach.c index 91d72d4bb9a2..d055582a3bf1 100644 --- a/sound/soc/amd/acp/acp-sdw-sof-mach.c +++ b/sound/soc/amd/acp/acp-sdw-sof-mach.c @@ -176,9 +176,9 @@ static int create_sdw_dailink(struct snd_soc_card *card, cpus->dai_name = devm_kasprintf(dev, GFP_KERNEL, "SDW%d Pin%d", link_num, cpu_pin_id); - dev_dbg(dev, "cpu->dai_name:%s\n", cpus->dai_name); if (!cpus->dai_name) return -ENOMEM; + dev_dbg(dev, "cpu->dai_name:%s\n", cpus->dai_name); codec_maps[j].cpu = 0; codec_maps[j].codec = j; From e6416c2dfe23c9a6fec881fda22ebb9ae486cfc5 Mon Sep 17 00:00:00 2001 From: Rong Zhang Date: Sat, 11 Oct 2025 00:59:58 +0800 Subject: [PATCH 219/330] x86/CPU/AMD: Prevent reset reasons from being retained across reboot The S5_RESET_STATUS register is parsed on boot and printed to kmsg. However, this could sometimes be misleading and lead to users wasting a lot of time on meaningless debugging for two reasons: * Some bits are never cleared by hardware. It's the software's responsibility to clear them as per the Processor Programming Reference (see [1]). * Some rare hardware-initiated platform resets do not update the register at all. In both cases, a previous reboot could leave its trace in the register, resulting in users seeing unrelated reboot reasons while debugging random reboots afterward. Write the read value back to the register in order to clear all reason bits since they are write-1-to-clear while the others must be preserved. [1]: https://bugzilla.kernel.org/show_bug.cgi?id=206537#attach_303991 [ bp: Massage commit message. ] Fixes: ab8131028710 ("x86/CPU/AMD: Print the reason for the last reset") Signed-off-by: Rong Zhang Signed-off-by: Borislav Petkov (AMD) Reviewed-by: Mario Limonciello (AMD) Reviewed-by: Yazen Ghannam Cc: Link: https://lore.kernel.org/all/20250913144245.23237-1-i@rong.moe/ --- arch/x86/kernel/cpu/amd.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c index 5398db4dedb4..ccaa51ce63f6 100644 --- a/arch/x86/kernel/cpu/amd.c +++ b/arch/x86/kernel/cpu/amd.c @@ -1355,11 +1355,23 @@ static __init int print_s5_reset_status_mmio(void) return 0; value = ioread32(addr); - iounmap(addr); /* Value with "all bits set" is an error response and should be ignored. */ - if (value == U32_MAX) + if (value == U32_MAX) { + iounmap(addr); return 0; + } + + /* + * Clear all reason bits so they won't be retained if the next reset + * does not update the register. Besides, some bits are never cleared by + * hardware so it's software's responsibility to clear them. + * + * Writing the value back effectively clears all reason bits as they are + * write-1-to-clear. + */ + iowrite32(value, addr); + iounmap(addr); for (i = 0; i < ARRAY_SIZE(s5_reset_reason_txt); i++) { if (!(value & BIT(i))) From 18d6b1743eafeb3fb1e0ea5a2b7fd0a773d525a8 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Wed, 15 Oct 2025 13:38:53 -0600 Subject: [PATCH 220/330] io_uring/rw: check for NULL io_br_sel when putting a buffer Both the read and write side use kiocb_done() to finish a request, and kiocb_done() will call io_put_kbuf() in case a provided buffer was used for the request. Provided buffers are not supported for writes, hence NULL is being passed in. This normally works fine, as io_put_kbuf() won't actually use the value unless REQ_F_BUFFER_RING or REQ_F_BUFFER_SELECTED is set in the request flags. But depending on compiler (or whether or not CONFIG_CC_OPTIMIZE_FOR_SIZE is set), that may be done even though the value is never used. This will then cause a NULL pointer dereference. Make it a bit more obvious and check for a NULL io_br_sel, and don't even bother calling io_put_kbuf() for that case. Fixes: 5fda51255439 ("io_uring/kbuf: switch to storing struct io_buffer_list locally") Reported-by: David Howells Tested-by: David Howells Signed-off-by: Jens Axboe --- io_uring/rw.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/io_uring/rw.c b/io_uring/rw.c index a0f9d2021e3f..5b2241a5813c 100644 --- a/io_uring/rw.c +++ b/io_uring/rw.c @@ -655,13 +655,17 @@ static int kiocb_done(struct io_kiocb *req, ssize_t ret, if (ret >= 0 && req->flags & REQ_F_CUR_POS) req->file->f_pos = rw->kiocb.ki_pos; if (ret >= 0 && !(req->ctx->flags & IORING_SETUP_IOPOLL)) { + u32 cflags = 0; + __io_complete_rw_common(req, ret); /* * Safe to call io_end from here as we're inline * from the submission path. */ io_req_io_end(req); - io_req_set_res(req, final_ret, io_put_kbuf(req, ret, sel->buf_list)); + if (sel) + cflags = io_put_kbuf(req, ret, sel->buf_list); + io_req_set_res(req, final_ret, cflags); io_req_rw_cleanup(req, issue_flags); return IOU_COMPLETE; } else { From c97513cddcfc235f2522617980838e500af21d01 Mon Sep 17 00:00:00 2001 From: Lance Yang Date: Tue, 9 Sep 2025 22:52:43 +0800 Subject: [PATCH 221/330] hung_task: fix warnings caused by unaligned lock pointers The blocker tracking mechanism assumes that lock pointers are at least 4-byte aligned to use their lower bits for type encoding. However, as reported by Eero Tamminen, some architectures like m68k only guarantee 2-byte alignment of 32-bit values. This breaks the assumption and causes two related WARN_ON_ONCE checks to trigger. To fix this, the runtime checks are adjusted to silently ignore any lock that is not 4-byte aligned, effectively disabling the feature in such cases and avoiding the related warnings. Thanks to Geert Uytterhoeven for bisecting! Link: https://lkml.kernel.org/r/20250909145243.17119-1-lance.yang@linux.dev Fixes: e711faaafbe5 ("hung_task: replace blocker_mutex with encoded blocker") Signed-off-by: Lance Yang Reported-by: Eero Tamminen Closes: https://lore.kernel.org/lkml/CAMuHMdW7Ab13DdGs2acMQcix5ObJK0O2dG_Fxzr8_g58Rc1_0g@mail.gmail.com Reviewed-by: Masami Hiramatsu (Google) Cc: John Paul Adrian Glaubitz Cc: Anna Schumaker Cc: Boqun Feng Cc: Finn Thain Cc: Geert Uytterhoeven Cc: Ingo Molnar Cc: Joel Granados Cc: John Stultz Cc: Kent Overstreet Cc: Lance Yang Cc: Mingzhe Yang Cc: Peter Zijlstra Cc: Sergey Senozhatsky Cc: Steven Rostedt Cc: Tomasz Figa Cc: Waiman Long Cc: Will Deacon Cc: Yongliang Gao Cc: Signed-off-by: Andrew Morton --- include/linux/hung_task.h | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/include/linux/hung_task.h b/include/linux/hung_task.h index 34e615c76ca5..c4403eeb7144 100644 --- a/include/linux/hung_task.h +++ b/include/linux/hung_task.h @@ -20,6 +20,10 @@ * always zero. So we can use these bits to encode the specific blocking * type. * + * Note that on architectures where this is not guaranteed, or for any + * unaligned lock, this tracking mechanism is silently skipped for that + * lock. + * * Type encoding: * 00 - Blocked on mutex (BLOCKER_TYPE_MUTEX) * 01 - Blocked on semaphore (BLOCKER_TYPE_SEM) @@ -45,7 +49,7 @@ static inline void hung_task_set_blocker(void *lock, unsigned long type) * If the lock pointer matches the BLOCKER_TYPE_MASK, return * without writing anything. */ - if (WARN_ON_ONCE(lock_ptr & BLOCKER_TYPE_MASK)) + if (lock_ptr & BLOCKER_TYPE_MASK) return; WRITE_ONCE(current->blocker, lock_ptr | type); @@ -53,8 +57,6 @@ static inline void hung_task_set_blocker(void *lock, unsigned long type) static inline void hung_task_clear_blocker(void) { - WARN_ON_ONCE(!READ_ONCE(current->blocker)); - WRITE_ONCE(current->blocker, 0UL); } From f0c5118ebb0eb7e4fd6f0d2ace3315ca141b317f Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Fri, 3 Oct 2025 13:14:54 -0700 Subject: [PATCH 222/330] mm/damon/sysfs: catch commit test ctx alloc failure Patch series "mm/damon/sysfs: fix commit test damon_ctx [de]allocation". DAMON sysfs interface dynamically allocates and uses a damon_ctx object for testing if given inputs for online DAMON parameters update is valid. The object is being used without an allocation failure check, and leaked when the test succeeds. Fix the two bugs. This patch (of 2): The damon_ctx for testing online DAMON parameters commit inputs is used without its allocation failure check. This could result in an invalid memory access. Fix it by directly returning an error when the allocation failed. Link: https://lkml.kernel.org/r/20251003201455.41448-1-sj@kernel.org Link: https://lkml.kernel.org/r/20251003201455.41448-2-sj@kernel.org Fixes: 4c9ea539ad59 ("mm/damon/sysfs: validate user inputs from damon_sysfs_commit_input()") Signed-off-by: SeongJae Park Cc: [6.15+] Signed-off-by: Andrew Morton --- mm/damon/sysfs.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/mm/damon/sysfs.c b/mm/damon/sysfs.c index 2fc722f998f8..703f55a91b3c 100644 --- a/mm/damon/sysfs.c +++ b/mm/damon/sysfs.c @@ -1473,6 +1473,8 @@ static int damon_sysfs_commit_input(void *data) if (IS_ERR(param_ctx)) return PTR_ERR(param_ctx); test_ctx = damon_new_ctx(); + if (!test_ctx) + return -ENOMEM; err = damon_commit_ctx(test_ctx, param_ctx); if (err) { damon_destroy_ctx(test_ctx); From 139e7a572af0b45f558b5e502121a768dc328ba8 Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Fri, 3 Oct 2025 13:14:55 -0700 Subject: [PATCH 223/330] mm/damon/sysfs: dealloc commit test ctx always The damon_ctx for testing online DAMON parameters commit inputs is deallocated only when the test fails. This means memory is leaked for every successful online DAMON parameters commit. Fix the leak by always deallocating it. Link: https://lkml.kernel.org/r/20251003201455.41448-3-sj@kernel.org Fixes: 4c9ea539ad59 ("mm/damon/sysfs: validate user inputs from damon_sysfs_commit_input()") Signed-off-by: SeongJae Park Cc: [6.15+] Signed-off-by: Andrew Morton --- mm/damon/sysfs.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/mm/damon/sysfs.c b/mm/damon/sysfs.c index 703f55a91b3c..cd6815ecc04e 100644 --- a/mm/damon/sysfs.c +++ b/mm/damon/sysfs.c @@ -1476,12 +1476,11 @@ static int damon_sysfs_commit_input(void *data) if (!test_ctx) return -ENOMEM; err = damon_commit_ctx(test_ctx, param_ctx); - if (err) { - damon_destroy_ctx(test_ctx); + if (err) goto out; - } err = damon_commit_ctx(kdamond->damon_ctx, param_ctx); out: + damon_destroy_ctx(test_ctx); damon_destroy_ctx(param_ctx); return err; } From 03521c892bb8d0712c23e158ae9bdf8705897df8 Mon Sep 17 00:00:00 2001 From: Marek Szyprowski Date: Thu, 9 Oct 2025 16:15:08 +0200 Subject: [PATCH 224/330] dma-debug: don't report false positives with DMA_BOUNCE_UNALIGNED_KMALLOC Commit 370645f41e6e ("dma-mapping: force bouncing if the kmalloc() size is not cache-line-aligned") introduced DMA_BOUNCE_UNALIGNED_KMALLOC feature and permitted architecture specific code configure kmalloc slabs with sizes smaller than the value of dma_get_cache_alignment(). When that feature is enabled, the physical address of some small kmalloc()-ed buffers might be not aligned to the CPU cachelines, thus not really suitable for typical DMA. To properly handle that case a SWIOTLB buffer bouncing is used, so no CPU cache corruption occurs. When that happens, there is no point reporting a false-positive DMA-API warning that the buffer is not properly aligned, as this is not a client driver fault. [m.szyprowski@samsung.com: replace is_swiotlb_allocated() with is_swiotlb_active(), per Catalin] Link: https://lkml.kernel.org/r/20251010173009.3916215-1-m.szyprowski@samsung.com Link: https://lkml.kernel.org/r/20251009141508.2342138-1-m.szyprowski@samsung.com Fixes: 370645f41e6e ("dma-mapping: force bouncing if the kmalloc() size is not cache-line-aligned") Signed-off-by: Marek Szyprowski Reviewed-by: Catalin Marinas Cc: Christoph Hellwig Cc: Inki Dae Cc: Robin Murohy Cc: "Isaac J. Manjarres" Cc: Signed-off-by: Andrew Morton --- kernel/dma/debug.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/kernel/dma/debug.c b/kernel/dma/debug.c index 1e5c64cb6a42..138ede653de4 100644 --- a/kernel/dma/debug.c +++ b/kernel/dma/debug.c @@ -23,6 +23,7 @@ #include #include #include +#include #include #include "debug.h" @@ -594,7 +595,9 @@ static void add_dma_entry(struct dma_debug_entry *entry, unsigned long attrs) if (rc == -ENOMEM) { pr_err_once("cacheline tracking ENOMEM, dma-debug disabled\n"); global_disable = true; - } else if (rc == -EEXIST && !(attrs & DMA_ATTR_SKIP_CPU_SYNC)) { + } else if (rc == -EEXIST && !(attrs & DMA_ATTR_SKIP_CPU_SYNC) && + !(IS_ENABLED(CONFIG_DMA_BOUNCE_UNALIGNED_KMALLOC) && + is_swiotlb_active(entry->dev))) { err_printk(entry->dev, entry, "cacheline tracking EEXIST, overlapping mappings aren't supported\n"); } From c83aab85e18103a6dc066b4939e2c92a02bb1b05 Mon Sep 17 00:00:00 2001 From: Alexei Starovoitov Date: Thu, 9 Oct 2025 17:15:13 -0700 Subject: [PATCH 225/330] mm: don't spin in add_stack_record when gfp flags don't allow syzbot was able to find the following path: add_stack_record_to_list mm/page_owner.c:182 [inline] inc_stack_record_count mm/page_owner.c:214 [inline] __set_page_owner+0x2c3/0x4a0 mm/page_owner.c:333 set_page_owner include/linux/page_owner.h:32 [inline] post_alloc_hook+0x240/0x2a0 mm/page_alloc.c:1851 prep_new_page mm/page_alloc.c:1859 [inline] get_page_from_freelist+0x21e4/0x22c0 mm/page_alloc.c:3858 alloc_pages_nolock_noprof+0x94/0x120 mm/page_alloc.c:7554 Don't spin in add_stack_record_to_list() when it is called from *_nolock() context. Link: https://lkml.kernel.org/r/CAADnVQK_8bNYEA7TJYgwTYR57=TTFagsvRxp62pFzS_z129eTg@mail.gmail.com Fixes: 97769a53f117 ("mm, bpf: Introduce try_alloc_pages() for opportunistic page allocation") Signed-off-by: Alexei Starovoitov Reported-by: syzbot+8259e1d0e3ae8ed0c490@syzkaller.appspotmail.com Reported-by: syzbot+665739f456b28f32b23d@syzkaller.appspotmail.com Acked-by: Vlastimil Babka Reviewed-by: Oscar Salvador Cc: Brendan Jackman Cc: Johannes Weiner Cc: Michal Hocko Cc: Suren Baghdasaryan Cc: Zi Yan Cc: Signed-off-by: Andrew Morton --- mm/page_owner.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/mm/page_owner.c b/mm/page_owner.c index c3ca21132c2c..589ec37c94aa 100644 --- a/mm/page_owner.c +++ b/mm/page_owner.c @@ -168,6 +168,9 @@ static void add_stack_record_to_list(struct stack_record *stack_record, unsigned long flags; struct stack *stack; + if (!gfpflags_allow_spinning(gfp_mask)) + return; + set_current_in_page_owner(); stack = kmalloc(sizeof(*stack), gfp_nested_mask(gfp_mask)); if (!stack) { From 78a63493f8e352296dbc7cb7b3f4973105e8679e Mon Sep 17 00:00:00 2001 From: Deepanshu Kartikey Date: Thu, 9 Oct 2025 21:19:03 +0530 Subject: [PATCH 226/330] ocfs2: clear extent cache after moving/defragmenting extents The extent map cache can become stale when extents are moved or defragmented, causing subsequent operations to see outdated extent flags. This triggers a BUG_ON in ocfs2_refcount_cal_cow_clusters(). The problem occurs when: 1. copy_file_range() creates a reflinked extent with OCFS2_EXT_REFCOUNTED 2. ioctl(FITRIM) triggers ocfs2_move_extents() 3. __ocfs2_move_extents_range() reads and caches the extent (flags=0x2) 4. ocfs2_move_extent()/ocfs2_defrag_extent() calls __ocfs2_move_extent() which clears OCFS2_EXT_REFCOUNTED flag on disk (flags=0x0) 5. The extent map cache is not invalidated after the move 6. Later write() operations read stale cached flags (0x2) but disk has updated flags (0x0), causing a mismatch 7. BUG_ON(!(rec->e_flags & OCFS2_EXT_REFCOUNTED)) triggers Fix by clearing the extent map cache after each extent move/defrag operation in __ocfs2_move_extents_range(). This ensures subsequent operations read fresh extent data from disk. Link: https://lore.kernel.org/all/20251009142917.517229-1-kartikey406@gmail.com/T/ Link: https://lkml.kernel.org/r/20251009154903.522339-1-kartikey406@gmail.com Fixes: 53069d4e7695 ("Ocfs2/move_extents: move/defrag extents within a certain range.") Signed-off-by: Deepanshu Kartikey Reported-by: syzbot+6fdd8fa3380730a4b22c@syzkaller.appspotmail.com Tested-by: syzbot+6fdd8fa3380730a4b22c@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?id=2959889e1f6e216585ce522f7e8bc002b46ad9e7 Reviewed-by: Mark Fasheh Reviewed-by: Joseph Qi Cc: Joel Becker Cc: Junxiao Bi Cc: Changwei Ge Cc: Jun Piao Cc: Signed-off-by: Andrew Morton --- fs/ocfs2/move_extents.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/fs/ocfs2/move_extents.c b/fs/ocfs2/move_extents.c index 86f2631e6360..10923bf7c8b8 100644 --- a/fs/ocfs2/move_extents.c +++ b/fs/ocfs2/move_extents.c @@ -867,6 +867,11 @@ static int __ocfs2_move_extents_range(struct buffer_head *di_bh, mlog_errno(ret); goto out; } + /* + * Invalidate extent cache after moving/defragging to prevent + * stale cached data with outdated extent flags. + */ + ocfs2_extent_map_trunc(inode, cpos); context->clusters_moved += alloc_size; next: From 841a8bfcbad94bb1ba60f59ce34f75259074ae0d Mon Sep 17 00:00:00 2001 From: Qiuxu Zhuo Date: Sat, 11 Oct 2025 15:55:19 +0800 Subject: [PATCH 227/330] mm: prevent poison consumption when splitting THP When performing memory error injection on a THP (Transparent Huge Page) mapped to userspace on an x86 server, the kernel panics with the following trace. The expected behavior is to terminate the affected process instead of panicking the kernel, as the x86 Machine Check code can recover from an in-userspace #MC. mce: [Hardware Error]: CPU 0: Machine Check Exception: f Bank 3: bd80000000070134 mce: [Hardware Error]: RIP 10: {memchr_inv+0x4c/0xf0} mce: [Hardware Error]: TSC afff7bbff88a ADDR 1d301b000 MISC 80 PPIN 1e741e77539027db mce: [Hardware Error]: PROCESSOR 0:d06d0 TIME 1758093249 SOCKET 0 APIC 0 microcode 80000320 mce: [Hardware Error]: Run the above through 'mcelog --ascii' mce: [Hardware Error]: Machine check: Data load in unrecoverable area of kernel Kernel panic - not syncing: Fatal local machine check The root cause of this panic is that handling a memory failure triggered by an in-userspace #MC necessitates splitting the THP. The splitting process employs a mechanism, implemented in try_to_map_unused_to_zeropage(), which reads the pages in the THP to identify zero-filled pages. However, reading the pages in the THP results in a second in-kernel #MC, occurring before the initial memory_failure() completes, ultimately leading to a kernel panic. See the kernel panic call trace on the two #MCs. First Machine Check occurs // [1] memory_failure() // [2] try_to_split_thp_page() split_huge_page() split_huge_page_to_list_to_order() __folio_split() // [3] remap_page() remove_migration_ptes() remove_migration_pte() try_to_map_unused_to_zeropage() // [4] memchr_inv() // [5] Second Machine Check occurs // [6] Kernel panic [1] Triggered by accessing a hardware-poisoned THP in userspace, which is typically recoverable by terminating the affected process. [2] Call folio_set_has_hwpoisoned() before try_to_split_thp_page(). [3] Pass the RMP_USE_SHARED_ZEROPAGE remap flag to remap_page(). [4] Try to map the unused THP to zeropage. [5] Re-access pages in the hw-poisoned THP in the kernel. [6] Triggered in-kernel, leading to a panic kernel. In Step[2], memory_failure() sets the poisoned flag on the page in the THP by TestSetPageHWPoison() before calling try_to_split_thp_page(). As suggested by David Hildenbrand, fix this panic by not accessing to the poisoned page in the THP during zeropage identification, while continuing to scan unaffected pages in the THP for possible zeropage mapping. This prevents a second in-kernel #MC that would cause kernel panic in Step[4]. Thanks to Andrew Zaborowski for his initial work on fixing this issue. Link: https://lkml.kernel.org/r/20251015064926.1887643-1-qiuxu.zhuo@intel.com Link: https://lkml.kernel.org/r/20251011075520.320862-1-qiuxu.zhuo@intel.com Fixes: b1f202060afe ("mm: remap unused subpages to shared zeropage when splitting isolated thp") Signed-off-by: Qiuxu Zhuo Reported-by: Farrah Chen Suggested-by: David Hildenbrand Acked-by: David Hildenbrand Tested-by: Farrah Chen Tested-by: Qiuxu Zhuo Acked-by: Lance Yang Reviewed-by: Wei Yang Acked-by: Zi Yan Reviewed-by: Miaohe Lin Cc: Barry Song Cc: Dev Jain Cc: Jiaqi Yan Cc: Liam Howlett Cc: Lorenzo Stoakes Cc: "Luck, Tony" Cc: Mariano Pache Cc: Miaohe Lin Cc: Naoya Horiguchi Cc: Ryan Roberts Cc: Signed-off-by: Andrew Morton --- mm/huge_memory.c | 3 +++ mm/migrate.c | 3 ++- 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 1b81680b4225..1d1b74950332 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -4109,6 +4109,9 @@ static bool thp_underused(struct folio *folio) if (khugepaged_max_ptes_none == HPAGE_PMD_NR - 1) return false; + if (folio_contain_hwpoisoned_page(folio)) + return false; + for (i = 0; i < folio_nr_pages(folio); i++) { if (pages_identical(folio_page(folio, i), ZERO_PAGE(0))) { if (++num_zero_pages > khugepaged_max_ptes_none) diff --git a/mm/migrate.c b/mm/migrate.c index e3065c9edb55..c0e9f15be2a2 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -301,8 +301,9 @@ static bool try_to_map_unused_to_zeropage(struct page_vma_mapped_walk *pvmw, struct page *page = folio_page(folio, idx); pte_t newpte; - if (PageCompound(page)) + if (PageCompound(page) || PageHWPoison(page)) return false; + VM_BUG_ON_PAGE(!PageAnon(page), page); VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE(pte_present(old_pte), page); From c2b77f42205ef485a647f62082c442c1cd69d3fc Mon Sep 17 00:00:00 2001 From: Shuhao Fu Date: Thu, 16 Oct 2025 02:52:55 +0000 Subject: [PATCH 228/330] smb: client: Fix refcount leak for cifs_sb_tlink Fix three refcount inconsistency issues related to `cifs_sb_tlink`. Comments for `cifs_sb_tlink` state that `cifs_put_tlink()` needs to be called after successful calls to `cifs_sb_tlink()`. Three calls fail to update refcount accordingly, leading to possible resource leaks. Fixes: 8ceb98437946 ("CIFS: Move rename to ops struct") Fixes: 2f1afe25997f ("cifs: Use smb 2 - 3 and cifsacl mount options getacl functions") Fixes: 366ed846df60 ("cifs: Use smb 2 - 3 and cifsacl mount options setacl function") Cc: stable@vger.kernel.org Signed-off-by: Shuhao Fu Signed-off-by: Steve French --- fs/smb/client/inode.c | 6 ++++-- fs/smb/client/smb2ops.c | 8 ++++---- 2 files changed, 8 insertions(+), 6 deletions(-) diff --git a/fs/smb/client/inode.c b/fs/smb/client/inode.c index 239dd84a336f..098a79b7a959 100644 --- a/fs/smb/client/inode.c +++ b/fs/smb/client/inode.c @@ -2431,8 +2431,10 @@ cifs_do_rename(const unsigned int xid, struct dentry *from_dentry, tcon = tlink_tcon(tlink); server = tcon->ses->server; - if (!server->ops->rename) - return -ENOSYS; + if (!server->ops->rename) { + rc = -ENOSYS; + goto do_rename_exit; + } /* try path-based rename first */ rc = server->ops->rename(xid, tcon, from_dentry, diff --git a/fs/smb/client/smb2ops.c b/fs/smb/client/smb2ops.c index 7c392cf5940b..95cd484cfbba 100644 --- a/fs/smb/client/smb2ops.c +++ b/fs/smb/client/smb2ops.c @@ -3212,8 +3212,7 @@ get_smb2_acl_by_path(struct cifs_sb_info *cifs_sb, utf16_path = cifs_convert_path_to_utf16(path, cifs_sb); if (!utf16_path) { rc = -ENOMEM; - free_xid(xid); - return ERR_PTR(rc); + goto put_tlink; } oparms = (struct cifs_open_parms) { @@ -3245,6 +3244,7 @@ get_smb2_acl_by_path(struct cifs_sb_info *cifs_sb, SMB2_close(xid, tcon, fid.persistent_fid, fid.volatile_fid); } +put_tlink: cifs_put_tlink(tlink); free_xid(xid); @@ -3285,8 +3285,7 @@ set_smb2_acl(struct smb_ntsd *pnntsd, __u32 acllen, utf16_path = cifs_convert_path_to_utf16(path, cifs_sb); if (!utf16_path) { rc = -ENOMEM; - free_xid(xid); - return rc; + goto put_tlink; } oparms = (struct cifs_open_parms) { @@ -3307,6 +3306,7 @@ set_smb2_acl(struct smb_ntsd *pnntsd, __u32 acllen, SMB2_close(xid, tcon, fid.persistent_fid, fid.volatile_fid); } +put_tlink: cifs_put_tlink(tlink); free_xid(xid); return rc; From 6447b0e355562a1ff748c4a2ffb89aae7e84d2c9 Mon Sep 17 00:00:00 2001 From: Eugene Korenevsky Date: Mon, 13 Oct 2025 21:39:30 +0300 Subject: [PATCH 229/330] cifs: parse_dfs_referrals: prevent oob on malformed input Malicious SMB server can send invalid reply to FSCTL_DFS_GET_REFERRALS - reply smaller than sizeof(struct get_dfs_referral_rsp) - reply with number of referrals smaller than NumberOfReferrals in the header Processing of such replies will cause oob. Return -EINVAL error on such replies to prevent oob-s. Signed-off-by: Eugene Korenevsky Cc: stable@vger.kernel.org Suggested-by: Nathan Chancellor Acked-by: Paulo Alcantara (Red Hat) Signed-off-by: Steve French --- fs/smb/client/misc.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) diff --git a/fs/smb/client/misc.c b/fs/smb/client/misc.c index dda6dece802a..e10123d8cd7d 100644 --- a/fs/smb/client/misc.c +++ b/fs/smb/client/misc.c @@ -916,6 +916,14 @@ parse_dfs_referrals(struct get_dfs_referral_rsp *rsp, u32 rsp_size, char *data_end; struct dfs_referral_level_3 *ref; + if (rsp_size < sizeof(*rsp)) { + cifs_dbg(VFS | ONCE, + "%s: header is malformed (size is %u, must be %zu)\n", + __func__, rsp_size, sizeof(*rsp)); + rc = -EINVAL; + goto parse_DFS_referrals_exit; + } + *num_of_nodes = le16_to_cpu(rsp->NumberOfReferrals); if (*num_of_nodes < 1) { @@ -925,6 +933,15 @@ parse_dfs_referrals(struct get_dfs_referral_rsp *rsp, u32 rsp_size, goto parse_DFS_referrals_exit; } + if (sizeof(*rsp) + *num_of_nodes * sizeof(REFERRAL3) > rsp_size) { + cifs_dbg(VFS | ONCE, + "%s: malformed buffer (size is %u, must be at least %zu)\n", + __func__, rsp_size, + sizeof(*rsp) + *num_of_nodes * sizeof(REFERRAL3)); + rc = -EINVAL; + goto parse_DFS_referrals_exit; + } + ref = (struct dfs_referral_level_3 *) &(rsp->referrals); if (ref->VersionNumber != cpu_to_le16(3)) { cifs_dbg(VFS, "Referrals of V%d version are not supported, should be V3\n", From af5fea51411224cae61d54064a55fe22020bd2b7 Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Sat, 11 Oct 2025 18:57:31 -0700 Subject: [PATCH 230/330] smb: client: Use SHA-512 library for SMB3.1.1 preauth hash Convert smb311_update_preauth_hash() to use the SHA-512 library instead of a "sha512" crypto_shash. This is simpler and faster. With the library there's no need to allocate memory, no need to handle errors, and the SHA-512 code is accessed directly without inefficient indirect calls and other unnecessary API overhead. Remove the call to smb311_crypto_shash_allocate() from smb311_update_preauth_hash(), since it appears to have been needed only to allocate the "sha512" crypto_shash. (It also had the side effect of allocating the "cmac(aes)" crypto_shash, but that's also done in generate_key() which is where the AES-CMAC key is initialized.) For now the "sha512" crypto_shash is still being allocated elsewhere. It will be removed in a later commit. Reviewed-by: Stefan Metzmacher Acked-by: Ard Biesheuvel Signed-off-by: Eric Biggers Signed-off-by: Steve French --- fs/smb/client/Kconfig | 1 + fs/smb/client/smb2misc.c | 53 +++++++++------------------------------ fs/smb/client/smb2proto.h | 6 ++--- 3 files changed, 16 insertions(+), 44 deletions(-) diff --git a/fs/smb/client/Kconfig b/fs/smb/client/Kconfig index a4c02199fef4..4ac79ff5649b 100644 --- a/fs/smb/client/Kconfig +++ b/fs/smb/client/Kconfig @@ -16,6 +16,7 @@ config CIFS select CRYPTO_ECB select CRYPTO_AES select CRYPTO_LIB_ARC4 + select CRYPTO_LIB_SHA512 select KEYS select DNS_RESOLVER select ASN1 diff --git a/fs/smb/client/smb2misc.c b/fs/smb/client/smb2misc.c index 89d933b4a8bc..96bfe4c63ccf 100644 --- a/fs/smb/client/smb2misc.c +++ b/fs/smb/client/smb2misc.c @@ -7,6 +7,7 @@ * Pavel Shilovsky (pshilovsky@samba.org) 2012 * */ +#include #include #include "cifsglob.h" #include "cifsproto.h" @@ -888,13 +889,13 @@ smb2_handle_cancelled_mid(struct mid_q_entry *mid, struct TCP_Server_Info *serve * @iov: array containing the SMB request we will send to the server * @nvec: number of array entries for the iov */ -int +void smb311_update_preauth_hash(struct cifs_ses *ses, struct TCP_Server_Info *server, struct kvec *iov, int nvec) { - int i, rc; + int i; struct smb2_hdr *hdr; - struct shash_desc *sha512 = NULL; + struct sha512_ctx sha_ctx; hdr = (struct smb2_hdr *)iov[0].iov_base; /* neg prot are always taken */ @@ -907,52 +908,22 @@ smb311_update_preauth_hash(struct cifs_ses *ses, struct TCP_Server_Info *server, * and we can test it. Preauth requires 3.1.1 for now. */ if (server->dialect != SMB311_PROT_ID) - return 0; + return; if (hdr->Command != SMB2_SESSION_SETUP) - return 0; + return; /* skip last sess setup response */ if ((hdr->Flags & SMB2_FLAGS_SERVER_TO_REDIR) && (hdr->Status == NT_STATUS_OK || (hdr->Status != cpu_to_le32(NT_STATUS_MORE_PROCESSING_REQUIRED)))) - return 0; + return; ok: - rc = smb311_crypto_shash_allocate(server); - if (rc) - return rc; - - sha512 = server->secmech.sha512; - rc = crypto_shash_init(sha512); - if (rc) { - cifs_dbg(VFS, "%s: Could not init sha512 shash\n", __func__); - return rc; - } - - rc = crypto_shash_update(sha512, ses->preauth_sha_hash, - SMB2_PREAUTH_HASH_SIZE); - if (rc) { - cifs_dbg(VFS, "%s: Could not update sha512 shash\n", __func__); - return rc; - } - - for (i = 0; i < nvec; i++) { - rc = crypto_shash_update(sha512, iov[i].iov_base, iov[i].iov_len); - if (rc) { - cifs_dbg(VFS, "%s: Could not update sha512 shash\n", - __func__); - return rc; - } - } - - rc = crypto_shash_final(sha512, ses->preauth_sha_hash); - if (rc) { - cifs_dbg(VFS, "%s: Could not finalize sha512 shash\n", - __func__); - return rc; - } - - return 0; + sha512_init(&sha_ctx); + sha512_update(&sha_ctx, ses->preauth_sha_hash, SMB2_PREAUTH_HASH_SIZE); + for (i = 0; i < nvec; i++) + sha512_update(&sha_ctx, iov[i].iov_base, iov[i].iov_len); + sha512_final(&sha_ctx, ses->preauth_sha_hash); } diff --git a/fs/smb/client/smb2proto.h b/fs/smb/client/smb2proto.h index b3f1398c9f79..e7cda885c39f 100644 --- a/fs/smb/client/smb2proto.h +++ b/fs/smb/client/smb2proto.h @@ -296,9 +296,9 @@ extern void smb2_copy_fs_info_to_kstatfs( struct smb2_fs_full_size_info *pfs_inf, struct kstatfs *kst); extern int smb311_crypto_shash_allocate(struct TCP_Server_Info *server); -extern int smb311_update_preauth_hash(struct cifs_ses *ses, - struct TCP_Server_Info *server, - struct kvec *iov, int nvec); +extern void smb311_update_preauth_hash(struct cifs_ses *ses, + struct TCP_Server_Info *server, + struct kvec *iov, int nvec); extern int smb2_query_info_compound(const unsigned int xid, struct cifs_tcon *tcon, const char *path, u32 desired_access, From 4b4c6fdb25de4edc0a34b1b93cccb439e00e1f35 Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Sat, 11 Oct 2025 18:57:32 -0700 Subject: [PATCH 231/330] smb: client: Use HMAC-SHA256 library for key generation Convert generate_key() to use the HMAC-SHA256 library instead of a "hmac(sha256)" crypto_shash. This is simpler and faster. With the library there's no need to allocate memory, no need to handle errors, and the HMAC-SHA256 code is accessed directly without inefficient indirect calls and other unnecessary API overhead. Also remove the unnecessary 'hashptr' variable. For now smb3_crypto_shash_allocate() still allocates a "hmac(sha256)" crypto_shash. It will be removed in a later commit. Reviewed-by: Stefan Metzmacher Acked-by: Ard Biesheuvel Signed-off-by: Eric Biggers Signed-off-by: Steve French --- fs/smb/client/Kconfig | 1 + fs/smb/client/smb2transport.c | 68 ++++++++--------------------------- 2 files changed, 15 insertions(+), 54 deletions(-) diff --git a/fs/smb/client/Kconfig b/fs/smb/client/Kconfig index 4ac79ff5649b..f0c1ff8544f6 100644 --- a/fs/smb/client/Kconfig +++ b/fs/smb/client/Kconfig @@ -16,6 +16,7 @@ config CIFS select CRYPTO_ECB select CRYPTO_AES select CRYPTO_LIB_ARC4 + select CRYPTO_LIB_SHA256 select CRYPTO_LIB_SHA512 select KEYS select DNS_RESOLVER diff --git a/fs/smb/client/smb2transport.c b/fs/smb/client/smb2transport.c index 33f33013b392..bde96eace8c9 100644 --- a/fs/smb/client/smb2transport.c +++ b/fs/smb/client/smb2transport.c @@ -19,6 +19,7 @@ #include #include #include +#include #include "cifsglob.h" #include "cifsproto.h" #include "smb2proto.h" @@ -336,8 +337,8 @@ static int generate_key(struct cifs_ses *ses, struct kvec label, __u8 L256[4] = {0, 0, 1, 0}; int rc = 0; unsigned char prfhash[SMB2_HMACSHA256_SIZE]; - unsigned char *hashptr = prfhash; struct TCP_Server_Info *server = ses->server; + struct hmac_sha256_ctx hmac_ctx; memset(prfhash, 0x0, SMB2_HMACSHA256_SIZE); memset(key, 0x0, key_size); @@ -345,67 +346,26 @@ static int generate_key(struct cifs_ses *ses, struct kvec label, rc = smb3_crypto_shash_allocate(server); if (rc) { cifs_server_dbg(VFS, "%s: crypto alloc failed\n", __func__); - goto smb3signkey_ret; + return rc; } - rc = crypto_shash_setkey(server->secmech.hmacsha256->tfm, - ses->auth_key.response, SMB2_NTLMV2_SESSKEY_SIZE); - if (rc) { - cifs_server_dbg(VFS, "%s: Could not set with session key\n", __func__); - goto smb3signkey_ret; - } - - rc = crypto_shash_init(server->secmech.hmacsha256); - if (rc) { - cifs_server_dbg(VFS, "%s: Could not init sign hmac\n", __func__); - goto smb3signkey_ret; - } - - rc = crypto_shash_update(server->secmech.hmacsha256, i, 4); - if (rc) { - cifs_server_dbg(VFS, "%s: Could not update with n\n", __func__); - goto smb3signkey_ret; - } - - rc = crypto_shash_update(server->secmech.hmacsha256, label.iov_base, label.iov_len); - if (rc) { - cifs_server_dbg(VFS, "%s: Could not update with label\n", __func__); - goto smb3signkey_ret; - } - - rc = crypto_shash_update(server->secmech.hmacsha256, &zero, 1); - if (rc) { - cifs_server_dbg(VFS, "%s: Could not update with zero\n", __func__); - goto smb3signkey_ret; - } - - rc = crypto_shash_update(server->secmech.hmacsha256, context.iov_base, context.iov_len); - if (rc) { - cifs_server_dbg(VFS, "%s: Could not update with context\n", __func__); - goto smb3signkey_ret; - } + hmac_sha256_init_usingrawkey(&hmac_ctx, ses->auth_key.response, + SMB2_NTLMV2_SESSKEY_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); + hmac_sha256_update(&hmac_ctx, context.iov_base, context.iov_len); if ((server->cipher_type == SMB2_ENCRYPTION_AES256_CCM) || (server->cipher_type == SMB2_ENCRYPTION_AES256_GCM)) { - rc = crypto_shash_update(server->secmech.hmacsha256, L256, 4); + hmac_sha256_update(&hmac_ctx, L256, 4); } else { - rc = crypto_shash_update(server->secmech.hmacsha256, L128, 4); - } - if (rc) { - cifs_server_dbg(VFS, "%s: Could not update with L\n", __func__); - goto smb3signkey_ret; + hmac_sha256_update(&hmac_ctx, L128, 4); } + hmac_sha256_final(&hmac_ctx, prfhash); - rc = crypto_shash_final(server->secmech.hmacsha256, hashptr); - if (rc) { - cifs_server_dbg(VFS, "%s: Could not generate sha256 hash\n", __func__); - goto smb3signkey_ret; - } - - memcpy(key, hashptr, key_size); - -smb3signkey_ret: - return rc; + memcpy(key, prfhash, key_size); + return 0; } struct derivation { From e05b3115e75381369be84abe5d46565ce0fcedc8 Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Sat, 11 Oct 2025 18:57:33 -0700 Subject: [PATCH 232/330] smb: client: Use HMAC-SHA256 library for SMB2 signature calculation Convert smb2_calc_signature() to use the HMAC-SHA256 library instead of a "hmac(sha256)" crypto_shash. This is simpler and faster. With the library there's no need to allocate memory, no need to handle errors, and the HMAC-SHA256 code is accessed directly without inefficient indirect calls and other unnecessary API overhead. To make this possible, make __cifs_calc_signature() support both the HMAC-SHA256 library and crypto_shash. (crypto_shash is still needed for HMAC-MD5 and AES-CMAC. A later commit will switch HMAC-MD5 from shash to the library. I'd like to eventually do the same for AES-CMAC, but it doesn't have a library API yet. So for now, shash is still needed.) Also remove the unnecessary 'sigptr' variable. For now smb3_crypto_shash_allocate() still allocates a "hmac(sha256)" crypto_shash. It will be removed in a later commit. Reviewed-by: Stefan Metzmacher Acked-by: Ard Biesheuvel Signed-off-by: Eric Biggers Signed-off-by: Steve French --- fs/smb/client/cifsencrypt.c | 52 +++++++++++++++++++++++----------- fs/smb/client/cifsproto.h | 9 ++++-- fs/smb/client/smb2transport.c | 53 ++++++++--------------------------- 3 files changed, 53 insertions(+), 61 deletions(-) diff --git a/fs/smb/client/cifsencrypt.c b/fs/smb/client/cifsencrypt.c index 7b7c8c38fdd0..9522088a1cfb 100644 --- a/fs/smb/client/cifsencrypt.c +++ b/fs/smb/client/cifsencrypt.c @@ -24,14 +24,34 @@ #include #include #include +#include -static size_t cifs_shash_step(void *iter_base, size_t progress, size_t len, - void *priv, void *priv2) +static int cifs_sig_update(struct cifs_calc_sig_ctx *ctx, + const u8 *data, size_t len) { - struct shash_desc *shash = priv; + if (ctx->hmac) { + hmac_sha256_update(ctx->hmac, data, len); + return 0; + } + return crypto_shash_update(ctx->shash, data, len); +} + +static int cifs_sig_final(struct cifs_calc_sig_ctx *ctx, u8 *out) +{ + if (ctx->hmac) { + hmac_sha256_final(ctx->hmac, out); + return 0; + } + return crypto_shash_final(ctx->shash, out); +} + +static size_t cifs_sig_step(void *iter_base, size_t progress, size_t len, + void *priv, void *priv2) +{ + struct cifs_calc_sig_ctx *ctx = priv; int ret, *pret = priv2; - ret = crypto_shash_update(shash, iter_base, len); + ret = cifs_sig_update(ctx, iter_base, len); if (ret < 0) { *pret = ret; return len; @@ -42,21 +62,20 @@ static size_t cifs_shash_step(void *iter_base, size_t progress, size_t len, /* * Pass the data from an iterator into a hash. */ -static int cifs_shash_iter(const struct iov_iter *iter, size_t maxsize, - struct shash_desc *shash) +static int cifs_sig_iter(const struct iov_iter *iter, size_t maxsize, + struct cifs_calc_sig_ctx *ctx) { struct iov_iter tmp_iter = *iter; int err = -EIO; - if (iterate_and_advance_kernel(&tmp_iter, maxsize, shash, &err, - cifs_shash_step) != maxsize) + if (iterate_and_advance_kernel(&tmp_iter, maxsize, ctx, &err, + cifs_sig_step) != maxsize) return err; return 0; } -int __cifs_calc_signature(struct smb_rqst *rqst, - struct TCP_Server_Info *server, char *signature, - struct shash_desc *shash) +int __cifs_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, + char *signature, struct cifs_calc_sig_ctx *ctx) { int i; ssize_t rc; @@ -82,8 +101,7 @@ int __cifs_calc_signature(struct smb_rqst *rqst, return -EIO; } - rc = crypto_shash_update(shash, - iov[i].iov_base, iov[i].iov_len); + rc = cifs_sig_update(ctx, iov[i].iov_base, iov[i].iov_len); if (rc) { cifs_dbg(VFS, "%s: Could not update with payload\n", __func__); @@ -91,11 +109,11 @@ int __cifs_calc_signature(struct smb_rqst *rqst, } } - rc = cifs_shash_iter(&rqst->rq_iter, iov_iter_count(&rqst->rq_iter), shash); + rc = cifs_sig_iter(&rqst->rq_iter, iov_iter_count(&rqst->rq_iter), ctx); if (rc < 0) return rc; - rc = crypto_shash_final(shash, signature); + rc = cifs_sig_final(ctx, signature); if (rc) cifs_dbg(VFS, "%s: Could not generate hash\n", __func__); @@ -134,7 +152,9 @@ static int cifs_calc_signature(struct smb_rqst *rqst, return rc; } - return __cifs_calc_signature(rqst, server, signature, server->secmech.md5); + return __cifs_calc_signature( + rqst, server, signature, + &(struct cifs_calc_sig_ctx){ .shash = server->secmech.md5 }); } /* must be called with server->srv_mutex held */ diff --git a/fs/smb/client/cifsproto.h b/fs/smb/client/cifsproto.h index e8fba98690ce..3bb74eea0e4f 100644 --- a/fs/smb/client/cifsproto.h +++ b/fs/smb/client/cifsproto.h @@ -632,9 +632,12 @@ int cifs_create_mf_symlink(unsigned int xid, struct cifs_tcon *tcon, struct cifs_sb_info *cifs_sb, const unsigned char *path, char *pbuf, unsigned int *pbytes_written); -int __cifs_calc_signature(struct smb_rqst *rqst, - struct TCP_Server_Info *server, char *signature, - struct shash_desc *shash); +struct cifs_calc_sig_ctx { + struct hmac_sha256_ctx *hmac; + struct shash_desc *shash; +}; +int __cifs_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, + char *signature, struct cifs_calc_sig_ctx *ctx); enum securityEnum cifs_select_sectype(struct TCP_Server_Info *, enum securityEnum); diff --git a/fs/smb/client/smb2transport.c b/fs/smb/client/smb2transport.c index bde96eace8c9..89258accc220 100644 --- a/fs/smb/client/smb2transport.c +++ b/fs/smb/client/smb2transport.c @@ -254,10 +254,9 @@ smb2_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, { int rc; unsigned char smb2_signature[SMB2_HMACSHA256_SIZE]; - unsigned char *sigptr = smb2_signature; struct kvec *iov = rqst->rq_iov; struct smb2_hdr *shdr = (struct smb2_hdr *)iov[0].iov_base; - struct shash_desc *shash = NULL; + struct hmac_sha256_ctx hmac_ctx; struct smb_rqst drqst; __u64 sid = le64_to_cpu(shdr->SessionId); u8 key[SMB2_NTLMV2_SESSKEY_SIZE]; @@ -272,30 +271,7 @@ smb2_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, memset(smb2_signature, 0x0, SMB2_HMACSHA256_SIZE); memset(shdr->Signature, 0x0, SMB2_SIGNATURE_SIZE); - if (allocate_crypto) { - rc = cifs_alloc_hash("hmac(sha256)", &shash); - if (rc) { - cifs_server_dbg(VFS, - "%s: sha256 alloc failed\n", __func__); - goto out; - } - } else { - shash = server->secmech.hmacsha256; - } - - rc = crypto_shash_setkey(shash->tfm, key, sizeof(key)); - if (rc) { - cifs_server_dbg(VFS, - "%s: Could not update with response\n", - __func__); - goto out; - } - - rc = crypto_shash_init(shash); - if (rc) { - cifs_server_dbg(VFS, "%s: Could not init sha256", __func__); - goto out; - } + hmac_sha256_init_usingrawkey(&hmac_ctx, key, sizeof(key)); /* * For SMB2+, __cifs_calc_signature() expects to sign only the actual @@ -306,25 +282,17 @@ smb2_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, */ drqst = *rqst; if (drqst.rq_nvec >= 2 && iov[0].iov_len == 4) { - rc = crypto_shash_update(shash, iov[0].iov_base, - iov[0].iov_len); - if (rc) { - cifs_server_dbg(VFS, - "%s: Could not update with payload\n", - __func__); - goto out; - } + hmac_sha256_update(&hmac_ctx, iov[0].iov_base, iov[0].iov_len); drqst.rq_iov++; drqst.rq_nvec--; } - rc = __cifs_calc_signature(&drqst, server, sigptr, shash); + rc = __cifs_calc_signature( + &drqst, server, smb2_signature, + &(struct cifs_calc_sig_ctx){ .hmac = &hmac_ctx }); if (!rc) - memcpy(shdr->Signature, sigptr, SMB2_SIGNATURE_SIZE); + memcpy(shdr->Signature, smb2_signature, SMB2_SIGNATURE_SIZE); -out: - if (allocate_crypto) - cifs_free_hash(&shash); return rc; } @@ -542,7 +510,6 @@ smb3_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, { int rc; unsigned char smb3_signature[SMB2_CMACAES_SIZE]; - unsigned char *sigptr = smb3_signature; struct kvec *iov = rqst->rq_iov; struct smb2_hdr *shdr = (struct smb2_hdr *)iov[0].iov_base; struct shash_desc *shash = NULL; @@ -603,9 +570,11 @@ smb3_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, drqst.rq_nvec--; } - rc = __cifs_calc_signature(&drqst, server, sigptr, shash); + rc = __cifs_calc_signature( + &drqst, server, smb3_signature, + &(struct cifs_calc_sig_ctx){ .shash = shash }); if (!rc) - memcpy(shdr->Signature, sigptr, SMB2_SIGNATURE_SIZE); + memcpy(shdr->Signature, smb3_signature, SMB2_SIGNATURE_SIZE); out: if (allocate_crypto) From ae04b1bb06f8c1738d01dc2f9b9391c4480544e4 Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Sat, 11 Oct 2025 18:57:34 -0700 Subject: [PATCH 233/330] smb: client: Use MD5 library for M-F symlink hashing Convert parse_mf_symlink() and format_mf_symlink() to use the MD5 library instead of a "md5" crypto_shash. This is simpler and faster. With the library there's no need to allocate memory, no need to handle errors, and the MD5 code is accessed directly without inefficient indirect calls and other unnecessary API overhead. This also fixes an issue where these functions did not work on kernels booted in FIPS mode. The use of MD5 here is for data integrity rather than a security purpose, so it can use a non-FIPS-approved algorithm. Reviewed-by: Stefan Metzmacher Acked-by: Ard Biesheuvel Signed-off-by: Eric Biggers Signed-off-by: Steve French --- fs/smb/client/Kconfig | 1 + fs/smb/client/link.c | 31 +++---------------------------- 2 files changed, 4 insertions(+), 28 deletions(-) diff --git a/fs/smb/client/Kconfig b/fs/smb/client/Kconfig index f0c1ff8544f6..f5a980bdfc93 100644 --- a/fs/smb/client/Kconfig +++ b/fs/smb/client/Kconfig @@ -16,6 +16,7 @@ config CIFS select CRYPTO_ECB select CRYPTO_AES select CRYPTO_LIB_ARC4 + select CRYPTO_LIB_MD5 select CRYPTO_LIB_SHA256 select CRYPTO_LIB_SHA512 select KEYS diff --git a/fs/smb/client/link.c b/fs/smb/client/link.c index fe80e711cd75..70f3c0c67eeb 100644 --- a/fs/smb/client/link.c +++ b/fs/smb/client/link.c @@ -5,6 +5,7 @@ * Author(s): Steve French (sfrench@us.ibm.com) * */ +#include #include #include #include @@ -36,23 +37,6 @@ #define CIFS_MF_SYMLINK_MD5_FORMAT "%16phN\n" #define CIFS_MF_SYMLINK_MD5_ARGS(md5_hash) md5_hash -static int -symlink_hash(unsigned int link_len, const char *link_str, u8 *md5_hash) -{ - int rc; - struct shash_desc *md5 = NULL; - - rc = cifs_alloc_hash("md5", &md5); - if (rc) - return rc; - - rc = crypto_shash_digest(md5, link_str, link_len, md5_hash); - if (rc) - cifs_dbg(VFS, "%s: Could not generate md5 hash\n", __func__); - cifs_free_hash(&md5); - return rc; -} - static int parse_mf_symlink(const u8 *buf, unsigned int buf_len, unsigned int *_link_len, char **_link_str) @@ -77,11 +61,7 @@ parse_mf_symlink(const u8 *buf, unsigned int buf_len, unsigned int *_link_len, if (link_len > CIFS_MF_SYMLINK_LINK_MAXLEN) return -EINVAL; - rc = symlink_hash(link_len, link_str, md5_hash); - if (rc) { - cifs_dbg(FYI, "%s: MD5 hash failure: %d\n", __func__, rc); - return rc; - } + md5(link_str, link_len, md5_hash); scnprintf(md5_str2, sizeof(md5_str2), CIFS_MF_SYMLINK_MD5_FORMAT, @@ -103,7 +83,6 @@ parse_mf_symlink(const u8 *buf, unsigned int buf_len, unsigned int *_link_len, static int format_mf_symlink(u8 *buf, unsigned int buf_len, const char *link_str) { - int rc; unsigned int link_len; unsigned int ofs; u8 md5_hash[16]; @@ -116,11 +95,7 @@ format_mf_symlink(u8 *buf, unsigned int buf_len, const char *link_str) if (link_len > CIFS_MF_SYMLINK_LINK_MAXLEN) return -ENAMETOOLONG; - rc = symlink_hash(link_len, link_str, md5_hash); - if (rc) { - cifs_dbg(FYI, "%s: MD5 hash failure: %d\n", __func__, rc); - return rc; - } + md5(link_str, link_len, md5_hash); scnprintf(buf, buf_len, CIFS_MF_SYMLINK_LEN_FORMAT CIFS_MF_SYMLINK_MD5_FORMAT, From c04e55b257b42f5eb5a2c5e92ebd043fd75fe3ab Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Sat, 11 Oct 2025 18:57:35 -0700 Subject: [PATCH 234/330] smb: client: Use MD5 library for SMB1 signature calculation Convert cifs_calc_signature() to use the MD5 library instead of a "md5" crypto_shash. This is simpler and faster. With the library there's no need to allocate memory, no need to handle errors, and the MD5 code is accessed directly without inefficient indirect calls and other unnecessary API overhead. To preserve the existing behavior of MD5 signature support being disabled when the kernel is booted with "fips=1", make cifs_calc_signature() check fips_enabled itself. Previously it relied on the error from cifs_alloc_hash("md5", &server->secmech.md5). Reviewed-by: Stefan Metzmacher Acked-by: Ard Biesheuvel Signed-off-by: Eric Biggers Signed-off-by: Steve French --- fs/smb/client/cifsencrypt.c | 34 +++++++++++++++++----------------- fs/smb/client/cifsproto.h | 1 + 2 files changed, 18 insertions(+), 17 deletions(-) diff --git a/fs/smb/client/cifsencrypt.c b/fs/smb/client/cifsencrypt.c index 9522088a1cfb..80215ba7a574 100644 --- a/fs/smb/client/cifsencrypt.c +++ b/fs/smb/client/cifsencrypt.c @@ -24,11 +24,16 @@ #include #include #include +#include #include static int cifs_sig_update(struct cifs_calc_sig_ctx *ctx, const u8 *data, size_t len) { + if (ctx->md5) { + md5_update(ctx->md5, data, len); + return 0; + } if (ctx->hmac) { hmac_sha256_update(ctx->hmac, data, len); return 0; @@ -38,6 +43,10 @@ static int cifs_sig_update(struct cifs_calc_sig_ctx *ctx, static int cifs_sig_final(struct cifs_calc_sig_ctx *ctx, u8 *out) { + if (ctx->md5) { + md5_final(ctx->md5, out); + return 0; + } if (ctx->hmac) { hmac_sha256_final(ctx->hmac, out); return 0; @@ -130,31 +139,22 @@ int __cifs_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, static int cifs_calc_signature(struct smb_rqst *rqst, struct TCP_Server_Info *server, char *signature) { - int rc; + struct md5_ctx ctx; if (!rqst->rq_iov || !signature || !server) return -EINVAL; - - rc = cifs_alloc_hash("md5", &server->secmech.md5); - if (rc) - return -1; - - rc = crypto_shash_init(server->secmech.md5); - if (rc) { - cifs_dbg(VFS, "%s: Could not init md5\n", __func__); - return rc; + if (fips_enabled) { + cifs_dbg(VFS, + "MD5 signature support is disabled due to FIPS\n"); + return -EOPNOTSUPP; } - rc = crypto_shash_update(server->secmech.md5, - server->session_key.response, server->session_key.len); - if (rc) { - cifs_dbg(VFS, "%s: Could not update with response\n", __func__); - return rc; - } + md5_init(&ctx); + md5_update(&ctx, server->session_key.response, server->session_key.len); return __cifs_calc_signature( rqst, server, signature, - &(struct cifs_calc_sig_ctx){ .shash = server->secmech.md5 }); + &(struct cifs_calc_sig_ctx){ .md5 = &ctx }); } /* must be called with server->srv_mutex held */ diff --git a/fs/smb/client/cifsproto.h b/fs/smb/client/cifsproto.h index 3bb74eea0e4f..4976be2c47c1 100644 --- a/fs/smb/client/cifsproto.h +++ b/fs/smb/client/cifsproto.h @@ -633,6 +633,7 @@ int cifs_create_mf_symlink(unsigned int xid, struct cifs_tcon *tcon, const unsigned char *path, char *pbuf, unsigned int *pbytes_written); struct cifs_calc_sig_ctx { + struct md5_ctx *md5; struct hmac_sha256_ctx *hmac; struct shash_desc *shash; }; From 395a77b030a878a353465386e8618b5272a480ca Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Sat, 11 Oct 2025 18:57:36 -0700 Subject: [PATCH 235/330] smb: client: Use HMAC-MD5 library for NTLMv2 For the HMAC-MD5 computations in NTLMv2, use the HMAC-MD5 library instead of a "hmac(md5)" crypto_shash. This is simpler and faster. With the library there's no need to allocate memory, no need to handle errors, and the HMAC-MD5 code is accessed directly without inefficient indirect calls and other unnecessary API overhead. To preserve the existing behavior of NTLMv2 support being disabled when the kernel is booted with "fips=1", make setup_ntlmv2_rsp() check fips_enabled itself. Previously it relied on the error from cifs_alloc_hash("hmac(md5)", &hmacmd5). Reviewed-by: Stefan Metzmacher Acked-by: Ard Biesheuvel Signed-off-by: Eric Biggers Signed-off-by: Steve French --- fs/smb/client/cifsencrypt.c | 114 +++++++----------------------------- 1 file changed, 22 insertions(+), 92 deletions(-) diff --git a/fs/smb/client/cifsencrypt.c b/fs/smb/client/cifsencrypt.c index 80215ba7a574..bbcf3b05c19a 100644 --- a/fs/smb/client/cifsencrypt.c +++ b/fs/smb/client/cifsencrypt.c @@ -425,11 +425,11 @@ static __le64 find_timestamp(struct cifs_ses *ses) } static int calc_ntlmv2_hash(struct cifs_ses *ses, char *ntlmv2_hash, - const struct nls_table *nls_cp, struct shash_desc *hmacmd5) + const struct nls_table *nls_cp) { - int rc = 0; int len; char nt_hash[CIFS_NTHASH_SIZE]; + struct hmac_md5_ctx hmac_ctx; __le16 *user; wchar_t *domain; wchar_t *server; @@ -437,17 +437,7 @@ static int calc_ntlmv2_hash(struct cifs_ses *ses, char *ntlmv2_hash, /* calculate md4 hash of password */ E_md4hash(ses->password, nt_hash, nls_cp); - rc = crypto_shash_setkey(hmacmd5->tfm, nt_hash, CIFS_NTHASH_SIZE); - if (rc) { - cifs_dbg(VFS, "%s: Could not set NT hash as a key, rc=%d\n", __func__, rc); - return rc; - } - - rc = crypto_shash_init(hmacmd5); - if (rc) { - cifs_dbg(VFS, "%s: Could not init HMAC-MD5, rc=%d\n", __func__, rc); - return rc; - } + hmac_md5_init_usingrawkey(&hmac_ctx, nt_hash, CIFS_NTHASH_SIZE); /* convert ses->user_name to unicode */ len = ses->user_name ? strlen(ses->user_name) : 0; @@ -462,12 +452,8 @@ static int calc_ntlmv2_hash(struct cifs_ses *ses, char *ntlmv2_hash, *(u16 *)user = 0; } - rc = crypto_shash_update(hmacmd5, (char *)user, 2 * len); + hmac_md5_update(&hmac_ctx, (const u8 *)user, 2 * len); kfree(user); - if (rc) { - cifs_dbg(VFS, "%s: Could not update with user, rc=%d\n", __func__, rc); - return rc; - } /* convert ses->domainName to unicode and uppercase */ if (ses->domainName) { @@ -479,12 +465,8 @@ static int calc_ntlmv2_hash(struct cifs_ses *ses, char *ntlmv2_hash, len = cifs_strtoUTF16((__le16 *)domain, ses->domainName, len, nls_cp); - rc = crypto_shash_update(hmacmd5, (char *)domain, 2 * len); + hmac_md5_update(&hmac_ctx, (const u8 *)domain, 2 * len); kfree(domain); - if (rc) { - cifs_dbg(VFS, "%s: Could not update with domain, rc=%d\n", __func__, rc); - return rc; - } } else { /* We use ses->ip_addr if no domain name available */ len = strlen(ses->ip_addr); @@ -494,25 +476,16 @@ static int calc_ntlmv2_hash(struct cifs_ses *ses, char *ntlmv2_hash, return -ENOMEM; len = cifs_strtoUTF16((__le16 *)server, ses->ip_addr, len, nls_cp); - rc = crypto_shash_update(hmacmd5, (char *)server, 2 * len); + hmac_md5_update(&hmac_ctx, (const u8 *)server, 2 * len); kfree(server); - if (rc) { - cifs_dbg(VFS, "%s: Could not update with server, rc=%d\n", __func__, rc); - return rc; - } } - rc = crypto_shash_final(hmacmd5, ntlmv2_hash); - if (rc) - cifs_dbg(VFS, "%s: Could not generate MD5 hash, rc=%d\n", __func__, rc); - - return rc; + hmac_md5_final(&hmac_ctx, ntlmv2_hash); + return 0; } -static int -CalcNTLMv2_response(const struct cifs_ses *ses, char *ntlmv2_hash, struct shash_desc *hmacmd5) +static void CalcNTLMv2_response(const struct cifs_ses *ses, char *ntlmv2_hash) { - int rc; struct ntlmv2_resp *ntlmv2 = (struct ntlmv2_resp *) (ses->auth_key.response + CIFS_SESS_KEY_SIZE); unsigned int hash_len; @@ -521,35 +494,15 @@ CalcNTLMv2_response(const struct cifs_ses *ses, char *ntlmv2_hash, struct shash_ hash_len = ses->auth_key.len - (CIFS_SESS_KEY_SIZE + offsetof(struct ntlmv2_resp, challenge.key[0])); - rc = crypto_shash_setkey(hmacmd5->tfm, ntlmv2_hash, CIFS_HMAC_MD5_HASH_SIZE); - if (rc) { - cifs_dbg(VFS, "%s: Could not set NTLMv2 hash as a key, rc=%d\n", __func__, rc); - return rc; - } - - rc = crypto_shash_init(hmacmd5); - if (rc) { - cifs_dbg(VFS, "%s: Could not init HMAC-MD5, rc=%d\n", __func__, rc); - return rc; - } - if (ses->server->negflavor == CIFS_NEGFLAVOR_EXTENDED) memcpy(ntlmv2->challenge.key, ses->ntlmssp->cryptkey, CIFS_SERVER_CHALLENGE_SIZE); else memcpy(ntlmv2->challenge.key, ses->server->cryptkey, CIFS_SERVER_CHALLENGE_SIZE); - rc = crypto_shash_update(hmacmd5, ntlmv2->challenge.key, hash_len); - if (rc) { - cifs_dbg(VFS, "%s: Could not update with response, rc=%d\n", __func__, rc); - return rc; - } - - /* Note that the MD5 digest over writes anon.challenge_key.key */ - rc = crypto_shash_final(hmacmd5, ntlmv2->ntlmv2_hash); - if (rc) - cifs_dbg(VFS, "%s: Could not generate MD5 hash, rc=%d\n", __func__, rc); - - return rc; + /* Note that the HMAC-MD5 value overwrites ntlmv2->challenge.key */ + hmac_md5_usingrawkey(ntlmv2_hash, CIFS_HMAC_MD5_HASH_SIZE, + ntlmv2->challenge.key, hash_len, + ntlmv2->ntlmv2_hash); } /* @@ -606,7 +559,6 @@ static int set_auth_key_response(struct cifs_ses *ses) int setup_ntlmv2_rsp(struct cifs_ses *ses, const struct nls_table *nls_cp) { - struct shash_desc *hmacmd5 = NULL; unsigned char *tiblob = NULL; /* target info blob */ struct ntlmv2_resp *ntlmv2; char ntlmv2_hash[16]; @@ -677,51 +629,29 @@ setup_ntlmv2_rsp(struct cifs_ses *ses, const struct nls_table *nls_cp) ntlmv2->client_chal = cc; ntlmv2->reserved2 = 0; - rc = cifs_alloc_hash("hmac(md5)", &hmacmd5); - if (rc) { - cifs_dbg(VFS, "Could not allocate HMAC-MD5, rc=%d\n", rc); + if (fips_enabled) { + cifs_dbg(VFS, "NTLMv2 support is disabled due to FIPS\n"); + rc = -EOPNOTSUPP; goto unlock; } /* calculate ntlmv2_hash */ - rc = calc_ntlmv2_hash(ses, ntlmv2_hash, nls_cp, hmacmd5); + rc = calc_ntlmv2_hash(ses, ntlmv2_hash, nls_cp); if (rc) { cifs_dbg(VFS, "Could not get NTLMv2 hash, rc=%d\n", rc); goto unlock; } /* calculate first part of the client response (CR1) */ - rc = CalcNTLMv2_response(ses, ntlmv2_hash, hmacmd5); - if (rc) { - cifs_dbg(VFS, "Could not calculate CR1, rc=%d\n", rc); - goto unlock; - } + CalcNTLMv2_response(ses, ntlmv2_hash); /* now calculate the session key for NTLMv2 */ - rc = crypto_shash_setkey(hmacmd5->tfm, ntlmv2_hash, CIFS_HMAC_MD5_HASH_SIZE); - if (rc) { - cifs_dbg(VFS, "%s: Could not set NTLMv2 hash as a key, rc=%d\n", __func__, rc); - goto unlock; - } - - rc = crypto_shash_init(hmacmd5); - if (rc) { - cifs_dbg(VFS, "%s: Could not init HMAC-MD5, rc=%d\n", __func__, rc); - goto unlock; - } - - rc = crypto_shash_update(hmacmd5, ntlmv2->ntlmv2_hash, CIFS_HMAC_MD5_HASH_SIZE); - if (rc) { - cifs_dbg(VFS, "%s: Could not update with response, rc=%d\n", __func__, rc); - goto unlock; - } - - rc = crypto_shash_final(hmacmd5, ses->auth_key.response); - if (rc) - cifs_dbg(VFS, "%s: Could not generate MD5 hash, rc=%d\n", __func__, rc); + hmac_md5_usingrawkey(ntlmv2_hash, CIFS_HMAC_MD5_HASH_SIZE, + ntlmv2->ntlmv2_hash, CIFS_HMAC_MD5_HASH_SIZE, + ses->auth_key.response); + rc = 0; unlock: cifs_server_unlock(ses->server); - cifs_free_hash(&hmacmd5); setup_ntlmv2_rsp_ret: kfree_sensitive(tiblob); From 2c09630d09c64b6b46e3d59a0031bc1807f742c4 Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Sat, 11 Oct 2025 18:57:37 -0700 Subject: [PATCH 236/330] smb: client: Remove obsolete crypto_shash allocations Now that the SMB client accesses MD5, HMAC-MD5, HMAC-SHA256, and SHA-512 only via the library API and not via crypto_shash, allocating crypto_shash objects for these algorithms is no longer necessary. Remove all these allocations, their corresponding kconfig selections, and their corresponding module soft dependencies. Reviewed-by: Stefan Metzmacher Acked-by: Ard Biesheuvel Signed-off-by: Eric Biggers Signed-off-by: Steve French --- fs/smb/client/Kconfig | 4 ---- fs/smb/client/cifsencrypt.c | 3 --- fs/smb/client/cifsfs.c | 4 ---- fs/smb/client/cifsglob.h | 3 --- fs/smb/client/smb2transport.c | 35 ++--------------------------------- 5 files changed, 2 insertions(+), 47 deletions(-) diff --git a/fs/smb/client/Kconfig b/fs/smb/client/Kconfig index f5a980bdfc93..17bd368574e9 100644 --- a/fs/smb/client/Kconfig +++ b/fs/smb/client/Kconfig @@ -5,11 +5,7 @@ config CIFS select NLS select NLS_UCS2_UTILS select CRYPTO - select CRYPTO_MD5 - select CRYPTO_SHA256 - select CRYPTO_SHA512 select CRYPTO_CMAC - select CRYPTO_HMAC select CRYPTO_AEAD2 select CRYPTO_CCM select CRYPTO_GCM diff --git a/fs/smb/client/cifsencrypt.c b/fs/smb/client/cifsencrypt.c index bbcf3b05c19a..801824825ecf 100644 --- a/fs/smb/client/cifsencrypt.c +++ b/fs/smb/client/cifsencrypt.c @@ -693,9 +693,6 @@ void cifs_crypto_secmech_release(struct TCP_Server_Info *server) { cifs_free_hash(&server->secmech.aes_cmac); - cifs_free_hash(&server->secmech.hmacsha256); - cifs_free_hash(&server->secmech.md5); - cifs_free_hash(&server->secmech.sha512); if (server->secmech.enc) { crypto_free_aead(server->secmech.enc); diff --git a/fs/smb/client/cifsfs.c b/fs/smb/client/cifsfs.c index 05b1fa76e8cc..4f959f1e08d2 100644 --- a/fs/smb/client/cifsfs.c +++ b/fs/smb/client/cifsfs.c @@ -2139,13 +2139,9 @@ MODULE_DESCRIPTION "also older servers complying with the SNIA CIFS Specification)"); MODULE_VERSION(CIFS_VERSION); MODULE_SOFTDEP("ecb"); -MODULE_SOFTDEP("hmac"); -MODULE_SOFTDEP("md5"); MODULE_SOFTDEP("nls"); MODULE_SOFTDEP("aes"); MODULE_SOFTDEP("cmac"); -MODULE_SOFTDEP("sha256"); -MODULE_SOFTDEP("sha512"); MODULE_SOFTDEP("aead2"); MODULE_SOFTDEP("ccm"); MODULE_SOFTDEP("gcm"); diff --git a/fs/smb/client/cifsglob.h b/fs/smb/client/cifsglob.h index c5034cf9ac9e..16a00a61fd2c 100644 --- a/fs/smb/client/cifsglob.h +++ b/fs/smb/client/cifsglob.h @@ -222,9 +222,6 @@ struct session_key { /* crypto hashing related structure/fields, not specific to a sec mech */ struct cifs_secmech { - struct shash_desc *md5; /* md5 hash function, for CIFS/SMB1 signatures */ - struct shash_desc *hmacsha256; /* hmac-sha256 hash function, for SMB2 signatures */ - struct shash_desc *sha512; /* sha512 hash function, for SMB3.1.1 preauth hash */ struct shash_desc *aes_cmac; /* block-cipher based MAC function, for SMB3 signatures */ struct crypto_aead *enc; /* smb3 encryption AEAD TFM (AES-CCM and AES-GCM) */ diff --git a/fs/smb/client/smb2transport.c b/fs/smb/client/smb2transport.c index 89258accc220..cd689bc27bfd 100644 --- a/fs/smb/client/smb2transport.c +++ b/fs/smb/client/smb2transport.c @@ -31,49 +31,18 @@ static int smb3_crypto_shash_allocate(struct TCP_Server_Info *server) { struct cifs_secmech *p = &server->secmech; - int rc; - rc = cifs_alloc_hash("hmac(sha256)", &p->hmacsha256); - if (rc) - goto err; - - rc = cifs_alloc_hash("cmac(aes)", &p->aes_cmac); - if (rc) - goto err; - - return 0; -err: - cifs_free_hash(&p->hmacsha256); - return rc; + return cifs_alloc_hash("cmac(aes)", &p->aes_cmac); } int smb311_crypto_shash_allocate(struct TCP_Server_Info *server) { struct cifs_secmech *p = &server->secmech; - int rc = 0; - rc = cifs_alloc_hash("hmac(sha256)", &p->hmacsha256); - if (rc) - return rc; - - rc = cifs_alloc_hash("cmac(aes)", &p->aes_cmac); - if (rc) - goto err; - - rc = cifs_alloc_hash("sha512", &p->sha512); - if (rc) - goto err; - - return 0; - -err: - cifs_free_hash(&p->aes_cmac); - cifs_free_hash(&p->hmacsha256); - return rc; + return cifs_alloc_hash("cmac(aes)", &p->aes_cmac); } - static int smb3_get_sign_key(__u64 ses_id, struct TCP_Server_Info *server, u8 *key) { From 3c15a6df61bab034b087f00181408b1537a535bb Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Sat, 11 Oct 2025 18:57:38 -0700 Subject: [PATCH 237/330] smb: client: Consolidate cmac(aes) shash allocation Now that smb3_crypto_shash_allocate() and smb311_crypto_shash_allocate() are identical and only allocate "cmac(aes)", delete the latter and replace the call to it with the former. Reviewed-by: Stefan Metzmacher Acked-by: Ard Biesheuvel Signed-off-by: Eric Biggers Signed-off-by: Steve French --- fs/smb/client/sess.c | 2 +- fs/smb/client/smb2proto.h | 2 +- fs/smb/client/smb2transport.c | 10 +--------- 3 files changed, 3 insertions(+), 11 deletions(-) diff --git a/fs/smb/client/sess.c b/fs/smb/client/sess.c index 0a8c2fcc9ded..ef3b498b0a02 100644 --- a/fs/smb/client/sess.c +++ b/fs/smb/client/sess.c @@ -584,7 +584,7 @@ cifs_ses_add_channel(struct cifs_ses *ses, * to sign packets before we generate the channel signing key * (we sign with the session key) */ - rc = smb311_crypto_shash_allocate(chan->server); + rc = smb3_crypto_shash_allocate(chan->server); if (rc) { cifs_dbg(VFS, "%s: crypto alloc failed\n", __func__); mutex_unlock(&ses->session_mutex); diff --git a/fs/smb/client/smb2proto.h b/fs/smb/client/smb2proto.h index e7cda885c39f..6eb86d134abc 100644 --- a/fs/smb/client/smb2proto.h +++ b/fs/smb/client/smb2proto.h @@ -295,7 +295,7 @@ extern int smb2_validate_and_copy_iov(unsigned int offset, extern void smb2_copy_fs_info_to_kstatfs( struct smb2_fs_full_size_info *pfs_inf, struct kstatfs *kst); -extern int smb311_crypto_shash_allocate(struct TCP_Server_Info *server); +extern int smb3_crypto_shash_allocate(struct TCP_Server_Info *server); extern void smb311_update_preauth_hash(struct cifs_ses *ses, struct TCP_Server_Info *server, struct kvec *iov, int nvec); diff --git a/fs/smb/client/smb2transport.c b/fs/smb/client/smb2transport.c index cd689bc27bfd..ad6068e17a2a 100644 --- a/fs/smb/client/smb2transport.c +++ b/fs/smb/client/smb2transport.c @@ -27,16 +27,8 @@ #include "../common/smb2status.h" #include "smb2glob.h" -static int -smb3_crypto_shash_allocate(struct TCP_Server_Info *server) -{ - struct cifs_secmech *p = &server->secmech; - - return cifs_alloc_hash("cmac(aes)", &p->aes_cmac); -} - int -smb311_crypto_shash_allocate(struct TCP_Server_Info *server) +smb3_crypto_shash_allocate(struct TCP_Server_Info *server) { struct cifs_secmech *p = &server->secmech; From 7987b93e3a11a7a95ddf2b21563d3286661b999c Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Thomas=20Hellstr=C3=B6m?= Date: Fri, 10 Oct 2025 12:41:48 +0200 Subject: [PATCH 238/330] drm/xe/svm: Ensure data will be migrated to system if indicated by madvise. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If the location madvise() is set to DRM_XE_PREFERRED_LOC_DEFAULT_SYSTEM, the drm_pagemap in the SVM gpu fault handler will be set to NULL. However there is nothing that explicitly migrates the data to system if it is already present in device memory. In that case, set the device memory owner to NULL to ensure data gets properly migrated to system on page-fault. v2: - Remove redundant dpagemap assignment (Himal Prasad Ghimiray) Signed-off-by: Thomas Hellström Reviewed-by: Matthew Brost #v1 Reviewed-by: Himal Prasad Ghimiray Link: https://lore.kernel.org/r/20251010104149.72783-2-thomas.hellstrom@linux.intel.com Fixes: 10aa5c806030 ("drm/gpusvm, drm/xe: Fix userptr to not allow device private pages") (cherry picked from commit 2cfcea7a745794f9b8e265a309717ca6ba335fc4) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_svm.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c index b268ee0d2271..da2a412f80c0 100644 --- a/drivers/gpu/drm/xe/xe_svm.c +++ b/drivers/gpu/drm/xe/xe_svm.c @@ -1034,6 +1034,9 @@ static int __xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma, if (err) return err; + dpagemap = xe_vma_resolve_pagemap(vma, tile); + if (!dpagemap && !ctx.devmem_only) + ctx.device_private_page_owner = NULL; range = xe_svm_range_find_or_insert(vm, fault_addr, vma, &ctx); if (IS_ERR(range)) @@ -1054,7 +1057,6 @@ static int __xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma, range_debug(range, "PAGE FAULT"); - dpagemap = xe_vma_resolve_pagemap(vma, tile); if (--migrate_try_count >= 0 && xe_svm_range_needs_migrate_to_vram(range, vma, !!dpagemap || ctx.devmem_only)) { ktime_t migrate_start = xe_svm_stats_ktime_get(); From 6d36f65ba551d28710c3e1aaceecacf19df0cd8f Mon Sep 17 00:00:00 2001 From: Matt Roper Date: Mon, 13 Oct 2025 08:30:15 -0700 Subject: [PATCH 239/330] drm/xe/kunit: Fix kerneldoc for parameterized tests Kunit's generate_params() was recently updated to take an additional test context parameter. Xe's IP and platform parameter generators were updated accordingly at the same time, but the new parameter was not added to the functions' kerneldoc, resulting in the following warnings: Warning: drivers/gpu/drm/xe/tests/xe_pci.c:78 function parameter 'test' not described in 'xe_pci_fake_data_gen_params' Warning: drivers/gpu/drm/xe/tests/xe_pci.c:254 function parameter 'test' not described in 'xe_pci_graphics_ip_gen_param' Warning: drivers/gpu/drm/xe/tests/xe_pci.c:278 function parameter 'test' not described in 'xe_pci_media_ip_gen_param' Warning: drivers/gpu/drm/xe/tests/xe_pci.c:302 function parameter 'test' not described in 'xe_pci_id_gen_param' Warning: drivers/gpu/drm/xe/tests/xe_pci.c:390 function parameter 'test' not described in 'xe_pci_live_device_gen_param' 5 warnings as errors Document the new parameter to eliminate the warnings and make CI happy. Fixes: b9a214b5f6aa ("kunit: Pass parameterized test context to generate_params()") Reviewed-by: Shuicheng Lin Link: https://lore.kernel.org/r/20251013153014.2362879-2-matthew.d.roper@intel.com Signed-off-by: Matt Roper (cherry picked from commit 89e347f8a70165d1e8d88a93d875da7742c902ce) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/tests/xe_pci.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/gpu/drm/xe/tests/xe_pci.c b/drivers/gpu/drm/xe/tests/xe_pci.c index 69e2840c7ef0..663a79ec960d 100644 --- a/drivers/gpu/drm/xe/tests/xe_pci.c +++ b/drivers/gpu/drm/xe/tests/xe_pci.c @@ -66,6 +66,7 @@ KUNIT_ARRAY_PARAM(platform, cases, xe_pci_fake_data_desc); /** * xe_pci_fake_data_gen_params - Generate struct xe_pci_fake_data parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -242,6 +243,7 @@ KUNIT_ARRAY_PARAM(pci_id, pciidlist, xe_pci_id_kunit_desc); /** * xe_pci_graphics_ip_gen_param - Generate graphics struct xe_ip parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -266,6 +268,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_graphics_ip_gen_param); /** * xe_pci_media_ip_gen_param - Generate media struct xe_ip parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -290,6 +293,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_media_ip_gen_param); /** * xe_pci_id_gen_param - Generate struct pci_device_id parameters + * @test: test context object * @prev: the pointer to the previous parameter to iterate from or NULL * @desc: output buffer with minimum size of KUNIT_PARAM_DESC_SIZE * @@ -376,6 +380,7 @@ EXPORT_SYMBOL_IF_KUNIT(xe_pci_fake_device_init); /** * xe_pci_live_device_gen_param - Helper to iterate Xe devices as KUnit parameters + * @test: test context object * @prev: the previously returned value, or NULL for the first iteration * @desc: the buffer for a parameter name * From 6a91af25cdbce2086d85cc4994cf791bda3a2c90 Mon Sep 17 00:00:00 2001 From: Matthew Auld Date: Fri, 10 Oct 2025 17:20:21 +0100 Subject: [PATCH 240/330] drm/xe/migrate: don't misalign current bytes If current bytes exceeds the max copy size, ensure the clamped size still accounts for the XE_CACHELINE_BYTES alignment, otherwise we trigger the assert in xe_migrate_vram with the size now being out of alignment. Fixes: 8c2d61e0e916 ("drm/xe/migrate: don't overflow max copy size") Link: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/6212 Signed-off-by: Matthew Auld Cc: Stuart Summers Cc: Matthew Brost Reviewed-by: Matthew Brost Link: https://lore.kernel.org/r/20251010162020.190962-2-matthew.auld@intel.com (cherry picked from commit 641bcf8731d21b56760e3646a39a65f471e9efd1) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_migrate.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c index 569869a2b339..a36ce7dce8cc 100644 --- a/drivers/gpu/drm/xe/xe_migrate.c +++ b/drivers/gpu/drm/xe/xe_migrate.c @@ -2113,7 +2113,9 @@ int xe_migrate_access_memory(struct xe_migrate *m, struct xe_bo *bo, if (current_bytes & ~PAGE_MASK) { int pitch = 4; - current_bytes = min_t(int, current_bytes, S16_MAX * pitch); + current_bytes = min_t(int, current_bytes, + round_down(S16_MAX * pitch, + XE_CACHELINE_BYTES)); } __fence = xe_migrate_vram(m, current_bytes, From 225bc03d85427e7e3821d6f99f4f2d4a09350dda Mon Sep 17 00:00:00 2001 From: Matthew Auld Date: Fri, 10 Oct 2025 16:24:58 +0100 Subject: [PATCH 241/330] drm/xe/evict: drop bogus assert MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This assert can trigger here with non pin_map users that select LATE_RESTORE, since the vmap is allowed to be NULL given that save/restore can now use the blitter instead. The check here doesn't seem to have much value anymore given that we no longer move pinned memory, so any existing vmap is left well alone, and doesn't need to be recreated upon restore, so just drop the assert here. Fixes: 86f69c26113c ("drm/xe: use backup object for pinned save/restore") Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/6213 Signed-off-by: Matthew Auld Cc: Thomas Hellström Cc: Matthew Brost Reviewed-by: Thomas Hellström Link: https://lore.kernel.org/r/20251010152457.177884-2-matthew.auld@intel.com (cherry picked from commit a10b4a69c7f8f596d2c5218fbe84430734fab3b2) Signed-off-by: Lucas De Marchi --- drivers/gpu/drm/xe/xe_bo_evict.c | 8 -------- 1 file changed, 8 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_bo_evict.c b/drivers/gpu/drm/xe/xe_bo_evict.c index d5dbc51e8612..bc5b4c5fab81 100644 --- a/drivers/gpu/drm/xe/xe_bo_evict.c +++ b/drivers/gpu/drm/xe/xe_bo_evict.c @@ -182,7 +182,6 @@ int xe_bo_evict_all(struct xe_device *xe) static int xe_bo_restore_and_map_ggtt(struct xe_bo *bo) { - struct xe_device *xe = xe_bo_device(bo); int ret; ret = xe_bo_restore_pinned(bo); @@ -201,13 +200,6 @@ static int xe_bo_restore_and_map_ggtt(struct xe_bo *bo) } } - /* - * We expect validate to trigger a move VRAM and our move code - * should setup the iosys map. - */ - xe_assert(xe, !(bo->flags & XE_BO_FLAG_PINNED_LATE_RESTORE) || - !iosys_map_is_null(&bo->vmap)); - return 0; } From 9af61fc91486c7ba93cf1ec3bd381978cac8308c Mon Sep 17 00:00:00 2001 From: Pauli Virtanen Date: Wed, 15 Oct 2025 22:47:10 +0300 Subject: [PATCH 242/330] ALSA: usb-audio: add volume quirks for MS LifeChat LX-3000 ID 045e:070f Microsoft Corp. LifeChat LX-3000 Headset has muted minimum Speaker Playback Volume, and 4 amixer steps were observed to produce 1 actual volume step. Apply min_mute quirk and correct res=48 -> 4*48. Tested with the device. Signed-off-by: Pauli Virtanen Signed-off-by: Takashi Iwai --- sound/usb/mixer.c | 8 ++++++++ sound/usb/quirks.c | 2 ++ 2 files changed, 10 insertions(+) diff --git a/sound/usb/mixer.c b/sound/usb/mixer.c index ae412e651faf..6f00e0d52382 100644 --- a/sound/usb/mixer.c +++ b/sound/usb/mixer.c @@ -1147,6 +1147,14 @@ static void volume_control_quirks(struct usb_mixer_elem_info *cval, } break; + case USB_ID(0x045e, 0x070f): /* MS LifeChat LX-3000 Headset */ + if (!strcmp(kctl->id.name, "Speaker Playback Volume")) { + usb_audio_info(chip, + "set volume quirk for MS LifeChat LX-3000\n"); + cval->res = 192; + } + break; + case USB_ID(0x0471, 0x0101): case USB_ID(0x0471, 0x0104): case USB_ID(0x0471, 0x0105): diff --git a/sound/usb/quirks.c b/sound/usb/quirks.c index 43793a5c3f51..dac469a8d07a 100644 --- a/sound/usb/quirks.c +++ b/sound/usb/quirks.c @@ -2153,6 +2153,8 @@ static const struct usb_audio_quirk_flags_table quirk_flags_table[] = { DEVICE_FLG(0x045e, 0x083c, /* MS USB Link headset */ QUIRK_FLAG_GET_SAMPLE_RATE | QUIRK_FLAG_CTL_MSG_DELAY | QUIRK_FLAG_DISABLE_AUTOSUSPEND), + DEVICE_FLG(0x045e, 0x070f, /* MS LifeChat LX-3000 Headset */ + QUIRK_FLAG_MIXER_PLAYBACK_MIN_MUTE), DEVICE_FLG(0x046d, 0x0807, /* Logitech Webcam C500 */ QUIRK_FLAG_CTL_MSG_DELAY_1M | QUIRK_FLAG_MIC_RES_384), DEVICE_FLG(0x046d, 0x0808, /* Logitech Webcam C600 */ From c6fceaf166479c05f7d3158ef08e78ae3e3dfa23 Mon Sep 17 00:00:00 2001 From: Pauli Virtanen Date: Thu, 16 Oct 2025 00:11:30 +0300 Subject: [PATCH 243/330] ALSA: usb-audio: fix vendor quirk for Logitech H390 Vendor quirk QUIRK_FLAG_CTL_MSG_DELAY_1M was inadvertently missing when adding quirk for Logitech H390. Add it back. Fixes: 2b929b6eec0c ("ALSA: usb-audio: add mixer_playback_min_mute quirk for Logitech H390") Signed-off-by: Pauli Virtanen Signed-off-by: Takashi Iwai --- sound/usb/quirks.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/usb/quirks.c b/sound/usb/quirks.c index dac469a8d07a..71638e6dfb20 100644 --- a/sound/usb/quirks.c +++ b/sound/usb/quirks.c @@ -2183,6 +2183,7 @@ static const struct usb_audio_quirk_flags_table quirk_flags_table[] = { DEVICE_FLG(0x046d, 0x09a4, /* Logitech QuickCam E 3500 */ QUIRK_FLAG_CTL_MSG_DELAY_1M | QUIRK_FLAG_IGNORE_CTL_ERROR), DEVICE_FLG(0x046d, 0x0a8f, /* Logitech H390 headset */ + QUIRK_FLAG_CTL_MSG_DELAY_1M | QUIRK_FLAG_MIXER_PLAYBACK_MIN_MUTE), DEVICE_FLG(0x0499, 0x1506, /* Yamaha THR5 */ QUIRK_FLAG_GENERIC_IMPLICIT_FB), From 5801e65206b065b0b2af032f7f1eef222aa2fd83 Mon Sep 17 00:00:00 2001 From: Tvrtko Ursulin Date: Wed, 15 Oct 2025 09:40:15 +0100 Subject: [PATCH 244/330] drm/sched: Fix potential double free in drm_sched_job_add_resv_dependencies MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When adding dependencies with drm_sched_job_add_dependency(), that function consumes the fence reference both on success and failure, so in the latter case the dma_fence_put() on the error path (xarray failed to expand) is a double free. Interestingly this bug appears to have been present ever since commit ebd5f74255b9 ("drm/sched: Add dependency tracking"), since the code back then looked like this: drm_sched_job_add_implicit_dependencies(): ... for (i = 0; i < fence_count; i++) { ret = drm_sched_job_add_dependency(job, fences[i]); if (ret) break; } for (; i < fence_count; i++) dma_fence_put(fences[i]); Which means for the failing 'i' the dma_fence_put was already a double free. Possibly there were no users at that time, or the test cases were insufficient to hit it. The bug was then only noticed and fixed after commit 9c2ba265352a ("drm/scheduler: use new iterator in drm_sched_job_add_implicit_dependencies v2") landed, with its fixup of commit 4eaf02d6076c ("drm/scheduler: fix drm_sched_job_add_implicit_dependencies"). At that point it was a slightly different flavour of a double free, which commit 963d0b356935 ("drm/scheduler: fix drm_sched_job_add_implicit_dependencies harder") noticed and attempted to fix. But it only moved the double free from happening inside the drm_sched_job_add_dependency(), when releasing the reference not yet obtained, to the caller, when releasing the reference already released by the former in the failure case. As such it is not easy to identify the right target for the fixes tag so lets keep it simple and just continue the chain. While fixing we also improve the comment and explain the reason for taking the reference and not dropping it. Signed-off-by: Tvrtko Ursulin Fixes: 963d0b356935 ("drm/scheduler: fix drm_sched_job_add_implicit_dependencies harder") Reported-by: Dan Carpenter Closes: https://lore.kernel.org/dri-devel/aNFbXq8OeYl3QSdm@stanley.mountain/ Cc: Christian König Cc: Rob Clark Cc: Daniel Vetter Cc: Matthew Brost Cc: Danilo Krummrich Cc: Philipp Stanner Cc: Christian König Cc: dri-devel@lists.freedesktop.org Cc: stable@vger.kernel.org # v5.16+ Signed-off-by: Philipp Stanner Link: https://lore.kernel.org/r/20251015084015.6273-1-tvrtko.ursulin@igalia.com --- drivers/gpu/drm/scheduler/sched_main.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 46119aacb809..c39f0245e3a9 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -965,13 +965,14 @@ int drm_sched_job_add_resv_dependencies(struct drm_sched_job *job, dma_resv_assert_held(resv); dma_resv_for_each_fence(&cursor, resv, usage, fence) { - /* Make sure to grab an additional ref on the added fence */ - dma_fence_get(fence); - ret = drm_sched_job_add_dependency(job, fence); - if (ret) { - dma_fence_put(fence); + /* + * As drm_sched_job_add_dependency always consumes the fence + * reference (even when it fails), and dma_resv_for_each_fence + * is not obtaining one, we need to grab one before calling. + */ + ret = drm_sched_job_add_dependency(job, dma_fence_get(fence)); + if (ret) return ret; - } } return 0; } From 86f54f9b6c17d6567c69e3a6fed52fdf5d7dbe93 Mon Sep 17 00:00:00 2001 From: Hao Ge Date: Wed, 15 Oct 2025 22:16:42 +0800 Subject: [PATCH 245/330] slab: reset slab->obj_ext when freeing and it is OBJEXTS_ALLOC_FAIL If obj_exts allocation failed, slab->obj_exts is set to OBJEXTS_ALLOC_FAIL, But we do not clear it when freeing the slab. Since OBJEXTS_ALLOC_FAIL and MEMCG_DATA_OBJEXTS currently share the same bit position, during the release of the associated folio, a VM_BUG_ON_FOLIO() check in folio_memcg_kmem() is triggered because the OBJEXTS_ALLOC_FAIL flag was not cleared, causing it to be interpreted as a kmem folio (non-slab) with MEMCG_OBJEXTS_DATA flag set, which is invalid because MEMCG_OBJEXTS_DATA is supposed to be set only on slabs. Another problem that predates sharing the OBJEXTS_ALLOC_FAIL and MEMCG_DATA_OBJEXTS bits is that on configurations with is_check_pages_enabled(), the non-cleared bit in page->memcg_data will trigger a free_page_is_bad() failure "page still charged to cgroup" When freeing a slab, we clear slab->obj_exts if the obj_ext array has been successfully allocated. So let's clear it also when the allocation has failed. Fixes: 09c46563ff6d ("codetag: debug: introduce OBJEXTS_ALLOC_FAIL to mark failed slab_ext allocations") Fixes: 7612833192d5 ("slab: Reuse first bit for OBJEXTS_ALLOC_FAIL") Link: https://lore.kernel.org/all/20251015141642.700170-1-hao.ge@linux.dev/ Cc: Signed-off-by: Hao Ge Reviewed-by: Suren Baghdasaryan Reviewed-by: Harry Yoo Signed-off-by: Vlastimil Babka --- mm/slub.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/mm/slub.c b/mm/slub.c index 13ae4491136a..a8fcc7e6f25a 100644 --- a/mm/slub.c +++ b/mm/slub.c @@ -2170,8 +2170,15 @@ static inline void free_slab_obj_exts(struct slab *slab) struct slabobj_ext *obj_exts; obj_exts = slab_obj_exts(slab); - if (!obj_exts) + if (!obj_exts) { + /* + * If obj_exts allocation failed, slab->obj_exts is set to + * OBJEXTS_ALLOC_FAIL. In this case, we end up here and should + * clear the flag. + */ + slab->obj_exts = 0; return; + } /* * obj_exts was created with __GFP_NO_OBJ_EXT flag, therefore its From ed80cc4667ac997b84546e6d35f0a0ae525d239c Mon Sep 17 00:00:00 2001 From: Stuart Hayhurst Date: Mon, 6 Oct 2025 02:05:49 +0100 Subject: [PATCH 246/330] HID: logitech-hidpp: Add HIDPP_QUIRK_RESET_HI_RES_SCROLL The Logitech G502 Hero Wireless's high resolution scrolling resets after being unplugged without notifying the driver, causing extremely slow scrolling. The only indication of this is a battery update packet, so add a quirk to detect when the device is unplugged and re-enable the scrolling. Link: https://bugzilla.kernel.org/show_bug.cgi?id=218037 Signed-off-by: Stuart Hayhurst Signed-off-by: Jiri Kosina --- drivers/hid/hid-logitech-hidpp.c | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) diff --git a/drivers/hid/hid-logitech-hidpp.c b/drivers/hid/hid-logitech-hidpp.c index aaef405a717e..5e763de4b94f 100644 --- a/drivers/hid/hid-logitech-hidpp.c +++ b/drivers/hid/hid-logitech-hidpp.c @@ -75,6 +75,7 @@ MODULE_PARM_DESC(disable_tap_to_click, #define HIDPP_QUIRK_HIDPP_CONSUMER_VENDOR_KEYS BIT(27) #define HIDPP_QUIRK_HI_RES_SCROLL_1P0 BIT(28) #define HIDPP_QUIRK_WIRELESS_STATUS BIT(29) +#define HIDPP_QUIRK_RESET_HI_RES_SCROLL BIT(30) /* These are just aliases for now */ #define HIDPP_QUIRK_KBD_SCROLL_WHEEL HIDPP_QUIRK_HIDPP_WHEELS @@ -193,6 +194,7 @@ struct hidpp_device { void *private_data; struct work_struct work; + struct work_struct reset_hi_res_work; struct kfifo delayed_work_fifo; struct input_dev *delayed_input; @@ -3836,6 +3838,7 @@ static int hidpp_raw_hidpp_event(struct hidpp_device *hidpp, u8 *data, struct hidpp_report *answer = hidpp->send_receive_buf; struct hidpp_report *report = (struct hidpp_report *)data; int ret; + int last_online; /* * If the mutex is locked then we have a pending answer from a @@ -3877,6 +3880,7 @@ static int hidpp_raw_hidpp_event(struct hidpp_device *hidpp, u8 *data, "See: https://gitlab.freedesktop.org/jwrdegoede/logitech-27mhz-keyboard-encryption-setup/\n"); } + last_online = hidpp->battery.online; if (hidpp->capabilities & HIDPP_CAPABILITY_HIDPP20_BATTERY) { ret = hidpp20_battery_event_1000(hidpp, data, size); if (ret != 0) @@ -3901,6 +3905,11 @@ static int hidpp_raw_hidpp_event(struct hidpp_device *hidpp, u8 *data, return ret; } + if (hidpp->quirks & HIDPP_QUIRK_RESET_HI_RES_SCROLL) { + if (last_online == 0 && hidpp->battery.online == 1) + schedule_work(&hidpp->reset_hi_res_work); + } + if (hidpp->quirks & HIDPP_QUIRK_HIDPP_WHEELS) { ret = hidpp10_wheel_raw_event(hidpp, data, size); if (ret != 0) @@ -4274,6 +4283,13 @@ static void hidpp_connect_event(struct work_struct *work) hidpp->delayed_input = input; } +static void hidpp_reset_hi_res_handler(struct work_struct *work) +{ + struct hidpp_device *hidpp = container_of(work, struct hidpp_device, reset_hi_res_work); + + hi_res_scroll_enable(hidpp); +} + static DEVICE_ATTR(builtin_power_supply, 0000, NULL, NULL); static struct attribute *sysfs_attrs[] = { @@ -4404,6 +4420,7 @@ static int hidpp_probe(struct hid_device *hdev, const struct hid_device_id *id) } INIT_WORK(&hidpp->work, hidpp_connect_event); + INIT_WORK(&hidpp->reset_hi_res_work, hidpp_reset_hi_res_handler); mutex_init(&hidpp->send_mutex); init_waitqueue_head(&hidpp->wait); @@ -4499,6 +4516,7 @@ static void hidpp_remove(struct hid_device *hdev) hid_hw_stop(hdev); cancel_work_sync(&hidpp->work); + cancel_work_sync(&hidpp->reset_hi_res_work); mutex_destroy(&hidpp->send_mutex); } @@ -4546,6 +4564,9 @@ static const struct hid_device_id hidpp_devices[] = { { /* Keyboard MX5500 (Bluetooth-receiver in HID proxy mode) */ LDJ_DEVICE(0xb30b), .driver_data = HIDPP_QUIRK_HIDPP_CONSUMER_VENDOR_KEYS }, + { /* Logitech G502 Lightspeed Wireless Gaming Mouse */ + LDJ_DEVICE(0x407f), + .driver_data = HIDPP_QUIRK_RESET_HI_RES_SCROLL }, { LDJ_DEVICE(HID_ANY_ID) }, From 0c1999ed33722f85476a248186d6e0eb2bf3dd2a Mon Sep 17 00:00:00 2001 From: Xing Guo Date: Thu, 16 Oct 2025 11:53:30 +0800 Subject: [PATCH 247/330] selftests: arg_parsing: Ensure data is flushed to disk before reading. test_parse_test_list_file writes some data to /tmp/bpf_arg_parsing_test.XXXXXX and parse_test_list_file() will read the data back. However, after writing data to that file, we forget to call fsync() and it's causing testing failure in my laptop. This patch helps fix it by adding the missing fsync() call. Fixes: 64276f01dce8 ("selftests/bpf: Test_progs can read test lists from file") Signed-off-by: Xing Guo Signed-off-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/20251016035330.3217145-1-higuoxing@gmail.com --- tools/testing/selftests/bpf/prog_tests/arg_parsing.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/tools/testing/selftests/bpf/prog_tests/arg_parsing.c b/tools/testing/selftests/bpf/prog_tests/arg_parsing.c index fbf0d9c2f58b..e27d66b75fb1 100644 --- a/tools/testing/selftests/bpf/prog_tests/arg_parsing.c +++ b/tools/testing/selftests/bpf/prog_tests/arg_parsing.c @@ -144,6 +144,9 @@ static void test_parse_test_list_file(void) if (!ASSERT_OK(ferror(fp), "prepare tmp")) goto out_fclose; + if (!ASSERT_OK(fsync(fileno(fp)), "fsync tmp")) + goto out_fclose; + init_test_filter_set(&set); if (!ASSERT_OK(parse_test_list_file(tmpfile, &set, true), "parse file")) From f6fddc6df3fc0cffce329b87927db4eb5989728d Mon Sep 17 00:00:00 2001 From: Shardul Bankar Date: Thu, 16 Oct 2025 12:03:30 +0530 Subject: [PATCH 248/330] bpf: Fix memory leak in __lookup_instance error path When __lookup_instance() allocates a func_instance structure but fails to allocate the must_write_set array, it returns an error without freeing the previously allocated func_instance. This causes a memory leak of 192 bytes (sizeof(struct func_instance)) each time this error path is triggered. Fix by freeing 'result' on must_write_set allocation failure. Fixes: b3698c356ad9 ("bpf: callchain sensitive stack liveness tracking using CFG") Reported-by: BPF Runtime Fuzzer (BRF) Signed-off-by: Shardul Bankar Signed-off-by: Martin KaFai Lau Acked-by: Eduard Zingerman Link: https://patch.msgid.link/20251016063330.4107547-1-shardulsb08@gmail.com --- kernel/bpf/liveness.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/kernel/bpf/liveness.c b/kernel/bpf/liveness.c index 3c611aba7f52..1e6538f59a78 100644 --- a/kernel/bpf/liveness.c +++ b/kernel/bpf/liveness.c @@ -195,8 +195,10 @@ static struct func_instance *__lookup_instance(struct bpf_verifier_env *env, return ERR_PTR(-ENOMEM); result->must_write_set = kvcalloc(subprog_sz, sizeof(*result->must_write_set), GFP_KERNEL_ACCOUNT); - if (!result->must_write_set) + if (!result->must_write_set) { + kvfree(result); return ERR_PTR(-ENOMEM); + } memcpy(&result->callchain, callchain, sizeof(*callchain)); result->insn_cnt = subprog_sz; hash_add(liveness->func_instances, &result->hl_node, key); From 5a869d017793399fd1d2609ff27e900534173eb3 Mon Sep 17 00:00:00 2001 From: Wilfred Mallawa Date: Fri, 10 Oct 2025 17:19:42 +1000 Subject: [PATCH 249/330] nvme/tcp: handle tls partially sent records in write_space() With TLS enabled, records that are encrypted and appended to TLS TX list can fail to see a retry if the underlying TCP socket is busy, for example, hitting an EAGAIN from tcp_sendmsg_locked(). This is not known to the NVMe TCP driver, as the TLS layer successfully generated a record. Typically, the TLS write_space() callback would ensure such records are retried, but in the NVMe TCP Host driver, write_space() invokes nvme_tcp_write_space(). This causes a partially sent record in the TLS TX list to timeout after not being retried. This patch fixes the above by calling queue->write_space(), which calls into the TLS layer to retry any pending records. Fixes: be8e82caa685 ("nvme-tcp: enable TLS handshake upcall") Signed-off-by: Wilfred Mallawa Reviewed-by: Hannes Reinecke Signed-off-by: Keith Busch --- drivers/nvme/host/tcp.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c index 1413788ca7d5..9a96df1a511c 100644 --- a/drivers/nvme/host/tcp.c +++ b/drivers/nvme/host/tcp.c @@ -1081,6 +1081,9 @@ static void nvme_tcp_write_space(struct sock *sk) queue = sk->sk_user_data; if (likely(queue && sk_stream_is_writeable(sk))) { clear_bit(SOCK_NOSPACE, &sk->sk_socket->flags); + /* Ensure pending TLS partial records are retried */ + if (nvme_tcp_queue_tls(queue)) + queue->write_space(sk); queue_work_on(queue->io_cpu, nvme_tcp_wq, &queue->io_work); } read_unlock_bh(&sk->sk_callback_lock); From 75cea9860aa6b2350d90a8d78fed114d27c7eca2 Mon Sep 17 00:00:00 2001 From: Michal Pecio Date: Tue, 14 Oct 2025 20:35:28 +0200 Subject: [PATCH 250/330] net: usb: rtl8150: Fix frame padding TX frames aren't padded and unknown memory is sent into the ether. Theoretically, it isn't even guaranteed that the extra memory exists and can be sent out, which could cause further problems. In practice, I found that plenty of tailroom exists in the skb itself (in my test with ping at least) and skb_padto() easily succeeds, so use it here. In the event of -ENOMEM drop the frame like other drivers do. The use of one more padding byte instead of a USB zero-length packet is retained to avoid regression. I have a dodgy Etron xHCI controller which doesn't seem to support sending ZLPs at all. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Signed-off-by: Michal Pecio Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251014203528.3f9783c4.michal.pecio@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/usb/rtl8150.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/drivers/net/usb/rtl8150.c b/drivers/net/usb/rtl8150.c index 92add3daadbb..278e6cb6f4d9 100644 --- a/drivers/net/usb/rtl8150.c +++ b/drivers/net/usb/rtl8150.c @@ -685,9 +685,16 @@ static netdev_tx_t rtl8150_start_xmit(struct sk_buff *skb, rtl8150_t *dev = netdev_priv(netdev); int count, res; + /* pad the frame and ensure terminating USB packet, datasheet 9.2.3 */ + count = max(skb->len, ETH_ZLEN); + if (count % 64 == 0) + count++; + if (skb_padto(skb, count)) { + netdev->stats.tx_dropped++; + return NETDEV_TX_OK; + } + netif_stop_queue(netdev); - count = (skb->len < 60) ? 60 : skb->len; - count = (count & 0x3f) ? count : count + 1; dev->tx_skb = skb; usb_fill_bulk_urb(dev->tx_urb, dev->udev, usb_sndbulkpipe(dev->udev, 2), skb->data, count, write_bulk_callback, dev); From aaf043a5688114703ae2c1482b92e7e0754d684e Mon Sep 17 00:00:00 2001 From: Nathan Chancellor Date: Tue, 14 Oct 2025 13:46:49 -0700 Subject: [PATCH 251/330] net/mlx5e: Return 1 instead of 0 in invalid case in mlx5e_mpwrq_umr_entry_size() When building with Clang 20 or newer, there are some objtool warnings from unexpected fallthroughs to other functions: vmlinux.o: warning: objtool: mlx5e_mpwrq_mtts_per_wqe() falls through to next function mlx5e_mpwrq_max_num_entries() vmlinux.o: warning: objtool: mlx5e_mpwrq_max_log_rq_size() falls through to next function mlx5e_get_linear_rq_headroom() LLVM 20 contains an (admittedly problematic [1]) optimization [2] to convert divide by zero into the equivalent of __builtin_unreachable(), which invokes undefined behavior and destroys code generation when it is encountered in a control flow graph. mlx5e_mpwrq_umr_entry_size() returns 0 in the default case of an unrecognized mlx5e_mpwrq_umr_mode value. mlx5e_mpwrq_mtts_per_wqe(), which is inlined into mlx5e_mpwrq_max_log_rq_size(), uses the result of mlx5e_mpwrq_umr_entry_size() in a divide operation without checking for zero, so LLVM is able to infer there will be a divide by zero in this case and invokes undefined behavior. While there is some proposed work to isolate this undefined behavior and avoid the destructive code generation that results in these objtool warnings, code should still be defensive against divide by zero. As the WARN_ONCE() implies that an invalid value should be handled gracefully, return 1 instead of 0 in the default case so that the results of this division operation is always valid. Fixes: 168723c1f8d6 ("net/mlx5e: xsk: Use umr_mode to calculate striding RQ parameters") Link: https://lore.kernel.org/CAGG=3QUk8-Ak7YKnRziO4=0z=1C_7+4jF+6ZeDQ9yF+kuTOHOQ@mail.gmail.com/ [1] Link: https://github.com/llvm/llvm-project/commit/37932643abab699e8bb1def08b7eb4eae7ff1448 [2] Closes: https://github.com/ClangBuiltLinux/linux/issues/2131 Closes: https://github.com/ClangBuiltLinux/linux/issues/2132 Signed-off-by: Nathan Chancellor Reviewed-by: Tariq Toukan Link: https://patch.msgid.link/20251014-mlx5e-avoid-zero-div-from-mlx5e_mpwrq_umr_entry_size-v1-1-dc186b8819ef@kernel.org Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/en/params.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c index 3692298e10f2..c9bdee9a8b30 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c @@ -100,7 +100,7 @@ u8 mlx5e_mpwrq_umr_entry_size(enum mlx5e_mpwrq_umr_mode mode) return sizeof(struct mlx5_ksm) * 4; } WARN_ONCE(1, "MPWRQ UMR mode %d is not known\n", mode); - return 0; + return 1; } u8 mlx5e_mpwrq_log_wqe_sz(struct mlx5_core_dev *mdev, u8 page_shift, From d0d3e9c2867b32c9c70e39e74b9425871cf0042a Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 15 Oct 2025 06:32:21 +0000 Subject: [PATCH 252/330] net: gro: clear skb_shinfo(skb)->hwtstamps in napi_reuse_skb() Some network drivers assume this field is zero after napi_get_frags(). We must clear it in napi_reuse_skb() otherwise the following can happen: 1) A packet is received, and skb_shinfo(skb)->hwtstamps is populated because a bit in the receive descriptor announced hwtstamp availability for this packet. 2) Packet is given to gro layer via napi_gro_frags(). 3) Packet is merged to a prior one held in GRO queues. 4) skb is saved after some cleanup in napi->skb via a call to napi_reuse_skb(). 5) Next packet is received 10 seconds later, gets the recycled skb from napi_get_frags(). 6) The receive descriptor does not announce hwtstamp availability. Driver does not clear shinfo->hwtstamps. 7) We have in shinfo->hwtstamps an old timestamp. Fixes: ac45f602ee3d ("net: infrastructure for hardware time stamping") Signed-off-by: Eric Dumazet Reviewed-by: Alexander Lobakin Link: https://patch.msgid.link/20251015063221.4171986-1-edumazet@google.com Signed-off-by: Jakub Kicinski --- net/core/gro.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/net/core/gro.c b/net/core/gro.c index 5ba4504cfd28..76f9c3712422 100644 --- a/net/core/gro.c +++ b/net/core/gro.c @@ -639,6 +639,8 @@ EXPORT_SYMBOL(gro_receive_skb); static void napi_reuse_skb(struct napi_struct *napi, struct sk_buff *skb) { + struct skb_shared_info *shinfo; + if (unlikely(skb->pfmemalloc)) { consume_skb(skb); return; @@ -655,8 +657,12 @@ static void napi_reuse_skb(struct napi_struct *napi, struct sk_buff *skb) skb->encapsulation = 0; skb->ip_summed = CHECKSUM_NONE; - skb_shinfo(skb)->gso_type = 0; - skb_shinfo(skb)->gso_size = 0; + + shinfo = skb_shinfo(skb); + shinfo->gso_type = 0; + shinfo->gso_size = 0; + shinfo->hwtstamps.hwtstamp = 0; + if (unlikely(skb->slow_gro)) { skb_orphan(skb); skb_ext_reset(skb); From 5348d6312446929edabced02bd6438bfe5220e31 Mon Sep 17 00:00:00 2001 From: Cosmin Ratiu Date: Wed, 15 Oct 2025 10:05:23 +0300 Subject: [PATCH 253/330] net/mlx5e: psp, avoid 'accel' NULL pointer dereference The 'accel' parameter of mlx5e_txwqe_build_eseg_csum() and the similar 'state' parameter of mlx5e_accel_tx_ids_len() were NULL when called from mlx5i_sq_xmit() and were causing kernel panics from that context. Fix that by passing in a local empty mlx5e_accel_tx_state variable, thus guaranteeing that 'accel' is never NULL. Also remove an unnecessary check from mlx5e_tx_wqe_inline_mode(). Fixes: e5a1861a298e ("net/mlx5e: Implement PSP Tx data path") Signed-off-by: Cosmin Ratiu Reviewed-by: Dragos Tatulea Signed-off-by: Tariq Toukan Reviewed-by: Michal Swiatkowski Link: https://patch.msgid.link/1760511923-890650-1-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/en_tx.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c index b7227afcb51d..2702b3885f06 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c @@ -256,7 +256,7 @@ mlx5e_tx_wqe_inline_mode(struct mlx5e_txqsq *sq, struct sk_buff *skb, u8 mode; #ifdef CONFIG_MLX5_EN_TLS - if (accel && accel->tls.tls_tisn) + if (accel->tls.tls_tisn) return MLX5_INLINE_MODE_TCP_UDP; #endif @@ -982,6 +982,7 @@ void mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb, struct mlx5e_tx_attr attr; struct mlx5i_tx_wqe *wqe; + struct mlx5e_accel_tx_state accel = {}; struct mlx5_wqe_datagram_seg *datagram; struct mlx5_wqe_ctrl_seg *cseg; struct mlx5_wqe_eth_seg *eseg; @@ -992,7 +993,7 @@ void mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb, int num_dma; u16 pi; - mlx5e_sq_xmit_prepare(sq, skb, NULL, &attr); + mlx5e_sq_xmit_prepare(sq, skb, &accel, &attr); mlx5i_sq_calc_wqe_attr(skb, &attr, &wqe_attr); pi = mlx5e_txqsq_get_next_pi(sq, wqe_attr.num_wqebbs); @@ -1009,7 +1010,7 @@ void mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb, mlx5i_txwqe_build_datagram(av, dqpn, dqkey, datagram); - mlx5e_txwqe_build_eseg_csum(sq, skb, NULL, eseg); + mlx5e_txwqe_build_eseg_csum(sq, skb, &accel, eseg); eseg->mss = attr.mss; From 1b0124ad5039678a9dfafb6aafef6f430a246b91 Mon Sep 17 00:00:00 2001 From: Bagas Sanjaya Date: Wed, 15 Oct 2025 16:25:41 +0700 Subject: [PATCH 254/330] net: rmnet: Fix checksum offload header v5 and aggregation packet formatting Packet format for checksum offload header v5 and aggregation, and header type table for the former, are shown in normal paragraphs instead. Use appropriate markup. Signed-off-by: Bagas Sanjaya Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251015092540.32282-2-bagasdotme@gmail.com Signed-off-by: Jakub Kicinski --- .../device_drivers/cellular/qualcomm/rmnet.rst | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) diff --git a/Documentation/networking/device_drivers/cellular/qualcomm/rmnet.rst b/Documentation/networking/device_drivers/cellular/qualcomm/rmnet.rst index 289c146a8291..6877a3260582 100644 --- a/Documentation/networking/device_drivers/cellular/qualcomm/rmnet.rst +++ b/Documentation/networking/device_drivers/cellular/qualcomm/rmnet.rst @@ -137,16 +137,20 @@ d. Checksum offload header v5 Checksum offload header fields are in big endian format. +Packet format:: + Bit 0 - 6 7 8-15 16-31 Function Header Type Next Header Checksum Valid Reserved Header Type is to indicate the type of header, this usually is set to CHECKSUM Header types -= ========================================== + += =============== 0 Reserved 1 Reserved 2 checksum header += =============== Checksum Valid is to indicate whether the header checksum is valid. Value of 1 implies that checksum is calculated on this packet and is valid, value of 0 @@ -183,9 +187,11 @@ rmnet in a single linear skb. rmnet will process the individual packets and either ACK the MAP command or deliver the IP packet to the network stack as needed -MAP header|IP Packet|Optional padding|MAP header|IP Packet|Optional padding.... +Packet format:: -MAP header|IP Packet|Optional padding|MAP header|Command Packet|Optional pad... + MAP header|IP Packet|Optional padding|MAP header|IP Packet|Optional padding.... + + MAP header|IP Packet|Optional padding|MAP header|Command Packet|Optional pad... 3. Userspace configuration ========================== From 4a9cb2eecc78fa9d388481762dd798fa770e1971 Mon Sep 17 00:00:00 2001 From: Miguel Ojeda Date: Fri, 10 Oct 2025 19:43:49 +0200 Subject: [PATCH 255/330] docs: rust: add section on imports formatting `rustfmt`, by default, formats imports in a way that is prone to conflicts while merging and rebasing, since in some cases it condenses several items into the same line. For instance, Linus mentioned [1] that the following case: use crate::{ fmt, page::AsPageIter, }; is compressed by `rustfmt` into: use crate::{fmt, page::AsPageIter}; which is undesirable. Similarly, `rustfmt` may put several items in the same line even if the braces span already multiple lines, e.g.: use kernel::{ acpi, c_str, device::{property, Core}, of, platform, }; The options that control the formatting behavior around imports are generally unstable, and `rustfmt` releases do not allow to use nightly features, unlike the compiler and other Rust tooling [2]. For the moment, we can introduce a workaround to prevent `rustfmt` from compressing the example above -- the "trailing empty comment": use crate::{ fmt, page::AsPageIter, // }; which is reminiscent of the trailing comma behavior in other formatters. We already used empty comments for formatting purposes in the past, e.g. in commit b9b701fce49a ("rust: clarify the language unstable features in use"). In addition, `rustfmt` actually reformats with a vertical layout (i.e. it does not put two items in the same line) when seeing such a comment, i.e. it doesn't just preserve the formatting, which is good in the sense that we can use it to easily reformat some imports, since it matches the style we generally want to have. A Git merge driver would help (suggested by Gary and Wedson), though maintainers would need to set it up, the diffs would still be larger and the formatting rules for imports would remain hard to predict. Thus document the style that we will follow in the coding guidelines by introducing a new section and explain how the trailing empty comment works there too. We discussed the issue with upstream Rust in our usual Rust <-> Rust for Linux meeting [3], and there have also been a few other discussions in parallel in issues [4][5] and Zulip [6]. We will see what happens, but upstream Rust has already created a subteam of `rustfmt` to try to overcome the bandwidth issue [7], which is a good signal, and some organization work has already started (e.g. tracking issues). We will continue our discussions with them about it. Cc: Caleb Cartwright Cc: Yacin Tmimi Cc: Manish Goregaokar Cc: Deadbeef Cc: Cameron Steffen Cc: Jieyou Xu Link: https://lore.kernel.org/all/CAHk-=wgO7S_FZUSBbngG5vtejWOpzDfTTBkVvP3_yjJmFddbzA@mail.gmail.com/ [1] Link: https://github.com/rust-lang/rustfmt/issues/4884 [2] Link: https://hackmd.io/iSCyY3JTTz-g8YM-nnzTTA [3] Link: https://github.com/rust-lang/rustfmt/issues/4991 [4] Link: https://github.com/rust-lang/rustfmt/issues/3361 [5] Link: https://rust-lang.zulipchat.com/#narrow/channel/392734-council/topic/rustfmt.20maintenance/near/543815381 [6] Link: https://github.com/rust-lang/team/pull/2017 [7] Reviewed-by: Benno Lossin Signed-off-by: Miguel Ojeda --- Documentation/rust/coding-guidelines.rst | 75 ++++++++++++++++++++++++ 1 file changed, 75 insertions(+) diff --git a/Documentation/rust/coding-guidelines.rst b/Documentation/rust/coding-guidelines.rst index 6ff9e754755d..3198be3a6d63 100644 --- a/Documentation/rust/coding-guidelines.rst +++ b/Documentation/rust/coding-guidelines.rst @@ -38,6 +38,81 @@ Like ``clang-format`` for the rest of the kernel, ``rustfmt`` works on individual files, and does not require a kernel configuration. Sometimes it may even work with broken code. +Imports +~~~~~~~ + +``rustfmt``, by default, formats imports in a way that is prone to conflicts +while merging and rebasing, since in some cases it condenses several items into +the same line. For instance: + +.. code-block:: rust + + // Do not use this style. + use crate::{ + example1, + example2::{example3, example4, example5}, + example6, example7, + example8::example9, + }; + +Instead, the kernel uses a vertical layout that looks like this: + +.. code-block:: rust + + use crate::{ + example1, + example2::{ + example3, + example4, + example5, // + }, + example6, + example7, + example8::example9, // + }; + +That is, each item goes into its own line, and braces are used as soon as there +is more than one item in a list. + +The trailing empty comment allows to preserve this formatting. Not only that, +``rustfmt`` will actually reformat imports vertically when the empty comment is +added. That is, it is possible to easily reformat the original example into the +expected style by running ``rustfmt`` on an input like: + +.. code-block:: rust + + // Do not use this style. + use crate::{ + example1, + example2::{example3, example4, example5, // + }, + example6, example7, + example8::example9, // + }; + +The trailing empty comment works for nested imports, as shown above, as well as +for single item imports -- this can be useful to minimize diffs within patch +series: + +.. code-block:: rust + + use crate::{ + example1, // + }; + +The trailing empty comment works in any of the lines within the braces, but it +is preferred to keep it in the last item, since it is reminiscent of the +trailing comma in other formatters. Sometimes it may be simpler to avoid moving +the comment several times within a patch series due to changes in the list. + +There may be cases where exceptions may need to be made, i.e. none of this is +a hard rule. There is also code that is not migrated to this style yet, but +please do not introduce code in other styles. + +Eventually, the goal is to get ``rustfmt`` to support this formatting style (or +a similar one) automatically in a stable release without requiring the trailing +empty comment. Thus, at some point, the goal is to remove those comments. + Comments -------- From 8a7c601e14576a22c2bbf7f67455ccf3f3d2737f Mon Sep 17 00:00:00 2001 From: Miguel Ojeda Date: Fri, 10 Oct 2025 19:43:50 +0200 Subject: [PATCH 256/330] rust: alloc: employ a trailing comment to keep vertical layout Apply the formatting guidelines introduced in the previous commit to make the file `rustfmt`-clean again. Reviewed-by: Benno Lossin Signed-off-by: Miguel Ojeda --- rust/kernel/alloc/kvec.rs | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/rust/kernel/alloc/kvec.rs b/rust/kernel/alloc/kvec.rs index e94aebd084c8..ac8d6f763ae8 100644 --- a/rust/kernel/alloc/kvec.rs +++ b/rust/kernel/alloc/kvec.rs @@ -9,7 +9,7 @@ }; use crate::{ fmt, - page::AsPageIter, + page::AsPageIter, // }; use core::{ borrow::{Borrow, BorrowMut}, From 32f072d9eaf9c31c2b0527a4a3370570a731e3cc Mon Sep 17 00:00:00 2001 From: Miguel Ojeda Date: Fri, 10 Oct 2025 19:43:51 +0200 Subject: [PATCH 257/330] rust: cpufreq: fix formatting We do our best to keep the repository `rustfmt`-clean, thus run the tool to fix the formatting issue. Link: https://docs.kernel.org/rust/coding-guidelines.html#style-formatting Link: https://rust-for-linux.com/contributing#submit-checklist-addendum Fixes: f97aef092e19 ("cpufreq: Make drivers using CPUFREQ_ETERNAL specify transition latency") Acked-by: Viresh Kumar Reviewed-by: Benno Lossin Signed-off-by: Miguel Ojeda --- rust/kernel/cpufreq.rs | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/rust/kernel/cpufreq.rs b/rust/kernel/cpufreq.rs index 21b5b9b8acc1..1a555fcb120a 100644 --- a/rust/kernel/cpufreq.rs +++ b/rust/kernel/cpufreq.rs @@ -38,8 +38,7 @@ const CPUFREQ_NAME_LEN: usize = bindings::CPUFREQ_NAME_LEN as usize; /// Default transition latency value in nanoseconds. -pub const DEFAULT_TRANSITION_LATENCY_NS: u32 = - bindings::CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; +pub const DEFAULT_TRANSITION_LATENCY_NS: u32 = bindings::CPUFREQ_DEFAULT_TRANSITION_LATENCY_NS; /// CPU frequency driver flags. pub mod flags { From bf29555f5bdc017bac22ca66fcb6c9f46ec8788f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Johannes=20Wiesb=C3=B6ck?= Date: Wed, 15 Oct 2025 22:15:43 +0200 Subject: [PATCH 258/330] rtnetlink: Allow deleting FDB entries in user namespace MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Creating FDB entries is possible from a non-initial user namespace when having CAP_NET_ADMIN, yet, when deleting FDB entries, processes receive an EPERM because the capability is always checked against the initial user namespace. This restricts the FDB management from unprivileged containers. Drop the netlink_capable check in rtnl_fdb_del as it was originally dropped in c5c351088ae7 and reintroduced in 1690be63a27b without intention. This patch was tested using a container on GyroidOS, where it was possible to delete FDB entries from an unprivileged user namespace and private network namespace. Fixes: 1690be63a27b ("bridge: Add vlan support to static neighbors") Reviewed-by: Michael Weiß Tested-by: Harshal Gohel Signed-off-by: Johannes Wiesböck Reviewed-by: Ido Schimmel Reviewed-by: Nikolay Aleksandrov Link: https://patch.msgid.link/20251015201548.319871-1-johannes.wiesboeck@aisec.fraunhofer.de Signed-off-by: Jakub Kicinski --- net/core/rtnetlink.c | 3 --- 1 file changed, 3 deletions(-) diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c index 8040ff7c356e..576d5ec3bb36 100644 --- a/net/core/rtnetlink.c +++ b/net/core/rtnetlink.c @@ -4715,9 +4715,6 @@ static int rtnl_fdb_del(struct sk_buff *skb, struct nlmsghdr *nlh, int err; u16 vid; - if (!netlink_capable(skb, CAP_NET_ADMIN)) - return -EPERM; - if (!del_bulk) { err = nlmsg_parse_deprecated(nlh, sizeof(*ndm), tb, NDA_MAX, NULL, extack); From 7f864458e9a6d2000b726d14b3d3a706ac92a3b0 Mon Sep 17 00:00:00 2001 From: Sebastian Reichel Date: Tue, 14 Oct 2025 17:49:34 +0200 Subject: [PATCH 259/330] net: stmmac: dwmac-rk: Fix disabling set_clock_selection On all platforms set_clock_selection() writes to a GRF register. This requires certain clocks running and thus should happen before the clocks are disabled. This has been noticed on RK3576 Sige5, which hangs during system suspend when trying to suspend the second network interface. Note, that suspending the first interface works, because the second device ensures that the necessary clocks for the GRF are enabled. Cc: stable@vger.kernel.org Fixes: 2f2b60a0ec28 ("net: ethernet: stmmac: dwmac-rk: Add gmac support for rk3588") Signed-off-by: Sebastian Reichel Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251014-rockchip-network-clock-fix-v1-1-c257b4afdf75@collabora.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c index 51ea0caf16c1..0786816e05f0 100644 --- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c +++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c @@ -1446,14 +1446,15 @@ static int gmac_clk_enable(struct rk_priv_data *bsp_priv, bool enable) } } else { if (bsp_priv->clk_enabled) { + if (bsp_priv->ops && bsp_priv->ops->set_clock_selection) { + bsp_priv->ops->set_clock_selection(bsp_priv, + bsp_priv->clock_input, false); + } + clk_bulk_disable_unprepare(bsp_priv->num_clks, bsp_priv->clks); clk_disable_unprepare(bsp_priv->clk_phy); - if (bsp_priv->ops && bsp_priv->ops->set_clock_selection) - bsp_priv->ops->set_clock_selection(bsp_priv, - bsp_priv->clock_input, false); - bsp_priv->clk_enabled = false; } } From a429b76114aaca3ef1aff4cd469dcf025431bd11 Mon Sep 17 00:00:00 2001 From: Gao Xiang Date: Sun, 12 Oct 2025 21:59:25 +0800 Subject: [PATCH 260/330] erofs: fix crafted invalid cases for encoded extents Robert recently reported two corrupted images that can cause system crashes, which are related to the new encoded extents introduced in Linux 6.15: - The first one [1] has plen != 0 (e.g. plen == 0x2000000) but (plen & Z_EROFS_EXTENT_PLEN_MASK) == 0. It is used to represent special extents such as sparse extents (!EROFS_MAP_MAPPED), but previously only plen == 0 was handled; - The second one [2] has pa 0xffffffffffdcffed and plen 0xb4000, then "cur [0xfffffffffffff000] += bvec.bv_len [0x1000]" in "} while ((cur += bvec.bv_len) < end);" wraps around, causing an out-of-bound access of pcl->compressed_bvecs[] in z_erofs_submit_queue(). EROFS only supports 48-bit physical block addresses (up to 1EiB for 4k blocks), so add a sanity check to enforce this. Fixes: 1d191b4ca51d ("erofs: implement encoded extent metadata") Reported-by: Robert Morris Closes: https://lore.kernel.org/r/75022.1759355830@localhost [1] Closes: https://lore.kernel.org/r/80524.1760131149@localhost [2] Reviewed-by: Hongbo Li Signed-off-by: Gao Xiang --- fs/erofs/zmap.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c index e5581dbeb4c2..8007814f721e 100644 --- a/fs/erofs/zmap.c +++ b/fs/erofs/zmap.c @@ -596,7 +596,7 @@ static int z_erofs_map_blocks_ext(struct inode *inode, vi->z_fragmentoff = map->m_plen; if (recsz > offsetof(struct z_erofs_extent, pstart_lo)) vi->z_fragmentoff |= map->m_pa << 32; - } else if (map->m_plen) { + } else if (map->m_plen & Z_EROFS_EXTENT_PLEN_MASK) { map->m_flags |= EROFS_MAP_MAPPED | EROFS_MAP_FULL_MAPPED | EROFS_MAP_ENCODED; fmt = map->m_plen >> Z_EROFS_EXTENT_PLEN_FMT_BIT; @@ -715,6 +715,7 @@ static int z_erofs_map_sanity_check(struct inode *inode, struct erofs_map_blocks *map) { struct erofs_sb_info *sbi = EROFS_I_SB(inode); + u64 pend; if (!(map->m_flags & EROFS_MAP_ENCODED)) return 0; @@ -732,6 +733,10 @@ static int z_erofs_map_sanity_check(struct inode *inode, if (unlikely(map->m_plen > Z_EROFS_PCLUSTER_MAX_SIZE || map->m_llen > Z_EROFS_PCLUSTER_MAX_DSIZE)) return -EOPNOTSUPP; + /* Filesystems beyond 48-bit physical block addresses are invalid */ + if (unlikely(check_add_overflow(map->m_pa, map->m_plen, &pend) || + (pend >> sbi->blkszbits) >= BIT_ULL(48))) + return -EFSCORRUPTED; return 0; } From 1f1d3e1d094db732d22b892227bf1e1ac3a8ca04 Mon Sep 17 00:00:00 2001 From: Miguel Ojeda Date: Fri, 17 Oct 2025 00:57:54 +0200 Subject: [PATCH 261/330] rust: bitmap: fix formatting We do our best to keep the repository `rustfmt`-clean, thus run the tool to fix the formatting issue. Link: https://docs.kernel.org/rust/coding-guidelines.html#style-formatting Link: https://rust-for-linux.com/contributing#submit-checklist-addendum Fixes: 0f5878834d6c ("rust: bitmap: clean Rust 1.92.0 `unused_unsafe` warning") Reviewed-by: Burak Emir Signed-off-by: Miguel Ojeda --- rust/kernel/bitmap.rs | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/rust/kernel/bitmap.rs b/rust/kernel/bitmap.rs index 711b8368b38f..aa8fc7bf06fc 100644 --- a/rust/kernel/bitmap.rs +++ b/rust/kernel/bitmap.rs @@ -167,7 +167,9 @@ fn deref(&self) -> &Bitmap { let ptr = if self.nbits <= BITS_PER_LONG { // SAFETY: Bitmap is represented inline. #[allow(unused_unsafe, reason = "Safe since Rust 1.92.0")] - unsafe { core::ptr::addr_of!(self.repr.bitmap) } + unsafe { + core::ptr::addr_of!(self.repr.bitmap) + } } else { // SAFETY: Bitmap is represented as array of `unsigned long`. unsafe { self.repr.ptr.as_ptr() } @@ -184,7 +186,9 @@ fn deref_mut(&mut self) -> &mut Bitmap { let ptr = if self.nbits <= BITS_PER_LONG { // SAFETY: Bitmap is represented inline. #[allow(unused_unsafe, reason = "Safe since Rust 1.92.0")] - unsafe { core::ptr::addr_of_mut!(self.repr.bitmap) } + unsafe { + core::ptr::addr_of_mut!(self.repr.bitmap) + } } else { // SAFETY: Bitmap is represented as array of `unsigned long`. unsafe { self.repr.ptr.as_ptr() } From e433110eb5bf067f74d3d15c5fb252206c66ae0b Mon Sep 17 00:00:00 2001 From: Inochi Amaoto Date: Tue, 14 Oct 2025 09:46:07 +0800 Subject: [PATCH 262/330] PCI: vmd: Override irq_startup()/irq_shutdown() in vmd_init_dev_msi_info() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Since commit 54f45a30c0d0 ("PCI/MSI: Add startup/shutdown for per device domains") set callback irq_startup() and irq_shutdown() of the struct pci_msi[x]_template, __irq_startup() will always invokes irq_startup() callback instead of irq_enable() callback overridden in vmd_init_dev_msi_info(). This will not start the IRQ correctly. Also override irq_startup()/irq_shutdown() in vmd_init_dev_msi_info(), so the irq_startup() can invoke the real logic. Fixes: 54f45a30c0d0 ("PCI/MSI: Add startup/shutdown for per device domains") Reported-by: Kenneth Crudup Closes: https://lore.kernel.org/r/8a923590-5b3a-406f-a324-7bd1cf894d8f@panix.com/ Reported-by: Genes Lists Closes: https://lore.kernel.org/r/4b392af8847cc19720ffcd53865f60ab3edc56b3.camel@sapience.com Reported-by: Todd Brandt Closes: https://bugzilla.kernel.org/show_bug.cgi?id=220658 Reported-by: Oliver Hartkopp Closes: https://lore.kernel.org/r/8d6887a5-60bc-423c-8f7a-87b4ab739f6a@hartkopp.net Reported-by: Hervé Signed-off-by: Inochi Amaoto Signed-off-by: Bjorn Helgaas Tested-by: Kenneth R. Crudup Tested-by: Genes Lists Tested-by: Oliver Hartkopp Tested-by: Todd Brandt Tested-by: Hervé Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20251014014607.612586-1-inochiama@gmail.com --- drivers/pci/controller/vmd.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/drivers/pci/controller/vmd.c b/drivers/pci/controller/vmd.c index 1bd5bf4a6097..b4b62b9ccc45 100644 --- a/drivers/pci/controller/vmd.c +++ b/drivers/pci/controller/vmd.c @@ -192,6 +192,12 @@ static void vmd_pci_msi_enable(struct irq_data *data) data->chip->irq_unmask(data); } +static unsigned int vmd_pci_msi_startup(struct irq_data *data) +{ + vmd_pci_msi_enable(data); + return 0; +} + static void vmd_irq_disable(struct irq_data *data) { struct vmd_irq *vmdirq = data->chip_data; @@ -210,6 +216,11 @@ static void vmd_pci_msi_disable(struct irq_data *data) vmd_irq_disable(data->parent_data); } +static void vmd_pci_msi_shutdown(struct irq_data *data) +{ + vmd_pci_msi_disable(data); +} + static struct irq_chip vmd_msi_controller = { .name = "VMD-MSI", .irq_compose_msi_msg = vmd_compose_msi_msg, @@ -309,6 +320,8 @@ static bool vmd_init_dev_msi_info(struct device *dev, struct irq_domain *domain, if (!msi_lib_init_dev_msi_info(dev, domain, real_parent, info)) return false; + info->chip->irq_startup = vmd_pci_msi_startup; + info->chip->irq_shutdown = vmd_pci_msi_shutdown; info->chip->irq_enable = vmd_pci_msi_enable; info->chip->irq_disable = vmd_pci_msi_disable; return true; From a78835b86a4414230e4cf9a9f16d22302cdb8388 Mon Sep 17 00:00:00 2001 From: "Mario Limonciello (AMD)" Date: Mon, 13 Oct 2025 17:08:26 -0500 Subject: [PATCH 263/330] PCI/VGA: Select SCREEN_INFO on X86 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit commit 337bf13aa9dda ("PCI/VGA: Replace vga_is_firmware_default() with a screen info check") introduced an implicit dependency upon SCREEN_INFO by removing the open coded implementation. If a user didn't have CONFIG_SCREEN_INFO set, vga_is_firmware_default() would now return false. SCREEN_INFO is only used on X86 so add a conditional select for SCREEN_INFO to ensure that the VGA arbiter works as intended. Fixes: 337bf13aa9dda ("PCI/VGA: Replace vga_is_firmware_default() with a screen info check") Reported-by: Eric Biggers Closes: https://lore.kernel.org/linux-pci/20251012182302.GA3412@sol/ Suggested-by: Thomas Zimmermann Signed-off-by: Mario Limonciello (AMD) Signed-off-by: Bjorn Helgaas Reviewed-by: Thomas Zimmermann Reviewed-by: Ilpo Järvinen Tested-by: Eric Biggers Link: https://patch.msgid.link/20251013220829.1536292-1-superm1@kernel.org --- drivers/pci/Kconfig | 1 + drivers/pci/vgaarb.c | 6 ++---- 2 files changed, 3 insertions(+), 4 deletions(-) diff --git a/drivers/pci/Kconfig b/drivers/pci/Kconfig index 7065a8e5f9b1..f94f5d384362 100644 --- a/drivers/pci/Kconfig +++ b/drivers/pci/Kconfig @@ -306,6 +306,7 @@ config VGA_ARB bool "VGA Arbitration" if EXPERT default y depends on (PCI && !S390) + select SCREEN_INFO if X86 help Some "legacy" VGA devices implemented on PCI typically have the same hard-decoded addresses as they did on ISA. When multiple PCI devices diff --git a/drivers/pci/vgaarb.c b/drivers/pci/vgaarb.c index b58f94ee4891..436fa7f4c387 100644 --- a/drivers/pci/vgaarb.c +++ b/drivers/pci/vgaarb.c @@ -556,10 +556,8 @@ EXPORT_SYMBOL(vga_put); static bool vga_is_firmware_default(struct pci_dev *pdev) { -#ifdef CONFIG_SCREEN_INFO - struct screen_info *si = &screen_info; - - return pdev == screen_info_pci_dev(si); +#if defined CONFIG_X86 + return pdev == screen_info_pci_dev(&screen_info); #else return false; #endif From 2a786348004b34c5f61235d51c40c1c718b1f8f9 Mon Sep 17 00:00:00 2001 From: Dawn Gardner Date: Thu, 16 Oct 2025 15:42:06 -0300 Subject: [PATCH 264/330] ALSA: hda/realtek: Fix mute led for HP Omen 17-cb0xxx This laptop uses the ALC285 codec, fixed by enabling the ALC285_FIXUP_HP_MUTE_LED quirk Signed-off-by: Dawn Gardner Link: https://patch.msgid.link/20251016184218.31508-3-dawn.auroali@gmail.com Signed-off-by: Takashi Iwai --- sound/hda/codecs/realtek/alc269.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/hda/codecs/realtek/alc269.c b/sound/hda/codecs/realtek/alc269.c index 6b9e5c091bf3..8ad5febd822a 100644 --- a/sound/hda/codecs/realtek/alc269.c +++ b/sound/hda/codecs/realtek/alc269.c @@ -6397,6 +6397,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = { SND_PCI_QUIRK(0x103c, 0x854a, "HP EliteBook 830 G6", ALC285_FIXUP_HP_GPIO_LED), SND_PCI_QUIRK(0x103c, 0x85c6, "HP Pavilion x360 Convertible 14-dy1xxx", ALC295_FIXUP_HP_MUTE_LED_COEFBIT11), SND_PCI_QUIRK(0x103c, 0x85de, "HP Envy x360 13-ar0xxx", ALC285_FIXUP_HP_ENVY_X360), + SND_PCI_QUIRK(0x103c, 0x8603, "HP Omen 17-cb0xxx", ALC285_FIXUP_HP_MUTE_LED), SND_PCI_QUIRK(0x103c, 0x860c, "HP ZBook 17 G6", ALC285_FIXUP_HP_GPIO_AMP_INIT), SND_PCI_QUIRK(0x103c, 0x860f, "HP ZBook 15 G6", ALC285_FIXUP_HP_GPIO_AMP_INIT), SND_PCI_QUIRK(0x103c, 0x861f, "HP Elite Dragonfly G1", ALC285_FIXUP_HP_GPIO_AMP_INIT), From 0fbbcab7f9082cdc233da5e5e353f69830f11956 Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Fri, 17 Oct 2025 00:07:42 -0700 Subject: [PATCH 265/330] cgroup/misc: fix misc_res_type kernel-doc warning Format the kernel-doc for SCALE_HW_CALIB_INVALID correctly to avoid a kernel-doc warning: Warning: include/linux/misc_cgroup.h:26 Enum value 'MISC_CG_RES_TDX' not described in enum 'misc_res_type' Fixes: 7c035bea9407 ("KVM: TDX: Register TDX host key IDs to cgroup misc controller") Signed-off-by: Randy Dunlap Signed-off-by: Tejun Heo --- include/linux/misc_cgroup.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/include/linux/misc_cgroup.h b/include/linux/misc_cgroup.h index 71cf5bfc6349..0cb36a3ffc47 100644 --- a/include/linux/misc_cgroup.h +++ b/include/linux/misc_cgroup.h @@ -19,7 +19,7 @@ enum misc_res_type { MISC_CG_RES_SEV_ES, #endif #ifdef CONFIG_INTEL_TDX_HOST - /* Intel TDX HKIDs resource */ + /** @MISC_CG_RES_TDX: Intel TDX HKIDs resource */ MISC_CG_RES_TDX, #endif /** @MISC_CG_RES_TYPES: count of enum misc_res_type constants */ From 17679ac6df6c4830ba711835aa8cf961be36cfa1 Mon Sep 17 00:00:00 2001 From: Dewei Meng Date: Thu, 16 Oct 2025 14:10:11 +0800 Subject: [PATCH 266/330] btrfs: directly free partially initialized fs_info in btrfs_check_leaked_roots() If fs_info->super_copy or fs_info->super_for_commit allocated failed in btrfs_get_tree_subvol(), then no need to call btrfs_free_fs_info(). Otherwise btrfs_check_leaked_roots() would access NULL pointer because fs_info->allocated_roots had not been initialised. syzkaller reported the following information: ------------[ cut here ]------------ BUG: unable to handle page fault for address: fffffffffffffbb0 #PF: supervisor read access in kernel mode #PF: error_code(0x0000) - not-present page PGD 64c9067 P4D 64c9067 PUD 64cb067 PMD 0 Oops: Oops: 0000 [#1] SMP KASAN PTI CPU: 0 UID: 0 PID: 1402 Comm: syz.1.35 Not tainted 6.15.8 #4 PREEMPT(lazy) Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), (...) RIP: 0010:arch_atomic_read arch/x86/include/asm/atomic.h:23 [inline] RIP: 0010:raw_atomic_read include/linux/atomic/atomic-arch-fallback.h:457 [inline] RIP: 0010:atomic_read include/linux/atomic/atomic-instrumented.h:33 [inline] RIP: 0010:refcount_read include/linux/refcount.h:170 [inline] RIP: 0010:btrfs_check_leaked_roots+0x18f/0x2c0 fs/btrfs/disk-io.c:1230 [...] Call Trace: btrfs_free_fs_info+0x310/0x410 fs/btrfs/disk-io.c:1280 btrfs_get_tree_subvol+0x592/0x6b0 fs/btrfs/super.c:2029 btrfs_get_tree+0x63/0x80 fs/btrfs/super.c:2097 vfs_get_tree+0x98/0x320 fs/super.c:1759 do_new_mount+0x357/0x660 fs/namespace.c:3899 path_mount+0x716/0x19c0 fs/namespace.c:4226 do_mount fs/namespace.c:4239 [inline] __do_sys_mount fs/namespace.c:4450 [inline] __se_sys_mount fs/namespace.c:4427 [inline] __x64_sys_mount+0x28c/0x310 fs/namespace.c:4427 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0x92/0x180 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x76/0x7e RIP: 0033:0x7f032eaffa8d [...] Fixes: 3bb17a25bcb0 ("btrfs: add get_tree callback for new mount API") CC: stable@vger.kernel.org # 6.12+ Reviewed-by: Daniel Vacek Reviewed-by: Qu Wenruo Signed-off-by: Dewei Meng Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/super.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c index aadc02374b2a..430e7419349c 100644 --- a/fs/btrfs/super.c +++ b/fs/btrfs/super.c @@ -2068,7 +2068,13 @@ static int btrfs_get_tree_subvol(struct fs_context *fc) fs_info->super_copy = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_KERNEL); fs_info->super_for_commit = kzalloc(BTRFS_SUPER_INFO_SIZE, GFP_KERNEL); if (!fs_info->super_copy || !fs_info->super_for_commit) { - btrfs_free_fs_info(fs_info); + /* + * Dont call btrfs_free_fs_info() to free it as it's still + * initialized partially. + */ + kfree(fs_info->super_copy); + kfree(fs_info->super_for_commit); + kvfree(fs_info); return -ENOMEM; } btrfs_init_fs_info(fs_info); From 1fabe43b4e1a97597ec5d5ffcd2b7cf96e654b8f Mon Sep 17 00:00:00 2001 From: Ting-Chang Hou Date: Thu, 16 Oct 2025 15:53:51 +0800 Subject: [PATCH 267/330] btrfs: send: fix duplicated rmdir operations when using extrefs Commit 29d6d30f5c8a ("Btrfs: send, don't send rmdir for same target multiple times") has fixed an issue that a send stream contained a rmdir operation for the same directory multiple times. After that fix we keep track of the last directory for which we sent a rmdir operation and compare with it before sending a rmdir for the parent inode of a deleted hardlink we are processing. But there is still a corner case that in between rmdir dir operations for the same inode we find deleted hardlinks for other parent inodes, so tracking just the last inode for which we sent a rmdir operation is not enough. Hardlinks of a file in the same directory are stored in the same INODE_REF item, but if the number of hardlinks is too large and can not fit in a leaf, we use INODE_EXTREF items to store them. The key of an INODE_EXTREF item is (inode_id, INODE_EXTREF, hash[name, parent ino]), so between two hardlinks for the same parent directory, we can find others for other parent directories. For example for the reproducer below we get the following (from a btrfs inspect-internal dump-tree output): item 0 key (259 INODE_EXTREF 2309449) itemoff 16257 itemsize 26 index 6925 parent 257 namelen 8 name: foo.6923 item 1 key (259 INODE_EXTREF 2311350) itemoff 16231 itemsize 26 index 6588 parent 258 namelen 8 name: foo.6587 item 2 key (259 INODE_EXTREF 2457395) itemoff 16205 itemsize 26 index 6611 parent 257 namelen 8 name: foo.6609 (...) So tracking the last directory's inode number does not work in this case since we process a link for parent inode 257, then for 258 and then back again for 257, and that second time we process a deleted link for 257 we think we have not yet sent a rmdir operation. Fix this by using a rbtree to keep track of all the directories for which we have already sent rmdir operations, and add those directories to the 'check_dirs' ref list in process_recorded_refs() only if the directory is not yet in the rbtree, otherwise skip it since it means we have already sent a rmdir operation for that directory. The following test script reproduces the problem: $ cat test.sh #!/bin/bash DEV=/dev/sdi MNT=/mnt/sdi mkfs.btrfs -f $DEV mount $DEV $MNT mkdir $MNT/a $MNT/b echo 123 > $MNT/a/foo for ((i = 1; i <= 1000; i++)); do ln $MNT/a/foo $MNT/a/foo.$i ln $MNT/a/foo $MNT/b/foo.$i done btrfs subvolume snapshot -r $MNT $MNT/snap1 btrfs send $MNT/snap1 -f /tmp/base.send rm -r $MNT/a $MNT/b btrfs subvolume snapshot -r $MNT $MNT/snap2 btrfs send -p $MNT/snap1 $MNT/snap2 -f /tmp/incremental.send umount $MNT mkfs.btrfs -f $DEV mount $DEV $MNT btrfs receive $MNT -f /tmp/base.send btrfs receive $MNT -f /tmp/incremental.send rm -f /tmp/base.send /tmp/incremental.send umount $MNT When running it, it fails like this: $ ./test.sh (...) At subvol snap1 At snapshot snap2 ERROR: rmdir o257-9-0 failed: No such file or directory CC: Reviewed-by: Filipe Manana Signed-off-by: Ting-Chang Hou [ Updated changelog ] Signed-off-by: Filipe Manana Signed-off-by: David Sterba --- fs/btrfs/send.c | 56 ++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 48 insertions(+), 8 deletions(-) diff --git a/fs/btrfs/send.c b/fs/btrfs/send.c index 6144e66661f5..96a030d28e09 100644 --- a/fs/btrfs/send.c +++ b/fs/btrfs/send.c @@ -4102,6 +4102,48 @@ static int refresh_ref_path(struct send_ctx *sctx, struct recorded_ref *ref) return ret; } +static int rbtree_check_dir_ref_comp(const void *k, const struct rb_node *node) +{ + const struct recorded_ref *data = k; + const struct recorded_ref *ref = rb_entry(node, struct recorded_ref, node); + + if (data->dir > ref->dir) + return 1; + if (data->dir < ref->dir) + return -1; + if (data->dir_gen > ref->dir_gen) + return 1; + if (data->dir_gen < ref->dir_gen) + return -1; + return 0; +} + +static bool rbtree_check_dir_ref_less(struct rb_node *node, const struct rb_node *parent) +{ + const struct recorded_ref *entry = rb_entry(node, struct recorded_ref, node); + + return rbtree_check_dir_ref_comp(entry, parent) < 0; +} + +static int record_check_dir_ref_in_tree(struct rb_root *root, + struct recorded_ref *ref, struct list_head *list) +{ + struct recorded_ref *tmp_ref; + int ret; + + if (rb_find(ref, root, rbtree_check_dir_ref_comp)) + return 0; + + ret = dup_ref(ref, list); + if (ret < 0) + return ret; + + tmp_ref = list_last_entry(list, struct recorded_ref, list); + rb_add(&tmp_ref->node, root, rbtree_check_dir_ref_less); + tmp_ref->root = root; + return 0; +} + static int rename_current_inode(struct send_ctx *sctx, struct fs_path *current_path, struct fs_path *new_path) @@ -4129,11 +4171,11 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move) struct recorded_ref *cur; struct recorded_ref *cur2; LIST_HEAD(check_dirs); + struct rb_root rbtree_check_dirs = RB_ROOT; struct fs_path *valid_path = NULL; u64 ow_inode = 0; u64 ow_gen; u64 ow_mode; - u64 last_dir_ino_rm = 0; bool did_overwrite = false; bool is_orphan = false; bool can_rename = true; @@ -4437,7 +4479,7 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move) goto out; } } - ret = dup_ref(cur, &check_dirs); + ret = record_check_dir_ref_in_tree(&rbtree_check_dirs, cur, &check_dirs); if (ret < 0) goto out; } @@ -4465,7 +4507,7 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move) } list_for_each_entry(cur, &sctx->deleted_refs, list) { - ret = dup_ref(cur, &check_dirs); + ret = record_check_dir_ref_in_tree(&rbtree_check_dirs, cur, &check_dirs); if (ret < 0) goto out; } @@ -4475,7 +4517,7 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move) * We have a moved dir. Add the old parent to check_dirs */ cur = list_first_entry(&sctx->deleted_refs, struct recorded_ref, list); - ret = dup_ref(cur, &check_dirs); + ret = record_check_dir_ref_in_tree(&rbtree_check_dirs, cur, &check_dirs); if (ret < 0) goto out; } else if (!S_ISDIR(sctx->cur_inode_mode)) { @@ -4509,7 +4551,7 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move) if (is_current_inode_path(sctx, cur->full_path)) fs_path_reset(&sctx->cur_inode_path); } - ret = dup_ref(cur, &check_dirs); + ret = record_check_dir_ref_in_tree(&rbtree_check_dirs, cur, &check_dirs); if (ret < 0) goto out; } @@ -4552,8 +4594,7 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move) ret = cache_dir_utimes(sctx, cur->dir, cur->dir_gen); if (ret < 0) goto out; - } else if (ret == inode_state_did_delete && - cur->dir != last_dir_ino_rm) { + } else if (ret == inode_state_did_delete) { ret = can_rmdir(sctx, cur->dir, cur->dir_gen); if (ret < 0) goto out; @@ -4565,7 +4606,6 @@ static int process_recorded_refs(struct send_ctx *sctx, int *pending_move) ret = send_rmdir(sctx, valid_path); if (ret < 0) goto out; - last_dir_ino_rm = cur->dir; } } } From e9ad390a4812fd60c1da46823f7a6f84f2411f0c Mon Sep 17 00:00:00 2001 From: Lorenzo Pieralisi Date: Tue, 7 Oct 2025 12:26:00 +0200 Subject: [PATCH 268/330] arm64/sysreg: Fix GIC CDEOI instruction encoding The GIC CDEOI system instruction requires the Rt field to be set to 0b11111 otherwise the instruction behaviour becomes CONSTRAINED UNPREDICTABLE. Currenly, its usage is encoded as a system register write, with a constant 0 value: write_sysreg_s(0, GICV5_OP_GIC_CDEOI) While compiling with GCC, the 0 constant value, through these asm constraints and modifiers ('x' modifier and 'Z' constraint combo): asm volatile(__msr_s(r, "%x0") : : "rZ" (__val)); forces the compiler to issue the XZR register for the MSR operation (ie that corresponds to Rt == 0b11111) issuing the right instruction encoding. Unfortunately LLVM does not yet understand that modifier/constraint combo so it ends up issuing a different register from XZR for the MSR source, which in turns means that it encodes the GIC CDEOI instruction wrongly and the instruction behaviour becomes CONSTRAINED UNPREDICTABLE that we must prevent. Add a conditional to write_sysreg_s() macro that detects whether it is passed a constant 0 value and issues an MSR write with XZR as source register - explicitly doing what the asm modifier/constraint is meant to achieve through constraints/modifiers, fixing the LLVM compilation issue. Fixes: 7ec80fb3f025 ("irqchip/gic-v5: Add GICv5 PPI support") Suggested-by: Catalin Marinas Signed-off-by: Lorenzo Pieralisi Acked-by: Marc Zyngier Cc: stable@vger.kernel.org Cc: Sascha Bischoff Cc: Will Deacon Cc: Mark Rutland Cc: Marc Zyngier Reviewed-by: Catalin Marinas Signed-off-by: Catalin Marinas --- arch/arm64/include/asm/sysreg.h | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/arch/arm64/include/asm/sysreg.h b/arch/arm64/include/asm/sysreg.h index 6455db1b54fd..c231d2a3e515 100644 --- a/arch/arm64/include/asm/sysreg.h +++ b/arch/arm64/include/asm/sysreg.h @@ -1220,10 +1220,19 @@ __val; \ }) +/* + * The "Z" constraint combined with the "%x0" template should be enough + * to force XZR generation if (v) is a constant 0 value but LLVM does not + * yet understand that modifier/constraint combo so a conditional is required + * to nudge the compiler into using XZR as a source for a 0 constant value. + */ #define write_sysreg_s(v, r) do { \ u64 __val = (u64)(v); \ u32 __maybe_unused __check_r = (u32)(r); \ - asm volatile(__msr_s(r, "%x0") : : "rZ" (__val)); \ + if (__builtin_constant_p(__val) && __val == 0) \ + asm volatile(__msr_s(r, "xzr")); \ + else \ + asm volatile(__msr_s(r, "%x0") : : "r" (__val)); \ } while (0) /* From ea0d55ae4b3207c33691a73da3443b1fd379f1d2 Mon Sep 17 00:00:00 2001 From: Ada Couprie Diaz Date: Tue, 14 Oct 2025 10:25:36 +0100 Subject: [PATCH 269/330] arm64: debug: always unmask interrupts in el0_softstp() We intend that EL0 exception handlers unmask all DAIF exceptions before calling exit_to_user_mode(). When completing single-step of a suspended breakpoint, we do not call local_daif_restore(DAIF_PROCCTX) before calling exit_to_user_mode(), leaving all DAIF exceptions masked. When pseudo-NMIs are not in use this is benign. When pseudo-NMIs are in use, this is unsound. At this point interrupts are masked by both DAIF.IF and PMR_EL1, and subsequent irq flag manipulation may not work correctly. For example, a subsequent local_irq_enable() within exit_to_user_mode_loop() will only unmask interrupts via PMR_EL1 (leaving those masked via DAIF.IF), and anything depending on interrupts being unmasked (e.g. delivery of signals) will not work correctly. This was detected by CONFIG_ARM64_DEBUG_PRIORITY_MASKING. Move the call to `try_step_suspended_breakpoints()` outside of the check so that interrupts can be unmasked even if we don't call the step handler. Fixes: 0ac7584c08ce ("arm64: debug: split single stepping exception entry") Cc: # 6.17 Signed-off-by: Ada Couprie Diaz Acked-by: Mark Rutland [catalin.marinas@arm.com: added Mark's rewritten commit log and some whitespace] Signed-off-by: Catalin Marinas --- arch/arm64/kernel/entry-common.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/arch/arm64/kernel/entry-common.c b/arch/arm64/kernel/entry-common.c index f546a914f041..a9c81715ce59 100644 --- a/arch/arm64/kernel/entry-common.c +++ b/arch/arm64/kernel/entry-common.c @@ -697,6 +697,8 @@ static void noinstr el0_breakpt(struct pt_regs *regs, unsigned long esr) static void noinstr el0_softstp(struct pt_regs *regs, unsigned long esr) { + bool step_done; + if (!is_ttbr0_addr(regs->pc)) arm64_apply_bp_hardening(); @@ -707,10 +709,10 @@ static void noinstr el0_softstp(struct pt_regs *regs, unsigned long esr) * If we are stepping a suspended breakpoint there's nothing more to do: * the single-step is complete. */ - if (!try_step_suspended_breakpoints(regs)) { - local_daif_restore(DAIF_PROCCTX); + step_done = try_step_suspended_breakpoints(regs); + local_daif_restore(DAIF_PROCCTX); + if (!step_done) do_el0_softstep(esr, regs); - } arm64_exit_to_user_mode(regs); } From 7c33e97a6ef5d84e98b892c3e00c6d1678d20395 Mon Sep 17 00:00:00 2001 From: Sahil Chandna Date: Wed, 15 Oct 2025 00:26:35 +0530 Subject: [PATCH 270/330] bpf: Do not disable preemption in bpf_test_run(). The timer mode is initialized to NO_PREEMPT mode by default, this disables preemption and force execution in atomic context causing issue on PREEMPT_RT configurations when invoking spin_lock_bh(), leading to the following warning: BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 6107, name: syz.0.17 preempt_count: 1, expected: 0 RCU nest depth: 1, expected: 1 Preemption disabled at: [] bpf_test_timer_enter+0xf8/0x140 net/bpf/test_run.c:42 Fix this, by removing NO_PREEMPT/NO_MIGRATE mode check. Also, the test timer context no longer needs explicit calls to migrate_disable()/migrate_enable() with rcu_read_lock()/rcu_read_unlock(). Use helpers rcu_read_lock_dont_migrate() and rcu_read_unlock_migrate() instead. Reported-by: syzbot+1f1fbecb9413cdbfbef8@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=1f1fbecb9413cdbfbef8 Suggested-by: Yonghong Song Suggested-by: Menglong Dong Acked-by: Yonghong Song Tested-by: syzbot+1f1fbecb9413cdbfbef8@syzkaller.appspotmail.com Co-developed-by: Brahmajit Das Signed-off-by: Brahmajit Das Signed-off-by: Sahil Chandna Link: https://lore.kernel.org/r/20251014185635.10300-1-chandna.sahil@gmail.com Signed-off-by: Alexei Starovoitov --- net/bpf/test_run.c | 23 ++++++----------------- 1 file changed, 6 insertions(+), 17 deletions(-) diff --git a/net/bpf/test_run.c b/net/bpf/test_run.c index 1782e83de2cb..8b7d0b90fea7 100644 --- a/net/bpf/test_run.c +++ b/net/bpf/test_run.c @@ -29,7 +29,6 @@ #include struct bpf_test_timer { - enum { NO_PREEMPT, NO_MIGRATE } mode; u32 i; u64 time_start, time_spent; }; @@ -37,12 +36,7 @@ struct bpf_test_timer { static void bpf_test_timer_enter(struct bpf_test_timer *t) __acquires(rcu) { - rcu_read_lock(); - if (t->mode == NO_PREEMPT) - preempt_disable(); - else - migrate_disable(); - + rcu_read_lock_dont_migrate(); t->time_start = ktime_get_ns(); } @@ -50,12 +44,7 @@ static void bpf_test_timer_leave(struct bpf_test_timer *t) __releases(rcu) { t->time_start = 0; - - if (t->mode == NO_PREEMPT) - preempt_enable(); - else - migrate_enable(); - rcu_read_unlock(); + rcu_read_unlock_migrate(); } static bool bpf_test_timer_continue(struct bpf_test_timer *t, int iterations, @@ -374,7 +363,7 @@ static int bpf_test_run_xdp_live(struct bpf_prog *prog, struct xdp_buff *ctx, { struct xdp_test_data xdp = { .batch_size = batch_size }; - struct bpf_test_timer t = { .mode = NO_MIGRATE }; + struct bpf_test_timer t = {}; int ret; if (!repeat) @@ -404,7 +393,7 @@ static int bpf_test_run(struct bpf_prog *prog, void *ctx, u32 repeat, struct bpf_prog_array_item item = {.prog = prog}; struct bpf_run_ctx *old_ctx; struct bpf_cg_run_ctx run_ctx; - struct bpf_test_timer t = { NO_MIGRATE }; + struct bpf_test_timer t = {}; enum bpf_cgroup_storage_type stype; int ret; @@ -1377,7 +1366,7 @@ int bpf_prog_test_run_flow_dissector(struct bpf_prog *prog, const union bpf_attr *kattr, union bpf_attr __user *uattr) { - struct bpf_test_timer t = { NO_PREEMPT }; + struct bpf_test_timer t = {}; u32 size = kattr->test.data_size_in; struct bpf_flow_dissector ctx = {}; u32 repeat = kattr->test.repeat; @@ -1445,7 +1434,7 @@ int bpf_prog_test_run_flow_dissector(struct bpf_prog *prog, int bpf_prog_test_run_sk_lookup(struct bpf_prog *prog, const union bpf_attr *kattr, union bpf_attr __user *uattr) { - struct bpf_test_timer t = { NO_PREEMPT }; + struct bpf_test_timer t = {}; struct bpf_prog_array *progs = NULL; struct bpf_sk_lookup_kern ctx = {}; u32 repeat = kattr->test.repeat; From a1e83d4c0361f4b0e3b7ef8b603bf5e5ef60af86 Mon Sep 17 00:00:00 2001 From: Brahmajit Das Date: Fri, 17 Oct 2025 22:45:51 +0530 Subject: [PATCH 271/330] selftests/bpf: Fix redefinition of 'off' as different kind of symbol This fixes the following build error CLNG-BPF [test_progs] verifier_global_ptr_args.bpf.o progs/verifier_global_ptr_args.c:228:5: error: redefinition of 'off' as different kind of symbol 228 | u32 off; | ^ The symbol 'off' was previously defined in tools/testing/selftests/bpf/tools/include/vmlinux.h, which includes an enum i40e_ptp_gpio_pin_state from drivers/net/ethernet/intel/i40e/i40e_ptp.c: enum i40e_ptp_gpio_pin_state { end = -2, invalid = -1, off = 0, in_A = 1, in_B = 2, out_A = 3, out_B = 4, }; This enum is included when CONFIG_I40E is enabled. As of commit 032676ff8217 ("LoongArch: Update Loongson-3 default config file"), CONFIG_I40E is set in the defconfig, which leads to the conflict. Renaming the local variable avoids the redefinition and allows the build to succeed. Suggested-by: Yonghong Song Signed-off-by: Brahmajit Das Acked-by: Yonghong Song Link: https://lore.kernel.org/r/20251017171551.53142-1-listout@listout.xyz Signed-off-by: Alexei Starovoitov --- .../selftests/bpf/progs/verifier_global_ptr_args.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/tools/testing/selftests/bpf/progs/verifier_global_ptr_args.c b/tools/testing/selftests/bpf/progs/verifier_global_ptr_args.c index 6630a92b1b47..1204fbc58178 100644 --- a/tools/testing/selftests/bpf/progs/verifier_global_ptr_args.c +++ b/tools/testing/selftests/bpf/progs/verifier_global_ptr_args.c @@ -225,7 +225,7 @@ int trusted_to_untrusted(void *ctx) } char mem[16]; -u32 off; +u32 offset; SEC("tp_btf/sys_enter") __success @@ -240,9 +240,9 @@ int anything_to_untrusted(void *ctx) /* scalar to untrusted */ subprog_untrusted(0); /* variable offset to untrusted (map) */ - subprog_untrusted((void *)mem + off); + subprog_untrusted((void *)mem + offset); /* variable offset to untrusted (trusted) */ - subprog_untrusted((void *)bpf_get_current_task_btf() + off); + subprog_untrusted((void *)bpf_get_current_task_btf() + offset); return 0; } @@ -298,12 +298,12 @@ int anything_to_untrusted_mem(void *ctx) /* scalar to untrusted mem */ subprog_void_untrusted(0); /* variable offset to untrusted mem (map) */ - subprog_void_untrusted((void *)mem + off); + subprog_void_untrusted((void *)mem + offset); /* variable offset to untrusted mem (trusted) */ - subprog_void_untrusted(bpf_get_current_task_btf() + off); + subprog_void_untrusted(bpf_get_current_task_btf() + offset); /* variable offset to untrusted char/enum (map) */ - subprog_char_untrusted(mem + off); - subprog_enum_untrusted((void *)mem + off); + subprog_char_untrusted(mem + offset); + subprog_enum_untrusted((void *)mem + offset); return 0; } From 50bd33f6b3922a6b760aa30d409cae891cec8fb5 Mon Sep 17 00:00:00 2001 From: Jianpeng Chang Date: Wed, 15 Oct 2025 10:14:27 +0800 Subject: [PATCH 272/330] net: enetc: fix the deadlock of enetc_mdio_lock After applying the workaround for err050089, the LS1028A platform experiences RCU stalls on RT kernel. This issue is caused by the recursive acquisition of the read lock enetc_mdio_lock. Here list some of the call stacks identified under the enetc_poll path that may lead to a deadlock: enetc_poll -> enetc_lock_mdio -> enetc_clean_rx_ring OR napi_complete_done -> napi_gro_receive -> enetc_start_xmit -> enetc_lock_mdio -> enetc_map_tx_buffs -> enetc_unlock_mdio -> enetc_unlock_mdio After enetc_poll acquires the read lock, a higher-priority writer attempts to acquire the lock, causing preemption. The writer detects that a read lock is already held and is scheduled out. However, readers under enetc_poll cannot acquire the read lock again because a writer is already waiting, leading to a thread hang. Currently, the deadlock is avoided by adjusting enetc_lock_mdio to prevent recursive lock acquisition. Fixes: 6d36ecdbc441 ("net: enetc: take the MDIO lock only once per NAPI poll cycle") Signed-off-by: Jianpeng Chang Acked-by: Wei Fang Link: https://patch.msgid.link/20251015021427.180757-1-jianpeng.chang.cn@windriver.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/freescale/enetc/enetc.c | 25 ++++++++++++++++---- 1 file changed, 21 insertions(+), 4 deletions(-) diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c index aae462a0cf5a..0535e92404e3 100644 --- a/drivers/net/ethernet/freescale/enetc/enetc.c +++ b/drivers/net/ethernet/freescale/enetc/enetc.c @@ -1595,6 +1595,8 @@ static int enetc_clean_rx_ring(struct enetc_bdr *rx_ring, /* next descriptor to process */ i = rx_ring->next_to_clean; + enetc_lock_mdio(); + while (likely(rx_frm_cnt < work_limit)) { union enetc_rx_bd *rxbd; struct sk_buff *skb; @@ -1630,7 +1632,9 @@ static int enetc_clean_rx_ring(struct enetc_bdr *rx_ring, rx_byte_cnt += skb->len + ETH_HLEN; rx_frm_cnt++; + enetc_unlock_mdio(); napi_gro_receive(napi, skb); + enetc_lock_mdio(); } rx_ring->next_to_clean = i; @@ -1638,6 +1642,8 @@ static int enetc_clean_rx_ring(struct enetc_bdr *rx_ring, rx_ring->stats.packets += rx_frm_cnt; rx_ring->stats.bytes += rx_byte_cnt; + enetc_unlock_mdio(); + return rx_frm_cnt; } @@ -1947,6 +1953,8 @@ static int enetc_clean_rx_ring_xdp(struct enetc_bdr *rx_ring, /* next descriptor to process */ i = rx_ring->next_to_clean; + enetc_lock_mdio(); + while (likely(rx_frm_cnt < work_limit)) { union enetc_rx_bd *rxbd, *orig_rxbd; struct xdp_buff xdp_buff; @@ -2010,7 +2018,9 @@ static int enetc_clean_rx_ring_xdp(struct enetc_bdr *rx_ring, */ enetc_bulk_flip_buff(rx_ring, orig_i, i); + enetc_unlock_mdio(); napi_gro_receive(napi, skb); + enetc_lock_mdio(); break; case XDP_TX: tx_ring = priv->xdp_tx_ring[rx_ring->index]; @@ -2045,7 +2055,9 @@ static int enetc_clean_rx_ring_xdp(struct enetc_bdr *rx_ring, } break; case XDP_REDIRECT: + enetc_unlock_mdio(); err = xdp_do_redirect(rx_ring->ndev, &xdp_buff, prog); + enetc_lock_mdio(); if (unlikely(err)) { enetc_xdp_drop(rx_ring, orig_i, i); rx_ring->stats.xdp_redirect_failures++; @@ -2065,8 +2077,11 @@ static int enetc_clean_rx_ring_xdp(struct enetc_bdr *rx_ring, rx_ring->stats.packets += rx_frm_cnt; rx_ring->stats.bytes += rx_byte_cnt; - if (xdp_redirect_frm_cnt) + if (xdp_redirect_frm_cnt) { + enetc_unlock_mdio(); xdp_do_flush(); + enetc_lock_mdio(); + } if (xdp_tx_frm_cnt) enetc_update_tx_ring_tail(tx_ring); @@ -2075,6 +2090,8 @@ static int enetc_clean_rx_ring_xdp(struct enetc_bdr *rx_ring, enetc_refill_rx_ring(rx_ring, enetc_bd_unused(rx_ring) - rx_ring->xdp.xdp_tx_in_flight); + enetc_unlock_mdio(); + return rx_frm_cnt; } @@ -2093,6 +2110,7 @@ static int enetc_poll(struct napi_struct *napi, int budget) for (i = 0; i < v->count_tx_rings; i++) if (!enetc_clean_tx_ring(&v->tx_ring[i], budget)) complete = false; + enetc_unlock_mdio(); prog = rx_ring->xdp.prog; if (prog) @@ -2104,10 +2122,8 @@ static int enetc_poll(struct napi_struct *napi, int budget) if (work_done) v->rx_napi_work = true; - if (!complete) { - enetc_unlock_mdio(); + if (!complete) return budget; - } napi_complete_done(napi, work_done); @@ -2116,6 +2132,7 @@ static int enetc_poll(struct napi_struct *napi, int budget) v->rx_napi_work = false; + enetc_lock_mdio(); /* enable interrupts */ enetc_wr_reg_hot(v->rbier, ENETC_RBIER_RXTIE); From e59bc32df2e989f034623a580e30a2a72af33b3f Mon Sep 17 00:00:00 2001 From: Wei Fang Date: Thu, 16 Oct 2025 16:01:31 +0800 Subject: [PATCH 273/330] net: enetc: correct the value of ENETC_RXB_TRUESIZE The ENETC RX ring uses the page halves flipping mechanism, each page is split into two halves for the RX ring to use. And ENETC_RXB_TRUESIZE is defined to 2048 to indicate the size of half a page. However, the page size is configurable, for ARM64 platform, PAGE_SIZE is default to 4K, but it could be configured to 16K or 64K. When PAGE_SIZE is set to 16K or 64K, ENETC_RXB_TRUESIZE is not correct, and the RX ring will always use the first half of the page. This is not consistent with the description in the relevant kernel doc and commit messages. This issue is invisible in most cases, but if users want to increase PAGE_SIZE to receive a Jumbo frame with a single buffer for some use cases, it will not work as expected, because the buffer size of each RX BD is fixed to 2048 bytes. Based on the above two points, we expect to correct ENETC_RXB_TRUESIZE to (PAGE_SIZE >> 1), as described in the comment. Fixes: d4fd0404c1c9 ("enetc: Introduce basic PF and VF ENETC ethernet drivers") Signed-off-by: Wei Fang Reviewed-by: Claudiu Manoil Link: https://patch.msgid.link/20251016080131.3127122-1-wei.fang@nxp.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/freescale/enetc/enetc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/freescale/enetc/enetc.h b/drivers/net/ethernet/freescale/enetc/enetc.h index 0ec010a7d640..f279fa597991 100644 --- a/drivers/net/ethernet/freescale/enetc/enetc.h +++ b/drivers/net/ethernet/freescale/enetc/enetc.h @@ -76,7 +76,7 @@ struct enetc_lso_t { #define ENETC_LSO_MAX_DATA_LEN SZ_256K #define ENETC_RX_MAXFRM_SIZE ENETC_MAC_MAXFRM_SIZE -#define ENETC_RXB_TRUESIZE 2048 /* PAGE_SIZE >> 1 */ +#define ENETC_RXB_TRUESIZE (PAGE_SIZE >> 1) #define ENETC_RXB_PAD NET_SKB_PAD /* add extra space if needed */ #define ENETC_RXB_DMA_SIZE \ (SKB_WITH_OVERHEAD(ENETC_RXB_TRUESIZE) - ENETC_RXB_PAD) From cb74f8c952508bc85ec9583fa7da31c9b1440f26 Mon Sep 17 00:00:00 2001 From: Bagas Sanjaya Date: Thu, 16 Oct 2025 16:39:37 +0700 Subject: [PATCH 274/330] Documentation: net: net_failover: Separate cloud-ifupdown-helper and reattach-vf.sh code blocks marker cloud-ifupdown-helper patch and reattach-vf.sh script are rendered in htmldocs output as normal paragraphs instead of literal code blocks due to missing separator from respective code block marker. Add it. Signed-off-by: Bagas Sanjaya Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251016093936.29442-2-bagasdotme@gmail.com Signed-off-by: Jakub Kicinski --- Documentation/networking/net_failover.rst | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/Documentation/networking/net_failover.rst b/Documentation/networking/net_failover.rst index f4e1b4e07adc..2f776e90d318 100644 --- a/Documentation/networking/net_failover.rst +++ b/Documentation/networking/net_failover.rst @@ -96,9 +96,8 @@ needed to these network configuration daemons to make sure that an IP is received only on the 'failover' device. Below is the patch snippet used with 'cloud-ifupdown-helper' script found on -Debian cloud images: +Debian cloud images:: -:: @@ -27,6 +27,8 @@ do_setup() { local working="$cfgdir/.$INTERFACE" local final="$cfgdir/$INTERFACE" @@ -172,9 +171,8 @@ appropriate FDB entry is added. The following script is executed on the destination hypervisor once migration completes, and it reattaches the VF to the VM and brings down the virtio-net -interface. +interface:: -:: # reattach-vf.sh #!/bin/bash From e0caeb24f538c3c9c94f471882ceeb43d9dc2739 Mon Sep 17 00:00:00 2001 From: Tonghao Zhang Date: Thu, 16 Oct 2025 20:51:36 +0800 Subject: [PATCH 275/330] net: bonding: update the slave array for broadcast mode This patch fixes ce7a381697cb ("net: bonding: add broadcast_neighbor option for 802.3ad"). Before this commit, on the broadcast mode, all devices were traversed using the bond_for_each_slave_rcu. This patch supports traversing devices by using all_slaves. Therefore, we need to update the slave array when enslave or release slave. Fixes: ce7a381697cb ("net: bonding: add broadcast_neighbor option for 802.3ad") Cc: Simon Horman Cc: Jonathan Corbet Cc: Andrew Lunn Cc: Reported-by: Jiri Slaby Tested-by: Jiri Slaby Link: https://lore.kernel.org/all/a97e6e1e-81bc-4a79-8352-9e4794b0d2ca@kernel.org/ Signed-off-by: Tonghao Zhang Reviewed-by: Hangbin Liu Reviewed-by: Nikolay Aleksandrov Acked-by: Jay Vosburgh Link: https://patch.msgid.link/20251016125136.16568-1-tonghao@bamaicloud.com Signed-off-by: Jakub Kicinski --- drivers/net/bonding/bond_main.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 4da619210c1f..67fdcbdd2764 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -2287,7 +2287,9 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev, unblock_netpoll_tx(); } - if (bond_mode_can_use_xmit_hash(bond)) + /* broadcast mode uses the all_slaves to loop through slaves. */ + if (bond_mode_can_use_xmit_hash(bond) || + BOND_MODE(bond) == BOND_MODE_BROADCAST) bond_update_slave_arr(bond, NULL); if (!slave_dev->netdev_ops->ndo_bpf || @@ -2463,7 +2465,8 @@ static int __bond_release_one(struct net_device *bond_dev, bond_upper_dev_unlink(bond, slave); - if (bond_mode_can_use_xmit_hash(bond)) + if (bond_mode_can_use_xmit_hash(bond) || + BOND_MODE(bond) == BOND_MODE_BROADCAST) bond_update_slave_arr(bond, slave); slave_info(bond_dev, slave_dev, "Releasing %s interface\n", From 902e81e679d86846a2404630d349709ad9372d0d Mon Sep 17 00:00:00 2001 From: Ioana Ciornei Date: Thu, 16 Oct 2025 16:58:07 +0300 Subject: [PATCH 276/330] dpaa2-eth: fix the pointer passed to PTR_ALIGN on Tx path The blamed commit increased the needed headroom to account for alignment. This means that the size required to always align a Tx buffer was added inside the dpaa2_eth_needed_headroom() function. By doing that, a manual adjustment of the pointer passed to PTR_ALIGN() was no longer correct since the 'buffer_start' variable was already pointing to the start of the skb's memory. The behavior of the dpaa2-eth driver without this patch was to drop frames on Tx even when the headroom was matching the 128 bytes necessary. Fix this by removing the manual adjust of 'buffer_start' from the PTR_MODE call. Closes: https://lore.kernel.org/netdev/70f0dcd9-1906-4d13-82df-7bbbbe7194c6@app.fastmail.com/T/#u Fixes: f422abe3f23d ("dpaa2-eth: increase the needed headroom to account for alignment") Signed-off-by: Ioana Ciornei Tested-by: Mathew McBride Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251016135807.360978-1-ioana.ciornei@nxp.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c index c96d1d6ba8fe..18d86badd6ea 100644 --- a/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c +++ b/drivers/net/ethernet/freescale/dpaa2/dpaa2-eth.c @@ -1077,8 +1077,7 @@ static int dpaa2_eth_build_single_fd(struct dpaa2_eth_priv *priv, dma_addr_t addr; buffer_start = skb->data - dpaa2_eth_needed_headroom(skb); - aligned_start = PTR_ALIGN(buffer_start - DPAA2_ETH_TX_BUF_ALIGN, - DPAA2_ETH_TX_BUF_ALIGN); + aligned_start = PTR_ALIGN(buffer_start, DPAA2_ETH_TX_BUF_ALIGN); if (aligned_start >= skb->head) buffer_start = aligned_start; else From ffff5c8fc2af2218a3332b3d5b97654599d50cde Mon Sep 17 00:00:00 2001 From: Aleksander Jan Bajkowski Date: Thu, 16 Oct 2025 21:22:52 +0200 Subject: [PATCH 277/330] net: phy: realtek: fix rtl8221b-vm-cg name When splitting the RTL8221B-VM-CG into C22 and C45 variants, the name was accidentally changed to RTL8221B-VN-CG. This patch brings back the previous part number. Fixes: ad5ce743a6b0 ("net: phy: realtek: Add driver instances for rtl8221b via Clause 45") Signed-off-by: Aleksander Jan Bajkowski Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251016192325.2306757-1-olek2@wp.pl Signed-off-by: Jakub Kicinski --- drivers/net/phy/realtek/realtek_main.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/drivers/net/phy/realtek/realtek_main.c b/drivers/net/phy/realtek/realtek_main.c index a724b21b4fe7..16a347084293 100644 --- a/drivers/net/phy/realtek/realtek_main.c +++ b/drivers/net/phy/realtek/realtek_main.c @@ -154,7 +154,7 @@ #define RTL_8211FVD_PHYID 0x001cc878 #define RTL_8221B 0x001cc840 #define RTL_8221B_VB_CG 0x001cc849 -#define RTL_8221B_VN_CG 0x001cc84a +#define RTL_8221B_VM_CG 0x001cc84a #define RTL_8251B 0x001cc862 #define RTL_8261C 0x001cc890 @@ -1523,16 +1523,16 @@ static int rtl8221b_vb_cg_c45_match_phy_device(struct phy_device *phydev, return rtlgen_is_c45_match(phydev, RTL_8221B_VB_CG, true); } -static int rtl8221b_vn_cg_c22_match_phy_device(struct phy_device *phydev, +static int rtl8221b_vm_cg_c22_match_phy_device(struct phy_device *phydev, const struct phy_driver *phydrv) { - return rtlgen_is_c45_match(phydev, RTL_8221B_VN_CG, false); + return rtlgen_is_c45_match(phydev, RTL_8221B_VM_CG, false); } -static int rtl8221b_vn_cg_c45_match_phy_device(struct phy_device *phydev, +static int rtl8221b_vm_cg_c45_match_phy_device(struct phy_device *phydev, const struct phy_driver *phydrv) { - return rtlgen_is_c45_match(phydev, RTL_8221B_VN_CG, true); + return rtlgen_is_c45_match(phydev, RTL_8221B_VM_CG, true); } static int rtl_internal_nbaset_match_phy_device(struct phy_device *phydev, @@ -1879,7 +1879,7 @@ static struct phy_driver realtek_drvs[] = { .suspend = genphy_c45_pma_suspend, .resume = rtlgen_c45_resume, }, { - .match_phy_device = rtl8221b_vn_cg_c22_match_phy_device, + .match_phy_device = rtl8221b_vm_cg_c22_match_phy_device, .name = "RTL8221B-VM-CG 2.5Gbps PHY (C22)", .probe = rtl822x_probe, .get_features = rtl822x_get_features, @@ -1892,8 +1892,8 @@ static struct phy_driver realtek_drvs[] = { .read_page = rtl821x_read_page, .write_page = rtl821x_write_page, }, { - .match_phy_device = rtl8221b_vn_cg_c45_match_phy_device, - .name = "RTL8221B-VN-CG 2.5Gbps PHY (C45)", + .match_phy_device = rtl8221b_vm_cg_c45_match_phy_device, + .name = "RTL8221B-VM-CG 2.5Gbps PHY (C45)", .probe = rtl822x_probe, .config_init = rtl822xb_config_init, .get_rate_matching = rtl822xb_get_rate_matching, From dbfdaeb381a49a7bc753d18e2876bc56a15e01cc Mon Sep 17 00:00:00 2001 From: Stuart Yoder Date: Sat, 18 Oct 2025 14:25:18 +0300 Subject: [PATCH 278/330] tpm_crb: Add idle support for the Arm FF-A start method According to the CRB over FF-A specification [1], a TPM that implements the ABI must comply with the TCG PTP specification. This requires support for the Idle and Ready states. This patch implements CRB control area requests for goIdle and cmdReady on FF-A based TPMs. The FF-A message used to notify the TPM of CRB updates includes a locality parameter, which provides a hint to the TPM about which locality modified the CRB. This patch adds a locality parameter to __crb_go_idle() and __crb_cmd_ready() to support this. [1] https://developer.arm.com/documentation/den0138/latest/ Signed-off-by: Stuart Yoder Reviewed-by: Jarkko Sakkinen Signed-off-by: Jarkko Sakkinen --- drivers/char/tpm/tpm_crb.c | 29 ++++++++++++++++++++--------- 1 file changed, 20 insertions(+), 9 deletions(-) diff --git a/drivers/char/tpm/tpm_crb.c b/drivers/char/tpm/tpm_crb.c index ed97344f2324..c75a531cfb98 100644 --- a/drivers/char/tpm/tpm_crb.c +++ b/drivers/char/tpm/tpm_crb.c @@ -133,8 +133,7 @@ static inline bool tpm_crb_has_idle(u32 start_method) { return !(start_method == ACPI_TPM2_START_METHOD || start_method == ACPI_TPM2_COMMAND_BUFFER_WITH_START_METHOD || - start_method == ACPI_TPM2_COMMAND_BUFFER_WITH_ARM_SMC || - start_method == ACPI_TPM2_CRB_WITH_ARM_FFA); + start_method == ACPI_TPM2_COMMAND_BUFFER_WITH_ARM_SMC); } static bool crb_wait_for_reg_32(u32 __iomem *reg, u32 mask, u32 value, @@ -191,7 +190,7 @@ static int crb_try_pluton_doorbell(struct crb_priv *priv, bool wait_for_complete * * Return: 0 always */ -static int __crb_go_idle(struct device *dev, struct crb_priv *priv) +static int __crb_go_idle(struct device *dev, struct crb_priv *priv, int loc) { int rc; @@ -200,6 +199,12 @@ static int __crb_go_idle(struct device *dev, struct crb_priv *priv) iowrite32(CRB_CTRL_REQ_GO_IDLE, &priv->regs_t->ctrl_req); + if (priv->sm == ACPI_TPM2_CRB_WITH_ARM_FFA) { + rc = tpm_crb_ffa_start(CRB_FFA_START_TYPE_COMMAND, loc); + if (rc) + return rc; + } + rc = crb_try_pluton_doorbell(priv, true); if (rc) return rc; @@ -220,7 +225,7 @@ static int crb_go_idle(struct tpm_chip *chip) struct device *dev = &chip->dev; struct crb_priv *priv = dev_get_drvdata(dev); - return __crb_go_idle(dev, priv); + return __crb_go_idle(dev, priv, chip->locality); } /** @@ -238,7 +243,7 @@ static int crb_go_idle(struct tpm_chip *chip) * * Return: 0 on success -ETIME on timeout; */ -static int __crb_cmd_ready(struct device *dev, struct crb_priv *priv) +static int __crb_cmd_ready(struct device *dev, struct crb_priv *priv, int loc) { int rc; @@ -247,6 +252,12 @@ static int __crb_cmd_ready(struct device *dev, struct crb_priv *priv) iowrite32(CRB_CTRL_REQ_CMD_READY, &priv->regs_t->ctrl_req); + if (priv->sm == ACPI_TPM2_CRB_WITH_ARM_FFA) { + rc = tpm_crb_ffa_start(CRB_FFA_START_TYPE_COMMAND, loc); + if (rc) + return rc; + } + rc = crb_try_pluton_doorbell(priv, true); if (rc) return rc; @@ -267,7 +278,7 @@ static int crb_cmd_ready(struct tpm_chip *chip) struct device *dev = &chip->dev; struct crb_priv *priv = dev_get_drvdata(dev); - return __crb_cmd_ready(dev, priv); + return __crb_cmd_ready(dev, priv, chip->locality); } static int __crb_request_locality(struct device *dev, @@ -444,7 +455,7 @@ static int crb_send(struct tpm_chip *chip, u8 *buf, size_t bufsiz, size_t len) /* Seems to be necessary for every command */ if (priv->sm == ACPI_TPM2_COMMAND_BUFFER_WITH_PLUTON) - __crb_cmd_ready(&chip->dev, priv); + __crb_cmd_ready(&chip->dev, priv, chip->locality); memcpy_toio(priv->cmd, buf, len); @@ -672,7 +683,7 @@ static int crb_map_io(struct acpi_device *device, struct crb_priv *priv, * PTT HW bug w/a: wake up the device to access * possibly not retained registers. */ - ret = __crb_cmd_ready(dev, priv); + ret = __crb_cmd_ready(dev, priv, 0); if (ret) goto out_relinquish_locality; @@ -744,7 +755,7 @@ static int crb_map_io(struct acpi_device *device, struct crb_priv *priv, if (!ret) priv->cmd_size = cmd_size; - __crb_go_idle(dev, priv); + __crb_go_idle(dev, priv, 0); out_relinquish_locality: From 211ddde0823f1442e4ad052a2f30f050145ccada Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 19 Oct 2025 15:19:16 -1000 Subject: [PATCH 279/330] Linux 6.18-rc2 --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 17cfa11ca716..d14824792227 100644 --- a/Makefile +++ b/Makefile @@ -2,7 +2,7 @@ VERSION = 6 PATCHLEVEL = 18 SUBLEVEL = 0 -EXTRAVERSION = -rc1 +EXTRAVERSION = -rc2 NAME = Baby Opossum Posse # *DOCUMENTATION* From 3a20c444cd123e820e10ae22eeaf00e189315aa1 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Fri, 17 Oct 2025 16:28:49 +0200 Subject: [PATCH 280/330] can: bxcan: bxcan_start_xmit(): use can_dev_dropped_skb() instead of can_dropped_invalid_skb() In addition to can_dropped_invalid_skb(), the helper function can_dev_dropped_skb() checks whether the device is in listen-only mode and discards the skb accordingly. Replace can_dropped_invalid_skb() by can_dev_dropped_skb() to also drop skbs in for listen-only mode. Reported-by: Marc Kleine-Budde Closes: https://lore.kernel.org/all/20251017-bizarre-enchanted-quokka-f3c704-mkl@pengutronix.de/ Fixes: f00647d8127b ("can: bxcan: add support for ST bxCAN controller") Link: https://patch.msgid.link/20251017-fix-skb-drop-check-v1-1-556665793fa4@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/bxcan.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/can/bxcan.c b/drivers/net/can/bxcan.c index bfc60eb33dc3..333ad42ea73b 100644 --- a/drivers/net/can/bxcan.c +++ b/drivers/net/can/bxcan.c @@ -842,7 +842,7 @@ static netdev_tx_t bxcan_start_xmit(struct sk_buff *skb, u32 id; int i, j; - if (can_dropped_invalid_skb(ndev, skb)) + if (can_dev_dropped_skb(ndev, skb)) return NETDEV_TX_OK; if (bxcan_tx_busy(priv)) From 0bee15a5caf36fe513fdeee07fd4f0331e61c064 Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Fri, 17 Oct 2025 16:28:49 +0200 Subject: [PATCH 281/330] can: esd: acc_start_xmit(): use can_dev_dropped_skb() instead of can_dropped_invalid_skb() In addition to can_dropped_invalid_skb(), the helper function can_dev_dropped_skb() checks whether the device is in listen-only mode and discards the skb accordingly. Replace can_dropped_invalid_skb() by can_dev_dropped_skb() to also drop skbs in for listen-only mode. Reported-by: Marc Kleine-Budde Closes: https://lore.kernel.org/all/20251017-bizarre-enchanted-quokka-f3c704-mkl@pengutronix.de/ Fixes: 9721866f07e1 ("can: esd: add support for esd GmbH PCIe/402 CAN interface family") Link: https://patch.msgid.link/20251017-fix-skb-drop-check-v1-2-556665793fa4@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/esd/esdacc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/can/esd/esdacc.c b/drivers/net/can/esd/esdacc.c index c80032bc1a52..73e66f9a3781 100644 --- a/drivers/net/can/esd/esdacc.c +++ b/drivers/net/can/esd/esdacc.c @@ -254,7 +254,7 @@ netdev_tx_t acc_start_xmit(struct sk_buff *skb, struct net_device *netdev) u32 acc_id; u32 acc_dlc; - if (can_dropped_invalid_skb(netdev, skb)) + if (can_dev_dropped_skb(netdev, skb)) return NETDEV_TX_OK; /* Access core->tx_fifo_tail only once because it may be changed From 3a3bc9bbb3a0287164a595787df0c70d91e77cfd Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Fri, 17 Oct 2025 16:28:49 +0200 Subject: [PATCH 282/330] can: rockchip-canfd: rkcanfd_start_xmit(): use can_dev_dropped_skb() instead of can_dropped_invalid_skb() In addition to can_dropped_invalid_skb(), the helper function can_dev_dropped_skb() checks whether the device is in listen-only mode and discards the skb accordingly. Replace can_dropped_invalid_skb() by can_dev_dropped_skb() to also drop skbs in for listen-only mode. Reported-by: Marc Kleine-Budde Closes: https://lore.kernel.org/all/20251017-bizarre-enchanted-quokka-f3c704-mkl@pengutronix.de/ Fixes: ff60bfbaf67f ("can: rockchip_canfd: add driver for Rockchip CAN-FD controller") Link: https://patch.msgid.link/20251017-fix-skb-drop-check-v1-3-556665793fa4@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/rockchip/rockchip_canfd-tx.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/can/rockchip/rockchip_canfd-tx.c b/drivers/net/can/rockchip/rockchip_canfd-tx.c index 865a15e033a9..12200dcfd338 100644 --- a/drivers/net/can/rockchip/rockchip_canfd-tx.c +++ b/drivers/net/can/rockchip/rockchip_canfd-tx.c @@ -72,7 +72,7 @@ netdev_tx_t rkcanfd_start_xmit(struct sk_buff *skb, struct net_device *ndev) int err; u8 i; - if (can_dropped_invalid_skb(ndev, skb)) + if (can_dev_dropped_skb(ndev, skb)) return NETDEV_TX_OK; if (!netif_subqueue_maybe_stop(priv->ndev, 0, From 8e93ac51e4c6dc399fad59ec21f55f2cfb46d27c Mon Sep 17 00:00:00 2001 From: Marc Kleine-Budde Date: Mon, 20 Oct 2025 11:51:03 +0200 Subject: [PATCH 283/330] can: netlink: can_changelink(): allow disabling of automatic restart Since the commit c1f3f9797c1f ("can: netlink: can_changelink(): fix NULL pointer deref of struct can_priv::do_set_mode"), the automatic restart delay can only be set for devices that implement the restart handler struct can_priv::do_set_mode. As it makes no sense to configure a automatic restart for devices that doesn't support it. However, since systemd commit 13ce5d4632e3 ("network/can: properly handle CAN.RestartSec=0") [1], systemd-networkd correctly handles a restart delay of "0" (i.e. the restart is disabled). Which means that a disabled restart is always configured in the kernel. On systems with both changes active this causes that CAN interfaces that don't implement a restart handler cannot be brought up by systemd-networkd. Solve this problem by allowing a delay of "0" to be configured, even if the device does not implement a restart handler. [1] https://github.com/systemd/systemd/commit/13ce5d4632e395521e6205c954493c7fc1c4c6e0 Cc: stable@vger.kernel.org Cc: Andrei Lalaev Reported-by: Marc Kleine-Budde Closes: https://lore.kernel.org/all/20251020-certain-arrogant-vole-of-sunshine-141841-mkl@pengutronix.de Fixes: c1f3f9797c1f ("can: netlink: can_changelink(): fix NULL pointer deref of struct can_priv::do_set_mode") Link: https://patch.msgid.link/20251020-netlink-fix-restart-v1-1-3f53c7f8520b@pengutronix.de Signed-off-by: Marc Kleine-Budde --- drivers/net/can/dev/netlink.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/net/can/dev/netlink.c b/drivers/net/can/dev/netlink.c index 0591406b6f32..6f83b87d54fc 100644 --- a/drivers/net/can/dev/netlink.c +++ b/drivers/net/can/dev/netlink.c @@ -452,7 +452,9 @@ static int can_changelink(struct net_device *dev, struct nlattr *tb[], } if (data[IFLA_CAN_RESTART_MS]) { - if (!priv->do_set_mode) { + unsigned int restart_ms = nla_get_u32(data[IFLA_CAN_RESTART_MS]); + + if (restart_ms != 0 && !priv->do_set_mode) { NL_SET_ERR_MSG(extack, "Device doesn't support restart from Bus Off"); return -EOPNOTSUPP; @@ -461,7 +463,7 @@ static int can_changelink(struct net_device *dev, struct nlattr *tb[], /* Do not allow changing restart delay while running */ if (dev->flags & IFF_UP) return -EBUSY; - priv->restart_ms = nla_get_u32(data[IFLA_CAN_RESTART_MS]); + priv->restart_ms = restart_ms; } if (data[IFLA_CAN_RESTART]) { From 10fad4012234a7dea621ae17c0c9486824f645a0 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Sat, 18 Oct 2025 14:27:15 +0200 Subject: [PATCH 284/330] Revert "cpuidle: menu: Avoid discarding useful information" It is reported that commit 85975daeaa4d ("cpuidle: menu: Avoid discarding useful information") led to a performance regression on Intel Jasper Lake systems because it reduced the time spent by CPUs in idle state C7 which is correlated to the maximum frequency the CPUs can get to because of an average running power limit [1]. Before that commit, get_typical_interval() would have returned UINT_MAX whenever it had been unable to make a high-confidence prediction which had led to selecting the deepest available idle state too often and both power and performance had been inadequate as a result of that on some systems. However, this had not been a problem on systems with relatively aggressive average running power limits, like the Jasper Lake systems in question, because on those systems it was compensated by the ability to run CPUs faster. It was addressed by causing get_typical_interval() to return a number based on the recent idle duration information available to it even if it could not make a high-confidence prediction, but that clearly did not take the possible correlation between idle power and available CPU capacity into account. For this reason, revert most of the changes made by commit 85975daeaa4d, except for one cosmetic cleanup, and add a comment explaining the rationale for returning UINT_MAX from get_typical_interval() when it is unable to make a high-confidence prediction. Fixes: 85975daeaa4d ("cpuidle: menu: Avoid discarding useful information") Closes: https://lore.kernel.org/linux-pm/36iykr223vmcfsoysexug6s274nq2oimcu55ybn6ww4il3g3cv@cohflgdbpnq7/ [1] Reported-by: Sergey Senozhatsky Cc: All applicable Signed-off-by: Rafael J. Wysocki Link: https://patch.msgid.link/3663603.iIbC2pHGDl@rafael.j.wysocki --- drivers/cpuidle/governors/menu.c | 21 +++++++++------------ 1 file changed, 9 insertions(+), 12 deletions(-) diff --git a/drivers/cpuidle/governors/menu.c b/drivers/cpuidle/governors/menu.c index 4d9aa5ce31f0..7d21fb5a72f4 100644 --- a/drivers/cpuidle/governors/menu.c +++ b/drivers/cpuidle/governors/menu.c @@ -188,20 +188,17 @@ static unsigned int get_typical_interval(struct menu_device *data) * * This can deal with workloads that have long pauses interspersed * with sporadic activity with a bunch of short pauses. + * + * However, if the number of remaining samples is too small to exclude + * any more outliers, allow the deepest available idle state to be + * selected because there are systems where the time spent by CPUs in + * deep idle states is correlated to the maximum frequency the CPUs + * can get to. On those systems, shallow idle states should be avoided + * unless there is a clear indication that the given CPU is most likley + * going to be woken up shortly. */ - if (divisor * 4 <= INTERVALS * 3) { - /* - * If there are sufficiently many data points still under - * consideration after the outliers have been eliminated, - * returning without a prediction would be a mistake because it - * is likely that the next interval will not exceed the current - * maximum, so return the latter in that case. - */ - if (divisor >= INTERVALS / 2) - return max; - + if (divisor * 4 <= INTERVALS * 3) return UINT_MAX; - } /* Update the thresholds for the next round. */ if (avg - min > max - avg) From f3f313c51148668d3c44f1119762325ce2c0715b Mon Sep 17 00:00:00 2001 From: Sunil V L Date: Fri, 17 Oct 2025 15:37:44 +0530 Subject: [PATCH 285/330] ACPI: property: Fix argument order in __acpi_node_get_property_reference() A refactoring bug introduced an argument order mistake in the call to acpi_fwnode_get_reference_args() from __acpi_node_get_property_reference(). This caused incorrect behavior when resolving ACPI property references. Fix the issue by correcting the argument order. Fixes: e121be784d35 ("ACPI: property: Refactor acpi_fwnode_get_reference_args() to support nargs_prop") Reported-by: Thomas Richard Closes: https://lore.kernel.org/all/1241f2b6-9b4e-4623-8a83-77db8774ac32@bootlin.com/ Tested-by: Thomas Richard Signed-off-by: Sunil V L Reviewed-by: Andy Shevchenko Link: https://patch.msgid.link/20251017100744.71871-1-sunilvl@ventanamicro.com Signed-off-by: Rafael J. Wysocki --- drivers/acpi/property.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/acpi/property.c b/drivers/acpi/property.c index 1b997a5497e7..43d5e457814e 100644 --- a/drivers/acpi/property.c +++ b/drivers/acpi/property.c @@ -1107,7 +1107,7 @@ int __acpi_node_get_property_reference(const struct fwnode_handle *fwnode, size_t num_args, struct fwnode_reference_args *args) { - return acpi_fwnode_get_reference_args(fwnode, propname, NULL, index, num_args, args); + return acpi_fwnode_get_reference_args(fwnode, propname, NULL, num_args, index, args); } EXPORT_SYMBOL_GPL(__acpi_node_get_property_reference); From b2c37c1168f537900158c860174001d055d8d583 Mon Sep 17 00:00:00 2001 From: Alex Williamson Date: Mon, 13 Oct 2025 09:26:11 -0600 Subject: [PATCH 286/330] MAINTAINERS: Update Alex Williamson's email address Switch to a personal email account as I'll be leaving Red Hat soon. Signed-off-by: Alex Williamson Reviewed-by: Kevin Tian Link: https://lore.kernel.org/r/20251013152613.3088777-1-alex.williamson@redhat.com Signed-off-by: Alex Williamson --- .mailmap | 1 + MAINTAINERS | 4 ++-- 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/.mailmap b/.mailmap index c371ba3fb845..717d754b378c 100644 --- a/.mailmap +++ b/.mailmap @@ -27,6 +27,7 @@ Alan Cox Alan Cox Aleksandar Markovic Aleksey Gorelov +Alex Williamson Alexander Lobakin Alexander Lobakin Alexander Lobakin diff --git a/MAINTAINERS b/MAINTAINERS index 545a4776795e..74389eddb954 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -26885,7 +26885,7 @@ S: Maintained F: drivers/vfio/cdx/* VFIO DRIVER -M: Alex Williamson +M: Alex Williamson L: kvm@vger.kernel.org S: Maintained T: git https://github.com/awilliam/linux-vfio.git @@ -27048,7 +27048,7 @@ T: git git://linuxtv.org/media.git F: drivers/media/test-drivers/vimc/* VIRT LIB -M: Alex Williamson +M: Alex Williamson M: Paolo Bonzini L: kvm@vger.kernel.org S: Supported From a73ca0449bcb7c238097cc6a1bf3fd82a78374df Mon Sep 17 00:00:00 2001 From: Xin Long Date: Fri, 17 Oct 2025 16:06:14 -0400 Subject: [PATCH 287/330] selftests: net: fix server bind failure in sctp_vrf.sh MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit sctp_vrf.sh could fail: TEST 12: bind vrf-2 & 1 in server, connect from client 1 & 2, N [FAIL] not ok 1 selftests: net: sctp_vrf.sh # exit=3 The failure happens when the server bind in a new run conflicts with an existing association from the previous run: [1] ip netns exec $SERVER_NS ./sctp_hello server ... [2] ip netns exec $CLIENT_NS ./sctp_hello client ... [3] ip netns exec $SERVER_NS pkill sctp_hello ... [4] ip netns exec $SERVER_NS ./sctp_hello server ... It occurs if the client in [2] sends a message and closes immediately. With the message unacked, no SHUTDOWN is sent. Killing the server in [3] triggers a SHUTDOWN the client also ignores due to the unacked message, leaving the old association alive. This causes the bind at [4] to fail until the message is acked and the client responds to a second SHUTDOWN after the server’s T2 timer expires (3s). This patch fixes the issue by preventing the client from sending data. Instead, the client blocks on recv() and waits for the server to close. It also waits until both the server and the client sockets are fully released in stop_server and wait_client before restarting. Additionally, replace 2>&1 >/dev/null with -q in sysctl and grep, and drop other redundant 2>&1 >/dev/null redirections, and fix a typo from N to Y (connect successfully) in the description of the last test. Fixes: a61bd7b9fef3 ("selftests: add a selftest for sctp vrf") Reported-by: Hangbin Liu Tested-by: Jakub Kicinski Signed-off-by: Xin Long Link: https://patch.msgid.link/be2dacf52d0917c4ba5e2e8c5a9cb640740ad2b6.1760731574.git.lucien.xin@gmail.com Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/sctp_hello.c | 17 +----- tools/testing/selftests/net/sctp_vrf.sh | 73 +++++++++++++++--------- 2 files changed, 47 insertions(+), 43 deletions(-) diff --git a/tools/testing/selftests/net/sctp_hello.c b/tools/testing/selftests/net/sctp_hello.c index f02f1f95d227..a04dac0b8027 100644 --- a/tools/testing/selftests/net/sctp_hello.c +++ b/tools/testing/selftests/net/sctp_hello.c @@ -29,7 +29,6 @@ static void set_addr(struct sockaddr_storage *ss, char *ip, char *port, int *len static int do_client(int argc, char *argv[]) { struct sockaddr_storage ss; - char buf[] = "hello"; int csk, ret, len; if (argc < 5) { @@ -56,16 +55,10 @@ static int do_client(int argc, char *argv[]) set_addr(&ss, argv[3], argv[4], &len); ret = connect(csk, (struct sockaddr *)&ss, len); - if (ret < 0) { - printf("failed to connect to peer\n"); + if (ret < 0) return -1; - } - ret = send(csk, buf, strlen(buf) + 1, 0); - if (ret < 0) { - printf("failed to send msg %d\n", ret); - return -1; - } + recv(csk, NULL, 0, 0); close(csk); return 0; @@ -75,7 +68,6 @@ int main(int argc, char *argv[]) { struct sockaddr_storage ss; int lsk, csk, ret, len; - char buf[20]; if (argc < 2 || (strcmp(argv[1], "server") && strcmp(argv[1], "client"))) { printf("%s server|client ...\n", argv[0]); @@ -125,11 +117,6 @@ int main(int argc, char *argv[]) return -1; } - ret = recv(csk, buf, sizeof(buf), 0); - if (ret <= 0) { - printf("failed to recv msg %d\n", ret); - return -1; - } close(csk); close(lsk); diff --git a/tools/testing/selftests/net/sctp_vrf.sh b/tools/testing/selftests/net/sctp_vrf.sh index c854034b6aa1..667b211aa8a1 100755 --- a/tools/testing/selftests/net/sctp_vrf.sh +++ b/tools/testing/selftests/net/sctp_vrf.sh @@ -20,9 +20,9 @@ setup() { modprobe sctp_diag setup_ns CLIENT_NS1 CLIENT_NS2 SERVER_NS - ip net exec $CLIENT_NS1 sysctl -w net.ipv6.conf.default.accept_dad=0 2>&1 >/dev/null - ip net exec $CLIENT_NS2 sysctl -w net.ipv6.conf.default.accept_dad=0 2>&1 >/dev/null - ip net exec $SERVER_NS sysctl -w net.ipv6.conf.default.accept_dad=0 2>&1 >/dev/null + ip net exec $CLIENT_NS1 sysctl -wq net.ipv6.conf.default.accept_dad=0 + ip net exec $CLIENT_NS2 sysctl -wq net.ipv6.conf.default.accept_dad=0 + ip net exec $SERVER_NS sysctl -wq net.ipv6.conf.default.accept_dad=0 ip -n $SERVER_NS link add veth1 type veth peer name veth1 netns $CLIENT_NS1 ip -n $SERVER_NS link add veth2 type veth peer name veth1 netns $CLIENT_NS2 @@ -62,17 +62,40 @@ setup() { } cleanup() { - ip netns exec $SERVER_NS pkill sctp_hello 2>&1 >/dev/null + wait_client $CLIENT_NS1 + wait_client $CLIENT_NS2 + stop_server cleanup_ns $CLIENT_NS1 $CLIENT_NS2 $SERVER_NS } -wait_server() { +start_server() { local IFACE=$1 local CNT=0 - until ip netns exec $SERVER_NS ss -lS src $SERVER_IP:$SERVER_PORT | \ - grep LISTEN | grep "$IFACE" 2>&1 >/dev/null; do - [ $((CNT++)) = "20" ] && { RET=3; return $RET; } + ip netns exec $SERVER_NS ./sctp_hello server $AF $SERVER_IP $SERVER_PORT $IFACE & + disown + until ip netns exec $SERVER_NS ss -SlH | grep -q "$IFACE"; do + [ $((CNT++)) -eq 30 ] && { RET=3; return $RET; } + sleep 0.1 + done +} + +stop_server() { + local CNT=0 + + ip netns exec $SERVER_NS pkill sctp_hello + while ip netns exec $SERVER_NS ss -SaH | grep -q .; do + [ $((CNT++)) -eq 30 ] && break + sleep 0.1 + done +} + +wait_client() { + local CLIENT_NS=$1 + local CNT=0 + + while ip netns exec $CLIENT_NS ss -SaH | grep -q .; do + [ $((CNT++)) -eq 30 ] && break sleep 0.1 done } @@ -81,14 +104,12 @@ do_test() { local CLIENT_NS=$1 local IFACE=$2 - ip netns exec $SERVER_NS pkill sctp_hello 2>&1 >/dev/null - ip netns exec $SERVER_NS ./sctp_hello server $AF $SERVER_IP \ - $SERVER_PORT $IFACE 2>&1 >/dev/null & - disown - wait_server $IFACE || return $RET + start_server $IFACE || return $RET timeout 3 ip netns exec $CLIENT_NS ./sctp_hello client $AF \ - $SERVER_IP $SERVER_PORT $CLIENT_IP $CLIENT_PORT 2>&1 >/dev/null + $SERVER_IP $SERVER_PORT $CLIENT_IP $CLIENT_PORT RET=$? + wait_client $CLIENT_NS + stop_server return $RET } @@ -96,25 +117,21 @@ do_testx() { local IFACE1=$1 local IFACE2=$2 - ip netns exec $SERVER_NS pkill sctp_hello 2>&1 >/dev/null - ip netns exec $SERVER_NS ./sctp_hello server $AF $SERVER_IP \ - $SERVER_PORT $IFACE1 2>&1 >/dev/null & - disown - wait_server $IFACE1 || return $RET - ip netns exec $SERVER_NS ./sctp_hello server $AF $SERVER_IP \ - $SERVER_PORT $IFACE2 2>&1 >/dev/null & - disown - wait_server $IFACE2 || return $RET + start_server $IFACE1 || return $RET + start_server $IFACE2 || return $RET timeout 3 ip netns exec $CLIENT_NS1 ./sctp_hello client $AF \ - $SERVER_IP $SERVER_PORT $CLIENT_IP $CLIENT_PORT 2>&1 >/dev/null && \ + $SERVER_IP $SERVER_PORT $CLIENT_IP $CLIENT_PORT && \ timeout 3 ip netns exec $CLIENT_NS2 ./sctp_hello client $AF \ - $SERVER_IP $SERVER_PORT $CLIENT_IP $CLIENT_PORT 2>&1 >/dev/null + $SERVER_IP $SERVER_PORT $CLIENT_IP $CLIENT_PORT RET=$? + wait_client $CLIENT_NS1 + wait_client $CLIENT_NS2 + stop_server return $RET } testup() { - ip netns exec $SERVER_NS sysctl -w net.sctp.l3mdev_accept=1 2>&1 >/dev/null + ip netns exec $SERVER_NS sysctl -wq net.sctp.l3mdev_accept=1 echo -n "TEST 01: nobind, connect from client 1, l3mdev_accept=1, Y " do_test $CLIENT_NS1 || { echo "[FAIL]"; return $RET; } echo "[PASS]" @@ -123,7 +140,7 @@ testup() { do_test $CLIENT_NS2 && { echo "[FAIL]"; return $RET; } echo "[PASS]" - ip netns exec $SERVER_NS sysctl -w net.sctp.l3mdev_accept=0 2>&1 >/dev/null + ip netns exec $SERVER_NS sysctl -wq net.sctp.l3mdev_accept=0 echo -n "TEST 03: nobind, connect from client 1, l3mdev_accept=0, N " do_test $CLIENT_NS1 && { echo "[FAIL]"; return $RET; } echo "[PASS]" @@ -160,7 +177,7 @@ testup() { do_testx vrf-1 vrf-2 || { echo "[FAIL]"; return $RET; } echo "[PASS]" - echo -n "TEST 12: bind vrf-2 & 1 in server, connect from client 1 & 2, N " + echo -n "TEST 12: bind vrf-2 & 1 in server, connect from client 1 & 2, Y " do_testx vrf-2 vrf-1 || { echo "[FAIL]"; return $RET; } echo "[PASS]" } From afd5ba577c10639f62e8120df67dc70ea4b61176 Mon Sep 17 00:00:00 2001 From: Amery Hung Date: Thu, 16 Oct 2025 22:55:39 +0300 Subject: [PATCH 288/330] net/mlx5e: RX, Fix generating skb from non-linear xdp_buff for legacy RQ XDP programs can release xdp_buff fragments when calling bpf_xdp_adjust_tail(). The driver currently assumes the number of fragments to be unchanged and may generate skb with wrong truesize or containing invalid frags. Fix the bug by generating skb according to xdp_buff after the XDP program runs. Fixes: ea5d49bdae8b ("net/mlx5e: Add XDP multi buffer support to the non-linear legacy RQ") Reviewed-by: Dragos Tatulea Signed-off-by: Amery Hung Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/1760644540-899148-2-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- .../net/ethernet/mellanox/mlx5/core/en_rx.c | 25 ++++++++++++++----- 1 file changed, 19 insertions(+), 6 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c index 263d5628ee44..17cab14b328b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c @@ -1794,14 +1794,27 @@ mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5e_wqe_frag_info *wi } prog = rcu_dereference(rq->xdp_prog); - if (prog && mlx5e_xdp_handle(rq, prog, mxbuf)) { - if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) { - struct mlx5e_wqe_frag_info *pwi; + if (prog) { + u8 nr_frags_free, old_nr_frags = sinfo->nr_frags; - for (pwi = head_wi; pwi < wi; pwi++) - pwi->frag_page->frags++; + if (mlx5e_xdp_handle(rq, prog, mxbuf)) { + if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, + rq->flags)) { + struct mlx5e_wqe_frag_info *pwi; + + wi -= old_nr_frags - sinfo->nr_frags; + + for (pwi = head_wi; pwi < wi; pwi++) + pwi->frag_page->frags++; + } + return NULL; /* page/packet was consumed by XDP */ + } + + nr_frags_free = old_nr_frags - sinfo->nr_frags; + if (unlikely(nr_frags_free)) { + wi -= nr_frags_free; + truesize -= nr_frags_free * frag_info->frag_stride; } - return NULL; /* page/packet was consumed by XDP */ } skb = mlx5e_build_linear_skb( From 87bcef158ac1faca1bd7e0104588e8e2956d10be Mon Sep 17 00:00:00 2001 From: Amery Hung Date: Thu, 16 Oct 2025 22:55:40 +0300 Subject: [PATCH 289/330] net/mlx5e: RX, Fix generating skb from non-linear xdp_buff for striding RQ XDP programs can change the layout of an xdp_buff through bpf_xdp_adjust_tail() and bpf_xdp_adjust_head(). Therefore, the driver cannot assume the size of the linear data area nor fragments. Fix the bug in mlx5 by generating skb according to xdp_buff after XDP programs run. Currently, when handling multi-buf XDP, the mlx5 driver assumes the layout of an xdp_buff to be unchanged. That is, the linear data area continues to be empty and fragments remain the same. This may cause the driver to generate erroneous skb or triggering a kernel warning. When an XDP program added linear data through bpf_xdp_adjust_head(), the linear data will be ignored as mlx5e_build_linear_skb() builds an skb without linear data and then pull data from fragments to fill the linear data area. When an XDP program has shrunk the non-linear data through bpf_xdp_adjust_tail(), the delta passed to __pskb_pull_tail() may exceed the actual nonlinear data size and trigger the BUG_ON in it. To fix the issue, first record the original number of fragments. If the number of fragments changes after the XDP program runs, rewind the end fragment pointer by the difference and recalculate the truesize. Then, build the skb with the linear data area matching the xdp_buff. Finally, only pull data in if there is non-linear data and fill the linear part up to 256 bytes. Fixes: f52ac7028bec ("net/mlx5e: RX, Add XDP multi-buffer support in Striding RQ") Signed-off-by: Amery Hung Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/1760644540-899148-3-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- .../net/ethernet/mellanox/mlx5/core/en_rx.c | 26 ++++++++++++++++--- 1 file changed, 23 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c index 17cab14b328b..1c79adc51a04 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c @@ -2040,6 +2040,7 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w u32 byte_cnt = cqe_bcnt; struct skb_shared_info *sinfo; unsigned int truesize = 0; + u32 pg_consumed_bytes; struct bpf_prog *prog; struct sk_buff *skb; u32 linear_frame_sz; @@ -2093,7 +2094,8 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w while (byte_cnt) { /* Non-linear mode, hence non-XSK, which always uses PAGE_SIZE. */ - u32 pg_consumed_bytes = min_t(u32, PAGE_SIZE - frag_offset, byte_cnt); + pg_consumed_bytes = + min_t(u32, PAGE_SIZE - frag_offset, byte_cnt); if (test_bit(MLX5E_RQ_STATE_SHAMPO, &rq->state)) truesize += pg_consumed_bytes; @@ -2109,10 +2111,15 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w } if (prog) { + u8 nr_frags_free, old_nr_frags = sinfo->nr_frags; + u32 len; + if (mlx5e_xdp_handle(rq, prog, mxbuf)) { if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) { struct mlx5e_frag_page *pfp; + frag_page -= old_nr_frags - sinfo->nr_frags; + for (pfp = head_page; pfp < frag_page; pfp++) pfp->frags++; @@ -2123,9 +2130,19 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w return NULL; /* page/packet was consumed by XDP */ } + nr_frags_free = old_nr_frags - sinfo->nr_frags; + if (unlikely(nr_frags_free)) { + frag_page -= nr_frags_free; + truesize -= (nr_frags_free - 1) * PAGE_SIZE + + ALIGN(pg_consumed_bytes, + BIT(rq->mpwqe.log_stride_sz)); + } + + len = mxbuf->xdp.data_end - mxbuf->xdp.data; + skb = mlx5e_build_linear_skb( rq, mxbuf->xdp.data_hard_start, linear_frame_sz, - mxbuf->xdp.data - mxbuf->xdp.data_hard_start, 0, + mxbuf->xdp.data - mxbuf->xdp.data_hard_start, len, mxbuf->xdp.data - mxbuf->xdp.data_meta); if (unlikely(!skb)) { mlx5e_page_release_fragmented(rq->page_pool, @@ -2150,8 +2167,11 @@ mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *w do pagep->frags++; while (++pagep < frag_page); + + headlen = min_t(u16, MLX5E_RX_MAX_HEAD - len, + skb->data_len); + __pskb_pull_tail(skb, headlen); } - __pskb_pull_tail(skb, headlen); } else { dma_addr_t addr; From f584239a9ed25057496bf397c370cc5163dde419 Mon Sep 17 00:00:00 2001 From: Wang Liang Date: Fri, 17 Oct 2025 10:48:27 +0800 Subject: [PATCH 290/330] net/smc: fix general protection fault in __smc_diag_dump The syzbot report a crash: Oops: general protection fault, probably for non-canonical address 0xfbd5a5d5a0000003: 0000 [#1] SMP KASAN NOPTI KASAN: maybe wild-memory-access in range [0xdead4ead00000018-0xdead4ead0000001f] CPU: 1 UID: 0 PID: 6949 Comm: syz.0.335 Not tainted syzkaller #0 PREEMPT(full) Hardware name: Google Compute Engine/Google Compute Engine, BIOS Google 08/18/2025 RIP: 0010:smc_diag_msg_common_fill net/smc/smc_diag.c:44 [inline] RIP: 0010:__smc_diag_dump.constprop.0+0x3ca/0x2550 net/smc/smc_diag.c:89 Call Trace: smc_diag_dump_proto+0x26d/0x420 net/smc/smc_diag.c:217 smc_diag_dump+0x27/0x90 net/smc/smc_diag.c:234 netlink_dump+0x539/0xd30 net/netlink/af_netlink.c:2327 __netlink_dump_start+0x6d6/0x990 net/netlink/af_netlink.c:2442 netlink_dump_start include/linux/netlink.h:341 [inline] smc_diag_handler_dump+0x1f9/0x240 net/smc/smc_diag.c:251 __sock_diag_cmd net/core/sock_diag.c:249 [inline] sock_diag_rcv_msg+0x438/0x790 net/core/sock_diag.c:285 netlink_rcv_skb+0x158/0x420 net/netlink/af_netlink.c:2552 netlink_unicast_kernel net/netlink/af_netlink.c:1320 [inline] netlink_unicast+0x5a7/0x870 net/netlink/af_netlink.c:1346 netlink_sendmsg+0x8d1/0xdd0 net/netlink/af_netlink.c:1896 sock_sendmsg_nosec net/socket.c:714 [inline] __sock_sendmsg net/socket.c:729 [inline] ____sys_sendmsg+0xa95/0xc70 net/socket.c:2614 ___sys_sendmsg+0x134/0x1d0 net/socket.c:2668 __sys_sendmsg+0x16d/0x220 net/socket.c:2700 do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline] do_syscall_64+0xcd/0x4e0 arch/x86/entry/syscall_64.c:94 entry_SYSCALL_64_after_hwframe+0x77/0x7f The process like this: (CPU1) | (CPU2) ---------------------------------|------------------------------- inet_create() | // init clcsock to NULL | sk = sk_alloc() | | // unexpectedly change clcsock | inet_init_csk_locks() | | // add sk to hash table | smc_inet_init_sock() | smc_sk_init() | smc_hash_sk() | | // traverse the hash table | smc_diag_dump_proto | __smc_diag_dump() | // visit wrong clcsock | smc_diag_msg_common_fill() // alloc clcsock | smc_create_clcsk | sock_create_kern | With CONFIG_DEBUG_LOCK_ALLOC=y, the smc->clcsock is unexpectedly changed in inet_init_csk_locks(). The INET_PROTOSW_ICSK flag is no need by smc, just remove it. After removing the INET_PROTOSW_ICSK flag, this patch alse revert commit 6fd27ea183c2 ("net/smc: fix lacks of icsk_syn_mss with IPPROTO_SMC") to avoid casting smc_sock to inet_connection_sock. Reported-by: syzbot+f775be4458668f7d220e@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=f775be4458668f7d220e Tested-by: syzbot+f775be4458668f7d220e@syzkaller.appspotmail.com Fixes: d25a92ccae6b ("net/smc: Introduce IPPROTO_SMC") Signed-off-by: Wang Liang Reviewed-by: Kuniyuki Iwashima Reviewed-by: Eric Dumazet Reviewed-by: D. Wythe Link: https://patch.msgid.link/20251017024827.3137512-1-wangliang74@huawei.com Signed-off-by: Jakub Kicinski --- net/smc/smc_inet.c | 13 ------------- 1 file changed, 13 deletions(-) diff --git a/net/smc/smc_inet.c b/net/smc/smc_inet.c index a944e7dcb8b9..a94084b4a498 100644 --- a/net/smc/smc_inet.c +++ b/net/smc/smc_inet.c @@ -56,7 +56,6 @@ static struct inet_protosw smc_inet_protosw = { .protocol = IPPROTO_SMC, .prot = &smc_inet_prot, .ops = &smc_inet_stream_ops, - .flags = INET_PROTOSW_ICSK, }; #if IS_ENABLED(CONFIG_IPV6) @@ -104,27 +103,15 @@ static struct inet_protosw smc_inet6_protosw = { .protocol = IPPROTO_SMC, .prot = &smc_inet6_prot, .ops = &smc_inet6_stream_ops, - .flags = INET_PROTOSW_ICSK, }; #endif /* CONFIG_IPV6 */ -static unsigned int smc_sync_mss(struct sock *sk, u32 pmtu) -{ - /* No need pass it through to clcsock, mss can always be set by - * sock_create_kern or smc_setsockopt. - */ - return 0; -} - static int smc_inet_init_sock(struct sock *sk) { struct net *net = sock_net(sk); /* init common smc sock */ smc_sk_init(net, sk, IPPROTO_SMC); - - inet_csk(sk)->icsk_sync_mss = smc_sync_mss; - /* create clcsock */ return smc_create_clcsk(net, sk, sk->sk_family); } From 49d34f3dd8519581030547eb7543a62f9ab5fa08 Mon Sep 17 00:00:00 2001 From: Aksh Garg Date: Thu, 16 Oct 2025 17:27:55 +0530 Subject: [PATCH 291/330] net: ethernet: ti: am65-cpts: fix timestamp loss due to race conditions Resolve race conditions in timestamp events list handling between TX and RX paths causing missed timestamps. The current implementation uses a single events list for both TX and RX timestamps. The am65_cpts_find_ts() function acquires the lock, splices all events (TX as well as RX events) to a temporary list, and releases the lock. This function performs matching of timestamps for TX packets only. Before it acquires the lock again to put the non-TX events back to the main events list, a concurrent RX processing thread could acquire the lock (as observed in practice), find an empty events list, and fail to attach timestamp to it, even though a relevant event exists in the spliced list which is yet to be restored to the main list. Fix this by creating separate events lists to handle TX and RX timestamps independently. Fixes: c459f606f66df ("net: ethernet: ti: am65-cpts: Enable RX HW timestamp for PTP packets using CPTS FIFO") Signed-off-by: Aksh Garg Reviewed-by: Siddharth Vadapalli Link: https://patch.msgid.link/20251016115755.1123646-1-a-garg7@ti.com Signed-off-by: Paolo Abeni --- drivers/net/ethernet/ti/am65-cpts.c | 63 ++++++++++++++++++++--------- 1 file changed, 43 insertions(+), 20 deletions(-) diff --git a/drivers/net/ethernet/ti/am65-cpts.c b/drivers/net/ethernet/ti/am65-cpts.c index 59d6ab989c55..8ffbfaa3ab18 100644 --- a/drivers/net/ethernet/ti/am65-cpts.c +++ b/drivers/net/ethernet/ti/am65-cpts.c @@ -163,7 +163,9 @@ struct am65_cpts { struct device_node *clk_mux_np; struct clk *refclk; u32 refclk_freq; - struct list_head events; + /* separate lists to handle TX and RX timestamp independently */ + struct list_head events_tx; + struct list_head events_rx; struct list_head pool; struct am65_cpts_event pool_data[AM65_CPTS_MAX_EVENTS]; spinlock_t lock; /* protects events lists*/ @@ -227,6 +229,24 @@ static void am65_cpts_disable(struct am65_cpts *cpts) am65_cpts_write32(cpts, 0, int_enable); } +static int am65_cpts_purge_event_list(struct am65_cpts *cpts, + struct list_head *events) +{ + struct list_head *this, *next; + struct am65_cpts_event *event; + int removed = 0; + + list_for_each_safe(this, next, events) { + event = list_entry(this, struct am65_cpts_event, list); + if (time_after(jiffies, event->tmo)) { + list_del_init(&event->list); + list_add(&event->list, &cpts->pool); + ++removed; + } + } + return removed; +} + static int am65_cpts_event_get_port(struct am65_cpts_event *event) { return (event->event1 & AM65_CPTS_EVENT_1_PORT_NUMBER_MASK) >> @@ -239,20 +259,12 @@ static int am65_cpts_event_get_type(struct am65_cpts_event *event) AM65_CPTS_EVENT_1_EVENT_TYPE_SHIFT; } -static int am65_cpts_cpts_purge_events(struct am65_cpts *cpts) +static int am65_cpts_purge_events(struct am65_cpts *cpts) { - struct list_head *this, *next; - struct am65_cpts_event *event; int removed = 0; - list_for_each_safe(this, next, &cpts->events) { - event = list_entry(this, struct am65_cpts_event, list); - if (time_after(jiffies, event->tmo)) { - list_del_init(&event->list); - list_add(&event->list, &cpts->pool); - ++removed; - } - } + removed += am65_cpts_purge_event_list(cpts, &cpts->events_tx); + removed += am65_cpts_purge_event_list(cpts, &cpts->events_rx); if (removed) dev_dbg(cpts->dev, "event pool cleaned up %d\n", removed); @@ -287,7 +299,7 @@ static int __am65_cpts_fifo_read(struct am65_cpts *cpts) struct am65_cpts_event, list); if (!event) { - if (am65_cpts_cpts_purge_events(cpts)) { + if (am65_cpts_purge_events(cpts)) { dev_err(cpts->dev, "cpts: event pool empty\n"); ret = -1; goto out; @@ -306,11 +318,21 @@ static int __am65_cpts_fifo_read(struct am65_cpts *cpts) cpts->timestamp); break; case AM65_CPTS_EV_RX: + event->tmo = jiffies + + msecs_to_jiffies(AM65_CPTS_EVENT_RX_TX_TIMEOUT); + + list_move_tail(&event->list, &cpts->events_rx); + + dev_dbg(cpts->dev, + "AM65_CPTS_EV_RX e1:%08x e2:%08x t:%lld\n", + event->event1, event->event2, + event->timestamp); + break; case AM65_CPTS_EV_TX: event->tmo = jiffies + msecs_to_jiffies(AM65_CPTS_EVENT_RX_TX_TIMEOUT); - list_move_tail(&event->list, &cpts->events); + list_move_tail(&event->list, &cpts->events_tx); dev_dbg(cpts->dev, "AM65_CPTS_EV_TX e1:%08x e2:%08x t:%lld\n", @@ -828,7 +850,7 @@ static bool am65_cpts_match_tx_ts(struct am65_cpts *cpts, return found; } -static void am65_cpts_find_ts(struct am65_cpts *cpts) +static void am65_cpts_find_tx_ts(struct am65_cpts *cpts) { struct am65_cpts_event *event; struct list_head *this, *next; @@ -837,7 +859,7 @@ static void am65_cpts_find_ts(struct am65_cpts *cpts) LIST_HEAD(events); spin_lock_irqsave(&cpts->lock, flags); - list_splice_init(&cpts->events, &events); + list_splice_init(&cpts->events_tx, &events); spin_unlock_irqrestore(&cpts->lock, flags); list_for_each_safe(this, next, &events) { @@ -850,7 +872,7 @@ static void am65_cpts_find_ts(struct am65_cpts *cpts) } spin_lock_irqsave(&cpts->lock, flags); - list_splice_tail(&events, &cpts->events); + list_splice_tail(&events, &cpts->events_tx); list_splice_tail(&events_free, &cpts->pool); spin_unlock_irqrestore(&cpts->lock, flags); } @@ -861,7 +883,7 @@ static long am65_cpts_ts_work(struct ptp_clock_info *ptp) unsigned long flags; long delay = -1; - am65_cpts_find_ts(cpts); + am65_cpts_find_tx_ts(cpts); spin_lock_irqsave(&cpts->txq.lock, flags); if (!skb_queue_empty(&cpts->txq)) @@ -905,7 +927,7 @@ static u64 am65_cpts_find_rx_ts(struct am65_cpts *cpts, u32 skb_mtype_seqid) spin_lock_irqsave(&cpts->lock, flags); __am65_cpts_fifo_read(cpts); - list_for_each_safe(this, next, &cpts->events) { + list_for_each_safe(this, next, &cpts->events_rx) { event = list_entry(this, struct am65_cpts_event, list); if (time_after(jiffies, event->tmo)) { list_move(&event->list, &cpts->pool); @@ -1155,7 +1177,8 @@ struct am65_cpts *am65_cpts_create(struct device *dev, void __iomem *regs, return ERR_PTR(ret); mutex_init(&cpts->ptp_clk_lock); - INIT_LIST_HEAD(&cpts->events); + INIT_LIST_HEAD(&cpts->events_tx); + INIT_LIST_HEAD(&cpts->events_rx); INIT_LIST_HEAD(&cpts->pool); spin_lock_init(&cpts->lock); skb_queue_head_init(&cpts->txq); From 6e3a4754717a74e931a9f00b5f953be708e07acb Mon Sep 17 00:00:00 2001 From: Xi Ruoyao Date: Tue, 21 Oct 2025 17:28:25 +0800 Subject: [PATCH 292/330] ACPICA: Work around bogus -Wstringop-overread warning since GCC 11 When ACPI_MISALIGNMENT_NOT_SUPPORTED is set, GCC can produce a bogus -Wstringop-overread warning, see [1]. To me, it's very clear that we have a compiler bug here, thus just disable the warning. Fixes: a9d13433fe17 ("LoongArch: Align ACPI structures if ARCH_STRICT_ALIGN enabled") Link: https://lore.kernel.org/all/899f2dec-e8b9-44f4-ab8d-001e160a2aed@roeck-us.net/ Link: https://github.com/acpica/acpica/commit/abf5b573 Link: https://gcc.gnu.org/PR122073 [1] Co-developed-by: Saket Dumbre Signed-off-by: Saket Dumbre Signed-off-by: Xi Ruoyao Acked-by: Huacai Chen Cc: All applicable [ rjw: Subject and changelog edits ] Link: https://patch.msgid.link/20251021092825.822007-1-xry111@xry111.site Signed-off-by: Rafael J. Wysocki --- drivers/acpi/acpica/tbprint.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/acpi/acpica/tbprint.c b/drivers/acpi/acpica/tbprint.c index 049f6c2f1e32..e5631027f7f1 100644 --- a/drivers/acpi/acpica/tbprint.c +++ b/drivers/acpi/acpica/tbprint.c @@ -95,6 +95,11 @@ acpi_tb_print_table_header(acpi_physical_address address, { struct acpi_table_header local_header; +#pragma GCC diagnostic push +#if defined(__GNUC__) && __GNUC__ >= 11 +#pragma GCC diagnostic ignored "-Wstringop-overread" +#endif + if (ACPI_COMPARE_NAMESEG(header->signature, ACPI_SIG_FACS)) { /* FACS only has signature and length fields */ @@ -143,4 +148,5 @@ acpi_tb_print_table_header(acpi_physical_address address, local_header.asl_compiler_id, local_header.asl_compiler_revision)); } +#pragma GCC diagnostic pop } From 0e59f47c15cec4cd88c51c5cda749607b719c82b Mon Sep 17 00:00:00 2001 From: Lorenzo Stoakes Date: Mon, 13 Oct 2025 17:58:36 +0100 Subject: [PATCH 293/330] mm/mremap: correctly account old mapping after MREMAP_DONTUNMAP remap Commit b714ccb02a76 ("mm/mremap: complete refactor of move_vma()") mistakenly introduced a new behaviour - clearing the VM_ACCOUNT flag of the old mapping when a mapping is mremap()'d with the MREMAP_DONTUNMAP flag set. While we always clear the VM_LOCKED and VM_LOCKONFAULT flags for the old mapping (the page tables have been moved, so there is no data that could possibly be locked in memory), there is no reason to touch any other VMA flags. This is because after the move the old mapping is in a state as if it were freshly mapped. This implies that the attributes of the mapping ought to remain the same, including whether or not the mapping is accounted. Link: https://lkml.kernel.org/r/20251013165836.273113-1-lorenzo.stoakes@oracle.com Signed-off-by: Lorenzo Stoakes Fixes: b714ccb02a76 ("mm/mremap: complete refactor of move_vma()") Reviewed-by: Pedro Falcato Cc: Jann Horn Cc: Liam Howlett Cc: Vlastimil Babka Cc: Signed-off-by: Andrew Morton --- mm/mremap.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) diff --git a/mm/mremap.c b/mm/mremap.c index 35de0a7b910e..bd7314898ec5 100644 --- a/mm/mremap.c +++ b/mm/mremap.c @@ -1237,10 +1237,10 @@ static int copy_vma_and_data(struct vma_remap_struct *vrm, } /* - * Perform final tasks for MADV_DONTUNMAP operation, clearing mlock() and - * account flags on remaining VMA by convention (it cannot be mlock()'d any - * longer, as pages in range are no longer mapped), and removing anon_vma_chain - * links from it (if the entire VMA was copied over). + * Perform final tasks for MADV_DONTUNMAP operation, clearing mlock() flag on + * remaining VMA by convention (it cannot be mlock()'d any longer, as pages in + * range are no longer mapped), and removing anon_vma_chain links from it if the + * entire VMA was copied over. */ static void dontunmap_complete(struct vma_remap_struct *vrm, struct vm_area_struct *new_vma) @@ -1250,11 +1250,8 @@ static void dontunmap_complete(struct vma_remap_struct *vrm, unsigned long old_start = vrm->vma->vm_start; unsigned long old_end = vrm->vma->vm_end; - /* - * We always clear VM_LOCKED[ONFAULT] | VM_ACCOUNT on the old - * vma. - */ - vm_flags_clear(vrm->vma, VM_LOCKED_MASK | VM_ACCOUNT); + /* We always clear VM_LOCKED[ONFAULT] on the old VMA. */ + vm_flags_clear(vrm->vma, VM_LOCKED_MASK); /* * anon_vma links of the old vma is no longer needed after its page From c3fa5b1bfd8380d935fa961f2ac166bdf000f418 Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Tue, 14 Oct 2025 13:59:36 -0700 Subject: [PATCH 294/330] mm/damon/core: fix list_add_tail() call on damon_call() Each damon_ctx maintains callback requests using a linked list (damon_ctx->call_controls). When a new callback request is received via damon_call(), the new request should be added to the list. However, the function is making a mistake at list_add_tail() invocation: putting the new item to add and the list head to add it before, in the opposite order. Because of the linked list manipulation implementation, the new request can still be reached from the context's list head. But the list items that were added before the new request are dropped from the list. As a result, the callbacks are unexpectedly not invocated. Worse yet, if the dropped callback requests were dynamically allocated, the memory is leaked. Actually DAMON sysfs interface is using a dynamically allocated repeat-mode callback request for automatic essential stats update. And because the online DAMON parameters commit is using a non-repeat-mode callback request, the issue can easily be reproduced, like below. # damo start --damos_action stat --refresh_stat 1s # damo tune --damos_action stat --refresh_stat 1s The first command dynamically allocates the repeat-mode callback request for automatic essential stat update. Users can see the essential stats are automatically updated for every second, using the sysfs interface. The second command calls damon_commit() with a new callback request that was made for the commit. As a result, the previously added repeat-mode callback request is dropped from the list. The automatic stats refresh stops working, and the memory for the repeat-mode callback request is leaked. It can be confirmed using kmemleak. Fix the mistake on the list_add_tail() call. Link: https://lkml.kernel.org/r/20251014205939.1206-1-sj@kernel.org Fixes: 004ded6bee11 ("mm/damon: accept parallel damon_call() requests") Signed-off-by: SeongJae Park Cc: [6.17+] Signed-off-by: Andrew Morton --- mm/damon/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/damon/core.c b/mm/damon/core.c index 93848b4c6944..4670d293bbf4 100644 --- a/mm/damon/core.c +++ b/mm/damon/core.c @@ -1450,7 +1450,7 @@ int damon_call(struct damon_ctx *ctx, struct damon_call_control *control) INIT_LIST_HEAD(&control->list); mutex_lock(&ctx->call_controls_lock); - list_add_tail(&ctx->call_controls, &control->list); + list_add_tail(&control->list, &ctx->call_controls); mutex_unlock(&ctx->call_controls_lock); if (!damon_is_running(ctx)) return -EINVAL; From 4ba5a8a7faa647ada8eae61a36517cf369f5bbe4 Mon Sep 17 00:00:00 2001 From: David Hildenbrand Date: Tue, 14 Oct 2025 14:44:55 +0200 Subject: [PATCH 295/330] vmw_balloon: indicate success when effectively deflating during migration When migrating a balloon page, we first deflate the old page to then inflate the new page. However, if inflating the new page succeeded, we effectively deflated the old page, reducing the balloon size. In that case, the migration actually worked: similar to migrating+ immediately deflating the new page. The old page will be freed back to the buddy. Right now, the core will leave the page be marked as isolated (as we returned an error). When later trying to putback that page, we will run into the WARN_ON_ONCE() in balloon_page_putback(). That handling was changed in commit 3544c4faccb8 ("mm/balloon_compaction: stop using __ClearPageMovable()"); before that change, we would have tolerated that way of handling it. To fix it, let's just return 0 in that case, making the core effectively just clear the "isolated" flag + freeing it back to the buddy as if the migration succeeded. Note that the new page will also get freed when the core puts the last reference. Note that this also makes it all be more consistent: we will no longer unisolate the page in the balloon driver while keeping it marked as being isolated in migration core. This was found by code inspection. Link: https://lkml.kernel.org/r/20251014124455.478345-1-david@redhat.com Fixes: 3544c4faccb8 ("mm/balloon_compaction: stop using __ClearPageMovable()") Signed-off-by: David Hildenbrand Cc: Jerrin Shaji George Cc: Broadcom internal kernel review list Cc: Arnd Bergmann Cc: Greg Kroah-Hartman Cc: Signed-off-by: Andrew Morton --- drivers/misc/vmw_balloon.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/misc/vmw_balloon.c b/drivers/misc/vmw_balloon.c index 6df51ee8db62..cc1d18b3df5c 100644 --- a/drivers/misc/vmw_balloon.c +++ b/drivers/misc/vmw_balloon.c @@ -1737,7 +1737,7 @@ static int vmballoon_migratepage(struct balloon_dev_info *b_dev_info, { unsigned long status, flags; struct vmballoon *b; - int ret; + int ret = 0; b = container_of(b_dev_info, struct vmballoon, b_dev_info); @@ -1796,17 +1796,15 @@ static int vmballoon_migratepage(struct balloon_dev_info *b_dev_info, * A failure happened. While we can deflate the page we just * inflated, this deflation can also encounter an error. Instead * we will decrease the size of the balloon to reflect the - * change and report failure. + * change. */ atomic64_dec(&b->size); - ret = -EBUSY; } else { /* * Success. Take a reference for the page, and we will add it to * the list after acquiring the lock. */ get_page(newpage); - ret = 0; } /* Update the balloon list under the @pages_lock */ @@ -1817,7 +1815,7 @@ static int vmballoon_migratepage(struct balloon_dev_info *b_dev_info, * If we succeed just insert it to the list and update the statistics * under the lock. */ - if (!ret) { + if (status == VMW_BALLOON_SUCCESS) { balloon_page_insert(&b->b_dev_info, newpage); __count_vm_event(BALLOON_MIGRATE); } From cec944dd329fbefee907da95c298719d900d4787 Mon Sep 17 00:00:00 2001 From: Deepanshu Kartikey Date: Tue, 14 Oct 2025 17:03:44 +0530 Subject: [PATCH 296/330] hugetlbfs: move lock assertions after early returns in huge_pmd_unshare() When hugetlb_vmdelete_list() processes VMAs during truncate operations, it may encounter VMAs where huge_pmd_unshare() is called without the required shareable lock. This triggers an assertion failure in hugetlb_vma_assert_locked(). The previous fix in commit dd83609b8898 ("hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list") skipped entire VMAs without shareable locks to avoid the assertion. However, this prevented pages from being unmapped and freed, causing a regression in fallocate(PUNCH_HOLE) operations where pages were not freed immediately, as reported by Mark Brown. Instead of checking locks in the caller or skipping VMAs, move the lock assertions in huge_pmd_unshare() to after the early return checks. The assertions are only needed when actual PMD unsharing work will be performed. If the function returns early because sz != PMD_SIZE or the PMD is not shared, no locks are required and assertions should not fire. This approach reverts the VMA skipping logic from commit dd83609b8898 ("hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list") while moving the assertions to avoid the assertion failure, keeping all the logic within huge_pmd_unshare() itself and allowing page unmapping and freeing to proceed for all VMAs. Link: https://lkml.kernel.org/r/20251014113344.21194-1-kartikey406@gmail.com Fixes: dd83609b8898 ("hugetlbfs: skip VMAs without shareable locks in hugetlb_vmdelete_list") Signed-off-by: Deepanshu Kartikey Reported-by: Reported-by: Mark Brown Closes: https://syzkaller.appspot.com/bug?extid=f26d7c75c26ec19790e7 Suggested-by: David Hildenbrand Suggested-by: Oscar Salvador Tested-by: Acked-by: David Hildenbrand Signed-off-by: Andrew Morton --- fs/hugetlbfs/inode.c | 9 --------- mm/hugetlb.c | 5 ++--- 2 files changed, 2 insertions(+), 12 deletions(-) diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index 9c94ed8c3ab0..f42548ee9083 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -478,14 +478,6 @@ hugetlb_vmdelete_list(struct rb_root_cached *root, pgoff_t start, pgoff_t end, if (!hugetlb_vma_trylock_write(vma)) continue; - /* - * Skip VMAs without shareable locks. Per the design in commit - * 40549ba8f8e0, these will be handled by remove_inode_hugepages() - * called after this function with proper locking. - */ - if (!__vma_shareable_lock(vma)) - goto skip; - v_start = vma_offset_start(vma, start); v_end = vma_offset_end(vma, end); @@ -496,7 +488,6 @@ hugetlb_vmdelete_list(struct rb_root_cached *root, pgoff_t start, pgoff_t end, * vmas. Therefore, lock is not held when calling * unmap_hugepage_range for private vmas. */ -skip: hugetlb_vma_unlock_write(vma); } } diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 795ee393eac0..0455119716ec 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -7614,13 +7614,12 @@ int huge_pmd_unshare(struct mm_struct *mm, struct vm_area_struct *vma, p4d_t *p4d = p4d_offset(pgd, addr); pud_t *pud = pud_offset(p4d, addr); - i_mmap_assert_write_locked(vma->vm_file->f_mapping); - hugetlb_vma_assert_locked(vma); if (sz != PMD_SIZE) return 0; if (!ptdesc_pmd_is_shared(virt_to_ptdesc(ptep))) return 0; - + i_mmap_assert_write_locked(vma->vm_file->f_mapping); + hugetlb_vma_assert_locked(vma); pud_clear(pud); /* * Once our caller drops the rmap lock, some other process might be From 7071537159be845a5c4ed5fb7d3db25aa4bd04a3 Mon Sep 17 00:00:00 2001 From: Enze Li Date: Tue, 14 Oct 2025 16:42:25 +0800 Subject: [PATCH 297/330] mm/damon/core: fix potential memory leak by cleaning ops_filter in damon_destroy_scheme Currently, damon_destroy_scheme() only cleans up the filter list but leaves ops_filter untouched, which could lead to memory leaks when a scheme is destroyed. This patch ensures both filter and ops_filter are properly freed in damon_destroy_scheme(), preventing potential memory leaks. Link: https://lkml.kernel.org/r/20251014084225.313313-1-lienze@kylinos.cn Fixes: ab82e57981d0 ("mm/damon/core: introduce damos->ops_filters") Signed-off-by: Enze Li Reviewed-by: SeongJae Park Tested-by: SeongJae Park Cc: Signed-off-by: Andrew Morton --- mm/damon/core.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/mm/damon/core.c b/mm/damon/core.c index 4670d293bbf4..083d314fc4e6 100644 --- a/mm/damon/core.c +++ b/mm/damon/core.c @@ -452,6 +452,9 @@ void damon_destroy_scheme(struct damos *s) damos_for_each_filter_safe(f, next, s) damos_destroy_filter(f); + damos_for_each_ops_filter_safe(f, next, s) + damos_destroy_filter(f); + kfree(s->migrate_dests.node_id_arr); kfree(s->migrate_dests.weight_arr); damon_del_scheme(s); From 7eca961dd7188f20fdf8ce9ed5018280f79b2438 Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Mon, 13 Oct 2025 17:18:44 -0700 Subject: [PATCH 298/330] mm/damon/core: use damos_commit_quota_goal() for new goal commit When damos_commit_quota_goals() is called for adding new DAMOS quota goals of DAMOS_QUOTA_USER_INPUT metric, current_value fields of the new goals should be also set as requested. However, damos_commit_quota_goals() is not updating the field for the case, since it is setting only metrics and target values using damos_new_quota_goal(), and metric-optional union fields using damos_commit_quota_goal_union(). As a result, users could see the first current_value parameter that committed online with a new quota goal is ignored. Users are assumed to commit the current_value for DAMOS_QUOTA_USER_INPUT quota goals, since it is being used as a feedback. Hence the real impact would be subtle. That said, this is obviously not intended behavior. Fix the issue by using damos_commit_quota_goal() which sets all quota goal parameters, instead of damos_commit_quota_goal_union(), which sets only the union fields. Link: https://lkml.kernel.org/r/20251014001846.279282-1-sj@kernel.org Fixes: 1aef9df0ee90 ("mm/damon/core: commit damos_quota_goal->nid") Signed-off-by: SeongJae Park Cc: [6.16+] Signed-off-by: Andrew Morton --- mm/damon/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/damon/core.c b/mm/damon/core.c index 083d314fc4e6..109b050c795a 100644 --- a/mm/damon/core.c +++ b/mm/damon/core.c @@ -835,7 +835,7 @@ int damos_commit_quota_goals(struct damos_quota *dst, struct damos_quota *src) src_goal->metric, src_goal->target_value); if (!new_goal) return -ENOMEM; - damos_commit_quota_goal_union(new_goal, src_goal); + damos_commit_quota_goal(new_goal, src_goal); damos_add_quota_goal(dst, new_goal); } return 0; From 9aa12167ef1149d9980713b120ddcb31cf17222d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Thomas=20Wei=C3=9Fschuh?= Date: Mon, 6 Oct 2025 14:13:37 +0200 Subject: [PATCH 299/330] csky: abiv2: adapt to new folio flags field MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Recent changes require the raw folio flags to be accessed via ".f". The merge commit introducing this change adapted most architecture code but forgot the csky abiv2. [rppt@kernel.org: add fix for arch/csky/abiv2/cacheflush.c] Link: https://lkml.kernel.org/r/aPCE238oxAB9QcZa@kernel.org Fixes: 53fbef56e07d ("mm: introduce memdesc_flags_t") Signed-off-by: Thomas Weißschuh Signed-off-by: Mike Rapoport (Microsoft) Acked-by: Guo Ren Acked-by: Zi Yan Cc: Guo Ren Cc: Matthew Wilcox (Oracle) Signed-off-by: Andrew Morton --- arch/csky/abiv2/cacheflush.c | 2 +- arch/csky/abiv2/inc/abi/cacheflush.h | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/arch/csky/abiv2/cacheflush.c b/arch/csky/abiv2/cacheflush.c index 876028b1083f..064b0f0f95ca 100644 --- a/arch/csky/abiv2/cacheflush.c +++ b/arch/csky/abiv2/cacheflush.c @@ -21,7 +21,7 @@ void update_mmu_cache_range(struct vm_fault *vmf, struct vm_area_struct *vma, folio = page_folio(pfn_to_page(pfn)); - if (test_and_set_bit(PG_dcache_clean, &folio->flags)) + if (test_and_set_bit(PG_dcache_clean, &folio->flags.f)) return; icache_inv_range(address, address + nr*PAGE_SIZE); diff --git a/arch/csky/abiv2/inc/abi/cacheflush.h b/arch/csky/abiv2/inc/abi/cacheflush.h index 6513ac5d2578..da51a0f02391 100644 --- a/arch/csky/abiv2/inc/abi/cacheflush.h +++ b/arch/csky/abiv2/inc/abi/cacheflush.h @@ -20,8 +20,8 @@ static inline void flush_dcache_folio(struct folio *folio) { - if (test_bit(PG_dcache_clean, &folio->flags)) - clear_bit(PG_dcache_clean, &folio->flags); + if (test_bit(PG_dcache_clean, &folio->flags.f)) + clear_bit(PG_dcache_clean, &folio->flags.f); } #define flush_dcache_folio flush_dcache_folio From e13d315ae077bb7c3c6027cc292401bc0f4ec683 Mon Sep 17 00:00:00 2001 From: Gao Xiang Date: Fri, 17 Oct 2025 15:05:38 +0800 Subject: [PATCH 300/330] erofs: avoid infinite loops due to corrupted subpage compact indexes Robert reported an infinite loop observed by two crafted images. The root cause is that `clusterofs` can be larger than `lclustersize` for !NONHEAD `lclusters` in corrupted subpage compact indexes, e.g.: blocksize = lclustersize = 512 lcn = 6 clusterofs = 515 Move the corresponding check for full compress indexes to `z_erofs_load_lcluster_from_disk()` to also cover subpage compact compress indexes. It also fixes the position of `m->type >= Z_EROFS_LCLUSTER_TYPE_MAX` check, since it should be placed right after `z_erofs_load_{compact,full}_lcluster()`. Fixes: 8d2517aaeea3 ("erofs: fix up compacted indexes for block size < 4096") Fixes: 1a5223c182fd ("erofs: do sanity check on m->type in z_erofs_load_compact_lcluster()") Reported-by: Robert Morris Closes: https://lore.kernel.org/r/35167.1760645886@localhost Reviewed-by: Hongbo Li Signed-off-by: Gao Xiang --- fs/erofs/zmap.c | 32 ++++++++++++++++++-------------- 1 file changed, 18 insertions(+), 14 deletions(-) diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c index 8007814f721e..08ec7b0e50a8 100644 --- a/fs/erofs/zmap.c +++ b/fs/erofs/zmap.c @@ -55,10 +55,6 @@ static int z_erofs_load_full_lcluster(struct z_erofs_maprecorder *m, } else { m->partialref = !!(advise & Z_EROFS_LI_PARTIAL_REF); m->clusterofs = le16_to_cpu(di->di_clusterofs); - if (m->clusterofs >= 1 << vi->z_lclusterbits) { - DBG_BUGON(1); - return -EFSCORRUPTED; - } m->pblk = le32_to_cpu(di->di_u.blkaddr); } return 0; @@ -240,21 +236,29 @@ static int z_erofs_load_compact_lcluster(struct z_erofs_maprecorder *m, static int z_erofs_load_lcluster_from_disk(struct z_erofs_maprecorder *m, unsigned int lcn, bool lookahead) { + struct erofs_inode *vi = EROFS_I(m->inode); + int err; + + if (vi->datalayout == EROFS_INODE_COMPRESSED_COMPACT) { + err = z_erofs_load_compact_lcluster(m, lcn, lookahead); + } else { + DBG_BUGON(vi->datalayout != EROFS_INODE_COMPRESSED_FULL); + err = z_erofs_load_full_lcluster(m, lcn); + } + if (err) + return err; + if (m->type >= Z_EROFS_LCLUSTER_TYPE_MAX) { erofs_err(m->inode->i_sb, "unknown type %u @ lcn %u of nid %llu", - m->type, lcn, EROFS_I(m->inode)->nid); + m->type, lcn, EROFS_I(m->inode)->nid); DBG_BUGON(1); return -EOPNOTSUPP; + } else if (m->type != Z_EROFS_LCLUSTER_TYPE_NONHEAD && + m->clusterofs >= (1 << vi->z_lclusterbits)) { + DBG_BUGON(1); + return -EFSCORRUPTED; } - - switch (EROFS_I(m->inode)->datalayout) { - case EROFS_INODE_COMPRESSED_FULL: - return z_erofs_load_full_lcluster(m, lcn); - case EROFS_INODE_COMPRESSED_COMPACT: - return z_erofs_load_compact_lcluster(m, lcn, lookahead); - default: - return -EINVAL; - } + return 0; } static int z_erofs_extent_lookback(struct z_erofs_maprecorder *m, From 2a13fc417f493e28bdd368785320dd4c2b3d732e Mon Sep 17 00:00:00 2001 From: Gao Xiang Date: Fri, 17 Oct 2025 15:05:39 +0800 Subject: [PATCH 301/330] erofs: consolidate z_erofs_extent_lookback() The initial m.delta[0] also needs to be checked against zero. In addition, also drop the redundant logic that errors out for lcn == 0 / m.delta[0] == 1 case. Signed-off-by: Gao Xiang --- fs/erofs/zmap.c | 20 ++++++-------------- 1 file changed, 6 insertions(+), 14 deletions(-) diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c index 08ec7b0e50a8..c8d8e129eb4b 100644 --- a/fs/erofs/zmap.c +++ b/fs/erofs/zmap.c @@ -272,20 +272,19 @@ static int z_erofs_extent_lookback(struct z_erofs_maprecorder *m, unsigned long lcn = m->lcn - lookback_distance; int err; + if (!lookback_distance) + break; + err = z_erofs_load_lcluster_from_disk(m, lcn, false); if (err) return err; - if (m->type == Z_EROFS_LCLUSTER_TYPE_NONHEAD) { lookback_distance = m->delta[0]; - if (!lookback_distance) - break; continue; - } else { - m->headtype = m->type; - m->map->m_la = (lcn << lclusterbits) | m->clusterofs; - return 0; } + m->headtype = m->type; + m->map->m_la = (lcn << lclusterbits) | m->clusterofs; + return 0; } erofs_err(sb, "bogus lookback distance %u @ lcn %lu of nid %llu", lookback_distance, m->lcn, vi->nid); @@ -435,13 +434,6 @@ static int z_erofs_map_blocks_fo(struct inode *inode, end = inode->i_size; } else { if (m.type != Z_EROFS_LCLUSTER_TYPE_NONHEAD) { - /* m.lcn should be >= 1 if endoff < m.clusterofs */ - if (!m.lcn) { - erofs_err(sb, "invalid logical cluster 0 at nid %llu", - vi->nid); - err = -EFSCORRUPTED; - goto unmap_out; - } end = (m.lcn << lclusterbits) | m.clusterofs; map->m_flags |= EROFS_MAP_FULL_MAPPED; m.delta[0] = 1; From e84cb860ac3ce67ec6ecc364433fd5b412c448bc Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Mon, 20 Oct 2025 22:53:26 +0200 Subject: [PATCH 302/330] mptcp: pm: in-kernel: C-flag: handle late ADD_ADDR The special C-flag case expects the ADD_ADDR to be received when switching to 'fully-established'. But for various reasons, the ADD_ADDR could be sent after the "4th ACK", and the special case doesn't work. On NIPA, the new test validating this special case for the C-flag failed a few times, e.g. 102 default limits, server deny join id 0 syn rx [FAIL] got 0 JOIN[s] syn rx expected 2 Server ns stats (...) MPTcpExtAddAddrTx 1 MPTcpExtEchoAdd 1 Client ns stats (...) MPTcpExtAddAddr 1 MPTcpExtEchoAddTx 1 synack rx [FAIL] got 0 JOIN[s] synack rx expected 2 ack rx [FAIL] got 0 JOIN[s] ack rx expected 2 join Rx [FAIL] see above syn tx [FAIL] got 0 JOIN[s] syn tx expected 2 join Tx [FAIL] see above I had a suspicion about what the issue could be: the ADD_ADDR might have been received after the switch to the 'fully-established' state. The issue was not easy to reproduce. The packet capture shown that the ADD_ADDR can indeed be sent with a delay, and the client would not try to establish subflows to it as expected. A simple fix is not to mark the endpoints as 'used' in the C-flag case, when looking at creating subflows to the remote initial IP address and port. In this case, there is no need to try. Note: newly added fullmesh endpoints will still continue to be used as expected, thanks to the conditions behind mptcp_pm_add_addr_c_flag_case. Fixes: 4b1ff850e0c1 ("mptcp: pm: in-kernel: usable client side with C-flag") Cc: stable@vger.kernel.org Reviewed-by: Geliang Tang Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20251020-net-mptcp-c-flag-late-add-addr-v1-1-8207030cb0e8@kernel.org Signed-off-by: Jakub Kicinski --- net/mptcp/pm_kernel.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/mptcp/pm_kernel.c b/net/mptcp/pm_kernel.c index e0f44dc232aa..2ae95476dba3 100644 --- a/net/mptcp/pm_kernel.c +++ b/net/mptcp/pm_kernel.c @@ -370,6 +370,10 @@ static void mptcp_pm_create_subflow_or_signal_addr(struct mptcp_sock *msk) } subflow: + /* No need to try establishing subflows to remote id0 if not allowed */ + if (mptcp_pm_add_addr_c_flag_case(msk)) + goto exit; + /* check if should create a new subflow */ while (msk->pm.local_addr_used < endp_subflow_max && msk->pm.extra_subflows < limit_extra_subflows) { @@ -401,6 +405,8 @@ static void mptcp_pm_create_subflow_or_signal_addr(struct mptcp_sock *msk) __mptcp_subflow_connect(sk, &local, &addrs[i]); spin_lock_bh(&msk->pm.lock); } + +exit: mptcp_pm_nl_check_work_pending(msk); } From d68460bc31f9c8c6fc81fbb56ec952bec18409f1 Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Mon, 20 Oct 2025 22:53:27 +0200 Subject: [PATCH 303/330] selftests: mptcp: join: mark 'flush re-add' as skipped if not supported The call to 'continue_if' was missing: it properly marks a subtest as 'skipped' if the attached condition is not valid. Without that, the test is wrongly marked as passed on older kernels. Fixes: e06959e9eebd ("selftests: mptcp: join: test for flush/re-add endpoints") Cc: stable@vger.kernel.org Reviewed-by: Geliang Tang Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20251020-net-mptcp-c-flag-late-add-addr-v1-2-8207030cb0e8@kernel.org Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/mptcp/mptcp_join.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh index c90d8e8b95cb..deba21ca5a97 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_join.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh @@ -4115,7 +4115,7 @@ endpoint_tests() # flush and re-add if reset_with_tcp_filter "flush re-add" ns2 10.0.3.2 REJECT OUTPUT && - mptcp_lib_kallsyms_has "subflow_rebuild_header$"; then + continue_if mptcp_lib_kallsyms_has "subflow_rebuild_header$"; then pm_nl_set_limits $ns1 0 2 pm_nl_set_limits $ns2 1 2 # broadcast IP: no packet for this address will be received on ns1 From 973f80d715bd2504b4db6e049f292e694145cd79 Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Mon, 20 Oct 2025 22:53:28 +0200 Subject: [PATCH 304/330] selftests: mptcp: join: mark implicit tests as skipped if not supported The call to 'continue_if' was missing: it properly marks a subtest as 'skipped' if the attached condition is not valid. Without that, the test is wrongly marked as passed on older kernels. Fixes: 36c4127ae8dd ("selftests: mptcp: join: skip implicit tests if not supported") Cc: stable@vger.kernel.org Reviewed-by: Geliang Tang Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20251020-net-mptcp-c-flag-late-add-addr-v1-3-8207030cb0e8@kernel.org Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/mptcp/mptcp_join.sh | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh index deba21ca5a97..d98f8f8905b9 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_join.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh @@ -3939,7 +3939,7 @@ endpoint_tests() # subflow_rebuild_header is needed to support the implicit flag # userspace pm type prevents add_addr if reset "implicit EP" && - mptcp_lib_kallsyms_has "subflow_rebuild_header$"; then + continue_if mptcp_lib_kallsyms_has "subflow_rebuild_header$"; then pm_nl_set_limits $ns1 2 2 pm_nl_set_limits $ns2 2 2 pm_nl_add_endpoint $ns1 10.0.2.1 flags signal @@ -3964,7 +3964,7 @@ endpoint_tests() fi if reset_with_tcp_filter "delete and re-add" ns2 10.0.3.2 REJECT OUTPUT && - mptcp_lib_kallsyms_has "subflow_rebuild_header$"; then + continue_if mptcp_lib_kallsyms_has "subflow_rebuild_header$"; then start_events pm_nl_set_limits $ns1 0 3 pm_nl_set_limits $ns2 0 3 From c3496c052ac36ea98ec4f8e95ae6285a425a2457 Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Mon, 20 Oct 2025 22:53:29 +0200 Subject: [PATCH 305/330] selftests: mptcp: join: mark 'delete re-add signal' as skipped if not supported The call to 'continue_if' was missing: it properly marks a subtest as 'skipped' if the attached condition is not valid. Without that, the test is wrongly marked as passed on older kernels. Fixes: b5e2fb832f48 ("selftests: mptcp: add explicit test case for remove/readd") Cc: stable@vger.kernel.org Reviewed-by: Geliang Tang Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20251020-net-mptcp-c-flag-late-add-addr-v1-4-8207030cb0e8@kernel.org Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/mptcp/mptcp_join.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh index d98f8f8905b9..b2a8c51a3969 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_join.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh @@ -4040,7 +4040,7 @@ endpoint_tests() # remove and re-add if reset_with_events "delete re-add signal" && - mptcp_lib_kallsyms_has "subflow_rebuild_header$"; then + continue_if mptcp_lib_kallsyms_has "subflow_rebuild_header$"; then ip netns exec $ns1 sysctl -q net.mptcp.add_addr_timeout=0 pm_nl_set_limits $ns1 0 3 pm_nl_set_limits $ns2 3 3 From a9649dfbe552a42a3781fb681d93a2f510565954 Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Mon, 20 Oct 2025 22:53:30 +0200 Subject: [PATCH 306/330] selftests: mptcp: join: mark laminar tests as skipped if not supported The call to 'continue_if' was missing: it properly marks a subtest as 'skipped' if the attached condition is not valid. Without that, the test is wrongly marked as passed on older kernels. Fixes: c912f935a5c7 ("selftests: mptcp: join: validate new laminar endp") Reviewed-by: Geliang Tang Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20251020-net-mptcp-c-flag-late-add-addr-v1-5-8207030cb0e8@kernel.org Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/mptcp/mptcp_join.sh | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh index b2a8c51a3969..78a1aa4ecff2 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_join.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh @@ -2324,7 +2324,7 @@ laminar_endp_tests() { # no laminar endpoints: routing rules are used if reset_with_tcp_filter "without a laminar endpoint" ns1 10.0.2.2 REJECT && - mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then + continue_if mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then pm_nl_set_limits $ns1 0 2 pm_nl_set_limits $ns2 2 2 pm_nl_add_endpoint $ns1 10.0.2.1 flags signal @@ -2336,7 +2336,7 @@ laminar_endp_tests() # laminar endpoints: this endpoint is used if reset_with_tcp_filter "with a laminar endpoint" ns1 10.0.2.2 REJECT && - mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then + continue_if mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then pm_nl_set_limits $ns1 0 2 pm_nl_set_limits $ns2 2 2 pm_nl_add_endpoint $ns1 10.0.2.1 flags signal @@ -2348,7 +2348,7 @@ laminar_endp_tests() # laminar endpoints: these endpoints are used if reset_with_tcp_filter "with multiple laminar endpoints" ns1 10.0.2.2 REJECT && - mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then + continue_if mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then pm_nl_set_limits $ns1 0 2 pm_nl_set_limits $ns2 2 2 pm_nl_add_endpoint $ns1 10.0.2.1 flags signal @@ -2363,7 +2363,7 @@ laminar_endp_tests() # laminar endpoints: only one endpoint is used if reset_with_tcp_filter "single laminar endpoint" ns1 10.0.2.2 REJECT && - mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then + continue_if mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then pm_nl_set_limits $ns1 0 2 pm_nl_set_limits $ns2 2 2 pm_nl_add_endpoint $ns1 10.0.2.1 flags signal @@ -2376,7 +2376,7 @@ laminar_endp_tests() # laminar endpoints: subflow and laminar flags if reset_with_tcp_filter "sublow + laminar endpoints" ns1 10.0.2.2 REJECT && - mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then + continue_if mptcp_lib_kallsyms_has "mptcp_pm_get_endp_laminar_max$"; then pm_nl_set_limits $ns1 0 4 pm_nl_set_limits $ns2 2 4 pm_nl_add_endpoint $ns1 10.0.2.1 flags signal From c5394b8b7a92c5013d2917591e28e938fe7ff2a2 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 20 Oct 2025 16:11:14 +0000 Subject: [PATCH 307/330] net: gro_cells: fix lock imbalance in gro_cells_receive() syzbot found that the local_unlock_nested_bh() call was missing in some cases. WARNING: possible recursive locking detected syzkaller #0 Not tainted -------------------------------------------- syz.2.329/7421 is trying to acquire lock: ffffe8ffffd48888 ((&cell->bh_lock)){+...}-{3:3}, at: spin_lock include/linux/spinlock_rt.h:44 [inline] ffffe8ffffd48888 ((&cell->bh_lock)){+...}-{3:3}, at: gro_cells_receive+0x404/0x790 net/core/gro_cells.c:30 but task is already holding lock: ffffe8ffffd48888 ((&cell->bh_lock)){+...}-{3:3}, at: spin_lock include/linux/spinlock_rt.h:44 [inline] ffffe8ffffd48888 ((&cell->bh_lock)){+...}-{3:3}, at: gro_cells_receive+0x404/0x790 net/core/gro_cells.c:30 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock((&cell->bh_lock)); lock((&cell->bh_lock)); *** DEADLOCK *** Given the introduction of @have_bh_lock variable, it seems the author intent was to have the local_unlock_nested_bh() after the @unlock label. Fixes: 25718fdcbdd2 ("net: gro_cells: Use nested-BH locking for gro_cell") Reported-by: syzbot+f9651b9a8212e1c8906f@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/68f65eb9.a70a0220.205af.0034.GAE@google.com/T/#u Signed-off-by: Eric Dumazet Cc: Sebastian Andrzej Siewior Reviewed-by: David Ahern Link: https://patch.msgid.link/20251020161114.1891141-1-edumazet@google.com Signed-off-by: Jakub Kicinski --- net/core/gro_cells.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/net/core/gro_cells.c b/net/core/gro_cells.c index b43911562f4d..fd57b845de33 100644 --- a/net/core/gro_cells.c +++ b/net/core/gro_cells.c @@ -43,12 +43,11 @@ int gro_cells_receive(struct gro_cells *gcells, struct sk_buff *skb) if (skb_queue_len(&cell->napi_skbs) == 1) napi_schedule(&cell->napi); - if (have_bh_lock) - local_unlock_nested_bh(&gcells->cells->bh_lock); - res = NET_RX_SUCCESS; unlock: + if (have_bh_lock) + local_unlock_nested_bh(&gcells->cells->bh_lock); rcu_read_unlock(); return res; } From 86c48f50bababbb45622616b48385aa94bfadf5f Mon Sep 17 00:00:00 2001 From: Randy Dunlap Date: Sun, 19 Oct 2025 22:27:16 -0700 Subject: [PATCH 308/330] Documentation: networking: ax25: update the mailing list info. Update the mailing list subscription information for the linux-hams mailing list. Signed-off-by: Randy Dunlap Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251020052716.3136773-1-rdunlap@infradead.org Signed-off-by: Jakub Kicinski --- Documentation/networking/ax25.rst | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/Documentation/networking/ax25.rst b/Documentation/networking/ax25.rst index 605e72c6c877..89c79dd6c6f9 100644 --- a/Documentation/networking/ax25.rst +++ b/Documentation/networking/ax25.rst @@ -11,6 +11,7 @@ found on https://linux-ax25.in-berlin.de. There is a mailing list for discussing Linux amateur radio matters called linux-hams@vger.kernel.org. To subscribe to it, send a message to -majordomo@vger.kernel.org with the words "subscribe linux-hams" in the body -of the message, the subject field is ignored. You don't need to be -subscribed to post but of course that means you might miss an answer. +linux-hams+subscribe@vger.kernel.org or use the web interface at +https://vger.kernel.org. The subject and body of the message are +ignored. You don't need to be subscribed to post but of course that +means you might miss an answer. From 5523508258d390fd55b25c1c0f99c092f23a39e9 Mon Sep 17 00:00:00 2001 From: Yeounsu Moon Date: Sun, 19 Oct 2025 16:55:40 +0900 Subject: [PATCH 309/330] net: dlink: use dev_kfree_skb_any instead of dev_kfree_skb Replace `dev_kfree_skb()` with `dev_kfree_skb_any()` in `start_xmit()` which can be called from netpoll (hard IRQ) and from other contexts. Also, `np->link_status` can be changed at any time by interrupt handler. -0 [011] dNh4. 4541.754603: start_xmit <-netpoll_start_xmit -0 [011] dNh4. 4541.754622: => [FTRACE TRAMPOLINE] => start_xmit => netpoll_start_xmit => netpoll_send_skb => write_msg => console_flush_all => console_unlock => vprintk_emit => _printk => rio_interrupt => __handle_irq_event_percpu => handle_irq_event => handle_fasteoi_irq => __common_interrupt => common_interrupt => asm_common_interrupt => mwait_idle => default_idle_call => do_idle => cpu_startup_entry => start_secondary => common_startup_64 This issue can occur when the link state changes from off to on (e.g., plugging or unplugging the LAN cable) while transmitting a packet. If the skb has a destructor, a warning message may be printed in this situation. -> consume_skb (dev_kfree_skb()) -> __kfree_skb() -> skb_release_all() -> skb_release_head_state(skb) if (skb->destructor) { DEBUG_NET_WARN_ON_ONCE(in_hardirq()); skb->destructor(skb); } Found by inspection. Signed-off-by: Yeounsu Moon Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Tested-on: D-Link DGE-550T Rev-A3 Reviewed-by: Simon Horman Link: https://patch.msgid.link/20251019075540.55697-1-yyyynoom@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/dlink/dl2k.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/dlink/dl2k.c b/drivers/net/ethernet/dlink/dl2k.c index 7077d705e471..6e4f17142519 100644 --- a/drivers/net/ethernet/dlink/dl2k.c +++ b/drivers/net/ethernet/dlink/dl2k.c @@ -733,7 +733,7 @@ start_xmit (struct sk_buff *skb, struct net_device *dev) u64 tfc_vlan_tag = 0; if (np->link_status == 0) { /* Link Down */ - dev_kfree_skb(skb); + dev_kfree_skb_any(skb); return NETDEV_TX_OK; } entry = np->cur_tx % TX_RING_SIZE; From d63f0391d6c7b75e1a847e1a26349fa8cad0004d Mon Sep 17 00:00:00 2001 From: Heiner Kallweit Date: Mon, 20 Oct 2025 08:54:54 +0200 Subject: [PATCH 310/330] net: hibmcge: select FIXED_PHY hibmcge uses fixed_phy_register() et al, but doesn't cater for the case that hibmcge is built-in and fixed_phy is a module. To solve this select FIXED_PHY. Fixes: 1d7cd7a9c69c ("net: hibmcge: support scenario without PHY") Signed-off-by: Heiner Kallweit Reviewed-by: Jijie Shao Link: https://patch.msgid.link/c4fc061f-b6d5-418b-a0dc-6b238cdbedce@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/hisilicon/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/hisilicon/Kconfig b/drivers/net/ethernet/hisilicon/Kconfig index 65302c41bfb1..38875c196cb6 100644 --- a/drivers/net/ethernet/hisilicon/Kconfig +++ b/drivers/net/ethernet/hisilicon/Kconfig @@ -148,6 +148,7 @@ config HIBMCGE tristate "Hisilicon BMC Gigabit Ethernet Device Support" depends on PCI && PCI_MSI select PHYLIB + select FIXED_PHY select MOTORCOMM_PHY select REALTEK_PHY help From 43c36a56ccf6d9b07b4b3f4f614756e687dcdc01 Mon Sep 17 00:00:00 2001 From: Dominique Martinet Date: Wed, 22 Oct 2025 06:33:42 +0900 Subject: [PATCH 311/330] Revert "fs/9p: Refresh metadata in d_revalidate for uncached mode too" This reverts commit 290434474c332a2ba9c8499fe699c7f2e1153280. That commit broke cache=mmap, a mode that doesn't cache metadata, but still has writeback cache. In commit 290434474c33 ("fs/9p: Refresh metadata in d_revalidate for uncached mode too") we considered metadata cache to be enough to not look at the server, but in writeback cache too looking at the server size would make the vfs consider the file has been truncated before the data has been flushed out, making the following repro fail (nothing is ever read back, the resulting file ends up with no data written) ``` #include #include #include #include char buf[4096]; int main(int argc, char *argv[]) { int ret, i; int fdw, fdr; if (argc < 2) return 1; fdw = openat(AT_FDCWD, argv[1], O_RDWR|O_CREAT|O_EXCL|O_CLOEXEC, 0600); if (fdw < 0) { fprintf(stderr, "cannot open fdw\n"); return 1; } write(fdw, buf, sizeof(buf)); fdr = openat(AT_FDCWD, argv[1], O_RDONLY|O_CLOEXEC); if (fdr < 0) { fprintf(stderr, "cannot open fdr\n"); close(fdw); return 1; } for (i = 0; i < 10; i++) { ret = read(fdr, buf, sizeof(buf)); fprintf(stderr, "i: %d, read returns %d\n", i, ret); } close(fdr); close(fdw); return 0; } ``` There is a fix for this particular reproducer but it looks like there are other problems around metadata refresh (e.g. around file rename), so revert this to avoid d_revalidate in uncached mode for now. Reported-by: Song Liu Link: https://lkml.kernel.org/r/CAHzjS_u_SYdt5=2gYO_dxzMKXzGMt-TfdE_ueowg-Hq5tRCAiw@mail.gmail.com Reported-by: Andrii Nakryiko Link: https://lore.kernel.org/bpf/CAEf4BzZbCE4tLoDZyUf_aASpgAGFj75QMfSXX4a4dLYixnOiLg@mail.gmail.com/ Fixes: 290434474c33 ("fs/9p: Refresh metadata in d_revalidate for uncached mode too") Signed-off-by: Dominique Martinet --- fs/9p/vfs_dentry.c | 10 ++-------- fs/9p/vfs_inode.c | 8 +------- fs/9p/vfs_inode_dotl.c | 8 +------- 3 files changed, 4 insertions(+), 22 deletions(-) diff --git a/fs/9p/vfs_dentry.c b/fs/9p/vfs_dentry.c index f3248a3e5402..c1acbc98465d 100644 --- a/fs/9p/vfs_dentry.c +++ b/fs/9p/vfs_dentry.c @@ -66,7 +66,6 @@ static int __v9fs_lookup_revalidate(struct dentry *dentry, unsigned int flags) struct p9_fid *fid; struct inode *inode; struct v9fs_inode *v9inode; - unsigned int cached; if (flags & LOOKUP_RCU) return -ECHILD; @@ -76,11 +75,7 @@ static int __v9fs_lookup_revalidate(struct dentry *dentry, unsigned int flags) goto out_valid; v9inode = V9FS_I(inode); - struct v9fs_session_info *v9ses = v9fs_inode2v9ses(inode); - - cached = v9ses->cache & (CACHE_META | CACHE_LOOSE); - - if (!cached || v9inode->cache_validity & V9FS_INO_INVALID_ATTR) { + if (v9inode->cache_validity & V9FS_INO_INVALID_ATTR) { int retval; struct v9fs_session_info *v9ses; @@ -114,6 +109,7 @@ static int __v9fs_lookup_revalidate(struct dentry *dentry, unsigned int flags) p9_debug(P9_DEBUG_VFS, "refresh inode: dentry = %pd (%p), got error %pe\n", dentry, dentry, ERR_PTR(retval)); + if (retval < 0) return retval; } } @@ -150,8 +146,6 @@ const struct dentry_operations v9fs_cached_dentry_operations = { }; const struct dentry_operations v9fs_dentry_operations = { - .d_revalidate = v9fs_lookup_revalidate, - .d_weak_revalidate = __v9fs_lookup_revalidate, .d_release = v9fs_dentry_release, .d_unalias_trylock = v9fs_dentry_unalias_trylock, .d_unalias_unlock = v9fs_dentry_unalias_unlock, diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 69f378a83775..d0c77ec31b1d 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -1339,14 +1339,8 @@ int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode) * Don't update inode if the file type is different */ umode = p9mode2unixmode(v9ses, st, &rdev); - if (inode_wrong_type(inode, umode)) { - /* - * Do this as a way of letting the caller know the inode should not - * be reused - */ - v9fs_invalidate_inode_attr(inode); + if (inode_wrong_type(inode, umode)) goto out; - } /* * We don't want to refresh inode->i_size, diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c index 0b404e8484d2..be297e335468 100644 --- a/fs/9p/vfs_inode_dotl.c +++ b/fs/9p/vfs_inode_dotl.c @@ -897,14 +897,8 @@ int v9fs_refresh_inode_dotl(struct p9_fid *fid, struct inode *inode) /* * Don't update inode if the file type is different */ - if (inode_wrong_type(inode, st->st_mode)) { - /* - * Do this as a way of letting the caller know the inode should not - * be reused - */ - v9fs_invalidate_inode_attr(inode); + if (inode_wrong_type(inode, st->st_mode)) goto out; - } /* * We don't want to refresh inode->i_size, From 0fd7e7a1ad7c5fb8801a9c1661fc4e1ae4d9b655 Mon Sep 17 00:00:00 2001 From: Leo Martins Date: Mon, 20 Oct 2025 16:16:15 -0700 Subject: [PATCH 312/330] btrfs: fix delayed_node ref_tracker use after free Move the print before releasing the delayed node. In my initial testing there was a bug that was causing delayed_nodes to not get freed which is why I put the print after the release. This obviously neglects the case where the delayed node is properly freed. Add condition to make sure we only print if we have more than one reference to the delayed_node to prevent printing when we only have the reference taken in btrfs_kill_all_delayed_nodes(). Fixes: b767a28d6154 ("btrfs: print leaked references in kill_all_delayed_nodes()") Tested-by: Christoph Hellwig Signed-off-by: Leo Martins Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/delayed-inode.c | 2 +- fs/btrfs/delayed-inode.h | 7 +++++++ 2 files changed, 8 insertions(+), 1 deletion(-) diff --git a/fs/btrfs/delayed-inode.c b/fs/btrfs/delayed-inode.c index 41e37f7f67cc..3df7b9d7fbe8 100644 --- a/fs/btrfs/delayed-inode.c +++ b/fs/btrfs/delayed-inode.c @@ -2110,9 +2110,9 @@ void btrfs_kill_all_delayed_nodes(struct btrfs_root *root) for (int i = 0; i < count; i++) { __btrfs_kill_delayed_node(delayed_nodes[i]); + btrfs_delayed_node_ref_tracker_dir_print(delayed_nodes[i]); btrfs_release_delayed_node(delayed_nodes[i], &delayed_node_trackers[i]); - btrfs_delayed_node_ref_tracker_dir_print(delayed_nodes[i]); } } } diff --git a/fs/btrfs/delayed-inode.h b/fs/btrfs/delayed-inode.h index 0d949edc0caf..b09d4ec8c77d 100644 --- a/fs/btrfs/delayed-inode.h +++ b/fs/btrfs/delayed-inode.h @@ -219,6 +219,13 @@ static inline void btrfs_delayed_node_ref_tracker_dir_print(struct btrfs_delayed if (!btrfs_test_opt(node->root->fs_info, REF_TRACKER)) return; + /* + * Only print if there are leaked references. The caller is + * holding one reference, so if refs == 1 there is no leak. + */ + if (refcount_read(&node->refs) == 1) + return; + ref_tracker_dir_print(&node->ref_dir.dir, BTRFS_DELAYED_NODE_REF_TRACKER_DISPLAY_LIMIT); } From ada7d45b568abe4f1fd9c53d66e05fbea300674b Mon Sep 17 00:00:00 2001 From: Amit Dhingra Date: Tue, 21 Oct 2025 07:07:20 -0500 Subject: [PATCH 313/330] btrfs: ref-verify: fix IS_ERR() vs NULL check in btrfs_build_ref_tree() btrfs_extent_root()/btrfs_global_root() does not return error pointers, it returns NULL on error. Reported-by: Dan Carpenter Link: https://lore.kernel.org/all/aNJfvxj0anEnk9Dm@stanley.mountain/ Fixes : ed4e6b5d644c ("btrfs: ref-verify: handle damaged extent root tree") CC: stable@vger.kernel.org # 6.17+ Signed-off-by: Amit Dhingra Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/ref-verify.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/btrfs/ref-verify.c b/fs/btrfs/ref-verify.c index de4cb0f3fbd0..e9224145d754 100644 --- a/fs/btrfs/ref-verify.c +++ b/fs/btrfs/ref-verify.c @@ -982,7 +982,7 @@ int btrfs_build_ref_tree(struct btrfs_fs_info *fs_info) extent_root = btrfs_extent_root(fs_info, 0); /* If the extent tree is damaged we cannot ignore it (IGNOREBADROOTS). */ - if (IS_ERR(extent_root)) { + if (!extent_root) { btrfs_warn(fs_info, "ref-verify: extent tree not available, disabling"); btrfs_clear_opt(fs_info->mount_opt, REF_VERIFY); return 0; From db82b8dbf5f06d7b1abec4e1326ed8c02fa16897 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Mon, 20 Oct 2025 17:03:28 +0200 Subject: [PATCH 314/330] PM: runtime: Fix conditional guard definitions Since pm_runtime_get_active() returns 0 on success, all of the DEFINE_GUARD_COND() macros in pm_runtime.h need the "_RET == 0" condition at the end of the argument list or they would not work correctly. Fixes: 9a0abc39450a ("PM: runtime: Add auto-cleanup macros for "resume and get" operations") Reported-by: kernel test robot Link: https://lore.kernel.org/linux-pm/202510191529.BCyjKlLQ-lkp@intel.com/ Signed-off-by: Rafael J. Wysocki Reviewed-by: Jonathan Cameron Reviewed-by: Dan Williams Tested-by: Farhan Ali Link: https://patch.msgid.link/5943878.DvuYhMxLoT@rafael.j.wysocki --- include/linux/pm_runtime.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index a3f44f6c2da1..0b436e15f4cd 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -629,13 +629,13 @@ DEFINE_GUARD(pm_runtime_active_auto, struct device *, * device. */ DEFINE_GUARD_COND(pm_runtime_active, _try, - pm_runtime_get_active(_T, RPM_TRANSPARENT)) + pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0) DEFINE_GUARD_COND(pm_runtime_active, _try_enabled, - pm_runtime_resume_and_get(_T)) + pm_runtime_resume_and_get(_T), _RET == 0) DEFINE_GUARD_COND(pm_runtime_active_auto, _try, - pm_runtime_get_active(_T, RPM_TRANSPARENT)) + pm_runtime_get_active(_T, RPM_TRANSPARENT), _RET == 0) DEFINE_GUARD_COND(pm_runtime_active_auto, _try_enabled, - pm_runtime_resume_and_get(_T)) + pm_runtime_resume_and_get(_T), _RET == 0) /** * pm_runtime_put_sync - Drop device usage counter and run "idle check" if 0. From 5370c31e84b0e0999c7b5ff949f4e104def35584 Mon Sep 17 00:00:00 2001 From: Lad Prabhakar Date: Fri, 17 Oct 2025 16:18:29 +0100 Subject: [PATCH 315/330] net: ravb: Enforce descriptor type ordering MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Ensure the TX descriptor type fields are published in a safe order so the DMA engine never begins processing a descriptor chain before all descriptor fields are fully initialised. For multi-descriptor transmits the driver writes DT_FEND into the last descriptor and DT_FSTART into the first. The DMA engine begins processing when it observes DT_FSTART. Move the dma_wmb() barrier so it executes immediately after DT_FEND and immediately before writing DT_FSTART (and before DT_FSINGLE in the single-descriptor case). This guarantees that all prior CPU writes to the descriptor memory are visible to the device before DT_FSTART is seen. This avoids a situation where compiler/CPU reordering could publish DT_FSTART ahead of DT_FEND or other descriptor fields, allowing the DMA to start on a partially initialised chain and causing corrupted transmissions or TX timeouts. Such a failure was observed on RZ/G2L with an RT kernel as transmit queue timeouts and device resets. Fixes: 2f45d1902acf ("ravb: minimize TX data copying") Cc: stable@vger.kernel.org Co-developed-by: Fabrizio Castro Signed-off-by: Fabrizio Castro Signed-off-by: Lad Prabhakar Reviewed-by: Niklas Söderlund Link: https://patch.msgid.link/20251017151830.171062-4-prabhakar.mahadev-lad.rj@bp.renesas.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/renesas/ravb_main.c | 16 ++++++++++++++-- 1 file changed, 14 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/renesas/ravb_main.c b/drivers/net/ethernet/renesas/ravb_main.c index 9d3bd65b85ff..044ee83c63bb 100644 --- a/drivers/net/ethernet/renesas/ravb_main.c +++ b/drivers/net/ethernet/renesas/ravb_main.c @@ -2211,13 +2211,25 @@ static netdev_tx_t ravb_start_xmit(struct sk_buff *skb, struct net_device *ndev) skb_tx_timestamp(skb); } - /* Descriptor type must be set after all the above writes */ - dma_wmb(); + if (num_tx_desc > 1) { desc->die_dt = DT_FEND; desc--; + /* When using multi-descriptors, DT_FEND needs to get written + * before DT_FSTART, but the compiler may reorder the memory + * writes in an attempt to optimize the code. + * Use a dma_wmb() barrier to make sure DT_FEND and DT_FSTART + * are written exactly in the order shown in the code. + * This is particularly important for cases where the DMA engine + * is already running when we are running this code. If the DMA + * sees DT_FSTART without the corresponding DT_FEND it will enter + * an error condition. + */ + dma_wmb(); desc->die_dt = DT_FSTART; } else { + /* Descriptor type must be set after all the above writes */ + dma_wmb(); desc->die_dt = DT_FSINGLE; } ravb_modify(ndev, TCCR, TCCR_TSRQ0 << q, TCCR_TSRQ0 << q); From 706136c5723626fcde8dd8f598a4dcd251e24927 Mon Sep 17 00:00:00 2001 From: Lad Prabhakar Date: Fri, 17 Oct 2025 16:18:30 +0100 Subject: [PATCH 316/330] net: ravb: Ensure memory write completes before ringing TX doorbell MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Add a final dma_wmb() barrier before triggering the transmit request (TCCR_TSRQ) to ensure all descriptor and buffer writes are visible to the DMA engine. According to the hardware manual, a read-back operation is required before writing to the doorbell register to guarantee completion of previous writes. Instead of performing a dummy read, a dma_wmb() is used to both enforce the same ordering semantics on the CPU side and also to ensure completion of writes. Fixes: c156633f1353 ("Renesas Ethernet AVB driver proper") Cc: stable@vger.kernel.org Co-developed-by: Fabrizio Castro Signed-off-by: Fabrizio Castro Signed-off-by: Lad Prabhakar Reviewed-by: Niklas Söderlund Link: https://patch.msgid.link/20251017151830.171062-5-prabhakar.mahadev-lad.rj@bp.renesas.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/renesas/ravb_main.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/net/ethernet/renesas/ravb_main.c b/drivers/net/ethernet/renesas/ravb_main.c index 044ee83c63bb..e2d7ce1a85e8 100644 --- a/drivers/net/ethernet/renesas/ravb_main.c +++ b/drivers/net/ethernet/renesas/ravb_main.c @@ -2232,6 +2232,14 @@ static netdev_tx_t ravb_start_xmit(struct sk_buff *skb, struct net_device *ndev) dma_wmb(); desc->die_dt = DT_FSINGLE; } + + /* Before ringing the doorbell we need to make sure that the latest + * writes have been committed to memory, otherwise it could delay + * things until the doorbell is rang again. + * This is in replacement of the read operation mentioned in the HW + * manuals. + */ + dma_wmb(); ravb_modify(ndev, TCCR, TCCR_TSRQ0 << q, TCCR_TSRQ0 << q); priv->cur_tx[q] += num_tx_desc; From a767957e7a83f9e742be196aa52a48de8ac5a7e4 Mon Sep 17 00:00:00 2001 From: Jiasheng Jiang Date: Tue, 21 Oct 2025 18:24:56 +0000 Subject: [PATCH 317/330] ptp: ocp: Fix typo using index 1 instead of i in SMA initialization loop In ptp_ocp_sma_fb_init(), the code mistakenly used bp->sma[1] instead of bp->sma[i] inside a for-loop, which caused only SMA[1] to have its DIRECTION_CAN_CHANGE capability cleared. This led to inconsistent capability flags across SMA pins. Fixes: 09eeb3aecc6c ("ptp_ocp: implement DPLL ops") Signed-off-by: Jiasheng Jiang Reviewed-by: Vadim Fedorenko Link: https://patch.msgid.link/20251021182456.9729-1-jiashengjiangcool@gmail.com Signed-off-by: Jakub Kicinski --- drivers/ptp/ptp_ocp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/ptp/ptp_ocp.c b/drivers/ptp/ptp_ocp.c index 794ec6e71990..a5c363252986 100644 --- a/drivers/ptp/ptp_ocp.c +++ b/drivers/ptp/ptp_ocp.c @@ -2548,7 +2548,7 @@ ptp_ocp_sma_fb_init(struct ptp_ocp *bp) for (i = 0; i < OCP_SMA_NUM; i++) { bp->sma[i].fixed_fcn = true; bp->sma[i].fixed_dir = true; - bp->sma[1].dpll_prop.capabilities &= + bp->sma[i].dpll_prop.capabilities &= ~DPLL_PIN_CAPABILITIES_DIRECTION_CAN_CHANGE; } return; From 441f0647f7673e0e64d4910ef61a5fb8f16bfb82 Mon Sep 17 00:00:00 2001 From: Alexey Simakov Date: Tue, 21 Oct 2025 16:00:36 +0300 Subject: [PATCH 318/330] sctp: avoid NULL dereference when chunk data buffer is missing chunk->skb pointer is dereferenced in the if-block where it's supposed to be NULL only. chunk->skb can only be NULL if chunk->head_skb is not. Check for frag_list instead and do it just before replacing chunk->skb. We're sure that otherwise chunk->skb is non-NULL because of outer if() condition. Fixes: 90017accff61 ("sctp: Add GSO support") Signed-off-by: Alexey Simakov Acked-by: Marcelo Ricardo Leitner Link: https://patch.msgid.link/20251021130034.6333-1-bigalex934@gmail.com Signed-off-by: Jakub Kicinski --- net/sctp/inqueue.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/net/sctp/inqueue.c b/net/sctp/inqueue.c index 5c1652181805..f5a7d5a38755 100644 --- a/net/sctp/inqueue.c +++ b/net/sctp/inqueue.c @@ -169,13 +169,14 @@ struct sctp_chunk *sctp_inq_pop(struct sctp_inq *queue) chunk->head_skb = chunk->skb; /* skbs with "cover letter" */ - if (chunk->head_skb && chunk->skb->data_len == chunk->skb->len) + if (chunk->head_skb && chunk->skb->data_len == chunk->skb->len) { + if (WARN_ON(!skb_shinfo(chunk->skb)->frag_list)) { + __SCTP_INC_STATS(dev_net(chunk->skb->dev), + SCTP_MIB_IN_PKT_DISCARDS); + sctp_chunk_free(chunk); + goto next_chunk; + } chunk->skb = skb_shinfo(chunk->skb)->frag_list; - - if (WARN_ON(!chunk->skb)) { - __SCTP_INC_STATS(dev_net(chunk->skb->dev), SCTP_MIB_IN_PKT_DISCARDS); - sctp_chunk_free(chunk); - goto next_chunk; } } From c0178eec8884231a5ae0592b9fce827bccb77e86 Mon Sep 17 00:00:00 2001 From: Fernando Fernandez Mancera Date: Mon, 20 Oct 2025 15:55:33 +0200 Subject: [PATCH 319/330] net: hsr: prevent creation of HSR device with slaves from another netns HSR/PRP driver does not handle correctly having slaves/interlink devices in a different net namespace. Currently, it is possible to create a HSR link in a different net namespace than the slaves/interlink with the following command: ip link add hsr0 netns hsr-ns type hsr slave1 eth1 slave2 eth2 As there is no use-case on supporting this scenario, enforce that HSR device link matches netns defined by IFLA_LINK_NETNSID. The iproute2 command mentioned above will throw the following error: Error: hsr: HSR slaves/interlink must be on the same net namespace than HSR link. Fixes: f421436a591d ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)") Signed-off-by: Fernando Fernandez Mancera Link: https://patch.msgid.link/20251020135533.9373-1-fmancera@suse.de Signed-off-by: Jakub Kicinski --- net/hsr/hsr_netlink.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/net/hsr/hsr_netlink.c b/net/hsr/hsr_netlink.c index b120470246cc..c96b63adf96f 100644 --- a/net/hsr/hsr_netlink.c +++ b/net/hsr/hsr_netlink.c @@ -34,12 +34,18 @@ static int hsr_newlink(struct net_device *dev, struct netlink_ext_ack *extack) { struct net *link_net = rtnl_newlink_link_net(params); + struct net_device *link[2], *interlink = NULL; struct nlattr **data = params->data; enum hsr_version proto_version; unsigned char multicast_spec; u8 proto = HSR_PROTOCOL_HSR; - struct net_device *link[2], *interlink = NULL; + if (!net_eq(link_net, dev_net(dev))) { + NL_SET_ERR_MSG_MOD(extack, + "HSR slaves/interlink must be on the same net namespace than HSR link"); + return -EINVAL; + } + if (!data) { NL_SET_ERR_MSG_MOD(extack, "No slave devices specified"); return -EINVAL; From 10843e1492e474c02b91314963161731fa92af91 Mon Sep 17 00:00:00 2001 From: Tonghao Zhang Date: Tue, 21 Oct 2025 13:09:33 +0800 Subject: [PATCH 320/330] net: bonding: fix possible peer notify event loss or dup issue If the send_peer_notif counter and the peer event notify are not synchronized. It may cause problems such as the loss or dup of peer notify event. Before this patch: - If should_notify_peers is true and the lock for send_peer_notif-- fails, peer event may be sent again in next mii_monitor loop, because should_notify_peers is still true. - If should_notify_peers is true and the lock for send_peer_notif-- succeeded, but the lock for peer event fails, the peer event will be lost. This patch locks the RTNL for send_peer_notif, events, and commit simultaneously. Fixes: 07a4ddec3ce9 ("bonding: add an option to specify a delay between peer notifications") Cc: Jay Vosburgh Cc: Andrew Lunn Cc: Eric Dumazet Cc: Jakub Kicinski Cc: Paolo Abeni Cc: Hangbin Liu Cc: Nikolay Aleksandrov Cc: Vincent Bernat Cc: Signed-off-by: Tonghao Zhang Acked-by: Jay Vosburgh Link: https://patch.msgid.link/20251021050933.46412-1-tonghao@bamaicloud.com Signed-off-by: Paolo Abeni --- drivers/net/bonding/bond_main.c | 40 +++++++++++++++------------------ 1 file changed, 18 insertions(+), 22 deletions(-) diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 67fdcbdd2764..e95e593cd12d 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -2874,7 +2874,7 @@ static void bond_mii_monitor(struct work_struct *work) { struct bonding *bond = container_of(work, struct bonding, mii_work.work); - bool should_notify_peers = false; + bool should_notify_peers; bool commit; unsigned long delay; struct slave *slave; @@ -2886,30 +2886,33 @@ static void bond_mii_monitor(struct work_struct *work) goto re_arm; rcu_read_lock(); + should_notify_peers = bond_should_notify_peers(bond); commit = !!bond_miimon_inspect(bond); - if (bond->send_peer_notif) { - rcu_read_unlock(); - if (rtnl_trylock()) { - bond->send_peer_notif--; - rtnl_unlock(); - } - } else { - rcu_read_unlock(); - } - if (commit) { + rcu_read_unlock(); + + if (commit || bond->send_peer_notif) { /* Race avoidance with bond_close cancel of workqueue */ if (!rtnl_trylock()) { delay = 1; - should_notify_peers = false; goto re_arm; } - bond_for_each_slave(bond, slave, iter) { - bond_commit_link_state(slave, BOND_SLAVE_NOTIFY_LATER); + if (commit) { + bond_for_each_slave(bond, slave, iter) { + bond_commit_link_state(slave, + BOND_SLAVE_NOTIFY_LATER); + } + bond_miimon_commit(bond); + } + + if (bond->send_peer_notif) { + bond->send_peer_notif--; + if (should_notify_peers) + call_netdevice_notifiers(NETDEV_NOTIFY_PEERS, + bond->dev); } - bond_miimon_commit(bond); rtnl_unlock(); /* might sleep, hold no other locks */ } @@ -2917,13 +2920,6 @@ static void bond_mii_monitor(struct work_struct *work) re_arm: if (bond->params.miimon) queue_delayed_work(bond->wq, &bond->mii_work, delay); - - if (should_notify_peers) { - if (!rtnl_trylock()) - return; - call_netdevice_notifiers(NETDEV_NOTIFY_PEERS, bond->dev); - rtnl_unlock(); - } } static int bond_upper_dev_walk(struct net_device *upper, From f6ceec6434b5efff62cecbaa2ff74fc29b96c0c6 Mon Sep 17 00:00:00 2001 From: Ralf Lici Date: Tue, 21 Oct 2025 12:09:40 +0200 Subject: [PATCH 321/330] net: datagram: introduce datagram_poll_queue for custom receive queues Some protocols using TCP encapsulation (e.g., espintcp, openvpn) deliver userspace-bound packets through a custom skb queue rather than the standard sk_receive_queue. Introduce datagram_poll_queue that accepts an explicit receive queue, and convert datagram_poll into a wrapper around datagram_poll_queue. This allows protocols with custom skb queues to reuse the core polling logic without relying on sk_receive_queue. Cc: Sabrina Dubroca Cc: Antonio Quartulli Signed-off-by: Ralf Lici Reviewed-by: Sabrina Dubroca Reviewed-by: Antonio Quartulli Link: https://patch.msgid.link/20251021100942.195010-2-ralf@mandelbit.com Signed-off-by: Paolo Abeni --- include/linux/skbuff.h | 3 +++ net/core/datagram.c | 44 ++++++++++++++++++++++++++++++++---------- 2 files changed, 37 insertions(+), 10 deletions(-) diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index fb3fec9affaa..a7cc3d1f4fd1 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -4204,6 +4204,9 @@ struct sk_buff *__skb_recv_datagram(struct sock *sk, struct sk_buff_head *sk_queue, unsigned int flags, int *off, int *err); struct sk_buff *skb_recv_datagram(struct sock *sk, unsigned int flags, int *err); +__poll_t datagram_poll_queue(struct file *file, struct socket *sock, + struct poll_table_struct *wait, + struct sk_buff_head *rcv_queue); __poll_t datagram_poll(struct file *file, struct socket *sock, struct poll_table_struct *wait); int skb_copy_datagram_iter(const struct sk_buff *from, int offset, diff --git a/net/core/datagram.c b/net/core/datagram.c index cb4b9ef2e4e3..c285c6465923 100644 --- a/net/core/datagram.c +++ b/net/core/datagram.c @@ -920,21 +920,22 @@ int skb_copy_and_csum_datagram_msg(struct sk_buff *skb, EXPORT_SYMBOL(skb_copy_and_csum_datagram_msg); /** - * datagram_poll - generic datagram poll + * datagram_poll_queue - same as datagram_poll, but on a specific receive + * queue * @file: file struct * @sock: socket * @wait: poll table + * @rcv_queue: receive queue to poll * - * Datagram poll: Again totally generic. This also handles - * sequenced packet sockets providing the socket receive queue - * is only ever holding data ready to receive. + * Performs polling on the given receive queue, handling shutdown, error, + * and connection state. This is useful for protocols that deliver + * userspace-bound packets through a custom queue instead of + * sk->sk_receive_queue. * - * Note: when you *don't* use this routine for this protocol, - * and you use a different write policy from sock_writeable() - * then please supply your own write_space callback. + * Return: poll bitmask indicating the socket's current state */ -__poll_t datagram_poll(struct file *file, struct socket *sock, - poll_table *wait) +__poll_t datagram_poll_queue(struct file *file, struct socket *sock, + poll_table *wait, struct sk_buff_head *rcv_queue) { struct sock *sk = sock->sk; __poll_t mask; @@ -956,7 +957,7 @@ __poll_t datagram_poll(struct file *file, struct socket *sock, mask |= EPOLLHUP; /* readable? */ - if (!skb_queue_empty_lockless(&sk->sk_receive_queue)) + if (!skb_queue_empty_lockless(rcv_queue)) mask |= EPOLLIN | EPOLLRDNORM; /* Connection-based need to check for termination and startup */ @@ -978,4 +979,27 @@ __poll_t datagram_poll(struct file *file, struct socket *sock, return mask; } +EXPORT_SYMBOL(datagram_poll_queue); + +/** + * datagram_poll - generic datagram poll + * @file: file struct + * @sock: socket + * @wait: poll table + * + * Datagram poll: Again totally generic. This also handles + * sequenced packet sockets providing the socket receive queue + * is only ever holding data ready to receive. + * + * Note: when you *don't* use this routine for this protocol, + * and you use a different write policy from sock_writeable() + * then please supply your own write_space callback. + * + * Return: poll bitmask indicating the socket's current state + */ +__poll_t datagram_poll(struct file *file, struct socket *sock, poll_table *wait) +{ + return datagram_poll_queue(file, sock, wait, + &sock->sk->sk_receive_queue); +} EXPORT_SYMBOL(datagram_poll); From 0fc3e32c2c069f541f2724d91f5e98480b640326 Mon Sep 17 00:00:00 2001 From: Ralf Lici Date: Tue, 21 Oct 2025 12:09:41 +0200 Subject: [PATCH 322/330] espintcp: use datagram_poll_queue for socket readiness espintcp uses a custom queue (ike_queue) to deliver packets to userspace. The polling logic relies on datagram_poll, which checks sk_receive_queue, which can lead to false readiness signals when that queue contains non-userspace packets. Switch espintcp_poll to use datagram_poll_queue with ike_queue, ensuring poll only signals readiness when userspace data is actually available. Fixes: e27cca96cd68 ("xfrm: add espintcp (RFC 8229)") Signed-off-by: Ralf Lici Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20251021100942.195010-3-ralf@mandelbit.com Signed-off-by: Paolo Abeni --- net/xfrm/espintcp.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-) diff --git a/net/xfrm/espintcp.c b/net/xfrm/espintcp.c index fc7a603b04f1..bf744ac9d5a7 100644 --- a/net/xfrm/espintcp.c +++ b/net/xfrm/espintcp.c @@ -555,14 +555,10 @@ static void espintcp_close(struct sock *sk, long timeout) static __poll_t espintcp_poll(struct file *file, struct socket *sock, poll_table *wait) { - __poll_t mask = datagram_poll(file, sock, wait); struct sock *sk = sock->sk; struct espintcp_ctx *ctx = espintcp_getctx(sk); - if (!skb_queue_empty(&ctx->ike_queue)) - mask |= EPOLLIN | EPOLLRDNORM; - - return mask; + return datagram_poll_queue(file, sock, wait, &ctx->ike_queue); } static void build_protos(struct proto *espintcp_prot, From efd729408bc7d57e0c8d027b9ff514187fc1a05b Mon Sep 17 00:00:00 2001 From: Ralf Lici Date: Tue, 21 Oct 2025 12:09:42 +0200 Subject: [PATCH 323/330] ovpn: use datagram_poll_queue for socket readiness in TCP openvpn TCP encapsulation uses a custom queue to deliver packets to userspace. Currently it relies on datagram_poll, which checks sk_receive_queue, leading to false readiness signals when that queue contains non-userspace packets. Switch ovpn_tcp_poll to use datagram_poll_queue with the peer's user_queue, ensuring poll only signals readiness when userspace data is actually available. Also refactor ovpn_tcp_poll in order to enforce the assumption we can make on the lifetime of ovpn_sock and peer. Fixes: 11851cbd60ea ("ovpn: implement TCP transport") Signed-off-by: Antonio Quartulli Signed-off-by: Ralf Lici Reviewed-by: Sabrina Dubroca Link: https://patch.msgid.link/20251021100942.195010-4-ralf@mandelbit.com Signed-off-by: Paolo Abeni --- drivers/net/ovpn/tcp.c | 26 ++++++++++++++++++++++---- 1 file changed, 22 insertions(+), 4 deletions(-) diff --git a/drivers/net/ovpn/tcp.c b/drivers/net/ovpn/tcp.c index 289f62c5d2c7..0d7f30360d87 100644 --- a/drivers/net/ovpn/tcp.c +++ b/drivers/net/ovpn/tcp.c @@ -560,16 +560,34 @@ static void ovpn_tcp_close(struct sock *sk, long timeout) static __poll_t ovpn_tcp_poll(struct file *file, struct socket *sock, poll_table *wait) { - __poll_t mask = datagram_poll(file, sock, wait); + struct sk_buff_head *queue = &sock->sk->sk_receive_queue; struct ovpn_socket *ovpn_sock; + struct ovpn_peer *peer = NULL; + __poll_t mask; rcu_read_lock(); ovpn_sock = rcu_dereference_sk_user_data(sock->sk); - if (ovpn_sock && ovpn_sock->peer && - !skb_queue_empty(&ovpn_sock->peer->tcp.user_queue)) - mask |= EPOLLIN | EPOLLRDNORM; + /* if we landed in this callback, we expect to have a + * meaningful state. The ovpn_socket lifecycle would + * prevent it otherwise. + */ + if (WARN(!ovpn_sock || !ovpn_sock->peer, + "ovpn: null state in ovpn_tcp_poll!")) { + rcu_read_unlock(); + return 0; + } + + if (ovpn_peer_hold(ovpn_sock->peer)) { + peer = ovpn_sock->peer; + queue = &peer->tcp.user_queue; + } rcu_read_unlock(); + mask = datagram_poll_queue(file, sock, wait, queue); + + if (peer) + ovpn_peer_put(peer); + return mask; } From f7c877e7535260cc7a21484c994e8ce7e8cb6780 Mon Sep 17 00:00:00 2001 From: Stefano Garzarella Date: Tue, 21 Oct 2025 14:17:18 +0200 Subject: [PATCH 324/330] vsock: fix lock inversion in vsock_assign_transport() Syzbot reported a potential lock inversion deadlock between vsock_register_mutex and sk_lock-AF_VSOCK when vsock_linger() is called. The issue was introduced by commit 687aa0c5581b ("vsock: Fix transport_* TOCTOU") which added vsock_register_mutex locking in vsock_assign_transport() around the transport->release() call, that can call vsock_linger(). vsock_assign_transport() can be called with sk_lock held. vsock_linger() calls sk_wait_event() that temporarily releases and re-acquires sk_lock. During this window, if another thread hold vsock_register_mutex while trying to acquire sk_lock, a circular dependency is created. Fix this by releasing vsock_register_mutex before calling transport->release() and vsock_deassign_transport(). This is safe because we don't need to hold vsock_register_mutex while releasing the old transport, and we ensure the new transport won't disappear by obtaining a module reference first via try_module_get(). Reported-by: syzbot+10e35716f8e4929681fa@syzkaller.appspotmail.com Tested-by: syzbot+10e35716f8e4929681fa@syzkaller.appspotmail.com Fixes: 687aa0c5581b ("vsock: Fix transport_* TOCTOU") Cc: mhal@rbox.co Cc: stable@vger.kernel.org Signed-off-by: Stefano Garzarella Link: https://patch.msgid.link/20251021121718.137668-1-sgarzare@redhat.com Signed-off-by: Paolo Abeni --- net/vmw_vsock/af_vsock.c | 38 +++++++++++++++++++------------------- 1 file changed, 19 insertions(+), 19 deletions(-) diff --git a/net/vmw_vsock/af_vsock.c b/net/vmw_vsock/af_vsock.c index 4c2db6cca557..76763247a377 100644 --- a/net/vmw_vsock/af_vsock.c +++ b/net/vmw_vsock/af_vsock.c @@ -487,12 +487,26 @@ int vsock_assign_transport(struct vsock_sock *vsk, struct vsock_sock *psk) goto err; } - if (vsk->transport) { - if (vsk->transport == new_transport) { - ret = 0; - goto err; - } + if (vsk->transport && vsk->transport == new_transport) { + ret = 0; + goto err; + } + /* We increase the module refcnt to prevent the transport unloading + * while there are open sockets assigned to it. + */ + if (!new_transport || !try_module_get(new_transport->module)) { + ret = -ENODEV; + goto err; + } + + /* It's safe to release the mutex after a successful try_module_get(). + * Whichever transport `new_transport` points at, it won't go away until + * the last module_put() below or in vsock_deassign_transport(). + */ + mutex_unlock(&vsock_register_mutex); + + if (vsk->transport) { /* transport->release() must be called with sock lock acquired. * This path can only be taken during vsock_connect(), where we * have already held the sock lock. In the other cases, this @@ -512,20 +526,6 @@ int vsock_assign_transport(struct vsock_sock *vsk, struct vsock_sock *psk) vsk->peer_shutdown = 0; } - /* We increase the module refcnt to prevent the transport unloading - * while there are open sockets assigned to it. - */ - if (!new_transport || !try_module_get(new_transport->module)) { - ret = -ENODEV; - goto err; - } - - /* It's safe to release the mutex after a successful try_module_get(). - * Whichever transport `new_transport` points at, it won't go away until - * the last module_put() below or in vsock_deassign_transport(). - */ - mutex_unlock(&vsock_register_mutex); - if (sk->sk_type == SOCK_SEQPACKET) { if (!new_transport->seqpacket_allow || !new_transport->seqpacket_allow(remote_cid)) { From 399d10934740ae8cdaa4e3245f7c5f6c332da844 Mon Sep 17 00:00:00 2001 From: Robert Marko Date: Tue, 21 Oct 2025 15:20:26 +0200 Subject: [PATCH 325/330] net: phy: micrel: always set shared->phydev for LAN8814 Currently, during the LAN8814 PTP probe shared->phydev is only set if PTP clock gets actually set, otherwise the function will return before setting it. This is an issue as shared->phydev is unconditionally being used when IRQ is being handled, especially in lan8814_gpio_process_cap and since it was not set it will cause a NULL pointer exception and crash the kernel. So, simply always set shared->phydev to avoid the NULL pointer exception. Fixes: b3f1a08fcf0d ("net: phy: micrel: Add support for PTP_PF_EXTTS for lan8814") Signed-off-by: Robert Marko Tested-by: Horatiu Vultur Link: https://patch.msgid.link/20251021132034.983936-1-robert.marko@sartura.hr Signed-off-by: Jakub Kicinski --- drivers/net/phy/micrel.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/net/phy/micrel.c b/drivers/net/phy/micrel.c index 79ce3eb6752b..604b5de0c158 100644 --- a/drivers/net/phy/micrel.c +++ b/drivers/net/phy/micrel.c @@ -4262,6 +4262,8 @@ static int __lan8814_ptp_probe_once(struct phy_device *phydev, char *pin_name, { struct lan8814_shared_priv *shared = phy_package_get_priv(phydev); + shared->phydev = phydev; + /* Initialise shared lock for clock*/ mutex_init(&shared->shared_lock); @@ -4317,8 +4319,6 @@ static int __lan8814_ptp_probe_once(struct phy_device *phydev, char *pin_name, phydev_dbg(phydev, "successfully registered ptp clock\n"); - shared->phydev = phydev; - /* The EP.4 is shared between all the PHYs in the package and also it * can be accessed by any of the PHYs */ From b2284768c6b32aa224ca7d0ef0741beb434f03aa Mon Sep 17 00:00:00 2001 From: Jason Wang Date: Wed, 22 Oct 2025 11:44:21 +0800 Subject: [PATCH 326/330] virtio-net: zero unused hash fields When GSO tunnel is negotiated virtio_net_hdr_tnl_from_skb() tries to initialize the tunnel metadata but forget to zero unused rxhash fields. This may leak information to another side. Fixing this by zeroing the unused hash fields. Acked-by: Michael S. Tsirkin Fixes: a2fb4bc4e2a6a ("net: implement virtio helpers to handle UDP GSO tunneling") Cc: Signed-off-by: Jason Wang Reviewed-by: Xuan Zhuo Link: https://patch.msgid.link/20251022034421.70244-1-jasowang@redhat.com Signed-off-by: Jakub Kicinski --- include/linux/virtio_net.h | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/include/linux/virtio_net.h b/include/linux/virtio_net.h index 20e0584db1dd..4d1780848d0e 100644 --- a/include/linux/virtio_net.h +++ b/include/linux/virtio_net.h @@ -401,6 +401,10 @@ virtio_net_hdr_tnl_from_skb(const struct sk_buff *skb, if (!tnl_hdr_negotiated) return -EINVAL; + vhdr->hash_hdr.hash_value = 0; + vhdr->hash_hdr.hash_report = 0; + vhdr->hash_hdr.padding = 0; + /* Let the basic parsing deal with plain GSO features. */ skb_shinfo(skb)->gso_type &= ~tnl_gso_type; ret = virtio_net_hdr_from_skb(skb, hdr, true, false, vlan_hlen); From bb65e0c141f879cdf54db11ae446ee3605fb54d5 Mon Sep 17 00:00:00 2001 From: Alexei Lazar Date: Wed, 22 Oct 2025 15:29:39 +0300 Subject: [PATCH 327/330] net/mlx5: Add PPHCR to PCAM supported registers mask Add the PPHCR bit to the port_access_reg_cap_mask field of PCAM register to indicate that the device supports the PPHCR register and the RS-FEC histogram feature. Signed-off-by: Alexei Lazar Reviewed-by: Yael Chemla Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/1761136182-918470-2-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- include/linux/mlx5/mlx5_ifc.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 07614cd95bed..1b0b36aa2a76 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -10833,7 +10833,9 @@ struct mlx5_ifc_pcam_regs_5000_to_507f_bits { u8 port_access_reg_cap_mask_127_to_96[0x20]; u8 port_access_reg_cap_mask_95_to_64[0x20]; - u8 port_access_reg_cap_mask_63_to_36[0x1c]; + u8 port_access_reg_cap_mask_63[0x1]; + u8 pphcr[0x1]; + u8 port_access_reg_cap_mask_61_to_36[0x1a]; u8 pplm[0x1]; u8 port_access_reg_cap_mask_34_to_32[0x3]; From d58a9a917aa39b10250ec16cb9c24e942cbc77d3 Mon Sep 17 00:00:00 2001 From: Alexei Lazar Date: Wed, 22 Oct 2025 15:29:40 +0300 Subject: [PATCH 328/330] net/mlx5e: Skip PPHCR register query if not supported by the device Check the PCAM supported registers mask before querying the PPHCR register, as it is not supported in older devices. Fixes: 44907e7c8fd0 ("net/mlx5e: Add logic to read RS-FEC histogram bin ranges from PPHCR") Signed-off-by: Alexei Lazar Reviewed-by: Yael Chemla Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/1761136182-918470-3-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/en_stats.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c index 7c029a7d0fd7..a2802cfc9b98 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_stats.c @@ -1614,7 +1614,9 @@ void mlx5e_stats_fec_get(struct mlx5e_priv *priv, fec_set_corrected_bits_total(priv, fec_stats); fec_set_block_stats(priv, mode, fec_stats); - fec_set_histograms_stats(priv, mode, hist); + + if (MLX5_CAP_PCAM_REG(priv->mdev, pphcr)) + fec_set_histograms_stats(priv, mode, hist); } #define PPORT_ETH_EXT_OFF(c) \ From 8f82f89550daafc8ca3ba74c389ae1b4afdd75c8 Mon Sep 17 00:00:00 2001 From: Patrisious Haddad Date: Wed, 22 Oct 2025 15:29:41 +0300 Subject: [PATCH 329/330] net/mlx5: Refactor devcom to return NULL on failure Devcom device and component registration isn't always critical to the functionality of the caller, hence the registration can fail and we can continue working with an ERR_PTR value saved inside a variable. In order to avoid that make sure all devcom failures return NULL. Signed-off-by: Patrisious Haddad Reviewed-by: Leon Romanovsky Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/1761136182-918470-4-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- .../net/ethernet/mellanox/mlx5/core/en_main.c | 6 +-- .../mellanox/mlx5/core/eswitch_offloads.c | 4 +- .../net/ethernet/mellanox/mlx5/core/lag/lag.c | 7 ++- .../ethernet/mellanox/mlx5/core/lib/clock.c | 2 +- .../ethernet/mellanox/mlx5/core/lib/devcom.c | 53 +++++++++---------- .../net/ethernet/mellanox/mlx5/core/lib/sd.c | 4 +- .../net/ethernet/mellanox/mlx5/core/main.c | 5 +- 7 files changed, 39 insertions(+), 42 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index a56825921c23..41fd5eee6306 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -242,8 +242,8 @@ static int mlx5e_devcom_init_mpv(struct mlx5e_priv *priv, u64 *data) &attr, mlx5e_devcom_event_mpv, priv); - if (IS_ERR(priv->devcom)) - return PTR_ERR(priv->devcom); + if (!priv->devcom) + return -EINVAL; if (mlx5_core_is_mp_master(priv->mdev)) { mlx5_devcom_send_event(priv->devcom, MPV_DEVCOM_MASTER_UP, @@ -256,7 +256,7 @@ static int mlx5e_devcom_init_mpv(struct mlx5e_priv *priv, u64 *data) static void mlx5e_devcom_cleanup_mpv(struct mlx5e_priv *priv) { - if (IS_ERR_OR_NULL(priv->devcom)) + if (!priv->devcom) return; if (mlx5_core_is_mp_master(priv->mdev)) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 4cf995be127d..34749814f19b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -3129,7 +3129,7 @@ void mlx5_esw_offloads_devcom_init(struct mlx5_eswitch *esw, attr, mlx5_esw_offloads_devcom_event, esw); - if (IS_ERR(esw->devcom)) + if (!esw->devcom) return; mlx5_devcom_send_event(esw->devcom, @@ -3140,7 +3140,7 @@ void mlx5_esw_offloads_devcom_init(struct mlx5_eswitch *esw, void mlx5_esw_offloads_devcom_cleanup(struct mlx5_eswitch *esw) { - if (IS_ERR_OR_NULL(esw->devcom)) + if (!esw->devcom) return; mlx5_devcom_send_event(esw->devcom, diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c index 59c00c911275..3db0387bf6dc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lag/lag.c @@ -1430,11 +1430,10 @@ static int mlx5_lag_register_hca_devcom_comp(struct mlx5_core_dev *dev) mlx5_devcom_register_component(dev->priv.devc, MLX5_DEVCOM_HCA_PORTS, &attr, NULL, dev); - if (IS_ERR(dev->priv.hca_devcom_comp)) { + if (!dev->priv.hca_devcom_comp) { mlx5_core_err(dev, - "Failed to register devcom HCA component, err: %ld\n", - PTR_ERR(dev->priv.hca_devcom_comp)); - return PTR_ERR(dev->priv.hca_devcom_comp); + "Failed to register devcom HCA component."); + return -EINVAL; } return 0; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c index d0ba83d77cd1..29e7fa09c32c 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/clock.c @@ -1444,7 +1444,7 @@ static void mlx5_shared_clock_register(struct mlx5_core_dev *mdev, u64 key) compd = mlx5_devcom_register_component(mdev->priv.devc, MLX5_DEVCOM_SHARED_CLOCK, &attr, NULL, mdev); - if (IS_ERR(compd)) + if (!compd) return; mdev->clock_state->compdev = compd; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c index faa2833602c8..e749618229bc 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/devcom.c @@ -76,20 +76,18 @@ mlx5_devcom_dev_alloc(struct mlx5_core_dev *dev) struct mlx5_devcom_dev * mlx5_devcom_register_device(struct mlx5_core_dev *dev) { - struct mlx5_devcom_dev *devc; + struct mlx5_devcom_dev *devc = NULL; mutex_lock(&dev_list_lock); if (devcom_dev_exists(dev)) { - devc = ERR_PTR(-EEXIST); + mlx5_core_err(dev, "devcom device already exists"); goto out; } devc = mlx5_devcom_dev_alloc(dev); - if (!devc) { - devc = ERR_PTR(-ENOMEM); + if (!devc) goto out; - } list_add_tail(&devc->list, &devcom_dev_list); out: @@ -110,8 +108,10 @@ mlx5_devcom_dev_release(struct kref *ref) void mlx5_devcom_unregister_device(struct mlx5_devcom_dev *devc) { - if (!IS_ERR_OR_NULL(devc)) - kref_put(&devc->ref, mlx5_devcom_dev_release); + if (!devc) + return; + + kref_put(&devc->ref, mlx5_devcom_dev_release); } static struct mlx5_devcom_comp * @@ -122,7 +122,7 @@ mlx5_devcom_comp_alloc(u64 id, const struct mlx5_devcom_match_attr *attr, comp = kzalloc(sizeof(*comp), GFP_KERNEL); if (!comp) - return ERR_PTR(-ENOMEM); + return NULL; comp->id = id; comp->key.key = attr->key; @@ -160,7 +160,7 @@ devcom_alloc_comp_dev(struct mlx5_devcom_dev *devc, devcom = kzalloc(sizeof(*devcom), GFP_KERNEL); if (!devcom) - return ERR_PTR(-ENOMEM); + return NULL; kref_get(&devc->ref); devcom->devc = devc; @@ -240,31 +240,28 @@ mlx5_devcom_register_component(struct mlx5_devcom_dev *devc, mlx5_devcom_event_handler_t handler, void *data) { - struct mlx5_devcom_comp_dev *devcom; + struct mlx5_devcom_comp_dev *devcom = NULL; struct mlx5_devcom_comp *comp; - if (IS_ERR_OR_NULL(devc)) - return ERR_PTR(-EINVAL); + if (!devc) + return NULL; mutex_lock(&comp_list_lock); comp = devcom_component_get(devc, id, attr, handler); - if (IS_ERR(comp)) { - devcom = ERR_PTR(-EINVAL); + if (IS_ERR(comp)) goto out_unlock; - } if (!comp) { comp = mlx5_devcom_comp_alloc(id, attr, handler); - if (IS_ERR(comp)) { - devcom = ERR_CAST(comp); + if (!comp) goto out_unlock; - } + list_add_tail(&comp->comp_list, &devcom_comp_list); } mutex_unlock(&comp_list_lock); devcom = devcom_alloc_comp_dev(devc, comp, data); - if (IS_ERR(devcom)) + if (!devcom) kref_put(&comp->ref, mlx5_devcom_comp_release); return devcom; @@ -276,8 +273,10 @@ mlx5_devcom_register_component(struct mlx5_devcom_dev *devc, void mlx5_devcom_unregister_component(struct mlx5_devcom_comp_dev *devcom) { - if (!IS_ERR_OR_NULL(devcom)) - devcom_free_comp_dev(devcom); + if (!devcom) + return; + + devcom_free_comp_dev(devcom); } int mlx5_devcom_comp_get_size(struct mlx5_devcom_comp_dev *devcom) @@ -296,7 +295,7 @@ int mlx5_devcom_send_event(struct mlx5_devcom_comp_dev *devcom, int err = 0; void *data; - if (IS_ERR_OR_NULL(devcom)) + if (!devcom) return -ENODEV; comp = devcom->comp; @@ -338,7 +337,7 @@ void mlx5_devcom_comp_set_ready(struct mlx5_devcom_comp_dev *devcom, bool ready) bool mlx5_devcom_comp_is_ready(struct mlx5_devcom_comp_dev *devcom) { - if (IS_ERR_OR_NULL(devcom)) + if (!devcom) return false; return READ_ONCE(devcom->comp->ready); @@ -348,7 +347,7 @@ bool mlx5_devcom_for_each_peer_begin(struct mlx5_devcom_comp_dev *devcom) { struct mlx5_devcom_comp *comp; - if (IS_ERR_OR_NULL(devcom)) + if (!devcom) return false; comp = devcom->comp; @@ -421,21 +420,21 @@ void *mlx5_devcom_get_next_peer_data_rcu(struct mlx5_devcom_comp_dev *devcom, void mlx5_devcom_comp_lock(struct mlx5_devcom_comp_dev *devcom) { - if (IS_ERR_OR_NULL(devcom)) + if (!devcom) return; down_write(&devcom->comp->sem); } void mlx5_devcom_comp_unlock(struct mlx5_devcom_comp_dev *devcom) { - if (IS_ERR_OR_NULL(devcom)) + if (!devcom) return; up_write(&devcom->comp->sem); } int mlx5_devcom_comp_trylock(struct mlx5_devcom_comp_dev *devcom) { - if (IS_ERR_OR_NULL(devcom)) + if (!devcom) return 0; return down_write_trylock(&devcom->comp->sem); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/lib/sd.c b/drivers/net/ethernet/mellanox/mlx5/core/lib/sd.c index f5c2701f6e87..8e17daae48af 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/lib/sd.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/lib/sd.c @@ -221,8 +221,8 @@ static int sd_register(struct mlx5_core_dev *dev) attr.net = mlx5_core_net(dev); devcom = mlx5_devcom_register_component(dev->priv.devc, MLX5_DEVCOM_SD_GROUP, &attr, NULL, dev); - if (IS_ERR(devcom)) - return PTR_ERR(devcom); + if (!devcom) + return -EINVAL; sd->devcom = devcom; diff --git a/drivers/net/ethernet/mellanox/mlx5/core/main.c b/drivers/net/ethernet/mellanox/mlx5/core/main.c index df93625c9dfa..70c156591b0b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/main.c @@ -978,9 +978,8 @@ static int mlx5_init_once(struct mlx5_core_dev *dev) int err; dev->priv.devc = mlx5_devcom_register_device(dev); - if (IS_ERR(dev->priv.devc)) - mlx5_core_warn(dev, "failed to register devcom device %pe\n", - dev->priv.devc); + if (!dev->priv.devc) + mlx5_core_warn(dev, "failed to register devcom device\n"); err = mlx5_query_board_id(dev); if (err) { From 664f76be38a18c61151d0ef248c7e2f3afb4f3c7 Mon Sep 17 00:00:00 2001 From: Patrisious Haddad Date: Wed, 22 Oct 2025 15:29:42 +0300 Subject: [PATCH 330/330] net/mlx5: Fix IPsec cleanup over MPV device When we do mlx5e_detach_netdev() we eventually disable blocking events notifier, among those events are IPsec MPV events from IB to core. So before disabling those blocking events, make sure to also unregister the devcom device and mark all this device operations as complete, in order to prevent the other device from using invalid netdev during future devcom events which could cause the trace below. BUG: kernel NULL pointer dereference, address: 0000000000000010 PGD 146427067 P4D 146427067 PUD 146488067 PMD 0 Oops: Oops: 0000 [#1] SMP CPU: 1 UID: 0 PID: 7735 Comm: devlink Tainted: GW 6.12.0-rc6_for_upstream_min_debug_2024_11_08_00_46 #1 Tainted: [W]=WARN Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014 RIP: 0010:mlx5_devcom_comp_set_ready+0x5/0x40 [mlx5_core] Code: 00 01 48 83 05 23 32 1e 00 01 41 b8 ed ff ff ff e9 60 ff ff ff 48 83 05 00 32 1e 00 01 eb e3 66 0f 1f 44 00 00 0f 1f 44 00 00 <48> 8b 47 10 48 83 05 5f 32 1e 00 01 48 8b 50 40 48 85 d2 74 05 40 RSP: 0018:ffff88811a5c35f8 EFLAGS: 00010206 RAX: ffff888106e8ab80 RBX: ffff888107d7e200 RCX: ffff88810d6f0a00 RDX: ffff88810d6f0a00 RSI: 0000000000000001 RDI: 0000000000000000 RBP: ffff88811a17e620 R08: 0000000000000040 R09: 0000000000000000 R10: ffff88811a5c3618 R11: 0000000de85d51bd R12: ffff88811a17e600 R13: ffff88810d6f0a00 R14: 0000000000000000 R15: ffff8881034bda80 FS: 00007f27bdf89180(0000) GS:ffff88852c880000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000010 CR3: 000000010f159005 CR4: 0000000000372eb0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: ? __die+0x20/0x60 ? page_fault_oops+0x150/0x3e0 ? exc_page_fault+0x74/0x130 ? asm_exc_page_fault+0x22/0x30 ? mlx5_devcom_comp_set_ready+0x5/0x40 [mlx5_core] mlx5e_devcom_event_mpv+0x42/0x60 [mlx5_core] mlx5_devcom_send_event+0x8c/0x170 [mlx5_core] blocking_event+0x17b/0x230 [mlx5_core] notifier_call_chain+0x35/0xa0 blocking_notifier_call_chain+0x3d/0x60 mlx5_blocking_notifier_call_chain+0x22/0x30 [mlx5_core] mlx5_core_mp_event_replay+0x12/0x20 [mlx5_core] mlx5_ib_bind_slave_port+0x228/0x2c0 [mlx5_ib] mlx5_ib_stage_init_init+0x664/0x9d0 [mlx5_ib] ? idr_alloc_cyclic+0x50/0xb0 ? __kmalloc_cache_noprof+0x167/0x340 ? __kmalloc_noprof+0x1a7/0x430 __mlx5_ib_add+0x34/0xd0 [mlx5_ib] mlx5r_probe+0xe9/0x310 [mlx5_ib] ? kernfs_add_one+0x107/0x150 ? __mlx5_ib_add+0xd0/0xd0 [mlx5_ib] auxiliary_bus_probe+0x3e/0x90 really_probe+0xc5/0x3a0 ? driver_probe_device+0x90/0x90 __driver_probe_device+0x80/0x160 driver_probe_device+0x1e/0x90 __device_attach_driver+0x7d/0x100 bus_for_each_drv+0x80/0xd0 __device_attach+0xbc/0x1f0 bus_probe_device+0x86/0xa0 device_add+0x62d/0x830 __auxiliary_device_add+0x3b/0xa0 ? auxiliary_device_init+0x41/0x90 add_adev+0xd1/0x150 [mlx5_core] mlx5_rescan_drivers_locked+0x21c/0x300 [mlx5_core] esw_mode_change+0x6c/0xc0 [mlx5_core] mlx5_devlink_eswitch_mode_set+0x21e/0x640 [mlx5_core] devlink_nl_eswitch_set_doit+0x60/0xe0 genl_family_rcv_msg_doit+0xd0/0x120 genl_rcv_msg+0x180/0x2b0 ? devlink_get_from_attrs_lock+0x170/0x170 ? devlink_nl_eswitch_get_doit+0x290/0x290 ? devlink_nl_pre_doit_port_optional+0x50/0x50 ? genl_family_rcv_msg_dumpit+0xf0/0xf0 netlink_rcv_skb+0x54/0x100 genl_rcv+0x24/0x40 netlink_unicast+0x1fc/0x2d0 netlink_sendmsg+0x1e4/0x410 __sock_sendmsg+0x38/0x60 ? sockfd_lookup_light+0x12/0x60 __sys_sendto+0x105/0x160 ? __sys_recvmsg+0x4e/0x90 __x64_sys_sendto+0x20/0x30 do_syscall_64+0x4c/0x100 entry_SYSCALL_64_after_hwframe+0x4b/0x53 RIP: 0033:0x7f27bc91b13a Code: bb 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 8b 05 fa 96 2c 00 45 89 c9 4c 63 d1 48 63 ff 85 c0 75 15 b8 2c 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 76 f3 c3 0f 1f 40 00 41 55 41 54 4d 89 c5 55 RSP: 002b:00007fff369557e8 EFLAGS: 00000246 ORIG_RAX: 000000000000002c RAX: ffffffffffffffda RBX: 0000000009c54b10 RCX: 00007f27bc91b13a RDX: 0000000000000038 RSI: 0000000009c54b10 RDI: 0000000000000006 RBP: 0000000009c54920 R08: 00007f27bd0030e0 R09: 000000000000000c R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000001 Modules linked in: mlx5_vdpa vringh vhost_iotlb vdpa xt_MASQUERADE nf_conntrack_netlink nfnetlink iptable_nat xt_addrtype xt_conntrack nf_nat br_netfilter rpcsec_gss_krb5 auth_rpcgss oid_registry overlay rpcrdma rdma_ucm ib_iser libiscsi ib_umad scsi_transport_iscsi ib_ipoib rdma_cm iw_cm ib_cm mlx5_fwctl mlx5_ib ib_uverbs ib_core mlx5_core CR2: 0000000000000010 Fixes: 82f9378c443c ("net/mlx5: Handle IPsec steering upon master unbind/bind") Signed-off-by: Patrisious Haddad Reviewed-by: Leon Romanovsky Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/1761136182-918470-5-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- .../mellanox/mlx5/core/en_accel/ipsec.h | 5 ++++ .../mellanox/mlx5/core/en_accel/ipsec_fs.c | 25 +++++++++++++++++-- .../net/ethernet/mellanox/mlx5/core/en_main.c | 2 ++ 3 files changed, 30 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h index 5d7c15abfcaf..f8eaaf37963b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec.h @@ -342,6 +342,7 @@ void mlx5e_ipsec_build_accel_xfrm_attrs(struct mlx5e_ipsec_sa_entry *sa_entry, void mlx5e_ipsec_handle_mpv_event(int event, struct mlx5e_priv *slave_priv, struct mlx5e_priv *master_priv); void mlx5e_ipsec_send_event(struct mlx5e_priv *priv, int event); +void mlx5e_ipsec_disable_events(struct mlx5e_priv *priv); static inline struct mlx5_core_dev * mlx5e_ipsec_sa2dev(struct mlx5e_ipsec_sa_entry *sa_entry) @@ -387,6 +388,10 @@ static inline void mlx5e_ipsec_handle_mpv_event(int event, struct mlx5e_priv *sl static inline void mlx5e_ipsec_send_event(struct mlx5e_priv *priv, int event) { } + +static inline void mlx5e_ipsec_disable_events(struct mlx5e_priv *priv) +{ +} #endif #endif /* __MLX5E_IPSEC_H__ */ diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c index bf1d2769d4f1..feef86fff4bf 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c @@ -2893,9 +2893,30 @@ void mlx5e_ipsec_handle_mpv_event(int event, struct mlx5e_priv *slave_priv, void mlx5e_ipsec_send_event(struct mlx5e_priv *priv, int event) { - if (!priv->ipsec) - return; /* IPsec not supported */ + if (!priv->ipsec || mlx5_devcom_comp_get_size(priv->devcom) < 2) + return; /* IPsec not supported or no peers */ mlx5_devcom_send_event(priv->devcom, event, event, priv); wait_for_completion(&priv->ipsec->comp); } + +void mlx5e_ipsec_disable_events(struct mlx5e_priv *priv) +{ + struct mlx5_devcom_comp_dev *tmp = NULL; + struct mlx5e_priv *peer_priv; + + if (!priv->devcom) + return; + + if (!mlx5_devcom_for_each_peer_begin(priv->devcom)) + goto out; + + peer_priv = mlx5_devcom_get_next_peer_data(priv->devcom, &tmp); + if (peer_priv) + complete_all(&peer_priv->ipsec->comp); + + mlx5_devcom_for_each_peer_end(priv->devcom); +out: + mlx5_devcom_unregister_component(priv->devcom); + priv->devcom = NULL; +} diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index 41fd5eee6306..9c46511e7b43 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -266,6 +266,7 @@ static void mlx5e_devcom_cleanup_mpv(struct mlx5e_priv *priv) } mlx5_devcom_unregister_component(priv->devcom); + priv->devcom = NULL; } static int blocking_event(struct notifier_block *nb, unsigned long event, void *data) @@ -6120,6 +6121,7 @@ static void mlx5e_nic_disable(struct mlx5e_priv *priv) if (mlx5e_monitor_counter_supported(priv)) mlx5e_monitor_counter_cleanup(priv); + mlx5e_ipsec_disable_events(priv); mlx5e_disable_blocking_events(priv); mlx5e_disable_async_events(priv); mlx5_lag_remove_netdev(mdev, priv->netdev);