Originally the addr != NULL check was meant to take care of the case
where __kfence_pool == NULL (KFENCE is disabled). However, this does
not work for addresses where addr > 0 && addr < KFENCE_POOL_SIZE.
This can be the case on NULL-deref where addr > 0 && addr < PAGE_SIZE or
any other faulting access with addr < KFENCE_POOL_SIZE. While the
kernel would likely crash, the stack traces and report might be
confusing due to double faults upon KFENCE's attempt to unprotect such
an address.
Fix it by just checking that __kfence_pool != NULL instead.
Link: https://lkml.kernel.org/r/20210818130300.2482437-1-elver@google.com
Fixes: 0ce20dd840 ("mm: add Kernel Electric-Fence infrastructure")
Signed-off-by: Marco Elver <elver@google.com>
Reported-by: Kuan-Ying Lee <Kuan-Ying.Lee@mediatek.com>
Acked-by: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: <stable@vger.kernel.org> [5.12+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
We've noticed occasional OOM killing when memory.low settings are in
effect for cgroups. This is unexpected and undesirable as memory.low is
supposed to express non-OOMing memory priorities between cgroups.
The reason for this is proportional memory.low reclaim. When cgroups
are below their memory.low threshold, reclaim passes them over in the
first round, and then retries if it couldn't find pages anywhere else.
But when cgroups are slightly above their memory.low setting, page scan
force is scaled down and diminished in proportion to the overage, to the
point where it can cause reclaim to fail as well - only in that case we
currently don't retry, and instead trigger OOM.
To fix this, hook proportional reclaim into the same retry logic we have
in place for when cgroups are skipped entirely. This way if reclaim
fails and some cgroups were scanned with diminished pressure, we'll try
another full-force cycle before giving up and OOMing.
[akpm@linux-foundation.org: coding-style fixes]
Link: https://lkml.kernel.org/r/20210817180506.220056-1-hannes@cmpxchg.org
Fixes: 9783aa9917 ("mm, memcg: proportional memory.{low,min} reclaim")
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: Leon Yang <lnyng@fb.com>
Reviewed-by: Rik van Riel <riel@surriel.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Roman Gushchin <guro@fb.com>
Acked-by: Chris Down <chris@chrisdown.name>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: <stable@vger.kernel.org> [5.4+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull networking fixes from Jakub Kicinski:
"Networking fixes, including fixes from bpf, wireless and mac80211
trees.
Current release - regressions:
- tipc: call tipc_wait_for_connect only when dlen is not 0
- mac80211: fix locking in ieee80211_restart_work()
Current release - new code bugs:
- bpf: add rcu_read_lock in bpf_get_current_[ancestor_]cgroup_id()
- ethernet: ice: fix perout start time rounding
- wwan: iosm: prevent underflow in ipc_chnl_cfg_get()
Previous releases - regressions:
- bpf: clear zext_dst of dead insns
- sch_cake: fix srchost/dsthost hashing mode
- vrf: reset skb conntrack connection on VRF rcv
- net/rds: dma_map_sg is entitled to merge entries
Previous releases - always broken:
- ethernet: bnxt: fix Tx path locking and races, add Rx path
barriers"
* tag 'net-5.14-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (42 commits)
net: dpaa2-switch: disable the control interface on error path
Revert "flow_offload: action should not be NULL when it is referenced"
iavf: Fix ping is lost after untrusted VF had tried to change MAC
i40e: Fix ATR queue selection
r8152: fix the maximum number of PLA bp for RTL8153C
r8152: fix writing USB_BP2_EN
mptcp: full fully established support after ADD_ADDR
mptcp: fix memory leak on address flush
net/rds: dma_map_sg is entitled to merge entries
net: mscc: ocelot: allow forwarding from bridge ports to the tag_8021q CPU port
net: asix: fix uninit value bugs
ovs: clear skb->tstamp in forwarding path
net: mdio-mux: Handle -EPROBE_DEFER correctly
net: mdio-mux: Don't ignore memory allocation errors
net: mdio-mux: Delete unnecessary devm_kfree
net: dsa: sja1105: fix use-after-free after calling of_find_compatible_node, or worse
sch_cake: fix srchost/dsthost hashing mode
ixgbe, xsk: clean up the resources in ixgbe_xsk_pool_enable error path
net: qlcnic: add missed unlock in qlcnic_83xx_flash_read32
mac80211: fix locking in ieee80211_restart_work()
...
This reverts commit 9ea3e52c5b.
Cited commit added a check to make sure 'action' is not NULL, but
'action' is already dereferenced before the check, when calling
flow_offload_has_one_action().
Therefore, the check does not make any sense and results in a smatch
warning:
include/net/flow_offload.h:322 flow_action_mixed_hw_stats_check() warn:
variable dereferenced before check 'action' (see line 319)
Fix by reverting this commit.
Cc: gushengxian <gushengxian@yulong.com>
Fixes: 9ea3e52c5b ("flow_offload: action should not be NULL when it is referenced")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Link: https://lore.kernel.org/r/20210819105842.1315705-1-idosch@idosch.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
I had forgotten just how sensitive hackbench is to extra pipe wakeups,
and commit 3a34b13a88 ("pipe: make pipe writes always wake up
readers") ended up causing a quite noticeable regression on larger
machines.
Now, hackbench isn't necessarily a hugely meaningful benchmark, and it's
not clear that this matters in real life all that much, but as Mel
points out, it's used often enough when comparing kernels and so the
performance regression shows up like a sore thumb.
It's easy enough to fix at least for the common cases where pipes are
used purely for data transfer, and you never have any exciting poll
usage at all. So set a special 'poll_usage' flag when there is polling
activity, and make the ugly "EPOLLET has crazy legacy expectations"
semantics explicit to only that case.
I would love to limit it to just the broken EPOLLET case, but the pipe
code can't see the difference between epoll and regular select/poll, so
any non-read/write waiting will trigger the extra wakeup behavior. That
is sufficient for at least the hackbench case.
Apart from making the odd extra wakeup cases more explicitly about
EPOLLET, this also makes the extra wakeup be at the _end_ of the pipe
write, not at the first write chunk. That is actually much saner
semantics (as much as you can call any of the legacy edge-triggered
expectations for EPOLLET "sane") since it means that you know the wakeup
will happen once the write is done, rather than possibly in the middle
of one.
[ For stable people: I'm putting a "Fixes" tag on this, but I leave it
up to you to decide whether you actually want to backport it or not.
It likely has no impact outside of synthetic benchmarks - Linus ]
Link: https://lore.kernel.org/lkml/20210802024945.GA8372@xsang-OptiPlex-9020/
Fixes: 3a34b13a88 ("pipe: make pipe writes always wake up readers")
Reported-by: kernel test robot <oliver.sang@intel.com>
Tested-by: Sandeep Patil <sspatil@android.com>
Tested-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull virtio fixes from Michael Tsirkin:
"Fixes in virtio, vhost, and vdpa drivers"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
vdpa/mlx5: Fix queue type selection logic
vdpa/mlx5: Avoid destroying MR on empty iotlb
tools/virtio: fix build
virtio_ring: pull in spinlock header
vringh: pull in spinlock header
virtio-blk: Add validation for block size in config space
vringh: Use wiov->used to check for read/write desc order
virtio_vdpa: reject invalid vq indices
vdpa: Add documentation for vdpa_alloc_device() macro
vDPA/ifcvf: Fix return value check for vdpa_alloc_device()
vp_vdpa: Fix return value check for vdpa_alloc_device()
vdpa_sim: Fix return value check for vdpa_alloc_device()
vhost: Fix the calculation in vhost_overflow()
vhost-vdpa: Fix integer overflow in vhost_vdpa_process_iotlb_update()
virtio_pci: Support surprise removal of virtio pci device
virtio: Protect vqs list access
virtio: Keep vring_del_virtqueue() mirror of VQ create
virtio: Improve vq->broken access to avoid any compiler optimization
Pull irq fixes from Thomas Gleixner:
"A set of fixes for PCI/MSI and x86 interrupt startup:
- Mask all MSI-X entries when enabling MSI-X otherwise stale unmasked
entries stay around e.g. when a crashkernel is booted.
- Enforce masking of a MSI-X table entry when updating it, which
mandatory according to speification
- Ensure that writes to MSI[-X} tables are flushed.
- Prevent invalid bits being set in the MSI mask register
- Properly serialize modifications to the mask cache and the mask
register for multi-MSI.
- Cure the violation of the affinity setting rules on X86 during
interrupt startup which can cause lost and stale interrupts. Move
the initial affinity setting ahead of actualy enabling the
interrupt.
- Ensure that MSI interrupts are completely torn down before freeing
them in the error handling case.
- Prevent an array out of bounds access in the irq timings code"
* tag 'irq-urgent-2021-08-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
driver core: Add missing kernel doc for device::msi_lock
genirq/msi: Ensure deactivation on teardown
genirq/timings: Prevent potential array overflow in __irq_timings_store()
x86/msi: Force affinity setup before startup
x86/ioapic: Force affinity setup before startup
genirq: Provide IRQCHIP_AFFINITY_PRE_STARTUP
PCI/MSI: Protect msi_desc::masked for multi-MSI
PCI/MSI: Use msi_mask_irq() in pci_msi_shutdown()
PCI/MSI: Correct misleading comments
PCI/MSI: Do not set invalid bits in MSI mask
PCI/MSI: Enforce MSI[X] entry updates to be visible
PCI/MSI: Enforce that MSI-X table entry is masked for update
PCI/MSI: Mask all unused MSI-X entries
PCI/MSI: Enable and mask MSI-X early
Pull networking fixes from Jakub Kicinski:
"Networking fixes, including fixes from netfilter, bpf, can and
ieee802154.
The size of this is pretty normal, but we got more fixes for 5.14
changes this week than last week. Nothing major but the trend is the
opposite of what we like. We'll see how the next week goes..
Current release - regressions:
- r8169: fix ASPM-related link-up regressions
- bridge: fix flags interpretation for extern learn fdb entries
- phy: micrel: fix link detection on ksz87xx switch
- Revert "tipc: Return the correct errno code"
- ptp: fix possible memory leak caused by invalid cast
Current release - new code bugs:
- bpf: add missing bpf_read_[un]lock_trace() for syscall program
- bpf: fix potentially incorrect results with bpf_get_local_storage()
- page_pool: mask the page->signature before the checking, avoid dma
mapping leaks
- netfilter: nfnetlink_hook: 5 fixes to information in netlink dumps
- bnxt_en: fix firmware interface issues with PTP
- mlx5: Bridge, fix ageing time
Previous releases - regressions:
- linkwatch: fix failure to restore device state across
suspend/resume
- bareudp: fix invalid read beyond skb's linear data
Previous releases - always broken:
- bpf: fix integer overflow involving bucket_size
- ppp: fix issues when desired interface name is specified via
netlink
- wwan: mhi_wwan_ctrl: fix possible deadlock
- dsa: microchip: ksz8795: fix number of VLAN related bugs
- dsa: drivers: fix broken backpressure in .port_fdb_dump
- dsa: qca: ar9331: make proper initial port defaults
Misc:
- bpf: add lockdown check for probe_write_user helper
- netfilter: conntrack: remove offload_pickup sysctl before 5.14 is
out
- netfilter: conntrack: collect all entries in one cycle,
heuristically slow down garbage collection scans on idle systems to
prevent frequent wake ups"
* tag 'net-5.14-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (87 commits)
vsock/virtio: avoid potential deadlock when vsock device remove
wwan: core: Avoid returning NULL from wwan_create_dev()
net: dsa: sja1105: unregister the MDIO buses during teardown
Revert "tipc: Return the correct errno code"
net: mscc: Fix non-GPL export of regmap APIs
net: igmp: increase size of mr_ifc_count
MAINTAINERS: switch to my OMP email for Renesas Ethernet drivers
tcp_bbr: fix u32 wrap bug in round logic if bbr_init() called after 2B packets
net: pcs: xpcs: fix error handling on failed to allocate memory
net: linkwatch: fix failure to restore device state across suspend/resume
net: bridge: fix memleak in br_add_if()
net: switchdev: zero-initialize struct switchdev_notifier_fdb_info emitted by drivers towards the bridge
net: bridge: fix flags interpretation for extern learn fdb entries
net: dsa: sja1105: fix broken backpressure in .port_fdb_dump
net: dsa: lantiq: fix broken backpressure in .port_fdb_dump
net: dsa: lan9303: fix broken backpressure in .port_fdb_dump
net: dsa: hellcreek: fix broken backpressure in .port_fdb_dump
bpf, core: Fix kernel-doc notation
net: igmp: fix data-race in igmp_ifc_timer_expire()
net: Fix memory leak in ieee802154_raw_deliver
...
A recent change in LLVM causes module_{c,d}tor sections to appear when
CONFIG_K{A,C}SAN are enabled, which results in orphan section warnings
because these are not handled anywhere:
ld.lld: warning: arch/x86/pci/built-in.a(legacy.o):(.text.asan.module_ctor) is being placed in '.text.asan.module_ctor'
ld.lld: warning: arch/x86/pci/built-in.a(legacy.o):(.text.asan.module_dtor) is being placed in '.text.asan.module_dtor'
ld.lld: warning: arch/x86/pci/built-in.a(legacy.o):(.text.tsan.module_ctor) is being placed in '.text.tsan.module_ctor'
Fangrui explains: "the function asan.module_ctor has the SHF_GNU_RETAIN
flag, so it is in a separate section even with -fno-function-sections
(default)".
Place them in the TEXT_TEXT section so that these technologies continue
to work with the newer compiler versions. All of the KASAN and KCSAN
KUnit tests continue to pass after this change.
Cc: stable@vger.kernel.org
Link: https://github.com/ClangBuiltLinux/linux/issues/1432
Link: 7b78956224
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Fangrui Song <maskray@google.com>
Acked-by: Marco Elver <elver@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20210731023107.1932981-1-nathan@kernel.org
get_queue_type() comments that splict virtqueue is preferred, however,
the actual logic preferred packed virtqueues. Since firmware has not
supported packed virtqueues we ended up using split virtqueues as was
desired.
Since we do not advertise support for packed virtqueues, we add a check
to verify split virtqueues are indeed supported.
Fixes: 1a86b377aa ("vdpa/mlx5: Add VDPA driver for supported mlx5 devices")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Link: https://lore.kernel.org/r/20210811053759.66752-1-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Ignore fdb flags when adding port extern learn entries and always set
BR_FDB_LOCAL flag when adding bridge extern learn entries. This is
closest to the behaviour we had before and avoids breaking any use cases
which were allowed.
This patch fixes iproute2 calls which assume NUD_PERMANENT and were
allowed before, example:
$ bridge fdb add 00:11:22:33:44:55 dev swp1 extern_learn
Extern learn entries are allowed to roam, but do not expire, so static
or dynamic flags make no sense for them.
Also add a comment for future reference.
Fixes: eb100e0e24 ("net: bridge: allow to add externally learned entries from user-space")
Fixes: 0541a62932 ("net: bridge: validate the NUD_PERMANENT bit when adding an extern_learn FDB entry")
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Tested-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Nikolay Aleksandrov <nikolay@nvidia.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20210810110010.43859-1-razor@blackwall.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Daniel Borkmann says:
====================
bpf 2021-08-10
We've added 5 non-merge commits during the last 2 day(s) which contain
a total of 7 files changed, 27 insertions(+), 15 deletions(-).
1) Fix missing bpf_read_lock_trace() context for BPF loader progs, from Yonghong Song.
2) Fix corner case where BPF prog retrieves wrong local storage, also from Yonghong Song.
3) Restrict availability of BPF write_user helper behind lockdown, from Daniel Borkmann.
4) Fix multiple kernel-doc warnings in BPF core, from Randy Dunlap.
* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
bpf, core: Fix kernel-doc notation
bpf: Fix potentially incorrect results with bpf_get_local_storage()
bpf: Add missing bpf_read_[un]lock_trace() for syscall program
bpf: Add lockdown check for probe_write_user helper
bpf: Add _kernel suffix to internal lockdown_bpf_read
====================
Link: https://lore.kernel.org/r/20210810144025.22814-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
X86 IO/APIC and MSI interrupts (when used without interrupts remapping)
require that the affinity setup on startup is done before the interrupt is
enabled for the first time as the non-remapped operation mode cannot safely
migrate enabled interrupts from arbitrary contexts. Provide a new irq chip
flag which allows affected hardware to request this.
This has to be opt-in because there have been reports in the past that some
interrupt chips cannot handle affinity setting before startup.
Fixes: 1840475676 ("genirq: Expose default irq affinity mask (take 3)")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Marc Zyngier <maz@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20210729222542.779791738@linutronix.de
Multi-MSI uses a single MSI descriptor and there is a single mask register
when the device supports per vector masking. To avoid reading back the mask
register the value is cached in the MSI descriptor and updates are done by
clearing and setting bits in the cache and writing it to the device.
But nothing protects msi_desc::masked and the mask register from being
modified concurrently on two different CPUs for two different Linux
interrupts which belong to the same multi-MSI descriptor.
Add a lock to struct device and protect any operation on the mask and the
mask register with it.
This makes the update of msi_desc::masked unconditional, but there is no
place which requires a modification of the hardware register without
updating the masked cache.
msi_mask_irq() is now an empty wrapper which will be cleaned up in follow
up changes.
The problem goes way back to the initial support of multi-MSI, but picking
the commit which introduced the mask cache is a valid cut off point
(2.6.30).
Fixes: f2440d9acb ("PCI MSI: Refactor interrupt masking code")
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Marc Zyngier <maz@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20210729222542.726833414@linutronix.de
Saeed Mahameed says:
====================
mlx5 fixes 2021-08-09
This series introduces fixes to mlx5 driver.
Please pull and let me know if there is any problem.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit b910eaaaa4 ("bpf: Fix NULL pointer dereference in bpf_get_local_storage()
helper") fixed a bug for bpf_get_local_storage() helper so different tasks
won't mess up with each other's percpu local storage.
The percpu data contains 8 slots so it can hold up to 8 contexts (same or
different tasks), for 8 different program runs, at the same time. This in
general is sufficient. But our internal testing showed the following warning
multiple times:
[...]
warning: WARNING: CPU: 13 PID: 41661 at include/linux/bpf-cgroup.h:193
__cgroup_bpf_run_filter_sock_ops+0x13e/0x180
RIP: 0010:__cgroup_bpf_run_filter_sock_ops+0x13e/0x180
<IRQ>
tcp_call_bpf.constprop.99+0x93/0xc0
tcp_conn_request+0x41e/0xa50
? tcp_rcv_state_process+0x203/0xe00
tcp_rcv_state_process+0x203/0xe00
? sk_filter_trim_cap+0xbc/0x210
? tcp_v6_inbound_md5_hash.constprop.41+0x44/0x160
tcp_v6_do_rcv+0x181/0x3e0
tcp_v6_rcv+0xc65/0xcb0
ip6_protocol_deliver_rcu+0xbd/0x450
ip6_input_finish+0x11/0x20
ip6_input+0xb5/0xc0
ip6_sublist_rcv_finish+0x37/0x50
ip6_sublist_rcv+0x1dc/0x270
ipv6_list_rcv+0x113/0x140
__netif_receive_skb_list_core+0x1a0/0x210
netif_receive_skb_list_internal+0x186/0x2a0
gro_normal_list.part.170+0x19/0x40
napi_complete_done+0x65/0x150
mlx5e_napi_poll+0x1ae/0x680
__napi_poll+0x25/0x120
net_rx_action+0x11e/0x280
__do_softirq+0xbb/0x271
irq_exit_rcu+0x97/0xa0
common_interrupt+0x7f/0xa0
</IRQ>
asm_common_interrupt+0x1e/0x40
RIP: 0010:bpf_prog_1835a9241238291a_tw_egress+0x5/0xbac
? __cgroup_bpf_run_filter_skb+0x378/0x4e0
? do_softirq+0x34/0x70
? ip6_finish_output2+0x266/0x590
? ip6_finish_output+0x66/0xa0
? ip6_output+0x6c/0x130
? ip6_xmit+0x279/0x550
? ip6_dst_check+0x61/0xd0
[...]
Using drgn [0] to dump the percpu buffer contents showed that on this CPU
slot 0 is still available, but slots 1-7 are occupied and those tasks in
slots 1-7 mostly don't exist any more. So we might have issues in
bpf_cgroup_storage_unset().
Further debugging confirmed that there is a bug in bpf_cgroup_storage_unset().
Currently, it tries to unset "current" slot with searching from the start.
So the following sequence is possible:
1. A task is running and claims slot 0
2. Running BPF program is done, and it checked slot 0 has the "task"
and ready to reset it to NULL (not yet).
3. An interrupt happens, another BPF program runs and it claims slot 1
with the *same* task.
4. The unset() in interrupt context releases slot 0 since it matches "task".
5. Interrupt is done, the task in process context reset slot 0.
At the end, slot 1 is not reset and the same process can continue to occupy
slots 2-7 and finally, when the above step 1-5 is repeated again, step 3 BPF
program won't be able to claim an empty slot and a warning will be issued.
To fix the issue, for unset() function, we should traverse from the last slot
to the first. This way, the above issue can be avoided.
The same reverse traversal should also be done in bpf_get_local_storage() helper
itself. Otherwise, incorrect local storage may be returned to BPF program.
[0] https://github.com/osandov/drgn
Fixes: b910eaaaa4 ("bpf: Fix NULL pointer dereference in bpf_get_local_storage() helper")
Signed-off-by: Yonghong Song <yhs@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210810010413.1976277-1-yhs@fb.com
Back then, commit 96ae522795 ("bpf: Add bpf_probe_write_user BPF helper
to be called in tracers") added the bpf_probe_write_user() helper in order
to allow to override user space memory. Its original goal was to have a
facility to "debug, divert, and manipulate execution of semi-cooperative
processes" under CAP_SYS_ADMIN. Write to kernel was explicitly disallowed
since it would otherwise tamper with its integrity.
One use case was shown in cf9b1199de ("samples/bpf: Add test/example of
using bpf_probe_write_user bpf helper") where the program DNATs traffic
at the time of connect(2) syscall, meaning, it rewrites the arguments to
a syscall while they're still in userspace, and before the syscall has a
chance to copy the argument into kernel space. These days we have better
mechanisms in BPF for achieving the same (e.g. for load-balancers), but
without having to write to userspace memory.
Of course the bpf_probe_write_user() helper can also be used to abuse
many other things for both good or bad purpose. Outside of BPF, there is
a similar mechanism for ptrace(2) such as PTRACE_PEEK{TEXT,DATA} and
PTRACE_POKE{TEXT,DATA}, but would likely require some more effort.
Commit 96ae522795 explicitly dedicated the helper for experimentation
purpose only. Thus, move the helper's availability behind a newly added
LOCKDOWN_BPF_WRITE_USER lockdown knob so that the helper is disabled under
the "integrity" mode. More fine-grained control can be implemented also
from LSM side with this change.
Fixes: 96ae522795 ("bpf: Add bpf_probe_write_user BPF helper to be called in tracers")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
The CQ destroy is performed based on the IRQ number that is stored in
cq->irqn. That number wasn't set explicitly during CQ creation and as
expected some of the API users of mlx5_core_create_cq() forgot to update
it.
This caused to wrong synchronization call of the wrong IRQ with a number
0 instead of the real one.
As a fix, set the IRQ number directly in the mlx5_core_create_cq() and
update all users accordingly.
Fixes: 1a86b377aa ("vdpa/mlx5: Add VDPA driver for supported mlx5 devices")
Fixes: ef1659ade3 ("IB/mlx5: Add DEVX support for CQ events")
Signed-off-by: Shay Drory <shayd@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Rename LOCKDOWN_BPF_READ into LOCKDOWN_BPF_READ_KERNEL so we have naming
more consistent with a LOCKDOWN_BPF_WRITE_USER option that we are adding.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Pull tty/serial fixes from Greg KH:
"Here are some small tty/serial driver fixes for 5.14-rc5 to resolve a
number of reported problems.
They include:
- mips serial driver fixes
- 8250 driver fixes for reported problems
- fsl_lpuart driver fixes
- other tiny driver fixes
All have been in linux-next for a while with no reported problems"
* tag 'tty-5.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
serial: 8250_pci: Avoid irq sharing for MSI(-X) interrupts.
serial: 8250_mtk: fix uart corruption issue when rx power off
tty: serial: fsl_lpuart: fix the wrong return value in lpuart32_get_mctrl
serial: 8250_pci: Enumerate Elkhart Lake UARTs via dedicated driver
serial: 8250: fix handle_irq locking
serial: tegra: Only print FIFO error message when an error occurs
MIPS: Malta: Do not byte-swap accesses to the CBUS UART
serial: 8250: Mask out floating 16/32-bit bus bits
serial: max310x: Unprepare and disable clock in error path
Pull USB driver fixes from Greg KH:
"Here are some small USB driver fixes for 5.14-rc5. They resolve a
number of small reported issues, including:
- cdnsp driver fixes
- usb serial driver fixes and device id updates
- usb gadget hid fixes
- usb host driver fixes
- usb dwc3 driver fixes
- other usb gadget driver fixes
All of these have been in linux-next for a while with no reported
issues"
* tag 'usb-5.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (21 commits)
usb: typec: tcpm: Keep other events when receiving FRS and Sourcing_vbus events
usb: dwc3: gadget: Avoid runtime resume if disabling pullup
usb: dwc3: gadget: Use list_replace_init() before traversing lists
USB: serial: ftdi_sio: add device ID for Auto-M3 OP-COM v2
USB: serial: pl2303: fix GT type detection
USB: serial: option: add Telit FD980 composition 0x1056
USB: serial: pl2303: fix HX type detection
USB: serial: ch341: fix character loss at high transfer rates
usb: cdnsp: Fix the IMAN_IE_SET and IMAN_IE_CLEAR macro
usb: cdnsp: Fixed issue with ZLP
usb: cdnsp: Fix incorrect supported maximum speed
usb: cdns3: Fixed incorrect gadget state
usb: gadget: f_hid: idle uses the highest byte for duration
Revert "thunderbolt: Hide authorized attribute if router does not support PCIe tunnels"
usb: otg-fsm: Fix hrtimer list corruption
usb: host: ohci-at91: suspend/resume ports after/before OHCI accesses
usb: musb: Fix suspend and resume issues for PHYs on I2C and SPI
usb: gadget: f_hid: added GET_IDLE and SET_IDLE handlers
usb: gadget: f_hid: fixed NULL pointer dereference
usb: gadget: remove leaked entry from udc driver list
...
Pull ARM SoC fixes from Arnd Bergmann:
"Lots of small fixes for Arm SoCs this time, nothing too worrying:
- omap/beaglebone boot regression fix in gpt12 timer
- revert for i.mx8 soc driver breaking as a platform_driver
- kexec/kdump fixes for op-tee
- various fixes for incorrect DT settings on imx, mvebu, omap, stm32,
and tegra causing problems.
- device tree fixes for static checks in nomadik, versatile, stm32
- code fixes for issues found in build testing and with static
checking on tegra, ixp4xx, imx, omap"
* tag 'soc-fixes-5.14-2' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (36 commits)
soc: ixp4xx/qmgr: fix invalid __iomem access
soc: ixp4xx: fix printing resources
ARM: ixp4xx: goramo_mlr depends on old PCI driver
ARM: ixp4xx: fix compile-testing soc drivers
soc/tegra: Make regulator couplers depend on CONFIG_REGULATOR
ARM: dts: nomadik: Fix up interrupt controller node names
ARM: dts: stm32: Fix touchscreen IRQ line assignment on DHCOM
ARM: dts: stm32: Disable LAN8710 EDPD on DHCOM
ARM: dts: stm32: Prefer HW RTC on DHCOM SoM
omap5-board-common: remove not physically existing vdds_1v8_main fixed-regulator
ARM: dts: am437x-l4: fix typo in can@0 node
ARM: dts: am43x-epos-evm: Reduce i2c0 bus speed for tps65218
bus: ti-sysc: AM3: RNG is GP only
ARM: omap2+: hwmod: fix potential NULL pointer access
arm64: dts: armada-3720-turris-mox: remove mrvl,i2c-fast-mode
arm64: dts: armada-3720-turris-mox: fixed indices for the SDHC controllers
ARM: dts: imx: Swap M53Menlo pinctrl_power_button/pinctrl_power_out pins
ARM: imx: fix missing 3rd argument in macro imx_mmdc_perf_init
ARM: dts: colibri-imx6ull: limit SDIO clock to 25MHz
arm64: dts: ls1028: sl28: fix networking for variant 2
...
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
The following patchset contains Netfilter fixes for net:
1) Restrict range element expansion in ipset to avoid soft lockup,
from Jozsef Kadlecsik.
2) Memleak in error path for nf_conntrack_bridge for IPv4 packets,
from Yajun Deng.
3) Simplify conntrack garbage collection strategy to avoid frequent
wake-ups, from Florian Westphal.
4) Fix NFNLA_HOOK_FUNCTION_NAME string, do not include module name.
5) Missing chain family netlink attribute in chain description
in nfnetlink_hook.
6) Incorrect sequence number on nfnetlink_hook dumps.
7) Use netlink request family in reply message for consistency.
8) Remove offload_pickup sysctl, use conntrack for established state
instead, from Florian Westphal.
9) Translate NFPROTO_INET/ingress to NFPROTO_NETDEV/ingress, since
NFPROTO_INET is not exposed through nfnetlink_hook.
* git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf:
netfilter: nfnetlink_hook: translate inet ingress to netdev
netfilter: conntrack: remove offload_pickup sysctl again
netfilter: nfnetlink_hook: Use same family as request message
netfilter: nfnetlink_hook: use the sequence number of the request message
netfilter: nfnetlink_hook: missing chain family
netfilter: nfnetlink_hook: strip off module name from hookfn
netfilter: conntrack: collect all entries in one cycle
netfilter: nf_conntrack_bridge: Fix memory leak when error
netfilter: ipset: Limit the maximal range of consecutive elements to add/delete
====================
Link: https://lore.kernel.org/r/20210806151149.6356-1-pablo@netfilter.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
These two sysctls were added because the hardcoded defaults (2 minutes,
tcp, 30 seconds, udp) turned out to be too low for some setups.
They appeared in 5.14-rc1 so it should be fine to remove it again.
Marcelo convinced me that there should be no difference between a flow
that was offloaded vs. a flow that was not wrt. timeout handling.
Thus the default is changed to those for TCP established and UDP stream,
5 days and 120 seconds, respectively.
Marcelo also suggested to account for the timeout value used for the
offloading, this avoids increase beyond the value in the conntrack-sysctl
and will also instantly expire the conntrack entry with altered sysctls.
Example:
nf_conntrack_udp_timeout_stream=60
nf_flowtable_udp_timeout=60
This will remove offloaded udp flows after one minute, rather than two.
An earlier version of this patch also cleared the ASSURED bit to
allow nf_conntrack to evict the entry via early_drop (i.e., table full).
However, it looks like we can safely assume that connection timed out
via HW is still in established state, so this isn't needed.
Quoting Oz:
[..] the hardware sends all packets with a set FIN flags to sw.
[..] Connections that are aged in hardware are expected to be in the
established state.
In case it turns out that back-to-sw-path transition can occur for
'dodgy' connections too (e.g., one side disappeared while software-path
would have been in RETRANS timeout), we can adjust this later.
Cc: Oz Shlomo <ozsh@nvidia.com>
Cc: Paul Blakey <paulb@nvidia.com>
Suggested-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The family is relevant for pseudo-families like NFPROTO_INET
otherwise the user needs to rely on the hook function name to
differentiate it from NFPROTO_IPV4 and NFPROTO_IPV6 names.
Add nfnl_hook_chain_desc_attributes instead of using the existing
NFTA_CHAIN_* attributes, since these do not provide a family number.
Fixes: e2cf17d377 ("netfilter: add new hook nfnl subsystem")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Pull networking fixes from Jakub Kicinski:
"Including fixes from ipsec.
Current release - regressions:
- sched: taprio: fix init procedure to avoid inf loop when dumping
- sctp: move the active_key update after sh_keys is added
Current release - new code bugs:
- sparx5: fix build with old GCC & bitmask on 32-bit targets
Previous releases - regressions:
- xfrm: redo the PREEMPT_RT RCU vs hash_resize_mutex deadlock fix
- xfrm: fixes for the compat netlink attribute translator
- phy: micrel: Fix detection of ksz87xx switch
Previous releases - always broken:
- gro: set inner transport header offset in tcp/udp GRO hook to avoid
crashes when such packets reach GSO
- vsock: handle VIRTIO_VSOCK_OP_CREDIT_REQUEST, as required by spec
- dsa: sja1105: fix static FDB entries on SJA1105P/Q/R/S and SJA1110
- bridge: validate the NUD_PERMANENT bit when adding an extern_learn
FDB entry
- usb: lan78xx: don't modify phy_device state concurrently
- usb: pegasus: check for errors of IO routines"
* tag 'net-5.14-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (48 commits)
net: vxge: fix use-after-free in vxge_device_unregister
net: fec: fix use-after-free in fec_drv_remove
net: pegasus: fix uninit-value in get_interrupt_interval
net: ethernet: ti: am65-cpsw: fix crash in am65_cpsw_port_offload_fwd_mark_update()
bnx2x: fix an error code in bnx2x_nic_load()
net: wwan: iosm: fix recursive lock acquire in unregister
net: wwan: iosm: correct data protocol mask bit
net: wwan: iosm: endianness type correction
net: wwan: iosm: fix lkp buildbot warning
net: usb: lan78xx: don't modify phy_device state concurrently
docs: networking: netdevsim rules
net: usb: pegasus: Remove the changelog and DRIVER_VERSION.
net: usb: pegasus: Check the return value of get_geristers() and friends;
net/prestera: Fix devlink groups leakage in error flow
net: sched: fix lockdep_set_class() typo error for sch->seqlock
net: dsa: qca: ar9331: reorder MDIO write sequence
VSOCK: handle VIRTIO_VSOCK_OP_CREDIT_REQUEST
mptcp: drop unused rcu member in mptcp_pm_addr_entry
net: ipv6: fix returned variable type in ip6_skb_dst_mtu
nfp: update ethtool reporting of pauseframe control
...
syzbot is hitting might_sleep() warning at hci_sock_dev_event() due to
calling lock_sock() with rw spinlock held [1].
It seems that history of this locking problem is a trial and error.
Commit b40df5743e ("[PATCH] bluetooth: fix socket locking in
hci_sock_dev_event()") in 2.6.21-rc4 changed bh_lock_sock() to
lock_sock() as an attempt to fix lockdep warning.
Then, commit 4ce61d1c7a ("[BLUETOOTH]: Fix locking in
hci_sock_dev_event().") in 2.6.22-rc2 changed lock_sock() to
local_bh_disable() + bh_lock_sock_nested() as an attempt to fix the
sleep in atomic context warning.
Then, commit 4b5dd696f8 ("Bluetooth: Remove local_bh_disable() from
hci_sock.c") in 3.3-rc1 removed local_bh_disable().
Then, commit e305509e67 ("Bluetooth: use correct lock to prevent UAF
of hdev object") in 5.13-rc5 again changed bh_lock_sock_nested() to
lock_sock() as an attempt to fix CVE-2021-3573.
This difficulty comes from current implementation that
hci_sock_dev_event(HCI_DEV_UNREG) is responsible for dropping all
references from sockets because hci_unregister_dev() immediately
reclaims resources as soon as returning from
hci_sock_dev_event(HCI_DEV_UNREG).
But the history suggests that hci_sock_dev_event(HCI_DEV_UNREG) was not
doing what it should do.
Therefore, instead of trying to detach sockets from device, let's accept
not detaching sockets from device at hci_sock_dev_event(HCI_DEV_UNREG),
by moving actual cleanup of resources from hci_unregister_dev() to
hci_cleanup_dev() which is called by bt_host_release() when all
references to this unregistered device (which is a kobject) are gone.
Since hci_sock_dev_event(HCI_DEV_UNREG) no longer resets
hci_pi(sk)->hdev, we need to check whether this device was unregistered
and return an error based on HCI_UNREGISTER flag. There might be subtle
behavioral difference in "monitor the hdev" functionality; please report
if you found something went wrong due to this patch.
Link: https://syzkaller.appspot.com/bug?extid=a5df189917e79d5e59c9 [1]
Reported-by: syzbot <syzbot+a5df189917e79d5e59c9@syzkaller.appspotmail.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Fixes: e305509e67 ("Bluetooth: use correct lock to prevent UAF of hdev object")
Acked-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Steffen Klassert says:
====================
pull request (net): ipsec 2021-08-04
1) Fix a sysbot reported memory leak in xfrm_user_rcv_msg.
From Pavel Skripkin.
2) Revert "xfrm: policy: Read seqcount outside of rcu-read side
in xfrm_policy_lookup_bytype". This commit tried to fix a
lockin bug, but only cured some of the symptoms. A proper
fix is applied on top of this revert.
3) Fix a locking bug on xfrm state hash resize. A recent change
on sequence counters accidentally repaced a spinlock by a mutex.
Fix from Frederic Weisbecker.
4) Fix possible user-memory-access in xfrm_user_rcv_msg_compat().
From Dmitry Safonov.
5) Add initialiation sefltest fot xfrm_spdattr_type_t.
From Dmitry Safonov.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The range size of consecutive elements were not limited. Thus one could
define a huge range which may result soft lockup errors due to the long
execution time. Now the range size is limited to 2^20 entries.
Reported-by: Brad Spengler <spender@grsecurity.net>
Signed-off-by: Jozsef Kadlecsik <kadlec@netfilter.org>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
The patch fixing the returned value of ip6_skb_dst_mtu (int -> unsigned
int) was rebased between its initial review and the version applied. In
the meantime fade56410c was applied, which added a new variable (int)
used as the returned value. This lead to a mismatch between the function
prototype and the variable used as the return value.
Fixes: 40fc3054b4 ("net: ipv6: fix return value of ip6_skb_dst_mtu")
Cc: Vadim Fedorenko <vfedorenko@novek.ru>
Signed-off-by: Antoine Tenart <atenart@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Provide missing kdoc of fields of struct tcf_pkt_info and tcf_ematch_ops.
Found using ./scripts/kernel-doc -none -Werror include/net/pkt_cls.h
Signed-off-by: Bijie Xu <bijie.xu@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Correct mismatch between the name of flow_offload_has_one_action()
and its kdoc entry.
Found using ./scripts/kernel-doc -Werror -none include/net/flow_offload.h
Signed-off-by: Bijie Xu <bijie.xu@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
tee: Improve support for kexec and kdump
This fixes several bugs uncovered while exercising the OP-TEE, ftpm
(firmware TPM), and tee_bnxt_fw (Broadcom BNXT firmware manager) drivers
with kexec and kdump (emergency kexec) based workflows.
* tag 'tee-kexec-fixes-for-v5.14' of git://git.linaro.org:/people/jens.wiklander/linux-tee:
firmware: tee_bnxt: Release TEE shm, session, and context during kexec
tpm_ftpm_tee: Free and unregister TEE shared memory during kexec
tee: Correct inappropriate usage of TEE_SHM_DMA_BUF flag
tee: add tee_shm_alloc_kernel_buf()
optee: Clear stale cache entries during initialization
optee: fix tee out of memory failure seen during kexec reboot
optee: Refuse to load the driver under the kdump kernel
optee: Fix memory leak when failing to register shm pages
Link: https://lore.kernel.org/r/20210726081039.GA2482361@jade
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Pull networking fixes from Jakub Kicinski:
"Networking fixes for 5.14-rc4, including fixes from bpf, can, WiFi
(mac80211) and netfilter trees.
Current release - regressions:
- mac80211: fix starting aggregation sessions on mesh interfaces
Current release - new code bugs:
- sctp: send pmtu probe only if packet loss in Search Complete state
- bnxt_en: add missing periodic PHC overflow check
- devlink: fix phys_port_name of virtual port and merge error
- hns3: change the method of obtaining default ptp cycle
- can: mcba_usb_start(): add missing urb->transfer_dma initialization
Previous releases - regressions:
- set true network header for ECN decapsulation
- mlx5e: RX, avoid possible data corruption w/ relaxed ordering and
LRO
- phy: re-add check for PHY_BRCM_DIS_TXCRXC_NOENRGY on the BCM54811
PHY
- sctp: fix return value check in __sctp_rcv_asconf_lookup
Previous releases - always broken:
- bpf:
- more spectre corner case fixes, introduce a BPF nospec
instruction for mitigating Spectre v4
- fix OOB read when printing XDP link fdinfo
- sockmap: fix cleanup related races
- mac80211: fix enabling 4-address mode on a sta vif after assoc
- can:
- raw: raw_setsockopt(): fix raw_rcv panic for sock UAF
- j1939: j1939_session_deactivate(): clarify lifetime of session
object, avoid UAF
- fix number of identical memory leaks in USB drivers
- tipc:
- do not blindly write skb_shinfo frags when doing decryption
- fix sleeping in tipc accept routine"
* tag 'net-5.14-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (91 commits)
gve: Update MAINTAINERS list
can: esd_usb2: fix memory leak
can: ems_usb: fix memory leak
can: usb_8dev: fix memory leak
can: mcba_usb_start(): add missing urb->transfer_dma initialization
can: hi311x: fix a signedness bug in hi3110_cmd()
MAINTAINERS: add Yasushi SHOJI as reviewer for the Microchip CAN BUS Analyzer Tool driver
bpf: Fix leakage due to insufficient speculative store bypass mitigation
bpf: Introduce BPF nospec instruction for mitigating Spectre v4
sis900: Fix missing pci_disable_device() in probe and remove
net: let flow have same hash in two directions
nfc: nfcsim: fix use after free during module unload
tulip: windbond-840: Fix missing pci_disable_device() in probe and remove
sctp: fix return value check in __sctp_rcv_asconf_lookup
nfc: s3fwrn5: fix undefined parameter values in dev_err()
net/mlx5: Fix mlx5_vport_tbl_attr chain from u16 to u32
net/mlx5e: Fix nullptr in mlx5e_hairpin_get_mdev()
net/mlx5: Unload device upon firmware fatal error
net/mlx5e: Fix page allocation failure for ptp-RQ over SF
net/mlx5e: Fix page allocation failure for trap-RQ over SF
...
Pull HID fixes from Jiri Kosina:
- resume timing fix for intel-ish driver (Ye Xiang)
- fix for using incorrect MMIO register in amd_sfh driver (Dylan
MacKenzie)
- Cintiq 24HDT / 27QHDT regression fix and touch processing fix for
Wacom driver (Jason Gerecke)
- device removal bugfix for ft260 driver (Michael Zaidman)
- other small assorted fixes
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:
HID: ft260: fix device removal due to USB disconnect
HID: wacom: Skip processing of touches with negative slot values
HID: wacom: Re-enable touch by default for Cintiq 24HDT / 27QHDT
HID: Kconfig: Fix spelling mistake "Uninterruptable" -> "Uninterruptible"
HID: apple: Add support for Keychron K1 wireless keyboard
HID: fix typo in Kconfig
HID: ft260: fix format type warning in ft260_word_show()
HID: amd_sfh: Use correct MMIO register for DMA address
HID: asus: Remove check for same LED brightness on set
HID: intel-ish-hid: use async resume function
Daniel Borkmann says:
====================
pull-request: bpf 2021-07-29
The following pull-request contains BPF updates for your *net* tree.
We've added 9 non-merge commits during the last 14 day(s) which contain
a total of 20 files changed, 446 insertions(+), 138 deletions(-).
The main changes are:
1) Fix UBSAN out-of-bounds splat for showing XDP link fdinfo, from Lorenz Bauer.
2) Fix insufficient Spectre v4 mitigation in BPF runtime, from Daniel Borkmann,
Piotr Krysiuk and Benedict Schlueter.
3) Batch of fixes for BPF sockmap found under stress testing, from John Fastabend.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Spectre v4 gadgets make use of memory disambiguation, which is a set of
techniques that execute memory access instructions, that is, loads and
stores, out of program order; Intel's optimization manual, section 2.4.4.5:
A load instruction micro-op may depend on a preceding store. Many
microarchitectures block loads until all preceding store addresses are
known. The memory disambiguator predicts which loads will not depend on
any previous stores. When the disambiguator predicts that a load does
not have such a dependency, the load takes its data from the L1 data
cache. Eventually, the prediction is verified. If an actual conflict is
detected, the load and all succeeding instructions are re-executed.
af86ca4e30 ("bpf: Prevent memory disambiguation attack") tried to mitigate
this attack by sanitizing the memory locations through preemptive "fast"
(low latency) stores of zero prior to the actual "slow" (high latency) store
of a pointer value such that upon dependency misprediction the CPU then
speculatively executes the load of the pointer value and retrieves the zero
value instead of the attacker controlled scalar value previously stored at
that location, meaning, subsequent access in the speculative domain is then
redirected to the "zero page".
The sanitized preemptive store of zero prior to the actual "slow" store is
done through a simple ST instruction based on r10 (frame pointer) with
relative offset to the stack location that the verifier has been tracking
on the original used register for STX, which does not have to be r10. Thus,
there are no memory dependencies for this store, since it's only using r10
and immediate constant of zero; hence af86ca4e30 /assumed/ a low latency
operation.
However, a recent attack demonstrated that this mitigation is not sufficient
since the preemptive store of zero could also be turned into a "slow" store
and is thus bypassed as well:
[...]
// r2 = oob address (e.g. scalar)
// r7 = pointer to map value
31: (7b) *(u64 *)(r10 -16) = r2
// r9 will remain "fast" register, r10 will become "slow" register below
32: (bf) r9 = r10
// JIT maps BPF reg to x86 reg:
// r9 -> r15 (callee saved)
// r10 -> rbp
// train store forward prediction to break dependency link between both r9
// and r10 by evicting them from the predictor's LRU table.
33: (61) r0 = *(u32 *)(r7 +24576)
34: (63) *(u32 *)(r7 +29696) = r0
35: (61) r0 = *(u32 *)(r7 +24580)
36: (63) *(u32 *)(r7 +29700) = r0
37: (61) r0 = *(u32 *)(r7 +24584)
38: (63) *(u32 *)(r7 +29704) = r0
39: (61) r0 = *(u32 *)(r7 +24588)
40: (63) *(u32 *)(r7 +29708) = r0
[...]
543: (61) r0 = *(u32 *)(r7 +25596)
544: (63) *(u32 *)(r7 +30716) = r0
// prepare call to bpf_ringbuf_output() helper. the latter will cause rbp
// to spill to stack memory while r13/r14/r15 (all callee saved regs) remain
// in hardware registers. rbp becomes slow due to push/pop latency. below is
// disasm of bpf_ringbuf_output() helper for better visual context:
//
// ffffffff8117ee20: 41 54 push r12
// ffffffff8117ee22: 55 push rbp
// ffffffff8117ee23: 53 push rbx
// ffffffff8117ee24: 48 f7 c1 fc ff ff ff test rcx,0xfffffffffffffffc
// ffffffff8117ee2b: 0f 85 af 00 00 00 jne ffffffff8117eee0 <-- jump taken
// [...]
// ffffffff8117eee0: 49 c7 c4 ea ff ff ff mov r12,0xffffffffffffffea
// ffffffff8117eee7: 5b pop rbx
// ffffffff8117eee8: 5d pop rbp
// ffffffff8117eee9: 4c 89 e0 mov rax,r12
// ffffffff8117eeec: 41 5c pop r12
// ffffffff8117eeee: c3 ret
545: (18) r1 = map[id:4]
547: (bf) r2 = r7
548: (b7) r3 = 0
549: (b7) r4 = 4
550: (85) call bpf_ringbuf_output#194288
// instruction 551 inserted by verifier \
551: (7a) *(u64 *)(r10 -16) = 0 | /both/ are now slow stores here
// storing map value pointer r7 at fp-16 | since value of r10 is "slow".
552: (7b) *(u64 *)(r10 -16) = r7 /
// following "fast" read to the same memory location, but due to dependency
// misprediction it will speculatively execute before insn 551/552 completes.
553: (79) r2 = *(u64 *)(r9 -16)
// in speculative domain contains attacker controlled r2. in non-speculative
// domain this contains r7, and thus accesses r7 +0 below.
554: (71) r3 = *(u8 *)(r2 +0)
// leak r3
As can be seen, the current speculative store bypass mitigation which the
verifier inserts at line 551 is insufficient since /both/, the write of
the zero sanitation as well as the map value pointer are a high latency
instruction due to prior memory access via push/pop of r10 (rbp) in contrast
to the low latency read in line 553 as r9 (r15) which stays in hardware
registers. Thus, architecturally, fp-16 is r7, however, microarchitecturally,
fp-16 can still be r2.
Initial thoughts to address this issue was to track spilled pointer loads
from stack and enforce their load via LDX through r10 as well so that /both/
the preemptive store of zero /as well as/ the load use the /same/ register
such that a dependency is created between the store and load. However, this
option is not sufficient either since it can be bypassed as well under
speculation. An updated attack with pointer spill/fills now _all_ based on
r10 would look as follows:
[...]
// r2 = oob address (e.g. scalar)
// r7 = pointer to map value
[...]
// longer store forward prediction training sequence than before.
2062: (61) r0 = *(u32 *)(r7 +25588)
2063: (63) *(u32 *)(r7 +30708) = r0
2064: (61) r0 = *(u32 *)(r7 +25592)
2065: (63) *(u32 *)(r7 +30712) = r0
2066: (61) r0 = *(u32 *)(r7 +25596)
2067: (63) *(u32 *)(r7 +30716) = r0
// store the speculative load address (scalar) this time after the store
// forward prediction training.
2068: (7b) *(u64 *)(r10 -16) = r2
// preoccupy the CPU store port by running sequence of dummy stores.
2069: (63) *(u32 *)(r7 +29696) = r0
2070: (63) *(u32 *)(r7 +29700) = r0
2071: (63) *(u32 *)(r7 +29704) = r0
2072: (63) *(u32 *)(r7 +29708) = r0
2073: (63) *(u32 *)(r7 +29712) = r0
2074: (63) *(u32 *)(r7 +29716) = r0
2075: (63) *(u32 *)(r7 +29720) = r0
2076: (63) *(u32 *)(r7 +29724) = r0
2077: (63) *(u32 *)(r7 +29728) = r0
2078: (63) *(u32 *)(r7 +29732) = r0
2079: (63) *(u32 *)(r7 +29736) = r0
2080: (63) *(u32 *)(r7 +29740) = r0
2081: (63) *(u32 *)(r7 +29744) = r0
2082: (63) *(u32 *)(r7 +29748) = r0
2083: (63) *(u32 *)(r7 +29752) = r0
2084: (63) *(u32 *)(r7 +29756) = r0
2085: (63) *(u32 *)(r7 +29760) = r0
2086: (63) *(u32 *)(r7 +29764) = r0
2087: (63) *(u32 *)(r7 +29768) = r0
2088: (63) *(u32 *)(r7 +29772) = r0
2089: (63) *(u32 *)(r7 +29776) = r0
2090: (63) *(u32 *)(r7 +29780) = r0
2091: (63) *(u32 *)(r7 +29784) = r0
2092: (63) *(u32 *)(r7 +29788) = r0
2093: (63) *(u32 *)(r7 +29792) = r0
2094: (63) *(u32 *)(r7 +29796) = r0
2095: (63) *(u32 *)(r7 +29800) = r0
2096: (63) *(u32 *)(r7 +29804) = r0
2097: (63) *(u32 *)(r7 +29808) = r0
2098: (63) *(u32 *)(r7 +29812) = r0
// overwrite scalar with dummy pointer; same as before, also including the
// sanitation store with 0 from the current mitigation by the verifier.
2099: (7a) *(u64 *)(r10 -16) = 0 | /both/ are now slow stores here
2100: (7b) *(u64 *)(r10 -16) = r7 | since store unit is still busy.
// load from stack intended to bypass stores.
2101: (79) r2 = *(u64 *)(r10 -16)
2102: (71) r3 = *(u8 *)(r2 +0)
// leak r3
[...]
Looking at the CPU microarchitecture, the scheduler might issue loads (such
as seen in line 2101) before stores (line 2099,2100) because the load execution
units become available while the store execution unit is still busy with the
sequence of dummy stores (line 2069-2098). And so the load may use the prior
stored scalar from r2 at address r10 -16 for speculation. The updated attack
may work less reliable on CPU microarchitectures where loads and stores share
execution resources.
This concludes that the sanitizing with zero stores from af86ca4e30 ("bpf:
Prevent memory disambiguation attack") is insufficient. Moreover, the detection
of stack reuse from af86ca4e30 where previously data (STACK_MISC) has been
written to a given stack slot where a pointer value is now to be stored does
not have sufficient coverage as precondition for the mitigation either; for
several reasons outlined as follows:
1) Stack content from prior program runs could still be preserved and is
therefore not "random", best example is to split a speculative store
bypass attack between tail calls, program A would prepare and store the
oob address at a given stack slot and then tail call into program B which
does the "slow" store of a pointer to the stack with subsequent "fast"
read. From program B PoV such stack slot type is STACK_INVALID, and
therefore also must be subject to mitigation.
2) The STACK_SPILL must not be coupled to register_is_const(&stack->spilled_ptr)
condition, for example, the previous content of that memory location could
also be a pointer to map or map value. Without the fix, a speculative
store bypass is not mitigated in such precondition and can then lead to
a type confusion in the speculative domain leaking kernel memory near
these pointer types.
While brainstorming on various alternative mitigation possibilities, we also
stumbled upon a retrospective from Chrome developers [0]:
[...] For variant 4, we implemented a mitigation to zero the unused memory
of the heap prior to allocation, which cost about 1% when done concurrently
and 4% for scavenging. Variant 4 defeats everything we could think of. We
explored more mitigations for variant 4 but the threat proved to be more
pervasive and dangerous than we anticipated. For example, stack slots used
by the register allocator in the optimizing compiler could be subject to
type confusion, leading to pointer crafting. Mitigating type confusion for
stack slots alone would have required a complete redesign of the backend of
the optimizing compiler, perhaps man years of work, without a guarantee of
completeness. [...]
From BPF side, the problem space is reduced, however, options are rather
limited. One idea that has been explored was to xor-obfuscate pointer spills
to the BPF stack:
[...]
// preoccupy the CPU store port by running sequence of dummy stores.
[...]
2106: (63) *(u32 *)(r7 +29796) = r0
2107: (63) *(u32 *)(r7 +29800) = r0
2108: (63) *(u32 *)(r7 +29804) = r0
2109: (63) *(u32 *)(r7 +29808) = r0
2110: (63) *(u32 *)(r7 +29812) = r0
// overwrite scalar with dummy pointer; xored with random 'secret' value
// of 943576462 before store ...
2111: (b4) w11 = 943576462
2112: (af) r11 ^= r7
2113: (7b) *(u64 *)(r10 -16) = r11
2114: (79) r11 = *(u64 *)(r10 -16)
2115: (b4) w2 = 943576462
2116: (af) r2 ^= r11
// ... and restored with the same 'secret' value with the help of AX reg.
2117: (71) r3 = *(u8 *)(r2 +0)
[...]
While the above would not prevent speculation, it would make data leakage
infeasible by directing it to random locations. In order to be effective
and prevent type confusion under speculation, such random secret would have
to be regenerated for each store. The additional complexity involved for a
tracking mechanism that prevents jumps such that restoring spilled pointers
would not get corrupted is not worth the gain for unprivileged. Hence, the
fix in here eventually opted for emitting a non-public BPF_ST | BPF_NOSPEC
instruction which the x86 JIT translates into a lfence opcode. Inserting the
latter in between the store and load instruction is one of the mitigations
options [1]. The x86 instruction manual notes:
[...] An LFENCE that follows an instruction that stores to memory might
complete before the data being stored have become globally visible. [...]
The latter meaning that the preceding store instruction finished execution
and the store is at minimum guaranteed to be in the CPU's store queue, but
it's not guaranteed to be in that CPU's L1 cache at that point (globally
visible). The latter would only be guaranteed via sfence. So the load which
is guaranteed to execute after the lfence for that local CPU would have to
rely on store-to-load forwarding. [2], in section 2.3 on store buffers says:
[...] For every store operation that is added to the ROB, an entry is
allocated in the store buffer. This entry requires both the virtual and
physical address of the target. Only if there is no free entry in the store
buffer, the frontend stalls until there is an empty slot available in the
store buffer again. Otherwise, the CPU can immediately continue adding
subsequent instructions to the ROB and execute them out of order. On Intel
CPUs, the store buffer has up to 56 entries. [...]
One small upside on the fix is that it lifts constraints from af86ca4e30
where the sanitize_stack_off relative to r10 must be the same when coming
from different paths. The BPF_ST | BPF_NOSPEC gets emitted after a BPF_STX
or BPF_ST instruction. This happens either when we store a pointer or data
value to the BPF stack for the first time, or upon later pointer spills.
The former needs to be enforced since otherwise stale stack data could be
leaked under speculation as outlined earlier. For non-x86 JITs the BPF_ST |
BPF_NOSPEC mapping is currently optimized away, but others could emit a
speculation barrier as well if necessary. For real-world unprivileged
programs e.g. generated by LLVM, pointer spill/fill is only generated upon
register pressure and LLVM only tries to do that for pointers which are not
used often. The program main impact will be the initial BPF_ST | BPF_NOSPEC
sanitation for the STACK_INVALID case when the first write to a stack slot
occurs e.g. upon map lookup. In future we might refine ways to mitigate
the latter cost.
[0] https://arxiv.org/pdf/1902.05178.pdf
[1] https://msrc-blog.microsoft.com/2018/05/21/analysis-and-mitigation-of-speculative-store-bypass-cve-2018-3639/
[2] https://arxiv.org/pdf/1905.05725.pdf
Fixes: af86ca4e30 ("bpf: Prevent memory disambiguation attack")
Fixes: f7cf25b202 ("bpf: track spill/fill of constants")
Co-developed-by: Piotr Krysiuk <piotras@gmail.com>
Co-developed-by: Benedict Schlueter <benedict.schlueter@rub.de>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Piotr Krysiuk <piotras@gmail.com>
Signed-off-by: Benedict Schlueter <benedict.schlueter@rub.de>
Acked-by: Alexei Starovoitov <ast@kernel.org>
In case of JITs, each of the JIT backends compiles the BPF nospec instruction
/either/ to a machine instruction which emits a speculation barrier /or/ to
/no/ machine instruction in case the underlying architecture is not affected
by Speculative Store Bypass or has different mitigations in place already.
This covers both x86 and (implicitly) arm64: In case of x86, we use 'lfence'
instruction for mitigation. In case of arm64, we rely on the firmware mitigation
as controlled via the ssbd kernel parameter. Whenever the mitigation is enabled,
it works for all of the kernel code with no need to provide any additional
instructions here (hence only comment in arm64 JIT). Other archs can follow
as needed. The BPF nospec instruction is specifically targeting Spectre v4
since i) we don't use a serialization barrier for the Spectre v1 case, and
ii) mitigation instructions for v1 and v4 might be different on some archs.
The BPF nospec is required for a future commit, where the BPF verifier does
annotate intermediate BPF programs with speculation barriers.
Co-developed-by: Piotr Krysiuk <piotras@gmail.com>
Co-developed-by: Benedict Schlueter <benedict.schlueter@rub.de>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Piotr Krysiuk <piotras@gmail.com>
Signed-off-by: Benedict Schlueter <benedict.schlueter@rub.de>
Acked-by: Alexei Starovoitov <ast@kernel.org>
