When TCP socket migration fails at inet_ehash_insert() in
reqsk_timer_handler(), we jump to the no_ownership: label
and free the new reqsk immediately with __reqsk_free().
Thus, we must stop the new reqsk's timer before jumping to the
label, but the timer might be missed since the cited commit,
resulting in UAF.
As we are in the original reqsk's timer context, we can safely
call timer_delete_sync() for the new reqsk.
Let's pass false to __inet_csk_reqsk_queue_drop() to stop
the new reqsk's timer.
PCI: Initialize temporary device in new_id_store()
When setting new_id of a PCI device driver using sysfs a lockdep splat
occurs. This is because new_id_store() builds a temporary pci_dev for
pci_match_device(), which calls device_match_driver_override(). That
depends on the driver_override.lock added by cb3d1049f4ea ("driver core:
generalize driver_override in struct device").
The new driver_override.lock was not initialized in the temporary pci_dev,
resulting in this lockdep splat.
Initialize the temporary pci_dev to fix this.
Repro:
Build with CONFIG_LOCKDEP=y, boot with QEMU, and add a new ID:
INFO: trying to register non-static key.
The code is fine but needs lockdep annotation, or maybe
you didn't initialize this object before use?
turning off the locking correctness validator.
CPU: 2 UID: 0 PID: 177 Comm: liveupdate-iomm Not tainted 7.0.0+ #9 PREEMPT(full)
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.16.3-0-ga6ed6b701f0a-prebuilt.qemu.org 04/01/2014
Call Trace:
<TASK>
dump_stack_lvl+0x5d/0x80
register_lock_class+0x77e/0x790
lock_acquire+0xbf/0x2e0
pci_match_device+0x24/0x180
new_id_store+0x189/0x1d0
kernfs_fop_write_iter+0x14f/0x210
vfs_write+0x263/0x5e0
ksys_write+0x79/0xf0
do_syscall_64+0x117/0xf80
Fixes: 10a4206a2401 ("PCI: use generic driver_override infrastructure") Fixes: 8895d3bcb8ba ("PCI: Fail new_id for vendor/device values already built into driver") Signed-off-by: Samiullah Khawaja <skhawaja@google.com>
[bhelgaas: add commit log details and repro, trim backtrace] Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Reviewed-by: Danilo Krummrich <dakr@kernel.org> Link: https://patch.msgid.link/20260505234327.716630-1-skhawaja@google.com
PCI: Update saved_config_space upon resource assignment
Bernd reports passthrough failure of a Digital Devices Cine S2 V6 DVB
adapter plugged into an ASRock X570S PG Riptide board with BIOS version
P5.41 (09/07/2023):
ddbridge 0000:05:00.0: detected Digital Devices Cine S2 V6 DVB adapter
ddbridge 0000:05:00.0: cannot read registers
ddbridge 0000:05:00.0: fail
BIOS assigns an incorrect BAR to the DVB adapter which doesn't fit into the
upstream bridge window. The kernel corrects the BAR assignment:
pci 0000:07:00.0: BAR 0 [mem 0xfffffffffc500000-0xfffffffffc50ffff 64bit]: can't claim; no compatible bridge window
pci 0000:07:00.0: BAR 0 [mem 0xfc500000-0xfc50ffff 64bit]: assigned
Correction of the BAR assignment happens in an x86-specific fs_initcall,
pcibios_assign_resources(), after device enumeration in a subsys_initcall.
This order was introduced at the behest of Linus in 2004:
No other architecture performs such a late BAR correction.
Bernd bisected the issue to commit a2f1e22390ac ("PCI/ERR: Ensure error
recoverability at all times"), but it only occurs in the absence of commit 4d4c10f763d7 ("PCI: Explicitly put devices into D0 when initializing").
This combination exists in stable kernel v6.12.70, but not in mainline,
hence Bernd cannot reproduce the issue with mainline.
Since a2f1e22390ac, config space is saved on enumeration, prior to BAR
correction. Upon passthrough, the corrected BAR is overwritten with the
incorrect saved value by:
But only if the device's current_state is PCI_UNKNOWN, as it was prior to
commit 4d4c10f763d7. Since the commit, it is PCI_D0, which changes the
behavior of vfio_pci_set_power_state() to no longer restore the state
without saving it first.
Alexandre is reporting the same issue as Bernd, but in his case, mainline
is affected as well. The difference is that on Alexandre's system, the
host kernel binds a driver to the device which is unbound prior to
passthrough, whereas on Bernd's system no driver gets bound by the host
kernel.
Unbinding sets current_state to PCI_UNKNOWN in pci_device_remove(), so when
vfio-pci is subsequently bound to the device, pci_restore_state() is once
again called without invoking pci_save_state() first.
To robustly fix the issue, always update saved_config_space upon resource
assignment.
Reported-by: Bernd Schumacher <bernd@bschu.de> Closes: https://lore.kernel.org/r/acfZrlP0Ua_5D3U4@eldamar.lan/ Reported-by: Alexandre N. <an.tech@mailo.com> Closes: https://lore.kernel.org/r/dd3c3358-de0f-4a56-9c81-04aceaab4058@mailo.com/ Fixes: a2f1e22390ac ("PCI/ERR: Ensure error recoverability at all times") Signed-off-by: Lukas Wunner <lukas@wunner.de> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Tested-by: Bernd Schumacher <bernd@bschu.de> Tested-by: Alexandre N. <an.tech@mailo.com> Cc: stable@vger.kernel.org # v6.12+ Link: https://patch.msgid.link/febc3f354e0c1f5a9f5b3ee9ffddaa44caccf651.1776268054.git.lukas@wunner.de
bpf: Free reuseport cBPF prog after RCU grace period.
Eulgyu Kim reported the splat below with a repro. [0]
The repro sets up a UDP reuseport group with a cBPF prog and
replaces it with a new one while another thread is sending
a UDP packet to the group.
The reuseport prog is freed by sk_reuseport_prog_free().
bpf_prog_put() is called for "e"BPF prog to destruct through
multiple stages while cBPF prog is freed immediately by
bpf_release_orig_filter() and bpf_prog_free().
If a reuseport prog is detached from the setsockopt() path
(reuseport_attach_prog() or reuseport_detach_prog()),
sk_reuseport_prog_free() is called without waiting for RCU
readers to complete, resulting in various bugs.
Let's defer freeing the reuseport cBPF prog after one RCU
grace period.
Note "e"BPF prog is safe as is unless the fast path starts
to touch fields destroyed in bpf_prog_put_deferred() and
__bpf_prog_put_noref().
Linus Torvalds [Fri, 8 May 2026 20:18:13 +0000 (13:18 -0700)]
Merge tag 'block-7.1-20260508' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux
Pull block fixes from Jens Axboe:
- Fix for ublk not doing an actual issue from the task_work fallback
path. Any request hitting that should be canceled automatically
- Fix for uring_cmd prep side handling, for the block side uring_cmd
discard handling
- Fix for missing validation of the io and physical block size shifts
- Fix for a use-after-free in ublk's cancel command handling
* tag 'block-7.1-20260508' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:
ublk: fix use-after-free in ublk_cancel_cmd()
ublk: validate physical_bs_shift, io_min_shift and io_opt_shift
block: only read from sqe on initial invocation of blkdev_uring_cmd()
ublk: don't issue uring_cmd from fallback task work
Linus Torvalds [Fri, 8 May 2026 20:12:48 +0000 (13:12 -0700)]
Merge tag 'io_uring-7.1-20260508' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux
Pull io_uring fixes from Jens Axboe:
- Ensure that the absolute timeouts for both the command side and the
waiting side honor the callers time namespace
- Ensure tracked NAPI entries are cleared at unregistration time, as
the NAPI polling loop checks the list state rather than the general
NAPI state. This can lead to NAPI polling even after unregistration
has been done. If unregistered, all NAPI polling should be disabled
- Fix for eventfd recursive invocation handling
* tag 'io_uring-7.1-20260508' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux:
io_uring/wait: honour caller's time namespace for IORING_ENTER_ABS_TIMER
io_uring/timeout: honour caller's time namespace for IORING_TIMEOUT_ABS
io_uring/eventfd: reset deferred signal state
io_uring/napi: clear tracked NAPI entries on unregister
Fixes: 21fb59ab4b976 ("ACPI: CPPC: Adjust debug messages in amd_set_max_freq_ratio() to warn") Suggested-by: Kim Phillips <kim.phillips@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Tested-by: Kim Phillips <kim.phillips@amd.com> Cc: All applicable <stable@vger.kernel.org> Link: https://patch.msgid.link/20260504230141.484743-2-mario.limonciello@amd.com Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
bpf: tcp: Fix type confusion in sol_tcp_sockopt().
sol_tcp_sockopt() only checks if sk->sk_protocol is IPPROTO_TCP,
but RAW socket can bypass it:
socket(AF_INET, SOCK_RAW, IPPROTO_TCP)
Let's use sk_is_tcp().
Note that initially sol_tcp_sockopt() checked sk->sk_prot->setsockopt.
Fixes: 2ab42c7b871f ("bpf: Check the protocol of a sock to agree the calls to bpf_setsockopt().") Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com> Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Link: https://patch.msgid.link/20260504210610.180150-7-kuniyu@google.com
mptcp: bpf: Fix type confusion in bpf_mptcp_sock_from_subflow()
bpf_mptcp_sock_from_subflow() only checks if sk->sk_protocol is
IPPROTO_TCP, but RAW socket can bypass it:
socket(AF_INET, SOCK_RAW, IPPROTO_TCP)
In this case, it would NOT be valid to call sk_is_mptcp() which will
assume sk is a pointer to a struct tcp_sock, and wrongly checks for:
tcp_sk(sk)->is_mptcp.
Breno Leitao [Fri, 8 May 2026 16:22:03 +0000 (09:22 -0700)]
workqueue: Fix wq->cpu_pwq leak in alloc_and_link_pwqs() WQ_UNBOUND path
For WQ_UNBOUND workqueues, alloc_and_link_pwqs() allocates wq->cpu_pwq
via alloc_percpu() and then calls apply_workqueue_attrs_locked(). On
failure it returns the error directly, bypassing the enomem: label
which holds the only free_percpu(wq->cpu_pwq) in this function.
The caller's error path kfree()s wq without touching wq->cpu_pwq,
leaking one percpu pointer table (nr_cpu_ids * sizeof(void *) bytes) per
failed call.
Breno Leitao [Thu, 7 May 2026 11:04:46 +0000 (04:04 -0700)]
workqueue: Release PENDING in __queue_work() drain/destroy reject path
The caller of __queue_work() owns WORK_STRUCT_PENDING, won via
test_and_set_bit() in queue_work_on()/__queue_delayed_work(). The
state machine documented above __queue_work() requires that owner
to either hand the token to a pwq (insert_work() -> set_work_pwq()),
hand it to a timer, or release it via set_work_pool_and_clear_pending().
try_to_grab_pending() relies on this: when it observes
"PENDING && off-queue" it busy-loops, trusting the current owner to
make progress.
The (__WQ_DESTROYING | __WQ_DRAINING) early-return path violates that
contract. It WARN_ONCE()s and bare-returns, leaving work->data with
PENDING set, WORK_STRUCT_PWQ clear, and work->entry empty.
The path is reachable without explicit API abuse: queue_delayed_work()
arms a timer with PENDING set; if drain_workqueue() runs while the
timer is still pending, delayed_work_timer_fn() -> __queue_work() in
softirq context hits the WARN, current is not a wq worker so
is_chained_work() is false, and the work is silently dropped with
PENDING leaked.
Mirror what clear_pending_if_disabled() already does on its analogous
reject path: unpack the off-queue data and call
set_work_pool_and_clear_pending() to release the token before
returning.
I was able to reproduce this by queueing several slow works on
a max_active=1 wq, arm a delayed_work whose timer fires while
drain_workqueue() is blocked, then call cancel_delayed_work_sync().
Without this patch the cancel livelocks at 100% CPU; with it the cancel
returns immediately.
Linus Torvalds [Fri, 8 May 2026 17:24:35 +0000 (10:24 -0700)]
Merge tag 'v7.1-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6
Pull smb client fixes from Steve French:
- Fix for two ACL issues (security fix to validate dacloffset better
and chmod fix)
- Fix out of bounds reads (in check_wsl_eas and smb2_check_msg for
symlinks)
- Two Kerberos fixes including an important one when AES-256 encryption
chosen
- Fix open_cached_dir problem when directory leases disabled
* tag 'v7.1-rc3-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
smb: client: validate dacloffset before building DACL pointers
smb/client: fix out-of-bounds read in smb2_compound_op()
smb/client: fix out-of-bounds read in symlink_data()
smb: client: Zero-pad short GSS session keys per MS-SMB2
smb: client: Use FullSessionKey for AES-256 encryption key derivation
smb: client: use kzalloc to zero-initialize security descriptor buffer
cifs: abort open_cached_dir if we don't request leases
Linus Torvalds [Fri, 8 May 2026 17:14:51 +0000 (10:14 -0700)]
Merge tag 'spi-fix-v7.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Pull spi fixes from Mark Brown:
"There's two main series here, fixing issues that came up in the
Microchip QSPI and Freescale i.MX drivers. Both of those could result
in some quite noticable issues if they were encountered in production.
We also have one minor documentation fix in the ch341 driver"
* tag 'spi-fix-v7.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
spi: ch341: correct company name in MODULE_DESCRIPTION
spi: microchip-core-qspi: remove some inline markings
spi: microchip-core-qspi: don't attempt to transmit during emulated read-only dual/quad operations
spi: microchip-core-qspi: control built-in cs manually
spi: imx: Propagate prepare_transfer() error from spi_imx_setupxfer()
spi: imx: Fix UAF on package-1 prepare failure in spi_imx_dma_data_prepare()
spi: imx: Fix precedence bug in spi_imx_dma_max_wml_find()
The buggy address belongs to the object at ffff88801083d280
which belongs to the cache RAW of size 1792
The buggy address is located 1248 bytes inside of
allocated 1792-byte region [ffff88801083d280, ffff88801083d980)
Fixes: 655a51e536c0 ("bpf: Add struct bpf_tcp_sock and BPF_FUNC_tcp_sock") Reported-by: Damiano Melotti <melotti@google.com> Signed-off-by: Kuniyuki Iwashima <kuniyu@google.com> Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Link: https://patch.msgid.link/20260504210610.180150-2-kuniyu@google.com
Linus Torvalds [Fri, 8 May 2026 15:23:06 +0000 (08:23 -0700)]
Merge tag 'drm-fixes-2026-05-08-1' of https://gitlab.freedesktop.org/drm/kernel
Pull drm fixes from Dave Airlie:
"Weekly fixes, lots of them but all pretty small, amdgpu and xe are the
usual but then a large amount of fixes all over.
xe:
- Add NULL check for media_gt in intel_hdcp_gsc_check_status
- Fix EAGAIN sign in pf_migration_consume
- Fix MMIO access using PF view instead of VF view during migration
- Exclude indirect ring state page from ADS engine state size
* tag 'drm-fixes-2026-05-08-1' of https://gitlab.freedesktop.org/drm/kernel: (37 commits)
drm: Set old handle to NULL before prime swap in change_handle
drm/bochs: Drop manual put on probe error path
drm/xe/guc: Exclude indirect ring state page from ADS engine state size
drm/xe/pf: Fix MMIO access using PF view instead of VF view during migration
drm/xe/pf: Fix EAGAIN sign in pf_migration_consume()
drm/xe/hdcp: Add NULL check for media_gt in intel_hdcp_gsc_check_status()
drm/exynos: remove bridge when component_add fails
drm/amdgpu: nuke amdgpu_userq_fence_slab v2
drm/amdgpu/userq: fix access to stale wptr mapping
drm/amdkfd: Check if there are kfd porcesses using adev by kfd_processes_count
drm/amdgpu: zero-initialize GART table on allocation
drm/amdgpu/sdma4: replace BUG_ON with WARN_ON in fence emission
drm/radeon: add missing revision check for CI
drm/amdgpu/pm: align Hawaii mclk workaround with radeon
drm/amdgpu/pm: add missing revision check for CI
drm/amdgpu/gfx9: drop unnecessary 64-bit fence flag check in KIQ
drm/amdkfd: Make all TLB-flushes heavy-weight
drm/panel: himax-hx83102: restore MODE_LPM after sending disable cmds
drm/panel: boe-tv101wum-nl6: restore MODE_LPM after sending disable cmds
drm/panel: feiyang-fy07024di26a30d: return display-on error
...
drm/ttm: Fix ttm_bo_swapout() infinite LRU walk on swapout failure
When ttm_tt_swapout() fails, the current code calls
ttm_resource_add_bulk_move() followed by ttm_resource_move_to_lru_tail()
to restore the resource's bulk_move membership.
However, ttm_resource_move_to_lru_tail() places the resource at the tail
of the LRU list which, relative to the walk cursor's hitch node (placed
immediately after the resource when it was yielded), puts the resource
*in front of the* the hitch. The next list_for_each_entry_continue() from
the hitch finds the same resource again, causing an infinite loop.
Fix by deferring del_bulk_move to the success path only.
On the success path, TTM_TT_FLAG_SWAPPED has just been set by
ttm_tt_swapout() but the resource is still tracked in the bulk_move range,
so ttm_resource_del_bulk_move()'s !ttm_resource_unevictable() guard would
incorrectly skip the removal. Introduce
ttm_resource_del_bulk_move_unevictable() which bypasses that guard.
Reported-by: Jatin Kataria <jkataria@netflix.com> Fixes: fc5d96670eb2 ("drm/ttm: Move swapped objects off the manager's LRU list") Cc: Christian König <christian.koenig@amd.com> Cc: Matthew Brost <matthew.brost@intel.com> Cc: <dri-devel@lists.freedesktop.org> Cc: <stable@vger.kernel.org> # v6.13+ Assisted-by: GitHub_Copilot:claude-sonnet-4.6 Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> Reviewed-by: Christian König <christian.koenig@amd.com> Tested-by: Boqun Feng <boqun@kernel.org> Link: https://patch.msgid.link/20260428094442.16985-1-thomas.hellstrom@linux.intel.com
Linus Torvalds [Fri, 8 May 2026 15:16:07 +0000 (08:16 -0700)]
Merge tag 'iommu-fixes-v7.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux
Pull iommu fixes from Joerg Roedel:
"Core:
- Cache-flushing fix for non-x86 platforms
AMD-Vi:
- Security fix when SEV-SNP is enabled
- Operator precedence fix in DTE setting"
* tag 'iommu-fixes-v7.1-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux:
iommu/amd: Fix precedence order in set_dte_passthrough()
iommu/pages: Fix iommu_pages_flush_incoherent() for non-x86
iommu/amd: Use maximum PPR log buffer size when SNP is enabled on Family 0x19
iommu/amd: Use maximum Event log buffer size when SNP is enabled on Family 0x19
Zqiang [Fri, 8 May 2026 11:50:45 +0000 (19:50 +0800)]
sched_ext: Use IRQ_WORK_INIT_HARD() to initialize sch->disable_irq_work
For built with PREEMPT_RT kernels, the scx_disable_irq_workfn() is
called from per-cpu irq_work kthreads context, this means that
when call the scx_dump_state() in the scx_disable_irq_workfn() to
output current->comm/pid, it always output current irq_work kthread's
comm/pid. this commit therefore use the IRQ_WORK_INIT_HARD() to
initialize sch->disable_irq_work to make scx_disable_irq_workfn() is
called from hardirq context.
Mark Rutland [Fri, 8 May 2026 14:20:23 +0000 (15:20 +0100)]
arm64/entry: Fix arm64-specific rseq brokenness
Mathias Stearn reports that since v6.19, there are two big issues
affecting rseq:
(1) On arm64 specifically, rseq critical sections aren't aborted when
they should be.
(2) The 'cpu_id_start' field is no longer written by the kernel in all
cases it used to be, including some cases where TCMalloc depends on
the kernel clobbering the field.
This patch fixes issue #1. This patch DOES NOT fix issue #2, which will
need to be addressed by other patches.
The arm64-specific brokenness is a result of commits:
2fc0e4b4126c ("rseq: Record interrupt from user space") 39a167560a61 ("rseq: Optimize event setting")
The first commit failed to add a call to rseq_note_user_irq_entry() on
arm64. Thus arm64 never sets rseq_event::user_irq to record that it may
be necessary to abort an active rseq critical section upon return to
userspace. On its own, this commit had no functional impact as the value
of rseq_event::user_irq was not consumed.
The second commit relied upon rseq_event::user_irq to determine whether
or not to bother to perform rseq work when returning to userspace. As
rseq_event::user_irq wasn't set on arm64, this work would be skipped,
and consequently an active rseq critical section would not be aborted.
Fix this by giving arm64 syscall-specific entry/exit paths, and
performing the relevant logic in syscall and non-syscall paths,
including calling rseq_note_user_irq_entry() for non-syscall entry.
Currently arm64 cannot use syscall_enter_from_user_mode(),
syscall_exit_to_user_mode(), and irqentry_exit_to_user_mode(), due to
ordering constraints with exception masking, and risk of ABI breakage
for syscall tracing/audit/etc. For the moment the entry/exit logic is
left as arm64-specific, directly using enter_from_user_mode() and
exit_to_user_mode(), but mirroring the generic code.
I intend to follow up with refactoring/cleanup, as we did for kernel
mode entry paths in commit:
041aa7a85390 ("entry: Split preemption from irqentry_exit_to_kernel_mode()")
... which will allow arm64 to use the GENERIC_IRQ_ENTRY functions directly.
David Woodhouse [Tue, 28 Apr 2026 20:59:52 +0000 (21:59 +0100)]
x86/kexec: Push kjump return address even for non-kjump kexec
The version of purgatory code shipped by kexec-tools attempts to look above
the top of its stack to find a return address for a kjump, even in a non-kjump
kexec.
After the commit in Fixes: the word above the stack might not be there,
leading to a fault (which is at least now caught by my exception-handling code
in kexec).
That commit fixed things for the actual kjump path, but no longer
"gratuitously" pushes the unused return address to the stack in the non-kjump
path. Put that *back* in the non-kjump path, to prevent purgatory from
crashing when trying to access it.
Fixes: 2cacf7f23a02 ("x86/kexec: Fix stack and handling of re-entry point for ::preserve_context") Reported-by: Rohan Kakulawaram <rohanka@google.com> Signed-off-by: David Woodhouse <dwmw@amazon.co.uk> Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de> Acked-by: Dave Hansen <dave.hansen@linux.intel.com> Tested-by: Rohan Kakulawaram <rohanka@google.com> Cc: <stable@kernel.org> Link: https://patch.msgid.link/32d627134143ffd957891cb697138e839c623211.camel@infradead.org
The buffer is then handed to ntfs_attr_add() and persisted as the
SECURITY_DESCRIPTOR attribute of the new MFT record. The descriptor
covers a relative security descriptor header, two SIDs (owner and
group), an ACL header, and a single ACE, but several fields inside
those structures are never written before the buffer is committed
to disk:
- struct security_descriptor_relative
@alignment (1 byte)
@sacl (4 bytes; SE_SACL_PRESENT is not set
but the offset still reaches disk)
- struct ntfs_sid (3 instances: owner, group, ACE.sid)
identifier_authority.value[0..4] (5 bytes per SID, 15 total
- only value[5] is set)
That is 23 bytes of uninitialised slab memory persisted to disk for
every new file or directory the legacy ntfs driver creates. The
"+ 4" trailing accounting in sd_len holds ace->sid.sub_authority[0],
which the existing code does explicitly write to zero, so it is
not part of the leak.
Anything later able to read the SECURITY_DESCRIPTOR attribute - the
same NTFS volume mounted on Windows or by another NTFS reader, an
offline forensics tool, an unprivileged user that ends up with read
access to the volume - can recover those bytes. The leak persists
for the lifetime of the file on disk, not just the lifetime of the
kernel that wrote it.
Switch the allocation to kzalloc() so every byte the on-disk
descriptor covers is zero before the explicit initialisations run.
While there, replace the bare "return -1" allocation-failure path
with a proper -ENOMEM so the error reaches userspace as a meaningful
errno instead of an unrelated -EPERM.
Found by inspection while auditing fs/ntfs new-inode paths.
Fixes: af0db57d4293 ("ntfs: update inode operations") Signed-off-by: DaeMyung Kang <charsyam@gmail.com> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
DaeMyung Kang [Thu, 7 May 2026 02:18:31 +0000 (11:18 +0900)]
ntfs: fix out-of-bounds write in ntfs_index_walk_down()
ntfs_index_walk_down() used to update the index traversal depth
directly before writing parent_pos[] and parent_vcn[]. A malformed
directory index with too many child-node levels can therefore advance
pindex past MAX_PARENT_VCN and write past the fixed arrays in struct
ntfs_index_context, corrupting context state used by later index
traversal.
Use ntfs_icx_parent_inc() for walk-down transitions so the existing
depth limit is enforced before the arrays are updated. Make the helper
check the limit before incrementing pindex so failed callers do not
leave the context at an out-of-range depth.
This is reachable by iterating a crafted NTFS directory after the volume
has been mounted, including read-only mounts. The reproducer uses
getdents64() on an index root that points to an excessively deep chain
of child index blocks.
A crafted directory index with a chain of child-node entries reproduced
UBSAN array-index-out-of-bounds reports in ntfs_index_walk_down() and
subsequent KASAN reports in ntfs_index_walk_up(). With this change, the
same image is rejected with "Index is over 32 level deep" and no KASAN
or UBSAN report is emitted.
DaeMyung Kang [Wed, 6 May 2026 09:24:48 +0000 (18:24 +0900)]
ntfs: fix out-of-bounds write in ntfs_rl_collapse_range() merge path
ntfs_rl_collapse_range() merges the run on the left of the collapsed
region with the run on its right when they are contiguous. The contiguous
check chooses a clamped index when @new_1st_cnt is 0:
i = new_1st_cnt == 0 ? 1 : new_1st_cnt;
if (ntfs_rle_lcn_contiguous(&new_rl[i - 1], &new_rl[i])) {
When @new_1st_cnt is 0 this computes &new_rl[-1] and writes 8 bytes
before the kvcalloc() runlist buffer. The path is reachable through
fallocate(FALLOC_FL_COLLAPSE_RANGE) starting at vcn 0 against an
attribute whose first run after the collapsed region and the following
run are holes. In that case ntfs_rle_lcn_contiguous() returns true
because both checked entries are LCN_HOLE, so the merge path is entered
with @new_1st_cnt still 0. Such consecutive holes do not occur on a
well-formed runlist (NTFS keeps runlists coalesced in memory), so this
OOB path is only reachable from a crafted volume.
A normal runlist has no element to the left of vcn 0, so the left/right
merge is not valid when @new_1st_cnt is 0. Require @new_1st_cnt to be
positive before checking or performing the merge. This skips the merge
entirely in that case instead of clamping the merge target.
The out-of-bounds write can corrupt an adjacent slab object. On a
non-KASAN kernel, it is reachable after a crafted NTFS volume has been
mounted read-write with the legacy fs/ntfs driver, by a local user that
has write access to the crafted file.
Fixes: 11ccc9107dc4 ("ntfs: update runlist handling and cluster allocator") Suggested-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: DaeMyung Kang <charsyam@gmail.com> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
DaeMyung Kang [Tue, 5 May 2026 13:07:52 +0000 (22:07 +0900)]
ntfs: fix default_upcase refcount underflow and UAF on fs_context teardown
ntfs_init_fs_context() allocates a fresh ntfs_volume with vol->upcase
left as NULL. ntfs_free_fs_context() unconditionally calls
ntfs_volume_free() during fs_context teardown, even when ntfs_fill_super()
never ran or already cleaned up. ntfs_volume_free() then executes:
When the global default_upcase is also NULL (very first mount attempt,
or all prior mounts have released the table), the comparison is
NULL == NULL, and ntfs_nr_upcase_users is decremented even though this
volume never claimed a reference. ntfs_nr_upcase_users is unsigned long,
so the decrement wraps to ULONG_MAX.
A subsequent successful mount can then free the shared table while the
mounted volume still points at it:
1. ntfs_fill_super() does the temporary ntfs_nr_upcase_users++ at the
"Generate the global default upcase table if necessary" block. With
the prior wraparound this brings the counter back to 0.
2. If the volume's $UpCase matches the default, the match path does
ntfs_nr_upcase_users++ and sets vol->upcase = default_upcase. The
counter is now 1.
3. On the success path, !--ntfs_nr_upcase_users evaluates true and
default_upcase is kvfree()'d while vol->upcase still points at it.
Subsequent upcase comparisons through that mount touch freed
memory.
This was reproduced with KASAN by closing a fresh fsopen("ntfs") context,
then mounting an NTFS image whose $UpCase table matches
generate_default_upcase(), and finally doing a case-insensitive lookup.
KASAN reports the dangling vol->upcase access:
BUG: KASAN: use-after-free in ntfs_collate_names+0x3b4/0x420
Read of size 2 at addr ffff888008d40048 by task init/1
ntfs_collate_names+0x3b4/0x420
ntfs_lookup_inode_by_name+0x1921/0x3130
ntfs_lookup+0x193/0xc40
vfs_statx+0xc7/0x190
vfs_fstatat+0x4b/0xa0
__do_sys_newfstatat+0x92/0xf0
The same QEMU reproducer was rerun after this change with KASAN
enabled. It reached "reproducer finished", and the log contained no
KASAN, use-after-free, Oops, or panic signatures.
Guard each comparison with an explicit vol->upcase non-NULL check so a
volume that never took a reference cannot decrement the global users
counter. Apply the same guard to the other default_upcase release sites
so all cleanup paths follow the same ownership rule: only volumes that
actually hold a default_upcase reference may drop one.
Fixes: 1e9ea7e04472 ("Revert "fs: Remove NTFS classic"") Signed-off-by: DaeMyung Kang <charsyam@gmail.com> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
DaeMyung Kang [Sat, 2 May 2026 00:49:16 +0000 (09:49 +0900)]
ntfs: avoid use-after-free of index inode in ntfs_inode_sync_filename()
ntfs_inode_sync_filename() walks every FILE_NAME attribute and, for
each one that points at a different parent, opens the parent index
inode with ntfs_iget() and locks index_ni->mrec_lock. All three error
branches (NInoBeingDeleted, ntfs_index_ctx_get failure, ntfs_index_lookup
failure) drop the parent reference before unlocking:
index_ni is NTFS_I(index_vi), so the ntfs_inode (and its mrec_lock) is
embedded in the inode allocation. If the parent directory is not held
outside the icache - no open dentry, recently evicted from dcache, no
other concurrent lookup - ntfs_iget() returns with i_count == 1 and
our iput() drops the last reference. evict_inode() then runs and
destroy_inode() schedules the slab object for RCU free, while
mutex_unlock() on the next line is still touching index_ni->mrec_lock.
Swap the order so the mutex is dropped while index_vi is still alive,
matching the success path at the bottom of the loop which already
unlocks before iput().
Reproduced under KASAN with a debug build that forces
ntfs_index_ctx_get() to fail when the parent index inode has been
opened with i_count == 1. KASAN reports a slab-use-after-free read
on the parent's mrec_lock from mutex_unlock() on the writeback worker:
BUG: KASAN: slab-use-after-free in __mutex_unlock_slowpath+0xb5/0x970
Read of size 8 at addr ffff8880014b7598 by task kworker/u8:0/12
Workqueue: writeback wb_workfn (flush-253:0)
Call Trace:
mutex_unlock
ntfs_inode_sync_filename
__ntfs_write_inode
ntfs_write_inode
__writeback_single_inode
Allocated by task 103:
ntfs_alloc_big_inode
ntfs_iget
ntfs_lookup
__x64_sys_mkdir
Freed by task 12:
ntfs_free_big_inode
i_callback
rcu_do_batch
Last potentially related work creation:
call_rcu
destroy_inode
evict
dispose_list
evict_inodes
ntfs_inode_sync_filename
__ntfs_write_inode
The buggy address belongs to the object at ffff8880014b7440
which belongs to the cache ntfs_big_inode_cache of size 1800
The freed object is the parent directory inode itself: allocated by
mkdir(2) via ntfs_iget(), then released through call_rcu(i_callback)
that destroy_inode() scheduled when evict_inodes() ran from inside
ntfs_inode_sync_filename(). Re-running the same workload with
mutex_unlock() moved before iput() runs cleanly under KASAN.
Fixes: af0db57d4293 ("ntfs: update inode operations") Signed-off-by: DaeMyung Kang <charsyam@gmail.com> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
DaeMyung Kang [Sat, 2 May 2026 00:48:52 +0000 (09:48 +0900)]
ntfs: fix copy length in ntfs_bdev_write() for non-page-aligned start
This is not a normal data I/O hot path. The single in-tree caller is
the $LogFile emptying path used during read-write mount/remount, and
the bug only becomes visible on NTFS volumes whose cluster_size is
strictly smaller than the kernel's PAGE_SIZE (typically 4 KiB on
x86_64). Per Microsoft's format command documentation, NTFS supports
allocation unit sizes starting at 512 bytes, so 512 B, 1 KiB and 2 KiB
clusters are uncommon but valid on-disk configurations. When
cluster_size >= PAGE_SIZE every "start" passed in is page-aligned and
the buggy "from != 0" path is never taken.
ntfs_bdev_write() splits the write across one or more block-device
folios. Inside the loop, "to" is computed as the *end byte offset*
within the current page (0..PAGE_SIZE), and "from" is the start byte
offset within the page (reset to 0 from the second iteration onward).
The copy length should therefore be "to - from", but the current code
uses "to" directly:
to = min_t(u32, end - offset, PAGE_SIZE);
memcpy_to_folio(folio, from, buf + buf_off, to);
buf_off += to;
When "from != 0" (i.e. "start" is not page-aligned) memcpy_to_folio()
copies "from" extra bytes:
- it reads "from" bytes past the source buffer into kernel heap;
- it writes "from" bytes past the requested range into the next part
of the block-device page (or, if "from + to > PAGE_SIZE", past the
folio boundary entirely, which trips the VM_BUG_ON inside
memcpy_to_folio() on CONFIG_DEBUG_VM=y kernels).
"buf_off" is then advanced by the wrong amount, so every subsequent
iteration also reads the source buffer at the wrong offset and writes
the wrong content to disk.
with empty_buf sized to vol->cluster_size. On a sub-PAGE_SIZE-cluster
volume, any $LogFile run whose LCN is not aligned to
PAGE_SIZE / cluster_size reaches the non-page-aligned path. The
over-copy can read beyond empty_buf and overwrite the sectors following
the requested cluster in the block-device page with unrelated kernel
heap contents while $LogFile is being emptied.
A userspace reducer of the same arithmetic and copy loop confirms the
bug under AddressSanitizer: ASan reports a heap-buffer-overflow read
past the source buffer for the buggy length, and the fixed version is
ASan-clean.
Compute the copy length as "to - from" and advance buf_off by the same
amount.
DaeMyung Kang [Thu, 30 Apr 2026 17:20:53 +0000 (02:20 +0900)]
ntfs: wait for sync mft writes to complete
ntfs_sync_mft_mirror() and write_mft_record_nolock() with @sync set
are both documented as synchronous, but neither actually waits for
the bio they submit nor inspects bi_status. write_inode() can
return success while dirty mft record bytes are still in flight, and
bio errors are silently dropped: the volume is not marked with
errors and the inode is not redirtied. This breaks fsync()/sync
metadata durability.
Switch ntfs_sync_mft_mirror() and the @sync path of
write_mft_record_nolock() to submit_bio_wait() and propagate the
returned error to the caller. Capture ntfs_sync_mft_mirror()'s
return value at its call sites in write_mft_record_nolock() so a
mirror write failure surfaces too.
The @sync parameter only controls the main MFT bio. The !@sync main
submission is therefore unchanged and still uses ntfs_bio_end_io() to
drop the folio reference taken before submission. The mirror call
has always been documented as performing synchronous I/O regardless
of @sync, so making it actually block restores the originally
intended contract for both @sync and !@sync callers.
Note this only fixes the synchronous mirror/main paths reachable
from write_mft_record_nolock(). The main MFT write submitted from
ntfs_write_mft_block() (the .writepages path) still does not wait
for completion or check bi_status; that requires a larger
restructuring and is left to a follow-up patch.
Fixes: 115380f9a2f9 ("ntfs: update mft operations") Signed-off-by: DaeMyung Kang <charsyam@gmail.com> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
DaeMyung Kang [Thu, 30 Apr 2026 17:20:55 +0000 (02:20 +0900)]
ntfs: capture mft mirror sync errors in ntfs_write_mft_block()
After ntfs_sync_mft_mirror() became able to return real I/O errors,
ntfs_write_mft_block() still discards its return value at the call
site inside the per-record loop. A failed $MFTMirr write therefore
leaves the volume looking clean from the writeback path even though
the on-disk mirror is now stale.
Capture the return value and feed it into the function's existing
@err variable using the same "first error wins" pattern already used
on other failure paths. The error is propagated to the caller and,
via the existing tail of the function, sets NVolErrors so umount and
chkdsk see the volume as inconsistent.
Fixes: 115380f9a2f9 ("ntfs: update mft operations") Signed-off-by: DaeMyung Kang <charsyam@gmail.com> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
DaeMyung Kang [Thu, 30 Apr 2026 17:20:54 +0000 (02:20 +0900)]
ntfs: redirty folio when ntfs_write_mft_block() runs out of memory
ntfs_write_mft_block() is called by writeback_iter() with the folio
locked. When the per-call allocations for @locked_nis or @ref_inos
fail, the function returns -ENOMEM directly without unlocking the
folio. Any later task that needs the folio's lock then stalls, and
the folio's dirty state is silently lost from the writeback
iterator's point of view.
Use folio_redirty_for_writepage() so the folio remains dirty for a
subsequent writeback pass, unlock it, and only then return -ENOMEM
so the caller can propagate the error to fsync()/sync_filesystem().
Fixes: f462fdf3d6a4 ("ntfs: reduce stack usage in ntfs_write_mft_block()") Signed-off-by: DaeMyung Kang <charsyam@gmail.com> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
DaeMyung Kang [Thu, 30 Apr 2026 11:54:47 +0000 (20:54 +0900)]
ntfs: use base mft_no when looking up base inode for extent record
When the mft record is an extent record, ntfs_may_write_mft_record()
looks up its base inode in the icache. The hash key passed to
find_inode_nowait() must be the base inode's mft number (na.mft_no,
set just above to MREF_LE(m->base_mft_record)), but the code passes
@mft_no, the extent record's own number.
find_inode_nowait() uses its second argument as the hashval, so the
lookup lands in the wrong bucket and almost always returns NULL.
ntfs_may_write_mft_record() then returns false and the writeback
path (ntfs_write_mft_block()) skips that extent record, leaving the
on-disk copy permanently out of sync with the in-memory one.
The original ilookup5_nowait() call this conversion replaced used
na.mft_no. Restore that.
Fixes: 115380f9a2f9 ("ntfs: update mft operations") Signed-off-by: DaeMyung Kang <charsyam@gmail.com> Reviewed-by: Hyunchul Lee <hyc.lee@gmail.com> Signed-off-by: Namjae Jeon <linkinjeon@kernel.org>
Arnd Bergmann [Fri, 8 May 2026 13:33:30 +0000 (15:33 +0200)]
Merge tag 'riscv-dt-fixes-for-v7.1-rc3' of https://git.kernel.org/pub/scm/linux/kernel/git/conor/linux into arm/fixes
RISC-V devicetrees fixes for v7.1-rc3
Microchip:
Fix a pinctrl misconfiguration caused by a erratum fixed between
engineering sample and production silicon, that causes settings for one
to not apply to the other.
Starfive:
Remove nodes relating to the "camss" video device that has been deleted
entirely from staging.
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
* tag 'riscv-dt-fixes-for-v7.1-rc3' of https://git.kernel.org/pub/scm/linux/kernel/git/conor/linux:
riscv: dts: microchip: fix icicle i2c pinctrl configuration
riscv: dts: starfive: jh7110: Drop CAMSS node
Ming Lei [Fri, 8 May 2026 12:37:46 +0000 (20:37 +0800)]
ublk: fix use-after-free in ublk_cancel_cmd()
When ublk_reset_ch_dev() clears io->cmd via ublk_queue_reinit()
concurrently with ublk_cancel_cmd(), ublk_cancel_cmd() can read a
stale pointer and pass it to io_uring_cmd_done(), causing a
use-after-free.
Fix by synchronizing the two paths with ubq->cancel_lock:
- ublk_cancel_cmd(): read and clear io->cmd under cancel_lock,
then call io_uring_cmd_done() on the saved local copy outside
the lock.
- ublk_reset_ch_dev(): hold cancel_lock across ublk_queue_reinit()
so that io->cmd and io->flags are cleared atomically with respect
to ublk_cancel_cmd().
Sven Eckelmann [Wed, 6 May 2026 20:20:52 +0000 (22:20 +0200)]
batman-adv: bla: put backbone reference on failed claim hash insert
When batadv_bla_add_claim() fails to insert a new claim into the hash, it
leaked a reference to the backbone_gw for which the claim was intended.
Call batadv_backbone_gw_put() on the error path to release the reference
and avoid leaking the backbone_gw object.
Sven Eckelmann [Wed, 6 May 2026 20:20:51 +0000 (22:20 +0200)]
batman-adv: bla: only purge non-released claims
When batadv_bla_purge_claims() goes through the list of claims, it is only
traversing the hash list with an rcu_read_lock(). Due to a potential
parallel batadv_claim_put(), it can happen that it encounters a claim which
was actually in the process of being released+freed by
batadv_claim_release(). In this case, backbone_gw is set to NULL before the
delayed RCU kfree is started. Calling batadv_bla_claim_get_backbone_gw() is
then no longer allowed because it would cause a NULL-ptr derefence.
To avoid this, only claims with a valid reference counter must be purged.
All others are already taken care of.
Sven Eckelmann [Wed, 6 May 2026 20:20:50 +0000 (22:20 +0200)]
batman-adv: bla: prevent use-after-free when deleting claims
When batadv_bla_del_backbone_claims() removes all claims for a backbone, it
does this by dropping the link entry in the hash list. This list entry
itself was one of the references which need to be dropped at the same time
via batadv_claim_put().
But the batadv_claim_put() must not be done before the last access to the
claim object in this function. Otherwise the claim might be freed already
by the batadv_claim_release() function before the list entry was dropped.
Sven Eckelmann [Wed, 6 May 2026 20:20:49 +0000 (22:20 +0200)]
batman-adv: tp_meter: fix tp_num leak on kmalloc failure
When batadv_tp_start() or batadv_tp_init_recv() fail to allocate a new
tp_vars object, the previously incremented bat_priv->tp_num counter is
never decremented. This causes tp_num to drift upward on each allocation
failure. Since only BATADV_TP_MAX_NUM sessions can be started and the count
is never reduced for these failed allocations, it causes to an exhaustion
of throughput meter sessions. In worst case, no new throughput meter
session can be started until the mesh interface is removed.
The error handling must decrement tp_num releasing the lock and aborting
the creation of an throughput meter session
Cc: stable@kernel.org Fixes: 33a3bb4a3345 ("batman-adv: throughput meter implementation") Signed-off-by: Sven Eckelmann <sven@narfation.org>
Jiexun Wang [Sun, 3 May 2026 04:28:58 +0000 (12:28 +0800)]
batman-adv: stop caching unowned originator pointers in BAT IV
BAT IV keeps the last-hop neighbor address in each neigh_node, but some
paths also cache an originator pointer derived from a temporary lookup.
That pointer is not owned by the neigh_node and may no longer refer to a
live originator entry after purge handling runs.
Stop storing the auxiliary originator pointer in the BAT IV neighbor
state. When BAT IV needs the neighbor originator data, resolve it from
the stored neighbor address and drop the reference again after use.
Fixes: c6c8fea29769 ("net: Add batman-adv meshing protocol") Cc: stable@kernel.org Reported-by: Yuan Tan <yuantan098@gmail.com> Reported-by: Yifan Wu <yifanwucs@gmail.com> Reported-by: Juefei Pu <tomapufckgml@gmail.com> Reported-by: Xin Liu <bird@lzu.edu.cn> Signed-off-by: Jiexun Wang <wangjiexun2025@gmail.com> Signed-off-by: Ren Wei <n05ec@lzu.edu.cn>
[sven: avoid bonding logic for outgoing OGM] Signed-off-by: Sven Eckelmann <sven@narfation.org>
Francis, David [Tue, 28 Apr 2026 19:25:50 +0000 (19:25 +0000)]
drm: Set old handle to NULL before prime swap in change_handle
There was a potential race condition in change_handle. The ioctl
briefly had a single object with two idr entries; a concurrent
gem_close could delete the object and remove one of the handles
while leaving the other one dangling, which could subsequently
be dereferenced for a use-after-free.
To fix this, do the same dance that gem_close itself does.
(f6cd7daecff5 drm: Release driver references to handle before making it available again)
First idr_replace the old handle to NULL. Later, if the prime
operations are successful, actually close it.
create_tail required a similar dance to avoid a similar problem.
(bd46cece51a3 drm/gem: Fix race in drm_gem_handle_create_tail())
It idr_allocs the new handle with NULL, then swaps in the correct
object later to avoid races. We don't need to do that here, since
the only operations that could race are drm_prime, and
change_handle holds the prime lock for the entire duration.
v2: cleanups of error paths
Signed-off-by: David Francis <David.Francis@amd.com> Co-authored-by: Dave Airlie <airlied@gmail.com> Reported-by: Puttimet Thammasaeng <pwn8official@gmail.com> Tested-by: Vitaly Prosyak <Vitaly.Prosyak@amd.com> Cc: Simona Vetter <simona@ffwll.ch> Cc: stable@vger.kernel.org Cc: Christian Koenig <Christian.Koenig@amd.com> Fixes: 53096728b8910 ("drm: Add DRM prime interface to reassign GEM handle") Signed-off-by: Dave Airlie <airlied@redhat.com>
John Walker [Thu, 7 May 2026 23:07:20 +0000 (17:07 -0600)]
wifi: cfg80211: advance loop vars in cfg80211_merge_profile()
cfg80211_merge_profile() reassembles a Multi-BSSID non-transmitted BSS
profile that has been split across multiple consecutive MBSSID elements.
Its while-loop calls
but never advances mbssid_elem or sub_elem inside the body. Each
iteration therefore searches for a continuation that follows the same
fixed pair; the helper returns the same next_mbssid; and the same
next_sub bytes are memcpy()'d into merged_ie at a growing offset until
the buffer fills.
Advance both mbssid_elem and sub_elem to the just-consumed continuation
so the next call to cfg80211_get_profile_continuation() searches for a
further continuation beyond it (or returns NULL when none exists).
A specially-crafted malicious beacon can take advantage of this bug
to cause the kernel to spend an excessive amount of time in
cfg80211_merge_profile (up to as much as 2ms per beacon received),
which could theoretically be abused in some way.
Cc: stable@vger.kernel.org Fixes: fe806e4992c9 ("cfg80211: support profile split between elements") Signed-off-by: John Walker <johnwalker0@gmail.com> Link: https://patch.msgid.link/20260507230720.64783-1-johnwalker0@gmail.com Signed-off-by: Johannes Berg <johannes.berg@intel.com>
K Prateek Nayak [Fri, 8 May 2026 05:17:48 +0000 (05:17 +0000)]
cpufreq/amd-pstate-ut: Drop policy reference before driver switch
Recent changes to the EPP unit test tries to perform a driver switch
with a cpufreq_policy reference held when the driver is loaded into
anything but the active mode which leads to a circular dependency and
the unit test hanging indefinitely.
Drop the reference before driver switch and grab it back once the driver
mode is stabilized for the test.
The EPP writes are only possible with CPUFREQ_POLICY_POWERSAVE policy.
Temporarily switch the cpudata->policy (while holding the write end of
the policy->rwsem) to CPUFREQ_POLICY_POWERSAVE and restore the original
policy once tests are done. To ensure the final EPP is correct in case
the driver started with CPUFREQ_POLICY_PERFORMANCE, EPP performance is
tested last.
The __free() based cleanup for cpufreq_policy is lost in the process.
Reported-by: Kalpana Shetty <kalpana.shetty@amd.com> Fixes: 7e173bc310d2b ("cpufreq/amd-pstate-ut: Add a unit test for raw EPP") Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com> Link: https://lore.kernel.org/r/20260508051748.10484-7-kprateek.nayak@amd.com Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
K Prateek Nayak [Fri, 8 May 2026 05:17:47 +0000 (05:17 +0000)]
cpufreq/amd-pstate: Use "epp_default_dc" as default when dynamic_epp is disabled
If "dynamic_epp" is disabled, the driver initialization and the default
EPP selection from sysfs currently sets the EPP based on the power
supply state of the system at that time but there is no power supply
callbacks registered to toggle it when the power supply state changes.
This can lead to faster battery drain on platforms that start off while
being plugged to the wall but later move to battery power since the EPP
stays at AMD_CPPC_EPP_PERFORMANCE.
Use "epp_default_dc" as the default EPP selection when dynamic_epp is
disabled, restoring older behavior. On servers, this defaults to
AMD_CPPC_EPP_PERFORMANCE and on other platforms, it defaults to
AMD_CPPC_EPP_BALANCE_PERFORMANCE.
Fixes: e30ca6dd5345 ("cpufreq/amd-pstate: Add dynamic energy performance preference") Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com> Link: https://lore.kernel.org/r/20260508051748.10484-6-kprateek.nayak@amd.com Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
K Prateek Nayak [Fri, 8 May 2026 05:17:46 +0000 (05:17 +0000)]
cpufreq/amd-pstate: Reorder notifier unregistration and floor perf reset
An active power supply notifier can race with amd_pstate_epp_cpu_exit()
trying to reset the floor perf and can overwrite the floor perf set in
MSR_AMD_CPPC_REQ.
Unregister the notifier before setting the floor perf to prevent the
rare race.
Fixes: e30ca6dd5345 ("cpufreq/amd-pstate: Add dynamic energy performance preference") Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com> Link: https://lore.kernel.org/r/20260508051748.10484-5-kprateek.nayak@amd.com Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
K Prateek Nayak [Fri, 8 May 2026 05:17:45 +0000 (05:17 +0000)]
cpufreq/amd-pstate: Allow writes to dynamic_epp when state isn't modified
Writing the current "dynamic_epp" state to sysfs fails with -EINVAL even
though the desired result was achieved. Allow writes to "dynamic_epp"
that does not modify the state.
Fixes: e30ca6dd5345 ("cpufreq/amd-pstate: Add dynamic energy performance preference") Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com> Link: https://lore.kernel.org/r/20260508051748.10484-4-kprateek.nayak@amd.com Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
K Prateek Nayak [Fri, 8 May 2026 05:17:44 +0000 (05:17 +0000)]
cpufreq/amd-pstate: Return -ENOMEM on failure to allocate profile_name
Failure to allocate profile name will return -EINVAL from
platform_profile_register() while in fact, it is a failure to allocate
memory for the profile_name string.
Return -ENOMEM when kasprintf() fails to allocate profile_name string.
Fixes: e30ca6dd5345 ("cpufreq/amd-pstate: Add dynamic energy performance preference") Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com> Link: https://lore.kernel.org/r/20260508051748.10484-3-kprateek.nayak@amd.com Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
hits the WARN_ON_ONCE() in static_key_disable_cpuslocked() and hangs the
system since both sysfs writes are trying to do
amd_pstate_change_driver_mode() without any synchronization.
Grab the "amd_pstate_driver_lock" mutex when modifying "dynamic_epp" to
prevent the two paths from racing with each other. Add a lockdep
assertion for "amd_pstate_driver_lock" in
amd_pstate_change_driver_mode() to formalize the dependency.
Since "cppc_mode" is stable under "amd_pstate_driver_lock", only reload
the driver when in "AMD_PSTATE_ACTIVE" mode and reject all writes when
in passive or guided mode, or if the driver is not loaded, since only
active mode operates on EPP.
Fixes: e30ca6dd5345 ("cpufreq/amd-pstate: Add dynamic energy performance preference") Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: K Prateek Nayak <kprateek.nayak@amd.com> Link: https://lore.kernel.org/r/20260508051748.10484-2-kprateek.nayak@amd.com Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org>
Hui Wang [Wed, 6 May 2026 13:21:52 +0000 (21:21 +0800)]
riscv: cpufeature: Use pre-defined ISA ext macros to index isa2hwcap
We have pre-defined ISA extension macros, here use those macros to
replace a magic number for isa2hwcap definition and some array
indexing for isa2hwcap access.
This doesn't change the original functionality, just improve the code
maintainability and readability.
The testsuite defines several kprobes and kretprobes as static variables
that are preserved across test runs.
After register_kprobe and unregister_kprobe, a kprobe contains some
leftover data that must be cleared before the kprobe can be registered
again. The tests are setting symbol_name to define the probe location.
Address and flags must be cleared.
The existing code clears some of the probes between subsequent tests, but
not between two test runs. The leftover data from a previous test run
makes the registrations fail in the next run.
Move the cleanups for all kprobes into kprobes_test_init, this function
is called before each single test (including the first test of a test
run).
Jianpeng Chang [Fri, 8 May 2026 00:56:36 +0000 (09:56 +0900)]
kprobes: skip non-symbol addresses in kprobe_add_ksym_blacklist()
When kprobe_add_area_blacklist() iterates through a section like
.kprobes.text, the start address may not correspond to a named symbol.
On ARM64 with CONFIG_DYNAMIC_FTRACE_WITH_CALL_OPS=y (introduced by
commit baaf553d3bc3 ("arm64: Implement
HAVE_DYNAMIC_FTRACE_WITH_CALL_OPS")), the compiler flag
-fpatchable-function-entry=4,2 inserts 2 NOPs before each function entry
point for ftrace call_ops. These pre-function NOPs sit at the section base
address, before the first named function symbol. The compiler emits a $x
mapping symbol at offset 0x00 to mark the start of code, but
find_kallsyms_symbol() ignores mapping symbols.
Without CONFIG_DYNAMIC_FTRACE_WITH_CALL_OPS (e.g. defconfig), no
pre-function NOPs are inserted, the first function starts at offset
0x00, and the bug does not trigger.
This only affects modules that have a .kprobes.text section (i.e. those
using the __kprobes annotation). Modules using NOKPROBE_SYMBOL() instead
(like kretprobe_example.ko) blacklist exact function addresses via the
_kprobe_blacklist section and are not affected.
For kprobe_example.ko on ARM64 with -fpatchable-function-entry=4,2,
the .kprobes.text section layout is:
kprobe_add_area_blacklist() starts iterating from the section base
address (offset 0x00), which only has the $x mapping symbol.
kprobe_add_ksym_blacklist() then calls kallsyms_lookup_size_offset()
for this address, which goes through:
find_kallsyms_symbol() scans all module symbols to find the closest
preceding symbol.
Since no named text symbol exists at offset 0x00,
find_kallsyms_symbol() picks __UNIQUE_ID_vermagic (a .modinfo symbol
whose address is in the temporary image) as the "best" match. The
computed "size" = next_text_symbol - modinfo_symbol spans across
these two unrelated memory regions, creating a blacklist entry with
a bogus range of tens of terabytes.
Whether this causes a visible failure depends on address randomization,
here is what happens on Raspberry Pi 4/5:
- On RPi5, the bogus size was ~35 TB. start + size stayed within
64-bit range, so the blacklist entry covered the entire kernel
text. register_kprobe() in the module's own init function failed
with -EINVAL.
- On RPi4, the bogus size was ~75 TB. start + size overflowed
64 bits and wrapped to a small address near zero. The range
check (addr >= start && addr < end) then failed because end
wrapped around, so the bogus entry was accidentally harmless
and kprobes worked by luck.
The same bug exists on both machines, but randomization determines whether
the integer overflow masks it or not.
Fix this by adding notrace to the __kprobes macro. Functions in
.kprobes.text are kprobe infrastructure handlers that should never be
traced by ftrace. With notrace, the compiler stops inserting them and the
non-symbol gap at the section start disappears entirely.
Linus Torvalds [Fri, 8 May 2026 00:26:43 +0000 (17:26 -0700)]
Merge tag 'selinux-pr-20260507' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux
Pull selinux fixes from Paul Moore:
- Allow for multiple opens of /sys/fs/selinux/policy
Prevent a single process from blocking others from reading the
SELinux policy loaded in the kernel. This does have the side effect
of potentially allowing userspace to trigger additional kernel memory
allocations as part of the open/read operation, but this is mitigated
by requiring the SELinux security/read_policy permission.
- Reduce the critical sections where the SELinux policy mutex is held
This includes the patch to the policy loader code where we move the
permission checks and an allocation outside the mutex as well as the
the patch to checkreqprot which drops the code/lock entirely.
While the checkreqprot code had effectively been dropped in an
earlier release, portions of the code still remained that would have
triggered the mutex to perform an IMA measurement. This finally drops
all of that while preserving the user visible behavior.
- Eliminate potential sources of log spamming
There were a few areas where processes could flood the system logs
and hide other, more critical events. The previously disabled
checkreqprot and runtime disable knobs in selinuxfs were two such
areas that have now been greatly simplified and a pr_err() replaced
with a pr_err_once().
The third such place is the /sys/fs/selinux/user file, which hasn't
been used by a userspace release since 2020 and was scheduled for
removal after 2025; this effectively disables this functionality, but
similar to checkreqprot, it is done in a way that should not break
old userspace.
* tag 'selinux-pr-20260507' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux:
selinux: shrink critical section in sel_write_load()
selinux: allow multiple opens of /sys/fs/selinux/policy
selinux: prune /sys/fs/selinux/user
selinux: prune /sys/fs/selinux/disable
selinux: prune /sys/fs/selinux/checkreqprot
Li Xiasong [Thu, 7 May 2026 14:04:22 +0000 (22:04 +0800)]
netfilter: nf_conntrack_sip: get helper before allocating expectation
process_register_request() allocates an expectation and then checks
whether a conntrack helper is available. If helper lookup fails, the
function returns early and the allocated expectation is left behind.
Reorder the code to fetch and validate helper before calling
nf_ct_expect_alloc(). This keeps the logic simpler and removes the leak
path while preserving existing behavior.
Fixes: e14575fa7529 ("netfilter: nf_conntrack: use rcu accessors where needed") Cc: stable@vger.kernel.org Signed-off-by: Li Xiasong <lixiasong1@huawei.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: nf_conntrack_expect: restore helper propagation via expectation
A recent series to fix expectations broke helper propagation via
expectation, this mechanism is used by the sip and h323 helper. This
also propagates the conntrack helper to expected connections. I changed
semantics of exp->helper which now tells us the actual helper that
created the expectation.
Add an explicit assign_helper field to expectations for this purpose
and update helpers to use it.
Restore this feature for userspace conntrack helper via ctnetlink
nfqueue integration so it is again possible to attach a helper to an
expectation, where it makes sense. This is not restored via ctnetlink
expectation creation as there is no client for such feature. Use the
expectation layer 4 protocol number for the helper lookup for
consistency.
Make sure the expectation using this helper propagation mechanism also
go away when the helper is unregistered.
netfilter: bridge: eb_tables: close module init race
sashiko reports for unrelated patch:
Does the core ebtables initialization in ebtables.c suffer from a similar race?
Once nf_register_sockopt() completes, the sockopts are exposed globally.
sockopt has to be registered last, just like in ip/ip6/arptables.
Fixes: 5b53951cfc85 ("netfilter: ebtables: use net_generic infra") Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: x_tables: close dangling table module init race
Similar to the previous ebtables patch:
template add exposes the table to userspace, we must do this last to
rnsure the pernet ops are set up (contain the destructors).
Fixes: fdacd57c79b7 ("netfilter: x_tables: never register tables by default") Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: ebtables: close dangling table module init race
sashiko reported for a related patch:
In modules like iptable_raw.c, [..], if register_pernet_subsys() fails,
the rollback might call kfree(rawtable_ops) before [..]
During this window, could a concurrent userspace process find the globally
visible template, trigger table_init(), [..]
The table init functions must always register the template last.
Otherwise, set/getsockopt can instantiate a table in a namespace
while the required pernet ops (contain the destructor) isn't available.
This change is also required in x_tables, handled in followup change.
Fixes: 87663c39f898 ("netfilter: ebtables: do not hook tables by default") Reviewed-by: Tristan Madani <tristan@talencesecurity.com> Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: x_tables: add and use xtables_unregister_table_exit
Previous change added xtables_unregister_table_pre_exit to detach the
table from the packetpath and to unlink it from the active table list.
In case of rmmod, userspace that is doing set/getsockopt for this table
will not be able to re-instantiate the table:
1. The larval table has been removed already
2. existing instantiated table is no longer on the xt pernet table list.
This adds the second stage helper:
unlink the table from the dying list, free the hook ops (if any) and do
the audit notification. It replaces xt_unregister_table().
Fixes: fdacd57c79b7 ("netfilter: x_tables: never register tables by default") Reported-by: Tristan Madani <tristan@talencesecurity.com> Reviewed-by: Tristan Madani <tristan@talencesecurity.com> Closes: https://lore.kernel.org/netfilter-devel/20260429175613.1459342-1-tristmd@gmail.com/ Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: x_tables: unregister the templates first
When the module is going away we need to zap the template
first. Else there is a small race window where userspace
could instantiate a new table after the pernet exit function
has removed the current table.
Fixes: fdacd57c79b7 ("netfilter: x_tables: never register tables by default") Reported-by: Tristan Madani <tristan@talencesecurity.com> Reviewed-by: Tristan Madani <tristan@talencesecurity.com> Closes: https://lore.kernel.org/netfilter-devel/20260429175613.1459342-1-tristmd@gmail.com/ Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
netfilter: x_tables: add and use xt_unregister_table_pre_exit
Remove the copypasted variants of _pre_exit and add one single
function in the xtables core. ebtables is not compatible with
x_tables and therefore unchanged.
This is a preparation patch to reduce noise in the followup
bug fixes.
netfilter: x_tables: allocate hook ops while under mutex
arp/ip(6)t_register_table() add the table to the per-netns list via
xt_register_table() before allocating the per-netns hook ops copy
via kmemdup_array(). This leaves a window where the table is
visible in the list with ops=NULL.
If the pernet exit happens runs concurrently the pre_exit callback finds
the table via xt_find_table() and passes the NULL ops pointer to
nf_unregister_net_hooks(), causing a NULL dereference:
general protection fault in nf_unregister_net_hooks+0xbc/0x150
RIP: nf_unregister_net_hooks (net/netfilter/core.c:613)
Call Trace:
ipt_unregister_table_pre_exit
iptable_mangle_net_pre_exit
ops_pre_exit_list
cleanup_net
Fix by moving the ops allocation into the xtables core so the table is
never in the list without valid ops. Also ensure the table is no longer
processing packets before its torn down on error unwind.
nf_register_net_hooks might have published at least one hook; call
synchronize_rcu() if there was an error.
audit log register message gets deferred until all operations have
passed, this avoids need to emit another ureg message in case of
error unwinding.
At the moment we emit the audit log a bit too early, which makes it
necessary to also emit an unregister log in case we have to unwind
errors after possible hook register failure.
Followup patch will be slightly simpler if we can delay the
register message until after the hooks have been wired up.
Signed-off-by: Florian Westphal <fw@strlen.de> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Tabrez Ahmed [Sat, 2 May 2026 02:08:42 +0000 (07:38 +0530)]
hwmon: (ads7871) Fix endianness bug in 16-bit register reads
The ads7871_read_reg16() function relies on spi_w8r16() to read the
16-bit sensor output. The ADS7871 device transmits the Least Significant
Byte (LSB) first.
On Little-Endian architectures, spi_w8r16() correctly reconstructs the
16-bit value. However, on Big-Endian architectures, the byte swapping
causes the first received byte (LSB) to be placed in the most significant
byte of the u16, resulting in corrupted voltage readings.
To fix this, cast the integer result of spi_w8r16() to a restricted
__le16 type and convert it to the host CPU's native byte order using
le16_to_cpu(). Negative error codes returned by the SPI core are caught
and returned prior to the conversion to avoid mangling the error status.
Reported-by: Sashiko <sashiko-bot@kernel.org> Closes: https://sashiko.dev/#/patchset/20260418034601.90226-1-tabreztalks@gmail.com Fixes: e0c70b8078629 ("hwmon: add TI ads7871 a/d converter driver") Suggested-by: David Laight <david.laight.linux@gmail.com> Signed-off-by: Tabrez Ahmed <tabreztalks@gmail.com> Link: https://lore.kernel.org/r/20260502020844.110038-2-tabreztalks@gmail.com Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Sensors configurations are defined by set and clear masks. These
do not follow a consistent "clear mask is a superset of set mask"
rule. This relaxed definition breaks lm75_write_config()
Basically all bits from set_mask that are not defined in clr_mask are
dropped. Fix that by enhancing the helper to always combine clr_mask
and set_mask into the mask bits of regmap_update_bits().
hwmon: (lm75) Fix AS6200 and TMP112 setup and alarm handling
The initialization of the AS6200 has two shortcomings
- The device-add-commit states "Conversion mode: continuous" but the
the lm75_params structure uses set_mask = 0x94c0. This activates
single shot mode (bit 15). According to the datasheet "The device
features a single shot measurement mode if the device is in sleep
mode (SM=1)". This is quite contradictionary.
- It is the only device that activates polarity active-high (bit 10)
All this is paired with a undefined clear mask bug in function
lm75_write_config() that was introduced with a later refactoring
commit.
[as6200] = {
.config_reg_16bits = true,
.set_mask = 0x94C0,
-> .clr_mask not defined here
.default_resolution = 12,
...
static inline int lm75_write_config(struct lm75_data *data, u16 set_mask,
u16 clr_mask)
{
return regmap_update_bits(data->regmap, LM75_REG_CONF,
clr_mask | LM75_SHUTDOWN, set_mask);
}
regmap_update_bits() requires clr_mask to be a superset of set_mask.
So basically all sensors with "wrong" masks like the AS6200 are not
initialized as intended.
Fix that by
- Change the set_mask to 0xc010 to reflect the current active-low
setup properly and to drive the sensor in continous mode. This
takes into account that the config register is little endian and
the first byte sent to the chip is the LSB.
- Adapt the alarm handling so it can report the alarm correctly
even if it is high active. This is done by comparing config register
bit 5 and 10 (translated to 2 and 13).
This commit does not introduce any ABI breakage as the mutliple bugs
effectly drive the AS6200 in standard active-low mode.
Fixes: 4b6358e1fe46 ("hwmon: (lm75) Add AMS AS6200 temperature sensor") Suggested-by: Guenter Roeck <linux@roeck-us.net> Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de> Link: https://lore.kernel.org/r/20260502173207.3567876-2-markus.stockhausen@gmx.de
[groeck: Update set_mask for as6200 further: As modeled, the upper bits
contain the conversion rate, so the config register needs to be set to
0xc010 instead of 0x10c0 to reflect 8 samples/s and 4 consecutive faults.
Fix the same problem for TMP112.] Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Dave Airlie [Thu, 7 May 2026 22:51:01 +0000 (08:51 +1000)]
Merge tag 'drm-xe-fixes-2026-05-07' of https://gitlab.freedesktop.org/drm/xe/kernel into drm-fixes
UAPI Changes:
Cross-subsystem Changes:
Core Changes:
Driver Changes:
- Add NULL check for media_gt in intel_hdcp_gsc_check_status (Gustavo)
- Fix EAGAIN sign in pf_migration_consume (Shuicheng)
- Fix MMIO access using PF view instead of VF view during migration (Shuicheng)
- Exclude indirect ring state page from ADS engine state size (Satya)
Dave Airlie [Thu, 7 May 2026 22:34:34 +0000 (08:34 +1000)]
Merge tag 'drm-rust-fixes-2026-05-07' of https://gitlab.freedesktop.org/drm/rust/kernel into drm-fixes
DRM Rust fixes for v7.1-rc3
- Fix unsound initialization in drm::Device::new(); if pinned
initialization of drm::Device::Data fails, make sure
drm::Device::release() isn't called, so we don't run the data's
destructor
- Fix missing GEM state cleanup in the init failure case; call
drm_gem_private_object_fini() if drm_gem_object_init() fails
- Fix wrong ARef import in the DRM shmem GEM helper abstraction
- Replace the nouveau mailing list with the new nova-gpu mailing list
for both nova-core and nova-drm, and remove unused patchwork entries
Robbie Ko [Fri, 1 May 2026 02:41:56 +0000 (10:41 +0800)]
btrfs: fix incorrect i_size after remount caused by KEEP_SIZE prealloc gap
When fallocate() with FALLOC_FL_KEEP_SIZE preallocates an extent past the
current i_size, the file_extent_tree of the inode is updated to cover
that range. However, on the next mount, btrfs_read_locked_inode() only
re-populates file_extent_tree with [0, round_up(i_size, sectorsize)),
losing the marks that belonged to the KEEP_SIZE prealloc extent beyond
i_size.
Later, when a non-KEEP_SIZE fallocate() extends i_size into / past that
old prealloc extent, the reservation loop in btrfs_fallocate() skips
already-prealloc segments and does not call into the path that marks the
file_extent_tree, so a gap remains inside the file_extent_tree across
[old_aligned_i_size, start_of_new_alloc). Then __btrfs_prealloc_file_range()
calls btrfs_inode_safe_disk_i_size_write(), which uses
find_contiguous_extent_bit() starting at offset 0 to derive disk_i_size.
The walk stops at the gap, so disk_i_size ends up smaller than i_size and
gets persisted. After the next mount, the file shows the wrong (smaller)
size.
# non-KEEP_SIZE fallocate that overlaps the previous prealloc tail
# and extends past it
fallocate -o 7M -l 2M $MNT/file1
ls -lh $MNT/file1
umount $MNT
mount $DEV $MNT
ls -lh $MNT/file1
umount $MNT
Running the reproducer gives the following result:
The size before the second mount is correct (9M), but after the
remount it drops to 7M, i.e. the start of the gap inside file_extent_tree.
Fix this in __btrfs_prealloc_file_range() by marking the entire range
[round_down(old_i_size, sectorsize), round_up(new_i_size, sectorsize))
in file_extent_tree before updating i_size and calling
btrfs_inode_safe_disk_i_size_write(). This ensures the contiguous bit
search starting from 0 is not truncated by a stale gap left behind by a
previous KEEP_SIZE prealloc that was not restored on inode load.
The fix has no effect when the NO_HOLES feature is enabled because
btrfs_inode_safe_disk_i_size_write() and
btrfs_inode_set_file_extent_range()
both take the fast path that directly tracks disk_i_size without
consulting file_extent_tree.
Fixes: 9ddc959e802b ("btrfs: use the file extent tree infrastructure") Reviewed-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Robbie Ko <robbieko@synology.com>
[ Minor updates to the change log ] Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
btrfs: only release the dirty pages io tree after successful writes
[WARNING]
With extra warning on dirty extent buffers at umount (aka, the next
patch in the series), test case generic/388 can trigger the following
warning about dirty extent buffers at unmount time:
BTRFS critical (device dm-2 state E): emergency shutdown
BTRFS error (device dm-2 state E): error while writing out transaction: -30
BTRFS warning (device dm-2 state E): Skipping commit of aborted transaction.
BTRFS error (device dm-2 state EA): Transaction 9 aborted (error -30)
BTRFS: error (device dm-2 state EA) in cleanup_transaction:2068: errno=-30 Readonly filesystem
BTRFS info (device dm-2 state EA): forced readonly
BTRFS info (device dm-2 state EA): last unmount of filesystem 4fbf2e15-f941-49a0-bc7c-716315d2777c
------------[ cut here ]------------
WARNING: disk-io.c:3311 at invalidate_and_check_btree_folios+0xfd/0x1ca [btrfs], CPU#8: umount/914368
CPU: 8 UID: 0 PID: 914368 Comm: umount Tainted: G OE 7.1.0-rc1-custom+ #372 PREEMPT(full) 2de38db8d1deae71fde295430a0ff3ab98ccf596
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS unknown 02/02/2022
RIP: 0010:invalidate_and_check_btree_folios+0xfd/0x1ca [btrfs]
Call Trace:
<TASK>
close_ctree+0x52e/0x574 [btrfs d2f0b1cd330d1287e7a9919d112eadfc0e914efd]
generic_shutdown_super+0x89/0x1a0
kill_anon_super+0x16/0x40
btrfs_kill_super+0x16/0x20 [btrfs d2f0b1cd330d1287e7a9919d112eadfc0e914efd]
deactivate_locked_super+0x2d/0xb0
cleanup_mnt+0xdc/0x140
task_work_run+0x5a/0xa0
exit_to_user_mode_loop+0x123/0x4b0
do_syscall_64+0x243/0x7c0
entry_SYSCALL_64_after_hwframe+0x4b/0x53
</TASK>
---[ end trace 0000000000000000 ]---
BTRFS warning (device dm-2 state EA): unable to release extent buffer 30539776 owner 9 gen 9 refs 2 flags 0x7
BTRFS warning (device dm-2 state EA): unable to release extent buffer 30621696 owner 257 gen 9 refs 2 flags 0x7
BTRFS warning (device dm-2 state EA): unable to release extent buffer 30638080 owner 258 gen 9 refs 2 flags 0x7
BTRFS warning (device dm-2 state EA): unable to release extent buffer 30654464 owner 7 gen 9 refs 2 flags 0x7
BTRFS warning (device dm-2 state EA): unable to release extent buffer 30703616 owner 2 gen 9 refs 2 flags 0x7
BTRFS warning (device dm-2 state EA): unable to release extent buffer 30720000 owner 10 gen 9 refs 2 flags 0x7
BTRFS warning (device dm-2 state EA): unable to release extent buffer 30736384 owner 4 gen 9 refs 2 flags 0x7
BTRFS warning (device dm-2 state EA): unable to release extent buffer 30752768 owner 11 gen 9 refs 2 flags 0x7
I'm using a stripped down version, which seems to trigger the warning
more reliably:
for (( i = 0; i < $runtime; i++ )); do
echo "=== $i/$runtime ==="
workload
done
[CAUSE]
Inside btrfs_write_and_wait_transaction(), we first try to write all
dirty ebs, then wait for them to finish.
After that we call btrfs_extent_io_tree_release() to free all
extent states from dirty_pages io tree.
However if we hit an error from btrfs_write_marked_extent(), then we
still call btrfs_extent_io_tree_release() to clear that dirty_pages io
tree, which may contain dirty records that we haven't yet submitted.
Furthermore, the later transaction cleanup path will utilize that
dirty_pages io tree to properly cleanup those dirty ebs, but since it's
already empty, no dirty ebs are properly cleaned up, thus will later
trigger the warnings inside invalidate_btree_folios().
[FIX]
Normally such dirty ebs won't cause problems, as when the iput() is
called on the btree inode, the dirty ebs will be forcibly written back,
and since the fs is already in an error status, such writeback will not
reach disk and finish immediately.
But it's still better to get rid of such dirty ebs, if we ended up with
dirty ebs but the fs is not in an error status, then such writeback at
iput() time will be too late, as all workers are already stopped but
writeback will utilize workers, which will lead to NULL pointer
dereferences.
Instead of unconditionally calling btrfs_extent_io_tree_release(), only
call it if btrfs_write_and_wait_transaction() finished successfully, so
that @dirty_pages extent io tree is kept untouched for transaction
cleanup.
CC: stable@vger.kernel.org # 6.1+ Reviewed-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: Qu Wenruo <wqu@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
Filipe Manana [Tue, 28 Apr 2026 15:58:56 +0000 (16:58 +0100)]
btrfs: tracepoints: fix sleep while in atomic context in btrfs_sync_file()
The trace event btrfs_sync_file() is called in an atomic context (all trace
events are) and its call to dput(), which is needed due to the call to
dget_parent(), can sleep, triggering a kernel splat.
This can be reproduced by enabling the trace event and running btrfs/056
from fstests for example. The splat shown in dmesg is the following:
So stop using dget_parent() and dput() and access the parent dentry
directly as dentry->d_parent. This is also what ext4 is doing in
its equivalent trace event ext4_sync_file_enter().
Fixes: a85b46db143f ("btrfs: tracepoints: get correct superblock from dentry in event btrfs_sync_file()") Reviewed-by: Boris Burkov <boris@bur.io> Signed-off-by: Filipe Manana <fdmanana@suse.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
(lldb) source list -a add_ra_bio_pages+0x13c
.../vmlinux.unstripped add_ra_bio_pages + 316 at .../fs/btrfs/compression.c:454:8
451
452 folio = filemap_alloc_folio(mapping_gfp_constraint(mapping, constraint_gfp),
453 0, NULL);
-> 454 if (!folio)
455 break;
I can reproduce this consistently by running a memory hog concurrently
with a buffered writer on a machine with a very large amount of swap.
Commit 7ae37b2c94ed ("btrfs: prevent direct reclaim during compressed
readahead") clearly intended to suppress these warnings. But because the
mask set in the address_space with mapping_set_gfp_mask() doesn't include
__GFP_NOWARN, mapping_gfp_constraint() removes it from constraint_gfp
before it is passed to filemap_alloc_folio().
Fix by refactoring the code to add __GFP_NOWARN after the call to
mapping_gfp_constraint().
Fixes: 7ae37b2c94ed ("btrfs: prevent direct reclaim during compressed readahead") Signed-off-by: Calvin Owens <calvin@wbinvd.org> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
ZhengYuan Huang [Wed, 25 Mar 2026 00:43:39 +0000 (08:43 +0800)]
btrfs: fix check_chunk_block_group_mappings() to iterate all chunk maps
[BUG]
A corrupted image with a chunk present in the chunk tree but whose
corresponding block group item is missing from the extent tree can be
mounted successfully, even though check_chunk_block_group_mappings()
is supposed to catch exactly this corruption at mount time. Once
mounted, running btrfs balance with a usage filter (-dusage=N or
-dusage=min..max) triggers a null-ptr-deref:
KASAN: null-ptr-deref in range [0x0000000000000070-0x0000000000000077]
RIP: 0010:chunk_usage_filter fs/btrfs/volumes.c:3874 [inline]
RIP: 0010:should_balance_chunk fs/btrfs/volumes.c:4018 [inline]
RIP: 0010:__btrfs_balance fs/btrfs/volumes.c:4172 [inline]
RIP: 0010:btrfs_balance+0x2024/0x42b0 fs/btrfs/volumes.c:4604
[CAUSE]
The crash occurs because __btrfs_balance() iterates the on-disk chunk
tree, finds the orphaned chunk, calls chunk_usage_filter() (or
chunk_usage_range_filter()), which queries the in-memory block group
cache via btrfs_lookup_block_group(). Since no block group was ever
inserted for this chunk, the lookup returns NULL, and the subsequent
dereference of cache->used crashes.
check_chunk_block_group_mappings() uses btrfs_find_chunk_map() to
iterate the in-memory chunk map (fs_info->mapping_tree):
map = btrfs_find_chunk_map(fs_info, start, 1);
With @start = 0 and @length = 1, btrfs_find_chunk_map() looks for a
chunk map that *contains* the logical address 0. If no chunk contains
logical address 0, btrfs_find_chunk_map(fs_info, 0, 1) returns NULL
immediately and the loop breaks after the very first iteration,
having checked zero chunks. The entire verification function is therefore
a no-op, and the corrupted image passes the mount-time check undetected.
[FIX]
Replace the btrfs_find_chunk_map() based loop with a direct in-order
walk of fs_info->mapping_tree using rb_first_cached() + rb_next().
This guarantees that every chunk map in the tree is visited regardless
of the logical addresses involved.
No lock is taken around the traversal. This function is called during
mount from btrfs_read_block_groups(), which is invoked from open_ctree()
before any background threads (cleaner, transaction kthread, etc.) are
started. There are therefore no concurrent writers that could modify
mapping_tree at this point. An analogous lockless direct traversal of
mapping_tree already exists in fill_dummy_bgs() in the same file.
Since we walk the rb-tree directly via rb_entry() without going through
btrfs_find_chunk_map(), no reference is taken on each map entry, so the
btrfs_free_chunk_map() calls are also removed.
Signed-off-by: ZhengYuan Huang <gality369@gmail.com> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com>
Tejun Heo [Thu, 7 May 2026 22:09:21 +0000 (12:09 -1000)]
sched_ext: Drop unused scx_find_sub_sched() stub
scx_find_sub_sched()'s only caller, scx_bpf_sub_dispatch(), is gated on
CONFIG_EXT_SUB_SCHED. When CONFIG_EXT_SUB_SCHED=n the caller compiles out
and the stub becomes dead code, tripping -Wunused-function on randconfigs.
Drop the stub.
Fixes: 25037af712eb ("sched_ext: Add rhashtable lookup for sub-schedulers") Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/all/202605080556.42PXw8U9-lkp@intel.com/ Signed-off-by: Tejun Heo <tj@kernel.org>
Chen Wandun [Thu, 7 May 2026 10:54:34 +0000 (18:54 +0800)]
cgroup/cpuset: move PF_EXITING check before __GFP_HARDWALL in cpuset_current_node_allowed()
Since prepare_alloc_pages() unconditionally adds __GFP_HARDWALL for the
fast path when cpusets are enabled, the __GFP_HARDWALL check in
cpuset_current_node_allowed() causes the PF_EXITING escape path to be
skipped on the first allocation attempt. This makes it unreachable in
the common case, so dying tasks can get stuck in direct reclaim or even
trigger OOM while trying to exit, despite being allowed to allocate from
any node.
Move the PF_EXITING check before __GFP_HARDWALL so that dying tasks
can allocate memory from any node to exit quickly, even when cpusets
are enabled.
Also update the function comment to reflect the actual behavior of
prepare_alloc_pages() and the corrected check ordering.
Signed-off-by: Chen Wandun <chenwandun@lixiang.com> Acked-by: Michal Koutný <mkoutny@suse.com> Acked-by: Waiman Long <longman@redhat.com> Signed-off-by: Tejun Heo <tj@kernel.org>
Tejun Heo [Thu, 7 May 2026 21:05:31 +0000 (11:05 -1000)]
sched_ext: Move scx_error() out of scx_link_sched()'s lock region
scx_link_sched() holds scx_sched_lock. The scx_error() calls inside take the
same lock through scx_claim_exit() and deadlock. Move them out of the guard.
Fixes: 6b4576b09714 ("sched_ext: Reject sub-sched attachment to a disabled parent") Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Andrea Righi <arighi@nvidia.com>
smb: client: validate dacloffset before building DACL pointers
parse_sec_desc(), build_sec_desc(), and the chown path in
id_mode_to_cifs_acl() all add the server-supplied dacloffset to pntsd
before proving a DACL header fits inside the returned security
descriptor.
On 32-bit builds a malicious server can return dacloffset near
U32_MAX, wrap the derived DACL pointer below end_of_acl, and then slip
past the later pointer-based bounds checks. build_sec_desc() and
id_mode_to_cifs_acl() can then dereference DACL fields from the wrapped
pointer in the chmod/chown rewrite paths.
Validate dacloffset numerically before building any DACL pointer and
reuse the same helper at the three DACL entry points.
Fixes: bc3e9dd9d104 ("cifs: Change SIDs in ACEs while transferring file ownership.") Cc: stable@vger.kernel.org Assisted-by: Claude:claude-opus-4-6 Signed-off-by: Michael Bommarito <michael.bommarito@gmail.com> Signed-off-by: Steve French <stfrench@microsoft.com>
Zisen Ye [Wed, 6 May 2026 03:49:08 +0000 (11:49 +0800)]
smb/client: fix out-of-bounds read in smb2_compound_op()
If a server sends a truncated response but a large OutputBufferLength, and
terminates the EA list early, check_wsl_eas() returns success without
validating that the entire OutputBufferLength fits within iov_len.
Then smb2_compound_op() does:
memcpy(idata->wsl.eas, data[0], size[0]);
Where size[0] is OutputBufferLength. If iov_len is smaller than size[0],
memcpy can read beyond the end of the rsp_iov allocation and leak adjacent
kernel heap memory.
Zisen Ye [Sat, 2 May 2026 10:48:36 +0000 (18:48 +0800)]
smb/client: fix out-of-bounds read in symlink_data()
Since smb2_check_message() returns success without length validation for
the symlink error response, in symlink_data() it is possible for
iov->iov_len to be smaller than sizeof(struct smb2_err_rsp). If the buffer
only contains the base SMB2 header (64 bytes), accessing
err->ErrorContextCount (at offset 66) or err->ByteCount later in
symlink_data() will cause an out-of-bounds read.
Piyush Sachdeva [Thu, 7 May 2026 16:52:14 +0000 (22:22 +0530)]
smb: client: Zero-pad short GSS session keys per MS-SMB2
Per MS-SMB2 section 3.2.5.3, Session.SessionKey is the first 16 bytes
of the GSS cryptographic key, right-padded with zero bytes if the key
is shorter than 16 bytes.
SMB2_auth_kerberos() copies the GSS session key from the cifs.upcall
response using kmemdup(msg->data, msg->sesskey_len, ...) and stores
the GSS-reported length verbatim in ses->auth_key.len. generate_key()
reads SMB2_NTLMV2_SESSKEY_SIZE bytes from this buffer when feeding the
HMAC-SHA256 KDF for signing key derivation. If a GSS mechanism returns
a session key shorter than 16 bytes (e.g. a deprecated single-DES
Kerberos enctype with an 8-byte session key), the KDF call performs an
out-of-bounds slab read and derives keys that do not match the server,
which pads per the spec.
Modern KDCs disable short-key enctypes by default, so this is latent
rather than reachable in production, but it is still a kernel heap
over-read.
Allocate auth_key.response with kzalloc() at a length of
max(msg->sesskey_len, SMB2_NTLMV2_SESSKEY_SIZE), copy the GSS key in,
and rely on kzalloc()'s zero initialization for the spec-mandated
padding. Set ses->auth_key.len to the padded length. Larger GSS keys
(e.g. the 32-byte aes256-cts-hmac-sha1-96 session key) continue to be
stored at their natural length, preserving the FullSessionKey path.
Emit a cifs_dbg(VFS, ...) message when a short key is encountered to
surface deprecated-enctype usage.
NTLMv2 and NTLMSSP code paths produce a 16-byte session key by
construction and are unaffected.
Signed-off-by: Piyush Sachdeva <psachdeva@microsoft.com> Signed-off-by: Piyush Sachdeva <s.piyush1024@gmail.com> Signed-off-by: Steve French <stfrench@microsoft.com>
Piyush Sachdeva [Thu, 7 May 2026 16:52:13 +0000 (22:22 +0530)]
smb: client: Use FullSessionKey for AES-256 encryption key derivation
When Kerberos authentication is used with AES-256 encryption (AES-256-CCM
or AES-256-GCM), the SMB3 encryption and decryption keys must be derived
using the full session key (Session.FullSessionKey) rather than just the
first 16 bytes (Session.SessionKey).
Per MS-SMB2 section 3.2.5.3.1, when Connection.Dialect is "3.1.1" and
Connection.CipherId is AES-256-CCM or AES-256-GCM, Session.FullSessionKey
must be set to the full cryptographic key from the GSS authentication
context. The encryption and decryption key derivation (SMBC2SCipherKey,
SMBS2CCipherKey) must use this FullSessionKey as the KDF input. The
signing key derivation continues to use Session.SessionKey (first 16
bytes) in all cases.
Previously, generate_key() hardcoded SMB2_NTLMV2_SESSKEY_SIZE (16) as the
HMAC-SHA256 key input length for all derivations. When Kerberos with
AES-256 provides a 32-byte session key, the KDF for encryption/decryption
was using only the first 16 bytes, producing keys that did not match the
server's, causing mount failures with sec=krb5 and require_gcm_256=1.
Add a full_key_size parameter to generate_key() and pass the appropriate
size from generate_smb3signingkey():
- Signing: always SMB2_NTLMV2_SESSKEY_SIZE (16 bytes)
- Encryption/Decryption: ses->auth_key.len when AES-256, otherwise 16
Also fix cifs_dump_full_key() to report the actual session key length for
AES-256 instead of hardcoded CIFS_SESS_KEY_SIZE, so that userspace tools
like Wireshark receive the correct key for decryption.
Cc: <stable@vger.kernel.org> Reviewed-by: Bharath SM <bharathsm@microsoft.com> Signed-off-by: Piyush Sachdeva <psachdeva@microsoft.com> Signed-off-by: Piyush Sachdeva <s.piyush1024@gmail.com> Signed-off-by: Steve French <stfrench@microsoft.com>
Dmitry Torokhov [Mon, 4 May 2026 18:54:46 +0000 (11:54 -0700)]
Input: atmel_mxt_ts - check mem_size before calculating config memory size
In mxt_update_cfg(), the driver calculates the memory size needed to store
the configuration as data->mem_size - cfg.start_ofs. If data->mem_size is
less than or equal to cfg.start_ofs, this calculation will underflow or
result in a zero-size buffer, neither of which is valid for a configuration
update.
Add a check to return -EINVAL if data->mem_size is too small. While at it,
change the types of start_ofs and mem_size in struct mxt_cfg to u16 to
match the device address space.
Dmitry Torokhov [Mon, 4 May 2026 18:54:45 +0000 (11:54 -0700)]
Input: atmel_mxt_ts - fix boundary check in mxt_prepare_cfg_mem
When a configuration file provides an object size that is larger than the
driver's known mxt_obj_size(object), the driver intends to discard the
extra bytes.
The loop iterates using for (i = 0; i < size; i++). Inside the loop, the
condition to skip processing extra bytes is:
if (i > mxt_obj_size(object))
continue;
Since i is a 0-based index, the valid indices for the object are 0 through
mxt_obj_size(object) - 1.
When i == mxt_obj_size(object), the condition evaluates to false, and the
code processes the byte instead of discarding it.
This causes the code to calculate byte_offset = reg + i - cfg->start_ofs
and writes the byte there, overwriting exactly one byte of the adjacent
instance or object.
Update the boundary check to skip extra bytes correctly by using >=.
projects / thirdparty / kernel / stable.git / log
