| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| In the Linux kernel, the following vulnerability has been resolved:
landlock: Handle weird files
A corrupted filesystem (e.g. bcachefs) might return weird files.
Instead of throwing a warning and allowing access to such file, treat
them as regular files. |
| In the Linux kernel, the following vulnerability has been resolved:
RDMA/rxe: Fix the warning "__rxe_cleanup+0x12c/0x170 [rdma_rxe]"
The Call Trace is as below:
"
<TASK>
? show_regs.cold+0x1a/0x1f
? __rxe_cleanup+0x12c/0x170 [rdma_rxe]
? __warn+0x84/0xd0
? __rxe_cleanup+0x12c/0x170 [rdma_rxe]
? report_bug+0x105/0x180
? handle_bug+0x46/0x80
? exc_invalid_op+0x19/0x70
? asm_exc_invalid_op+0x1b/0x20
? __rxe_cleanup+0x12c/0x170 [rdma_rxe]
? __rxe_cleanup+0x124/0x170 [rdma_rxe]
rxe_destroy_qp.cold+0x24/0x29 [rdma_rxe]
ib_destroy_qp_user+0x118/0x190 [ib_core]
rdma_destroy_qp.cold+0x43/0x5e [rdma_cm]
rtrs_cq_qp_destroy.cold+0x1d/0x2b [rtrs_core]
rtrs_srv_close_work.cold+0x1b/0x31 [rtrs_server]
process_one_work+0x21d/0x3f0
worker_thread+0x4a/0x3c0
? process_one_work+0x3f0/0x3f0
kthread+0xf0/0x120
? kthread_complete_and_exit+0x20/0x20
ret_from_fork+0x22/0x30
</TASK>
"
When too many rdma resources are allocated, rxe needs more time to
handle these rdma resources. Sometimes with the current timeout, rxe
can not release the rdma resources correctly.
Compared with other rdma drivers, a bigger timeout is used. |
| In the Linux kernel, the following vulnerability has been resolved:
wifi: mac80211: don't flush non-uploaded STAs
If STA state is pre-moved to AUTHORIZED (such as in IBSS
scenarios) and insertion fails, the station is freed.
In this case, the driver never knew about the station,
so trying to flush it is unexpected and may crash.
Check if the sta was uploaded to the driver before and
fix this. |
| In the Linux kernel, the following vulnerability has been resolved:
Bluetooth: btusb: mediatek: Add locks for usb_driver_claim_interface()
The documentation for usb_driver_claim_interface() says that "the
device lock" is needed when the function is called from places other
than probe(). This appears to be the lock for the USB interface
device. The Mediatek btusb code gets called via this path:
Workqueue: hci0 hci_power_on [bluetooth]
Call trace:
usb_driver_claim_interface
btusb_mtk_claim_iso_intf
btusb_mtk_setup
hci_dev_open_sync
hci_power_on
process_scheduled_works
worker_thread
kthread
With the above call trace the device lock hasn't been claimed. Claim
it.
Without this fix, we'd sometimes see the error "Failed to claim iso
interface". Sometimes we'd even see worse errors, like a NULL pointer
dereference (where `intf->dev.driver` was NULL) with a trace like:
Call trace:
usb_suspend_both
usb_runtime_suspend
__rpm_callback
rpm_suspend
pm_runtime_work
process_scheduled_works
Both errors appear to be fixed with the proper locking. |
| In the Linux kernel, the following vulnerability has been resolved:
netfilter: nf_tables: reject mismatching sum of field_len with set key length
The field length description provides the length of each separated key
field in the concatenation, each field gets rounded up to 32-bits to
calculate the pipapo rule width from pipapo_init(). The set key length
provides the total size of the key aligned to 32-bits.
Register-based arithmetics still allows for combining mismatching set
key length and field length description, eg. set key length 10 and field
description [ 5, 4 ] leading to pipapo width of 12. |
| In the Linux kernel, the following vulnerability has been resolved:
bpf: Cancel the running bpf_timer through kworker for PREEMPT_RT
During the update procedure, when overwrite element in a pre-allocated
htab, the freeing of old_element is protected by the bucket lock. The
reason why the bucket lock is necessary is that the old_element has
already been stashed in htab->extra_elems after alloc_htab_elem()
returns. If freeing the old_element after the bucket lock is unlocked,
the stashed element may be reused by concurrent update procedure and the
freeing of old_element will run concurrently with the reuse of the
old_element. However, the invocation of check_and_free_fields() may
acquire a spin-lock which violates the lockdep rule because its caller
has already held a raw-spin-lock (bucket lock). The following warning
will be reported when such race happens:
BUG: scheduling while atomic: test_progs/676/0x00000003
3 locks held by test_progs/676:
#0: ffffffff864b0240 (rcu_read_lock_trace){....}-{0:0}, at: bpf_prog_test_run_syscall+0x2c0/0x830
#1: ffff88810e961188 (&htab->lockdep_key){....}-{2:2}, at: htab_map_update_elem+0x306/0x1500
#2: ffff8881f4eac1b8 (&base->softirq_expiry_lock){....}-{2:2}, at: hrtimer_cancel_wait_running+0xe9/0x1b0
Modules linked in: bpf_testmod(O)
Preemption disabled at:
[<ffffffff817837a3>] htab_map_update_elem+0x293/0x1500
CPU: 0 UID: 0 PID: 676 Comm: test_progs Tainted: G ... 6.12.0+ #11
Tainted: [W]=WARN, [O]=OOT_MODULE
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996)...
Call Trace:
<TASK>
dump_stack_lvl+0x57/0x70
dump_stack+0x10/0x20
__schedule_bug+0x120/0x170
__schedule+0x300c/0x4800
schedule_rtlock+0x37/0x60
rtlock_slowlock_locked+0x6d9/0x54c0
rt_spin_lock+0x168/0x230
hrtimer_cancel_wait_running+0xe9/0x1b0
hrtimer_cancel+0x24/0x30
bpf_timer_delete_work+0x1d/0x40
bpf_timer_cancel_and_free+0x5e/0x80
bpf_obj_free_fields+0x262/0x4a0
check_and_free_fields+0x1d0/0x280
htab_map_update_elem+0x7fc/0x1500
bpf_prog_9f90bc20768e0cb9_overwrite_cb+0x3f/0x43
bpf_prog_ea601c4649694dbd_overwrite_timer+0x5d/0x7e
bpf_prog_test_run_syscall+0x322/0x830
__sys_bpf+0x135d/0x3ca0
__x64_sys_bpf+0x75/0xb0
x64_sys_call+0x1b5/0xa10
do_syscall_64+0x3b/0xc0
entry_SYSCALL_64_after_hwframe+0x4b/0x53
...
</TASK>
It seems feasible to break the reuse and refill of per-cpu extra_elems
into two independent parts: reuse the per-cpu extra_elems with bucket
lock being held and refill the old_element as per-cpu extra_elems after
the bucket lock is unlocked. However, it will make the concurrent
overwrite procedures on the same CPU return unexpected -E2BIG error when
the map is full.
Therefore, the patch fixes the lock problem by breaking the cancelling
of bpf_timer into two steps for PREEMPT_RT:
1) use hrtimer_try_to_cancel() and check its return value
2) if the timer is running, use hrtimer_cancel() through a kworker to
cancel it again
Considering that the current implementation of hrtimer_cancel() will try
to acquire a being held softirq_expiry_lock when the current timer is
running, these steps above are reasonable. However, it also has
downside. When the timer is running, the cancelling of the timer is
delayed when releasing the last map uref. The delay is also fixable
(e.g., break the cancelling of bpf timer into two parts: one part in
locked scope, another one in unlocked scope), it can be revised later if
necessary.
It is a bit hard to decide the right fix tag. One reason is that the
problem depends on PREEMPT_RT which is enabled in v6.12. Considering the
softirq_expiry_lock lock exists since v5.4 and bpf_timer is introduced
in v5.15, the bpf_timer commit is used in the fixes tag and an extra
depends-on tag is added to state the dependency on PREEMPT_RT.
Depends-on: v6.12+ with PREEMPT_RT enabled |
| In the Linux kernel, the following vulnerability has been resolved:
kernel: be more careful about dup_mmap() failures and uprobe registering
If a memory allocation fails during dup_mmap(), the maple tree can be left
in an unsafe state for other iterators besides the exit path. All the
locks are dropped before the exit_mmap() call (in mm/mmap.c), but the
incomplete mm_struct can be reached through (at least) the rmap finding
the vmas which have a pointer back to the mm_struct.
Up to this point, there have been no issues with being able to find an
mm_struct that was only partially initialised. Syzbot was able to make
the incomplete mm_struct fail with recent forking changes, so it has been
proven unsafe to use the mm_struct that hasn't been initialised, as
referenced in the link below.
Although 8ac662f5da19f ("fork: avoid inappropriate uprobe access to
invalid mm") fixed the uprobe access, it does not completely remove the
race.
This patch sets the MMF_OOM_SKIP to avoid the iteration of the vmas on the
oom side (even though this is extremely unlikely to be selected as an oom
victim in the race window), and sets MMF_UNSTABLE to avoid other potential
users from using a partially initialised mm_struct.
When registering vmas for uprobe, skip the vmas in an mm that is marked
unstable. Modifying a vma in an unstable mm may cause issues if the mm
isn't fully initialised. |
| In the Linux kernel, the following vulnerability has been resolved:
gtp: Destroy device along with udp socket's netns dismantle.
gtp_newlink() links the device to a list in dev_net(dev) instead of
src_net, where a udp tunnel socket is created.
Even when src_net is removed, the device stays alive on dev_net(dev).
Then, removing src_net triggers the splat below. [0]
In this example, gtp0 is created in ns2, and the udp socket is created
in ns1.
ip netns add ns1
ip netns add ns2
ip -n ns1 link add netns ns2 name gtp0 type gtp role sgsn
ip netns del ns1
Let's link the device to the socket's netns instead.
Now, gtp_net_exit_batch_rtnl() needs another netdev iteration to remove
all gtp devices in the netns.
[0]:
ref_tracker: net notrefcnt@000000003d6e7d05 has 1/2 users at
sk_alloc (./include/net/net_namespace.h:345 net/core/sock.c:2236)
inet_create (net/ipv4/af_inet.c:326 net/ipv4/af_inet.c:252)
__sock_create (net/socket.c:1558)
udp_sock_create4 (net/ipv4/udp_tunnel_core.c:18)
gtp_create_sock (./include/net/udp_tunnel.h:59 drivers/net/gtp.c:1423)
gtp_create_sockets (drivers/net/gtp.c:1447)
gtp_newlink (drivers/net/gtp.c:1507)
rtnl_newlink (net/core/rtnetlink.c:3786 net/core/rtnetlink.c:3897 net/core/rtnetlink.c:4012)
rtnetlink_rcv_msg (net/core/rtnetlink.c:6922)
netlink_rcv_skb (net/netlink/af_netlink.c:2542)
netlink_unicast (net/netlink/af_netlink.c:1321 net/netlink/af_netlink.c:1347)
netlink_sendmsg (net/netlink/af_netlink.c:1891)
____sys_sendmsg (net/socket.c:711 net/socket.c:726 net/socket.c:2583)
___sys_sendmsg (net/socket.c:2639)
__sys_sendmsg (net/socket.c:2669)
do_syscall_64 (arch/x86/entry/common.c:52 arch/x86/entry/common.c:83)
WARNING: CPU: 1 PID: 60 at lib/ref_tracker.c:179 ref_tracker_dir_exit (lib/ref_tracker.c:179)
Modules linked in:
CPU: 1 UID: 0 PID: 60 Comm: kworker/u16:2 Not tainted 6.13.0-rc5-00147-g4c1224501e9d #5
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
Workqueue: netns cleanup_net
RIP: 0010:ref_tracker_dir_exit (lib/ref_tracker.c:179)
Code: 00 00 00 fc ff df 4d 8b 26 49 bd 00 01 00 00 00 00 ad de 4c 39 f5 0f 85 df 00 00 00 48 8b 74 24 08 48 89 df e8 a5 cc 12 02 90 <0f> 0b 90 48 8d 6b 44 be 04 00 00 00 48 89 ef e8 80 de 67 ff 48 89
RSP: 0018:ff11000009a07b60 EFLAGS: 00010286
RAX: 0000000000002bd3 RBX: ff1100000f4e1aa0 RCX: 1ffffffff0e40ac6
RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffff8423ee3c
RBP: ff1100000f4e1af0 R08: 0000000000000001 R09: fffffbfff0e395ae
R10: 0000000000000001 R11: 0000000000036001 R12: ff1100000f4e1af0
R13: dead000000000100 R14: ff1100000f4e1af0 R15: dffffc0000000000
FS: 0000000000000000(0000) GS:ff1100006ce80000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007f9b2464bd98 CR3: 0000000005286005 CR4: 0000000000771ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe07f0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
<TASK>
? __warn (kernel/panic.c:748)
? ref_tracker_dir_exit (lib/ref_tracker.c:179)
? report_bug (lib/bug.c:201 lib/bug.c:219)
? handle_bug (arch/x86/kernel/traps.c:285)
? exc_invalid_op (arch/x86/kernel/traps.c:309 (discriminator 1))
? asm_exc_invalid_op (./arch/x86/include/asm/idtentry.h:621)
? _raw_spin_unlock_irqrestore (./arch/x86/include/asm/irqflags.h:42 ./arch/x86/include/asm/irqflags.h:97 ./arch/x86/include/asm/irqflags.h:155 ./include/linux/spinlock_api_smp.h:151 kernel/locking/spinlock.c:194)
? ref_tracker_dir_exit (lib/ref_tracker.c:179)
? __pfx_ref_tracker_dir_exit (lib/ref_tracker.c:158)
? kfree (mm/slub.c:4613 mm/slub.c:4761)
net_free (net/core/net_namespace.c:476 net/core/net_namespace.c:467)
cleanup_net (net/core/net_namespace.c:664 (discriminator 3))
process_one_work (kernel/workqueue.c:3229)
worker_thread (kernel/workqueue.c:3304 kernel/workqueue.c:3391
---truncated--- |
| In the Linux kernel, the following vulnerability has been resolved:
net: fec: handle page_pool_dev_alloc_pages error
The fec_enet_update_cbd function calls page_pool_dev_alloc_pages but did
not handle the case when it returned NULL. There was a WARN_ON(!new_page)
but it would still proceed to use the NULL pointer and then crash.
This case does seem somewhat rare but when the system is under memory
pressure it can happen. One case where I can duplicate this with some
frequency is when writing over a smbd share to a SATA HDD attached to an
imx6q.
Setting /proc/sys/vm/min_free_kbytes to higher values also seems to solve
the problem for my test case. But it still seems wrong that the fec driver
ignores the memory allocation error and can crash.
This commit handles the allocation error by dropping the current packet. |
| In the Linux kernel, the following vulnerability has been resolved:
smb: client: fix double free of TCP_Server_Info::hostname
When shutting down the server in cifs_put_tcp_session(), cifsd thread
might be reconnecting to multiple DFS targets before it realizes it
should exit the loop, so @server->hostname can't be freed as long as
cifsd thread isn't done. Otherwise the following can happen:
RIP: 0010:__slab_free+0x223/0x3c0
Code: 5e 41 5f c3 cc cc cc cc 4c 89 de 4c 89 cf 44 89 44 24 08 4c 89
1c 24 e8 fb cf 8e 00 44 8b 44 24 08 4c 8b 1c 24 e9 5f fe ff ff <0f>
0b 41 f7 45 08 00 0d 21 00 0f 85 2d ff ff ff e9 1f ff ff ff 80
RSP: 0018:ffffb26180dbfd08 EFLAGS: 00010246
RAX: ffff8ea34728e510 RBX: ffff8ea34728e500 RCX: 0000000000800068
RDX: 0000000000800068 RSI: 0000000000000000 RDI: ffff8ea340042400
RBP: ffffe112041ca380 R08: 0000000000000001 R09: 0000000000000000
R10: 6170732e31303000 R11: 70726f632e786563 R12: ffff8ea34728e500
R13: ffff8ea340042400 R14: ffff8ea34728e500 R15: 0000000000800068
FS: 0000000000000000(0000) GS:ffff8ea66fd80000(0000)
000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007ffc25376080 CR3: 000000012a2ba001 CR4:
PKRU: 55555554
Call Trace:
<TASK>
? show_trace_log_lvl+0x1c4/0x2df
? show_trace_log_lvl+0x1c4/0x2df
? __reconnect_target_unlocked+0x3e/0x160 [cifs]
? __die_body.cold+0x8/0xd
? die+0x2b/0x50
? do_trap+0xce/0x120
? __slab_free+0x223/0x3c0
? do_error_trap+0x65/0x80
? __slab_free+0x223/0x3c0
? exc_invalid_op+0x4e/0x70
? __slab_free+0x223/0x3c0
? asm_exc_invalid_op+0x16/0x20
? __slab_free+0x223/0x3c0
? extract_hostname+0x5c/0xa0 [cifs]
? extract_hostname+0x5c/0xa0 [cifs]
? __kmalloc+0x4b/0x140
__reconnect_target_unlocked+0x3e/0x160 [cifs]
reconnect_dfs_server+0x145/0x430 [cifs]
cifs_handle_standard+0x1ad/0x1d0 [cifs]
cifs_demultiplex_thread+0x592/0x730 [cifs]
? __pfx_cifs_demultiplex_thread+0x10/0x10 [cifs]
kthread+0xdd/0x100
? __pfx_kthread+0x10/0x10
ret_from_fork+0x29/0x50
</TASK> |
| FFmpeg through 8.1.2, fixed in commit 5d7112c, contains a heap out-of-bounds write vulnerability in the vf_hqdn3d filter that allows attackers to corrupt heap memory by supplying a crafted video whose frame resolution increases between frames when filtergraph reinitialization is disabled via the -reinit_filter 0 option. Attackers can provide a malicious video input where vf_hqdn3d.config_input() allocates undersized per-plane line-history buffers based on the initial frame width, and subsequent larger frames cause denoise_spatial() to write beyond the allocation boundary, resulting in heap memory corruption. |
| FFmpeg through 8.1.2, fixed in commit 5d7112c, contains an uncontrolled resource consumption vulnerability in the IAMF demuxer that allows an unauthenticated attacker to cause multi-gigabyte memory allocation from a 17-byte input file by supplying a crafted count_label field. The mix_presentation_obu() function in libavformat/iamf_parse.c calls av_calloc(count_label, sizeof(*language_label)) with an attacker-controlled value before validating available OBU data, enabling an allocation amplification of approximately 126 million bytes per input byte that exhausts process memory or triggers an OOM-kill during format probing. |
| FFmpeg through 8.1.2, fixed in commit aafb5c6, contains a signed integer overflow vulnerability in the MACE6 audio decoder that allows attackers to corrupt heap memory by supplying a crafted CAF file with a malicious bytes_per_packet value. Attackers can craft a CAF file with oversized bytes_per_packet and frames_per_packet values in the desc chunk to trigger an integer overflow in mace_decode_frame() during output sample count computation, resulting in an undersized buffer allocation and heap out-of-bounds write that could enable code execution. |
| FFmpeg through 8.1.2, fixed in commit b506faf, contains a heap out-of-bounds write vulnerability in the native PNG and APNG encoders that allows remote attackers to corrupt heap memory by supplying a crafted PNG image with a malicious eXIf chunk. Attackers can craft an eXIf chunk where multiple IFD entries reference the same large value payload, causing canonical serialization to expand the output far beyond the undersized allocation estimated by add_exif_profile_size(), resulting in png_write_chunk() writing tens of thousands of bytes past the buffer boundary, leading to deterministic heap corruption, process crash, and potentially arbitrary code execution. |
| FFmpeg 7.0 through 8.1.2, fixed in commit 4da9812, contains a heap out-of-bounds write vulnerability in the vf_quirc filter that allows an attacker to corrupt heap memory by supplying a crafted PGS/SUP subtitle file with mismatched frame dimensions. Attackers can provide a subtitle file whose second presentation has larger dimensions than its first, causing av_image_copy_plane() to copy data exceeding the initial allocation size into the undersized libquirc grayscale image buffer, resulting in heap corruption and process crash with potential for code execution. |
| In the Linux kernel, the following vulnerability has been resolved:
fuse-uring: Avoid use-after-free in fuse_uring_async_stop_queues
fuse_uring_async_stop_queues() might run when the last reference
on ring->queue_refs was already dropped.
In order to avoid an early destruction a reference on struct fuse_conn
is now taken before starting fuse_uring_async_stop_queues() and that
reference is only released when that delayed work queue terminates. |
| In the Linux kernel, the following vulnerability has been resolved:
RDMA/rtrs-srv: Bound RDMA-Write length to chunk size in rdma_write_sg
When the server answers an RTRS READ, rdma_write_sg() builds the source
scatter/gather entry for the IB_WR_RDMA_WRITE that returns data to the
peer. Its length is taken directly from the wire descriptor:
plist->length = le32_to_cpu(id->rd_msg->desc[0].len);
rd_msg points into the chunk buffer that the remote peer filled via
RDMA-WRITE-WITH-IMM (rtrs_srv_rdma_done() -> process_io_req() ->
process_read()), so desc[0].len is attacker-controlled and, before this
change, was only rejected when zero. The source address is the fixed
chunk start (dma_addr[msg_id]) and the source lkey is the PD-wide
local_dma_lkey, which is not tied to the chunk's MR mapping, so the verbs
layer does not constrain the transfer length to max_chunk_size. msg_id
and off are bounded against queue_depth and max_chunk_size in
rtrs_srv_rdma_done(), but desc[0].len is a separate field that was not
checked against the chunk size.
A peer that advertises desc[0].len larger than max_chunk_size can make
the posted RDMA write read past the chunk's mapped region. The resulting
behaviour depends on the IOMMU configuration: with no IOMMU or in
passthrough mode the read may extend into memory adjacent to the chunk
and be returned to the peer, which can disclose host memory; with a
translating IOMMU the out-of-range access is expected to fault and abort
the connection. In either case the transfer exceeds what the protocol
permits and is driven by a remote peer.
Reject a descriptor length above max_chunk_size, mirroring the existing
off >= max_chunk_size bound in rtrs_srv_rdma_done(). Legitimate clients
do not exceed it: the client sets desc[0].len to its MR length, which is
capped at the negotiated max_io_size (max_chunk_size - MAX_HDR_SIZE). |
| In the Linux kernel, the following vulnerability has been resolved:
iommufd: Set veventq_depth upper bound
iommufd_veventq_alloc() accepts any !0 veventq_depth from userspace, with
an upper bound at U32_MAX.
This leaves a vulnerability where userspace can allocate excessively large
queues to exhaust kernel memory reserves.
Cap the veventq_depth (maximum number of entries) to 1 << 19, matching the
maximum number of entries in the SMMUv3 EVTQ (the largest use case today). |
| In the Linux kernel, the following vulnerability has been resolved:
nvmet-auth: validate reply message payload bounds against transfer length
nvmet_auth_reply() accesses the variable-length rval[] array using
attacker-controlled hl (hash length) and dhvlen (DH value length) fields
without verifying they fit within the allocated buffer of tl bytes.
A malicious NVMe-oF initiator can craft a DHCHAP_REPLY message with a
small transfer length but large hl/dhvlen values, causing out-of-bounds
heap reads when the target processes the DH public key (rval + 2*hl) or
performs the host response memcmp.
With DH authentication configured, the OOB pointer is passed directly to
sg_init_one() and read by crypto_kpp_compute_shared_secret(), reaching
up to 526 bytes past the buffer. This is exploitable pre-authentication.
Add bounds validation ensuring sizeof(*data) + 2*hl + dhvlen <= tl before
any access to the variable-length fields.
Discovered by Atuin - Automated Vulnerability Discovery Engine. |
| In the Linux kernel, the following vulnerability has been resolved:
block: skip sync_blockdev() on surprise removal in bdev_mark_dead()
bdev_mark_dead()'s @surprise == true means the device is already gone.
The filesystem callback fs_bdev_mark_dead() honours this and skips
sync_filesystem(), but the bare block device path (no ->mark_dead op)
lost its !surprise guard when the holder ->mark_dead callback was wired
up (see Fixes), and now calls sync_blockdev() unconditionally, which can
hang forever waiting on writeback that can no longer complete.
syzkaller hit this via nvme_reset_work()'s "I/O queues lost" path:
nvme_mark_namespaces_dead() -> blk_mark_disk_dead() ->
bdev_mark_dead(bdev, true) -> sync_blockdev() blocks in
folio_wait_writeback(), wedging the reset worker and every task waiting
on it.
Skip the sync on surprise removal, matching fs_bdev_mark_dead();
invalidate_bdev() still runs. Orderly removal (surprise == false) is
unchanged.
Found by FuzzNvme(Syzkaller with FEMU fuzzing framework). |
