| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| In the Linux kernel, the following vulnerability has been resolved:
bcachefs: bch2_ioctl_subvolume_destroy() fixes
bch2_evict_subvolume_inodes() was getting stuck - due to incorrectly
pruning the dcache.
Also, fix missing permissions checks. |
| In the Linux kernel, the following vulnerability has been resolved:
nfsd: fix legacy client tracking initialization
Get rid of the nfsd4_legacy_tracking_ops->init() call in
check_for_legacy_methods(). That will be handled in the caller
(nfsd4_client_tracking_init()). Otherwise, we'll wind up calling
nfsd4_legacy_tracking_ops->init() twice, and the second time we'll
trigger the BUG_ON() in nfsd4_init_recdir(). |
| In the Linux kernel, the following vulnerability has been resolved:
mm: zswap: fix crypto_free_acomp() deadlock in zswap_cpu_comp_dead()
Currently, zswap_cpu_comp_dead() calls crypto_free_acomp() while holding
the per-CPU acomp_ctx mutex. crypto_free_acomp() then holds scomp_lock
(through crypto_exit_scomp_ops_async()).
On the other hand, crypto_alloc_acomp_node() holds the scomp_lock (through
crypto_scomp_init_tfm()), and then allocates memory. If the allocation
results in reclaim, we may attempt to hold the per-CPU acomp_ctx mutex.
The above dependencies can cause an ABBA deadlock. For example in the
following scenario:
(1) Task A running on CPU #1:
crypto_alloc_acomp_node()
Holds scomp_lock
Enters reclaim
Reads per_cpu_ptr(pool->acomp_ctx, 1)
(2) Task A is descheduled
(3) CPU #1 goes offline
zswap_cpu_comp_dead(CPU #1)
Holds per_cpu_ptr(pool->acomp_ctx, 1))
Calls crypto_free_acomp()
Waits for scomp_lock
(4) Task A running on CPU #2:
Waits for per_cpu_ptr(pool->acomp_ctx, 1) // Read on CPU #1
DEADLOCK
Since there is no requirement to call crypto_free_acomp() with the per-CPU
acomp_ctx mutex held in zswap_cpu_comp_dead(), move it after the mutex is
unlocked. Also move the acomp_request_free() and kfree() calls for
consistency and to avoid any potential sublte locking dependencies in the
future.
With this, only setting acomp_ctx fields to NULL occurs with the mutex
held. This is similar to how zswap_cpu_comp_prepare() only initializes
acomp_ctx fields with the mutex held, after performing all allocations
before holding the mutex.
Opportunistically, move the NULL check on acomp_ctx so that it takes place
before the mutex dereference. |
| In the Linux kernel, the following vulnerability has been resolved:
media: vimc: skip .s_stream() for stopped entities
Syzbot reported [1] a warning prompted by a check in call_s_stream()
that checks whether .s_stream() operation is warranted for unstarted
or stopped subdevs.
Add a simple fix in vimc_streamer_pipeline_terminate() ensuring that
entities skip a call to .s_stream() unless they have been previously
properly started.
[1] Syzbot report:
------------[ cut here ]------------
WARNING: CPU: 0 PID: 5933 at drivers/media/v4l2-core/v4l2-subdev.c:460 call_s_stream+0x2df/0x350 drivers/media/v4l2-core/v4l2-subdev.c:460
Modules linked in:
CPU: 0 UID: 0 PID: 5933 Comm: syz-executor330 Not tainted 6.13.0-rc2-syzkaller-00362-g2d8308bf5b67 #0
...
Call Trace:
<TASK>
vimc_streamer_pipeline_terminate+0x218/0x320 drivers/media/test-drivers/vimc/vimc-streamer.c:62
vimc_streamer_pipeline_init drivers/media/test-drivers/vimc/vimc-streamer.c:101 [inline]
vimc_streamer_s_stream+0x650/0x9a0 drivers/media/test-drivers/vimc/vimc-streamer.c:203
vimc_capture_start_streaming+0xa1/0x130 drivers/media/test-drivers/vimc/vimc-capture.c:256
vb2_start_streaming+0x15f/0x5a0 drivers/media/common/videobuf2/videobuf2-core.c:1789
vb2_core_streamon+0x2a7/0x450 drivers/media/common/videobuf2/videobuf2-core.c:2348
vb2_streamon drivers/media/common/videobuf2/videobuf2-v4l2.c:875 [inline]
vb2_ioctl_streamon+0xf4/0x170 drivers/media/common/videobuf2/videobuf2-v4l2.c:1118
__video_do_ioctl+0xaf0/0xf00 drivers/media/v4l2-core/v4l2-ioctl.c:3122
video_usercopy+0x4d2/0x1620 drivers/media/v4l2-core/v4l2-ioctl.c:3463
v4l2_ioctl+0x1ba/0x250 drivers/media/v4l2-core/v4l2-dev.c:366
vfs_ioctl fs/ioctl.c:51 [inline]
__do_sys_ioctl fs/ioctl.c:906 [inline]
__se_sys_ioctl fs/ioctl.c:892 [inline]
__x64_sys_ioctl+0x190/0x200 fs/ioctl.c:892
do_syscall_x64 arch/x86/entry/common.c:52 [inline]
do_syscall_64+0xcd/0x250 arch/x86/entry/common.c:83
entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7f2b85c01b19
... |
| In the Linux kernel, the following vulnerability has been resolved:
wifi: ath11k: add srng->lock for ath11k_hal_srng_* in monitor mode
ath11k_hal_srng_* should be used with srng->lock to protect srng data.
For ath11k_dp_rx_mon_dest_process() and ath11k_dp_full_mon_process_rx(),
they use ath11k_hal_srng_* for many times but never call srng->lock.
So when running (full) monitor mode, warning will occur:
RIP: 0010:ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k]
Call Trace:
? ath11k_hal_srng_dst_peek+0x18/0x30 [ath11k]
ath11k_dp_rx_process_mon_status+0xc45/0x1190 [ath11k]
? idr_alloc_u32+0x97/0xd0
ath11k_dp_rx_process_mon_rings+0x32a/0x550 [ath11k]
ath11k_dp_service_srng+0x289/0x5a0 [ath11k]
ath11k_pcic_ext_grp_napi_poll+0x30/0xd0 [ath11k]
__napi_poll+0x30/0x1f0
net_rx_action+0x198/0x320
__do_softirq+0xdd/0x319
So add srng->lock for them to avoid such warnings.
Inorder to fetch the srng->lock, should change srng's definition from
'void' to 'struct hal_srng'. And initialize them elsewhere to prevent
one line of code from being too long. This is consistent with other ring
process functions, such as ath11k_dp_process_rx().
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30
Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1 |
| In the Linux kernel, the following vulnerability has been resolved:
jfs: add check read-only before txBeginAnon() call
Added a read-only check before calling `txBeginAnon` in `extAlloc`
and `extRecord`. This prevents modification attempts on a read-only
mounted filesystem, avoiding potential errors or crashes.
Call trace:
txBeginAnon+0xac/0x154
extAlloc+0xe8/0xdec fs/jfs/jfs_extent.c:78
jfs_get_block+0x340/0xb98 fs/jfs/inode.c:248
__block_write_begin_int+0x580/0x166c fs/buffer.c:2128
__block_write_begin fs/buffer.c:2177 [inline]
block_write_begin+0x98/0x11c fs/buffer.c:2236
jfs_write_begin+0x44/0x88 fs/jfs/inode.c:299 |
| In the Linux kernel, the following vulnerability has been resolved:
jfs: add check read-only before truncation in jfs_truncate_nolock()
Added a check for "read-only" mode in the `jfs_truncate_nolock`
function to avoid errors related to writing to a read-only
filesystem.
Call stack:
block_write_begin() {
jfs_write_failed() {
jfs_truncate() {
jfs_truncate_nolock() {
txEnd() {
...
log = JFS_SBI(tblk->sb)->log;
// (log == NULL)
If the `isReadOnly(ip)` condition is triggered in
`jfs_truncate_nolock`, the function execution will stop, and no
further data modification will occur. Instead, the `xtTruncate`
function will be called with the "COMMIT_WMAP" flag, preventing
modifications in "read-only" mode. |
| In the Linux kernel, the following vulnerability has been resolved:
PCI/ASPM: Fix link state exit during switch upstream function removal
Before 456d8aa37d0f ("PCI/ASPM: Disable ASPM on MFD function removal to
avoid use-after-free"), we would free the ASPM link only after the last
function on the bus pertaining to the given link was removed.
That was too late. If function 0 is removed before sibling function,
link->downstream would point to free'd memory after.
After above change, we freed the ASPM parent link state upon any function
removal on the bus pertaining to a given link.
That is too early. If the link is to a PCIe switch with MFD on the upstream
port, then removing functions other than 0 first would free a link which
still remains parent_link to the remaining downstream ports.
The resulting GPFs are especially frequent during hot-unplug, because
pciehp removes devices on the link bus in reverse order.
On that switch, function 0 is the virtual P2P bridge to the internal bus.
Free exactly when function 0 is removed -- before the parent link is
obsolete, but after all subordinate links are gone.
[kwilczynski: commit log] |
| In the Linux kernel, the following vulnerability has been resolved:
usb: xhci: Don't skip on Stopped - Length Invalid
Up until commit d56b0b2ab142 ("usb: xhci: ensure skipped isoc TDs are
returned when isoc ring is stopped") in v6.11, the driver didn't skip
missed isochronous TDs when handling Stoppend and Stopped - Length
Invalid events. Instead, it erroneously cleared the skip flag, which
would cause the ring to get stuck, as future events won't match the
missed TD which is never removed from the queue until it's cancelled.
This buggy logic seems to have been in place substantially unchanged
since the 3.x series over 10 years ago, which probably speaks first
and foremost about relative rarity of this case in normal usage, but
by the spec I see no reason why it shouldn't be possible.
After d56b0b2ab142, TDs are immediately skipped when handling those
Stopped events. This poses a potential problem in case of Stopped -
Length Invalid, which occurs either on completed TDs (likely already
given back) or Link and No-Op TRBs. Such event won't be recognized
as matching any TD (unless it's the rare Link TRB inside a TD) and
will result in skipping all pending TDs, giving them back possibly
before they are done, risking isoc data loss and maybe UAF by HW.
As a compromise, don't skip and don't clear the skip flag on this
kind of event. Then the next event will skip missed TDs. A downside
of not handling Stopped - Length Invalid on a Link inside a TD is
that if the TD is cancelled, its actual length will not be updated
to account for TRBs (silently) completed before the TD was stopped.
I had no luck producing this sequence of completion events so there
is no compelling demonstration of any resulting disaster. It may be
a very rare, obscure condition. The sole motivation for this patch
is that if such unlikely event does occur, I'd rather risk reporting
a cancelled partially done isoc frame as empty than gamble with UAF.
This will be fixed more properly by looking at Stopped event's TRB
pointer when making skipping decisions, but such rework is unlikely
to be backported to v6.12, which will stay around for a few years. |
| In the Linux kernel, the following vulnerability has been resolved:
nvmem: core: fix cleanup after dev_set_name()
If dev_set_name() fails, we leak nvmem->wp_gpio as the cleanup does not
put this. While a minimal fix for this would be to add the gpiod_put()
call, we can do better if we split device_register(), and use the
tested nvmem_release() cleanup code by initialising the device early,
and putting the device.
This results in a slightly larger fix, but results in clear code.
Note: this patch depends on "nvmem: core: initialise nvmem->id early"
and "nvmem: core: remove nvmem_config wp_gpio".
[Srini: Fixed subject line and error code handing with wp_gpio while applying.] |
| In the Linux kernel, the following vulnerability has been resolved:
Squashfs: fix handling and sanity checking of xattr_ids count
A Sysbot [1] corrupted filesystem exposes two flaws in the handling and
sanity checking of the xattr_ids count in the filesystem. Both of these
flaws cause computation overflow due to incorrect typing.
In the corrupted filesystem the xattr_ids value is 4294967071, which
stored in a signed variable becomes the negative number -225.
Flaw 1 (64-bit systems only):
The signed integer xattr_ids variable causes sign extension.
This causes variable overflow in the SQUASHFS_XATTR_*(A) macros. The
variable is first multiplied by sizeof(struct squashfs_xattr_id) where the
type of the sizeof operator is "unsigned long".
On a 64-bit system this is 64-bits in size, and causes the negative number
to be sign extended and widened to 64-bits and then become unsigned. This
produces the very large number 18446744073709548016 or 2^64 - 3600. This
number when rounded up by SQUASHFS_METADATA_SIZE - 1 (8191 bytes) and
divided by SQUASHFS_METADATA_SIZE overflows and produces a length of 0
(stored in len).
Flaw 2 (32-bit systems only):
On a 32-bit system the integer variable is not widened by the unsigned
long type of the sizeof operator (32-bits), and the signedness of the
variable has no effect due it always being treated as unsigned.
The above corrupted xattr_ids value of 4294967071, when multiplied
overflows and produces the number 4294963696 or 2^32 - 3400. This number
when rounded up by SQUASHFS_METADATA_SIZE - 1 (8191 bytes) and divided by
SQUASHFS_METADATA_SIZE overflows again and produces a length of 0.
The effect of the 0 length computation:
In conjunction with the corrupted xattr_ids field, the filesystem also has
a corrupted xattr_table_start value, where it matches the end of
filesystem value of 850.
This causes the following sanity check code to fail because the
incorrectly computed len of 0 matches the incorrect size of the table
reported by the superblock (0 bytes).
len = SQUASHFS_XATTR_BLOCK_BYTES(*xattr_ids);
indexes = SQUASHFS_XATTR_BLOCKS(*xattr_ids);
/*
* The computed size of the index table (len bytes) should exactly
* match the table start and end points
*/
start = table_start + sizeof(*id_table);
end = msblk->bytes_used;
if (len != (end - start))
return ERR_PTR(-EINVAL);
Changing the xattr_ids variable to be "usigned int" fixes the flaw on a
64-bit system. This relies on the fact the computation is widened by the
unsigned long type of the sizeof operator.
Casting the variable to u64 in the above macro fixes this flaw on a 32-bit
system.
It also means 64-bit systems do not implicitly rely on the type of the
sizeof operator to widen the computation.
[1] https://lore.kernel.org/lkml/000000000000cd44f005f1a0f17f@google.com/ |
| In the Linux kernel, the following vulnerability has been resolved:
mm/MADV_COLLAPSE: catch !none !huge !bad pmd lookups
In commit 34488399fa08 ("mm/madvise: add file and shmem support to
MADV_COLLAPSE") we make the following change to find_pmd_or_thp_or_none():
- if (!pmd_present(pmde))
- return SCAN_PMD_NULL;
+ if (pmd_none(pmde))
+ return SCAN_PMD_NONE;
This was for-use by MADV_COLLAPSE file/shmem codepaths, where
MADV_COLLAPSE might identify a pte-mapped hugepage, only to have
khugepaged race-in, free the pte table, and clear the pmd. Such codepaths
include:
A) If we find a suitably-aligned compound page of order HPAGE_PMD_ORDER
already in the pagecache.
B) In retract_page_tables(), if we fail to grab mmap_lock for the target
mm/address.
In these cases, collapse_pte_mapped_thp() really does expect a none (not
just !present) pmd, and we want to suitably identify that case separate
from the case where no pmd is found, or it's a bad-pmd (of course, many
things could happen once we drop mmap_lock, and the pmd could plausibly
undergo multiple transitions due to intervening fault, split, etc).
Regardless, the code is prepared install a huge-pmd only when the existing
pmd entry is either a genuine pte-table-mapping-pmd, or the none-pmd.
However, the commit introduces a logical hole; namely, that we've allowed
!none- && !huge- && !bad-pmds to be classified as genuine
pte-table-mapping-pmds. One such example that could leak through are swap
entries. The pmd values aren't checked again before use in
pte_offset_map_lock(), which is expecting nothing less than a genuine
pte-table-mapping-pmd.
We want to put back the !pmd_present() check (below the pmd_none() check),
but need to be careful to deal with subtleties in pmd transitions and
treatments by various arch.
The issue is that __split_huge_pmd_locked() temporarily clears the present
bit (or otherwise marks the entry as invalid), but pmd_present() and
pmd_trans_huge() still need to return true while the pmd is in this
transitory state. For example, x86's pmd_present() also checks the
_PAGE_PSE , riscv's version also checks the _PAGE_LEAF bit, and arm64 also
checks a PMD_PRESENT_INVALID bit.
Covering all 4 cases for x86 (all checks done on the same pmd value):
1) pmd_present() && pmd_trans_huge()
All we actually know here is that the PSE bit is set. Either:
a) We aren't racing with __split_huge_page(), and PRESENT or PROTNONE
is set.
=> huge-pmd
b) We are currently racing with __split_huge_page(). The danger here
is that we proceed as-if we have a huge-pmd, but really we are
looking at a pte-mapping-pmd. So, what is the risk of this
danger?
The only relevant path is:
madvise_collapse() -> collapse_pte_mapped_thp()
Where we might just incorrectly report back "success", when really
the memory isn't pmd-backed. This is fine, since split could
happen immediately after (actually) successful madvise_collapse().
So, it should be safe to just assume huge-pmd here.
2) pmd_present() && !pmd_trans_huge()
Either:
a) PSE not set and either PRESENT or PROTNONE is.
=> pte-table-mapping pmd (or PROT_NONE)
b) devmap. This routine can be called immediately after
unlocking/locking mmap_lock -- or called with no locks held (see
khugepaged_scan_mm_slot()), so previous VMA checks have since been
invalidated.
3) !pmd_present() && pmd_trans_huge()
Not possible.
4) !pmd_present() && !pmd_trans_huge()
Neither PRESENT nor PROTNONE set
=> not present
I've checked all archs that implement pmd_trans_huge() (arm64, riscv,
powerpc, longarch, x86, mips, s390) and this logic roughly translates
(though devmap treatment is unique to x86 and powerpc, and (3) doesn't
necessarily hold in general -- but that doesn't matter since
!pmd_present() always takes failure path).
Also, add a comment above find_pmd_or_thp_or_none()
---truncated--- |
| In the Linux kernel, the following vulnerability has been resolved:
mm: multi-gen LRU: fix crash during cgroup migration
lru_gen_migrate_mm() assumes lru_gen_add_mm() runs prior to itself. This
isn't true for the following scenario:
CPU 1 CPU 2
clone()
cgroup_can_fork()
cgroup_procs_write()
cgroup_post_fork()
task_lock()
lru_gen_migrate_mm()
task_unlock()
task_lock()
lru_gen_add_mm()
task_unlock()
And when the above happens, kernel crashes because of linked list
corruption (mm_struct->lru_gen.list). |
| In the Linux kernel, the following vulnerability has been resolved:
can: isotp: split tx timer into transmission and timeout
The timer for the transmission of isotp PDUs formerly had two functions:
1. send two consecutive frames with a given time gap
2. monitor the timeouts for flow control frames and the echo frames
This led to larger txstate checks and potentially to a problem discovered
by syzbot which enabled the panic_on_warn feature while testing.
The former 'txtimer' function is split into 'txfrtimer' and 'txtimer'
to handle the two above functionalities with separate timer callbacks.
The two simplified timers now run in one-shot mode and make the state
transitions (especially with isotp_rcv_echo) better understandable. |
| In the Linux kernel, the following vulnerability has been resolved:
cgroup/cpuset: Fix wrong check in update_parent_subparts_cpumask()
It was found that the check to see if a partition could use up all
the cpus from the parent cpuset in update_parent_subparts_cpumask()
was incorrect. As a result, it is possible to leave parent with no
effective cpu left even if there are tasks in the parent cpuset. This
can lead to system panic as reported in [1].
Fix this probem by updating the check to fail the enabling the partition
if parent's effective_cpus is a subset of the child's cpus_allowed.
Also record the error code when an error happens in update_prstate()
and add a test case where parent partition and child have the same cpu
list and parent has task. Enabling partition in the child will fail in
this case.
[1] https://www.spinics.net/lists/cgroups/msg36254.html |
| In the Linux kernel, the following vulnerability has been resolved:
block: ublk: extending queue_size to fix overflow
When validating drafted SPDK ublk target, in a case that
assigning large queue depth to multiqueue ublk device,
ublk target would run into a weird incorrect state. During
rounds of review and debug, An overflow bug was found
in ublk driver.
In ublk_cmd.h, UBLK_MAX_QUEUE_DEPTH is 4096 which means
each ublk queue depth can be set as large as 4096. But
when setting qd for a ublk device,
sizeof(struct ublk_queue) + depth * sizeof(struct ublk_io)
will be larger than 65535 if qd is larger than 2728.
Then queue_size is overflowed, and ublk_get_queue()
references a wrong pointer position. The wrong content of
ublk_queue elements will lead to out-of-bounds memory
access.
Extend queue_size in ublk_device as "unsigned int". |
| In the Linux kernel, the following vulnerability has been resolved:
drm/i915: Fix request ref counting during error capture & debugfs dump
When GuC support was added to error capture, the reference counting
around the request object was broken. Fix it up.
The context based search manages the spinlocking around the search
internally. So it needs to grab the reference count internally as
well. The execlist only request based search relies on external
locking, so it needs an external reference count but within the
spinlock not outside it.
The only other caller of the context based search is the code for
dumping engine state to debugfs. That code wasn't previously getting
an explicit reference at all as it does everything while holding the
execlist specific spinlock. So, that needs updaing as well as that
spinlock doesn't help when using GuC submission. Rather than trying to
conditionally get/put depending on submission model, just change it to
always do the get/put.
v2: Explicitly document adding an extra blank line in some dense code
(Andy Shevchenko). Fix multiple potential null pointer derefs in case
of no request found (some spotted by Tvrtko, but there was more!).
Also fix a leaked request in case of !started and another in
__guc_reset_context now that intel_context_find_active_request is
actually reference counting the returned request.
v3: Add a _get suffix to intel_context_find_active_request now that it
grabs a reference (Daniele).
v4: Split the intel_guc_find_hung_context change to a separate patch
and rename intel_context_find_active_request_get to
intel_context_get_active_request (Tvrtko).
v5: s/locking/reference counting/ in commit message (Tvrtko)
(cherry picked from commit 3700e353781e27f1bc7222f51f2cc36cbeb9b4ec) |
| In the Linux kernel, the following vulnerability has been resolved:
fscache: Use wait_on_bit() to wait for the freeing of relinquished volume
The freeing of relinquished volume will wake up the pending volume
acquisition by using wake_up_bit(), however it is mismatched with
wait_var_event() used in fscache_wait_on_volume_collision() and it will
never wake up the waiter in the wait-queue because these two functions
operate on different wait-queues.
According to the implementation in fscache_wait_on_volume_collision(),
if the wake-up of pending acquisition is delayed longer than 20 seconds
(e.g., due to the delay of on-demand fd closing), the first
wait_var_event_timeout() will timeout and the following wait_var_event()
will hang forever as shown below:
FS-Cache: Potential volume collision new=00000024 old=00000022
......
INFO: task mount:1148 blocked for more than 122 seconds.
Not tainted 6.1.0-rc6+ #1
task:mount state:D stack:0 pid:1148 ppid:1
Call Trace:
<TASK>
__schedule+0x2f6/0xb80
schedule+0x67/0xe0
fscache_wait_on_volume_collision.cold+0x80/0x82
__fscache_acquire_volume+0x40d/0x4e0
erofs_fscache_register_volume+0x51/0xe0 [erofs]
erofs_fscache_register_fs+0x19c/0x240 [erofs]
erofs_fc_fill_super+0x746/0xaf0 [erofs]
vfs_get_super+0x7d/0x100
get_tree_nodev+0x16/0x20
erofs_fc_get_tree+0x20/0x30 [erofs]
vfs_get_tree+0x24/0xb0
path_mount+0x2fa/0xa90
do_mount+0x7c/0xa0
__x64_sys_mount+0x8b/0xe0
do_syscall_64+0x30/0x60
entry_SYSCALL_64_after_hwframe+0x46/0xb0
Considering that wake_up_bit() is more selective, so fix it by using
wait_on_bit() instead of wait_var_event() to wait for the freeing of
relinquished volume. In addition because waitqueue_active() is used in
wake_up_bit() and clear_bit() doesn't imply any memory barrier, use
clear_and_wake_up_bit() to add the missing memory barrier between
cursor->flags and waitqueue_active(). |
| In the Linux kernel, the following vulnerability has been resolved:
can: j1939: fix errant WARN_ON_ONCE in j1939_session_deactivate
The conclusion "j1939_session_deactivate() should be called with a
session ref-count of at least 2" is incorrect. In some concurrent
scenarios, j1939_session_deactivate can be called with the session
ref-count less than 2. But there is not any problem because it
will check the session active state before session putting in
j1939_session_deactivate_locked().
Here is the concurrent scenario of the problem reported by syzbot
and my reproduction log.
cpu0 cpu1
j1939_xtp_rx_eoma
j1939_xtp_rx_abort_one
j1939_session_get_by_addr [kref == 2]
j1939_session_get_by_addr [kref == 3]
j1939_session_deactivate [kref == 2]
j1939_session_put [kref == 1]
j1939_session_completed
j1939_session_deactivate
WARN_ON_ONCE(kref < 2)
=====================================================
WARNING: CPU: 1 PID: 21 at net/can/j1939/transport.c:1088 j1939_session_deactivate+0x5f/0x70
CPU: 1 PID: 21 Comm: ksoftirqd/1 Not tainted 5.14.0-rc7+ #32
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1 04/01/2014
RIP: 0010:j1939_session_deactivate+0x5f/0x70
Call Trace:
j1939_session_deactivate_activate_next+0x11/0x28
j1939_xtp_rx_eoma+0x12a/0x180
j1939_tp_recv+0x4a2/0x510
j1939_can_recv+0x226/0x380
can_rcv_filter+0xf8/0x220
can_receive+0x102/0x220
? process_backlog+0xf0/0x2c0
can_rcv+0x53/0xf0
__netif_receive_skb_one_core+0x67/0x90
? process_backlog+0x97/0x2c0
__netif_receive_skb+0x22/0x80 |
| In the Linux kernel, the following vulnerability has been resolved:
KVM: arm64: Unconditionally save+flush host FPSIMD/SVE/SME state
There are several problems with the way hyp code lazily saves the host's
FPSIMD/SVE state, including:
* Host SVE being discarded unexpectedly due to inconsistent
configuration of TIF_SVE and CPACR_ELx.ZEN. This has been seen to
result in QEMU crashes where SVE is used by memmove(), as reported by
Eric Auger:
https://issues.redhat.com/browse/RHEL-68997
* Host SVE state is discarded *after* modification by ptrace, which was an
unintentional ptrace ABI change introduced with lazy discarding of SVE state.
* The host FPMR value can be discarded when running a non-protected VM,
where FPMR support is not exposed to a VM, and that VM uses
FPSIMD/SVE. In these cases the hyp code does not save the host's FPMR
before unbinding the host's FPSIMD/SVE/SME state, leaving a stale
value in memory.
Avoid these by eagerly saving and "flushing" the host's FPSIMD/SVE/SME
state when loading a vCPU such that KVM does not need to save any of the
host's FPSIMD/SVE/SME state. For clarity, fpsimd_kvm_prepare() is
removed and the necessary call to fpsimd_save_and_flush_cpu_state() is
placed in kvm_arch_vcpu_load_fp(). As 'fpsimd_state' and 'fpmr_ptr'
should not be used, they are set to NULL; all uses of these will be
removed in subsequent patches.
Historical problems go back at least as far as v5.17, e.g. erroneous
assumptions about TIF_SVE being clear in commit:
8383741ab2e773a9 ("KVM: arm64: Get rid of host SVE tracking/saving")
... and so this eager save+flush probably needs to be backported to ALL
stable trees. |
