| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| In the Linux kernel, the following vulnerability has been resolved:
net/rds: Fix circular locking dependency in rds_tcp_tune
syzbot reported a circular locking dependency in rds_tcp_tune() where
sk_net_refcnt_upgrade() is called while holding the socket lock:
======================================================
WARNING: possible circular locking dependency detected
======================================================
kworker/u10:8/15040 is trying to acquire lock:
ffffffff8e9aaf80 (fs_reclaim){+.+.}-{0:0},
at: __kmalloc_cache_noprof+0x4b/0x6f0
but task is already holding lock:
ffff88805a3c1ce0 (k-sk_lock-AF_INET6){+.+.}-{0:0},
at: rds_tcp_tune+0xd7/0x930
The issue occurs because sk_net_refcnt_upgrade() performs memory
allocation (via get_net_track() -> ref_tracker_alloc()) while the
socket lock is held, creating a circular dependency with fs_reclaim.
Fix this by moving sk_net_refcnt_upgrade() outside the socket lock
critical section. This is safe because the fields modified by the
sk_net_refcnt_upgrade() call (sk_net_refcnt, ns_tracker) are not
accessed by any concurrent code path at this point.
v2:
- Corrected fixes tag
- check patch line wrap nits
- ai commentary nits |
| In the Linux kernel, the following vulnerability has been resolved:
exfat: add cluster chain loop check for dir
An infinite loop may occur if the following conditions occur due to
file system corruption.
(1) Condition for exfat_count_dir_entries() to loop infinitely.
- The cluster chain includes a loop.
- There is no UNUSED entry in the cluster chain.
(2) Condition for exfat_create_upcase_table() to loop infinitely.
- The cluster chain of the root directory includes a loop.
- There are no UNUSED entry and up-case table entry in the cluster
chain of the root directory.
(3) Condition for exfat_load_bitmap() to loop infinitely.
- The cluster chain of the root directory includes a loop.
- There are no UNUSED entry and bitmap entry in the cluster chain
of the root directory.
(4) Condition for exfat_find_dir_entry() to loop infinitely.
- The cluster chain includes a loop.
- The unused directory entries were exhausted by some operation.
(5) Condition for exfat_check_dir_empty() to loop infinitely.
- The cluster chain includes a loop.
- The unused directory entries were exhausted by some operation.
- All files and sub-directories under the directory are deleted.
This commit adds checks to break the above infinite loop. |
| In the Linux kernel, the following vulnerability has been resolved:
workqueue: Do not warn when cancelling WQ_MEM_RECLAIM work from !WQ_MEM_RECLAIM worker
After commit
746ae46c1113 ("drm/sched: Mark scheduler work queues with WQ_MEM_RECLAIM")
amdgpu started seeing the following warning:
[ ] workqueue: WQ_MEM_RECLAIM sdma0:drm_sched_run_job_work [gpu_sched] is flushing !WQ_MEM_RECLAIM events:amdgpu_device_delay_enable_gfx_off [amdgpu]
...
[ ] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched]
...
[ ] Call Trace:
[ ] <TASK>
...
[ ] ? check_flush_dependency+0xf5/0x110
...
[ ] cancel_delayed_work_sync+0x6e/0x80
[ ] amdgpu_gfx_off_ctrl+0xab/0x140 [amdgpu]
[ ] amdgpu_ring_alloc+0x40/0x50 [amdgpu]
[ ] amdgpu_ib_schedule+0xf4/0x810 [amdgpu]
[ ] ? drm_sched_run_job_work+0x22c/0x430 [gpu_sched]
[ ] amdgpu_job_run+0xaa/0x1f0 [amdgpu]
[ ] drm_sched_run_job_work+0x257/0x430 [gpu_sched]
[ ] process_one_work+0x217/0x720
...
[ ] </TASK>
The intent of the verifcation done in check_flush_depedency is to ensure
forward progress during memory reclaim, by flagging cases when either a
memory reclaim process, or a memory reclaim work item is flushed from a
context not marked as memory reclaim safe.
This is correct when flushing, but when called from the
cancel(_delayed)_work_sync() paths it is a false positive because work is
either already running, or will not be running at all. Therefore
cancelling it is safe and we can relax the warning criteria by letting the
helper know of the calling context.
References: 746ae46c1113 ("drm/sched: Mark scheduler work queues with WQ_MEM_RECLAIM") |
| In the Linux kernel, the following vulnerability has been resolved:
bpf: Prevent tailcall infinite loop caused by freplace
There is a potential infinite loop issue that can occur when using a
combination of tail calls and freplace.
In an upcoming selftest, the attach target for entry_freplace of
tailcall_freplace.c is subprog_tc of tc_bpf2bpf.c, while the tail call in
entry_freplace leads to entry_tc. This results in an infinite loop:
entry_tc -> subprog_tc -> entry_freplace --tailcall-> entry_tc.
The problem arises because the tail_call_cnt in entry_freplace resets to
zero each time entry_freplace is executed, causing the tail call mechanism
to never terminate, eventually leading to a kernel panic.
To fix this issue, the solution is twofold:
1. Prevent updating a program extended by an freplace program to a
prog_array map.
2. Prevent extending a program that is already part of a prog_array map
with an freplace program.
This ensures that:
* If a program or its subprogram has been extended by an freplace program,
it can no longer be updated to a prog_array map.
* If a program has been added to a prog_array map, neither it nor its
subprograms can be extended by an freplace program.
Moreover, an extension program should not be tailcalled. As such, return
-EINVAL if the program has a type of BPF_PROG_TYPE_EXT when adding it to a
prog_array map.
Additionally, fix a minor code style issue by replacing eight spaces with a
tab for proper formatting. |
| In the Linux kernel, the following vulnerability has been resolved:
bpf: Defer work in bpf_timer_cancel_and_free
Currently, the same case as previous patch (two timer callbacks trying
to cancel each other) can be invoked through bpf_map_update_elem as
well, or more precisely, freeing map elements containing timers. Since
this relies on hrtimer_cancel as well, it is prone to the same deadlock
situation as the previous patch.
It would be sufficient to use hrtimer_try_to_cancel to fix this problem,
as the timer cannot be enqueued after async_cancel_and_free. Once
async_cancel_and_free has been done, the timer must be reinitialized
before it can be armed again. The callback running in parallel trying to
arm the timer will fail, and freeing bpf_hrtimer without waiting is
sufficient (given kfree_rcu), and bpf_timer_cb will return
HRTIMER_NORESTART, preventing the timer from being rearmed again.
However, there exists a UAF scenario where the callback arms the timer
before entering this function, such that if cancellation fails (due to
timer callback invoking this routine, or the target timer callback
running concurrently). In such a case, if the timer expiration is
significantly far in the future, the RCU grace period expiration
happening before it will free the bpf_hrtimer state and along with it
the struct hrtimer, that is enqueued.
Hence, it is clear cancellation needs to occur after
async_cancel_and_free, and yet it cannot be done inline due to deadlock
issues. We thus modify bpf_timer_cancel_and_free to defer work to the
global workqueue, adding a work_struct alongside rcu_head (both used at
_different_ points of time, so can share space).
Update existing code comments to reflect the new state of affairs. |
| In the Linux kernel, the following vulnerability has been resolved:
net/mlx5: Reload only IB representors upon lag disable/enable
On lag disable, the bond IB device along with all of its
representors are destroyed, and then the slaves' representors get reloaded.
In case the slave IB representor load fails, the eswitch error flow
unloads all representors, including ethernet representors, where the
netdevs get detached and removed from lag bond. Such flow is inaccurate
as the lag driver is not responsible for loading/unloading ethernet
representors. Furthermore, the flow described above begins by holding
lag lock to prevent bond changes during disable flow. However, when
reaching the ethernet representors detachment from lag, the lag lock is
required again, triggering the following deadlock:
Call trace:
__switch_to+0xf4/0x148
__schedule+0x2c8/0x7d0
schedule+0x50/0xe0
schedule_preempt_disabled+0x18/0x28
__mutex_lock.isra.13+0x2b8/0x570
__mutex_lock_slowpath+0x1c/0x28
mutex_lock+0x4c/0x68
mlx5_lag_remove_netdev+0x3c/0x1a0 [mlx5_core]
mlx5e_uplink_rep_disable+0x70/0xa0 [mlx5_core]
mlx5e_detach_netdev+0x6c/0xb0 [mlx5_core]
mlx5e_netdev_change_profile+0x44/0x138 [mlx5_core]
mlx5e_netdev_attach_nic_profile+0x28/0x38 [mlx5_core]
mlx5e_vport_rep_unload+0x184/0x1b8 [mlx5_core]
mlx5_esw_offloads_rep_load+0xd8/0xe0 [mlx5_core]
mlx5_eswitch_reload_reps+0x74/0xd0 [mlx5_core]
mlx5_disable_lag+0x130/0x138 [mlx5_core]
mlx5_lag_disable_change+0x6c/0x70 [mlx5_core] // hold ldev->lock
mlx5_devlink_eswitch_mode_set+0xc0/0x410 [mlx5_core]
devlink_nl_cmd_eswitch_set_doit+0xdc/0x180
genl_family_rcv_msg_doit.isra.17+0xe8/0x138
genl_rcv_msg+0xe4/0x220
netlink_rcv_skb+0x44/0x108
genl_rcv+0x40/0x58
netlink_unicast+0x198/0x268
netlink_sendmsg+0x1d4/0x418
sock_sendmsg+0x54/0x60
__sys_sendto+0xf4/0x120
__arm64_sys_sendto+0x30/0x40
el0_svc_common+0x8c/0x120
do_el0_svc+0x30/0xa0
el0_svc+0x20/0x30
el0_sync_handler+0x90/0xb8
el0_sync+0x160/0x180
Thus, upon lag enable/disable, load and unload only the IB representors
of the slaves preventing the deadlock mentioned above.
While at it, refactor the mlx5_esw_offloads_rep_load() function to have
a static helper method for its internal logic, in symmetry with the
representor unload design. |
| In the Linux kernel, the following vulnerability has been resolved:
Bluetooth: af_bluetooth: Fix deadlock
Attemting to do sock_lock on .recvmsg may cause a deadlock as shown
bellow, so instead of using sock_sock this uses sk_receive_queue.lock
on bt_sock_ioctl to avoid the UAF:
INFO: task kworker/u9:1:121 blocked for more than 30 seconds.
Not tainted 6.7.6-lemon #183
Workqueue: hci0 hci_rx_work
Call Trace:
<TASK>
__schedule+0x37d/0xa00
schedule+0x32/0xe0
__lock_sock+0x68/0xa0
? __pfx_autoremove_wake_function+0x10/0x10
lock_sock_nested+0x43/0x50
l2cap_sock_recv_cb+0x21/0xa0
l2cap_recv_frame+0x55b/0x30a0
? psi_task_switch+0xeb/0x270
? finish_task_switch.isra.0+0x93/0x2a0
hci_rx_work+0x33a/0x3f0
process_one_work+0x13a/0x2f0
worker_thread+0x2f0/0x410
? __pfx_worker_thread+0x10/0x10
kthread+0xe0/0x110
? __pfx_kthread+0x10/0x10
ret_from_fork+0x2c/0x50
? __pfx_kthread+0x10/0x10
ret_from_fork_asm+0x1b/0x30
</TASK> |
| In the Linux kernel, the following vulnerability has been resolved:
scsi: hisi_sas: Fix a deadlock issue related to automatic dump
If we issue a disabling PHY command, the device attached with it will go
offline, if a 2 bit ECC error occurs at the same time, a hung task may be
found:
[ 4613.652388] INFO: task kworker/u256:0:165233 blocked for more than 120 seconds.
[ 4613.666297] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 4613.674809] task:kworker/u256:0 state:D stack: 0 pid:165233 ppid: 2 flags:0x00000208
[ 4613.683959] Workqueue: 0000:74:02.0_disco_q sas_revalidate_domain [libsas]
[ 4613.691518] Call trace:
[ 4613.694678] __switch_to+0xf8/0x17c
[ 4613.698872] __schedule+0x660/0xee0
[ 4613.703063] schedule+0xac/0x240
[ 4613.706994] schedule_timeout+0x500/0x610
[ 4613.711705] __down+0x128/0x36c
[ 4613.715548] down+0x240/0x2d0
[ 4613.719221] hisi_sas_internal_abort_timeout+0x1bc/0x260 [hisi_sas_main]
[ 4613.726618] sas_execute_internal_abort+0x144/0x310 [libsas]
[ 4613.732976] sas_execute_internal_abort_dev+0x44/0x60 [libsas]
[ 4613.739504] hisi_sas_internal_task_abort_dev.isra.0+0xbc/0x1b0 [hisi_sas_main]
[ 4613.747499] hisi_sas_dev_gone+0x174/0x250 [hisi_sas_main]
[ 4613.753682] sas_notify_lldd_dev_gone+0xec/0x2e0 [libsas]
[ 4613.759781] sas_unregister_common_dev+0x4c/0x7a0 [libsas]
[ 4613.765962] sas_destruct_devices+0xb8/0x120 [libsas]
[ 4613.771709] sas_do_revalidate_domain.constprop.0+0x1b8/0x31c [libsas]
[ 4613.778930] sas_revalidate_domain+0x60/0xa4 [libsas]
[ 4613.784716] process_one_work+0x248/0x950
[ 4613.789424] worker_thread+0x318/0x934
[ 4613.793878] kthread+0x190/0x200
[ 4613.797810] ret_from_fork+0x10/0x18
[ 4613.802121] INFO: task kworker/u256:4:316722 blocked for more than 120 seconds.
[ 4613.816026] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 4613.824538] task:kworker/u256:4 state:D stack: 0 pid:316722 ppid: 2 flags:0x00000208
[ 4613.833670] Workqueue: 0000:74:02.0 hisi_sas_rst_work_handler [hisi_sas_main]
[ 4613.841491] Call trace:
[ 4613.844647] __switch_to+0xf8/0x17c
[ 4613.848852] __schedule+0x660/0xee0
[ 4613.853052] schedule+0xac/0x240
[ 4613.856984] schedule_timeout+0x500/0x610
[ 4613.861695] __down+0x128/0x36c
[ 4613.865542] down+0x240/0x2d0
[ 4613.869216] hisi_sas_controller_prereset+0x58/0x1fc [hisi_sas_main]
[ 4613.876324] hisi_sas_rst_work_handler+0x40/0x8c [hisi_sas_main]
[ 4613.883019] process_one_work+0x248/0x950
[ 4613.887732] worker_thread+0x318/0x934
[ 4613.892204] kthread+0x190/0x200
[ 4613.896118] ret_from_fork+0x10/0x18
[ 4613.900423] INFO: task kworker/u256:1:348985 blocked for more than 121 seconds.
[ 4613.914341] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 4613.922852] task:kworker/u256:1 state:D stack: 0 pid:348985 ppid: 2 flags:0x00000208
[ 4613.931984] Workqueue: 0000:74:02.0_event_q sas_port_event_worker [libsas]
[ 4613.939549] Call trace:
[ 4613.942702] __switch_to+0xf8/0x17c
[ 4613.946892] __schedule+0x660/0xee0
[ 4613.951083] schedule+0xac/0x240
[ 4613.955015] schedule_timeout+0x500/0x610
[ 4613.959725] wait_for_common+0x200/0x610
[ 4613.964349] wait_for_completion+0x3c/0x5c
[ 4613.969146] flush_workqueue+0x198/0x790
[ 4613.973776] sas_porte_broadcast_rcvd+0x1e8/0x320 [libsas]
[ 4613.979960] sas_port_event_worker+0x54/0xa0 [libsas]
[ 4613.985708] process_one_work+0x248/0x950
[ 4613.990420] worker_thread+0x318/0x934
[ 4613.994868] kthread+0x190/0x200
[ 4613.998800] ret_from_fork+0x10/0x18
This is because when the device goes offline, we obtain the hisi_hba
semaphore and send the ABORT_DEV command to the device. However, the
internal abort timed out due to the 2 bit ECC error and triggers automatic
dump. In addition, since the hisi_hba semaphore has been obtained, the dump
cannot be executed and the controller cannot be reset.
Therefore, the deadlocks occur on the following circular dependencies
---truncated--- |
| In the Linux kernel, the following vulnerability has been resolved:
net/sched: act_mirred: use the backlog for mirred ingress
The test Davide added in commit ca22da2fbd69 ("act_mirred: use the backlog
for nested calls to mirred ingress") hangs our testing VMs every 10 or so
runs, with the familiar tcp_v4_rcv -> tcp_v4_rcv deadlock reported by
lockdep.
The problem as previously described by Davide (see Link) is that
if we reverse flow of traffic with the redirect (egress -> ingress)
we may reach the same socket which generated the packet. And we may
still be holding its socket lock. The common solution to such deadlocks
is to put the packet in the Rx backlog, rather than run the Rx path
inline. Do that for all egress -> ingress reversals, not just once
we started to nest mirred calls.
In the past there was a concern that the backlog indirection will
lead to loss of error reporting / less accurate stats. But the current
workaround does not seem to address the issue. |
| MariaDB Connector/J is used to connect applications developed in Java to MariaDB and MySQL databases. From 3.0.0 until 3.3.7, 3.4.5, and 3.5.11, the connector assumed result bytes were UTF-8 but guarded only changes to character_set_client. It did not reject non-UTF-8 character_set_results changes and relied on the server's global session_track_system_variables configuration to report character-set changes. A malicious or compromised server, or a machine-in-the-middle attacker, could suppress tracking or change the result charset so application strings differed from the returned bytes, potentially defeating application-side validation or escaping that depended on the decoded values. This issue is fixed in versions 3.3.7, 3.4.5, and 3.5.11. |
| In the Linux kernel, the following vulnerability has been resolved:
drm/amdgpu: cap ATOM command table nesting depth
Cap nesting at 32 levels with execute_depth and
return -ELOOP when exceeded. |
| In the Linux kernel, the following vulnerability has been resolved:
x86/mm: Drop unnecessary PMD page copy when freeing
On a box with a discrete GPU, lockdep reports a possible deadlock as
soon as kswapd shrinks the TTM page pool. The immediate cause is an
x86 commit that added an mmap_read_lock() to kernel page protection
munging code.
The huge vmap code holds the same lock over a GFP_KERNEL allocation,
which is a no-no now that reclaim can take it. That allocation is in a
page table *free* path and ends up being for dubious purposes[1].
Basically, it tries to avoid hardware setting Accessed=1 in page table
entries that are unreachable by the hardware, a non-issue.
Remove the PMD copy. Detach the original PMD page at the PUD, flush
the mid-level caches, and free the PTE tables straight from the
detached PMD page. With no allocation left, the locking issue is gone.
Lockdep splat/analysis:
WARNING: possible circular locking dependency detected
7.3.0-rc3-f6e7b42bf05b+ #183 Tainted: G U
------------------------------------------------------
kswapd0/269 is trying to acquire lock:
((init_mm).mmap_lock){++++}-{4:4}, at: change_page_attr_set_clr+0x29a/0x4a0
but task is already holding lock:
(pool_shrink_rwsem){.+.+}-{4:4}, at: ttm_pool_shrink+0xb2/0x330 [ttm]
Chain exists of:
(init_mm).mmap_lock --> fs_reclaim --> pool_shrink_rwsem
The cycle is built from three edges:
1) pool_shrink_rwsem -> (init_mm).mmap_lock
The TTM shrinker restores the caching attribute of every page it
frees, while holding pool_shrink_rwsem:
ttm_pool_shrink()
-> ttm_pool_dispose_list()
-> ttm_pool_free_page()
-> set_pages_wb()
-> change_page_attr_set_clr() [ init_mm mmap read lock ]
2) fs_reclaim -> pool_shrink_rwsem
The same shrinker, called from reclaim.
3) (init_mm).mmap_lock -> fs_reclaim
ioremap() installing a huge PUD mapping over an existing PMD table:
ioremap_page_range()
-> vmap_range_noflush()
-> vmap_try_huge_pud() [ init_mm mmap read lock ]
-> pud_free_pmd_page()
-> __get_free_page(GFP_KERNEL) [ enters reclaim ]
[ dhansen: Lots of changelog munging/trimming and merged comments from my
version of the fix. ] |
| By permanently omitting the final control-response line, a malicious FTP,
IMAP, POP3, or SMTP server can exploit libcurl and completely bypass its
configured transfer timeout mechanism, indefinitely blocking all concurrent
transfers bound to the same multi-handle. |
| In the Linux kernel, the following vulnerability has been resolved:
scsi: scsi_transport_fc: Widen FPIN pname walker counter to u32
An adjacent Fibre Channel fabric actor that can deliver an FPIN ELS
frame to an lpfc or qla2xxx Linux initiator can trigger a non-return in
the generic FC transport. This is not a local userspace or IP network
path; the attacker must be able to inject fabric traffic, for example as
a compromised switch or fabric controller, or as a same-zone N_Port on a
fabric that permits source spoofing.
The Link-Integrity and Peer-Congestion FPIN walkers used a u8 loop
counter against the 32-bit on-wire pname_count field, and did not bound
pname_count by the descriptor body already validated by the TLV walker.
A pname_count of 256 therefore wraps the counter and keeps the loop
condition true indefinitely.
Factor the shared pname_list[] walk into one helper, widen the counter
to u32, and clamp pname_count against the entries that fit in the
descriptor body before iterating. |
| In the Linux kernel, the following vulnerability has been resolved:
net: stmmac: dwmac1000: clear the RGMII status interrupt
Since commit 2e2c878a3141 ("net: stmmac: remove SGMII/RGMII/SMII
interrupt handling") nothing on the RGMII path reads GMAC_RGSMIIIS, so
RGSMIIIS stays pending behind GMAC_INT_MASK from the first link change
on. stmmac_resume() resets the MAC in stmmac_hw_setup(); the reset sets
GMAC_INT_MASK to 0 until dwmac1000_core_init() writes the mask again,
and the pending status storms the line:
dwmac_dma_interrupt: unexpected status 04000000
A register dump at the first message on ODROID-HC4 showed it:
core int_status 00000001 int_mask 00000000 pmt 00000000 rgsmiiis 0000000d
When the resume thread runs on the CPU that takes the MAC interrupt,
the storm starves it: a soft lockup, then a watchdog reset or a hang.
Read GMAC_RGSMIIIS in dwmac1000_irq_status() when the status is not
masked, as the driver did before 2e2c878a3141; the read clears it.
Wake-on-LAN resume from s2idle with the resume thread and the MAC
interrupt on one CPU: Helios64 (dwmac-rk), 7.3-rc6, soft-locked in 3 of
3 resumes without a fix and in 0 of 10 with this patch; it ran with an
out-of-tree dwmac-rk change that keeps the GMAC powered across suspend.
ODROID-HC4 (dwmac-meson8b): 0 of 10 with this patch, 2 of 2 without a
fix on a 7.3-rc5 build. irq_rgmii_n went up by one per resume. |
| P-Net before 1.0.3 enters an infinite loop when handling a read request for the PDPortDataCheck record with subslot 0 (whole module). The loop over the ports never advanced to the next port, so the request never completes and the device stops responding. A remote unauthenticated attacker can use this to cause a denial of service. |
| In the Linux kernel, the following vulnerability has been resolved:
scsi: target: Disable interrupts while holding delayed_cmd_lock
target_do_delayed_work() takes dev->delayed_cmd_lock with a plain
spin_lock(). It runs from system_percpu_wq, so interrupts stay enabled
while the lock is held. Every other acquisition of that lock uses
spin_lock_irqsave(): target_handle_task_attr() and
transport_complete_ordered_sync() in target_core_transport.c, and
target_non_ordered_release() in target_core_device.c.
That was safe while every caller of transport_complete_task_attr() ran
in process context: target_complete_ok_work() from target_completion_wq,
transport_complete_qf() from dev->qf_work_queue, and
transport_generic_request_failure() from the submit and execute paths.
Since commit 06933066d88a ("scsi: target: Add support for completing
commands from backend context") that is no longer true. With
complete_type=1 on a fabric that sets direct_compl_supp,
target_complete() runs target_complete_ok_work() inline in the backend's
completion context, which target_complete_cmd_with_sense() documents as
"May be called from interrupt context". For iblock that is the bio
end_io.
target_complete_ok_work() starts with transport_complete_task_attr(),
which takes delayed_cmd_lock in two ways:
- Every ORDERED command is put on dev->delayed_cmd_list by
target_handle_task_attr() and dispatched from there by
target_do_delayed_work(), which marks it SCF_TASK_ORDERED_SYNC. Its
completion therefore goes through transport_complete_ordered_sync().
- A SIMPLE command that completes while an ORDERED command is pending
drops the last reference to the killed dev->non_ordered, and
percpu_ref_put() then runs target_non_ordered_release() in the
completing context.
Both take the lock and, if commands are waiting, call
schedule_work(&dev->delayed_cmd_work), which queues on system_percpu_wq,
i.e. on the CPU that took the completion. Further completions for the
same queue are typically delivered there too, so one arriving while the
work holds the lock spins on it forever.
Reaching this needs complete_type=1, which is not the default
(vhost-scsi, the only fabric with direct_compl_supp, defaults to
TARGET_QUEUE_COMPL), a backend whose I/O completes from interrupt
context, and ORDERED commands from the initiator. Linux virtio_scsi
guests send only VIRTIO_SCSI_S_SIMPLE and never take the ordered path;
any guest that uses the ORDERED task attribute, which vhost-scsi
accepts, does.
target_do_delayed_work() is a work item and always runs in process
context, so spin_lock_irq() is sufficient.
Found by inspection; no runtime report. A single ORDERED command goes
through both acquisitions, so with CONFIG_PROVE_LOCKING the first one
completed through direct completion should produce an inconsistent lock
state report on delayed_cmd_lock (hardirq or softirq, depending on where
the backend completes). |
| Twisted is an event-based framework for internet applications, supporting Python 3.6+. Prior to 26.10.0rc1, twisted.mail.imap4.IMAP4Server reaches twisted.mail.imap4.parseNestedParens with a negative literal size in an authenticated SEARCH command after the client enters the selected mailbox state. The negative size moves the parser cursor below zero, and Python's negative indexing prevents the synchronous scanning loop from terminating. The loop freezes the single-threaded reactor, stopping other IMAP connections and any SMTP or POP3 services hosted in the same process until restart. This issue is fixed in version 26.10.0rc1. |
| Kafka Streams derives its stream-time clock from record timestamps. PunctuationSchedule.next() computed the next punctuation time as timestamp + (intervalsMissed + 1) * interval with no overflow check. A record with a timestamp at or near Long.MAX_VALUE advances stream time to that value, so the next stream-time punctuation overflows to a negative time. PunctuationQueue then re-queues that time and loops on it indefinitely while holding a lock, stalling the StreamThread.
Impact:
A single record is enough. The affected StreamThread stops processing, and because the record stays in the topic the hang returns on restart; an application whose partitions are all affected becomes unresponsive. Only the Streams application is affected, not the brokers or the cluster. Triggering it needs write access to a source topic whose topology uses stream-time punctuation, and a broker that stores the record: brokers from Apache Kafka 4.0.0 reject such a timestamp by default (log.message.timestamp.after.max.ms, one hour), so it is reachable only on older brokers or where that bound has been relaxed.
Mitigation:
Upgrade Kafka Streams to 4.2.2, 4.3.2, or 4.4.0. As defense in depth, keep log.message.timestamp.after.max.ms bounded on source topics and restrict write access to them.
Resolution:
From Apache Kafka 4.2.2, 4.3.2, and 4.4.0: the next punctuation time is computed with overflow-checked arithmetic (clamped to Long.MAX_VALUE), and a rescheduled punctuation is re-queued only when its time moves forward. |
| In the Linux kernel, the following vulnerability has been resolved:
fs: avoid repeated scans in evict_inodes()
We observed hung tasks when users attempted to unmount a filesystem
after its disk had been removed while still in use. During device
removal, fs_bdev_mark_dead() calls evict_inodes() while holding s_umount.
Each time evict_inodes() drops s_inode_list_lock to reschedule, it
restarts the walk from the head of s_inodes. With many referenced inodes
at the head of the list, these restarts repeatedly scan the same inodes
without reclaiming them. This can keep s_umount held for a long time,
blocking concurrent umount attempts and triggering hung-task reports.
Keep the current inode, already marked I_FREEING, out of the disposal
batch until s_inode_list_lock is reacquired. Resume the walk from this
inode and dispose of it in a later batch or at the end of the walk.
The zero-refcount and state checks under i_lock allow this walker to
claim the inode by setting I_FREEING and removing it from the LRU.
Other reclaimers skip the inode, leaving this walker responsible for
eviction. Only evict() removes it from s_inodes, so keeping it out of
the disposal batch ensures that it remains on the list while the lock
is dropped. After reacquiring the lock, reading its current next pointer
accounts for concurrent removal of following inodes.
The existing inode lifetime rules prohibit acquiring a reference to an
inode marked I_FREEING or I_WILL_FREE. __iget() requires its caller to
hold i_lock and establish that taking a reference is valid. Inode lookup
and igrab() check these flags under i_lock when acquiring a reference
from zero. ihold() requires an existing reference, which would keep
i_count nonzero and prevent this walker from claiming the inode. These
rules already allow iput_final() and the inode shrinker to release
i_lock after setting I_FREEING and before eviction completes.
A temporary __iget() reference would also keep the inode on the list,
but its release must preserve last-reference handling. Another user can
acquire a reference, update lazy timestamps and drop its reference while
the pin is held. If the pin becomes the last reference, dropping it with
atomic_dec_and_test() and evicting directly bypasses iput()'s lazytime
handling and can lose those timestamp updates.
Releasing the pin with iput() preserves that handling, but does not
guarantee eviction. fs_bdev_mark_dead() runs with SB_ACTIVE set, so iput()
may retain the inode in cache, whereas evict_inodes() must evict eligible
zero-reference inodes. The inode may also have been freed when iput()
returns, so the walker cannot then use it to force eviction. Using
I_FREEING preserves the existing eviction behavior without introducing
an additional last-reference transition.
The xfstests auto group passed on ext4 and XFS with known unrelated
failures excluded. No new issues were observed, and the previously
reproducible hung task no longer occurs with this patch. |
