| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| In the Linux kernel, the following vulnerability has been resolved:
virtio_net: Fix napi_skb_cache_put warning
After the commit bdacf3e34945 ("net: Use nested-BH locking for
napi_alloc_cache.") was merged, the following warning began to appear:
WARNING: CPU: 5 PID: 1 at net/core/skbuff.c:1451 napi_skb_cache_put+0x82/0x4b0
__warn+0x12f/0x340
napi_skb_cache_put+0x82/0x4b0
napi_skb_cache_put+0x82/0x4b0
report_bug+0x165/0x370
handle_bug+0x3d/0x80
exc_invalid_op+0x1a/0x50
asm_exc_invalid_op+0x1a/0x20
__free_old_xmit+0x1c8/0x510
napi_skb_cache_put+0x82/0x4b0
__free_old_xmit+0x1c8/0x510
__free_old_xmit+0x1c8/0x510
__pfx___free_old_xmit+0x10/0x10
The issue arises because virtio is assuming it's running in NAPI context
even when it's not, such as in the netpoll case.
To resolve this, modify virtnet_poll_tx() to only set NAPI when budget
is available. Same for virtnet_poll_cleantx(), which always assumed that
it was in a NAPI context. |
| In the Linux kernel, the following vulnerability has been resolved:
Revert "ALSA: firewire-lib: operate for period elapse event in process context"
Commit 7ba5ca32fe6e ("ALSA: firewire-lib: operate for period elapse event
in process context") removed the process context workqueue from
amdtp_domain_stream_pcm_pointer() and update_pcm_pointers() to remove
its overhead.
With RME Fireface 800, this lead to a regression since
Kernels 5.14.0, causing an AB/BA deadlock competition for the
substream lock with eventual system freeze under ALSA operation:
thread 0:
* (lock A) acquire substream lock by
snd_pcm_stream_lock_irq() in
snd_pcm_status64()
* (lock B) wait for tasklet to finish by calling
tasklet_unlock_spin_wait() in
tasklet_disable_in_atomic() in
ohci_flush_iso_completions() of ohci.c
thread 1:
* (lock B) enter tasklet
* (lock A) attempt to acquire substream lock,
waiting for it to be released:
snd_pcm_stream_lock_irqsave() in
snd_pcm_period_elapsed() in
update_pcm_pointers() in
process_ctx_payloads() in
process_rx_packets() of amdtp-stream.c
? tasklet_unlock_spin_wait
</NMI>
<TASK>
ohci_flush_iso_completions firewire_ohci
amdtp_domain_stream_pcm_pointer snd_firewire_lib
snd_pcm_update_hw_ptr0 snd_pcm
snd_pcm_status64 snd_pcm
? native_queued_spin_lock_slowpath
</NMI>
<IRQ>
_raw_spin_lock_irqsave
snd_pcm_period_elapsed snd_pcm
process_rx_packets snd_firewire_lib
irq_target_callback snd_firewire_lib
handle_it_packet firewire_ohci
context_tasklet firewire_ohci
Restore the process context work queue to prevent deadlock
AB/BA deadlock competition for ALSA substream lock of
snd_pcm_stream_lock_irq() in snd_pcm_status64()
and snd_pcm_stream_lock_irqsave() in snd_pcm_period_elapsed().
revert commit 7ba5ca32fe6e ("ALSA: firewire-lib: operate for period
elapse event in process context")
Replace inline description to prevent future deadlock. |
| In the Linux kernel, the following vulnerability has been resolved:
net/mlx5: Fix missing lock on sync reset reload
On sync reset reload work, when remote host updates devlink on reload
actions performed on that host, it misses taking devlink lock before
calling devlink_remote_reload_actions_performed() which results in
triggering lock assert like the following:
WARNING: CPU: 4 PID: 1164 at net/devlink/core.c:261 devl_assert_locked+0x3e/0x50
…
CPU: 4 PID: 1164 Comm: kworker/u96:6 Tainted: G S W 6.10.0-rc2+ #116
Hardware name: Supermicro SYS-2028TP-DECTR/X10DRT-PT, BIOS 2.0 12/18/2015
Workqueue: mlx5_fw_reset_events mlx5_sync_reset_reload_work [mlx5_core]
RIP: 0010:devl_assert_locked+0x3e/0x50
…
Call Trace:
<TASK>
? __warn+0xa4/0x210
? devl_assert_locked+0x3e/0x50
? report_bug+0x160/0x280
? handle_bug+0x3f/0x80
? exc_invalid_op+0x17/0x40
? asm_exc_invalid_op+0x1a/0x20
? devl_assert_locked+0x3e/0x50
devlink_notify+0x88/0x2b0
? mlx5_attach_device+0x20c/0x230 [mlx5_core]
? __pfx_devlink_notify+0x10/0x10
? process_one_work+0x4b6/0xbb0
process_one_work+0x4b6/0xbb0
[…] |
| In the Linux kernel, the following vulnerability has been resolved:
Revert "sched/fair: Make sure to try to detach at least one movable task"
This reverts commit b0defa7ae03ecf91b8bfd10ede430cff12fcbd06.
b0defa7ae03ec changed the load balancing logic to ignore env.max_loop if
all tasks examined to that point were pinned. The goal of the patch was
to make it more likely to be able to detach a task buried in a long list
of pinned tasks. However, this has the unfortunate side effect of
creating an O(n) iteration in detach_tasks(), as we now must fully
iterate every task on a cpu if all or most are pinned. Since this load
balance code is done with rq lock held, and often in softirq context, it
is very easy to trigger hard lockups. We observed such hard lockups with
a user who affined O(10k) threads to a single cpu.
When I discussed this with Vincent he initially suggested that we keep
the limit on the number of tasks to detach, but increase the number of
tasks we can search. However, after some back and forth on the mailing
list, he recommended we instead revert the original patch, as it seems
likely no one was actually getting hit by the original issue. |
| In the Linux kernel, the following vulnerability has been resolved:
i2c: pnx: Fix potential deadlock warning from del_timer_sync() call in isr
When del_timer_sync() is called in an interrupt context it throws a warning
because of potential deadlock. The timer is used only to exit from
wait_for_completion() after a timeout so replacing the call with
wait_for_completion_timeout() allows to remove the problematic timer and
its related functions altogether. |
| In the Linux kernel, the following vulnerability has been resolved:
riscv: kexec: Avoid deadlock in kexec crash path
If the kexec crash code is called in the interrupt context, the
machine_kexec_mask_interrupts() function will trigger a deadlock while
trying to acquire the irqdesc spinlock and then deactivate irqchip in
irq_set_irqchip_state() function.
Unlike arm64, riscv only requires irq_eoi handler to complete EOI and
keeping irq_set_irqchip_state() will only leave this possible deadlock
without any use. So we simply remove it. |
| In the Linux kernel, the following vulnerability has been resolved:
usb: dwc3: core: remove lock of otg mode during gadget suspend/resume to avoid deadlock
When config CONFIG_USB_DWC3_DUAL_ROLE is selected, and trigger system
to enter suspend status with below command:
echo mem > /sys/power/state
There will be a deadlock issue occurring. Detailed invoking path as
below:
dwc3_suspend_common()
spin_lock_irqsave(&dwc->lock, flags); <-- 1st
dwc3_gadget_suspend(dwc);
dwc3_gadget_soft_disconnect(dwc);
spin_lock_irqsave(&dwc->lock, flags); <-- 2nd
This issue is exposed by commit c7ebd8149ee5 ("usb: dwc3: gadget: Fix
NULL pointer dereference in dwc3_gadget_suspend") that removes the code
of checking whether dwc->gadget_driver is NULL or not. It causes the
following code is executed and deadlock occurs when trying to get the
spinlock. In fact, the root cause is the commit 5265397f9442("usb: dwc3:
Remove DWC3 locking during gadget suspend/resume") that forgot to remove
the lock of otg mode. So, remove the redundant lock of otg mode during
gadget suspend/resume. |
| In the Linux kernel, the following vulnerability has been resolved:
io_uring: fix possible deadlock in io_register_iowq_max_workers()
The io_register_iowq_max_workers() function calls io_put_sq_data(),
which acquires the sqd->lock without releasing the uring_lock.
Similar to the commit 009ad9f0c6ee ("io_uring: drop ctx->uring_lock
before acquiring sqd->lock"), this can lead to a potential deadlock
situation.
To resolve this issue, the uring_lock is released before calling
io_put_sq_data(), and then it is re-acquired after the function call.
This change ensures that the locks are acquired in the correct
order, preventing the possibility of a deadlock. |
| In the Linux kernel, the following vulnerability has been resolved:
cachefiles: Set object to close if ondemand_id < 0 in copen
If copen is maliciously called in the user mode, it may delete the request
corresponding to the random id. And the request may have not been read yet.
Note that when the object is set to reopen, the open request will be done
with the still reopen state in above case. As a result, the request
corresponding to this object is always skipped in select_req function, so
the read request is never completed and blocks other process.
Fix this issue by simply set object to close if its id < 0 in copen. |
| In the Linux kernel, the following vulnerability has been resolved:
wifi: mt76: mt7921s: fix potential hung tasks during chip recovery
During chip recovery (e.g. chip reset), there is a possible situation that
kernel worker reset_work is holding the lock and waiting for kernel thread
stat_worker to be parked, while stat_worker is waiting for the release of
the same lock.
It causes a deadlock resulting in the dumping of hung tasks messages and
possible rebooting of the device.
This patch prevents the execution of stat_worker during the chip recovery. |
| In the Linux kernel, the following vulnerability has been resolved:
riscv: rewrite __kernel_map_pages() to fix sleeping in invalid context
__kernel_map_pages() is a debug function which clears the valid bit in page
table entry for deallocated pages to detect illegal memory accesses to
freed pages.
This function set/clear the valid bit using __set_memory(). __set_memory()
acquires init_mm's semaphore, and this operation may sleep. This is
problematic, because __kernel_map_pages() can be called in atomic context,
and thus is illegal to sleep. An example warning that this causes:
BUG: sleeping function called from invalid context at kernel/locking/rwsem.c:1578
in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 2, name: kthreadd
preempt_count: 2, expected: 0
CPU: 0 PID: 2 Comm: kthreadd Not tainted 6.9.0-g1d4c6d784ef6 #37
Hardware name: riscv-virtio,qemu (DT)
Call Trace:
[<ffffffff800060dc>] dump_backtrace+0x1c/0x24
[<ffffffff8091ef6e>] show_stack+0x2c/0x38
[<ffffffff8092baf8>] dump_stack_lvl+0x5a/0x72
[<ffffffff8092bb24>] dump_stack+0x14/0x1c
[<ffffffff8003b7ac>] __might_resched+0x104/0x10e
[<ffffffff8003b7f4>] __might_sleep+0x3e/0x62
[<ffffffff8093276a>] down_write+0x20/0x72
[<ffffffff8000cf00>] __set_memory+0x82/0x2fa
[<ffffffff8000d324>] __kernel_map_pages+0x5a/0xd4
[<ffffffff80196cca>] __alloc_pages_bulk+0x3b2/0x43a
[<ffffffff8018ee82>] __vmalloc_node_range+0x196/0x6ba
[<ffffffff80011904>] copy_process+0x72c/0x17ec
[<ffffffff80012ab4>] kernel_clone+0x60/0x2fe
[<ffffffff80012f62>] kernel_thread+0x82/0xa0
[<ffffffff8003552c>] kthreadd+0x14a/0x1be
[<ffffffff809357de>] ret_from_fork+0xe/0x1c
Rewrite this function with apply_to_existing_page_range(). It is fine to
not have any locking, because __kernel_map_pages() works with pages being
allocated/deallocated and those pages are not changed by anyone else in the
meantime. |
| In the Linux kernel, the following vulnerability has been resolved:
USB: class: cdc-wdm: Fix CPU lockup caused by excessive log messages
The syzbot fuzzer found that the interrupt-URB completion callback in
the cdc-wdm driver was taking too long, and the driver's immediate
resubmission of interrupt URBs with -EPROTO status combined with the
dummy-hcd emulation to cause a CPU lockup:
cdc_wdm 1-1:1.0: nonzero urb status received: -71
cdc_wdm 1-1:1.0: wdm_int_callback - 0 bytes
watchdog: BUG: soft lockup - CPU#0 stuck for 26s! [syz-executor782:6625]
CPU#0 Utilization every 4s during lockup:
#1: 98% system, 0% softirq, 3% hardirq, 0% idle
#2: 98% system, 0% softirq, 3% hardirq, 0% idle
#3: 98% system, 0% softirq, 3% hardirq, 0% idle
#4: 98% system, 0% softirq, 3% hardirq, 0% idle
#5: 98% system, 1% softirq, 3% hardirq, 0% idle
Modules linked in:
irq event stamp: 73096
hardirqs last enabled at (73095): [<ffff80008037bc00>] console_emit_next_record kernel/printk/printk.c:2935 [inline]
hardirqs last enabled at (73095): [<ffff80008037bc00>] console_flush_all+0x650/0xb74 kernel/printk/printk.c:2994
hardirqs last disabled at (73096): [<ffff80008af10b00>] __el1_irq arch/arm64/kernel/entry-common.c:533 [inline]
hardirqs last disabled at (73096): [<ffff80008af10b00>] el1_interrupt+0x24/0x68 arch/arm64/kernel/entry-common.c:551
softirqs last enabled at (73048): [<ffff8000801ea530>] softirq_handle_end kernel/softirq.c:400 [inline]
softirqs last enabled at (73048): [<ffff8000801ea530>] handle_softirqs+0xa60/0xc34 kernel/softirq.c:582
softirqs last disabled at (73043): [<ffff800080020de8>] __do_softirq+0x14/0x20 kernel/softirq.c:588
CPU: 0 PID: 6625 Comm: syz-executor782 Tainted: G W 6.10.0-rc2-syzkaller-g8867bbd4a056 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 04/02/2024
Testing showed that the problem did not occur if the two error
messages -- the first two lines above -- were removed; apparently adding
material to the kernel log takes a surprisingly large amount of time.
In any case, the best approach for preventing these lockups and to
avoid spamming the log with thousands of error messages per second is
to ratelimit the two dev_err() calls. Therefore we replace them with
dev_err_ratelimited(). |
| In the Linux kernel, the following vulnerability has been resolved:
net: fec: remove .ndo_poll_controller to avoid deadlocks
There is a deadlock issue found in sungem driver, please refer to the
commit ac0a230f719b ("eth: sungem: remove .ndo_poll_controller to avoid
deadlocks"). The root cause of the issue is that netpoll is in atomic
context and disable_irq() is called by .ndo_poll_controller interface
of sungem driver, however, disable_irq() might sleep. After analyzing
the implementation of fec_poll_controller(), the fec driver should have
the same issue. Due to the fec driver uses NAPI for TX completions, the
.ndo_poll_controller is unnecessary to be implemented in the fec driver,
so fec_poll_controller() can be safely removed. |
| In the Linux kernel, the following vulnerability has been resolved:
tty: xilinx_uartps: split sysrq handling
lockdep detects the following circular locking dependency:
CPU 0 CPU 1
========================== ============================
cdns_uart_isr() printk()
uart_port_lock(port) console_lock()
cdns_uart_console_write()
if (!port->sysrq)
uart_port_lock(port)
uart_handle_break()
port->sysrq = ...
uart_handle_sysrq_char()
printk()
console_lock()
The fixed commit attempts to avoid this situation by only taking the
port lock in cdns_uart_console_write if port->sysrq unset. However, if
(as shown above) cdns_uart_console_write runs before port->sysrq is set,
then it will try to take the port lock anyway. This may result in a
deadlock.
Fix this by splitting sysrq handling into two parts. We use the prepare
helper under the port lock and defer handling until we release the lock. |
| In the Linux kernel, the following vulnerability has been resolved:
irqchip/gic-v3-its: Don't enable interrupts in its_irq_set_vcpu_affinity()
The following call-chain leads to enabling interrupts in a nested interrupt
disabled section:
irq_set_vcpu_affinity()
irq_get_desc_lock()
raw_spin_lock_irqsave() <--- Disable interrupts
its_irq_set_vcpu_affinity()
guard(raw_spinlock_irq) <--- Enables interrupts when leaving the guard()
irq_put_desc_unlock() <--- Warns because interrupts are enabled
This was broken in commit b97e8a2f7130, which replaced the original
raw_spin_[un]lock() pair with guard(raw_spinlock_irq).
Fix the issue by using guard(raw_spinlock).
[ tglx: Massaged change log ] |
| In the Linux kernel, the following vulnerability has been resolved:
virtio-blk: don't keep queue frozen during system suspend
Commit 4ce6e2db00de ("virtio-blk: Ensure no requests in virtqueues before
deleting vqs.") replaces queue quiesce with queue freeze in virtio-blk's
PM callbacks. And the motivation is to drain inflight IOs before suspending.
block layer's queue freeze looks very handy, but it is also easy to cause
deadlock, such as, any attempt to call into bio_queue_enter() may run into
deadlock if the queue is frozen in current context. There are all kinds
of ->suspend() called in suspend context, so keeping queue frozen in the
whole suspend context isn't one good idea. And Marek reported lockdep
warning[1] caused by virtio-blk's freeze queue in virtblk_freeze().
[1] https://lore.kernel.org/linux-block/ca16370e-d646-4eee-b9cc-87277c89c43c@samsung.com/
Given the motivation is to drain in-flight IOs, it can be done by calling
freeze & unfreeze, meantime restore to previous behavior by keeping queue
quiesced during suspend. |
| In the Linux kernel, the following vulnerability has been resolved:
net: restrict SO_REUSEPORT to inet sockets
After blamed commit, crypto sockets could accidentally be destroyed
from RCU call back, as spotted by zyzbot [1].
Trying to acquire a mutex in RCU callback is not allowed.
Restrict SO_REUSEPORT socket option to inet sockets.
v1 of this patch supported TCP, UDP and SCTP sockets,
but fcnal-test.sh test needed RAW and ICMP support.
[1]
BUG: sleeping function called from invalid context at kernel/locking/mutex.c:562
in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 24, name: ksoftirqd/1
preempt_count: 100, expected: 0
RCU nest depth: 0, expected: 0
1 lock held by ksoftirqd/1/24:
#0: ffffffff8e937ba0 (rcu_callback){....}-{0:0}, at: rcu_lock_acquire include/linux/rcupdate.h:337 [inline]
#0: ffffffff8e937ba0 (rcu_callback){....}-{0:0}, at: rcu_do_batch kernel/rcu/tree.c:2561 [inline]
#0: ffffffff8e937ba0 (rcu_callback){....}-{0:0}, at: rcu_core+0xa37/0x17a0 kernel/rcu/tree.c:2823
Preemption disabled at:
[<ffffffff8161c8c8>] softirq_handle_begin kernel/softirq.c:402 [inline]
[<ffffffff8161c8c8>] handle_softirqs+0x128/0x9b0 kernel/softirq.c:537
CPU: 1 UID: 0 PID: 24 Comm: ksoftirqd/1 Not tainted 6.13.0-rc3-syzkaller-00174-ga024e377efed #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/13/2024
Call Trace:
<TASK>
__dump_stack lib/dump_stack.c:94 [inline]
dump_stack_lvl+0x241/0x360 lib/dump_stack.c:120
__might_resched+0x5d4/0x780 kernel/sched/core.c:8758
__mutex_lock_common kernel/locking/mutex.c:562 [inline]
__mutex_lock+0x131/0xee0 kernel/locking/mutex.c:735
crypto_put_default_null_skcipher+0x18/0x70 crypto/crypto_null.c:179
aead_release+0x3d/0x50 crypto/algif_aead.c:489
alg_do_release crypto/af_alg.c:118 [inline]
alg_sock_destruct+0x86/0xc0 crypto/af_alg.c:502
__sk_destruct+0x58/0x5f0 net/core/sock.c:2260
rcu_do_batch kernel/rcu/tree.c:2567 [inline]
rcu_core+0xaaa/0x17a0 kernel/rcu/tree.c:2823
handle_softirqs+0x2d4/0x9b0 kernel/softirq.c:561
run_ksoftirqd+0xca/0x130 kernel/softirq.c:950
smpboot_thread_fn+0x544/0xa30 kernel/smpboot.c:164
kthread+0x2f0/0x390 kernel/kthread.c:389
ret_from_fork+0x4b/0x80 arch/x86/kernel/process.c:147
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
</TASK> |
| In the Linux kernel, the following vulnerability has been resolved:
pinctrl: mcp23s08: Fix sleeping in atomic context due to regmap locking
If a device uses MCP23xxx IO expander to receive IRQs, the following
bug can happen:
BUG: sleeping function called from invalid context
at kernel/locking/mutex.c:283
in_atomic(): 1, irqs_disabled(): 1, non_block: 0, ...
preempt_count: 1, expected: 0
...
Call Trace:
...
__might_resched+0x104/0x10e
__might_sleep+0x3e/0x62
mutex_lock+0x20/0x4c
regmap_lock_mutex+0x10/0x18
regmap_update_bits_base+0x2c/0x66
mcp23s08_irq_set_type+0x1ae/0x1d6
__irq_set_trigger+0x56/0x172
__setup_irq+0x1e6/0x646
request_threaded_irq+0xb6/0x160
...
We observed the problem while experimenting with a touchscreen driver which
used MCP23017 IO expander (I2C).
The regmap in the pinctrl-mcp23s08 driver uses a mutex for protection from
concurrent accesses, which is the default for regmaps without .fast_io,
.disable_locking, etc.
mcp23s08_irq_set_type() calls regmap_update_bits_base(), and the latter
locks the mutex.
However, __setup_irq() locks desc->lock spinlock before calling these
functions. As a result, the system tries to lock the mutex whole holding
the spinlock.
It seems, the internal regmap locks are not needed in this driver at all.
mcp->lock seems to protect the regmap from concurrent accesses already,
except, probably, in mcp_pinconf_get/set.
mcp23s08_irq_set_type() and mcp23s08_irq_mask/unmask() are called under
chip_bus_lock(), which calls mcp23s08_irq_bus_lock(). The latter takes
mcp->lock and enables regmap caching, so that the potentially slow I2C
accesses are deferred until chip_bus_unlock().
The accesses to the regmap from mcp23s08_probe_one() do not need additional
locking.
In all remaining places where the regmap is accessed, except
mcp_pinconf_get/set(), the driver already takes mcp->lock.
This patch adds locking in mcp_pinconf_get/set() and disables internal
locking in the regmap config. Among other things, it fixes the sleeping
in atomic context described above. |
| In the Linux kernel, the following vulnerability has been resolved:
powerpc/pseries: Fix dtl_access_lock to be a rw_semaphore
The dtl_access_lock needs to be a rw_sempahore, a sleeping lock, because
the code calls kmalloc() while holding it, which can sleep:
# echo 1 > /proc/powerpc/vcpudispatch_stats
BUG: sleeping function called from invalid context at include/linux/sched/mm.h:337
in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 199, name: sh
preempt_count: 1, expected: 0
3 locks held by sh/199:
#0: c00000000a0743f8 (sb_writers#3){.+.+}-{0:0}, at: vfs_write+0x324/0x438
#1: c0000000028c7058 (dtl_enable_mutex){+.+.}-{3:3}, at: vcpudispatch_stats_write+0xd4/0x5f4
#2: c0000000028c70b8 (dtl_access_lock){+.+.}-{2:2}, at: vcpudispatch_stats_write+0x220/0x5f4
CPU: 0 PID: 199 Comm: sh Not tainted 6.10.0-rc4 #152
Hardware name: IBM pSeries (emulated by qemu) POWER9 (raw) 0x4e1202 0xf000005 of:SLOF,HEAD hv:linux,kvm pSeries
Call Trace:
dump_stack_lvl+0x130/0x148 (unreliable)
__might_resched+0x174/0x410
kmem_cache_alloc_noprof+0x340/0x3d0
alloc_dtl_buffers+0x124/0x1ac
vcpudispatch_stats_write+0x2a8/0x5f4
proc_reg_write+0xf4/0x150
vfs_write+0xfc/0x438
ksys_write+0x88/0x148
system_call_exception+0x1c4/0x5a0
system_call_common+0xf4/0x258 |
| In the Linux kernel, the following vulnerability has been resolved:
usb: musb: Fix hardware lockup on first Rx endpoint request
There is a possibility that a request's callback could be invoked from
usb_ep_queue() (call trace below, supplemented with missing calls):
req->complete from usb_gadget_giveback_request
(drivers/usb/gadget/udc/core.c:999)
usb_gadget_giveback_request from musb_g_giveback
(drivers/usb/musb/musb_gadget.c:147)
musb_g_giveback from rxstate
(drivers/usb/musb/musb_gadget.c:784)
rxstate from musb_ep_restart
(drivers/usb/musb/musb_gadget.c:1169)
musb_ep_restart from musb_ep_restart_resume_work
(drivers/usb/musb/musb_gadget.c:1176)
musb_ep_restart_resume_work from musb_queue_resume_work
(drivers/usb/musb/musb_core.c:2279)
musb_queue_resume_work from musb_gadget_queue
(drivers/usb/musb/musb_gadget.c:1241)
musb_gadget_queue from usb_ep_queue
(drivers/usb/gadget/udc/core.c:300)
According to the docstring of usb_ep_queue(), this should not happen:
"Note that @req's ->complete() callback must never be called from within
usb_ep_queue() as that can create deadlock situations."
In fact, a hardware lockup might occur in the following sequence:
1. The gadget is initialized using musb_gadget_enable().
2. Meanwhile, a packet arrives, and the RXPKTRDY flag is set, raising an
interrupt.
3. If IRQs are enabled, the interrupt is handled, but musb_g_rx() finds an
empty queue (next_request() returns NULL). The interrupt flag has
already been cleared by the glue layer handler, but the RXPKTRDY flag
remains set.
4. The first request is enqueued using usb_ep_queue(), leading to the call
of req->complete(), as shown in the call trace above.
5. If the callback enables IRQs and another packet is waiting, step (3)
repeats. The request queue is empty because usb_g_giveback() removes the
request before invoking the callback.
6. The endpoint remains locked up, as the interrupt triggered by hardware
setting the RXPKTRDY flag has been handled, but the flag itself remains
set.
For this scenario to occur, it is only necessary for IRQs to be enabled at
some point during the complete callback. This happens with the USB Ethernet
gadget, whose rx_complete() callback calls netif_rx(). If called in the
task context, netif_rx() disables the bottom halves (BHs). When the BHs are
re-enabled, IRQs are also enabled to allow soft IRQs to be processed. The
gadget itself is initialized at module load (or at boot if built-in), but
the first request is enqueued when the network interface is brought up,
triggering rx_complete() in the task context via ioctl(). If a packet
arrives while the interface is down, it can prevent the interface from
receiving any further packets from the USB host.
The situation is quite complicated with many parties involved. This
particular issue can be resolved in several possible ways:
1. Ensure that callbacks never enable IRQs. This would be difficult to
enforce, as discovering how netif_rx() interacts with interrupts was
already quite challenging and u_ether is not the only function driver.
Similar "bugs" could be hidden in other drivers as well.
2. Disable MUSB interrupts in musb_g_giveback() before calling the callback
and re-enable them afterwars (by calling musb_{dis,en}able_interrupts(),
for example). This would ensure that MUSB interrupts are not handled
during the callback, even if IRQs are enabled. In fact, it would allow
IRQs to be enabled when releasing the lock. However, this feels like an
inelegant hack.
3. Modify the interrupt handler to clear the RXPKTRDY flag if the request
queue is empty. While this approach also feels like a hack, it wastes
CPU time by attempting to handle incoming packets when the software is
not ready to process them.
4. Flush the Rx FIFO instead of calling rxstate() in musb_ep_restart().
This ensures that the hardware can receive packets when there is at
least one request in the queue. Once I
---truncated--- |
