[PATCH] usbip: fix use-after-free in event_handler()

[syzbot]

From: Aleksandr Nogikh <nog...@google.com>

The event_handler function in drivers/usb/usbip/usbip_event.c is a
workqueue item responsible for processing events for a struct usbip_device.
During device teardown, usbip_stop_eh() is called to wait for the event
handler to finish processing. However, usbip_stop_eh() uses
wait_event_interruptible() and ignores its return value. If the process
unbinding the driver receives a signal, wait_event_interruptible() returns
immediately, causing the teardown process to falsely assume the event
handler has finished. The teardown process then proceeds to free the
usbip_device memory. Meanwhile, the event_handler workqueue is still
running and attempts to access the freed usbip_device, resulting in a KASAN
slab-use-after-free crash.

BUG: KASAN: slab-use-after-free in __mutex_lock_common
kernel/locking/rtmutex_api.c:559 [inline]
BUG: KASAN: slab-use-after-free in mutex_lock_nested+0x5a/0x1d0
kernel/locking/rtmutex_api.c:578
Read of size 1 at addr ffff8881145245b0 by task kworker/u8:5/6177

Call Trace:
<TASK>
lock_acquire+0x84/0x350 kernel/locking/lockdep.c:5842
__mutex_lock_common kernel/locking/rtmutex_api.c:559 [inline]
mutex_lock_nested+0x5a/0x1d0 kernel/locking/rtmutex_api.c:578
event_handler+0x1e3/0x4a0 drivers/usb/usbip/usbip_event.c:73
process_one_work kernel/workqueue.c:3322 [inline]
process_scheduled_works+0xa8e/0x14e0 kernel/workqueue.c:3405
worker_thread+0xa47/0xfb0 kernel/workqueue.c:3486
kthread+0x388/0x470 kernel/kthread.c:436
ret_from_fork+0x514/0xb70 arch/x86/kernel/process.c:158
ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:245
</TASK>

To fix this, change wait_event_interruptible() to wait_event() in
usbip_stop_eh(). This ensures that the teardown process strictly waits for
the event handler to finish its execution and drop all references to the
usbip_device before the memory is freed, preventing the use-after-free.
This change is safe from deadlocks because usbip_stop_eh() is never called
with locks held that the event_handler would need to acquire.

Fixes: bb7871ad99ea ("usbip: event handler as one thread")
Assisted-by: Gemini:gemini-3.5-flash Gemini:gemini-3.1-pro-preview syzbot
Reported-by: syzbot+f4518a...@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=f4518ac0a3dd3aafde38
Link: https://syzkaller.appspot.com/ai_job?id=88101c8e-4da9-45de-b879-0549d6fe5649
Signed-off-by: Aleksandr Nogikh <nog...@google.com>

---
diff --git a/drivers/usb/usbip/usbip_event.c b/drivers/usb/usbip/usbip_event.c
index 0e00c2d00..81d761891 100644
--- a/drivers/usb/usbip/usbip_event.c
+++ b/drivers/usb/usbip/usbip_event.c
@@ -115,7 +115,7 @@ void usbip_stop_eh(struct usbip_device *ud)
if (pending)
usbip_dbg_eh("usbip_eh waiting completion %lx\n", pending);

- wait_event_interruptible(ud->eh_waitq, !(ud->event & ~USBIP_EH_BYE));
+ wait_event(ud->eh_waitq, !(ud->event & ~USBIP_EH_BYE));
usbip_dbg_eh("usbip_eh has stopped\n");
}
EXPORT_SYMBOL_GPL(usbip_stop_eh);


base-commit: 075b74841bd0065a3bda3440873c747938e69b68
--
See https://goo.gle/syzbot-ai-patches for information about AI-generated patches.
You can comment on the patch as usual, syzbot will try to address
the comments and send a new version of the patch if necessary.
syzbot engineers can be reached at syzk...@googlegroups.com.

[Shuah Khan]

On 8/3/26 09:05, syzbot wrote:
> From: Aleksandr Nogikh <nog...@google.com>
>
> The event_handler function in drivers/usb/usbip/usbip_event.c is a
> workqueue item responsible for processing events for a struct usbip_device.
> During device teardown, usbip_stop_eh() is called to wait for the event
> handler to finish processing. However, usbip_stop_eh() uses
> wait_event_interruptible() and ignores its return value. If the process
> unbinding the driver receives a signal, wait_event_interruptible() returns
> immediately, causing the teardown process to falsely assume the event
> handler has finished.

Does this mean unbinding didn't happen?

Correct - usbip_stop_eh() is called without lock hold after updating
the shutdown_busid status to true. However I am curious what happens
to the unbinding? Should usbip_stop_eh() check the return value of
wait_event_interruptible() and handle the error instead?

thanks,
-- Shuah

[Aleksandr Nogikh]

Hi Shuah,

Thanks for reviewing the patch!

On Tue, Aug 4, 2026 at 7:31 PM Shuah Khan <sk...@linuxfoundation.org> wrote:
>
> On 8/3/26 09:05, syzbot wrote:
> > From: Aleksandr Nogikh <nog...@google.com>
> >
> > The event_handler function in drivers/usb/usbip/usbip_event.c is a
> > workqueue item responsible for processing events for a struct usbip_device.
> > During device teardown, usbip_stop_eh() is called to wait for the event
> > handler to finish processing. However, usbip_stop_eh() uses
> > wait_event_interruptible() and ignores its return value. If the process
> > unbinding the driver receives a signal, wait_event_interruptible() returns
> > immediately, causing the teardown process to falsely assume the event
> > handler has finished.
>
> Does this mean unbinding didn't happen?

The unbinding did happen (and finished freeing the memory), but
without actually waiting for event_handler() in
drivers/usb/usbip/usbip_event.c to finish processing the removal
event. As the device memory was prematurely freed, we got the
use-after-free crash in event_handler().

For reference, here's a C reproducer for the original bug:
https://syzkaller.appspot.com/text?tag=ReproC&x=16ffb7b9580000
It sets up a timer to deliver a signal and wakes up the
wait_event_interruptible() call.

During unbinding, we must wait for the scheduled removal event to
finish before we can safely free the device structures. From what I
see in the code, once we have reached wait_event_interruptible(),
there's no way to abort the process or somehow gracefully handle the
error.

--
Aleksandr

>
> thanks,
> -- Shuah

[Shuah Khan]

On 8/7/26 08:18, Aleksandr Nogikh wrote:
> Hi Shuah,
>
> Thanks for reviewing the patch!
>
> On Tue, Aug 4, 2026 at 7:31 PM Shuah Khan <sk...@linuxfoundation.org> wrote:
>>
>> On 8/3/26 09:05, syzbot wrote:
>>> From: Aleksandr Nogikh <nog...@google.com>
>>>
>>> The event_handler function in drivers/usb/usbip/usbip_event.c is a
>>> workqueue item responsible for processing events for a struct usbip_device.
>>> During device teardown, usbip_stop_eh() is called to wait for the event
>>> handler to finish processing. However, usbip_stop_eh() uses
>>> wait_event_interruptible() and ignores its return value. If the process
>>> unbinding the driver receives a signal, wait_event_interruptible() returns
>>> immediately, causing the teardown process to falsely assume the event
>>> handler has finished.
>>
>> Does this mean unbinding didn't happen?
>
> The unbinding did happen (and finished freeing the memory), but
> without actually waiting for event_handler() in
> drivers/usb/usbip/usbip_event.c to finish processing the removal
> event. As the device memory was prematurely freed, we got the
> use-after-free crash in event_handler().
>
> For reference, here's a C reproducer for the original bug:
> https://syzkaller.appspot.com/text?tag=ReproC&x=16ffb7b9580000
> It sets up a timer to deliver a signal and wakes up the
> wait_event_interruptible() call.

Thanks.

Yes I agree with you on rewinding being hard.

Care to explain the scope of this assist?

Assisted-by: Gemini:gemini-3.5-flash Gemini:gemini-3.1-pro-preview syzbot

thanks,
-- Shuah

[Aleksandr Nogikh]

On Mon, Aug 10, 2026 at 9:33 PM Shuah Khan <sk...@linuxfoundation.org> wrote:
>
< ... >

> >
> > During unbinding, we must wait for the scheduled removal event to
> > finish before we can safely free the device structures. From what I
> > see in the code, once we have reached wait_event_interruptible(),
> > there's no way to abort the process or somehow gracefully handle the
> > error.
> >
>
> Yes I agree with you on rewinding being hard.
>
> Care to explain the scope of this assist?
>
> Assisted-by: Gemini:gemini-3.5-flash Gemini:gemini-3.1-pro-preview syzbot

It was generated by the syzbot's AI patch generation agent:
https://github.com/google/syzkaller/blob/master/docs/syzbot_ai_patches.md

We manually review the patches it suggests and upstream those we
believe are fully correct.