On Thu, Mar 20, 2025 at 12:39:24PM +0800, 白烁冉 wrote: > Dear Maintainers, > > When using our customized Syzkaller to fuzz the latest Linux kernel, the following crash (94th)was triggered. > > > HEAD commit: 6537cfb395f352782918d8ee7b7f10ba2cc3cbf2 > git tree: upstream > Output:https://github.com/pghk13/Kernel-Bug/tree/main/0305_6.14rc5/94-INFO_%20rcu%20detected%20stall%20in%20dcache_dir_open > Kernel config:https://github.com/pghk13/Kernel-Bug/blob/main/0305_6.14rc5/config.txt > C reproducer:https://github.com/pghk13/Kernel-Bug/blob/main/0305_6.14rc5/94-INFO_%20rcu%20detected%20stall%20in%20dcache_dir_open/94repro.c > Syzlang reproducer: https://github.com/pghk13/Kernel-Bug/blob/main/0305_6.14rc5/94-INFO_%20rcu%20detected%20stall%20in%20dcache_dir_open/94report > > > The error occurs around line 379 of the urb.c file. The problem ends up in the cm109_urb_irq_callback function in the cm109.c file:In the cm109_urb_irq_callback function, the driver attempts to resubmit a URB that has not yet been processed. There may be a race condition in the driver that resubmits the URB in the URB completion callback, but the same URB may have already been committed to another location in the system. This issue seems to involve the creation of USB devices, the operation of TTY devices, and file descriptor copying. This complex interaction resulted in duplicate commits of the URB. > We have reproduced this issue several times on 6.14-rc5 again. Great! Can you submit a fix for this as you have a reproducer you can use to prove that it resolves the issue? thanks, greg k-h
