On Wed, 9 Jan 2008, Jiri Slaby wrote:
BUG: soft lockup - CPU#1 stuck for 11s! [X:2887][ ... ]Call Trace:
[<ffffffff804e7648>] __mutex_lock_slowpath+0x38/0xd0
[<ffffffff804e747e>] mutex_lock+0x1e/0x30
[<ffffffff8040c497>] input_release_device+0x27/0x50
[<ffffffff804102aa>] evdev_ungrab+0x3a/0x50
[<ffffffff804103eb>] evdev_release+0xcb/0xd0
[<ffffffff80293460>] __fput+0xc0/0x230
[<ffffffff802938a6>] fput+0x16/0x20
[<ffffffff80290296>] filp_close+0x56/0x90
[<ffffffff80291afa>] sys_close+0x9a/0xf0
[<ffffffff8020ba4e>] system_call+0x7e/0x83
I was not able to use sysrq keys, the keyboard is obviously defunct. If this is not known, I'll try to turn lockdep on and maybe connect through ssh next time it happens.
Is this reproducible on your side?
Running with lockdep would be really helpful, so that we possibly know who is holding the other instance of dev->mutex.