Hello,
As I originally reported on March 9th in the "2.6.3 userspace freeze"
thread, I'm having some problems with 2.6.3 and 2.6.4 locking up and
being rebooted after 60 seconds by the software watchdog.
I have since then been experimenting with various kernel options to
narrow this down, and I recently tried enabling CONFIG_DEBUG_SPINLOCK,
CONFIG_DEBUG_SPINLOCK_SLEEP and the NMI watchdog, which so far has given
me the following kernel output, when the lockup happens:
1st time (full output attached as 2.6.4-lockup1.log):
Debug: sleeping function called from invalid context at mm/slab.c:1900
in_atomic():1, irqs_disabled():0
Call Trace:
[<c011ce06>] __might_sleep+0xa6/0xb0
[<c0144ac5>] kmem_cache_alloc+0x25/0x70
[<c01b74c9>] nfs_sync_file+0x79/0x90
[<c016adae>] getname+0x1e/0xa0
[<c016c65e>] __user_walk+0xe/0x40
[<c015b1a7>] sys_chdir+0x17/0x70
[<c015c78f>] sys_write+0x2f/0x50
[<c0117410>] do_page_fault+0x0/0x537
[<c0109344>] do_notify_resume+0x34/0x48
[<c010953f>] syscall_call+0x7/0xb
2nd time (2.6.4-lockup2.log):
Debug: sleeping function called from invalid context at mm/slab.c:1900
in_atomic():1, irqs_disabled():0
Call Trace:
[<c011ce06>] __might_sleep+0xa6/0xb0
[<c0144ac5>] kmem_cache_alloc+0x25/0x70
[<c0119caa>] scheduler_tick+0x4a/0x620
[<c01239cf>] sys_setitimer+0x9f/0x100
[<c016adae>] getname+0x1e/0xa0
[<c016c65e>] __user_walk+0xe/0x40
[<c015b1a7>] sys_chdir+0x17/0x70
[<c015c78f>] sys_write+0x2f/0x50
[<c0117410>] do_page_fault+0x0/0x537
[<c0109344>] do_notify_resume+0x34/0x48
[<c010953f>] syscall_call+0x7/0xb
3rd time (2.6.4-lockup3.log):
Debug: sleeping function called from invalid context at
include/asm/semaphore.h:119
in_atomic():1, irqs_disabled():0
Call Trace:
[<c011ce06>] __might_sleep+0xa6/0xb0
[<c013efbc>] generic_file_aio_write+0x5c/0xa0
[<c0195f5a>] ext3_file_write+0x2a/0xb0
[<c015c5fe>] do_sync_write+0x7e/0xb0
[<c01b552e>] nfs_scan_commit+0x2e/0x70
[<c01b7372>] nfs_commit_file+0x52/0x130
[<c01b74c9>] nfs_sync_file+0x79/0x90
[<c015c6da>] vfs_write+0xaa/0xe0
[<c0110fb1>] sys_ipc+0x51/0x2c0
[<c015c78f>] sys_write+0x2f/0x50
[<c010953f>] syscall_call+0x7/0xb
And a 4th (2.6.4-lockup4.log), which was almost identical to the 3rd.
.config and dmesg is attached as 2.6.4-config and 2.6.4-dmesg.
Any suggestions on how to fix this would be much appreciated.
--
Med venlig hilsen - Best regards
Anders K. Pedersen
Network Engineer
------------------------------------------------
Cohaesio A/S - Maglebjergvej 5D - DK-2800 Lyngby
Phone: +45 45 880 888 - Fax: +45 45 880 777
Mail: [email protected] - http://www.cohaesio.com
------------------------------------------------