Hi,
When swapping (~2mb/sec swapin rate, ~80mb into swap, 95%cpu usage), IKD
semaphore lockup detection triggers on proc access. I'm using multiple
instances of xboard in two machine mode to initiate swapping. It seems
that any process accessing proc can sit there trying to grab a semaphore
long enough for IKD's timeout to trigger. No trouble when not under load.
-Mike
oops isn't really very useful, but....
ksymoops 2.3.3 on i686 2.3.99-pre6. Options used
-V (default)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.3.99-pre6/ (default)
-m /boot/2.3.99-pre6/System.map (specified)
invalid operand: 0000
CPU: 0
EIP: 0010:[<c0109672>]
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010282
eax: 0000001f ebx: c2ac4c1c ecx: ffffffff edx: c3259be0
esi: c2b79eec edi: c2ac4c24 ebp: c2b79f04 esp: c2b79ea0
ds: 0018 es: 0018 ss: 0018
Process w (pid: 595, stackpage=c2b79000)
Stack: c021908a c02190e8 0000006c c2ac4c1c c4146000 c2ac4c38 c2b78000 01234567
c2b78000 00000000 00000000 c2b79ee4 00000000 01234567 c2b78000 c2ac4c28
c2ac4c28 c2b79ee4 c01094e5 00000000 00000000 00026686 c2ac4c1c c01094c4
Call Trace: [<c0109930>] (0) [<c0152280>] (96) [<c01325dc>] (48) [<c010b155>] (40) [<ffffffff>]
[<c010e811>] (485) [<c010e891>] (485) [<c010c725>] (485) [<c011e814>] (485)
[<c011e978>] (485) [<c011ea6a>] (485) [<c012189e>] (485) [<c01cff3c>] (485)
[<c01cfd13>] (485) [<c01cfdd5>] (485) [<c0116383>] (485) [<c011ea6a>] (485)
[<c0121e4b>] (485) [<c0121c5a>] (485) [<c0121b9d>] (485) [<c0121d97>] (485)
[<c0121cac>] (485) [<c01094d1>] (485) [<c0116383>] (485) [<c0115f97>] (485)
[<c0109280>] (595) [<c0118fdb>] (595) [<c0211f7b>] (595) [<c0211d47>] (595)
[<c01d5a53>] (595) [<c01d2511>] (595) [<c0116383>] (595) [<c010b8ea>] (595)
[<c010b752>] (595) [<c01143e4>] (595)
Code: 0f 0b 83 c4 0c 56 e8 cb 81 01 00 8b 4d b4 83 c4 04 c7 01 22
>>EIP; c0109672 <__down+186/214> <=====
Trace; c0109930 <__down_failed+8/c>
Trace; c0152280 <proc_info_read+198/318>
Trace; c01325dc <sys_read+cc/f0>
Trace; c010b155 <system_call+3d/44>
Trace; ffffffff <BEGIN_EMERGENCY_TRACE> (how ikd makes semaphore oops:)
Trace; c010e811 <end_8259A_irq+d/28>
Trace; c010e891 <enable_8259A_irq+d/3c>
Trace; c010c725 <do_IRQ+a9/c4>
Trace; c011e814 <do_softirq+10/98>
Trace; c011e978 <tasklet_hi_action+10/78>
Trace; c011ea6a <bh_action+e/6c>
Trace; c012189e <tqueue_bh+e/60>
Trace; c01cff3c <batch_entropy_process+10/b8>
Trace; c01cfd13 <add_entropy_words+13/c8>
Trace; c01cfdd5 <credit_entropy_store+d/40>
Trace; c0116383 <__wake_up+13/1c0>
Trace; c011ea6a <bh_action+e/6c>
Trace; c0121e4b <timer_bh+13/2a4>
Trace; c0121c5a <update_wall_time+e/50>
Trace; c0121b9d <update_wall_time_one_tick+d/bc>
Trace; c0121d97 <update_process_times+f/70>
Trace; c0121cac <update_one_process+10/ec>
Trace; c01094d1 <generate_oops+d/28>
Trace; c0116383 <__wake_up+13/1c0>
Trace; c0115f97 <schedule+13/3ec>
Trace; c0109280 <__switch_to+10/c4>
Trace; c0118fdb <printk+13/16c>
Trace; c0211f7b <vsprintf+13/348>
Trace; c0211d47 <number+13/234>
Trace; c01d5a53 <vt_console_print+13/2dc>
Trace; c01d2511 <vc_cons_allocated+d/30>
Trace; c0116383 <__wake_up+13/1c0>
Trace; c010b8ea <do_invalid_op+e/4c>
Trace; c010b752 <die_if_no_fixup+e/48>
Trace; c01143e4 <search_exception_table+10/7c>
Code; c0109672 <__down+186/214>
00000000 <_EIP>:
Code; c0109672 <__down+186/214> <=====
0: 0f 0b ud2a <=====
Code; c0109674 <__down+188/214>
2: 83 c4 0c add $0xc,%esp
Code; c0109677 <__down+18b/214>
5: 56 push %esi
Code; c0109678 <__down+18c/214>
6: e8 cb 81 01 00 call 181d6 <_EIP+0x181d6> c0121848 <del_timer+0/48>
Code; c010967d <__down+191/214>
b: 8b 4d b4 mov 0xffffffb4(%ebp),%ecx
Code; c0109680 <__down+194/214>
e: 83 c4 04 add $0x4,%esp
Code; c0109683 <__down+197/214>
11: c7 01 22 00 00 00 movl $0x22,(%ecx)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Apr 23 2000 - 21:00:08 EST