Re: Linux 2.6.23-rc5

From: Gabriel C
Date: Mon Sep 03 2007 - 14:34:40 EST


Rafael J. Wysocki wrote:
> On Sunday, 2 September 2007 09:54, Prakash Punnoor wrote:
>> Hi,
>>
>> 2.6.23-rc5 locks up hard (Magic Syskeys won't even work) after a few minutes
>> of work on x86_64. 2.6.23-rc4 was fine. I'll try git-bisect to find out what
>> is causing trouble. Yes, I am using nvidia binary but it didn't make troubles
>> since ages... When I found the bugger, I'll try whether it works w/o nvidia
>> binary.
>
> Just for the record, I'm running -rc5 on two x86_64 boxes without any visible
> issue.
>
> I don't use any binary graphics drivers, though.

2.6.23-rc5 dies here random[1] too .. on 2 Dell boxes ( i686 not x86_64 here ) which I upgraded to rc5 yesterday.

When dies 'Caps Lock and Scroll Lock' blinks and the box lockups hard .. rc4 got an 7 days uptime on both with the same config ,
kernel compiled with same gcc/glibc/binutils etc.

I don't have any Oops/BUG or whatever messages in any of my logs nor I'm able to capture everything with netconsole :/
Any idea how else I could capture the Oops/BUG messages ?

Ohh well all I have is this : ( captured with a script which I started some seconds before a box died , no idea it has something to
do with this bug but I post anyway )

..

[ 5414.309269] SysRq : Show Pending Timers
[ 5414.309359] Timer List Version: v0.3
[ 5414.309363] HRTIMER_MAX_CLOCK_BASES: 2
[ 5414.309366] now at 5391544457987 nsecs
[ 5414.309371]
[ 5414.309372] cpu: 0
[ 5414.309377] clock 0:
[ 5414.309390] .index: 0
[ 5414.309394] .resolution: 1 nsecs
[ 5414.309399] .get_time: ktime_get_real
[ 5414.309429] .offset: 1188833557343399512 nsecs
[ 5414.309435] active timers:
[ 5414.309443] clock 1:
[ 5414.309448] .index: 1
[ 5414.309454] .resolution: 1 nsecs
[ 5414.309458] .get_time: ktime_get
[ 5414.309473] .offset: 0 nsecs
[ 5414.309477] active timers:
[ 5414.309482] #0: <3>BUG: scheduling while atomic: swapper/0x10010000/0
[ 5414.309500] [<c0317120>] __sched_text_start+0x5b0/0x6c0
[ 5414.309530] [<c0115f3c>] __cond_resched+0x18/0x29
[ 5414.309548] [<c031729a>] cond_resched+0x2a/0x31
[ 5414.309562] [<c0317ad4>] mutex_lock+0xb/0x1c
[ 5414.309573] [<c013879e>] lookup_module_symbol_name+0x14/0xbb
[ 5414.309613] [<c0130018>] print_name_offset+0x1a/0x8e
[ 5414.309630] [<c0119fd7>] release_console_sem+0x1b1/0x1d1
[ 5414.309656] [<c012ed00>] timekeeping_resume+0x7d/0xa1
[ 5414.309668] [<c012d566>] ktime_get_ts+0x1b/0x4e
[ 5414.309681] [<c01309aa>] timer_list_show+0x661/0xa55
[ 5414.309700] [<c01320c2>] tick_sched_timer+0x0/0xbd
[ 5414.309715] [<c022e2a3>] __handle_sysrq+0x7a/0x115
[ 5414.309730] [<c0228abc>] kbd_event+0x330/0x5d4
[ 5414.309741] [<c022878c>] kbd_event+0x0/0x5d4
[ 5414.309755] [<c0279c62>] input_event+0x410/0x44f
[ 5414.309790] [<c027d8de>] atkbd_interrupt+0x42c/0x56e
[ 5414.309817] [<c0276a78>] serio_interrupt+0x2f/0x69
[ 5414.309832] [<c02775f1>] i8042_interrupt+0x108/0x23e
[ 5414.309849] [<c0121429>] do_timer+0x1c/0xed
[ 5414.309867] [<c013a271>] handle_IRQ_event+0x25/0x4a
[ 5414.309885] [<c013b706>] handle_edge_irq+0xb2/0x102
[ 5414.309911] [<c0104ec3>] do_IRQ+0x3b/0x78
[ 5414.309923] [<c010db93>] smp_apic_timer_interrupt+0x59/0x84
[ 5414.309937] [<c010323f>] common_interrupt+0x23/0x28
[ 5414.309947] [<c0101565>] default_idle+0x0/0x3e
[ 5414.309959] [<c0101594>] default_idle+0x2f/0x3e
[ 5414.309973] [<c0100c19>] cpu_idle+0x5e/0x78
[ 5414.309982] [<c03f2c5a>] start_kernel+0x2c2/0x348
[ 5414.309996] [<c03f23dd>] unknown_bootoption+0x0/0x1de
[ 5414.310010] =======================

...

I don't use any *binary* drivers.


[1] It dies after minutes or after hour(s) , dies while doing something ( like compiling , copying stuff ) or while
doing nothing ( box idles no activity at all ).

Right now I got an 3h uptime ( heh ) while died before in about 2 hours 5 times..

>
> Greetings,
> Rafael


Gabriel
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/