Re: Bizarre oops after suspend to RAM (was: Re: [ACPI] Resume from Suspend to RAM)

[Martin Michlmayr <tbm@cyrius.com>]

* Pavel Machek <pavel@ucw.cz> [2005-06-06 15:10]:
> > machine (an HP nc4000) to resume. Since then, it simply freezes before 
> > hitting "Back to C" despite having had no kernel or configuration 
> > changes. The behaviour is very non-deterministic, which makes me wonder 
> > about something in the suspend or resume process damaging state.
> 
> Hmm, strange. You may want to test if length of sleep affects it...

I tried to resume immediately when the screen goes blank and got the
following results.  Out of 9 trials I got:

 - 4x: it would say "Restarting tasks..." but then oops and hang
   immediately.  See the oops below.
 - 4x: come back and give me a shell - but it hangs
 - 1x: come back, give me a shell and let me type; built-ins worked
   but commands produced the oops posted by Matthew a few days ago.

So far it has never come back when I don't resume immediately.

The new oops I got is:

root@(none):/# mount -t proc none /proc
root@(none):/# echo 3 > /proc/acpi/sleep
Stopping tasks: ===|
    ACPI-0286: *** Error: No installed handler for fixed event [00000002]
Restarting tasks... done
Unable to handle kernel paging request at virtual address 00008265
 printing eip:
c015ac62
*pde = 00000000
Oops: 0002 [#2]
PREEMPT
Modules linked in:
CPU:    0
EIP:    0060:[<c015ac62>]    Not tainted VLI
EFLAGS: 00010202   (2.6.12-rc5)
EIP is at filp_open+0x32/0x70
eax: f7c7a000   ebx: 00008241   ecx: 00000003   edx: 00000180
esi: 00008241   edi: f7c7a000   ebp: c191c000   esp: c191df48
ds: 007b   es: 007b   ss: 0068
Process bash (pid: 1, threadinfo=c191c000 task=c18bfa20)
Stack: bfeafcf4 00008242 c191df64 c191df58 c016690b 080f1f48 c191df64 c191df6c
       00000014 00000000 f7c7a000 00000003 c18c0e40 c191c000 c015af95 c18c0e40
       00000003 00000003 f7c7a000 00008241 00000003 c015b109 f7c7a000 00008241
Call Trace:
 [<c016690b>] sys_stat64+0x1b/0x40
 [<c015af95>] get_unused_fd+0xa5/0xe0
 [<c015b109>] sys_open+0x49/0x90
 [<c0103245>] syscall_call+0x7/0xb
Code: 8b 5c 24 5c 8d 43 01 a8 03 0f 44 c3 89 c2 83 ca 02 f6 c4 02 0f 45 c2 8d 54 24 10 89 54 24 0c 8b 54 24 60 89 44 24 04 8b 44 24 58 <89> 56 24 08 89 04 24 e8 62 19 01 00 85 c0 74 0e 8b 5c 24 50 83
 <0>Kernel panic - not syncing: Attempted to kill init!

-- 
Martin Michlmayr
http://www.cyrius.com/