[Pm-utils] System unstable after resume if not setting NEED_CLOCK_SYNC for pm-utils

[Grissiom]

 On Fri, Apr 17, 2009 at 15:23, Robby Workman <rworkman at slackware.com>wrote:

> This is on the stock 2.6.27.7 kernel, correct?
> Honestly, I'm not sure what else would be needed on the pm-utils end
> of things; this seems more like a kernel bug.  You might consider building
> the latest in the 2.6.27.x series (.21 last I checked) to see
> if it still happens there; this might have already been fixed upstream,
> and I'd hate to waste someone's time with a bug report that's already
> fixed.
>

Sorry, I forgot to tell you that I am using slackware-current, all things
up-to-date, i.e., linux-2.6.29.1 . Although I have configured and complied
kernel myself, the configuration should be very close to the official one in
this aspect.

Unfortunately, the problem is more complicated. The truth is:

1) With NEED_CLOCK_SYNC off, 2.6.29.1 break about 10min after resume. dmesgs
attached in the LQ post.

2) With NEED_CLOCK_SYNC on, 2.6.29.1 will break about 1hour or so after
resume, at least this case has occurred two times on my box. Symptom alike.
dmesg snip here: (the attachment is full dmesg)
======================================
[ 6413.351177] BUG: unable to handle kernel paging request at f76f6000
[ 6413.351193] IP: [<c016f919>] get_page_from_freelist+0x2a9/0x470
[ 6413.351213] *pde = 00007067 *pte = e5cb4839
[ 6413.351224] Oops: 0003 [#2] SMP
[ 6413.351230] last sysfs file:
/sys/devices/system/cpu/cpu1/cache/index2/shared_cpu_map
[ 6413.351239] Modules linked in: radeon drm vboxnetflt vboxdrv
snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device
snd_pcm_oss snd_mixer_oss ipv6 nls_cp936 vfat fat ext4 jbd2 crc16 fuse b43
mac80211 dell_laptop cfg80211 sdhci_pci dcdbas ricoh_mmc joydev led_class
ohci_hcd ati_agp input_polldev shpchp sg agpgart video output
[ 6413.351287]
[ 6413.351294] Pid: 3145, comm: plasma Tainted: G      D
(2.6.29.1-slk-based-2 #4) Inspiron 1501
[ 6413.351302] EIP: 0060:[<c016f919>] EFLAGS: 00210246 CPU: 1
[ 6413.351311] EIP is at get_page_from_freelist+0x2a9/0x470
[ 6413.351317] EAX: 00000000 EBX: c16edec0 ECX: 00000400 EDX: 00000003
[ 6413.351323] ESI: 00000000 EDI: f76f6000 EBP: c16edec0 ESP: d7411d78
[ 6413.351328]  DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
[ 6413.351334] Process plasma (pid: 3145, ti=d7410000 task=f6b389f0
task.ti=d7410000)
[ 6413.351340] Stack:
[ 6413.351342]  00000002 00000044 00200246 f76f6000 c0617920 00000002
00000000 00000000
[ 6413.351354]  001280d2 00000002 00200246 00000001 c0587980 00000000
00000000 c058930c
[ 6413.351365]  c0587c80 c0587140 c0587c80 f6b389f0 c9b1fff8 001280d2
c016fddd c0589300
[ 6413.351378] Call Trace:
[ 6413.351384]  [<c016fddd>] __alloc_pages_internal+0xbd/0x470
[ 6413.351395]  [<c017eae5>] handle_mm_fault+0x385/0x610
[ 6413.351409]  [<c011bc8e>] do_page_fault+0x23e/0x6a0
[ 6413.351420]  [<c0101d65>] __switch_to+0x175/0x190
[ 6413.351427]  [<c044f0e4>] schedule+0x294/0x930
[ 6413.351435]  [<c014cad2>] tick_dev_program_event+0x32/0xb0
[ 6413.351446]  [<c014cbca>] tick_program_event+0x2a/0x40
[ 6413.351453]  [<c013261c>] irq_exit+0x3c/0x80
[ 6413.351461]  [<c011685a>] smp_apic_timer_interrupt+0x5a/0x90
[ 6413.351468]  [<c011ba50>] do_page_fault+0x0/0x6a0
[ 6413.351475]  [<c0451292>] error_code+0x72/0x78
[ 6413.351487] Code: 24 20 00 79 40 8b 7c 24 2c 85 ff 7e 38 89 dd 31 f6 90
ba 03 00 00 00 89 e8 e8 94 03 fb ff b9 00 04 00 00 89 44 24 0c 89 c7 31 c0
<f3> ab 8b 44 24 0c ba 03 00 00 00 46 83 c5 20 e8 b3 01 fb ff 3b
[ 6413.351542] EIP: [<c016f919>] get_page_from_freelist+0x2a9/0x470 SS:ESP
0068:d7411d78
[ 6413.351556] ---[ end trace 1c05764e1d5f5ab3 ]---
[ 6413.351566] note: plasma[3145] exited with preempt_count 1
======================================

3) 2.6.27.21 kernel won't break. With or Without NEED_CLOCK_SYNC. I
suspend/resume 3 times and run happily 2+ hours after resume.

I think it maybe a regression in kernel...

-- 
Cheers,
Grissiom
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freedesktop.org/archives/pm-utils/attachments/20090418/74755acf/attachment.htm 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dmesg-long
Type: application/octet-stream
Size: 31417 bytes
Desc: not available
Url : http://lists.freedesktop.org/archives/pm-utils/attachments/20090418/74755acf/attachment-0001.obj