Bug 469770

As an extra data point, I have tried hibernating the system and resuming it and cannot get it to crash or otherwise fail to resume from a hibernate.

Looks like memory corruption? Is the backtrace always the same?

it probably is memory corruption.. got a different backtrace this time:

BUG: unable to handle kernel paging request at virtual address 06fd05fd
 printing eip:
06fd05fd
*pde = 00000000
Oops: 0000 [#1]
SMP 
last sysfs file: /devices/pci0000:00/0000:00:00.0/resource
Modules linked in: i915 drm autofs4 hidp rfcomm l2cap bluetooth sunrpc ipv6 xfrm_nalgo crypto_api cpufreq_ondemand acpi_cpufreq dm_multipath scsi_dh video backlight sbs i2c_ec button battery asus_acpi ac parport_pc lp parport joydev sg snd_hda_intel snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq tg3 serio_raw snd_seq_device sdhci ide_cd libphy snd_pcm_oss mmc_core pcspkr snd_mixer_oss i2c_i801 snd_pcm cdrom i2c_core snd_timer snd_page_alloc snd_hwdep snd soundcore dm_snapshot dm_zero dm_mirror dm_log dm_mod ata_piix ahci libata sd_mod scsi_mod ext3 jbd uhci_hcd ohci_hcd ehci_hcd
CPU:    0
EIP:    0060:[<06fd05fd>]    Not tainted VLI
EFLAGS: 00010246   (2.6.18-121.el5 #1) 
EIP is at 0x6fd05fd
eax: f6058efc   ebx: c062b040   ecx: f6058eb4   edx: f58d6680
esi: f6058eb4   edi: f58d6680   ebp: 0000001e   esp: f6058e6c
ds: 007b   es: 007b   ss: 0068
Process acpid (pid: 3236, ti=f6058000 task=f7d86000 task.ti=f6058000)
Stack: c05af004 0000001e f6058efc f6058efc 0000001e f6058e94 00000000 c05af48a 
       f6058ed0 00000001 f6058e98 ffffffff ffffffff 0000001e f58d6680 00000001 
       00000000 f6058eb4 00000000 00000000 f6058ed0 00000001 00000000 00000000 
Call Trace:
 [<c05af004>] do_sock_write+0xa3/0xaa
 [<c05af48a>] sock_aio_write+0x53/0x61
 [<c047240e>] do_sync_write+0xb6/0xf1
 [<c043466b>] autoremove_wake_function+0x0/0x2d
 [<c0472cd8>] vfs_write+0xb2/0x143
 [<c04732b9>] sys_write+0x3c/0x63
 [<c0404f17>] syscall_call+0x7/0xb
 =======================
Code:  Bad EIP value.
EIP: [<06fd05fd>] 0x6fd05fd SS:ESP 0068:f6058e6c
 

dmesg log and crash dump available at:
http://test185.test.redhat.com/crashdumps/Z61t/2008-11-5.1/

I'll try and reproduce it with the -122 kernel, any other ideas?

If you try suspending from single user mode, does it resume reliably? Best way to test this is:

dbus-send --system --print-reply --dest=org.freedesktop.Hal /org/freedesktop/Hal/devices/computer org.freedesktop.Hal.Device.SystemPowerManagement.Suspend int32:0

I can't get that dbus command to work.. looks like the word wrapping may have mangled it. In single user mode it looks like one of the services it needs isn't running and in runlevel 3 I just get back the usage summary. 

I tried 'echo mem > /sys/power/state' and that suspended the system but it did not return from suspend nor did it appear to crash no matter what I did.

Can you attach the output of lshal?

Created attachment 322626 [details]
lshal output

Hm. Did this ever work? From X (as root) does:

pm-suspend --quirk-s3-bios --quirk-s3-mode

result in any improvement?

I have a Z61t here, too. There is a regression in 5.3, filed as bug 468289, where the screen is blank on resume. VT switching works around it, but if I Ctrl-Alt-Backspace instead, I do get a kernel panic.

Zack, can you try the same thing?

I'll try it, but this laptop never needed any quirks before.

Er, actually it does seem to help. Resumes fine, and there's no kernel panic if I Ctrl-Alt-Backspace.

Hm. Ok, in that case if it's a regression it's in the Intel graphics driver - presumably it's now doing slightly less reprogramming than it was before, and then managed to stomp over inappropriate memory when killed. We can add quirks for the Z61t easily enough, but there's the potential for other systems to have been affected.

*** This bug has been marked as a duplicate of bug 468289 ***

This has been marked a duplicate of 468289 but attempting to access that thread returns "You are not authorized to access bug #468289."