Re: Very rare resume bug

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Sunday 01 March 2009, Maxim Levitsky wrote:
> I currently am testing suspend/resume on my desktop system.
> I already kept it in suspend to ram state in 2 weeks, and now
> I resumed it, and screen didn't came back, I tried switching vts
> it did seem to help, but then everything hung.

Well, it's difficult to say what problem this is.

It may be the problem with shared interrupts coming in too early during resume
that should be fixed in the mainline.

> I got this in logs:
> 
> 
> 
> > Feb 27 21:45:12 MAIN kernel: [47277.954656] WARNING: at /build/buildd/linux-2.6.27/kernel/power/main.c:176 suspend_test_finish+0x74/0x80()
> > Feb 27 21:45:12 MAIN kernel: [47277.954658] Modules linked in: videodev v4l1_compat af_packet binfmt_misc ipv6 nfsd auth_rpcgss exportfs ppdev video output battery container wmi sbs pci_slot sbshc nfs lockd nfs_acl sunrpc iptable_filter ip_tables x_tables ext2 mbcache vfat fat ac kvm_intel kvm i2c_dev snd_virmidi snd_seq_virmidi coretemp cpufreq_ondemand cpufreq_powersave cpufreq_userspace acpi_cpufreq freq_table sbp2 eeprom lp snd_hda_intel snd_seq_dummy snd_seq_oss snd_seq_midi nvidia(P) snd_pcm_oss joydev snd_mixer_oss snd_rawmidi snd_pcm parport_pc snd_seq_midi_event snd_seq psmouse intel_agp shpchp evdev iTCO_wdt serio_raw pci_hotplug i2c_i801 parport button iTCO_vendor_support agpgart snd_timer snd_seq_device snd soundcore i2c_core snd_page_alloc heci reiserfs sd_mod crc_t10dif s
 r_mod cdrom sg usbhid hid ata_piix pata_marvell ata_generic ohci1394 pata_acpi ieee1394 libata scsi_mod dock uhci_hcd ehci_hcd usbcore e1000e dm_mirror dm_log dm_snapshot dm_mod thermal pro!
 ce!
>  ssor fan fuse vesafb fbcon tileblit f
> > Feb 27 21:45:12 MAIN kernel: nt bitblit softcursor [last unloaded: microcode]
> > Feb 27 21:45:12 MAIN kernel: [47277.954720] Pid: 27868, comm: pm-suspend Tainted: P        W 2.6.27-11-generic #1
> > Feb 27 21:45:12 MAIN kernel: [47277.954722]  [<c037d236>] ? printk+0x1d/0x1f
> > Feb 27 21:45:12 MAIN kernel: [47277.954726]  [<c0131e99>] warn_on_slowpath+0x59/0x90
> > Feb 27 21:45:12 MAIN kernel: [47277.954730]  [<c014d305>] ? sched_clock_cpu+0xd5/0x170
> > Feb 27 21:45:12 MAIN kernel: [47277.954734]  [<c014c15f>] ? down_trylock+0x2f/0x40
> > Feb 27 21:45:12 MAIN kernel: [47277.954737]  [<c0132572>] ? try_acquire_console_sem+0x12/0x40
> > Feb 27 21:45:12 MAIN kernel: [47277.954740]  [<c024ed00>] ? kobject_put+0x20/0x50
> > Feb 27 21:45:12 MAIN kernel: [47277.954744]  [<c015d2b4>] suspend_test_finish+0x74/0x80
> > Feb 27 21:45:12 MAIN kernel: [47277.954747]  [<c015d3a6>] suspend_devices_and_enter+0xe6/0x190
> > Feb 27 21:45:12 MAIN kernel: [47277.954750]  [<c015d621>] enter_state+0xd1/0x100
> > Feb 27 21:45:12 MAIN kernel: [47277.954752]  [<c015d6d5>] state_store+0x85/0xd0
> > Feb 27 21:45:12 MAIN kernel: [47277.954754]  [<c015d650>] ? state_store+0x0/0xd0
> > Feb 27 21:45:12 MAIN kernel: [47277.954757]  [<c024ebc4>] kobj_attr_store+0x24/0x30
> > Feb 27 21:45:12 MAIN kernel: [47277.954759]  [<c01ffac7>] sysfs_write_file+0x97/0x100
> > Feb 27 21:45:12 MAIN kernel: [47277.954763]  [<c01b2760>] vfs_write+0xa0/0x110
> > Feb 27 21:45:12 MAIN kernel: [47277.954765]  [<c01ffa30>] ? sysfs_write_file+0x0/0x100
> > Feb 27 21:45:12 MAIN kernel: [47277.954768]  [<c01b28a2>] sys_write+0x42/0x70
> > Feb 27 21:45:12 MAIN kernel: [47277.954771]  [<c0103f7b>] sysenter_do_call+0x12/0x2f
> > Feb 27 21:45:12 MAIN kernel: [47277.954774]  [<c0370000>] ? enable_nonboot_cpus+0x60/0xc0
> > Feb 27 21:45:12 MAIN kernel: [47277.954777]  =======================
> > Feb 27 21:45:12 MAIN kernel: [47277.954779] ---[ end trace f9ae6cac97d5cf9f ]---
> 
> 
> I have nvidia driver loaded - otherwise nothing suspends.

Yes, the NVidia driver is necessary for resume to work with the newer NVidia
adapters.  Unfortunately.

Thanks,
Rafael