You are not logged in.
When starting games through wine they sort of start. Music plays, but there's no picture whatsoever. Then boom goes the dynamite and filesystem gets corrupted. No files exist in any bin directories, so no reboot and REISUB does not work either.
After power off-fsck-reboot everything works again.
kernel.log shows a lot of Corrupted low memory at c0005ffc (5ffc phys) = 00ffffff messages (memory addresses change) and after that:
Nov 3 05:21:16 arch kernel: ------------[ cut here ]------------
Nov 3 05:21:16 arch kernel: WARNING: at arch/x86/kernel/check.c:134 check_for_bios_corruption+0xd2/0xe0()
Nov 3 05:21:16 arch kernel: Hardware name: 23748TG
Nov 3 05:21:16 arch kernel: Memory corruption detected in low memory
Nov 3 05:21:16 arch kernel: Modules linked in: ipv6 ext2 pcmcia snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss fan radeon ttm drm i2c_algo_bit joydev yenta_socket rsrc_nonstatic pcmcia_core snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm e1000 snd_timer snd soundcore snd_page_alloc video output sr_mod cdrom battery uhci_hcd irtty_sir thinkpad_acpi nsc_ircc sir_dev ehci_hcd usbcore rfkill led_class nvram shpchp pci_hotplug thermal ppdev ac button sg ipw2200 parport_pc psmouse i2c_i801 irda crc_ccitt i2c_core floppy iTCO_wdt iTCO_vendor_support evdev lp parport serio_raw intel_agp agpgart libipw lib80211 cpufreq_ondemand acpi_cpufreq freq_table processor rtc_cmos rtc_core rtc_lib ext4 mbcache jbd2 crc16 sd_mod ata_piix ata_generic pata_acpi libata scsi_mod
Nov 3 05:21:16 arch kernel: Pid: 6, comm: events/0 Tainted: G W 2.6.31-ARCH #1
Nov 3 05:21:16 arch kernel: Call Trace:
Nov 3 05:21:16 arch kernel: [<c104654a>] ? warn_slowpath_common+0x7a/0xc0
Nov 3 05:21:16 arch kernel: [<c102a092>] ? check_for_bios_corruption+0xd2/0xe0
Nov 3 05:21:16 arch kernel: [<c1046607>] ? warn_slowpath_fmt+0x37/0x60
Nov 3 05:21:16 arch kernel: [<c102a092>] ? check_for_bios_corruption+0xd2/0xe0
Nov 3 05:21:16 arch kernel: [<c102a0a0>] ? check_corruption+0x0/0x50
Nov 3 05:21:16 arch kernel: [<c102a0b3>] ? check_corruption+0x13/0x50
Nov 3 05:21:16 arch kernel: [<c105d03f>] ? worker_thread+0x11f/0x280
Nov 3 05:21:16 arch kernel: [<c1062de0>] ? autoremove_wake_function+0x0/0x60
Nov 3 05:21:16 arch kernel: [<c105cf20>] ? worker_thread+0x0/0x280
Nov 3 05:21:16 arch kernel: [<c106298c>] ? kthread+0x8c/0xa0
Nov 3 05:21:16 arch kernel: [<c1062900>] ? kthread+0x0/0xa0
Nov 3 05:21:16 arch kernel: [<c10048e7>] ? kernel_thread_helper+0x7/0x10
Nov 3 05:21:16 arch kernel: ---[ end trace a7919e7f17c0a727 ]---
Nov 3 05:21:42 arch kernel: EXT4-fs error (device sda3): htree_dirblock_to_tree: bad entry in directory #4228: inode out of bounds - offset=0, inode=1315463744, rec_len=4096, name_len=34
Nov 3 05:21:42 arch kernel: EXT4-fs error (device sda3): htree_dirblock_to_tree: bad entry in directory #2762: inode out of bounds - offset=0, inode=1315463744, rec_len=4096, name_len=34
Nov 3 05:21:43 arch kernel: EXT4-fs error (device sda3): htree_dirblock_to_tree: bad entry in directory #2: inode out of bounds - offset=0, inode=1315463744, rec_len=4096, name_len=34
Nov 3 05:21:43 arch kernel: EXT4-fs error (device sda3): htree_dirblock_to_tree: bad entry in directory #2: inode out of bounds - offset=0, inode=1315463744, rec_len=4096, name_len=34
Nov 3 05:21:45 arch kernel: EXT4-fs error (device sda3): htree_dirblock_to_tree: bad entry in directory #4228: inode out of bounds - offset=0, inode=1315463744, rec_len=4096, name_len=34
Nov 3 05:21:45 arch kernel: EXT4-fs error (device sda3): htree_dirblock_to_tree: bad entry in directory #2762: inode out of bounds - offset=0, inode=1315463744, rec_len=4096, name_len=34
Nov 3 05:21:45 arch kernel: EXT4-fs error (device sda3): htree_dirblock_to_tree: bad entry in directory #2: inode out of bounds - offset=0, inode=1315463744, rec_len=4096, name_len=34system: Thinkpad T42 with Radeon 7500
kernel: 2.6.31-5
Xorg: 1.6.3.901
xf86-video-ati: 6.12.4
Offline
Woah. Looks like a kernel panic. Did anything trigger this? A package upgrade, a Wine upgrade, or something? If so, try going back to the previous revision.
I'd say report that, http://bugzilla.kernel.org/
Offline
I'm not sure as I don't play really often. Yesterday I ran HOMMIV through wine, but it was really slow. Then I found out that my xorg.conf was responsible for not starting hardware acceleration. As I didn't need it, I just deleted the file and that reenabled HW acceleration (glxgears went up from 70 to 700fps) and after that I noticed this problem.
Also couple of days ago I upgraded to newer xorg and xf86-video-ati, but that made xorg hang once in a while so I downgraded.
EDIT: as it turns out this problem is not related to wine. I installed extreme tux racer, ran it and same things happened. So either my graphics chip has decided not to cooperate anymore or something's messed up in kernel/drivers.
Upgraded xorg and xf86-video-ati again, this time sort of successfully. With KMS disabled same problems as before. With KMS enabled games start, but colors are corrupted, everything's red.
Last edited by DoS (2009-11-03 05:20:43)
Offline
Fancy thing this "Corrupted Low Memory", as I get it as well on my T42...
I downgraded to kernel 2.6.30.1 and still had the Corrupt message in my dmesg. I then downgraded to 2.6.29.1 and voila the Corrupt messages went away, and so did the random module crashes (as you seem to be getting too after the Corrupt messages).
The problem seems to be with:
CONFIG_X86_CHECK_BIOS_CORRUPTION: Check for low memory corruption
Which is a new feature ever since 2.6.28.x kernel. It was working properly for us T42 ppl, until 2.6.30.x came along, and still hasn't been fixed in 2.6.31.x. To fix this and get system stability back, add the following to the end of your kernel line (in either Lilo or Grub):
memory_corruption_check=0
That stops the kernel from checking for corrupted low memory, but, since the check procedure messes up the system and causes modules to crash, it's better to disable it until the problem is worked out.
To be on the safe side, I ran a few memory check programs, and they all ran without errors. So my low memory isn't corrupt after all, just a bug in a feature that can be disabled.
Offline