Skip to contentSkip to conversation

Linux 2.6.11-rc1

DISCUSSION Voluspa 2 participants 5 messages Jan 12, 2005 – Jan 13, 2005

lore
Share & tools
lore

Voluspa

author
#0 lore raw

Yes, tainted. X black screen, no keyboard. Power button to turn off. I really don't feel like compiling lots of debug-kernels to chase this, unless someone is really interested, which I doubt.

Signature · 2 lines
Mvh
Mats Johannesson
Jan 12 08:28:45 loke kernel: ------------[ cut here ]------------
Jan 12 08:28:45 loke kernel: kernel BUG at <bad filename>:395!
Jan 12 08:28:45 loke kernel: invalid operand: 0000 [#1]
Jan 12 08:28:45 loke kernel: PREEMPT 
Jan 12 08:28:45 loke kernel: Modules linked in: nvidia 8139too crc32 snd_seq_oss snd_emu10k1_synth snd_emux_synth snd_seq_virmidi snd_seq_midi_event snd_seq_midi_emul snd_seq snd_pcm_oss snd_mixer_oss snd_emu10k1 snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd_page_alloc snd_util_mem snd_hwdep snd soundcore
Jan 12 08:28:45 loke kernel: CPU:    0
Jan 12 08:28:45 loke kernel: EIP:    0060:[__change_page_attr+169/286]    Tainted: P      VLI
Jan 12 08:28:45 loke kernel: EFLAGS: 00213002   (2.6.11-rc1) 
Jan 12 08:28:45 loke kernel: eax: 050001e3   ebx: 05150000   ecx: c10067c0   edx: 050001e3
Jan 12 08:28:45 loke kernel: esi: c033ec50   edi: 00000163   ebp: c1000000   esp: c61d1db4
Jan 12 08:28:45 loke kernel: ds: 007b   es: 007b   ss: 0068
Jan 12 08:28:45 loke kernel: Process X (pid: 653, threadinfo=c61d1000 task=cfdaa0e0)
Jan 12 08:28:45 loke kernel: Stack: c5150000 c10a2a00 00000010 00000000 00203246 c010db7d 00000163 00000011 
Jan 12 08:28:45 loke kernel:        c10a2800 cdbcd120 cdbac800 c61d1e20 c010d863 00010000 d0a40000 d0f48bec 
Jan 12 08:28:45 loke kernel:        d0d5d180 d0a40000 00010000 cdbac800 d0d5d173 cff08400 ca0f4c00 c61d1e60 
Jan 12 08:28:45 loke kernel: Call Trace:
Jan 12 08:28:45 loke kernel:  [change_page_attr+48/97]
Jan 12 08:28:45 loke kernel:  [iounmap+101/118]
Jan 12 08:28:45 loke kernel:  [pg0+280890348/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278876544/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278876531/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278861617/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278825594/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278898228/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278903180/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278825594/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+280882328/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278885835/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278828198/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278888316/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278888407/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278890097/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278890110/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278877374/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278900967/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278901015/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278901003/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+280873510/1070162944]
Jan 12 08:28:45 loke kernel:  [__fput+61/228]
Jan 12 08:28:45 loke kernel:  [filp_close+89/95]
Jan 12 08:28:45 loke kernel:  [sysenter_past_esp+82/117]
Jan 12 08:28:45 loke kernel: Code: 16 35 c0 c1 f9 05 c1 e1 0c 0b 0d c8 52 2c c0 e8 10 ff ff ff 89 d9 ff 41 04 eb 12 a9 80 00 00 00 75 09 09 fb 89 1e ff 49 04 eb 02 <0f> 0b 8b 01 f6 c4 08 75 64 8b 41 04 40 75 02 0f 0b a1 0c 3e 2c 
Jan 12 08:28:45 loke kernel:  <6>note: X[653] exited with preempt_count 1
Jan 12 08:28:45 loke kernel: scheduling while atomic: X/0x00000001/653
Jan 12 08:28:45 loke kernel:  [schedule+64/1068]
Jan 12 08:28:45 loke kernel:  [autoremove_wake_function+0/45]
Jan 12 08:28:45 loke kernel:  [__sched_text_start+134/237]
Jan 12 08:28:45 loke kernel:  [default_wake_function+0/12]
Jan 12 08:28:45 loke kernel:  [release_mem+446/458]
Jan 12 08:28:45 loke kernel:  [__down_failed+7/12]
Jan 12 08:28:45 loke kernel:  [pg0+280890751/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278877374/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278902522/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278902546/1070162944]
Jan 12 08:28:45 loke kernel:  [dput+27/459]
Jan 12 08:28:45 loke kernel:  [pg0+280877129/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+280873401/1070162944]
Jan 12 08:28:45 loke kernel:  [destroy_inode+34/49]
Jan 12 08:28:45 loke kernel:  [__fput+61/228]
Jan 12 08:28:45 loke kernel:  [filp_close+89/95]
Jan 12 08:28:45 loke kernel:  [put_files_struct+86/169]
Jan 12 08:28:45 loke kernel:  [do_exit+257/703]
Jan 12 08:28:45 loke kernel:  [do_trap+0/162]
Jan 12 08:28:45 loke kernel:  [do_invalid_op+0/139]
Jan 12 08:28:45 loke kernel:  [do_invalid_op+127/139]
Jan 12 08:28:45 loke kernel:  [__change_page_attr+169/286]
Jan 12 08:28:45 loke kernel:  [kobject_get+15/19]
Jan 12 08:28:45 loke kernel:  [get_device+14/20]
Jan 12 08:28:45 loke kernel:  [pci_dev_get+15/19]
Jan 12 08:28:45 loke kernel:  [pci_get_subsys+174/206]
Jan 12 08:28:45 loke kernel:  [pci_get_device+11/14]
Jan 12 08:28:45 loke kernel:  [pg0+280889510/1070162944]
Jan 12 08:28:45 loke kernel:  [pci_read+28/33]
Jan 12 08:28:45 loke kernel:  [error_code+43/48]
Jan 12 08:28:45 loke kernel:  [__change_page_attr+169/286]
Jan 12 08:28:45 loke kernel:  [change_page_attr+48/97]
Jan 12 08:28:45 loke kernel:  [iounmap+101/118]
Jan 12 08:28:45 loke kernel:  [pg0+280890348/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278876544/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278876531/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278861617/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278825594/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278898228/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278903180/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278825594/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+280882328/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278885835/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278828198/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278888316/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278888407/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278890097/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278890110/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278877374/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278900967/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278901015/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+278901003/1070162944]
Jan 12 08:28:45 loke kernel:  [pg0+280873510/1070162944]
Jan 12 08:28:45 loke kernel:  [__fput+61/228]
Jan 12 08:28:45 loke kernel:  [filp_close+89/95]
Jan 12 08:28:45 loke kernel:  [sysenter_past_esp+82/117]
author
#1 lore raw
Quoted 3 lines · On Wed, 12 Jan 2005 09:52:38 +0100 Voluspa wrote:

Yes, tainted. X black screen, no keyboard. Power button to turn off. I really don't feel like compiling lots of debug-kernels to chase this, unless someone is really interested, which I doubt.

A fyi directed to other users: It was the nvidia module. Xorg nv is fine. Did a CONFIG_KALLSYMS=y kernel and am attaching the output.

Turning on relevant debugging options in "Kernel hacking" hides the bug (ie prevents it from happening). Very annoying and is a behaviour I've seen before with other bugs. The price of kernel complexity, I guess.

Signature · 2 lines
Mvh
Mats Johannesson
Jan 13 01:09:42 loke kernel: ------------[ cut here ]------------
Jan 13 01:09:42 loke kernel: kernel BUG at <bad filename>:395!
Jan 13 01:09:42 loke kernel: invalid operand: 0000 [#1]
Jan 13 01:09:42 loke kernel: PREEMPT 
Jan 13 01:09:42 loke kernel: Modules linked in: nvidia 8139too crc32 snd_seq_oss snd_emu10k1_synth snd_emux_synth snd_seq_virmidi snd_seq_midi_event snd_seq_midi_emul snd_seq snd_pcm_oss snd_mixer_oss snd_emu10k1 snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd_page_alloc snd_util_mem snd_hwdep snd soundcore
Jan 13 01:09:42 loke kernel: CPU:    0
Jan 13 01:09:42 loke kernel: EIP:    0060:[__change_page_attr+169/286]    Tainted: P      VLI
Jan 13 01:09:42 loke kernel: EFLAGS: 00213002   (2.6.11-rc1) 
Jan 13 01:09:42 loke kernel: EIP is at __change_page_attr+0xa9/0x11e
Jan 13 01:09:42 loke kernel: eax: 054001e3   ebx: 05530000   ecx: c1006c40   edx: 054001e3
Jan 13 01:09:42 loke kernel: esi: c0362c54   edi: 00000163   ebp: c1000000   esp: ca147db4
Jan 13 01:09:42 loke kernel: ds: 007b   es: 007b   ss: 0068
Jan 13 01:09:42 loke kernel: Process X (pid: 575, threadinfo=ca147000 task=cd4b4060)
Jan 13 01:09:42 loke kernel: Stack: c5530000 c10aa600 00000010 00000000 00203246 c010db9d 00000163 00000011 
Jan 13 01:09:42 loke kernel:        c10aa400 ce2eda60 cd54fc00 ca147e20 c010d883 00010000 d0a80000 d0f75bec 
Jan 13 01:09:42 loke kernel:        d0d8a180 d0a80000 00010000 cd54fc00 d0d8a173 cff08400 cd54f400 ca147e60 
Jan 13 01:09:42 loke kernel: Call Trace:
Jan 13 01:09:42 loke kernel:  [change_page_attr+48/97] change_page_attr+0x30/0x61
Jan 13 01:09:42 loke kernel:  [iounmap+101/118] iounmap+0x65/0x76
Jan 13 01:09:42 loke kernel:  [pg0+280927212/1070015488] os_unmap_kernel_space+0x9/0xa [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278913408/1070015488] _nv001706rm+0x20/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278913395/1070015488] _nv001706rm+0x13/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278898481/1070015488] _nv002359rm+0xe9/0x184 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278862458/1070015488] _nv001955rm+0x36/0xe0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278935092/1070015488] _nv001297rm+0x9c/0xa8 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278940044/1070015488] rm_teardown_agp+0x48/0x50 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278862458/1070015488] _nv001955rm+0x36/0xe0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280919192/1070015488] nv_agp_teardown+0x4c/0x7c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278922699/1070015488] _nv001708rm+0x73/0xa0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278865062/1070015488] _nv001847rm+0x26/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278925180/1070015488] _nv000650rm+0x58/0xcc [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278925271/1070015488] _nv000650rm+0xb3/0xcc [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278926961/1070015488] _nv001362rm+0x71/0xb0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278926974/1070015488] _nv001362rm+0x7e/0xb0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278914238/1070015488] _nv001820rm+0x12/0x18 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937831/1070015488] rm_disable_adapter+0x2f/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937879/1070015488] rm_disable_adapter+0x5f/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937867/1070015488] rm_disable_adapter+0x53/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280910374/1070015488] nv_kern_close+0x168/0x1e1 [nvidia]
Jan 13 01:09:42 loke kernel:  [__fput+61/228] __fput+0x3d/0xe4
Jan 13 01:09:42 loke kernel:  [filp_close+89/95] filp_close+0x59/0x5f
Jan 13 01:09:42 loke kernel:  [sysenter_past_esp+82/117] sysenter_past_esp+0x52/0x75
Jan 13 01:09:42 loke kernel: Code: 56 37 c0 c1 f9 05 c1 e1 0c 0b 0d c8 62 2c c0 e8 10 ff ff ff 89 d9 ff 41 04 eb 12 a9 80 00 00 00 75 09 09 fb 89 1e ff 49 04 eb 02 <0f> 0b 8b 01 f6 c4 08 75 64 8b 41 04 40 75 02 0f 0b a1 0c 4e 2c 
Jan 13 01:09:42 loke kernel:  <6>note: X[575] exited with preempt_count 1
Jan 13 01:09:42 loke kernel: scheduling while atomic: X/0x00000001/575
Jan 13 01:09:42 loke kernel:  [schedule+64/1068] schedule+0x40/0x42c
Jan 13 01:09:42 loke kernel:  [autoremove_wake_function+0/45] autoremove_wake_function+0x0/0x2d
Jan 13 01:09:42 loke kernel:  [__sched_text_start+134/237] __down+0x86/0xed
Jan 13 01:09:42 loke kernel:  [default_wake_function+0/12] default_wake_function+0x0/0xc
Jan 13 01:09:42 loke kernel:  [release_mem+446/458] release_mem+0x1be/0x1ca
Jan 13 01:09:42 loke kernel:  [__down_failed+7/12] __down_failed+0x7/0xc
Jan 13 01:09:42 loke kernel:  [pg0+280927615/1070015488] .text.lock.os_interface+0x7/0x18 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278914238/1070015488] _nv001820rm+0x12/0x18 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278939386/1070015488] rm_free_unused_clients+0x2e/0x88 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278939410/1070015488] rm_free_unused_clients+0x46/0x88 [nvidia]
Jan 13 01:09:42 loke kernel:  [dput+27/459] dput+0x1b/0x1cb
Jan 13 01:09:42 loke kernel:  [pg0+280913993/1070015488] nv_kern_ctl_close+0x77/0xa1 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280910265/1070015488] nv_kern_close+0xfb/0x1e1 [nvidia]
Jan 13 01:09:42 loke kernel:  [destroy_inode+34/49] destroy_inode+0x22/0x31
Jan 13 01:09:42 loke kernel:  [__fput+61/228] __fput+0x3d/0xe4
Jan 13 01:09:42 loke kernel:  [filp_close+89/95] filp_close+0x59/0x5f
Jan 13 01:09:42 loke kernel:  [put_files_struct+86/169] put_files_struct+0x56/0xa9
Jan 13 01:09:42 loke kernel:  [do_exit+257/703] do_exit+0x101/0x2bf
Jan 13 01:09:42 loke kernel:  [do_trap+0/162] do_trap+0x0/0xa2
Jan 13 01:09:42 loke kernel:  [do_invalid_op+0/139] do_invalid_op+0x0/0x8b
Jan 13 01:09:42 loke kernel:  [do_invalid_op+127/139] do_invalid_op+0x7f/0x8b
Jan 13 01:09:42 loke kernel:  [__change_page_attr+169/286] __change_page_attr+0xa9/0x11e
Jan 13 01:09:42 loke kernel:  [kobject_get+15/19] kobject_get+0xf/0x13
Jan 13 01:09:42 loke kernel:  [get_device+14/20] get_device+0xe/0x14
Jan 13 01:09:42 loke kernel:  [pci_dev_get+15/19] pci_dev_get+0xf/0x13
Jan 13 01:09:42 loke kernel:  [pci_get_subsys+174/206] pci_get_subsys+0xae/0xce
Jan 13 01:09:42 loke kernel:  [pci_get_device+11/14] pci_get_device+0xb/0xe
Jan 13 01:09:42 loke kernel:  [pg0+280926374/1070015488] os_pci_init_handle+0x7d/0x86 [nvidia]
Jan 13 01:09:42 loke kernel:  [pci_read+28/33] pci_read+0x1c/0x21
Jan 13 01:09:42 loke kernel:  [error_code+43/48] error_code+0x2b/0x30
Jan 13 01:09:42 loke kernel:  [__change_page_attr+169/286] __change_page_attr+0xa9/0x11e
Jan 13 01:09:42 loke kernel:  [change_page_attr+48/97] change_page_attr+0x30/0x61
Jan 13 01:09:42 loke kernel:  [iounmap+101/118] iounmap+0x65/0x76
Jan 13 01:09:42 loke kernel:  [pg0+280927212/1070015488] os_unmap_kernel_space+0x9/0xa [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278913408/1070015488] _nv001706rm+0x20/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278913395/1070015488] _nv001706rm+0x13/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278898481/1070015488] _nv002359rm+0xe9/0x184 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278862458/1070015488] _nv001955rm+0x36/0xe0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278935092/1070015488] _nv001297rm+0x9c/0xa8 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278940044/1070015488] rm_teardown_agp+0x48/0x50 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278862458/1070015488] _nv001955rm+0x36/0xe0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280919192/1070015488] nv_agp_teardown+0x4c/0x7c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278922699/1070015488] _nv001708rm+0x73/0xa0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278865062/1070015488] _nv001847rm+0x26/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278925180/1070015488] _nv000650rm+0x58/0xcc [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278925271/1070015488] _nv000650rm+0xb3/0xcc [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278926961/1070015488] _nv001362rm+0x71/0xb0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278926974/1070015488] _nv001362rm+0x7e/0xb0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278914238/1070015488] _nv001820rm+0x12/0x18 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937831/1070015488] rm_disable_adapter+0x2f/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937879/1070015488] rm_disable_adapter+0x5f/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937867/1070015488] rm_disable_adapter+0x53/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280910374/1070015488] nv_kern_close+0x168/0x1e1 [nvidia]
Jan 13 01:09:42 loke kernel:  [__fput+61/228] __fput+0x3d/0xe4
Jan 13 01:09:42 loke kernel:  [filp_close+89/95] filp_close+0x59/0x5f
Jan 13 01:09:42 loke kernel:  [sysenter_past_esp+82/117] sysenter_past_esp+0x52/0x75
#2 lore raw

this looks suspiciously like a kernel bug Andi Kleen and I are currently investigating. He has fixes in the bk branches. I would have expected them to be in such a new kernel, but I noticed that you're using an x86 kernel, and his initial fixes were only in the x86_64 arch files.

change_page_attr has a book-keeping bug that surprisingly hasn't caused problems until recently (on my todo list is to track down what caused this problem to suddenly start triggering).

the x86_64 change in bk is here, but the only thing you really need is the 'get_page' fix. you should be able to manually edit linux/arch/i386/mm/pageattr.c:__change_page_attr(), update that single line and be fine:

http://linux.bkbits.net:8080/linux-2.6/diffs/arch/x86_64/mm/[email protected]?nav=index.html|src/|src/arch|src/arch/x86_64|src/arch/x86_64/mm|hist/arch/x86_64/mm/pageattr.c

Thanks,
Terence

Quoted 122 lines · On Thu, Jan 13, 2005 at 02:13:28AM +0100, [email protected] wrote:
Quoted 3 lines · On Wed, 12 Jan 2005 09:52:38 +0100 Voluspa wrote:

Yes, tainted. X black screen, no keyboard. Power button to turn off. I really don't feel like compiling lots of debug-kernels to chase this, unless someone is really interested, which I doubt.

A fyi directed to other users: It was the nvidia module. Xorg nv is fine. Did a CONFIG_KALLSYMS=y kernel and am attaching the output.

Turning on relevant debugging options in "Kernel hacking" hides the bug (ie prevents it from happening). Very annoying and is a behaviour I've seen before with other bugs. The price of kernel complexity, I guess.

Signature · 107 lines
Mvh
Mats Johannesson

[-- mutt.octet.filter file type: "ASCII text, with very long lines" --]


Jan 13 01:09:42 loke kernel: ------------[ cut here ]------------
Jan 13 01:09:42 loke kernel: kernel BUG at <bad filename>:395!
Jan 13 01:09:42 loke kernel: invalid operand: 0000 [#1]
Jan 13 01:09:42 loke kernel: PREEMPT 
Jan 13 01:09:42 loke kernel: Modules linked in: nvidia 8139too crc32 snd_seq_oss snd_emu10k1_synth snd_emux_synth snd_seq_virmidi snd_seq_midi_event snd_seq_midi_emul snd_seq snd_pcm_oss snd_mixer_oss snd_emu10k1 snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd_page_alloc snd_util_mem snd_hwdep snd soundcore
Jan 13 01:09:42 loke kernel: CPU:    0
Jan 13 01:09:42 loke kernel: EIP:    0060:[__change_page_attr+169/286]    Tainted: P      VLI
Jan 13 01:09:42 loke kernel: EFLAGS: 00213002   (2.6.11-rc1) 
Jan 13 01:09:42 loke kernel: EIP is at __change_page_attr+0xa9/0x11e
Jan 13 01:09:42 loke kernel: eax: 054001e3   ebx: 05530000   ecx: c1006c40   edx: 054001e3
Jan 13 01:09:42 loke kernel: esi: c0362c54   edi: 00000163   ebp: c1000000   esp: ca147db4
Jan 13 01:09:42 loke kernel: ds: 007b   es: 007b   ss: 0068
Jan 13 01:09:42 loke kernel: Process X (pid: 575, threadinfo=ca147000 task=cd4b4060)
Jan 13 01:09:42 loke kernel: Stack: c5530000 c10aa600 00000010 00000000 00203246 c010db9d 00000163 00000011 
Jan 13 01:09:42 loke kernel:        c10aa400 ce2eda60 cd54fc00 ca147e20 c010d883 00010000 d0a80000 d0f75bec 
Jan 13 01:09:42 loke kernel:        d0d8a180 d0a80000 00010000 cd54fc00 d0d8a173 cff08400 cd54f400 ca147e60 
Jan 13 01:09:42 loke kernel: Call Trace:
Jan 13 01:09:42 loke kernel:  [change_page_attr+48/97] change_page_attr+0x30/0x61
Jan 13 01:09:42 loke kernel:  [iounmap+101/118] iounmap+0x65/0x76
Jan 13 01:09:42 loke kernel:  [pg0+280927212/1070015488] os_unmap_kernel_space+0x9/0xa [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278913408/1070015488] _nv001706rm+0x20/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278913395/1070015488] _nv001706rm+0x13/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278898481/1070015488] _nv002359rm+0xe9/0x184 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278862458/1070015488] _nv001955rm+0x36/0xe0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278935092/1070015488] _nv001297rm+0x9c/0xa8 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278940044/1070015488] rm_teardown_agp+0x48/0x50 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278862458/1070015488] _nv001955rm+0x36/0xe0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280919192/1070015488] nv_agp_teardown+0x4c/0x7c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278922699/1070015488] _nv001708rm+0x73/0xa0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278865062/1070015488] _nv001847rm+0x26/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278925180/1070015488] _nv000650rm+0x58/0xcc [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278925271/1070015488] _nv000650rm+0xb3/0xcc [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278926961/1070015488] _nv001362rm+0x71/0xb0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278926974/1070015488] _nv001362rm+0x7e/0xb0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278914238/1070015488] _nv001820rm+0x12/0x18 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937831/1070015488] rm_disable_adapter+0x2f/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937879/1070015488] rm_disable_adapter+0x5f/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937867/1070015488] rm_disable_adapter+0x53/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280910374/1070015488] nv_kern_close+0x168/0x1e1 [nvidia]
Jan 13 01:09:42 loke kernel:  [__fput+61/228] __fput+0x3d/0xe4
Jan 13 01:09:42 loke kernel:  [filp_close+89/95] filp_close+0x59/0x5f
Jan 13 01:09:42 loke kernel:  [sysenter_past_esp+82/117] sysenter_past_esp+0x52/0x75
Jan 13 01:09:42 loke kernel: Code: 56 37 c0 c1 f9 05 c1 e1 0c 0b 0d c8 62 2c c0 e8 10 ff ff ff 89 d9 ff 41 04 eb 12 a9 80 00 00 00 75 09 09 fb 89 1e ff 49 04 eb 02 <0f> 0b 8b 01 f6 c4 08 75 64 8b 41 04 40 75 02 0f 0b a1 0c 4e 2c 
Jan 13 01:09:42 loke kernel:  <6>note: X[575] exited with preempt_count 1
Jan 13 01:09:42 loke kernel: scheduling while atomic: X/0x00000001/575
Jan 13 01:09:42 loke kernel:  [schedule+64/1068] schedule+0x40/0x42c
Jan 13 01:09:42 loke kernel:  [autoremove_wake_function+0/45] autoremove_wake_function+0x0/0x2d
Jan 13 01:09:42 loke kernel:  [__sched_text_start+134/237] __down+0x86/0xed
Jan 13 01:09:42 loke kernel:  [default_wake_function+0/12] default_wake_function+0x0/0xc
Jan 13 01:09:42 loke kernel:  [release_mem+446/458] release_mem+0x1be/0x1ca
Jan 13 01:09:42 loke kernel:  [__down_failed+7/12] __down_failed+0x7/0xc
Jan 13 01:09:42 loke kernel:  [pg0+280927615/1070015488] .text.lock.os_interface+0x7/0x18 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278914238/1070015488] _nv001820rm+0x12/0x18 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278939386/1070015488] rm_free_unused_clients+0x2e/0x88 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278939410/1070015488] rm_free_unused_clients+0x46/0x88 [nvidia]
Jan 13 01:09:42 loke kernel:  [dput+27/459] dput+0x1b/0x1cb
Jan 13 01:09:42 loke kernel:  [pg0+280913993/1070015488] nv_kern_ctl_close+0x77/0xa1 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280910265/1070015488] nv_kern_close+0xfb/0x1e1 [nvidia]
Jan 13 01:09:42 loke kernel:  [destroy_inode+34/49] destroy_inode+0x22/0x31
Jan 13 01:09:42 loke kernel:  [__fput+61/228] __fput+0x3d/0xe4
Jan 13 01:09:42 loke kernel:  [filp_close+89/95] filp_close+0x59/0x5f
Jan 13 01:09:42 loke kernel:  [put_files_struct+86/169] put_files_struct+0x56/0xa9
Jan 13 01:09:42 loke kernel:  [do_exit+257/703] do_exit+0x101/0x2bf
Jan 13 01:09:42 loke kernel:  [do_trap+0/162] do_trap+0x0/0xa2
Jan 13 01:09:42 loke kernel:  [do_invalid_op+0/139] do_invalid_op+0x0/0x8b
Jan 13 01:09:42 loke kernel:  [do_invalid_op+127/139] do_invalid_op+0x7f/0x8b
Jan 13 01:09:42 loke kernel:  [__change_page_attr+169/286] __change_page_attr+0xa9/0x11e
Jan 13 01:09:42 loke kernel:  [kobject_get+15/19] kobject_get+0xf/0x13
Jan 13 01:09:42 loke kernel:  [get_device+14/20] get_device+0xe/0x14
Jan 13 01:09:42 loke kernel:  [pci_dev_get+15/19] pci_dev_get+0xf/0x13
Jan 13 01:09:42 loke kernel:  [pci_get_subsys+174/206] pci_get_subsys+0xae/0xce
Jan 13 01:09:42 loke kernel:  [pci_get_device+11/14] pci_get_device+0xb/0xe
Jan 13 01:09:42 loke kernel:  [pg0+280926374/1070015488] os_pci_init_handle+0x7d/0x86 [nvidia]
Jan 13 01:09:42 loke kernel:  [pci_read+28/33] pci_read+0x1c/0x21
Jan 13 01:09:42 loke kernel:  [error_code+43/48] error_code+0x2b/0x30
Jan 13 01:09:42 loke kernel:  [__change_page_attr+169/286] __change_page_attr+0xa9/0x11e
Jan 13 01:09:42 loke kernel:  [change_page_attr+48/97] change_page_attr+0x30/0x61
Jan 13 01:09:42 loke kernel:  [iounmap+101/118] iounmap+0x65/0x76
Jan 13 01:09:42 loke kernel:  [pg0+280927212/1070015488] os_unmap_kernel_space+0x9/0xa [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278913408/1070015488] _nv001706rm+0x20/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278913395/1070015488] _nv001706rm+0x13/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278898481/1070015488] _nv002359rm+0xe9/0x184 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278862458/1070015488] _nv001955rm+0x36/0xe0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278935092/1070015488] _nv001297rm+0x9c/0xa8 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278940044/1070015488] rm_teardown_agp+0x48/0x50 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278862458/1070015488] _nv001955rm+0x36/0xe0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280919192/1070015488] nv_agp_teardown+0x4c/0x7c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278922699/1070015488] _nv001708rm+0x73/0xa0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278865062/1070015488] _nv001847rm+0x26/0x2c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278925180/1070015488] _nv000650rm+0x58/0xcc [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278925271/1070015488] _nv000650rm+0xb3/0xcc [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278926961/1070015488] _nv001362rm+0x71/0xb0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278926974/1070015488] _nv001362rm+0x7e/0xb0 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278914238/1070015488] _nv001820rm+0x12/0x18 [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937831/1070015488] rm_disable_adapter+0x2f/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937879/1070015488] rm_disable_adapter+0x5f/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+278937867/1070015488] rm_disable_adapter+0x53/0x8c [nvidia]
Jan 13 01:09:42 loke kernel:  [pg0+280910374/1070015488] nv_kern_close+0x168/0x1e1 [nvidia]
Jan 13 01:09:42 loke kernel:  [__fput+61/228] __fput+0x3d/0xe4
Jan 13 01:09:42 loke kernel:  [filp_close+89/95] filp_close+0x59/0x5f
Jan 13 01:09:42 loke kernel:  [sysenter_past_esp+82/117] sysenter_past_esp+0x52/0x75
author
#3 lore raw
Quoted 4 lines · On Wed, 12 Jan 2005 19:21:59 -0600 Terence Ripperda wrote:

the x86_64 change in bk is here, but the only thing you really need is the 'get_page' fix. you should be able to manually edit linux/arch/i386/mm/pageattr.c:__change_page_attr(), update that single line and be fine:

Not being a coder I was unable to change that function correctly. replacing "struct page *page" with a "unsigned long pfn" craved other changes further down - if that even is how you begin hacking... I'll just wait for the real implementation.

Signature · 2 lines
Mvh
Mats Johannesson
author
#4 lore raw
In reply to Terence Ripperda
Quoted 3 lines · On Wed, 12 Jan 2005 19:21:59 -0600 Terence Ripperda wrote:

change_page_attr has a book-keeping bug that surprisingly hasn't caused problems until recently (on my todo list is to track down what caused this problem to suddenly start triggering).

If it can help, I'm writing this on a 2.6.10-bk14 kernel. So it's just the final bits going into 2.6.11-rc1 that pushed the bug on. At least on my machine.

Can't find any incremental bk14->rc1 to finalize that statement, though.

Ok kernels, from 2.6.10, that I've compiled and run now is: -bk7, -bk11, -bk13 and -bk14.

Signature · 2 lines
Mvh
Mats Johannesson