The letter itself is in an attached file.
Intended for: linux-kernel@vger.kernel.org
Hello dear Linux Developer
(sorry, I do not know precisely whom I'm writing to),
PROBLEM: Opaque kernel bug in work with bzip2(?)
It occures every time I try to bunzip2 (1.0.2) the linux-2.6.11.10.tar.bz2
tarball. Of this action kernel panics with an enviable stability. At the first
time I was amazed and rebooted. Then, in the same case, I put down (on the
piece of paper) all that, what I saw on the screen (fig.1), and rebooted.
After restart I removed all modules, I think concerns the timer, and tried
to bunzip2 once again with the same result (fig.2).
Indeed I used to "tar -xpjv -C . -f linux-2.6.11.10.tar.bz2", i.e. used to
run bunzip2 implicitly. Guessing anything about bzip2, I tried to buzip2
my tarball strictly with "bunzip2 linux-2.6.11.10.tar.bz2". And kernel
was calm! But extracting was not successful. It said, the archive was
corrupted. I found linux-2.6.9.tar.bz2 on one of my forgotten CD. The result
was the same. Bad bzip2? I reinstalled it, but nothing extraordinary happend.
And also, rpm works (it uses bzip2)! But not very well... The kernel sources
were not fully installed. But in common they were installed (kernel 2.4.20).
Documentation for that kernel (it is in another rpm file) was installed
without any problems!
Another mistery.
Being logged in as usual user I tried to uninstall two rpm packages. I should
do it as super-user only, but I forgot. What should happen? Some errors and
nothing else. What really was? Charset modified, I saw pseudographical and some
else funny symbols instead of any letters.
Then, playing with bzip2, I got the fig.3 when kernel does not panic. I copied
it with mouse and it is an only precise screen shot. The first two figures
were scratched from screen to the paper, and rewritten from paper again.
May be bzip2 is not guilt and the above cases are circumstances only. But
when I do not use bzip2 there are now problems.
And more, problem was noticed when kernel felt in panic when there is a time
for to run rc.sysinit (at the booting). It said something about traces and did
nothing else. It was linux-2.6.11.12. I decided to recompile kernel and loaded
the one I used to work with for a long time (for a couple months) -
linux-2.6.11.10. Then I tried to bunzip2 tarball and ...
*******
*fig.1*
******************************************************************************
Kernel bug at kernel/timer.c:416!
invalid operand: 0000 [#2]
Modules linked in: usb_storage scsi_mod usblp uhci_hcd usbcore snd_seq_midi
snd_emu10k1_synth snd_emux_synth snd_seq_virmidi snd_seq_midi_event
snd_seq_midi_emul snd_seq snd_emu10k1 snd_rawmidi snd_seq_device
snd_ac97_codec snd_pcm snd_timer snd_page_alloc snd_util_mem snd_hwdep snd
soundcore psmouse af_packet xfs exports nls_koi8_r nls_cp866 vfat fat
nls_base rtc
CPU: 0
EIP: 0060: [<c0117842>] Not tainted VLI
EFLAGS: 00010007 (2.6.11.12)
EIP is at cascade+0x42/0x50
eax: c119afc0 ebx: c119afc0 ecx: 0002aa00 edx: c10d3eb0
esi: c02faa58 edi: c02fa100 ebp: 0000002a esp: c02f1fa8
ds: 007b es: 007b ss: 0068
Process swapper (pid: 0, threadinfo=c02f1000 task=c0269b20)
Stack: 00000000 c02f9e88 c02f1fcc c02f1000 c0117dd4 c02fa100 c02fa908 0000002a
c02f1fcc c02f1fcc c02f1fcc c02fb9c0 00000001 c02f9e88 0000000a 0032b007
c01144f2 c02f9e88 c02d1fa0 00000046 c02ee120 c01039e9
Call Trace:
[<c0117dd4>] run_timer_softirq+0x114/0x190
[<c01144f2>] __do_softirq+0x42/0xa0
[<c01039e9>] do_softirq+0x39/0x40
===========================================
[<c0103908>] do_IRQ+0x48/0x60
[<c010241a>] common_interrupt+0x1a/0x20
[<c0100550>] default_idle+0x0/0x30
[<c0100574>] default_idle+0x24/0x30
[<c01005dc>] cpu_idle+0x1c/0x60
[<c02d26e6>] start_kernel+0x136/0x150
Code:
01 ce 8b 1e eb 15 8d 76 00 89 d8 39 7b 18 75 1b 8b 1b 50 57 e8 90 fd ff ff
58 5a 39 f3 75 ea 89 36 89 76 04 89 e8 5b 5e 5f 5d c3 <0f> 0b a0 01 36 82 24 c0
eb db 8d 74 26 00 0f bf 05 f6 eb 26 c0
<0> Kernel panic - not syncing: Fatal exception in interrupt
*******
*fig.2*
*******************************************************************************
Kernel bug at kernel/timer.c:416!
invalid operand: 0000 [#2]
Modules linked in: floppy usbcore snd_page_alloc psmouse af_packet xfs exports
nls_koi8_r nls_cp866 nls_base
CPU: 0
EIP: 0060: [<c0117842>] Not tainted VLI
EFLAGS: 00010007 (2.6.11.12)
EIP is at cascade+0x42/0x50
eax: c119afc0 ebx: c119afc0 ecx: 0004b000 edx: c10d3eb0
esi: c02faa88 edi: c02fa100 ebp: 00000030 esp: c02f1fa8
ds: 007b es: 007b ss: 0068
Process bzip2 (pid: 4715, threadinfo=c02f1000 task=c36c0a20)
Stack: 00000000 c02f9e88 c02f1fcc c02f1000 c0117dd4 c02fa100 c02fa908 00000030
c02f1fcc c02f1fcc c02f1fcc 0000000a 00000001 c02f9e88 0000000a bfffd0e8
c01144f2 c02f9e88 c48e9fac 00000046 00000000 c01039e9
Call Trace:
[<c0117dd4>] run_timer_softirq+0x114/0x190
[<c01144f2>] __do_softirq+0x42/0xa0
[<c01039e9>] do_softirq+0x39/0x40
===========================================
[<c0103908>] do_IRQ+0x48/0x60
[<c010241a>] common_interrupt+0x1a/0x20
Code:
01 ce 8b 1e eb 15 8d 76 00 89 d8 39 7b 18 75 1b 8b 1b 50 57 e8 90 fd ff ff
58 5a 39 f3 75 ea 89 36 89 76 04 89 e8 5b 5e 5f 5d c3 <0f> 0b a0 01 36 82 24 c0
eb db 8d 74 26 00 0f bf 05 f6 eb 26 c0
<0> Kernel panic - not syncing: Fatal exception in interrupt
*******
*fig.3*
*******************************************************************************
Unable to handle kernel paging request at virtual address a274c285
printing eip:
a274c285
*pde = 00000000
Oops: 0000 [#1]
Modules linked in: isofs zlib_inflate nls_cp1251 udf ide_cd cdrom usb_storage
scsi_mod uhci_hcd usbcore snd_seq_midi snd_emu10k1_synth snd_emux_synth
snd_seq_virmidi snd_seq_midi_event snd_seq_midi_emul snd_seq snd_emu10k1
snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd_page_alloc
snd_util_mem snd_hwdep snd soundcore psmouse af_packet xfs exportfs nls_koi8_r
nls_cp866 vfat fat nls_base rtc
CPU: 0
EIP: 0060:[<a274c285>] Not tainted VLI
EFLAGS: 00010286 (2.6.11.12)
EIP is at 0xa274c285
eax: 00000000 ebx: c46d7000 ecx: 00000003 edx: c10bf460
esi: c46d7000 edi: 0813f6c8 ebp: c1d9a000 esp: c1d9afa8
ds: 007b es: 007b ss: 0068
Process setterm (pid: 2565, threadinfo=c1d9a000 task=c3f82aa0)
Stack: c46d7000 57500424 08140088 c1d9afc4 0813f6c8 00000000 c01021f7 00000000
00000000 00000000 00000000 00000000 00000000 00000000 0000007b 0000007b
0000000b b7feea70 00000073 00000246 bffffa80 0000007b
Call Trace:
[<c01021f7>] syscall_call+0x7/0xb
Code: Bad EIP value.
/dev/tty2
Online: 2
óÕ éÀÌ 23 00:16:05 EEST 2005
Uptime: 10 hours
<1>Unable to handle kernel paging request at virtual address a274c285
printing eip:
a274c285
*pde = 00000000
Oops: 0000 [#2]
Modules linked in: isofs zlib_inflate nls_cp1251 udf ide_cd cdrom usb_storage
scsi_mod uhci_hcd usbcore snd_seq_midi snd_emu10k1_synth snd_emux_synth
snd_seq_virmidi snd_seq_midi_event snd_seq_midi_emul snd_seq snd_emu10k1
snd_rawmidi snd_seq_device snd_ac97_codec snd_pcm snd_timer snd_page_alloc
snd_util_mem snd_hwdep snd soundcore psmouse af_packet xfs exportfs nls_koi8_r
nls_cp866 vfat fat nls_base rtc
CPU: 0
EIP: 0060:[<a274c285>] Not tainted VLI
EFLAGS: 00010286 (2.6.11.12)
EIP is at 0xa274c285
eax: 00000000 ebx: c1a06000 ecx: 00000003 edx: c10bf460
esi: c1a06000 edi: 0813ffc8 ebp: c1d9a000 esp: c1d9afa8
ds: 007b es: 007b ss: 0068
Process procinfo (pid: 2589, threadinfo=c1d9a000 task=c3f82aa0)
Stack: c1a06000 57500424 08140088 c1d9afc4 0813ffc8 00000000 c01021f7 00000000
00000000 00000000 00000000 00000000 00000000 00000000 0000007b 0000007b
0000000b b7feea70 00000073 00000246 bffffa40 0000007b
Call Trace:
[<c01021f7>] syscall_call+0x7/0xb
Code: Bad EIP value.
*******************************************************************************
May be I wrote too much excessive information, apologies please, for I do not
know what is useful and what is not.
P.S.
If I can help somehow else you are welcome to e-mail me (but I will answer
after a couple days).
P.S.S.
Now I'm going to download linux-2.6.12 (gzipped:) and the last bzip2 and test
them together. I'll inform you of any kernel crashes.
Obligedly yours
Igor
