[paranoia] cdparanoia 9.7 crash on redhat 6.1
Ian Collinson
icollinson at imerge.co.uk
Mon Dec 3 03:46:11 PST 2001
Doug,
Thanks for that. We installed the version of the sg driver you suggested.
Unfortunately, after a few hours of ripping, using postgres, and doing other
things, both kswapd and cdparanoia provoked kernel problems. See below for
ksymoops output. Once kswapd was killed off, the box became unusable until
reboot.
ksymoops 2.4.3 on i686 2.4.7. Options used
-V (default)
-k /proc/ksyms (default)
-l /proc/modules (default)
-o /lib/modules/2.4.7/ (default)
-m /boot/System.map (specified)
Warning (compare_maps): ksyms_base symbol ivideo_R__ver_ivideo not found in
System.map. Ignoring ksyms_base entry
Warning (compare_maps): ksyms_base symbol sis_free_R__ver_sis_free not found
in System.map. Ignoring ksyms_base entry
Warning (compare_maps): ksyms_base symbol sis_malloc_R__ver_sis_malloc not
found in System.map. Ignoring ksyms_base entry
Warning (compare_maps): mismatch on symbol V32U96eyeLocation , pctel says
c4ca93f0, /lib/modules/2.4.7/misc/pctel.o says c4cc6ed4. Ignoring
/lib/modules/2.4.7/misc/pctel.o entry
kernel BUG at page_alloc.c:75!
invalid operand: 0000
CPU: 0
EIP: 0010:[<c012991e>]
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010296
eax: 0000001f ebx: c10a655c ecx: 00000000 edx: ffffffff
esi: c10a6584 edi: 00000102 ebp: 00000000 esp: c3bc7f64
ds: 0018 es: 0018 ss: 0018
Process kswapd (pid: 4, stackpage=c3bc7000)
Stack: c024a86c c024a960 0000004b c10a655c c10a6584 00000102 00000004
00000101
00000004 c21477b0 00000003 c0128a46 c012a25e c0128c79 00000000
000000c0
00000000 0008e000 c10a655c 00000040 00000000 000002a1 00000001
c0129349
Call Trace: [<c0128a46>] [<c012a25e>] [<c0128c79>] [<c0129349>] [<c01293de>]
[<c0105000>] [<c0105663>]
Code: 0f 0b 83 c4 0c 89 da 2b 15 8c 00 34 c0 89 d0 c1 e0 04 29 d0
>>EIP; c012991e <__free_pages_ok+3e/338> <=====
Trace; c0128a46 <page_launder+36a/8e0>
Trace; c012a25e <__free_pages+1a/1c>
Trace; c0128c78 <page_launder+59c/8e0>
Trace; c0129348 <do_try_to_free_pages+1c/58>
Trace; c01293de <kswapd+5a/e4>
Trace; c0105000 <_stext+0/0>
Trace; c0105662 <kernel_thread+22/30>
Code; c012991e <__free_pages_ok+3e/338>
00000000 <_EIP>:
Code; c012991e <__free_pages_ok+3e/338> <=====
0: 0f 0b ud2a <=====
Code; c0129920 <__free_pages_ok+40/338>
2: 83 c4 0c addl $0xc,%esp
Code; c0129922 <__free_pages_ok+42/338>
5: 89 da movl %ebx,%edx
Code; c0129924 <__free_pages_ok+44/338>
7: 2b 15 8c 00 34 c0 subl 0xc034008c,%edx
Code; c012992a <__free_pages_ok+4a/338>
d: 89 d0 movl %edx,%eax
Code; c012992c <__free_pages_ok+4c/338>
f: c1 e0 04 shll $0x4,%eax
Code; c0129930 <__free_pages_ok+50/338>
12: 29 d0 subl %edx,%eax
kernel BUG at vmscan.c:395!
invalid operand: 0000
CPU: 0
EIP: 0010:[<c012866b>]
EFLAGS: 00010296
eax: 0000001c ebx: c10a6578 ecx: 00000001 edx: 00000001
esi: c0297950 edi: c10a655c ebp: c0297934 esp: c2d83ec4
ds: 0018 es: 0018 ss: 0018
Process cdparanoia-9.7- (pid: 16263, stackpage=c2d83000)
Stack: c024a3ec c024a50a 0000018b c0297934 c0297ac0 00000001 00000000
00000001
c0297950 00000000 c0129f44 c0297934 00000000 c0297ac8 00000001
fffffff4
c012a061 c0297abc 00000000 00000001 00000001 000000d2 c111a154
c2d83f98
Call Trace: [<c0129f44>] [<c012a061>] [<c0129f80>] [<c0124e68>] [<c012e9e9>]
[<c0106d9b>]
Code: 0f 0b 83 c4 0c 31 c0 0f b3 47 18 19 c0 85 c0 75 19 68 8c 01
>>EIP; c012866a <reclaim_page+372/3e4> <=====
Trace; c0129f44 <__alloc_pages_limit+6c/90>
Trace; c012a060 <__alloc_pages+dc/27c>
Trace; c0129f80 <_alloc_pages+18/1c>
Trace; c0124e68 <generic_file_write+390/5b8>
Trace; c012e9e8 <sys_write+8c/c0>
Trace; c0106d9a <system_call+32/38>
Code; c012866a <reclaim_page+372/3e4>
00000000 <_EIP>:
Code; c012866a <reclaim_page+372/3e4> <=====
0: 0f 0b ud2a <=====
Code; c012866c <reclaim_page+374/3e4>
2: 83 c4 0c addl $0xc,%esp
Code; c012866e <reclaim_page+376/3e4>
5: 31 c0 xorl %eax,%eax
Code; c0128670 <reclaim_page+378/3e4>
7: 0f b3 47 18 btrl %eax,0x18(%edi)
Code; c0128674 <reclaim_page+37c/3e4>
b: 19 c0 sbbl %eax,%eax
Code; c0128676 <reclaim_page+37e/3e4>
d: 85 c0 testl %eax,%eax
Code; c0128678 <reclaim_page+380/3e4>
f: 75 19 jne 2a <_EIP+0x2a> c0128694
<reclaim_page+39c/3e4>
Code; c012867a <reclaim_page+382/3e4>
11: 68 8c 01 00 00 pushl $0x18c
4 warnings issued. Results may not be reliable.
Cheers,
Ian
> -----Original Message-----
> From: Douglas Gilbert [mailto:dgilbert at interlog.com]
> Sent: 30 November 2001 04:39
> To: paranoia at xiph.org; Ian Collinson
> Subject: Re: [paranoia] cdparanoia 9.7 crash on redhat 6.1
>
>
> Monty wrote:
> >
> > On Thu, Nov 29, 2001 at 02:00:14PM -0000, Ian Collinson wrote:
> > >
> > > cdparanoia 9.7 is crashing often (mid rip) on our boxes
> (RedHat 6.1). Once
> > > it has crashed, then running it again blocks forerver
> (strace shows it
> > > blocks on open("/dev/sg0"...)).
> >
> > Cdparanoia crashed because of a kernel fault, possibly in
> SG, that has
> > left the kernel with a gaping head wound. As the fault
> left SG in an
> > inconsistent state, it's unsurprising that SG would not be
> usable from
> > that point on.
> >
> > OS bug. I'll let the good SG maintainers chase this further
> as they've
> > already spoken up.
> >
>
> Monty,
> Well thanks. Unfortunately your analysis is correct. In my
> defence, I didn't add that particular bug into sg. The good
> news, however, is that it has been fixed. Since Ian seems to
> know how to compile a module (as he is using lk 2.4.7 on
> RH 6.1) go to http://www.torque.net/sg and fetch sg3120.tgz .
> This is the sg version found in lk 2.4.10 through to lk 2.4.16 .
>
> FYI Monty, sg version 3.1.22 [currently submitted for lk 2.4.17]
> adds the ability to memory map (via the mmap() system
> call) sg's reserve buffer. It is faster than direct IO and
> very frugal on CPU cycles. That's more cycles for paranoia's
> crunching.
>
> Ian,
> Thanks for the ksymoops output. That is exactly what I
> wanted/expected to see.
>
> Doug Gilbert
>
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Imerge Limited Tel :- +44 (0)1954 783600
Unit 6 Bar Hill Business Park Fax :- +44 (0)1954 783601
Saxon Way Web :- http://www.imerge.co.uk
Bar Hill
Cambridge
CB3 8SL
United Kingdom
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
--- >8 ----
List archives: http://www.xiph.org/archives/
Paranoia homepage: http://www.xiph.org/paranoia/
To unsubscribe from this list, send a message to 'paranoia-request at xiph.org'
containing only the word 'unsubscribe' in the body. No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.
More information about the Paranoia
mailing list