Hi,
while booting 2.5.2-dj7 I get the attached oops when syslogd is loading.
The system only responds to sysrq so the oops is hand copied!
Last boot I had a very similar oops while starting proftpd, so it doesn't seem to be a syslogd bug.
My filesystem seems to be ok after booting a working kernel.
If you have any questions, suggestions, fixes, patches, ... ask me :)
System.map:
c0190fdc T reiserfs_panic
c0191028 t print_leaf
Bye
---------------------------------------------
Starting system log daemon...VS-500: unknown uniqueness 536870912
VS-500: unknown uniqueness 536870912
VS-500: unknown uniqueness 268435456
VS-500: unknown uniqueness 268435456
======================================================================
LEAF NODE (26720) contains level=1, nr_items=16, free_space=1452 rd_key
-------------------------------------------------------------------------------------------------------------------------------------------------------------
|##| type | key | ilen | free_space | version | loc |
0 | *3.6* [58952 21828 0x0 SD], item_len 44, item_location 4052, free_space (entry_count) 65535
mode | size | nlinks | first direct | mtime
0100644 | 3 | 1 | 2038999 | 1012397516
======================================================================
PAP-5580: reiserfs_cut_from_item: item to convert does not exist ([58952 21828 0x1 IND]) invalid operant: 0000
CPU: 0
EIP: 0010:[<c0191005>] Not tainted
EFLAGS: 60010282
eax: 0000005b ebx: c02b9640 ecx: 00000001 edx: 00000001
esi: x15a6800 edi: c15a6800 ebp: cdd95e80 esp: cdd95c30
ds: 0018 es: 0018 ss: 0018
Process syslogd (pid: 92, stackpage=cdd950000)
Stack: c02b819a c0351ba0 c02b9640 cdd95c54 00000000 00000000 c0198065 c15a6800
c02b9640 cdd95e80 ceece980 00000003 ffffffff 00000001 00001000 00000000
cdd95ed8 cdd95ed8 000028a3 cdd95c9c cdd95c98 00000ff8 00000001 00000ff8
Call Trace: [<c0198065>] [<c0198649>] [<c01893fb>] [<c018a65b>] [<c01335748>] [<c0132378>] [<c010864b>]
Code: 0f 0b 68 a0 1b 35 c0 b8 a0 81 2b c0 8d 96 cc 00 00 00 85 fb
Hello!
Can you please feed this entire oops to a ksymoops?
Just 2 first entries are not enough.
Thank you.
BTW, you are running on a IDE system, right?
Bye,
Oleg
On Wed, Jan 30, 2002 at 03:14:20PM +0100, Sebastian Dr?ge wrote:
> CPU: 0
> EIP: 0010:[<c0191005>] Not tainted
> EFLAGS: 60010282
> eax: 0000005b ebx: c02b9640 ecx: 00000001 edx: 00000001
> esi: x15a6800 edi: c15a6800 ebp: cdd95e80 esp: cdd95c30
> ds: 0018 es: 0018 ss: 0018
> Process syslogd (pid: 92, stackpage=cdd950000)
> Stack: c02b819a c0351ba0 c02b9640 cdd95c54 00000000 00000000 c0198065 c15a6800
> c02b9640 cdd95e80 ceece980 00000003 ffffffff 00000001 00001000 00000000
> cdd95ed8 cdd95ed8 000028a3 cdd95c9c cdd95c98 00000ff8 00000001 00000ff8
> Call Trace: [<c0198065>] [<c0198649>] [<c01893fb>] [<c018a65b>] [<c01335748>] [<c0132378>] [<c010864b>]
> Code: 0f 0b 68 a0 1b 35 c0 b8 a0 81 2b c0 8d 96 cc 00 00 00 85 fb
Hi,
On Wed, 30 Jan 2002 17:37:15 +0300
Oleg Drokin <[email protected]> wrote:
> Hello!
>
> Can you please feed this entire oops to a ksymoops?
Here it is (I can't use /proc/ksyms because I'm running an other kernel right now and I can't run ksymoops with 2.5.2-dj7... vmlinux and System.map are the ones from 2.5.2-dj7):
slomo:/usr/src/linux2$ ksymoops -v vmlinux -m System.map -K
ksymoops 2.4.3 on i686 2.5.2-dj6. Options used
-v vmlinux (specified)
-K (specified)
-l /proc/modules (default)
-o /lib/modules/2.5.2-dj6/ (default)
-m System.map (specified)
No modules in ksyms, skipping objects
No ksyms, skipping lsmod
Reading Oops report from the terminal
PAP-5580: reiserfs_cut_from_item: item to convert does not exist ([58952 21828 0x1 IND]) invalid operant: 0000
CPU: 0
EIP: 0010:[<c0191005>] Not tainted
EFLAGS: 60010282
eax: 0000005b ebx: c02b9640 ecx: 00000001 edx: 00000001
esi: x15a6800 edi: c15a6800 ebp: cdd95e80 esp: cdd95c30
ds: 0018 es: 0018 ss: 0018
Process syslogd (pid: 92, stackpage=cdd950000)
Stack: c02b819a c0351ba0 c02b9640 cdd95c54 00000000 00000000 c0198065 c15a6800
c02b9640 cdd95e80 ceece980 00000003 ffffffff 00000001 00001000 00000000
cdd95ed8 cdd95ed8 000028a3 cdd95c9c cdd95c98 00000ff8 00000001 00000ff8
Call Trace: [<c0198065>] [<c0198649>] [<c01893fb>] [<c018a65b>] [<c0133574>] [<c0132378>] [<c010864b>]
Code: 0f 0b 68 a0 1b 35 c0 b8 a0 81 2b c0 8d 96 cc 00 00 00 85 fbCPU: 0
EIP: 0010:[<c0191005>] Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 60010282
eax: 0000005b ebx: c02b9640 ecx: 00000001 edx: 00000001
esi: x15a6800 edi: c15a6800 ebp: cdd95e80 esp: cdd95c30
ds: 0018 es: 0018 ss: 0018
Process syslogd (pid: 92, stackpage=cdd950000)
Stack: c02b819a c0351ba0 c02b9640 cdd95c54 00000000 00000000 c0198065 c15a6800
c02b9640 cdd95e80 ceece980 00000003 ffffffff 00000001 00001000 00000000
cdd95ed8 cdd95ed8 000028a3 cdd95c9c cdd95c98 00000ff8 00000001 00000ff8
Call Trace: [<c0198065>] [<c0198649>] [<c01893fb>] [<c018a65b>] [<c0133574>] [<c0132378>] [<c010864b>]
Code: 0f 0b 68 a0 1b 35 c0 b8 a0 81 2b c0 8d 96 cc 00 00 00 85
>>EIP; c0191004 <reiserfs_panic+28/4c> <=====
Trace; c0198064 <reiserfs_cut_from_item+1ec/468>
Trace; c0198648 <reiserfs_do_truncate+320/4ac>
Trace; c01893fa <reiserfs_truncate_file+1f2/294>
Trace; c018a65a <reiserfs_file_release+31a/340>
Trace; c0133574 <fput+4c/d0>
Trace; c0132378 <sys_close+90/a4>
Trace; c010864a <system_call+32/38>
Code; c0191004 <reiserfs_panic+28/4c>
00000000 <_EIP>:
Code; c0191004 <reiserfs_panic+28/4c> <=====
0: 0f 0b ud2a <=====
Code; c0191006 <reiserfs_panic+2a/4c>
2: 68 a0 1b 35 c0 push $0xc0351ba0
Code; c019100a <reiserfs_panic+2e/4c>
7: b8 a0 81 2b c0 mov $0xc02b81a0,%eax
Code; c0191010 <reiserfs_panic+34/4c>
c: 8d 96 cc 00 00 00 lea 0xcc(%esi),%edx
Code; c0191016 <reiserfs_panic+3a/4c>
12: 85 fb
> BTW, you are running on a IDE system, right?
Yes.
I'll provide some more system informations now ;)
Pentium II 350 / 256 MB SD-RAM
MSI-6151 Mobo (Intel BX chipset)
gcc version 2.95.3 20010315 (release)
GNU binutils version 2.11.92.0.7
If you need more informations just ask
Bye
Hello!
On Wed, Jan 30, 2002 at 04:39:51PM +0100, Sebastian Dr?ge wrote:
> > BTW, you are running on a IDE system, right?
> Yes.
Ok, I see.
I can reproduce this problem on IDE only.
Our SCSI boxes run without a single problem.
Hm, may be this is IDE corruption thing, Andre Hendrick spoke about,
or was it fixed already?
I am looking into it anyway.
Bye,
Oleg
On Wed, Jan 30, 2002 at 07:09:05PM +0300, Oleg Drokin wrote:
> I can reproduce this problem on IDE only.
> Hm, may be this is IDE corruption thing, Andre Hendrick spoke about,
> or was it fixed already?
> I am looking into it anyway.
There were no IDE changes in my tree recently, and its strange
that this only shows up in reiserfs since the new set of patches
went in. I've no reports from users of other filesystems with any
problems, so I'm suspecting a rogue change in your last update.
Finding a common factor seems tricky, as it works flawlessly here
on IDE [*], but dies instantly for others.
[*] Even with stress tools.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
Hello!
On Wed, Jan 30, 2002 at 05:40:11PM +0100, Dave Jones wrote:
> > I can reproduce this problem on IDE only.
> > Hm, may be this is IDE corruption thing, Andre Hendrick spoke about,
> > or was it fixed already?
> > I am looking into it anyway.
> There were no IDE changes in my tree recently, and its strange
> that this only shows up in reiserfs since the new set of patches
> went in. I've no reports from users of other filesystems with any
> problems, so I'm suspecting a rogue change in your last update.
> Finding a common factor seems tricky, as it works flawlessly here
> on IDE [*], but dies instantly for others.
You do not play with a hdparm in your boot scripts, do you?
I do (will retry without this now).
How about others?
Bye,
Oleg
On Wed, Jan 30, 2002 at 07:44:08PM +0300, Oleg Drokin wrote:
> You do not play with a hdparm in your boot scripts, do you?
> I do (will retry without this now).
> How about others?
No, my testboxes autoconfigure with the right settings.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
Hello!
On Wed, Jan 30, 2002 at 05:55:23PM +0100, Dave Jones wrote:
> > You do not play with a hdparm in your boot scripts, do you?
> > I do (will retry without this now).
> > How about others?
> No, my testboxes autoconfigure with the right settings.
In fact I am able to reproduce with just bare booting into /bin/bash
remounting reiserfs into rw mode and do depmod -a
Hmmm. Interesting thing is may be only those who have reiserfs as root are
affected? Going to check it now (though scsi system with reiserfs root still
looks fine running stressetsts).
Bye,
Oleg
On Wed, 30 Jan 2002 17:40:11 +0100
Dave Jones <[email protected]> wrote:
> On Wed, Jan 30, 2002 at 07:09:05PM +0300, Oleg Drokin wrote:
>
> > I can reproduce this problem on IDE only.
> > Hm, may be this is IDE corruption thing, Andre Hendrick spoke about,
> > or was it fixed already?
> > I am looking into it anyway.
>
> There were no IDE changes in my tree recently, and its strange
> that this only shows up in reiserfs since the new set of patches
> went in. I've no reports from users of other filesystems with any
> problems, so I'm suspecting a rogue change in your last update.
>
> Finding a common factor seems tricky, as it works flawlessly here
> on IDE [*], but dies instantly for others.
OK here are some facts ;)
I run hdparm but only after syslogd... so hdparm isn't the bad boy
I have 3 partitions. Two reiserfs partitions, one mounted on /, one on /home
One ext2 /boot partition
The oops happens in 3 of 4 tests on syslogd. Only devfsd is loaded before it.
The other time it happens when loading proftpd. This is one of the last processes I started by the boot scripts
It happens with the IDE layer version as in the dj tree and with acb-io-2.5.3-p2.01212002 update (why haven't you included this in your tree, Dave?)
I have really no idea the oops comes from
Bye
On Wed, Jan 30, 2002 at 08:10:54PM +0100, Sebastian Dr?ge wrote:
> I have 3 partitions. Two reiserfs partitions, one mounted on /, one on /home
Ok, my test box for reiserfs uses ext3 root, and reiser on a scratch disk.
It could be Oleg's earlier guess that it may be reiser-on-root related.
> It happens with the IDE layer version as in the dj tree and with
> acb-io-2.5.3-p2.01212002 update (why haven't you included this in your tree,
> Dave?)
I never saw Andre pushing it on Linux-kernel (which is unusual for Andre 8)
If it didn't reach my inbox, it didn't happen in my world.
Its possible I overlooked the thread it was mentioned, or it got
lost in noise.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
Hello!
On Wed, Jan 30, 2002 at 08:17:57PM +0100, Dave Jones wrote:
> On Wed, Jan 30, 2002 at 08:10:54PM +0100, Sebastian Dr?ge wrote:
> > I have 3 partitions. Two reiserfs partitions, one mounted on /, one on /home
> Ok, my test box for reiserfs uses ext3 root, and reiser on a scratch disk.
> It could be Oleg's earlier guess that it may be reiser-on-root related.
No. I can reproduce with ext2 root & reiserfs mounted separately.
depmod -a -b /mnt causes a crash for me 100% of time.
(of course you need correct /lib/modules/... for that)
> > It happens with the IDE layer version as in the dj tree and with
> > acb-io-2.5.3-p2.01212002 update (why haven't you included this in your tree,
> > Dave?)
> I never saw Andre pushing it on Linux-kernel (which is unusual for Andre 8)
I dug the original message,
it's subject is "DO NOT USE IT (Re: linux-2.5.3-pre1 and IDE Driver Trouble) FATAL"
msgid is <[email protected]>
It warns about 2.5.3-pre1 and 2.5.2 is unstable with IDE because there are
conflicts between BIO & ACB. He promised to come up with the fix later.
After that there were no messages from him with patches.
At least not in the lkml.
Bye,
Oleg
Hello!
On Wed, Jan 30, 2002 at 08:17:57PM +0100, Dave Jones wrote:
Ok, as of now, I tried vanilla 2.5.3 and it works.
2.5.2-dj7 breaks instantly on the first truncate call to reiserfs.
I tried to dig up the difference between these 2 kernels but have not found
anything that will change that behaviour yet. And resierfs code is identical.
But dj7 seems to have a lot of modifications in the mm/* and fs/* stuff
compared to 2.5.3
Bye,
Oleg
Hi!
Oleg Drokin wrote:
> Ok, as of now, I tried vanilla 2.5.3 and it works.
> 2.5.2-dj7 breaks instantly on the first truncate call to reiserfs.
> I tried to dig up the difference between these 2 kernels but have not
> found anything that will change that behaviour yet. And resierfs code is
> identical.
> But dj7 seems to have a lot of modifications in the mm/* and fs/* stuff
> compared to 2.5.3
I have exactly the same problems you mentioned earlier in this thread. I
get the Ooops at various steps in the boot process. Sometimes the system
hangs directly after depmod, sometimes it can calculate the dependencies
and freezes when loading the first module (here: vfat.o)
This happens with 2.5.3 on a system with an IDE harddisk and root fs on
reiserfs.
bye
Martin
--
Martin Bahlinger <[email protected]> (PGP-ID: 0x98C32AC5)
On Thu, Jan 31, 2002 at 12:24:24PM +0300, Oleg Drokin wrote:
> Ok, as of now, I tried vanilla 2.5.3 and it works.
That's something I had hoped wouldn't be the case.
> 2.5.2-dj7 breaks instantly on the first truncate call to reiserfs.
> I tried to dig up the difference between these 2 kernels but have not found
> anything that will change that behaviour yet. And resierfs code is identical.
> But dj7 seems to have a lot of modifications in the mm/* and fs/* stuff
> compared to 2.5.3
One possible is that I've goofed whilst merging Andrew Mortons
"out of disk space during truncate" fixes from 2.4. Andrew, could
have a quick scan through the fs/ changes in -dj6 and see if anything
jumps out at you ?
I'll take a look myself later too, but right now, it's a head-scratcher.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
Hello!
On Thu, Jan 31, 2002 at 01:49:31PM +0100, Dave Jones wrote:
> > 2.5.2-dj7 breaks instantly on the first truncate call to reiserfs.
> > I tried to dig up the difference between these 2 kernels but have not found
> > anything that will change that behaviour yet. And resierfs code is identical.
> > But dj7 seems to have a lot of modifications in the mm/* and fs/* stuff
> > compared to 2.5.3
> One possible is that I've goofed whilst merging Andrew Mortons
> "out of disk space during truncate" fixes from 2.4. Andrew, could
> have a quick scan through the fs/ changes in -dj6 and see if anything
> jumps out at you ?
Hm, but I remember dj6 was reported as "working"?
> I'll take a look myself later too, but right now, it's a head-scratcher.
Do you have some place where one can see all separate patches 2.5.2-dj7
consist of?
Bye,
Oleg
Hello!
On Thu, Jan 31, 2002 at 11:10:28AM +0100, Martin Bahlinger wrote:
> I have exactly the same problems you mentioned earlier in this thread. I
> get the Ooops at various steps in the boot process. Sometimes the system
> hangs directly after depmod, sometimes it can calculate the dependencies
> and freezes when loading the first module (here: vfat.o)
> This happens with 2.5.3 on a system with an IDE harddisk and root fs on
> reiserfs.
Hm, weird.
I will try to reproduce further.
your oops is prepended with PAP-5760 message, right?
Bye,
Oleg
On Thu, Jan 31, 2002 at 03:53:25PM +0300, Oleg Drokin wrote:
> > One possible is that I've goofed whilst merging Andrew Mortons
> > "out of disk space during truncate" fixes from 2.4. Andrew, could
> > have a quick scan through the fs/ changes in -dj6 and see if anything
> > jumps out at you ?
> Hm, but I remember dj6 was reported as "working"?
-ENOTAWAKEYET. Yes, I meant -dj7 there. oops.
> > I'll take a look myself later too, but right now, it's a head-scratcher.
> Do you have some place where one can see all separate patches 2.5.2-dj7
> consist of?
I keep a directory named after each revision, patches get dumped there
as I merge them, and also I dump bits of ~/Mail/merged there.
These patches however are not always identical to what gets merged,
as they sometimes need to be bent into shape to live with something
else that got merged, I hand-edit the diffs to make them apply,
sometimes removing complete hunks, and grafting that hunks intention
into place in the source instead.
So they're of little use to anyone in their current state, given that
there are 'holes' where I chopped bits out. Sometimes I rediff the
patch afterwards, but not always.
--
| Dave Jones. http://www.codemonkey.org.uk
| SuSE Labs
Hello!
On Thu, Jan 31, 2002 at 02:11:01PM +0100, Dave Jones wrote:
Ok, I think we got it. And yes it it was reiserfs fault.
What I really cannot understand is how it was working before???
Ok, so anybody who sees the oopses should try 2 patches attached.
prealloc_init_list_head.diff is just forgotten initialisation
and pick_correct_key_version.diff is the real fix.
I wonder is anybody will be able to reproduce a bug with these 2 fixes
(I hope not).
Chris: Can you also take a look?
Bye,
Oleg
On Thu, 31 Jan 2002 19:44:01 +0300
Oleg Drokin <[email protected]> wrote:
> Hello!
>
> On Thu, Jan 31, 2002 at 02:11:01PM +0100, Dave Jones wrote:
>
> Ok, I think we got it. And yes it it was reiserfs fault.
> What I really cannot understand is how it was working before???
>
> Ok, so anybody who sees the oopses should try 2 patches attached.
> prealloc_init_list_head.diff is just forgotten initialisation
> and pick_correct_key_version.diff is the real fix.
>
> I wonder is anybody will be able to reproduce a bug with these 2 fixes
> (I hope not).
Hi,
everything seems to work perfect :)
my system has booted without any problems
Thanks and Bye
Hi!
On Thu, 2002-01-31 at 14:03, Oleg Drokin wrote:
> Hello!
>
> On Thu, Jan 31, 2002 at 11:10:28AM +0100, Martin Bahlinger wrote:
>
> > I have exactly the same problems you mentioned earlier in this thread. I
> > get the Ooops at various steps in the boot process. Sometimes the system
> > hangs directly after depmod, sometimes it can calculate the dependencies
> > and freezes when loading the first module (here: vfat.o)
> > This happens with 2.5.3 on a system with an IDE harddisk and root fs on
> > reiserfs.
> Hm, weird.
> I will try to reproduce further.
> your oops is prepended with PAP-5760 message, right?
No. It's PAP-14030 here.
--
Martin Bahlinger <[email protected]> (PGP-ID: 0x98C32AC5)
Hi!
Sebastian Dr?ge wrote:
> Oleg Drokin <[email protected]> wrote:
> > Ok, I think we got it. And yes it it was reiserfs fault.
> > What I really cannot understand is how it was working before???
> > Ok, so anybody who sees the oopses should try 2 patches attached.
> > prealloc_init_list_head.diff is just forgotten initialisation
> > and pick_correct_key_version.diff is the real fix.
> > I wonder is anybody will be able to reproduce a bug with these 2
> > fixes
> > (I hope not).
> everything seems to work perfect :)
> my system has booted without any problems
After applying those patches to 2.5.3 I still got an Oops after a
PAP-14030 message. I will try to catch the Oops (have never done this
before, may take some time) and feed it to ksymoops.
bye
Martin
--
Martin Bahlinger <[email protected]> (PGP-ID: 0x98C32AC5)
On Thursday, January 31, 2002 07:44:01 PM +0300 Oleg Drokin <[email protected]> wrote:
> Hello!
>
> On Thu, Jan 31, 2002 at 02:11:01PM +0100, Dave Jones wrote:
>
> Ok, I think we got it. And yes it it was reiserfs fault.
> What I really cannot understand is how it was working before???
>
> Ok, so anybody who sees the oopses should try 2 patches attached.
> prealloc_init_list_head.diff is just forgotten initialisation
> and pick_correct_key_version.diff is the real fix.
>
> I wonder is anybody will be able to reproduce a bug with these 2 fixes
> (I hope not).
>
> Chris: Can you also take a look?
Both fixes look right. I'm a little worried about
pick_correct_key_version.diff, that bug looks like it could get the keys
into the tree in the wrong order. We need to reproduce it with an
unpatched kernel, and then apply the fix to make sure the sky doesn't
fall on us.
-chris
On 31 Jan 2002, Martin Bahlinger wrote:
> After applying those patches to 2.5.3 I still got an Oops after a
> PAP-14030 message. I will try to catch the Oops (have never done this
> before, may take some time) and feed it to ksymoops.
I actually had PAP-5760. And after applying the patches it was the
PAP-14030. During all the tests today my reiserfs got currupted. A
reiserfsck ran into a segfault when checking the semantic tree. And this
happened exactly while checking /var/log/ksymoops/20020131.log ;-) After
deleting this file reiserfsck did it's job and my 2.5.3 works now.
bye and thanks for the fix,
Martin
--
Martin Bahlinger <[email protected]> (PGP-ID: 0x98C32AC5)
Hello!
On Thu, Jan 31, 2002 at 09:44:32PM +0100, Martin Bahlinger wrote:
> On 31 Jan 2002, Martin Bahlinger wrote:
> > After applying those patches to 2.5.3 I still got an Oops after a
> > PAP-14030 message. I will try to catch the Oops (have never done this
> > before, may take some time) and feed it to ksymoops.
> I actually had PAP-5760. And after applying the patches it was the
> PAP-14030. During all the tests today my reiserfs got currupted. A
You are the only who reporting these errors for now.
Can you reboot into vanilla 2.5.3 and capture PAP-5760 oops
and all reiserfs-specific output around it? Thank you.
> reiserfsck ran into a segfault when checking the semantic tree. And this
This means you need updated reiserfsprogs.
Bye,
Oleg
Hi!
On Fri, 2002-02-01 at 06:55, Oleg Drokin wrote:
> You are the only who reporting these errors for now.
> Can you reboot into vanilla 2.5.3 and capture PAP-5760 oops
> and all reiserfs-specific output around it? Thank you.
Because I don't have my first build of 2.5.3 any more, I had to compile
a new one without any patches applied to it. But now I get a completely
different behaviour: I can boot without problems and get those PAP-5580
oops after a few minutes uptime. I attached the oops message and the
output of ksymoops at the end of this mail.
I think, the other oops (PAP-14030) happened because of those filesystem
corruptions I mentioned before. Maybe the fs corrupted during the first
PAP-5580 oops, maybe it has been so for some days or even weeks. I just
don't know.
> > reiserfsck ran into a segfault when checking the semantic tree. And this
>
> This means you need updated reiserfsprogs.
I use reiserfsprogs 3.x.0j as recommended by linux/Documentation/Changes
bye
Martin
-----------------------------------------------------------------------------
vs-500: unknown uniqueness 536870912
vs-500: unknown uniqueness 536870912
vs-500: unknown uniqueness 268435456
vs-500: unknown uniqueness 268435456
===================================================================
LEAF NODE (28440) contains level=1, nr_items=6, free_space=2384 rdkey
-------------------------------------------------------------------------------
|##| type | key | ilen | free_space | version |
loc |
-------------------------------------------------------------------------------
| 1| *3.6* [67650 1815 0x1 DIRECT], item_len 600, item_location 3452,
free_space(entry_count) 65535 |
"^_^^U<^B^Cmainlog.0VKj^\A^P^Da=OQ^W^PTfu=^F<t
|^B^C4pFXB:?g^L^^\^Y^?Gh^?t^DU^Uc%xS1'oESC^WK^Y^Wb[ww^_oo[o^O^W^O^Wmmcu2}zy|5;w/^Nz[k36cs?x=^Tz^R<^Zb
N4^^^?q(^N^Xr3mu<7^UD^TESChU^O^^PSn"~~=S^\1df*z^]^Sn""_Y1e~OgVd ^^{@L9
*~[CL9
*~WEL9
~W1D^PESC))~^WUD^TESC(hw9!&\Do^UyIneL9 *~w
^Xr^CuYosD^TESC(j7ESCbJMT^\_5*bJMTu;^]1e^F*__5
bJMTu;^VbLTtk%^LT^P^\-I^Wg4^N^X^W.R5^\GgJDK)<q#ND8tG*o\X;bJMT^1&bJMT-Fq
"U^K<'bJT_8c<^X:q *ESCg^E1.Q$jE3
^Xs#U/^ZwD8pG:K~kB^L^K^Wi-2^?TB]qc^E;RUoqa"U}v^B^X^WnH[_J|S^KwG^O^WnJUy^]^F^X^W.Rzi^E;cG^Kw%*_5cnvu~-^X!F^EKV4|U^K5'G+W9n(19Y&>9lL?/9Y!{^N'~(^^U=[^Y()9[qo]lx_^?.19YU3m5&^G;eOW ^Z^R3=5WO^RC/|'19Yu^^]^WjN^NVu<{^N^Z^R#]2g=:jLNV^]g=<7K^\glUOm1&dlUO}^_()yf^?vK|l^T^T"
-------------------------------------------------------------------------------
| 2| *3.6* [67650 2233 0x0 SD], item_len 44, item_location 3408,
free_space(entry_count) 65535 |
mode | size | nlinks | first direct | mtime
0100640 | 808 | 1 | 785786 | 1012598582
===================================================================
PAP-5580: reiserfs_cut_from_item: item to convert does not exist ([67650
2233 0x1 IND])invalid operand: 0000
CPU: 0
EIP: 0010:[<c0167549>] Not tainted
EFLAGS: 00010282
eax: 0000005a ebx: c0224e80 ecx: 00000001 edx: df256000
esi: c1971c00 edi: 00000000 ebp: d2507e20 esp: d2507c24
ds: 0018 es: 0018 ss: 0018
Process exim (pid: 605, stackpage=d2507000)
Stack: c02239fa c02aa0a0 c0224e80 d2507c48 d2507c80 00000000 c016e85c
c1971c00
c0224e80 d2507e60 d5504cc0 00000003 00000001 00000003 00001000
00000000
00000001 d2507c84 00000cd8 00000001 00000cd8 c1971c00 63000000
00000000
Call Trace: [<c016e85c>] [<c016ee59>] [<c0160506>] [<c01612fb>]
[<c013113c>]
[<c013017c>] [<c0117998>] [<c0117f8e>] [<c011813e>] [<c0108837>]
Code: 0f 0b 68 a0 a0 2a c0 b8 00 3a 22 c0 8d 96 cc 00 00 00 85 f6
-----------------------------------------------------------------------------
ksymoops 2.4.3 on i686 2.5.3. Options used
-V (default)
-k /var/log/ksymoops/20020201222518.ksyms (specified)
-l /var/log/ksymoops/20020201222518.modules (specified)
-o /lib/modules/2.5.3/ (default)
-m /usr/src/linux/System.map (specified)
8139too Fast Ethernet driver 0.9.22
ac97_codec: AC97 Audio codec, id: 0x414c:0x4326 (Unknown)
0100640 | 808 | 1 | 785786 | 1012598582
CPU: 0
EIP: 0010:[<c0167549>] Not tainted
Using defaults from ksymoops -t elf32-i386 -a i386
EFLAGS: 00010282
eax: 0000005a ebx: c0224e80 ecx: 00000001 edx: df256000
esi: c1971c00 edi: 00000000 ebp: d2507e20 esp: d2507c24
ds: 0018 es: 0018 ss: 0018
Process exim (pid: 605, stackpage=d2507000)
Stack: c02239fa c02aa0a0 c0224e80 d2507c48 d2507c80 00000000 c016e85c
c1971c00
c0224e80 d2507e60 d5504cc0 00000003 00000001 00000003 00001000
00000000
00000001 d2507c84 00000cd8 00000001 00000cd8 c1971c00 63000000
00000000
Call Trace: [<c016e85c>] [<c016ee59>] [<c0160506>] [<c01612fb>]
[<c013113c>]
[<c013017c>] [<c0117998>] [<c0117f8e>] [<c011813e>] [<c0108837>]
Code: 0f 0b 68 a0 a0 2a c0 b8 00 3a 22 c0 8d 96 cc 00 00 00 85 f6
>>EIP; c0167548 <reiserfs_panic+28/50> <=====
Trace; c016e85c <reiserfs_cut_from_item+1ac/450>
Trace; c016ee58 <reiserfs_do_truncate+308/440>
Trace; c0160506 <reiserfs_truncate_file+c6/160>
Trace; c01612fa <reiserfs_file_release+31a/340>
Trace; c013113c <fput+4c/e0>
Trace; c013017c <filp_close+5c/70>
Trace; c0117998 <put_files_struct+58/c0>
Trace; c0117f8e <do_exit+ae/230>
Trace; c011813e <sys_exit+e/10>
Trace; c0108836 <syscall_traced+6/a>
Code; c0167548 <reiserfs_panic+28/50>
00000000 <_EIP>:
Code; c0167548 <reiserfs_panic+28/50> <=====
0: 0f 0b ud2a <=====
Code; c016754a <reiserfs_panic+2a/50>
2: 68 a0 a0 2a c0 push $0xc02aa0a0
Code; c016754e <reiserfs_panic+2e/50>
7: b8 00 3a 22 c0 mov $0xc0223a00,%eax
Code; c0167554 <reiserfs_panic+34/50>
c: 8d 96 cc 00 00 00 lea 0xcc(%esi),%edx
Code; c016755a <reiserfs_panic+3a/50>
12: 85 f6 test %esi,%esi
-----------------------------------------------------------------------------
--
Martin Bahlinger <[email protected]> (PGP-ID: 0x98C32AC5)
Martin Bahlinger <[email protected]> wrote:
MB> Because I don't have my first build of 2.5.3 any more, I had to compile
MB> a new one without any patches applied to it. But now I get a completely
MB> different behaviour: I can boot without problems and get those PAP-5580
MB> oops after a few minutes uptime. I attached the oops message and the
MB> output of ksymoops at the end of this mail.
Ok, thank you. Though this one is already fixed.
MB> I think, the other oops (PAP-14030) happened because of those filesystem
MB> corruptions I mentioned before. Maybe the fs corrupted during the first
Sad, you cannot reproduce it anymore.
>> > reiserfsck ran into a segfault when checking the semantic tree. And this
>> This means you need updated reiserfsprogs.
MB> I use reiserfsprogs 3.x.0j as recommended by linux/Documentation/Changes
We will eventually change that to something more recent.
Bye,
Oleg