Hang after adding swap in 2.4.19-2.4.22

Hang after adding swap in 2.4.19-2.4.22

[Lev A. Melnikovsky]

Hi there!

I have this problem quite long already, I first noticed it when switched
from 2.4.18 to 2.4.19 and it is still here with 2.4.22 (have not tried
2.4.23 yet). The system "hangs" during the early stages of initscripts
execution. In my case these are RedHat initscripts (several versions
tried) but other people reported similar problems with Mandrake (look for
the words like "hang finding module dependencies" or "hang mounting local
filesystems"). There was a similar post to the LKML (Brad Tilley, Jun 23
2003, "OS Fails to Load"). Unfortunately the problem is not 100%
reproduceable in the sense the system "hangs" some 10%-50% (your mileage
may vary) of times. On the other side the problem was seen here at three
separate computers with quite different hardware (well, all CPUs are i386,
but they are Celeron 533, Athlon XP 2200+ and XP 1700+).

I have collected all possible stack traces, PCs, memory info etc. The
bunch is available from http://kapitza.ras.ru/~leva/ops.tar.bz2

[leva@ari ops]$ tar -tjf ops.tar.bz2
addresses.txt
oooooops.txt
ops-improved.txt
readme
.config

If you ask, yes, it does include ksymoops (ops-improved.txt) and addr2line
(addresses.txt) outputs. And the kernel is just a 2.4.22 with a patch from
http://w.ods.org/tools/kmsgdump/0.4.4/patch-2.4.23p6-kmsgdump-0.4.4.gz
just to store the dump after the hang.

I would really appreciate if someone could tell me what the problem is
and, if this is my own misconfiguration, how I should avoid it. Otherwise,
if this is a bug, some fix would be available, probably... I will be glad
to supply any additional information if needed.

If replying, please CC me, as I am not subscribed. Thanks in advance
-L.

+2.4.23 Re: Hang after adding swap in 2.4.19-2.4.22

[Lev A. Melnikovsky]

  On Wed, 3 Dec 2003 at 3:37pm, Lev A. Melnikovsky wrote:

> I have this problem quite long already, I first noticed it when switched
> from 2.4.18 to 2.4.19 and it is still here with 2.4.22 (have not tried
> 2.4.23 yet). The system "hangs" during the early stages of initscripts
> execution. In my case these are RedHat initscripts (several versions
> tried) but other people reported similar problems with Mandrake (look for
> the words like "hang finding module dependencies" or "hang mounting local
> filesystems"). There was a similar post to the LKML (Brad Tilley, Jun 23
> 2003, "OS Fails to Load"). Unfortunately the problem is not 100%
> reproduceable in the sense the system "hangs" some 10%-50% (your mileage
> may vary) of times. On the other side the problem was seen here at three
> separate computers with quite different hardware (well, all CPUs are i386,
> but they are Celeron 533, Athlon XP 2200+ and XP 1700+).
have to reply to my own message since got no answer to the original post.
The problem is already reproduced with 2.4.23. I have managed to catch the
"hang" in two slightly different situations, both kernel message logs
(along with ksymoops and addr2line outputs) are available at
http://kapitza.ras.ru/~leva/ops.2.4.23.tar.bz2

The things common to all the computers involved in my experiments are
reiserfs root filesystem and devfs (see .config)

The kernel used is plain vanilla 2.4.23 with a patch from
http://w.ods.org/tools/kmsgdump/0.4.4/patch-2.4.23p6-kmsgdump-0.4.4.gz

As I have written earlier, the problem occurs with different computers and
I doubt that the hardware details may be important. Still, here's info for
my desktop where the logs were obtained (actually this stuff can be found
in the top of the log):

[leva@desk leva]$ cat /proc/cpuinfo
processor       : 0
vendor_id       : AuthenticAMD
cpu family      : 6
model           : 8
model name      : AMD Athlon(tm) XP 1700+
stepping        : 1
cpu MHz         : 1470.012
cache size      : 256 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 1
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 mmx fxsr sse syscall mmxext 3dnowext 3dnow
bogomips        : 2936.01

Not overclocked, no vmware (mostly no modules at all - the hang usually
happens quite early in the boot), no nothing. And, by the way, Seagate
harddrives in all boxes, if this matters.

As before, if replying, please CC me, as I am not subscribed.
I will appreciate any help
-L.