lkml.org 
[lkml]   [1997]   [May]   [12]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectSMP
Okay. Latest status on testing the latest 2.1.37:

While taring one fs to another (about 1 gb).

panic VFS: Buffer LRU pointers corrupted.
panic VFS: LRU block list corrupted.
panic VFS: Free block list corrupted.

(all from fs/buffer.c)

This does NOT happen if the kernel is compiled without SMP.
The kernel uses aic7xxx SCSI driver.

This is repeatable. Sometimes I get a different panic as the first
message. I usually get two panics (two CPUs, two panics??).

Anyhow, I hacked up a copy of buffer.c and put save_flags() Cli(), restore
flags() about everywhere and it did not help. I was going to shotgun this,
hoping for a "fix", then remove some pairs until it broke. It didn't work.

The kernel seems to work okay under "normal" usage. SYSCTL even works. This
had been broken for awhile.

Question? When SMP is finally robust and stable, if there is a crash, shall
we expect two panics <grin>? Gotta halt both CPUs in panic. This might
be the reason why SMP crashes seem to be more prone to trashing file-systems.
The second CPU might be writing some bad stuff to the file-system before
it panics.

Cheers,
Dick Johnson
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Richard B. Johnson
Project Engineer
Analogic Corporation
Voice : (508) 977-3000 ext. 3754
Fax : (508) 532-6097
Modem : (508) 977-6870
Ftp : ftp@boneserver.analogic.com
Email : rjohnson@analogic.com, johnson@analogic.com
Penguin : Linux version 2.1.35 on an i586 machine (66.15 BogoMips).
Warning : I read unsolicited mail for $350.00 per hour. Supply billing address.
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-


\
 
 \ /
  Last update: 2005-03-22 13:39    [from the cache]
©2003-2011 Jasper Spaans. Advertise on this site