LOR or panic with ips driver (was Re: cvs commit: src/sys/dev/ips ips.c ips.h ips_commands.c)

Hello.

On 2003/11/27 00:37:36, Martin Blapp wrote:
> mbr         2003/11/27 00:37:36 PST
> 
>   FreeBSD src repository
> 
>   Modified files:
>     sys/dev/ips          ips.c ips.h ips_commands.c 
>   Log:
>   ServeRaid (at least 5i) didn't initialize correctly. To get
>   them working (cache, automatic rebuild and hotswap) the FFDC
>   info (First Failure Data Capture) on the adapter must be
>   initialised.
>   
>   Logical drives in critical/degraded states weren't added to
>   the drive list. FreeBSD was not able to see a degraded array
>   after a reboot. Degraded drives are now also added to the drivelist
>   and the state of the logical drive is given at boottime.
>   
>   The adapter type is detected from informations in nvram page 5
>   and displayed at boottime.
>   
>   Change IPS_OS_FREEBSD definition from 10 to 8 according to IBM
>   specs.
>   
>   Submitted by:   <Patrick Guelat> pgfb_at_imp.ch
>   Reviewed by:    mbr, scottl
>   Approved by:    re
>   
>   Revision  Changes    Path
>   1.6       +64 -2     src/sys/dev/ips/ips.c
>   1.5       +57 -2     src/sys/dev/ips/ips.h
>   1.7       +90 -0     src/sys/dev/ips/ips_commands.c

After seeing this commit message, I've upgraded -CURRENT installed on
NetFinity 6000R(with ServeRAID 4H) to catch up with the recent changes.
However, LOR or panic still persists. You can panic your -CURRENT machine
by creating a log of files in a directory. You can do it with a simple
shell script like this:

$ mkdir foo; i=0; while :; do echo $i > foo/$i; i=$(($i+1)); done

- If you have WITNESS_* options in the kernel config file, you'll see
  a lot of LOR messages on the console or in /var/log/messages, along with
  a bunch of `iobuf error 5', but the kernel doesn't panic(or takes very
  long time before it panics).
- If you don't have WITNESS_* options in the kernel config file and run
  the code above, the kernel panics in a few minutes after seeing
  `iobuf error 5' messages.

I've already submitted a PR for this as kern/57832, but still have no
idea how to fix this. The traceback looks almost the same as it was
a few months ago.

Regards.
-- 
YONETANI Tomokazu / Ergo-Brains Inc.