SMP + SCSI + MD + 2.0.x bad?

Johannes Erdfelt (
Mon, 17 Nov 1997 13:19:25 -0500

First a little background about my system:

Dual PPro 200
Tyan Titan Pro
128meg of EDO memory
Adaptec 2940UW
3 4.3gig Wide Seagates
1 1gig Narrow Seagate

The machine has been very unstable. It crashes/reboots on the average of
once every other day. I boot off the 1gig drive and then use MD to make
3 raid0 partitions off of the 3 other drives (/usr, /var and /home).

The first time the machine crashed was when a friend was ssh'd to
another box when a couple of SCSI errors popped up, an oops and then
this scrolled forever until a reboot:

scsi0: Warning!! Interrupt routine called reentrantly!

Upon rebooting, nothing in the logs (suprise). The other 4 times have
been reboots when I wasn't near the computer. Nothing in the logs any of
the times.

Between the error message above and the lack of anything in the logs
it looks like there's something wrong with SCSI or MD. I think I
remember seeing a reference to SMP not mixing with MD at all in the 2.0

I've tried 2.0.29, 2.0.31 and 2.0.32-pre2. I'm about to try the latest
2.0.32-pre but I don't have much hope for it fixing anything.

Are there known problems with SMP + MD?
Will they be fixed in 2.0.x?
Am I better off running a 2.1.x kernel?
Am I leaving off any information?