I'm running RedHat 6.0 on a PC164 with SRM 5.5-1 firmware. When I reboot or shutdown, all processes terminate as normal, and then "reset_for_srm" gets called in the kernel causing SRM to panic. I must then hit the reset button. This is obviously not normal behavior, as it renders rebooting impossible, especially remotely. This same machine also has NetBSD on it, with no such problem; so I know there's nothing wrong with my machine or firmware. I've tried both 2.2.5-16 and the 2.2.5-22 update and both behave the same. Is there some fix for this????? Using the ARC firmware is not an option. The "reset_for_srm" kernel code is obviously broken. Please help.
I've fixed the problem! Or at least, I can work around it. The problem is with the md (multiple disk) driver. Apparently one of its threads is still running when reset_for_srm is called in the shutdown process, confusing SRM into panicking. I've disabled the md driver altogether, since I'm not using any kind of RAID or drive concatenation. Now everything works fine. However, the md driver SHOULD get fixed, since this behavior is definitely broken.
Assigned to dledford
Delinquent record (old) Assuming that 6.2/7.0 fixes the problem