There are some known lockup conditions under the latest 2.1.x block driver
area. It's being worked on, but depending on how hard you push the system,
it could never lock up or lock up very frequently. My current code can lock
the system up within about an hour of heavy use. Or, if I change one line
in the driver so that my driver is under a spin lock at the time I call
cmd->done(cmd) on the completed SCSI commands, I can lock it up in seconds.
> So I downloaded a plain 2.0.33 and applied the original patch without
> any modifications, but just had the same results. BTW, the bus
> activity LED was constantly on and after rebooting with the big red
> button the Adaptec BIOS hung without identifying any of my devices.
>
> Power-cycling the machine just cured _that_ problem.
>
> I'm using an AHA-3940 with a pretty old BIOS version (1.12 I think).
It sounds like an interrupt problem. Aka, the bus activity LED was stuck on
because it was waiting for us to answer some interrupt we couldn't answer
because the machine was hung with interrupts off. In any case, I need more
description of the event in order to be able to track anything down.
--Doug Ledford <dledford@dialnet.net> Opinions expressed are my own, but they should be everybody's.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu