On Mon, 14 Feb 2000 moyer@mclinux.com wrote:
>
> ==> Regarding SMP spin_lock...; Robert Dinse <nanook@eskimo.com> adds:
>
> nanook> I wish there were an option to recover from spin_lock deadlock
> nanook> conditions, when one times out, release the lock and get on with
> nanook> life or something. Yes, I realize the potential for file or memory
> nanook> corruption exists, but file corruption potential also exists when
> nanook> data cached and not yet written to disk is lost due to the need to
> nanook> hard-boot the machine when it hangs in spin_lock hell, and memory
> nanook> corruption might crash the machine, well it's going to lock up and
> nanook> die otherwise anyway.
>
> Well, I believe SGI has a lockmetering package out there, and it
> can be modified to do spin lock timeouts quite easily.
> Unfortunately, this adds a bit of overhead to the kernel.
>
> nanook> And if that can't be done, at the very least a panic and
> nanook> reboot would be preferable to infinite hang.
>
> Exactly. You may want to look into a watchdog card, or even the
> software watchdog timer that already exists in the kernel.
>
> Warmest Regards,
>
> Jeff Moyer
> Mission Critical Linux
> http://www.missioncriticallinux.com
On Sparc-32 at least; with Linux, neither the software watchdog nor the
hardware seem to result in a reboot after spin_lock hell lock-up.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Feb 15 2000 - 21:00:28 EST