> This is a followup of a post to the rt-linux mailing list. Daniele
> Lugli first spotted a problem with interupt handling in RT tasks under X
> and did some digging. The X server is apparently doing global cli/sti's.
certain (broken) videocards need to be probed very conservatively,
otherwise the hw might lock up (and/or get damaged). It _has_ to be done
that way.
i suppose there are other pieces of software too that use iopl(3) to do
cli/sti, i think U-Net?
i think RTLinux should disallow iopl(3) and ioperm(), user-code allowed to
manipulate hardware can lead to other, less obvious latency
problems/hw lockups (scanpci comes to mind, etc.).
alternatively, a /proc/sys/kernel/user-io control can be provided too,
thus the box can be configured through X, but can be made safe when it
goes mission-critical...
-- mingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu