That's a very interesting note. I'm skeptical of the mechanism for
failure tht you suggested at first because the RTL patch should
_reduce_ the worst case time between inta cycles

Linux
cli time pases sti interrupt arrives

Rtlinux
soft_cli  interrupt is acked and marked pending soft_sti interrupt simulated

But Michael did find a place where I was redirecting interrupts to 
a single processor which could really confuse things.
This is part of a not yet completed feature allowing RTL handlers to
ask that the irq only be deliverd to a specific processor.



On Thu, Aug 12, 1999 at 08:29:40PM -0400, nunca wrote:
> This message just appeared on the l-k mailing list:
> http://boudicca.tux.org/hypermail/linux-kernel/this-week/0553.html
> It covers what I think is happening in rtlinux.  I'm currently trying to
> figure out the rtlinux kernel patch enough to fix this.
> -- 
> Nolan Leake
> Section 8
> --- [rtl] ---
> To unsubscribe:
> echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
> echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
> ----
> For more information on Real-Time Linux see:
> http://www.rtlinux.org/~rtlinux/
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/

Reply via email to