> On Fri, 19 Sep 1997, Sachin Garg wrote:
>
> doesn't scale very much, because it seems that interrupts are done only on
> one processor. Linux 2.0/SMP sucks probably even more :(, as it has a
> global spin lock :(((((
>
> Andreas
>
I don't think the spin-lock hurts. Look what happens if you only have
one interrupting processor and a new interrupt occurs while another
handler is in progress. The ISR isn't executed until the previous one
has completed.
Now look what happens if a new interrupt occurs during the execution of
a previous ISR under Linux SMP. The new interrupt is handled by another
CPU after a few hundred CPU cycles. If there are no other interrupts
being handled, you don't waste any CPU cycles.
Cheers,
DJ
Richard B. Johnson
Analogic Corporation
Penguin : Linux version 2.1.55 on an i586 machine (66.15 BogoMips).
Warning : It's hard to stay on the trailing edge of technology.
Linux : Engineering tool
Windows : Child's computer interface written in BASIC.
Win-NT : Child's computer interface rewritten in C.