> On Wed, 28 Apr 1999, Ingo Molnar wrote:
>
> >
> > On Tue, 27 Apr 1999, Richard B. Johnson wrote:
> >
> > > Hello IO-APIC/SMP gurus,
> > >
> > > I have finally upgraded all my CPUs to linux 2.2.6. They all work
> > > except this one! This one refuses to use IRQ0 on both CPUs. The
> > > kernel declares that the timer is not connected to the APIC. However,
> > > Linux version 2.1.127 (the last version running on that machine)
> > > found and used the timer without any problem!
> > >
> > > Perhaps there still needs to be some work done here?
> > >
> > > CPU0 CPU1
> > > 0: 54716 0 XT-PIC timer
> >
> > i would not be worried too much about the timer interrupt. It only does
> > timekeeping. even in 2.1.127 it was not 'perfect' IO-APIC interrupt, but a
> > special 'ExtINT' interrupt, delivered to all CPUs. This proved to give
> > problems on certain boxes. Moving timer-list handling from IRQ0 to the
> > local per-CPU interrupt in 2.3 should solve this problem. (then IRQ0 will
> > only be used for timekeeping) [the timer IRQ does not execute the timer
> > lists, but it's the IRQ mostly likely to trigger timer_bh processing)
> >
> > if you want to experiment wether your system is able to distribute all
> > ExtINT interrupts to all CPUs, just change setup_ExtINT_pin() in
> > io_apic.c:
> >
> > - entry.dest.physical.physical_dest = boot_cpu_id;
> > + entry.dest.physical.physical_dest = 0xff;
> >
Okay, I did this. At least _some_ timer ticks are now handled by the
other CPU. Guess this is "non-symmetrical multiprocessing" ... <grin>
Cheers,
Dick Johnson
***** FILE SYSTEM WAS MODIFIED *****
Penguin : Linux version 2.2.6 on an i686 machine (400.59 BogoMips).
Warning : It's hard to remain at the trailing edge of technology.
-
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/