[lkml]   [2002]   [Aug]   [25]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [PATCH] 2.5.31 Summit NUMA patch with dynamic IRQ balancing
On Friday 23 August 2002 05:29 pm, Grover, Andrew wrote:
> > From: James Cleverdon []
> >
> > > This should be moved to acpi.h
> >
> > Will be, once I'm sure this is the right way to go. As
> > mentioned earlier, I'm
> > having ACPI problems that seem to imply ACPI isn't building
> > the full IRQ
> > table. In 2.4 we could let MPS do this. Maybe 2.5 will need
> > to revert to
> > that behavior.
> What happens when you use the FULL ACPI support? I suspect that you really
> do want the interpreter, in order to evaluate _PRTs properly.
> ISTR that the reason you are thinking that ACPI only is programming some of
> the ioapic entries is because whatever is printing them is looking at the
> mp_irqs array. Which is MPS specific. So ACPI doesn't bother filling it all
> in. :)
> Is that a bug? Should ACPI fill it in completely, or maybe not at all?
> Don't know. But it is strictly unnecessary.
> Regards -- Andy

Bingo! With full ACPI turned on, the system does indeed boot. The extra I/O
APIC entries are being programmed from the PRT.

(Call chain is: pci_acpi_init --> acpi_pci_irq_init --> mp_parse_prt -->

So, given that quite a number of our customers would like to run with
hyperthreading turned on, but do not want full ACPI, what is the right thing
to do in the HT-only case? Add extra code to process the PRT? Fall back on
MPS's IRQ records? Something else entirely?

James Cleverdon
IBM xSeries Linux Solutions
{jamesclv(Unix, preferred), cleverdj(Notes)} at us dot ibm dot com

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

 \ /
  Last update: 2005-03-22 13:22    [W:0.041 / U:1.740 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site