Re: PATCH/RFC: [kdump] fix APIC shutdown sequence

[Martin Wilck <martin.wilck@fujitsu-siemens.com>]

Vivek Goyal wrote:

> Did you also check IRR bits on LAPIC. May be some interrupt is already
> being served and your new interrupts has been queued on LAPIC and IRR bit
> on LAPIC is set?

That's it. Whenever the IO-APIC IRR bit is set, I see the LAPIC IRR bit set, too.
I never see any ISR bits set. Very rarely I see that IRQs are IRQ_PENDING or
IRQ_INPROGRESS, but that's apparently unrelated to the problem.

Actually, I often see APIC IRR bits set, but it only seems to matter for
IRQs coming from the secondary IO-APIC.

Unfortunately, just writng APIC_EOI in this situation (when an IRR bit is set)
has no effect.

I have tried to set the IRQ_DISABLED flag for all IRQs. From my understanding,
if I re-enable IRQs, that disables the actual IRQ handlers, but the ack()
function of the IRQ chip (which, in our case, sends EOI) is called anyway.
*That appears to work*, in the kdump kernel the IRR is cleared.

I'll run an overnight test with that technique tonight.

Martin

-- 
Martin Wilck
PRIMERGY System Software Engineer
FSC IP ESP DE6

Fujitsu Siemens Computers GmbH
Heinz-Nixdorf-Ring 1
33106 Paderborn
Germany

Tel:			++49 5251 8 15113
Fax:			++49 5251 8 20409
Email:			mailto:martin.wilck@fujitsu-siemens.com
Internet:		http://www.fujitsu-siemens.com
Company Details:	http://www.fujitsu-siemens.com/imprint.html