Red Hat Bugzilla – Bug 61045
"Irq routing conflict" reported with UP kernel on a 6450
Last modified: 2007-04-18 12:40:53 EDT
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:0.9.4) Gecko/20011019
Description of problem:
On a 6450 with hampton beta 2 (2.4.9-0.1up), w/ bios A10, 5GB ram, 2GB swap, and
Perc3/SC boot controller.
Dmesg reports Irq routing conflicts with megaraid boot controller and onboard
Occurs only on UP kernel. Smp and bigmem report no errors.
Attached is dmesg from UP kernel showing error and bigmem kernel showing no error.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Install hampton beta2 on a 6450 w/ Perc3SC as boot controller
2. Boot to UP kernel
3. Check dmesg for Irq routing conflict error
Actual Results: Irq routing conflict with UP kernel
Expected Results: No error
Created attachment 48287 [details]
Up kernel dmesg w/ Perc3/SC and onboard adaptec 7899
Created attachment 48288 [details]
Bigmem kernel dmesg w/ Perc3/SC and onboard adaptec 7899
Clay, need the output of 'dump_pirq' too.
Created attachment 48304 [details]
Up kernel dump_pirq
Created attachment 48305 [details]
Bigmem kernel dump_pirq
is it possible to get an lspci -vv from the up kernel too ?
Created attachment 48325 [details]
It's interesting that the BIOS (I think) has assigned the interrupts according
to what's found in the pirq table (so that matches), but values requested in
the ServerWorks router are different. How are those values originally set?
Not fixed in beta3
Is this one of the bugs that goes away with the BOOT kernel and
the "apic" boot option?
Any word on this, Clay?
Using kernel 2.4.18-0.16, and loading with the "apic" option, error message
The apic option is ignored by UP kernels since it is assumed that the buggy
bioses are all on SMP machines.
Reproduced in Hampton Gold.
Adding the "apic" option to the UP kernel is known to break
laptops, so that is not an option. We are able to install on
those laptops with the "apic" option available, but that's
different from normal operation.
Closing. Good in Milan beta 5