Am 02.02.2012 04:12, schrieb Ross Hamblin:
On 02/02/12 15:56, mn0 wrote:
IRQ fiddling is fun. Does anyone want to employ a
perfectly trained pci
card dis- and re-mounter?
I can also restore BIOSes with corrupted checksums...
I don't know what to do... one of the cards' interrupts is magically
attracted by hdd controllers, either it's ata and usb or it's ide.
BIOS options are totally pointless, yet changing something but in an
unforeseeable manner.
I could disable the IDE controller if there wasn't the os on an ide hdd.
This would also be a stupid approach.
I'll try again in a few hours...
/mn0
I probably missed something as I lost track with the split threads, but
did you check your interrupts in linux to see whether there are any free
IRQs? If there are free IRQs you could try using the BIOS option to
reassign the card to a free interrupt. If Linux is running in PIC
instead of APIC mode there will be way less interrupts to choose from -
if is APIC then cat /proc/interrrupts will show IRQs above 24 (IIRC)
otherwise you could look at enabling APIC which should allow the card to
take a higher and hopefully free IRQ.
Sorry for the noise if this was already covered.
HTH
Ross.
Not covered, yet.
APIC is enabled, but thanks for "cat /proc/interrrupts", makes it a
little easier.
There are indeed free interrupts/numbers not listed in /proc interrupts.
Only my BIOS settings are ignored when trying to assign certain devices
to IRQs. The BIOS options are also the wrong way: I can assign an IRQ#
to one or a group of devices... one of the devices is always grouped
with hdd. So I could assign a different IRQ# to the group, but don't
split it, to have the card on a single IRQ.
/mn0