Re: IDE kernel traces in 2.6.9 - Bad IRQ ?

[Phil Dibowitz <phil@ipom.com>]

[-- Attachment #1: Type: text/plain, Size: 3060 bytes --]

Ericisko wrote:
> Hi,
> 
> I have similar problem. I have MSI Neo2-FIS2R board
> with 865PE chipset. I have 1 PATA hard drive connected
> to ICH5, IDE channel 1 and DVD and CD-RW connected to
> IDE channel 2. I use IDE driver piix, not libata. When
> IDE is in legacy mode (BIOS setup), everything works
> fine - each channel has its own IRQ. When I change
> mode to native, I get the message "Disabling IRQ" with
> trace stack showing 2 ide_intr handlers for IRQ (1 for
> each channel?) and both channels share the same IRQ.
> It looks like interrupt is not handled and it's level
> triggered so /proc/interrupts shows huge number of
> interrupts and soon IRQ is disabled.
> If ide-cd is compiled as module, everything works fine
> unless I load ide-cd module. Than I get the same
> errors.
> As that particular interrupt was shared also by USB
> and eth0, I built the minimal kernel so ide0 and ide1
> were the only devices using the interrupt - the same
> problem.
> I'm not at my computer right now so I can't send any
> additional info at the moment.

So I played with this a little more. And I think I understand better. 
With this BIOS there's a few possible combinations:

Enhanced: S-ATA
           P-ATA+S-ATA
           P-ATA

or

Compatability: Primary P-ATA+S-ATA
                Secondary P-ATA+S-ATA
                P-ATA Only

Basically, there are way more ATA chains on this mobo than others. Yuo 
have your SATA controller (1 chain, 2 drives), your PATA controller (2 
chains, 4 drives), and your RAID PATA controller (1 chain, 2 drives).

So here's what happens:

- I chose Enhanced P-ATA+S-ATA which meant that SATA drives came first, 
then PATA. This caused the kernel some confusion... not exactly sure 
why, but something to the effect of /dev/hda was first, but wasn't first 
(that chainw as listed as the 2nd chain in bios with nothing on the 
first). This causes traces
- I was able to fix it with Compatibilty P-ATA only which puts things in 
the right order
- I was *also* able to fix it, it turns out, with Enhanced S-ATA. Now 
that seems counter intuitive until you break down and read the manual 
that comes with the board. This basically says "if you have SATA, we 
ignore all PATA and just use SATA. If you don't have SATA, we ignore all 
SATA and just use PATA" -- that thus puts things in the right order 
since I don't have any SATA at this time.

Ther other compatiblity modes mix which Primaries you want to use and 
which you want to throw away and replace with SATA. The other enhances 
modes are exactly what they say.

So that's some more information based on expiramentation, reading the 
fine manual, and a bit of speculating... doesn't solve everything, but I 
hope it helps.

-- 
Phil Dibowitz                             phil@ipom.com
Freeware and Technical Pages              Insanity Palace of Metallica
http://www.phildev.net/                   http://www.ipom.com/

"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety."
  - Benjamin Franklin, 1759


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 256 bytes --]