2002-07-26 15:42:20

by Aaron Gaudio

[permalink] [raw]
Subject: Problems assigning IRQs on Sony R505ECK laptop

Hi all. I'm having trouble with linux finding IRQs for several
devices on my Sony R505ECK laptop (just got it yesterday :)

I added the "pci=biosirq" and I've disabled the "PnP OS" in the
BIOS, however there is nothing in the BIOS that allows me to
manually allocate IRQs to devices, and the problem has not
gone away.

Attached to this is the output of dmesg, which shows the
devices that are not getting IRQs, and the output of /proc/pci,
which shows all the PCI devices that Linux detected. Any help
would be much appreciated.



The laptop came with Windows 2000, and looking through the devices
there, it appears that all the ones causing problems in Linux are
assigned IRQ 9. According to Windows, I have the following devices:

00:1d.0 - USB host controller (IRQ 9)
00:1d.1 - USB host controller (IRQ 9)
00:1d.2 - USB host controller (IRQ 9)

-- linux loads two usb drivers (both usb-uhci) with IRQ9, the third it
can't detect an IRQ for. I haven't tried any USB devices with
any of the ports yet

00:1f.5 - Yamaha AC-XG sound controller (IRQ 9) - this uses the i810
audio driver... sound will play, but is choppy (sox
says "Unable to set audio speed to 44100 (set to 48000)" and
i810_audio reports "drain_dac, dma timeout?") -- but I haven't
gotten around to setting up any module options yet

02:0b.0 - Texas Instruments CardBus controller (IRQ 9)
02:05.0 - Ricoh CardBus controller (IRQ 9) - in linux the Ricoh is
loaded with IRQ 3

-- i don't have any PC cards, so I haven't even tried setting up
the CardBus stuff

02:08.0 - Intel 82801CAM ICH3 ethernet controller (IRQ 9)- works fine in
Linux

00:1f.1 - IDE ATA/ATAPI controller... it appears to work fine in Linux,
which detects the UDMA(100) hard drive, but I haven't tried
any hdparm benchmarks

02:02.0 - Texas Instruments OHCI-compliant 1394 (firewire) adapter (IRQ 9)-
This device is one of the ones which Linux can't detect an IRQ for;
when I add "irq=9" to the module option, the module loads, but
I'm not sure what to do with it after that (I have the laptop
connected to the cradle, which uses the Firewire, and has a DVD/CD-RW
and floppy on it).

Orinoco MiniPCI Card (wireless LAN)- this uses "CardBus Slot 0" and
IRQ 9. When the module tries to load, it reports an IRQ conflict. Maybe
there's a way to load the module and tell it to grab a different IRQ?

COM1 (IRQ 4)
LPT1 (IRQ 7 DMA 3)
mouse (IRQ 12)
keyboard (IRQ 1)
Conexant-Ambit softmodem (IRQ 9)


--

Aaron Gaudio prothontar @ tarnation.dyndns.org
http://tarnation.dyndns.org/~aaron
----------------
"From fullness, aspect. From aspect, being. From being, emptiness."


Attachments:
(No filename) (0.00 B)
(No filename) (240.00 B)
Download all attachments

2002-07-27 10:44:28

by Daniel Egger

[permalink] [raw]
Subject: Re: Problems assigning IRQs on Sony R505ECK laptop

Am Fre, 2002-07-26 um 17.41 schrieb Aaron Gaudio:

> Hi all. I'm having trouble with linux finding IRQs for several
> devices on my Sony R505ECK laptop (just got it yesterday :)

I've seen similar troubles on a different Vaio (unfortunately I don't
have the model number handy). Basically what happens is that the
Local-APIC seemingly isn't initialised correctly, all bridges in the
system are treated transparently and all devices get the same IRQ (9)
assigned. The system works just fine but I guess not as efficient as
it could go because of the interrupt storm on IRQ 9. I can get get all
data from this machine if someone wants to debug this problem.

--
Servus,
Daniel


Attachments:
signature.asc (189.00 B)
Dies ist ein digital signierter Nachrichtenteil

2002-07-27 13:58:00

by Aaron Gaudio

[permalink] [raw]
Subject: Re: Problems assigning IRQs on Sony R505ECK laptop

And lo, upon Sat, Jul 27, 2002 at 12:48:23PM +0200, Daniel Egger spaketh thusly:
> Am Fre, 2002-07-26 um 17.41 schrieb Aaron Gaudio:
>
> > Hi all. I'm having trouble with linux finding IRQs for several
> > devices on my Sony R505ECK laptop (just got it yesterday :)
>
> I've seen similar troubles on a different Vaio (unfortunately I don't
> have the model number handy). Basically what happens is that the
> Local-APIC seemingly isn't initialised correctly, all bridges in the
> system are treated transparently and all devices get the same IRQ (9)
> assigned. The system works just fine but I guess not as efficient as
> it could go because of the interrupt storm on IRQ 9. I can get get all
> data from this machine if someone wants to debug this problem.
>

Unfortunately, it doesn't seem like everything is working just fine.
In Windows, yes, all the PCI devices get assigned IRQ 9.. maybe that's
not the most efficient way, but that does seem to work. In Linux, only
5 of the devices get the IRQ, and the others don't (firewire, one of the
USB controllers, one of the CardBus controllers (an internal one,
apparently, because the built-in Orinoco card won't load due to an
IRQ conflict with IRQ 9).

I'd like to figure out how to get these devices working, whether or
not they have to use IRQ 9... then I can worry about making it
as efficient as possible. :)


--

Aaron Gaudio prothontar @ tarnation.dyndns.org
http://tarnation.dyndns.org/~aaron
----------------
"From fullness, aspect. From aspect, being. From being, emptiness."


Attachments:
(No filename) (1.58 kB)
(No filename) (240.00 B)
Download all attachments