Option error
Peter Stuge
stuge-linux-pcmcia at cdy.org
Sun Oct 8 18:35:08 EDT 2006
On Sun, Oct 08, 2006 at 09:51:50PM +0200, Dariusz Dwornikowski wrote:
> > Software has to deal with the problem (checking which of several
> > IO devices actually yanked the IRQ line) but not all of the Linux
> > PCMCIA card drivers in 2.6 do this yet.
>
> yes, i am using yenta driver.
yenta is the socket driver. That's not the problem. The warning comes
from the serial_cs driver, which is the card driver for serial port
cards. (Which is what the option cards look like.)
> is there a even dirty workaround for this problem ? maybe is there a
> way to check which interrupts cause the collision ?
Well, if you can disable all other sources of interrupts on irq 11
the card will work fine. Most laptop designs just connect all devices
to the same interrupt however. :\
You can check how interrupts are connected in your system. This is
what mine looks like:
--8<--
$ cat /proc/interrupts
CPU0
0: 69131347 XT-PIC timer
1: 144940 XT-PIC i8042
2: 0 XT-PIC cascade
3: 2 XT-PIC Intel ICH6 Modem
5: 4137359 XT-PIC yenta, ehci_hcd:usb1, uhci_hcd:usb2,
uhci_hcd:usb5, radeon at pci:0000:01:00.0
7: 836937 XT-PIC Intel ICH6
8: 62683 XT-PIC rtc
9: 20771246 XT-PIC acpi
10: 8236 XT-PIC uhci_hcd:usb3
11: 4333622 XT-PIC uhci_hcd:usb4
12: 1059462 XT-PIC i8042
14: 91751 XT-PIC libata
15: 126 XT-PIC libata
NMI: 0
ERR: 0
-->8--
As you can see I have cardbus, USB and graphics sharing irq 5.
In order to change what goes where you'll have to fiddle with
interrupt settings in your BIOS, but it's not really possible to
change the fact that yenta and radeon are wired together in my case.
> i could then not use this hardware during using option
> card. It is crucial to make it work for me as I administrate
> high-availibility servers.
Just a quick reminder; Linux comes with no warranty. But please do
feel free to help fix the problem. I know I would appreciate it. :)
//Peter
More information about the linux-pcmcia
mailing list