Re: 2.6.22-rc: regression: no irda0 interface (2.6.21 was OK), smsc does not find chip

From: Andrey Borzenkov
Date: Sun Jun 10 2007 - 04:03:33 EST


On Thursday 07 June 2007, Bjorn Helgaas wrote:
> On Wednesday 06 June 2007 02:45:01 pm Andrey Borzenkov wrote:
> > I am beginning to doubt whether drier
> > works on my system at all (i.e. before PnP change); have to find time to
> > test.
>
> In 2.6.21, smsc-ircc2 at least found the device. So we definitely have
> a problem because in 2.6.22-rc, we don't find the device.
>
> What laptop do you have? Maybe I can find one to play with.
>
> I think Windows always calls _SRS, even if it doesn't need to change
> the current resource settings. Linux doesn't do that, and my HP nw8240
> firmware has a bug that leaves the IR device partly-enabled until _SRS
> is called. Maybe yours has a similar bug.
>
> Can you set CONFIG_ACPI_DEBUG=y, make it so smsc-ircc2 isn't loaded
> automatically, and try this (along with my previous patch to swap
> FIR and SIR):
>
> # dmesg -n 8
> # echo 0x200 > /sys/module/acpi/parameters/debug_level
> # echo disable > /sys/bus/pnp/devices/00:0a/resources
> # echo activate > /sys/bus/pnp/devices/00:0a/resources
> # modprobe smsc-ircc2
>
> If you try this, can you collect the whole dmesg log?
>

well, dmesg (/var/log/messages) is around 35MB :) Even bz2 compressed it is
over 700KB. I put it here:
http://dump.ru:80/loadfile.php?filename=fir_acpi_debug.bz2&id=379192

This is from /var/log/messages. I do not have serial console and cannot
capture output directly; I can only hope nothing is lost at this rate.

> It's also possible that some other driver like i2c is messing with
> the ISA/LPC bridge configuration and breaking the IR device config.
> Your lspnp output doesn't show any other ISA devices down there,
> but I know i2c likes to poke around and discover things that kinda,
> sorta work, but the BIOS didn't expose.
>
> Bjorn


Attachment: signature.asc
Description: This is a digitally signed message part.