Re: Port of 3c575_cb to 2.3.43pre8

[Olaf Titz]

> IRQ9 usually works, but gives me the ``suprious interrupt on IRQ7'' error.
> IRQ7 with the new driver works the best, but still far from perfect.
>
> Is this a good indication that I should go see my vendor and eventually
> beat him up with the notebook, or could the spurious interrupt on IRQ7 be
> a kernel issue ?   It does happen both on 2.2 and 2.3 though...

It looks like either you have to shuffle around some IRQ settings in
the BIOS, or your HW/BIOS is really broken.

IRQ7 was used by the traditional PC interrupt controller to indicate a
spurious interrupt that the interrupt controller did not know how to
handle, perhaps by a glitch in the IRQ line. Not sure if this applies
to modern HW too, but the rule has always been "expect unexpected
IRQ7" and for the same reason it is usually assigned to the printer
port, because that counts as unimportant to handle IRQs for. :-)
I've not checked but perhaps this needs to be handled in the kernel
better.

But your problem could also be due to a simple resource conflict.
There are non-obvious cases, like the Bt848 card that would reliably
lock up my PC when given IRQ14 or 15. Turns out that somehow the
onboard IDE controllers wanted those for themselves despite them both
being disabled in the BIOS (and thus the BIOS thought those were
available for PCI). I could only solve that by assigning every PCI
card its IRQ by hand. A similar problem usually exists with the PS/2
mouse port on IRQ12.

Olaf