[PATCH] octeon: request_irq() failure ignored in cvm_oct_configure_common_hw()

[Jiri Slaby]

On 09/04/2009 12:47 AM, David Daney wrote:
> This is all a bit whacky, at this point the only sane thing I can think
> of doing is to panic if we cannot get the irq.

Panic from driver which fails on request_irq is not a sane thing at all.
There is just /proc entry creation, tasklet init and the things done in
cvm_oct_configure_common_hw. I believe it's not that hard to revert the
initialization properly.