Re: [PATCH 4/4] staging: octeon: handle rx/tx initialization failures in probe

[Dan Carpenter]

On Mon, Jun 29, 2026 at 09:49:49PM +0530, Prashant Rahul wrote:
> On Mon Jun 29, 2026 at 1:14 PM IST, Dan Carpenter wrote:
> > cvmx_pko_shutdown() has a built in call to
> > cvmx_pko_disable() so why are we calling that separately?
> 
> Didn't realize cvmx_pko_shutdown did that. Will remove that. Also, Should
> I remove the cvmx_pko_disable() call in the remove function in this patch
> series? Or should I do that in a new patch?
> 
> > Also I feel like probe() should fail if alloc_netdev() fails
> > or register_netdev(), but that's not introduced by this patch
> > either.
> 
> My main goal with the patch series was to get rid of panic() calls in
> some of the initialization error paths. I will look into properly handling
> alloc_etherdev and register_netdev errors but in new patch(es).

To be honest, I would keep your patch as-is instead of being more
ambitious.  If you get much more ambitious then we'd need to test
it.  Right now, we're only affecting paths which would have caused
a panic in the original code, but if we added error checking for
register_netdev() then there is a chance of breaking something.

I was just saying that this might not be 100% correct, but it
doesn't break anything that wasn't broken in the original so
it's fine.

regards,
dan carpenter