Re: [PATCH 6/8] Use SLAB_PANIC when creating critical slab cache

[Chris Wright <chrisw@xxxxxxxx>]

* Arnaldo Carvalho de Melo (acme@xxxxxxxxxxxxxxxx) wrote:
> Chris,
> 
>       I think somebody proposed this in the past and it was refused
> because we better get rid of the panics and return an error, propagate
> it and refuse to load the module, better have the machine still alive
> but without networking than panic it.

Sounds good too.  Part of my thinking was that these patches would be a
good chance to review if these are necessarily panic conditions.  What was
the plan on error return, since I expect some caches are interdependent?

thanks,
-chris