Re: [RFC][Patch] fix for irq_affinity_write_proc v2.5

[Alan Cox]

On Mer, 2003-05-07 at 03:27, William Lee Irwin III wrote:
> It's basically not working as specified for clustered hierarchical, and
> in truth the specification can never be met. As it stands most calls to
> it are lethal on such systems, especially those using physical destmod.
> 
> I'd prefer to have it redesigned for some validity checking and error
> returns as on such systems the impossible destinations serve no purpose
> but raising MCE's and/or deadlocking the box.

In which case I agree it does indeed make sense