Messages in this thread | | | Date | Wed, 15 Jun 2011 09:37:26 +0200 | From | Ingo Molnar <> | Subject | Re: [PATCH 4 of 6] x86, UV: correct reset_with_ipi() |
| |
* Cliff Wickman <cpw@sgi.com> wrote:
> From: Cliff Wickman <cpw@sgi.com> > > Fix reset_with_ipi() to look up a cpu for a blade based on the > distribution map being indexed by the potentially sparsely numbered pnode > (relative to the partition base pnode). (It had been doing it based on > 0-based, consecutive blade id.)
That's one fix.
> Rename 'hubmask' to 'pnmask' for clarity.
That's another one.
> And remove the stack-resident cpumask_t from reset_with_ipi() by allocating > one per uvhub.
That's a third one.
Why isnt this done in three separate patches?
Ingo
| |