[openib-general] [PATCH v2] CM: check cm_id state before handling a REP

[Davis, Arlin R]

I was pushing way beyond the normal parameters of the normal MPI test
suites so I would say 2.6.17 is fine.

>-----Original Message-----
>From: Sean Hefty [mailto:mshefty at ichips.intel.com]
>Sent: Thursday, March 02, 2006 3:12 PM
>To: Roland Dreier; Davis, Arlin R
>Cc: Hefty, Sean; openib-general at openib.org
>Subject: Re: [openib-general] [PATCH v2] CM: check cm_id state before
handling a REP
>
>Roland Dreier wrote:
>>     Roland> BTW, should this be queued for kernel 2.6.17?
>>
>> Actually, do you think it's important enough for 2.6.16?
>
>My initial feeling is no.  I never hit this in any of my testing, and
Arlin was
>only able to hit it with some MPI scale-up tests.  However, he was able
to hit
>it consistently at a certain point.
>
>Arlin, what are you thoughts?
>
>- Sean