[Pvfs2-developers] Re: double-free in BMI_ib_cancel

[Pete Wyckoff]

slang at mcs.anl.gov wrote on Tue, 08 Apr 2008 15:03 -0500:
>
> On Apr 8, 2008, at 2:21 PM, Kyle Schochenmaier wrote:
>> I think this is a fairly critical fix and will affect the majority of
>> the ib users I know of, who are all running cvs head for various
>> reasons.
>> The patch only moves two #defines around so I'm not sure why we should
>> hold off on it, maybe I'm missing something?
>
> Nope, I just need to know whether to merge it to the release branch or not.
> -sam

My version of Kyle's bugfix is in, but untested.  I was hoping for
confirmation from him first before poking it into -stable.

If/when he does confirm, and if you happened to have sucked in the
precursor that caused the bug, please push to -stable.  It was:

    Author: pw <pw>
    Date:   Fri Feb 15 15:10:14 2008 +0000

    bmi_ib: cancel memcache deregister states
    
    In BMI_ib_cancel, fix up the list of states during which memory should
    be deregistered.  This may avoid some assert() errors during BMI message
    cancellation.

Thanks,

		-- Pete