Re: cpuset_t

[Nicolas Joly <njoly%pasteur.fr@localhost>]

On Mon, May 16, 2016 at 09:15:55AM -0500, Jason Bacon wrote:

Hi,

> That makes sense...

cpuset chqnge looks good, except that you miss cpuset_destroy() call
for NetBSD. Without it, there is a small memory leak.

> I've now generated a patch I feel comfortable sending upstream and SLURM 
> 15.08.11 has passed basic testing on NetBSD and CentOS.  The patches 
> should make it into the next release.
> 
> Thanks much,
> 
>     JB
> 
> On 05/14/16 14:57, Kamil Rytarowski wrote:
> >On 14.05.2016 21:20, Jason Bacon wrote:
> >>Not sure if this is the right list, but I'm guessing someone here might
> >>be in the know.
> >>
> >>Porting the latest SLURM resource manager and ran into an off issue.
> >>
> >>sys/sched.h has the following typedef:
> >>
> >>     typedef struct _cpuset cpuset_t;
> >>
> >>The problem is, struct _cpuset is not defined anywhere in the headers.
> >>I discovered that it's actually defined in
> >>
> >>     /usr/src/common/lib/libc/sys/cpuset.c
> >>
> >>Is cpuset_t not intended for public consumption at this point, or was
> >>this just an oversight?
> >>
> >>Regards,
> >>
> >>     Jason
> >>
> >>
> >cpuset_t is to be used as an opaque pointer.
> >
> >Check for example:
> >pkgsrc-wip/rdsn-git/patches/patch-src_core_core_task__worker.cpp
> 
> 
> -- 
> All wars are civil wars, because all men are brothers ... Each one owes
> infinitely more to the human race than to the particular country in
> which he was born.
>                 -- Francois Fenelon

-- 
Nicolas Joly

Cluster & Computing Group
Biology IT Center
Institut Pasteur, Paris.