Re: [PATCH 2.6.16-mm1 1/2] sched_domain: handle kmalloc failure

[Srivatsa Vaddagiri <vatsa@in.ibm.com>]

On Sat, Mar 25, 2006 at 06:44:41PM -0800, Andrew Morton wrote:
> Well, when is this code called?  It would be at boot time, in which case
> the allocations will succeed (if not, the boot fails) or at cpu/node
> hot-add, in which case the appropriate response is to fail to bring up the
> new cpu/node.


Hmm ..dont follow you here. Are you saying the above is the current
behavior (say in 2.6.16-mm1)? AFAICS that is not true because 
arch_init_sched_domains (which handles both bootup & cpu hot-add cases) 
doesnt return any error back to its caller.

Also note that build_sched_domains can be called from CPUset code too
(partition_sched_domains) when we modify the exclusive property of a CPUset.

> It's better to send the administrator back to work out why we ran out of
> memory than to appear to have brought the new cpu/node online, only to have
> it run funny.
> 
> I think?

I don't know. Is load balancing an absolute critical feature that
without it system bootup and cpu hot-add should be failed? Also note
that we may be able to do partial load balancing (between threads of a
CPU) even when there is memory allocation failure.

Regarding giving administrator hints of allocation failure, the printks in 
build_sched_domain do give those hints.

> 
> > > build_sched_domains() should be static and __cpuinit, btw.
> > 
> > Ok ..Will take care of that in the next version of the patch.
> > 
> 
> umm, it's probably best to not bother.  I think Ashok is looking into all
> the memory we're presently wasting on non-cpu_hotplug builds.  There's
> quite a lot in there.

Ok ..Will let Ashok take care of adding __cpuinit prefix.

-- 
Regards,
vatsa
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/