Re: Huge memory takes too long time to initialize on 4TB ?

[David Rientjes]

On Thu, 24 May 2012, Andrea Arcangeli wrote:

> > This has nothing to do with mm/huge_memory.c, which is transparent 
> > hugepages code, it's the lengthy iteration done for a very large 
> > ZONE_NORMAL in setup_zone_migrate_reserve().  
> > 
> > This should be fixed by 938929f14cb5 ("mm: reduce the amount of work done 
> > when updating min_free_kbytes") and was merged in 3.3.  Let us know if 
> > there's still a problem after upgrading to either that or 3.4
> 
> I thought that was only a performance optimization, so if the oops is
> generated by a timeout it can fix it, so good idea to try it. But a
> timeout watchdog triggering exactly in spin_unlock sounds a bit
> unlikely occurence, so I assumed too long time here meant "wait forever
> because it crashed".
> 

The bug report is ambiguous because it doesn't include the full set of 
messages, so I was assuming it was timing out because irqs were disabled 
for too long while iterating over a large ZONE_NORMAL based on Ethan's 
description of what he was seeing.  Ethan, if the problem persists with a 
3.3 or later kernel please follow-up with the full dmesg, thanks.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/