Re: [patch -mmotm] mm: invoke oom killer for __GFP_NOFAIL

[David Rientjes]

On Mon, 11 May 2009, Mel Gorman wrote:

> Can this get a comment explaining the intention like the code you are
> removing had? Something like this?
> 
> /*
>  * The OOM killer only triggers for lower order allocation
>  * requests and when __GFP_NOFAIL is specified to prevent
>  * endlessly looping. Only retry the allocation if the OOM
>  * killer was used.
>  */
> 
> Otherwise, it looks ok.
> 

Sure, I'll add a comment to patch 11 in my latest series that changes the 
code again in this area.

> One related question, what is doing a costly-high-order allocation request
> with __GFP_NOFAIL specified?
> 

Nothing to my knowledge, this is simply a bug fix before the big change in 
my series where we fail page allocations if reclaim fails and the oom 
killer fails, which previously hasn't been done (we'd loop forever oom 
killing tasks that may or may not respond).  The definition of 
__GFP_NOFAIL does not make the guarantee of infinitely retrying for only 
specific ranges of orders.
--
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/