Re: [BUG] Commit d065bd81 severely regresses huge page allocation success rates

[Mel Gorman <mel@csn.ul.ie>]

On Thu, Nov 11, 2010 at 04:25:00AM -0800, Michel Lespinasse wrote:
> On Thu, Nov 11, 2010 at 4:15 AM, Mel Gorman <mel@csn.ul.ie> wrote:
> > When testing 2.6.37-rc1, I noticed that huge page allocation success
> > rates were severely impaired. Bisection showed that commit [d065bd81: mm:
> > retry page fault when blocking on disk transfer] was the biggest factor.
> > Reverting the patch confirmed this. Here are the results of a high-order
> > allocation stress test. The vanilla kernel is 2.6.37-rc1 and the revert
> > kernel has this commit removed with minor conflicts cleaned up.
> [...]
> > I have not digested what the patch is doing but am reporting it in case
> > people familiar with the patch spot the problem quickly.
> 
> There was a bug in that commit, which was fixed in
> d88c0922fa0e2c021a028b310a641126c6d4b7dc
> 
> I think this may explain the issue; compaction won't work well if
> extra references are kept :/
> 

Reclaim would not work either as it is effectively a leak.

> Can you check that this indeed solves your test case ?
> 

It does. Thanks very much.

-- 
Mel Gorman
Part-time Phd Student                          Linux Technology Center
University of Limerick                         IBM Dublin Software Lab