On Fri, Sep 07, 2012 at 10:32:03AM +0100, Mel Gorman wrote:

On Fri, Sep 07, 2012 at 09:21:45AM +0100, Mel Gorman wrote:

quoted

So other than the mix up of order parameters I think this should work.

But I'd be wrong, isolated page accounting is not fixed up so it will
eventually hang on too_many_isolated. It turns out it is necessary

Good spot.

to pass in zone after all. The following patch passed a high order
allocation stress test. To actually exercise the path I had compaction
call reclaim_clean_pages_from_list() in a separate debugging patch.

Minchan, can you test your CMA allocation latency test with this patch?
If the figures are satisfactory could you add them to the changelog and
consider replacing the MIGRATE_DISCARD pair of patches with this version
please?

Of course, I will do next week.
Thanks!

--
Kind Regards,
Minchan Kim

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>