On Wed, 18 May 2011, Pekka Enberg wrote:

On 5/17/11 12:10 AM, David Rientjes wrote:

quoted

On Fri, 13 May 2011, Mel Gorman wrote:

quoted

To avoid locking and per-cpu overhead, SLUB optimisically uses
high-order allocations and falls back to lower allocations if they
fail.  However, by simply trying to allocate, kswapd is woken up to
start reclaiming at that order. On a desktop system, two users report
that the system is getting locked up with kswapd using large amounts
of CPU.  Using SLAB instead of SLUB made this problem go away.

This patch prevents kswapd being woken up for high-order allocations.
Testing indicated that with this patch applied, the system was much
harder to hang and even when it did, it eventually recovered.

Signed-off-by: Mel Gorman<mgorman@suse.de>

Acked-by: David Rientjes<rientjes@google.com>

Christoph? I think this patch is sane although the original rationale was to
workaround kswapd problems.

I am mostly fine with it. The concerns that I have is if there is a
large series of high order allocs then at some point you would want
kswapd to be triggered instead of high order allocs constantly failing.

Can we have a "trigger once in a while" functionality?