Messages in this thread | | | Date | Mon, 23 Mar 2009 09:38:38 -0400 (EDT) | From | Christoph Lameter <> | Subject | Re: [PATCH] fix vmscan to take care of nodemask |
| |
On Mon, 23 Mar 2009, Mel Gorman wrote:
> try_to_free_pages() is used for the direct reclaim of up to > SWAP_CLUSTER_MAX pages when watermarks are low. The caller to > alloc_pages_nodemask() can specify a nodemask of nodes that are allowed > to be used but this is not passed to try_to_free_pages(). This can lead > to the unnecessary reclaim of pages that are unusable by the caller and > in the worst case lead to allocation failure as progress was not been > made where it is needed. > > This patch passes the nodemask used for alloc_pages_nodemask() to > try_to_free_pages().
This is only useful for MPOL_BIND. Direct reclaim within a cpuset already honors the boundaries of the cpuset.
| |