mm: numa: Sanitize task_numa_fault() callsites
authorMel Gorman <mgorman@suse.de>
Mon, 7 Oct 2013 10:28:45 +0000 (11:28 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 13 Nov 2013 03:05:34 +0000 (12:05 +0900)
commit174dfa40d63b2250e299d5376937d200e4662b7c
tree8e8eb3fdaa8eb818be95938f3591fedfae6b6cf6
parent299723f2297726ca4c0a9d8ef3548ece5466431d
mm: numa: Sanitize task_numa_fault() callsites

commit c61109e34f60f6e85bb43c5a1cd51c0e3db40847 upstream.

There are three callers of task_numa_fault():

 - do_huge_pmd_numa_page():
     Accounts against the current node, not the node where the
     page resides, unless we migrated, in which case it accounts
     against the node we migrated to.

 - do_numa_page():
     Accounts against the current node, not the node where the
     page resides, unless we migrated, in which case it accounts
     against the node we migrated to.

 - do_pmd_numa_page():
     Accounts not at all when the page isn't migrated, otherwise
     accounts against the node we migrated towards.

This seems wrong to me; all three sites should have the same
sementaics, furthermore we should accounts against where the page
really is, we already know where the task is.

So modify all three sites to always account; we did after all receive
the fault; and always account to where the page is after migration,
regardless of success.

They all still differ on when they clear the PTE/PMD; ideally that
would get sorted too.

Signed-off-by: Mel Gorman <mgorman@suse.de>
Reviewed-by: Rik van Riel <riel@redhat.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1381141781-10992-8-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
mm/huge_memory.c
mm/memory.c