sched/numa: Set preferred NUMA node based on number of private faults

Ideally it would be possible to distinguish between NUMA hinting faults that
are private to a task and those that are shared. If treated identically
there is a risk that shared pages bounce between nodes depending on
the order they are referenced by tasks. Ultimately what is desirable is
that task private pages remain local to the task while shared pages are
interleaved between sharing tasks running on different nodes to give good
average performance. This is further complicated by THP as even
applications that partition their data may not be partitioning on a huge
page boundary.

To start with, this patch assumes that multi-threaded or multi-process
applications partition their data and that in general the private accesses
are more important for cpu->memory locality in the general case. Also,
no new infrastructure is required to treat private pages properly but
interleaving for shared pages requires additional infrastructure.

To detect private accesses the pid of the last accessing task is required
but the storage requirements are a high. This patch borrows heavily from
Ingo Molnar's patch "numa, mm, sched: Implement last-CPU+PID hash tracking"
to encode some bits from the last accessing task in the page flags as
well as the node information. Collisions will occur but it is better than
just depending on the node information. Node information is then used to
determine if a page needs to migrate. The PID information is used to detect
private/shared accesses. The preferred NUMA node is selected based on where
the maximum number of approximately private faults were measured. Shared
faults are not taken into consideration for a few reasons.

First, if there are many tasks sharing the page then they'll all move
towards the same node. The node will be compute overloaded and then
scheduled away later only to bounce back again. Alternatively the shared
tasks would just bounce around nodes because the fault information is
effectively noise. Either way accounting for shared faults the same as
private faults can result in lower performance overall.

The second reason is based on a hypothetical workload that has a small
number of very important, heavily accessed private pages but a large shared
array. The shared array would dominate the number of faults and be selected
as a preferred node even though it's the wrong decision.

The third reason is that multiple threads in a process will race each
other to fault the shared page making the fault information unreliable.

Signed-off-by: Mel Gorman <mgorman@suse.de>
[ Fix complication error when !NUMA_BALANCING. ]
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/1381141781-10992-30-git-send-email-mgorman@suse.de
Signed-off-by: Ingo Molnar <mingo@kernel.org>

1 files changed, 8 insertions, 8 deletions

diff --git a/mm/memory.c b/mm/memory.c
index 3e3b4b8b6c41..cc7f20691c82 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -69,8 +69,8 @@
 
 #include "internal.h"
 
-#ifdef LAST_NID_NOT_IN_PAGE_FLAGS
-#warning Unfortunate NUMA and NUMA Balancing config, growing page-frame for last_nid.
+#ifdef LAST_NIDPID_NOT_IN_PAGE_FLAGS
+#warning Unfortunate NUMA and NUMA Balancing config, growing page-frame for last_nidpid.
 #endif
 
 #ifndef CONFIG_NEED_MULTIPLE_NODES
@@ -3536,7 +3536,7 @@ int do_numa_page(struct mm_struct *mm, struct vm_area_struct *vma,
         struct page *page = NULL;
         spinlock_t *ptl;
         int page_nid = -1;
-        int last_nid;
+        int last_nidpid;
         int target_nid;
         bool migrated = false;
 
@@ -3567,7 +3567,7 @@ int do_numa_page(struct mm_struct *mm, struct vm_area_struct *vma,
         }
         BUG_ON(is_zero_pfn(page_to_pfn(page)));
 
-        last_nid = page_nid_last(page);
+        last_nidpid = page_nidpid_last(page);
         page_nid = page_to_nid(page);
         target_nid = numa_migrate_prep(page, vma, addr, page_nid);
         pte_unmap_unlock(ptep, ptl);
@@ -3583,7 +3583,7 @@ int do_numa_page(struct mm_struct *mm, struct vm_area_struct *vma,
 
 out:
         if (page_nid != -1)
-                task_numa_fault(last_nid, page_nid, 1, migrated);
+                task_numa_fault(last_nidpid, page_nid, 1, migrated);
         return 0;
 }
 
@@ -3598,7 +3598,7 @@ static int do_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma,
         unsigned long offset;
         spinlock_t *ptl;
         bool numa = false;
-        int last_nid;
+        int last_nidpid;
 
         spin_lock(&mm->page_table_lock);
         pmd = *pmdp;
@@ -3643,7 +3643,7 @@ static int do_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma,
                 if (unlikely(!page))
                         continue;
 
-                last_nid = page_nid_last(page);
+                last_nidpid = page_nidpid_last(page);
                 page_nid = page_to_nid(page);
                 target_nid = numa_migrate_prep(page, vma, addr, page_nid);
                 pte_unmap_unlock(pte, ptl);
@@ -3656,7 +3656,7 @@ static int do_pmd_numa_page(struct mm_struct *mm, struct vm_area_struct *vma,
                 }
 
                 if (page_nid != -1)
-                        task_numa_fault(last_nid, page_nid, 1, migrated);
+                        task_numa_fault(last_nidpid, page_nid, 1, migrated);
 
                 pte = pte_offset_map_lock(mm, pmdp, addr, &ptl);
         }