lkml.org 
[lkml]   [2009]   [May]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[patch 11/11 -mmotm] oom: fail allocations if oom killer can't free memory
The oom killer now reports the potential number of pages that will be
freed when an oom kill task finally exits. If the oom killer can't make
any progress in freeing memory, however, there is no reason to continue
looping endlessly in the page allocator. Instead, the allocation is
failed if it is not __GFP_NOFAIL.

Cc: Mel Gorman <mel@csn.ul.ie>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Nick Piggin <npiggin@suse.de>
Cc: Christoph Lameter <cl@linux-foundation.org>
Cc: Dave Hansen <dave@linux.vnet.ibm.com>
Signed-off-by: David Rientjes <rientjes@google.com>
---
mm/page_alloc.c | 35 ++++++++++++++++++++---------------
1 files changed, 20 insertions(+), 15 deletions(-)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1524,7 +1524,7 @@ static inline struct page *
__alloc_pages_may_oom(gfp_t gfp_mask, unsigned int order,
struct zonelist *zonelist, enum zone_type high_zoneidx,
nodemask_t *nodemask, struct zone *preferred_zone,
- int migratetype)
+ int migratetype, unsigned long *did_some_progress)
{
struct page *page;

@@ -1551,7 +1551,7 @@ __alloc_pages_may_oom(gfp_t gfp_mask, unsigned int order,
goto out;

/* Exhausted what can be done so it's blamo time */
- out_of_memory(zonelist, gfp_mask, order);
+ *did_some_progress += out_of_memory(zonelist, gfp_mask, order);

out:
clear_zonelist_oom(zonelist, gfp_mask);
@@ -1716,7 +1716,6 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order,
*/
alloc_flags = gfp_to_alloc_flags(gfp_mask);

-restart:
/* This is the last chance, in general, before the goto nopage. */
page = get_page_from_freelist(gfp_mask, nodemask, order, zonelist,
high_zoneidx, alloc_flags & ~ALLOC_NO_WATERMARKS,
@@ -1760,21 +1759,27 @@ rebalance:
page = __alloc_pages_may_oom(gfp_mask, order,
zonelist, high_zoneidx,
nodemask, preferred_zone,
- migratetype);
+ migratetype, &did_some_progress);
if (page)
goto got_pg;

- /*
- * The OOM killer does not trigger for high-order
- * ~__GFP_NOFAIL allocations so if no progress is being
- * made, there are no other options and retrying is
- * unlikely to help.
- */
- if (order > PAGE_ALLOC_COSTLY_ORDER &&
- !(gfp_mask & __GFP_NOFAIL))
- goto nopage;
-
- goto restart;
+ if (!(gfp_mask & __GFP_NOFAIL)) {
+ /*
+ * The OOM killer does not trigger for
+ * high-order allocations so if no progress is
+ * being made, there are no other options and
+ * retrying is unlikely to help.
+ */
+ if (order > PAGE_ALLOC_COSTLY_ORDER)
+ goto nopage;
+
+ /*
+ * If the oom killer could not free any memory,
+ * there is no reason to endlessly loop.
+ */
+ if (!did_some_progress)
+ goto nopage;
+ }
}
}


\
 
 \ /
  Last update: 2009-05-11 00:13    [W:0.455 / U:0.296 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site