summaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorDavid Rientjes <rientjes@google.com>2007-10-16 23:25:55 -0700
committerLinus Torvalds <torvalds@woody.linux-foundation.org>2007-10-17 08:42:45 -0700
commit098d7f128a4e53cb64930628915ac767785e0e60 (patch)
treeed3cab1daecab7f2a64b27deed190df3ec218789
parente815af95f94914993bbad279c71cf5fef9f4eaac (diff)
downloadop-kernel-dev-098d7f128a4e53cb64930628915ac767785e0e60.zip
op-kernel-dev-098d7f128a4e53cb64930628915ac767785e0e60.tar.gz
oom: add per-zone locking
OOM killer synchronization should be done with zone granularity so that memory policy and cpuset allocations may have their corresponding zones locked and allow parallel kills for other OOM conditions that may exist elsewhere in the system. DMA allocations can be targeted at the zone level, which would not be possible if locking was done in nodes or globally. Synchronization shall be done with a variation of "trylocks." The goal is to put the current task to sleep and restart the failed allocation attempt later if the trylock fails. Otherwise, the OOM killer is invoked. Each zone in the zonelist that __alloc_pages() was called with is checked for the newly-introduced ZONE_OOM_LOCKED flag. If any zone has this flag present, the "trylock" to serialize the OOM killer fails and returns zero. Otherwise, all the zones have ZONE_OOM_LOCKED set and the try_set_zone_oom() function returns non-zero. Cc: Andrea Arcangeli <andrea@suse.de> Cc: Christoph Lameter <clameter@sgi.com> Signed-off-by: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
-rw-r--r--include/linux/mmzone.h5
-rw-r--r--include/linux/oom.h3
-rw-r--r--mm/oom_kill.c52
3 files changed, 60 insertions, 0 deletions
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index bad9486..9011505 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -344,6 +344,7 @@ struct zone {
typedef enum {
ZONE_ALL_UNRECLAIMABLE, /* all pages pinned */
ZONE_RECLAIM_LOCKED, /* prevents concurrent reclaim */
+ ZONE_OOM_LOCKED, /* zone is in OOM killer zonelist */
} zone_flags_t;
static inline void zone_set_flag(struct zone *zone, zone_flags_t flag)
@@ -363,6 +364,10 @@ static inline int zone_is_reclaim_locked(const struct zone *zone)
{
return test_bit(ZONE_RECLAIM_LOCKED, &zone->flags);
}
+static inline int zone_is_oom_locked(const struct zone *zone)
+{
+ return test_bit(ZONE_OOM_LOCKED, &zone->flags);
+}
/*
* The "priority" of VM scanning is how much of the queues we will scan in one
diff --git a/include/linux/oom.h b/include/linux/oom.h
index cf6ebf5..e908120 100644
--- a/include/linux/oom.h
+++ b/include/linux/oom.h
@@ -20,6 +20,9 @@ enum oom_constraint {
CONSTRAINT_MEMORY_POLICY,
};
+extern int try_set_zone_oom(struct zonelist *zonelist);
+extern void clear_zonelist_oom(struct zonelist *zonelist);
+
extern void out_of_memory(struct zonelist *zonelist, gfp_t gfp_mask, int order);
extern int register_oom_notifier(struct notifier_block *nb);
extern int unregister_oom_notifier(struct notifier_block *nb);
diff --git a/mm/oom_kill.c b/mm/oom_kill.c
index 1a7a4ef..6e999c8 100644
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c
@@ -27,6 +27,7 @@
#include <linux/notifier.h>
int sysctl_panic_on_oom;
+static DEFINE_MUTEX(zone_scan_mutex);
/* #define DEBUG */
/**
@@ -374,6 +375,57 @@ int unregister_oom_notifier(struct notifier_block *nb)
}
EXPORT_SYMBOL_GPL(unregister_oom_notifier);
+/*
+ * Try to acquire the OOM killer lock for the zones in zonelist. Returns zero
+ * if a parallel OOM killing is already taking place that includes a zone in
+ * the zonelist. Otherwise, locks all zones in the zonelist and returns 1.
+ */
+int try_set_zone_oom(struct zonelist *zonelist)
+{
+ struct zone **z;
+ int ret = 1;
+
+ z = zonelist->zones;
+
+ mutex_lock(&zone_scan_mutex);
+ do {
+ if (zone_is_oom_locked(*z)) {
+ ret = 0;
+ goto out;
+ }
+ } while (*(++z) != NULL);
+
+ /*
+ * Lock each zone in the zonelist under zone_scan_mutex so a parallel
+ * invocation of try_set_zone_oom() doesn't succeed when it shouldn't.
+ */
+ z = zonelist->zones;
+ do {
+ zone_set_flag(*z, ZONE_OOM_LOCKED);
+ } while (*(++z) != NULL);
+out:
+ mutex_unlock(&zone_scan_mutex);
+ return ret;
+}
+
+/*
+ * Clears the ZONE_OOM_LOCKED flag for all zones in the zonelist so that failed
+ * allocation attempts with zonelists containing them may now recall the OOM
+ * killer, if necessary.
+ */
+void clear_zonelist_oom(struct zonelist *zonelist)
+{
+ struct zone **z;
+
+ z = zonelist->zones;
+
+ mutex_lock(&zone_scan_mutex);
+ do {
+ zone_clear_flag(*z, ZONE_OOM_LOCKED);
+ } while (*(++z) != NULL);
+ mutex_unlock(&zone_scan_mutex);
+}
+
/**
* out_of_memory - kill the "best" process when we run out of memory
*
OpenPOWER on IntegriCloud