Hello,
On platforms with PAGE_SIZE set to rather big values (e.g. 256KB) the
amount of memory necessary for for the stripe cache may be significant,
so we can't rely on the hard-coded value of initial max_nr_stripes
(this may cause invoking OOM-killer for terminating RAID creation process).
What this patch does is just checks the size of available memory,
and assigns the appropriate, safe, value to initial max_nr_stripes:
either the hard-coded NR_STRIPES value if it's safe in the sense that
we'll have some free memory available using stripe cache with such size,
or the calculated, lesser, value.
Signed-off-by: Yuri Tikhonov <yur@xxxxxxxxxxx>
--
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index d59e1bc..1d2b064 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -4934,10 +4934,11 @@ static struct attribute_group raid5_attrs_group = {
static int run(mddev_t *mddev)
{
raid5_conf_t *conf;
- int raid_disk, memory;
+ int raid_disk, memory, avail_stripes;
mdk_rdev_t *rdev;
struct disk_info *disk;
struct list_head *tmp;
+ struct sysinfo val;
int working_disks = 0;
if (mddev->level != 5 && mddev->level != 4 && mddev->level != 6) {
@@ -5074,7 +5075,14 @@ static int run(mddev_t *mddev)
else
conf->max_degraded = 1;
conf->algorithm = mddev->layout;
- conf->max_nr_stripes = NR_STRIPES;
+
+ /* Set the safe stripe cache size */
+ si_meminfo(&val);
+ avail_stripes = (val.freeram + val.freehigh) * val.mem_unit /
+ (sizeof(struct stripe_head) + conf->raid_disks *
+ ((sizeof(struct bio) + PAGE_SIZE))) - 1;
+ conf->max_nr_stripes = (avail_stripes < NR_STRIPES) ? avail_stripes :
+ NR_STRIPES;
conf->expand_progress = mddev->reshape_position;
/* device size must be a multiple of chunk size */
--
Yuri Tikhonov, Senior Software Engineer
Emcraft Systems, www.emcraft.com
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
