On Wed, Apr 27, 2011 at 4:47 PM, KAMEZAWA Hiroyuki
<kamezawa.hiroyu@xxxxxxxxxxxxxx> wrote:
> At memory reclaim, we determine the number of pages to be scanned
> per zone as
> Â Â Â Â(anon + file) >> priority.
> Assume
> Â Â Â Âscan = (anon + file) >> priority.
>
> If scan < SWAP_CLUSTER_MAX, the scan will be skipped for this time
> and priority gets higher. This has some problems.
>
> Â1. This increases priority as 1 without any scan.
> Â Â To do scan in this priority, amount of pages should be larger than 512M.
> Â Â If pages>>priority < SWAP_CLUSTER_MAX, it's recorded and scan will be
> Â Â batched, later. (But we lose 1 priority.)
> Â Â If memory size is below 16M, pages >> priority is 0 and no scan in
> Â Â DEF_PRIORITY forever.
>
> Â2. If zone->all_unreclaimabe==true, it's scanned only when priority==0.
> Â Â So, x86's ZONE_DMA will never be recoverred until the user of pages
> Â Â frees memory by itself.
>
> Â3. With memcg, the limit of memory can be small. When using small memcg,
> Â Â it gets priority < DEF_PRIORITY-2 very easily and need to call
> Â Â wait_iff_congested().
> Â Â For doing scan before priorty=9, 64MB of memory should be used.
>
> Then, this patch tries to scan SWAP_CLUSTER_MAX of pages in force...when
>
> Â1. the target is enough small.
> Â2. it's kswapd or memcg reclaim.
>
> Then we can avoid rapid priority drop and may be able to recover
> all_unreclaimable in a small zones. And this patch removes nr_saved_scan.
> This will allow scanning in this priority even when pages >> priority
> is very small.
>
> Changelog v2->v3
> Â- removed nr_saved_scan completely.
>
> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@xxxxxxxxxxxxxx>
Reviewed-by: Minchan Kim <minchan.kim@xxxxxxxxx>
The patch looks good to me but I have a nitpick about just coding style.
How about this? I think below looks better but it's just my private
opinion and I can't insist on my style. If you don't mind it, ignore.
barrios@barrios-desktop:~/linux-2.6$ git diff
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 6771ea7..268e7d4 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1817,8 +1817,28 @@ out:
scan >>= priority;
scan = div64_u64(scan * fraction[file], denominator);
}
- nr[l] = nr_scan_try_batch(scan,
- &reclaim_stat->nr_saved_scan[l]);
+
+ nr[l] = scan;
+ if (scan)
+ continue;
+ /*
+ * If zone is small or memcg is small, nr[l] can be 0.
+ * This results no-scan on this priority and priority drop down.
+ * For global direct reclaim, it can visit next zone and tend
+ * not to have problems. For global kswapd, it's for zone
+ * balancing and it need to scan a small amounts. When using
+ * memcg, priority drop can cause big latency. So, it's better
+ * to scan small amount. See may_noscan above.
+ */
+ if (((anon + file) >> priority) < SWAP_CLUSTER_MAX) {
+ /* kswapd does zone balancing and need to scan
this zone */
+ /* memcg may have small limit and need to
avoid priority drop */
+ if ((scanning_global_lru(sc) && current_is_kswapd())
+ || !scanning_global_lru(sc)) {
+ if (file || !noswap)
+ nr[l] = SWAP_CLUSTER_MAX;
+ }
+ }
}
}
--
Kind regards,
Minchan Kim
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href