[PATCH] swap: Delete the "last_in_cluster < scan_base" loop in the body of scan_swap_map()

[PATCH] swap: Delete the "last_in_cluster < scan_base" loop in the body of scan_swap_map()

[Chen Yucong]

>From commit ebc2a1a69111, we can find that all SWP_SOLIDSTATE "seek is cheap"(SSD case) 
has already gone to si->cluster_info scan_swap_map_try_ssd_cluster() route. So that the
"last_in_cluster < scan_base" loop in the body of scan_swap_map() has already become a 
dead code snippet, and it should have been deleted.

This patch is to delete the redundant loop as Hugh and Shaohua suggested.

Signed-off-by: Chen Yucong <slaoub@gmail.com>
---
 mm/swapfile.c |   20 --------------------
 1 file changed, 20 deletions(-)

diff --git a/mm/swapfile.c b/mm/swapfile.c
index beeeef8..1b44bd9 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -549,26 +549,6 @@ static unsigned long scan_swap_map(struct swap_info_struct *si,
 			}
 		}
 
-		offset = si->lowest_bit;
-		last_in_cluster = offset + SWAPFILE_CLUSTER - 1;
-
-		/* Locate the first empty (unaligned) cluster */
-		for (; last_in_cluster < scan_base; offset++) {
-			if (si->swap_map[offset])
-				last_in_cluster = offset + SWAPFILE_CLUSTER;
-			else if (offset == last_in_cluster) {
-				spin_lock(&si->lock);
-				offset -= SWAPFILE_CLUSTER - 1;
-				si->cluster_next = offset;
-				si->cluster_nr = SWAPFILE_CLUSTER - 1;
-				goto checks;
-			}
-			if (unlikely(--latency_ration < 0)) {
-				cond_resched();
-				latency_ration = LATENCY_LIMIT;
-			}
-		}
-
 		offset = scan_base;
 		spin_lock(&si->lock);
 		si->cluster_nr = SWAPFILE_CLUSTER - 1;
-- 
1.7.10.4

Re: [PATCH] swap: Delete the "last_in_cluster < scan_base" loop in the body of scan_swap_map()

[Hugh Dickins]

On Mon, 2 Jun 2014, Chen Yucong wrote:

> From commit ebc2a1a69111, we can find that all SWP_SOLIDSTATE "seek is cheap"(SSD case) 
> has already gone to si->cluster_info scan_swap_map_try_ssd_cluster() route. So that the
> "last_in_cluster < scan_base" loop in the body of scan_swap_map() has already become a 
> dead code snippet, and it should have been deleted.
> 
> This patch is to delete the redundant loop as Hugh and Shaohua suggested.
> 
> Signed-off-by: Chen Yucong <slaoub@gmail.com>

That is very nice, thank you.

Acked-by: Hugh Dickins <hughd@google.com>

But it does beg for just a little more: perhaps Andrew can kindly fold in:
---

 mm/swapfile.c |    9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

--- chen/mm/swapfile.c	2014-06-02 15:55:44.812368186 -0700
+++ hugh/mm/swapfile.c	2014-06-02 16:15:20.344396124 -0700
@@ -505,13 +505,10 @@ static unsigned long scan_swap_map(struc
 		/*
 		 * If seek is expensive, start searching for new cluster from
 		 * start of partition, to minimize the span of allocated swap.
-		 * But if seek is cheap, search from our current position, so
-		 * that swap is allocated from all over the partition: if the
-		 * Flash Translation Layer only remaps within limited zones,
-		 * we don't want to wear out the first zone too quickly.
+		 * If seek is cheap, that is the SWP_SOLIDSTATE si->cluster_info
+		 * case, just handled by scan_swap_map_try_ssd_cluster() above.
 		 */
-		if (!(si->flags & SWP_SOLIDSTATE))
-			scan_base = offset = si->lowest_bit;
+		scan_base = offset = si->lowest_bit;
 		last_in_cluster = offset + SWAPFILE_CLUSTER - 1;
 
 		/* Locate the first empty (unaligned) cluster */