1 files changed, 56 insertions, 5 deletions

diff --git a/mm/swapfile.c b/mm/swapfile.c
index 18dfccb91123..c531d2687dd0 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -613,10 +613,12 @@ new_cluster:
 		} else if (!cluster_list_empty(&si->discard_clusters)) {
 			/*
 			 * we don't have free cluster but have some clusters in
-			 * discarding, do discard now and reclaim them
+			 * discarding, do discard now and reclaim them, then
+			 * reread cluster_next_cpu since we dropped si->lock
 			 */
 			swap_do_scheduled_discard(si);
-			*scan_base = *offset = si->cluster_next;
+			*scan_base = this_cpu_read(*si->cluster_next_cpu);
+			*offset = *scan_base;
 			goto new_cluster;
 		} else
 			return false;
@@ -722,6 +724,34 @@ static void swap_range_free(struct swap_info_struct *si, unsigned long offset,
 	}
 }
 
+static void set_cluster_next(struct swap_info_struct *si, unsigned long next)
+{
+	unsigned long prev;
+
+	if (!(si->flags & SWP_SOLIDSTATE)) {
+		si->cluster_next = next;
+		return;
+	}
+
+	prev = this_cpu_read(*si->cluster_next_cpu);
+	/*
+	 * Cross the swap address space size aligned trunk, choose
+	 * another trunk randomly to avoid lock contention on swap
+	 * address space if possible.
+	 */
+	if ((prev >> SWAP_ADDRESS_SPACE_SHIFT) !=
+	    (next >> SWAP_ADDRESS_SPACE_SHIFT)) {
+		/* No free swap slots available */
+		if (si->highest_bit <= si->lowest_bit)
+			return;
+		next = si->lowest_bit +
+			prandom_u32_max(si->highest_bit - si->lowest_bit + 1);
+		next = ALIGN_DOWN(next, SWAP_ADDRESS_SPACE_PAGES);
+		next = max_t(unsigned int, next, si->lowest_bit);
+	}
+	this_cpu_write(*si->cluster_next_cpu, next);
+}
+
 static int scan_swap_map_slots(struct swap_info_struct *si,
 			       unsigned char usage, int nr,
 			       swp_entry_t slots[])
@@ -746,7 +776,16 @@ static int scan_swap_map_slots(struct swap_info_struct *si,
 	 */
 
 	si->flags += SWP_SCANNING;
-	scan_base = offset = si->cluster_next;
+	/*
+	 * Use percpu scan base for SSD to reduce lock contention on
+	 * cluster and swap cache.  For HDD, sequential access is more
+	 * important.
+	 */
+	if (si->flags & SWP_SOLIDSTATE)
+		scan_base = this_cpu_read(*si->cluster_next_cpu);
+	else
+		scan_base = si->cluster_next;
+	offset = scan_base;
 
 	/* SSD algorithm */
 	if (si->cluster_info) {
@@ -835,7 +874,6 @@ checks:
 	unlock_cluster(ci);
 
 	swap_range_alloc(si, offset, 1);
-	si->cluster_next = offset + 1;
 	slots[n_ret++] = swp_entry(si->type, offset);
 
 	/* got enough slots or reach max slots? */
@@ -884,6 +922,7 @@ checks:
 	}
 
 done:
+	set_cluster_next(si, offset + 1);
 	si->flags -= SWP_SCANNING;
 	return n_ret;
 
@@ -2653,6 +2692,8 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 	mutex_unlock(&swapon_mutex);
 	free_percpu(p->percpu_cluster);
 	p->percpu_cluster = NULL;
+	free_percpu(p->cluster_next_cpu);
+	p->cluster_next_cpu = NULL;
 	vfree(swap_map);
 	kvfree(cluster_info);
 	kvfree(frontswap_map);
@@ -3205,11 +3246,19 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 		unsigned long ci, nr_cluster;
 
 		p->flags |= SWP_SOLIDSTATE;
+		p->cluster_next_cpu = alloc_percpu(unsigned int);
+		if (!p->cluster_next_cpu) {
+			error = -ENOMEM;
+			goto bad_swap_unlock_inode;
+		}
 		/*
 		 * select a random position to start with to help wear leveling
 		 * SSD
 		 */
-		p->cluster_next = 1 + prandom_u32_max(p->highest_bit);
+		for_each_possible_cpu(cpu) {
+			per_cpu(*p->cluster_next_cpu, cpu) =
+				1 + prandom_u32_max(p->highest_bit);
+		}
 		nr_cluster = DIV_ROUND_UP(maxpages, SWAPFILE_CLUSTER);
 
 		cluster_info = kvcalloc(nr_cluster, sizeof(*cluster_info),
@@ -3325,6 +3374,8 @@ bad_swap_unlock_inode:
 bad_swap:
 	free_percpu(p->percpu_cluster);
 	p->percpu_cluster = NULL;
+	free_percpu(p->cluster_next_cpu);
+	p->cluster_next_cpu = NULL;
 	if (inode && S_ISBLK(inode->i_mode) && p->bdev) {
 		set_blocksize(p->bdev, p->old_block_size);
 		blkdev_put(p->bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);