[v2,1/4] sbitmap: optimise sbitmap_deferred_clear()

Commit Message

Pavel Begunkov Nov. 22, 2020, 3:35 p.m. UTC

Because of spinlocks and atomics sbitmap_deferred_clear() have to reload
&sb->map[index] on each access even though the map address won't change.
Pass in sbitmap_word instead of {sb, index}, so it's cached in a
variable. It also improves code generation of
sbitmap_find_bit_in_index().

Signed-off-by: Pavel Begunkov <[e-mail]>
---
 lib/sbitmap.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

Comments

John Garry Nov. 24, 2020, 2:11 p.m. UTC | #1

On 22/11/2020 15:35, Pavel Begunkov wrote:
> Because of spinlocks and atomics sbitmap_deferred_clear() have to reload
> &sb->map[index] on each access even though the map address won't change.
> Pass in sbitmap_word instead of {sb, index}, so it's cached in a
> variable. It also improves code generation of
> sbitmap_find_bit_in_index().
> 
> Signed-off-by: Pavel Begunkov<[e-mail]>

Looks ok, even though a bit odd not be passing a struct sbitmap * now

Reviewed-by: John Garry <[e-mail]>

Pavel Begunkov Nov. 24, 2020, 3:01 p.m. UTC | #2

On 24/11/2020 14:11, John Garry wrote:
> On 22/11/2020 15:35, Pavel Begunkov wrote:
>> Because of spinlocks and atomics sbitmap_deferred_clear() have to reload
>> &sb->map[index] on each access even though the map address won't change.
>> Pass in sbitmap_word instead of {sb, index}, so it's cached in a
>> variable. It also improves code generation of
>> sbitmap_find_bit_in_index().
>>
>> Signed-off-by: Pavel Begunkov<[e-mail]>
> 
> Looks ok, even though a bit odd not be passing a struct sbitmap * now

IMHO, narrower context is better, so looks more natural to me.

> Reviewed-by: John Garry <[e-mail]>

Thanks!

Patch

diff --git a/lib/sbitmap.c b/lib/sbitmap.c
index 267aa7709416..c1c8a4e69325 100644
--- a/lib/sbitmap.c
+++ b/lib/sbitmap.c
@@ -12,32 +12,32 @@ 
 /*
  * See if we have deferred clears that we can batch move
  */
-static inline bool sbitmap_deferred_clear(struct sbitmap *sb, int index)
+static inline bool sbitmap_deferred_clear(struct sbitmap_word *map)
 {
 	unsigned long mask, val;
 	bool ret = false;
 	unsigned long flags;
 
-	spin_lock_irqsave(&sb->map[index].swap_lock, flags);
+	spin_lock_irqsave(&map->swap_lock, flags);
 
-	if (!sb->map[index].cleared)
+	if (!map->cleared)
 		goto out_unlock;
 
 	/*
 	 * First get a stable cleared mask, setting the old mask to 0.
 	 */
-	mask = xchg(&sb->map[index].cleared, 0);
+	mask = xchg(&map->cleared, 0);
 
 	/*
 	 * Now clear the masked bits in our free word
 	 */
 	do {
-		val = sb->map[index].word;
-	} while (cmpxchg(&sb->map[index].word, val, val & ~mask) != val);
+		val = map->word;
+	} while (cmpxchg(&map->word, val, val & ~mask) != val);
 
 	ret = true;
 out_unlock:
-	spin_unlock_irqrestore(&sb->map[index].swap_lock, flags);
+	spin_unlock_irqrestore(&map->swap_lock, flags);
 	return ret;
 }
 
@@ -92,7 +92,7 @@  void sbitmap_resize(struct sbitmap *sb, unsigned int depth)
 	unsigned int i;
 
 	for (i = 0; i < sb->map_nr; i++)
-		sbitmap_deferred_clear(sb, i);
+		sbitmap_deferred_clear(&sb->map[i]);
 
 	sb->depth = depth;
 	sb->map_nr = DIV_ROUND_UP(sb->depth, bits_per_word);
@@ -139,15 +139,15 @@  static int __sbitmap_get_word(unsigned long *word, unsigned long depth,
 static int sbitmap_find_bit_in_index(struct sbitmap *sb, int index,
 				     unsigned int alloc_hint, bool round_robin)
 {
+	struct sbitmap_word *map = &sb->map[index];
 	int nr;
 
 	do {
-		nr = __sbitmap_get_word(&sb->map[index].word,
-					sb->map[index].depth, alloc_hint,
+		nr = __sbitmap_get_word(&map->word, map->depth, alloc_hint,
 					!round_robin);
 		if (nr != -1)
 			break;
-		if (!sbitmap_deferred_clear(sb, index))
+		if (!sbitmap_deferred_clear(map))
 			break;
 	} while (1);
 
@@ -207,7 +207,7 @@  int sbitmap_get_shallow(struct sbitmap *sb, unsigned int alloc_hint,
 			break;
 		}
 
-		if (sbitmap_deferred_clear(sb, index))
+		if (sbitmap_deferred_clear(&sb->map[index]))
 			goto again;
 
 		/* Jump to next index. */