[PATCH] z3fold: make pages_nr atomic

[PATCH] z3fold: make pages_nr atomic

[Vitaly Wool]

This patch converts pages_nr per-pool counter to atomic64_t.

Signed-off-by: Vitaly Wool <vitalywool@gmail.com>
---
 mm/z3fold.c | 26 +++++++++++++++-----------
 1 file changed, 15 insertions(+), 11 deletions(-)

diff --git a/mm/z3fold.c b/mm/z3fold.c
index 8f9e89c..4d02280 100644
--- a/mm/z3fold.c
+++ b/mm/z3fold.c
@@ -80,7 +80,7 @@ struct z3fold_pool {
 	struct list_head unbuddied[NCHUNKS];
 	struct list_head buddied;
 	struct list_head lru;
-	u64 pages_nr;
+	atomic64_t pages_nr;
 	const struct z3fold_ops *ops;
 	struct zpool *zpool;
 	const struct zpool_ops *zpool_ops;
@@ -234,7 +234,7 @@ static struct z3fold_pool *z3fold_create_pool(gfp_t gfp,
 		INIT_LIST_HEAD(&pool->unbuddied[i]);
 	INIT_LIST_HEAD(&pool->buddied);
 	INIT_LIST_HEAD(&pool->lru);
-	pool->pages_nr = 0;
+	atomic64_set(&pool->pages_nr, 0);
 	pool->ops = ops;
 	return pool;
 }
@@ -346,7 +346,7 @@ static int z3fold_alloc(struct z3fold_pool *pool, size_t size, gfp_t gfp,
 	if (!page)
 		return -ENOMEM;
 	spin_lock(&pool->lock);
-	pool->pages_nr++;
+	atomic64_inc(&pool->pages_nr);
 	zhdr = init_z3fold_page(page);
 
 	if (bud == HEADLESS) {
@@ -439,10 +439,9 @@ static void z3fold_free(struct z3fold_pool *pool, unsigned long handle)
 		return;
 	}
 
-	if (bud != HEADLESS) {
-		/* Remove from existing buddy list */
+	/* Remove from existing buddy list */
+	if (bud != HEADLESS)
 		list_del(&zhdr->buddy);
-	}
 
 	if (bud == HEADLESS ||
 	    (zhdr->first_chunks == 0 && zhdr->middle_chunks == 0 &&
@@ -451,7 +451,7 @@ static void z3fold_free(struct z3fold_pool *pool, unsigned long handle)
 		list_del(&page->lru);
 		clear_bit(PAGE_HEADLESS, &page->private);
 		free_z3fold_page(zhdr);
-		pool->pages_nr--;
+		atomic64_dec(&pool->pages_nr);
 	} else {
 		z3fold_compact_page(zhdr);
 		/* Add to the unbuddied list */
@@ -569,7 +569,7 @@ static int z3fold_reclaim_page(struct z3fold_pool *pool, unsigned int retries)
 			 */
 			clear_bit(PAGE_HEADLESS, &page->private);
 			free_z3fold_page(zhdr);
-			pool->pages_nr--;
+			atomic64_dec(&pool->pages_nr);
 			spin_unlock(&pool->lock);
 			return 0;
 		}  else if (!test_bit(PAGE_HEADLESS, &page->private)) {
@@ -672,12 +672,11 @@ static void z3fold_unmap(struct z3fold_pool *pool, unsigned long handle)
  * z3fold_get_pool_size() - gets the z3fold pool size in pages
  * @pool:	pool whose size is being queried
  *
- * Returns: size in pages of the given pool.  The pool lock need not be
- * taken to access pages_nr.
+ * Returns: size in pages of the given pool.
  */
 static u64 z3fold_get_pool_size(struct z3fold_pool *pool)
 {
-	return pool->pages_nr;
+	return atomic64_read(&pool->pages_nr);
 }
 
 /*****************
-- 
2.4.2

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] z3fold: make pages_nr atomic

[Andrew Morton]

On Thu, 3 Nov 2016 22:00:58 +0100 Vitaly Wool <vitalywool@gmail.com> wrote:

> This patch converts pages_nr per-pool counter to atomic64_t.

Which is slower.

Presumably there is a reason for making this change.  This reason
should be described in the changelog.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] z3fold: make pages_nr atomic

[Vitaly Wool]

On Thu, Nov 3, 2016 at 10:14 PM, Andrew Morton
<akpm@linux-foundation.org> wrote:
> On Thu, 3 Nov 2016 22:00:58 +0100 Vitaly Wool <vitalywool@gmail.com> wrote:
>
>> This patch converts pages_nr per-pool counter to atomic64_t.
>
> Which is slower.
>
> Presumably there is a reason for making this change.  This reason
> should be described in the changelog.

The reason [which I thought was somewhat obvious :) ] is that there
won't be a need to take a per-pool lock to read or modify that
counter.

Best regards,
   Vitaly

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] z3fold: make pages_nr atomic

[Andrew Morton]

On Thu, 3 Nov 2016 22:24:07 +0100 Vitaly Wool <vitalywool@gmail.com> wrote:

> On Thu, Nov 3, 2016 at 10:14 PM, Andrew Morton
> <akpm@linux-foundation.org> wrote:
> > On Thu, 3 Nov 2016 22:00:58 +0100 Vitaly Wool <vitalywool@gmail.com> wrote:
> >
> >> This patch converts pages_nr per-pool counter to atomic64_t.
> >
> > Which is slower.
> >
> > Presumably there is a reason for making this change.  This reason
> > should be described in the changelog.
> 
> The reason [which I thought was somewhat obvious :) ] is that there
> won't be a need to take a per-pool lock to read or modify that
> counter.

But the patch didn't change the locking.  And as far as I can tell,
neither does "z3fold: extend compaction function".

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] z3fold: make pages_nr atomic

[Vitaly Wool]

On Thu, Nov 3, 2016 at 11:17 PM, Andrew Morton
<akpm@linux-foundation.org> wrote:
> On Thu, 3 Nov 2016 22:24:07 +0100 Vitaly Wool <vitalywool@gmail.com> wrote:
>
>> On Thu, Nov 3, 2016 at 10:14 PM, Andrew Morton
>> <akpm@linux-foundation.org> wrote:
>> > On Thu, 3 Nov 2016 22:00:58 +0100 Vitaly Wool <vitalywool@gmail.com> wrote:
>> >
>> >> This patch converts pages_nr per-pool counter to atomic64_t.
>> >
>> > Which is slower.
>> >
>> > Presumably there is a reason for making this change.  This reason
>> > should be described in the changelog.
>>
>> The reason [which I thought was somewhat obvious :) ] is that there
>> won't be a need to take a per-pool lock to read or modify that
>> counter.
>
> But the patch didn't change the locking.  And as far as I can tell,
> neither does "z3fold: extend compaction function".

Right. I'll come up with the locking rework shortly, but it will be a
RFC so I wanted to send it separately.

~vitaly

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH] z3fold: make pages_nr atomic

[Dan Streetman]

On Fri, Nov 4, 2016 at 3:27 AM, Vitaly Wool <vitalywool@gmail.com> wrote:
> On Thu, Nov 3, 2016 at 11:17 PM, Andrew Morton
> <akpm@linux-foundation.org> wrote:
>> On Thu, 3 Nov 2016 22:24:07 +0100 Vitaly Wool <vitalywool@gmail.com> wrote:
>>
>>> On Thu, Nov 3, 2016 at 10:14 PM, Andrew Morton
>>> <akpm@linux-foundation.org> wrote:
>>> > On Thu, 3 Nov 2016 22:00:58 +0100 Vitaly Wool <vitalywool@gmail.com> wrote:
>>> >
>>> >> This patch converts pages_nr per-pool counter to atomic64_t.
>>> >
>>> > Which is slower.
>>> >
>>> > Presumably there is a reason for making this change.  This reason
>>> > should be described in the changelog.
>>>
>>> The reason [which I thought was somewhat obvious :) ] is that there
>>> won't be a need to take a per-pool lock to read or modify that
>>> counter.
>>
>> But the patch didn't change the locking.  And as far as I can tell,
>> neither does "z3fold: extend compaction function".
>
> Right. I'll come up with the locking rework shortly, but it will be a
> RFC so I wanted to send it separately.

this is still in mmotm, and it seems the later patches rely on it, so
while i agree that the changelog should be clearer about why it's
needed, the change itself looks ok.

Acked-by: Dan Streetman <ddstreet@ieee.org>

>
> ~vitaly

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>