[PATCH RFC net-next/mm V1 2/3] page_pool: Use static_key for shutdown phase

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Performance is very important for page pool (PP). This add the use of
static_key APIs for regaining a single instruction, which makes the
new PP shutdown scheme zero impact.

We are uncertain if this is 100% correct, because static_key APIs uses
a mutex lock and it is uncertain if all contexts that can return pages
can support this. We could spawn a workqueue (like we just removed) to
workaround this issue.

Seeking input if this is worth the complexity.

Signed-off-by: Jesper Dangaard Brouer <brouer@xxxxxxxxxx>
---
 net/core/page_pool.c |   20 ++++++++++++++++----
 1 file changed, 16 insertions(+), 4 deletions(-)

diff --git a/net/core/page_pool.c b/net/core/page_pool.c
index ce7e8dda6403..3821d8874b15 100644
--- a/net/core/page_pool.c
+++ b/net/core/page_pool.c
@@ -25,6 +25,8 @@
 
 #define BIAS_MAX	LONG_MAX
 
+DEFINE_STATIC_KEY_FALSE(pp_shutdown_phase);
+
 #ifdef CONFIG_PAGE_POOL_STATS
 /* alloc_stat_inc is intended to be used in softirq context */
 #define alloc_stat_inc(pool, __stat)	(pool->alloc_stats.__stat++)
@@ -378,7 +380,8 @@ static struct page *__page_pool_alloc_pages_slow(struct page_pool *pool,
 	int i, nr_pages;
 
 	/* API usage BUG: PP in shutdown phase, cannot alloc new pages */
-	if (WARN_ON(pool->p.flags & PP_FLAG_SHUTDOWN))
+	if (static_key_enabled(&pp_shutdown_phase) &&
+	    WARN_ON(pool->p.flags & PP_FLAG_SHUTDOWN))
 		return NULL;
 
 	/* Don't support bulk alloc for high-order pages */
@@ -609,7 +612,7 @@ void page_pool_put_defragged_page(struct page_pool *pool, struct page *page,
 		recycle_stat_inc(pool, ring_full);
 		page_pool_return_page(pool, page);
 	}
-	if (pool->p.flags & PP_FLAG_SHUTDOWN)
+	if (static_branch_unlikely(&pp_shutdown_phase))
 		page_pool_shutdown_attempt(pool);
 }
 EXPORT_SYMBOL(page_pool_put_defragged_page);
@@ -659,7 +662,7 @@ void page_pool_put_page_bulk(struct page_pool *pool, void **data,
 		page_pool_return_page(pool, data[i]);
 
 out:
-	if (pool->p.flags & PP_FLAG_SHUTDOWN)
+	if (static_branch_unlikely(&pp_shutdown_phase))
 		page_pool_shutdown_attempt(pool);
 }
 EXPORT_SYMBOL(page_pool_put_page_bulk);
@@ -817,7 +820,15 @@ static int page_pool_release(struct page_pool *pool)
 noinline
 static void page_pool_shutdown_attempt(struct page_pool *pool)
 {
-	page_pool_release(pool);
+	int inflight;
+
+	if (!(pool->p.flags & PP_FLAG_SHUTDOWN))
+		return;
+
+	inflight = page_pool_release(pool);
+
+	if (static_key_enabled(&pp_shutdown_phase) && !inflight)
+		static_branch_dec(&pp_shutdown_phase);
 }
 
 void page_pool_use_xdp_mem(struct page_pool *pool, void (*disconnect)(void *),
@@ -861,6 +872,7 @@ void page_pool_destroy(struct page_pool *pool)
 	 * Enter into shutdown phase, and retry release to handle races.
 	 */
 	pool->p.flags |= PP_FLAG_SHUTDOWN;
+	static_branch_inc(&pp_shutdown_phase);
 	page_pool_shutdown_attempt(pool);
 }
 EXPORT_SYMBOL(page_pool_destroy);





[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux