net: core: page_pool: add user refcnt and reintroduce page_pool_destroy

Jesper recently removed page_pool_destroy() (from driver invocation)
and moved shutdown and free of page_pool into xdp_rxq_info_unreg(),
in-order to handle in-flight packets/pages. This created an asymmetry
in drivers create/destroy pairs.

This patch reintroduce page_pool_destroy and add page_pool user
refcnt. This serves the purpose to simplify drivers error handling as
driver now drivers always calls page_pool_destroy() and don't need to
track if xdp_rxq_info_reg_mem_model() was unsuccessful.

This could be used for a special cases where a single RX-queue (with a
single page_pool) provides packets for two net_device'es, and thus
needs to register the same page_pool twice with two xdp_rxq_info
structures.

This patch is primarily to ease API usage for drivers. The recently
merged netsec driver, actually have a bug in this area, which is
solved by this API change.

This patch is a modified version of Ivan Khoronzhuk's original patch.

Link: https://lore.kernel.org/netdev/20190625175948.24771-2-ivan.khoronzhuk@linaro.org/
Fixes: 5c67bf0ec4 ("net: netsec: Use page_pool API")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Acked-by: Jesper Dangaard Brouer <brouer@redhat.com>
Reviewed-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Ivan Khoronzhuk <ivan.khoronzhuk@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/mellanox/mlx5/core/en_main.c

@@ -577,8 +577,6 @@ static int mlx5e_alloc_rq(struct mlx5e_channel *c,
 		}
 		err = xdp_rxq_info_reg_mem_model(&rq->xdp_rxq,
 						 MEM_TYPE_PAGE_POOL, rq->page_pool);
-		if (err)
-			page_pool_free(rq->page_pool);
 	}
 	if (err)
 		goto err_free;
@@ -646,6 +644,7 @@ err_rq_wq_destroy:
 	if (rq->xdp_prog)
 		bpf_prog_put(rq->xdp_prog);
 	xdp_rxq_info_unreg(&rq->xdp_rxq);
+	page_pool_destroy(rq->page_pool);
 	mlx5_wq_destroy(&rq->wq_ctrl);
 
 	return err;
@@ -680,6 +679,7 @@ static void mlx5e_free_rq(struct mlx5e_rq *rq)
 	}
 
 	xdp_rxq_info_unreg(&rq->xdp_rxq);
+	page_pool_destroy(rq->page_pool);
 	mlx5_wq_destroy(&rq->wq_ctrl);
 }
 

drivers/net/ethernet/socionext/netsec.c

@@ -1212,15 +1212,11 @@ static void netsec_uninit_pkt_dring(struct netsec_priv *priv, int id)
 		}
 	}
 
-	/* Rx is currently using page_pool
-	 * since the pool is created during netsec_setup_rx_dring(), we need to
-	 * free the pool manually if the registration failed
-	 */
+	/* Rx is currently using page_pool */
 	if (id == NETSEC_RING_RX) {
 		if (xdp_rxq_info_is_reg(&dring->xdp_rxq))
 			xdp_rxq_info_unreg(&dring->xdp_rxq);
-		else
-			page_pool_free(dring->page_pool);
+		page_pool_destroy(dring->page_pool);
 	}
 
 	memset(dring->desc, 0, sizeof(struct netsec_desc) * DESC_NUM);

include/net/page_pool.h

@@ -101,6 +101,12 @@ struct page_pool {
 	struct ptr_ring ring;
 
 	atomic_t pages_state_release_cnt;
+
+	/* A page_pool is strictly tied to a single RX-queue being
+	 * protected by NAPI, due to above pp_alloc_cache. This
+	 * refcnt serves purpose is to simplify drivers error handling.
+	 */
+	refcount_t user_cnt;
 };
 
 struct page *page_pool_alloc_pages(struct page_pool *pool, gfp_t gfp);
@@ -134,6 +140,15 @@ static inline void page_pool_free(struct page_pool *pool)
 #endif
 }
 
+/* Drivers use this instead of page_pool_free */
+static inline void page_pool_destroy(struct page_pool *pool)
+{
+	if (!pool)
+		return;
+
+	page_pool_free(pool);
+}
+
 /* Never call this directly, use helpers below */
 void __page_pool_put_page(struct page_pool *pool,
 			  struct page *page, bool allow_direct);
@@ -201,4 +216,14 @@ static inline bool is_page_pool_compiled_in(void)
 #endif
 }
 
+static inline void page_pool_get(struct page_pool *pool)
+{
+	refcount_inc(&pool->user_cnt);
+}
+
+static inline bool page_pool_put(struct page_pool *pool)
+{
+	return refcount_dec_and_test(&pool->user_cnt);
+}
+
 #endif /* _NET_PAGE_POOL_H */

net/core/page_pool.c

@@ -49,6 +49,9 @@ static int page_pool_init(struct page_pool *pool,
 
 	atomic_set(&pool->pages_state_release_cnt, 0);
 
+	/* Driver calling page_pool_create() also call page_pool_destroy() */
+	refcount_set(&pool->user_cnt, 1);
+
 	if (pool->p.flags & PP_FLAG_DMA_MAP)
 		get_device(pool->p.dev);
 
@@ -70,6 +73,7 @@ struct page_pool *page_pool_create(const struct page_pool_params *params)
 		kfree(pool);
 		return ERR_PTR(err);
 	}
+
 	return pool;
 }
 EXPORT_SYMBOL(page_pool_create);
@@ -356,6 +360,10 @@ static void __warn_in_flight(struct page_pool *pool)
 
 void __page_pool_free(struct page_pool *pool)
 {
+	/* Only last user actually free/release resources */
+	if (!page_pool_put(pool))
+		return;
+
 	WARN(pool->alloc.count, "API usage violation");
 	WARN(!ptr_ring_empty(&pool->ring), "ptr_ring is not empty");
 

net/core/xdp.c

@@ -370,6 +370,9 @@ int xdp_rxq_info_reg_mem_model(struct xdp_rxq_info *xdp_rxq,
 		goto err;
 	}
 
+	if (type == MEM_TYPE_PAGE_POOL)
+		page_pool_get(xdp_alloc->page_pool);
+
 	mutex_unlock(&mem_id_lock);
 
 	trace_mem_connect(xdp_alloc, xdp_rxq);