net/mlx5e: Do not modify the TX SKB

If the SKB is cloned, or has an elevated users count, someone else
can be looking at it at the same time.

Signed-off-by: Achiad Shochat <[email protected]>
Signed-off-by: Saeed Mahameed <[email protected]>
Signed-off-by: David S. Miller <[email protected]>

Author: Achiad Shochat

drivers/net/ethernet/mellanox/mlx5/core/en.h

@@ -328,14 +328,12 @@ struct mlx5e_rq {
 	struct mlx5e_priv     *priv;
 } ____cacheline_aligned_in_smp;
 
-struct mlx5e_tx_skb_cb {
+struct mlx5e_tx_wqe_info {
 	u32 num_bytes;
 	u8  num_wqebbs;
 	u8  num_dma;
 };
 
-#define MLX5E_TX_SKB_CB(__skb) ((struct mlx5e_tx_skb_cb *)__skb->cb)
-
 enum mlx5e_dma_map_type {
 	MLX5E_DMA_MAP_SINGLE,
 	MLX5E_DMA_MAP_PAGE
@@ -371,6 +369,7 @@ struct mlx5e_sq {
 	/* pointers to per packet info: write@xmit, read@completion */
 	struct sk_buff           **skb;
 	struct mlx5e_sq_dma       *dma_fifo;
+	struct mlx5e_tx_wqe_info  *wqe_info;
 
 	/* read only */
 	struct mlx5_wq_cyc         wq;

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

@@ -507,6 +507,7 @@ static void mlx5e_close_rq(struct mlx5e_rq *rq)
 
 static void mlx5e_free_sq_db(struct mlx5e_sq *sq)
 {
+	kfree(sq->wqe_info);
 	kfree(sq->dma_fifo);
 	kfree(sq->skb);
 }
@@ -519,8 +520,10 @@ static int mlx5e_alloc_sq_db(struct mlx5e_sq *sq, int numa)
 	sq->skb = kzalloc_node(wq_sz * sizeof(*sq->skb), GFP_KERNEL, numa);
 	sq->dma_fifo = kzalloc_node(df_sz * sizeof(*sq->dma_fifo), GFP_KERNEL,
 				    numa);
+	sq->wqe_info = kzalloc_node(wq_sz * sizeof(*sq->wqe_info), GFP_KERNEL,
+				    numa);
 
-	if (!sq->skb || !sq->dma_fifo) {
+	if (!sq->skb || !sq->dma_fifo || !sq->wqe_info) {
 		mlx5e_free_sq_db(sq);
 		return -ENOMEM;
 	}

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

@@ -92,11 +92,11 @@ static inline struct mlx5e_sq_dma *mlx5e_dma_get(struct mlx5e_sq *sq, u32 i)
 	return &sq->dma_fifo[i & sq->dma_fifo_mask];
 }
 
-static void mlx5e_dma_unmap_wqe_err(struct mlx5e_sq *sq, struct sk_buff *skb)
+static void mlx5e_dma_unmap_wqe_err(struct mlx5e_sq *sq, u8 num_dma)
 {
 	int i;
 
-	for (i = 0; i < MLX5E_TX_SKB_CB(skb)->num_dma; i++) {
+	for (i = 0; i < num_dma; i++) {
 		struct mlx5e_sq_dma *last_pushed_dma =
 			mlx5e_dma_get(sq, --sq->dma_fifo_pc);
 
@@ -139,19 +139,28 @@ static inline u16 mlx5e_get_inline_hdr_size(struct mlx5e_sq *sq,
 	return MLX5E_MIN_INLINE;
 }
 
-static inline void mlx5e_insert_vlan(void *start, struct sk_buff *skb, u16 ihs)
+static inline void mlx5e_tx_skb_pull_inline(unsigned char **skb_data,
+					    unsigned int *skb_len,
+					    unsigned int len)
+{
+	*skb_len -= len;
+	*skb_data += len;
+}
+
+static inline void mlx5e_insert_vlan(void *start, struct sk_buff *skb, u16 ihs,
+				     unsigned char **skb_data,
+				     unsigned int *skb_len)
 {
 	struct vlan_ethhdr *vhdr = (struct vlan_ethhdr *)start;
 	int cpy1_sz = 2 * ETH_ALEN;
 	int cpy2_sz = ihs - cpy1_sz;
 
-	skb_copy_from_linear_data(skb, vhdr, cpy1_sz);
-	skb_pull_inline(skb, cpy1_sz);
+	memcpy(vhdr, *skb_data, cpy1_sz);
+	mlx5e_tx_skb_pull_inline(skb_data, skb_len, cpy1_sz);
 	vhdr->h_vlan_proto = skb->vlan_proto;
 	vhdr->h_vlan_TCI = cpu_to_be16(skb_vlan_tag_get(skb));
-	skb_copy_from_linear_data(skb, &vhdr->h_vlan_encapsulated_proto,
-				  cpy2_sz);
-	skb_pull_inline(skb, cpy2_sz);
+	memcpy(&vhdr->h_vlan_encapsulated_proto, *skb_data, cpy2_sz);
+	mlx5e_tx_skb_pull_inline(skb_data, skb_len, cpy2_sz);
 }
 
 static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
@@ -160,11 +169,14 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
 
 	u16 pi = sq->pc & wq->sz_m1;
 	struct mlx5e_tx_wqe      *wqe  = mlx5_wq_cyc_get_wqe(wq, pi);
+	struct mlx5e_tx_wqe_info *wi   = &sq->wqe_info[pi];
 
 	struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
 	struct mlx5_wqe_eth_seg  *eseg = &wqe->eth;
 	struct mlx5_wqe_data_seg *dseg;
 
+	unsigned char *skb_data = skb->data;
+	unsigned int skb_len = skb->len;
 	u8  opcode = MLX5_OPCODE_SEND;
 	dma_addr_t dma_addr = 0;
 	bool bf = false;
@@ -192,25 +204,25 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
 		opcode       = MLX5_OPCODE_LSO;
 		ihs          = skb_transport_offset(skb) + tcp_hdrlen(skb);
 		payload_len  = skb->len - ihs;
-		MLX5E_TX_SKB_CB(skb)->num_bytes = skb->len +
-					(skb_shinfo(skb)->gso_segs - 1) * ihs;
+		wi->num_bytes = skb->len +
+				(skb_shinfo(skb)->gso_segs - 1) * ihs;
 		sq->stats.tso_packets++;
 		sq->stats.tso_bytes += payload_len;
 	} else {
 		bf = sq->bf_budget &&
 		     !skb->xmit_more &&
 		     !skb_shinfo(skb)->nr_frags;
 		ihs = mlx5e_get_inline_hdr_size(sq, skb, bf);
-		MLX5E_TX_SKB_CB(skb)->num_bytes = max_t(unsigned int, skb->len,
-							ETH_ZLEN);
+		wi->num_bytes = max_t(unsigned int, skb->len, ETH_ZLEN);
 	}
 
 	if (skb_vlan_tag_present(skb)) {
-		mlx5e_insert_vlan(eseg->inline_hdr_start, skb, ihs);
+		mlx5e_insert_vlan(eseg->inline_hdr_start, skb, ihs, &skb_data,
+				  &skb_len);
 		ihs += VLAN_HLEN;
 	} else {
-		skb_copy_from_linear_data(skb, eseg->inline_hdr_start, ihs);
-		skb_pull_inline(skb, ihs);
+		memcpy(eseg->inline_hdr_start, skb_data, ihs);
+		mlx5e_tx_skb_pull_inline(&skb_data, &skb_len, ihs);
 	}
 
 	eseg->inline_hdr_sz = cpu_to_be16(ihs);
@@ -220,11 +232,11 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
 			       MLX5_SEND_WQE_DS);
 	dseg    = (struct mlx5_wqe_data_seg *)cseg + ds_cnt;
 
-	MLX5E_TX_SKB_CB(skb)->num_dma = 0;
+	wi->num_dma = 0;
 
-	headlen = skb_headlen(skb);
+	headlen = skb_len - skb->data_len;
 	if (headlen) {
-		dma_addr = dma_map_single(sq->pdev, skb->data, headlen,
+		dma_addr = dma_map_single(sq->pdev, skb_data, headlen,
 					  DMA_TO_DEVICE);
 		if (unlikely(dma_mapping_error(sq->pdev, dma_addr)))
 			goto dma_unmap_wqe_err;
@@ -234,7 +246,7 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
 		dseg->byte_count = cpu_to_be32(headlen);
 
 		mlx5e_dma_push(sq, dma_addr, headlen, MLX5E_DMA_MAP_SINGLE);
-		MLX5E_TX_SKB_CB(skb)->num_dma++;
+		wi->num_dma++;
 
 		dseg++;
 	}
@@ -253,23 +265,22 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
 		dseg->byte_count = cpu_to_be32(fsz);
 
 		mlx5e_dma_push(sq, dma_addr, fsz, MLX5E_DMA_MAP_PAGE);
-		MLX5E_TX_SKB_CB(skb)->num_dma++;
+		wi->num_dma++;
 
 		dseg++;
 	}
 
-	ds_cnt += MLX5E_TX_SKB_CB(skb)->num_dma;
+	ds_cnt += wi->num_dma;
 
 	cseg->opmod_idx_opcode = cpu_to_be32((sq->pc << 8) | opcode);
 	cseg->qpn_ds           = cpu_to_be32((sq->sqn << 8) | ds_cnt);
 
 	sq->skb[pi] = skb;
 
-	MLX5E_TX_SKB_CB(skb)->num_wqebbs = DIV_ROUND_UP(ds_cnt,
-							MLX5_SEND_WQEBB_NUM_DS);
-	sq->pc += MLX5E_TX_SKB_CB(skb)->num_wqebbs;
+	wi->num_wqebbs = DIV_ROUND_UP(ds_cnt, MLX5_SEND_WQEBB_NUM_DS);
+	sq->pc += wi->num_wqebbs;
 
-	netdev_tx_sent_queue(sq->txq, MLX5E_TX_SKB_CB(skb)->num_bytes);
+	netdev_tx_sent_queue(sq->txq, wi->num_bytes);
 
 	if (unlikely(!mlx5e_sq_has_room_for(sq, MLX5E_SQ_STOP_ROOM))) {
 		netif_tx_stop_queue(sq->txq);
@@ -280,7 +291,7 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
 		int bf_sz = 0;
 
 		if (bf && sq->uar_bf_map)
-			bf_sz = MLX5E_TX_SKB_CB(skb)->num_wqebbs << 3;
+			bf_sz = wi->num_wqebbs << 3;
 
 		cseg->fm_ce_se = MLX5_WQE_CTRL_CQ_UPDATE;
 		mlx5e_tx_notify_hw(sq, wqe, bf_sz);
@@ -297,7 +308,7 @@ static netdev_tx_t mlx5e_sq_xmit(struct mlx5e_sq *sq, struct sk_buff *skb)
 
 dma_unmap_wqe_err:
 	sq->stats.dropped++;
-	mlx5e_dma_unmap_wqe_err(sq, skb);
+	mlx5e_dma_unmap_wqe_err(sq, wi->num_dma);
 
 	dev_kfree_skb_any(skb);
 
@@ -352,6 +363,7 @@ bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq)
 		wqe_counter = be16_to_cpu(cqe->wqe_counter);
 
 		do {
+			struct mlx5e_tx_wqe_info *wi;
 			struct sk_buff *skb;
 			u16 ci;
 			int j;
@@ -360,23 +372,24 @@ bool mlx5e_poll_tx_cq(struct mlx5e_cq *cq)
 
 			ci = sqcc & sq->wq.sz_m1;
 			skb = sq->skb[ci];
+			wi = &sq->wqe_info[ci];
 
 			if (unlikely(!skb)) { /* nop */
 				sq->stats.nop++;
 				sqcc++;
 				continue;
 			}
 
-			for (j = 0; j < MLX5E_TX_SKB_CB(skb)->num_dma; j++) {
+			for (j = 0; j < wi->num_dma; j++) {
 				struct mlx5e_sq_dma *dma =
 					mlx5e_dma_get(sq, dma_fifo_cc++);
 
 				mlx5e_tx_dma_unmap(sq->pdev, dma);
 			}
 
 			npkts++;
-			nbytes += MLX5E_TX_SKB_CB(skb)->num_bytes;
-			sqcc += MLX5E_TX_SKB_CB(skb)->num_wqebbs;
+			nbytes += wi->num_bytes;
+			sqcc += wi->num_wqebbs;
 			dev_kfree_skb(skb);
 		} while (!last_wqe);
 	}