net/mlx5e: RX, Enhance legacy Receive Queue memory scheme

Enhance the memory scheme of the legacy RQ, such that
only order-0 pages are used.

Whenever possible, prefer using a linear SKB, and build it
wrapping the WQE buffer.

Otherwise (for example, jumbo frames on x86), use non-linear SKB,
with as many frags as needed. In this case, multiple WQE
scatter entries are used, up to a maximum of 4 frags and 10KB of MTU.

This implied to remove support of HW LRO in legacy RQ, as it would
require large number of page allocations and scatter entries per WQE
on archs with PAGE_SIZE = 4KB, yielding bad performance.

In earlier patches, we guaranteed that all completions are in-order,
and that we use a cyclic WQ.
This creates an oppurtunity for a performance optimization:
The mapping between a "struct mlx5e_dma_info", and the
WQEs (struct mlx5e_wqe_frag_info) pointing to it, is constant
across different cycles of a WQ. This allows initializing
the mapping in the time of RQ creation, and not handle it
in datapath.

A struct mlx5e_dma_info that is shared between different WQEs
is allocated by the first WQE, and freed by the last one.
This implies an important requirement: WQEs that share the same
struct mlx5e_dma_info must be posted within the same NAPI.
Otherwise, upon completion, struct mlx5e_wqe_frag_info would mistakenly
point to the new struct mlx5e_dma_info, not the one that was posted
(and the HW wrote to).
This bulking requirement is actually good also for performance reasons,
hence we extend the bulk beyong the minimal requirement above.

With this memory scheme, the RQs memory footprint is reduce by a
factor of 2 on x86, and by a factor of 32 on PowerPC.
Same factors apply for the number of pages in a GRO session.

Performance tests:
ConnectX-4, single core, single RX ring, default MTU.

x86:
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz

Packet rate (early drop in TC): no degradation
TCP streams: ~5% improvement

PowerPC:
CPU: POWER8 (raw), altivec supported

Packet rate (early drop in TC): 20% gain
TCP streams: 25% gain

Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

1 files changed, 147 insertions, 60 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
index 3cdf2c097356..d3a1dd20e41d 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rx.c
@@ -164,8 +164,6 @@ static inline u32 mlx5e_decompress_cqes_start(struct mlx5e_rq *rq,
 	return mlx5e_decompress_cqes_cont(rq, cq, 1, budget_rem) - 1;
 }
 
-#define RQ_PAGE_SIZE(rq) ((1 << rq->buff.page_order) << PAGE_SHIFT)
-
 static inline bool mlx5e_page_is_reserved(struct page *page)
 {
 	return page_is_pfmemalloc(page) || page_to_nid(page) != numa_mem_id();
@@ -214,7 +212,7 @@ static inline bool mlx5e_rx_cache_get(struct mlx5e_rq *rq,
 	stats->cache_reuse++;
 
 	dma_sync_single_for_device(rq->pdev, dma_info->addr,
-				   RQ_PAGE_SIZE(rq),
+				   PAGE_SIZE,
 				   DMA_FROM_DEVICE);
 	return true;
 }
@@ -230,7 +228,7 @@ static inline int mlx5e_page_alloc_mapped(struct mlx5e_rq *rq,
 		return -ENOMEM;
 
 	dma_info->addr = dma_map_page(rq->pdev, dma_info->page, 0,
-				      RQ_PAGE_SIZE(rq), rq->buff.map_dir);
+				      PAGE_SIZE, rq->buff.map_dir);
 	if (unlikely(dma_mapping_error(rq->pdev, dma_info->addr))) {
 		put_page(dma_info->page);
 		dma_info->page = NULL;
@@ -243,8 +241,7 @@ static inline int mlx5e_page_alloc_mapped(struct mlx5e_rq *rq,
 static void mlx5e_page_dma_unmap(struct mlx5e_rq *rq,
 					struct mlx5e_dma_info *dma_info)
 {
-	dma_unmap_page(rq->pdev, dma_info->addr, RQ_PAGE_SIZE(rq),
-		       rq->buff.map_dir);
+	dma_unmap_page(rq->pdev, dma_info->addr, PAGE_SIZE, rq->buff.map_dir);
 }
 
 void mlx5e_page_release(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info,
@@ -262,58 +259,96 @@ void mlx5e_page_release(struct mlx5e_rq *rq, struct mlx5e_dma_info *dma_info,
 	}
 }
 
-static inline bool mlx5e_page_reuse(struct mlx5e_rq *rq,
-				    struct mlx5e_wqe_frag_info *wi)
+static inline int mlx5e_get_rx_frag(struct mlx5e_rq *rq,
+				    struct mlx5e_wqe_frag_info *frag)
+{
+	int err = 0;
+
+	if (!frag->offset)
+		/* On first frag (offset == 0), replenish page (dma_info actually).
+		 * Other frags that point to the same dma_info (with a different
+		 * offset) should just use the new one without replenishing again
+		 * by themselves.
+		 */
+		err = mlx5e_page_alloc_mapped(rq, frag->di);
+
+	return err;
+}
+
+static inline void mlx5e_put_rx_frag(struct mlx5e_rq *rq,
+				     struct mlx5e_wqe_frag_info *frag)
 {
-	return rq->wqe.page_reuse && wi->di.page &&
-		(wi->offset + rq->wqe.frag_sz <= RQ_PAGE_SIZE(rq)) &&
-		!mlx5e_page_is_reserved(wi->di.page);
+	if (frag->last_in_page)
+		mlx5e_page_release(rq, frag->di, true);
 }
 
 static inline struct mlx5e_wqe_frag_info *get_frag(struct mlx5e_rq *rq, u16 ix)
 {
-	return &rq->wqe.frag_info[ix];
+	return &rq->wqe.frags[ix << rq->wqe.info.log_num_frags];
 }
 
-static int mlx5e_alloc_rx_wqe(struct mlx5e_rq *rq, struct mlx5e_rx_wqe_cyc *wqe, u16 ix)
+static int mlx5e_alloc_rx_wqe(struct mlx5e_rq *rq, struct mlx5e_rx_wqe_cyc *wqe,
+			      u16 ix)
 {
-	struct mlx5e_wqe_frag_info *wi = &rq->wqe.frag_info[ix];
+	struct mlx5e_wqe_frag_info *frag = get_frag(rq, ix);
+	int err;
+	int i;
 
-	/* check if page exists, hence can be reused */
-	if (!wi->di.page) {
-		if (unlikely(mlx5e_page_alloc_mapped(rq, &wi->di)))
-			return -ENOMEM;
-		wi->offset = 0;
+	for (i = 0; i < rq->wqe.info.num_frags; i++, frag++) {
+		err = mlx5e_get_rx_frag(rq, frag);
+		if (unlikely(err))
+			goto free_frags;
+
+		wqe->data[i].addr = cpu_to_be64(frag->di->addr +
+						frag->offset + rq->buff.headroom);
 	}
 
-	wqe->data[0].addr = cpu_to_be64(wi->di.addr + wi->offset + rq->buff.headroom);
 	return 0;
+
+free_frags:
+	while (--i >= 0)
+		mlx5e_put_rx_frag(rq, --frag);
+
+	return err;
 }
 
 static inline void mlx5e_free_rx_wqe(struct mlx5e_rq *rq,
 				     struct mlx5e_wqe_frag_info *wi)
 {
-	mlx5e_page_release(rq, &wi->di, true);
-	wi->di.page = NULL;
+	int i;
+
+	for (i = 0; i < rq->wqe.info.num_frags; i++, wi++)
+		mlx5e_put_rx_frag(rq, wi);
 }
 
-static inline void mlx5e_free_rx_wqe_reuse(struct mlx5e_rq *rq,
-					   struct mlx5e_wqe_frag_info *wi)
+void mlx5e_dealloc_rx_wqe(struct mlx5e_rq *rq, u16 ix)
 {
-	if (mlx5e_page_reuse(rq, wi)) {
-		rq->stats->page_reuse++;
-		return;
-	}
+	struct mlx5e_wqe_frag_info *wi = get_frag(rq, ix);
 
 	mlx5e_free_rx_wqe(rq, wi);
 }
 
-void mlx5e_dealloc_rx_wqe(struct mlx5e_rq *rq, u16 ix)
+static int mlx5e_alloc_rx_wqes(struct mlx5e_rq *rq, u16 ix, u8 wqe_bulk)
 {
-	struct mlx5e_wqe_frag_info *wi = &rq->wqe.frag_info[ix];
+	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+	int err;
+	int i;
 
-	if (wi->di.page)
-		mlx5e_free_rx_wqe(rq, wi);
+	for (i = 0; i < wqe_bulk; i++) {
+		struct mlx5e_rx_wqe_cyc *wqe = mlx5_wq_cyc_get_wqe(wq, ix + i);
+
+		err = mlx5e_alloc_rx_wqe(rq, wqe, ix + i);
+		if (unlikely(err))
+			goto free_wqes;
+	}
+
+	return 0;
+
+free_wqes:
+	while (--i >= 0)
+		mlx5e_dealloc_rx_wqe(rq, ix + i);
+
+	return err;
 }
 
 static inline void
@@ -476,26 +511,28 @@ void mlx5e_dealloc_rx_mpwqe(struct mlx5e_rq *rq, u16 ix)
 bool mlx5e_post_rx_wqes(struct mlx5e_rq *rq)
 {
 	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
+	u8 wqe_bulk;
 	int err;
 
 	if (unlikely(!test_bit(MLX5E_RQ_STATE_ENABLED, &rq->state)))
 		return false;
 
-	if (mlx5_wq_cyc_is_full(wq))
+	wqe_bulk = rq->wqe.info.wqe_bulk;
+
+	if (mlx5_wq_cyc_missing(wq) < wqe_bulk)
 		return false;
 
 	do {
 		u16 head = mlx5_wq_cyc_get_head(wq);
-		struct mlx5e_rx_wqe_cyc *wqe = mlx5_wq_cyc_get_wqe(wq, head);
 
-		err = mlx5e_alloc_rx_wqe(rq, wqe, head);
+		err = mlx5e_alloc_rx_wqes(rq, head, wqe_bulk);
 		if (unlikely(err)) {
 			rq->stats->buff_alloc_err++;
 			break;
 		}
 
-		mlx5_wq_cyc_push(wq);
-	} while (!mlx5_wq_cyc_is_full(wq));
+		mlx5_wq_cyc_push_n(wq, wqe_bulk);
+	} while (mlx5_wq_cyc_missing(wq) >= wqe_bulk);
 
 	/* ensure wqes are visible to device before updating doorbell record */
 	dma_wmb();
@@ -949,11 +986,11 @@ struct sk_buff *mlx5e_build_linear_skb(struct mlx5e_rq *rq, void *va,
 	return skb;
 }
 
-static inline
-struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
-			     struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt)
+struct sk_buff *
+mlx5e_skb_from_cqe_linear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			  struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt)
 {
-	struct mlx5e_dma_info *di = &wi->di;
+	struct mlx5e_dma_info *di = wi->di;
 	u16 rx_headroom = rq->buff.headroom;
 	struct sk_buff *skb;
 	void *va, *data;
@@ -968,7 +1005,6 @@ struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
 				      frag_size, DMA_FROM_DEVICE);
 	prefetchw(va); /* xdp_frame data area */
 	prefetch(data);
-	wi->offset += frag_size;
 
 	if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) {
 		rq->stats->wqe_err++;
@@ -991,6 +1027,56 @@ struct sk_buff *skb_from_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
 	return skb;
 }
 
+struct sk_buff *
+mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			     struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt)
+{
+	struct mlx5e_rq_frag_info *frag_info = &rq->wqe.info.arr[0];
+	struct mlx5e_wqe_frag_info *head_wi = wi;
+	u16 headlen      = min_t(u32, MLX5E_RX_MAX_HEAD, cqe_bcnt);
+	u16 frag_headlen = headlen;
+	u16 byte_cnt     = cqe_bcnt - headlen;
+	struct sk_buff *skb;
+
+	if (unlikely((cqe->op_own >> 4) != MLX5_CQE_RESP_SEND)) {
+		rq->stats->wqe_err++;
+		return NULL;
+	}
+
+	/* XDP is not supported in this configuration, as incoming packets
+	 * might spread among multiple pages.
+	 */
+	skb = napi_alloc_skb(rq->cq.napi,
+			     ALIGN(MLX5E_RX_MAX_HEAD, sizeof(long)));
+	if (unlikely(!skb)) {
+		rq->stats->buff_alloc_err++;
+		return NULL;
+	}
+
+	prefetchw(skb->data);
+
+	while (byte_cnt) {
+		u16 frag_consumed_bytes =
+			min_t(u16, frag_info->frag_size - frag_headlen, byte_cnt);
+
+		mlx5e_add_skb_frag(rq, skb, wi->di, wi->offset + frag_headlen,
+				   frag_consumed_bytes, frag_info->frag_stride);
+		byte_cnt -= frag_consumed_bytes;
+		frag_headlen = 0;
+		frag_info++;
+		wi++;
+	}
+
+	/* copy header */
+	mlx5e_copy_skb_header(rq->pdev, skb, head_wi->di, head_wi->offset,
+			      0, headlen);
+	/* skb linear part was allocated with headlen and aligned to long */
+	skb->tail += headlen;
+	skb->len  += headlen;
+
+	return skb;
+}
+
 void mlx5e_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 {
 	struct mlx5_wq_cyc *wq = &rq->wqe.wq;
@@ -1003,23 +1089,23 @@ void mlx5e_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	wi       = get_frag(rq, ci);
 	cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
 
-	skb = skb_from_cqe(rq, cqe, wi, cqe_bcnt);
+	skb = rq->wqe.skb_from_cqe(rq, cqe, wi, cqe_bcnt);
 	if (!skb) {
 		/* probably for XDP */
 		if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) {
-			wi->di.page = NULL;
-			/* do not return page to cache, it will be returned on XDP_TX completion */
+			/* do not return page to cache,
+			 * it will be returned on XDP_TX completion.
+			 */
 			goto wq_cyc_pop;
 		}
-		/* probably an XDP_DROP, save the page-reuse checks */
-		mlx5e_free_rx_wqe(rq, wi);
-		goto wq_cyc_pop;
+		goto free_wqe;
 	}
 
 	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
 	napi_gro_receive(rq->cq.napi, skb);
 
-	mlx5e_free_rx_wqe_reuse(rq, wi);
+free_wqe:
+	mlx5e_free_rx_wqe(rq, wi);
 wq_cyc_pop:
 	mlx5_wq_cyc_pop(wq);
 }
@@ -1041,16 +1127,16 @@ void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	wi       = get_frag(rq, ci);
 	cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
 
-	skb = skb_from_cqe(rq, cqe, wi, cqe_bcnt);
+	skb = rq->wqe.skb_from_cqe(rq, cqe, wi, cqe_bcnt);
 	if (!skb) {
+		/* probably for XDP */
 		if (__test_and_clear_bit(MLX5E_RQ_FLAG_XDP_XMIT, rq->flags)) {
-			wi->di.page = NULL;
-			/* do not return page to cache, it will be returned on XDP_TX completion */
+			/* do not return page to cache,
+			 * it will be returned on XDP_TX completion.
+			 */
 			goto wq_cyc_pop;
 		}
-		/* probably an XDP_DROP, save the page-reuse checks */
-		mlx5e_free_rx_wqe(rq, wi);
-		goto wq_cyc_pop;
+		goto free_wqe;
 	}
 
 	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
@@ -1060,7 +1146,8 @@ void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 
 	napi_gro_receive(rq->cq.napi, skb);
 
-	mlx5e_free_rx_wqe_reuse(rq, wi);
+free_wqe:
+	mlx5e_free_rx_wqe(rq, wi);
 wq_cyc_pop:
 	mlx5_wq_cyc_pop(wq);
 }
@@ -1409,7 +1496,7 @@ void mlx5i_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	wi       = get_frag(rq, ci);
 	cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
 
-	skb = skb_from_cqe(rq, cqe, wi, cqe_bcnt);
+	skb = rq->wqe.skb_from_cqe(rq, cqe, wi, cqe_bcnt);
 	if (!skb)
 		goto wq_free_wqe;
 
@@ -1421,7 +1508,7 @@ void mlx5i_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	napi_gro_receive(rq->cq.napi, skb);
 
 wq_free_wqe:
-	mlx5e_free_rx_wqe_reuse(rq, wi);
+	mlx5e_free_rx_wqe(rq, wi);
 	mlx5_wq_cyc_pop(wq);
 }
 
@@ -1441,7 +1528,7 @@ void mlx5e_ipsec_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	wi       = get_frag(rq, ci);
 	cqe_bcnt = be32_to_cpu(cqe->byte_cnt);
 
-	skb = skb_from_cqe(rq, cqe, wi, cqe_bcnt);
+	skb = rq->wqe.skb_from_cqe(rq, cqe, wi, cqe_bcnt);
 	if (unlikely(!skb)) {
 		/* a DROP, save the page-reuse checks */
 		mlx5e_free_rx_wqe(rq, wi);
@@ -1456,7 +1543,7 @@ void mlx5e_ipsec_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
 	napi_gro_receive(rq->cq.napi, skb);
 
-	mlx5e_free_rx_wqe_reuse(rq, wi);
+	mlx5e_free_rx_wqe(rq, wi);
 wq_cyc_pop:
 	mlx5_wq_cyc_pop(wq);
 }