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>

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

@@ -101,11 +101,15 @@ struct page_pool;
 	(MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE_MPW + \
 	 (MLX5_MPWRQ_LOG_WQE_SZ - MLX5E_ORDER2_MAX_PACKET_MTU))
 
+#define MLX5E_MIN_SKB_FRAG_SZ		(MLX5_SKB_FRAG_SZ(MLX5_RX_HEADROOM))
+#define MLX5E_LOG_MAX_RX_WQE_BULK	\
+	(ilog2(PAGE_SIZE / roundup_pow_of_two(MLX5E_MIN_SKB_FRAG_SZ)))
+
 #define MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE                0x6
 #define MLX5E_PARAMS_DEFAULT_LOG_SQ_SIZE                0xa
 #define MLX5E_PARAMS_MAXIMUM_LOG_SQ_SIZE                0xd
 
-#define MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE                0x1
+#define MLX5E_PARAMS_MINIMUM_LOG_RQ_SIZE (1 + MLX5E_LOG_MAX_RX_WQE_BULK)
 #define MLX5E_PARAMS_DEFAULT_LOG_RQ_SIZE                0xa
 #define MLX5E_PARAMS_MAXIMUM_LOG_RQ_SIZE min_t(u8, 0xd,	\
 					       MLX5E_LOG_MAX_RQ_NUM_PACKETS_MPW)
@@ -462,8 +466,9 @@ struct mlx5e_dma_info {
 };
 
 struct mlx5e_wqe_frag_info {
-	struct mlx5e_dma_info di;
+	struct mlx5e_dma_info *di;
 	u32 offset;
+	bool last_in_page;
 };
 
 struct mlx5e_umr_dma_info {
@@ -476,6 +481,8 @@ struct mlx5e_mpw_info {
 	DECLARE_BITMAP(xdp_xmit_bitmap, MLX5_MPWRQ_PAGES_PER_WQE);
 };
 
+#define MLX5E_MAX_RX_FRAGS 4
+
 /* a single cache unit is capable to serve one napi call (for non-striding rq)
  * or a MPWQE (for striding rq).
  */
@@ -493,6 +500,9 @@ typedef void (*mlx5e_fp_handle_rx_cqe)(struct mlx5e_rq*, struct mlx5_cqe64*);
 typedef struct sk_buff *
 (*mlx5e_fp_skb_from_cqe_mpwrq)(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
 			       u16 cqe_bcnt, u32 head_offset, u32 page_idx);
+typedef struct sk_buff *
+(*mlx5e_fp_skb_from_cqe)(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			 struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt);
 typedef bool (*mlx5e_fp_post_rx_wqes)(struct mlx5e_rq *rq);
 typedef void (*mlx5e_fp_dealloc_wqe)(struct mlx5e_rq*, u16);
 
@@ -500,16 +510,27 @@ enum mlx5e_rq_flag {
 	MLX5E_RQ_FLAG_XDP_XMIT = BIT(0),
 };
 
+struct mlx5e_rq_frag_info {
+	int frag_size;
+	int frag_stride;
+};
+
+struct mlx5e_rq_frags_info {
+	struct mlx5e_rq_frag_info arr[MLX5E_MAX_RX_FRAGS];
+	u8 num_frags;
+	u8 log_num_frags;
+	u8 wqe_bulk;
+};
+
 struct mlx5e_rq {
 	/* data path */
 	union {
 		struct {
-			struct mlx5_wq_cyc     wq;
-			struct mlx5e_wqe_frag_info *frag_info;
-			u32 frag_sz;	/* max possible skb frag_sz */
-			union {
-				bool page_reuse;
-			};
+			struct mlx5_wq_cyc          wq;
+			struct mlx5e_wqe_frag_info *frags;
+			struct mlx5e_dma_info      *di;
+			struct mlx5e_rq_frags_info  info;
+			mlx5e_fp_skb_from_cqe       skb_from_cqe;
 		} wqe;
 		struct {
 			struct mlx5_wq_ll      wq;
@@ -523,7 +544,6 @@ struct mlx5e_rq {
 	};
 	struct {
 		u16            headroom;
-		u8             page_order;
 		u8             map_dir;   /* dma map direction */
 	} buff;
 
@@ -879,6 +899,12 @@ mlx5e_skb_from_cqe_mpwrq_linear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
 struct sk_buff *
 mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
 				   u16 cqe_bcnt, u32 head_offset, u32 page_idx);
+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 sk_buff *
+mlx5e_skb_from_cqe_nonlinear(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe,
+			     struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt);
 
 void mlx5e_update_stats(struct mlx5e_priv *priv);