From 7f7183af448ac55da9cc72bdbef6d140909fee83 Mon Sep 17 00:00:00 2001 From: Matteo Croce Date: Thu, 24 Oct 2019 19:24:56 +0200 Subject: [PATCH 1/3] mvpp2: refactor frame drop routine Move some code down to remove a backward goto. Signed-off-by: Matteo Croce Signed-off-by: David S. Miller --- drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c index 5fea65256b9d4..b61e0ad0978af 100644 --- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c +++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c @@ -2956,14 +2956,8 @@ static int mvpp2_rx(struct mvpp2_port *port, struct napi_struct *napi, * by the hardware, and the information about the buffer is * comprised by the RX descriptor. */ - if (rx_status & MVPP2_RXD_ERR_SUMMARY) { -err_drop_frame: - dev->stats.rx_errors++; - mvpp2_rx_error(port, rx_desc); - /* Return the buffer to the pool */ - mvpp2_bm_pool_put(port, pool, dma_addr, phys_addr); - continue; - } + if (rx_status & MVPP2_RXD_ERR_SUMMARY) + goto err_drop_frame; if (bm_pool->frag_size > PAGE_SIZE) frag_size = 0; @@ -2994,6 +2988,13 @@ static int mvpp2_rx(struct mvpp2_port *port, struct napi_struct *napi, mvpp2_rx_csum(port, rx_status, skb); napi_gro_receive(napi, skb); + continue; + +err_drop_frame: + dev->stats.rx_errors++; + mvpp2_rx_error(port, rx_desc); + /* Return the buffer to the pool */ + mvpp2_bm_pool_put(port, pool, dma_addr, phys_addr); } if (rcvd_pkts) { From e1921168bbd4810de4197446e52f652cd0dd9541 Mon Sep 17 00:00:00 2001 From: Matteo Croce Date: Thu, 24 Oct 2019 19:24:57 +0200 Subject: [PATCH 2/3] mvpp2: sync only the received frame In the RX path we always sync against the maximum frame size for that pool. Do the DMA sync and the unmap separately, so we can only sync by the size of the received frame. Signed-off-by: Matteo Croce Signed-off-by: David S. Miller --- drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c index b61e0ad0978af..c66943219e1b1 100644 --- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c +++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c @@ -2959,6 +2959,10 @@ static int mvpp2_rx(struct mvpp2_port *port, struct napi_struct *napi, if (rx_status & MVPP2_RXD_ERR_SUMMARY) goto err_drop_frame; + dma_sync_single_for_cpu(dev->dev.parent, dma_addr, + rx_bytes + MVPP2_MH_SIZE, + DMA_FROM_DEVICE); + if (bm_pool->frag_size > PAGE_SIZE) frag_size = 0; else @@ -2976,8 +2980,9 @@ static int mvpp2_rx(struct mvpp2_port *port, struct napi_struct *napi, goto err_drop_frame; } - dma_unmap_single(dev->dev.parent, dma_addr, - bm_pool->buf_size, DMA_FROM_DEVICE); + dma_unmap_single_attrs(dev->dev.parent, dma_addr, + bm_pool->buf_size, DMA_FROM_DEVICE, + DMA_ATTR_SKIP_CPU_SYNC); rcvd_pkts++; rcvd_bytes += rx_bytes; From a0c78337dd3a4900bc8628a511131b5a0b1db42a Mon Sep 17 00:00:00 2001 From: Matteo Croce Date: Thu, 24 Oct 2019 19:24:58 +0200 Subject: [PATCH 3/3] mvpp2: prefetch frame header When receiving traffic, eth_type_trans() is high up on the perf top list, because it's the first function which access the packet data. Move the DMA unmap a bit higher, and put a prefetch just after it, so we have more time to load the data into the cache. The packet rate increase is about 14% with a tc drop test: 1620 => 1853 kpps Signed-off-by: Matteo Croce Signed-off-by: David S. Miller --- drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c index c66943219e1b1..17e24c1e1c2bd 100644 --- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c +++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c @@ -2962,6 +2962,7 @@ static int mvpp2_rx(struct mvpp2_port *port, struct napi_struct *napi, dma_sync_single_for_cpu(dev->dev.parent, dma_addr, rx_bytes + MVPP2_MH_SIZE, DMA_FROM_DEVICE); + prefetch(data); if (bm_pool->frag_size > PAGE_SIZE) frag_size = 0;