net: ethernet: mediatek: enhance RX path by reducing the frequency of…

… the memory barrier used

The patch makes move wmb() to outside the loop that could help
RX path handling more faster although that RX descriptors aren't
freed for DMA to use as soon as possible, but based on my experiment
and the result shows it still can reach about 943Mbpis without
performance drop that is tested based on the setup with one port
using Giga PHY and 256 RX descriptors for DMA to move.

Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/mediatek/mtk_eth_soc.c

@@ -895,13 +895,14 @@ static int mtk_poll_rx(struct napi_struct *napi, int budget,
 		rxd->rxd2 = RX_DMA_PLEN0(ring->buf_size);
 
 		ring->calc_idx = idx;
-		/* make sure that all changes to the dma ring are flushed before
-		 * we continue
-		 */
-		wmb();
-		mtk_w32(eth, ring->calc_idx, MTK_PRX_CRX_IDX0);
+
 		done++;
 	}
+	/* make sure that all changes to the dma ring are flushed before
+	 * we continue
+	 */
+	wmb();
+	mtk_w32(eth, ring->calc_idx, MTK_PRX_CRX_IDX0);
 
 	if (done < budget)
 		mtk_w32(eth, MTK_RX_DONE_INT, MTK_PDMA_INT_STATUS);