Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1031463AbbEEJ0f (ORCPT ); Tue, 5 May 2015 05:26:35 -0400 Received: from mail-wi0-f172.google.com ([209.85.212.172]:35696 "EHLO mail-wi0-f172.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1031409AbbEEJ0R (ORCPT ); Tue, 5 May 2015 05:26:17 -0400 From: Michal Simek To: netdev@vger.kernel.org Cc: Peter Crosthwaite , =?UTF-8?q?S=C3=B6ren=20Brinkmann?= , monstr@monstr.eu, John Linn , Anirudha Sarangi , linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org Subject: [PATCH 02/12] net: axienet: Handle 0 packet receive gracefully Date: Tue, 5 May 2015 11:25:55 +0200 Message-Id: <55b3c89b3549c61d62b8440636516fd572870842.1430817941.git.michal.simek@xilinx.com> X-Mailer: git-send-email 2.3.5 In-Reply-To: <7fb84f65a61bbe0fdb4b61a871cf4d4f7910955d.1430817941.git.michal.simek@xilinx.com> References: <7fb84f65a61bbe0fdb4b61a871cf4d4f7910955d.1430817941.git.michal.simek@xilinx.com> In-Reply-To: <7fb84f65a61bbe0fdb4b61a871cf4d4f7910955d.1430817941.git.michal.simek@xilinx.com> References: <7fb84f65a61bbe0fdb4b61a871cf4d4f7910955d.1430817941.git.michal.simek@xilinx.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2834 Lines: 73 From: Peter Crosthwaite The AXI-DMA rx-delay interrupt can sometimes be triggered when there are 0 outstanding packets received. This is due to the fact that the receive function will greedily consume as many packets as possible on interrupt. So if two packets (with a very particular timing) arrive in succession they will each cause the rx-delay interrupt, but the first interrupt will consume both packets. This means the second interrupt is a 0 packet receive. This is mostly OK, except that the tail pointer register is updated unconditionally on receive. Currently the tail pointer is always set to the current bd-ring descriptor under the assumption that the hardware has moved onto the next descriptor. What this means for length 0 recv is the current descriptor that the hardware is potentially yet to use will be marked as the tail. This causes the hardware to think its run out of descriptors deadlocking the whole rx path. Fixed by updating the tail pointer to the most recent successfully consumed descriptor. Reported-by: Wendy Liang Signed-off-by: Peter Crosthwaite Tested-by: Jason Wu Acked-by: Michal Simek Signed-off-by: Michal Simek --- drivers/net/ethernet/xilinx/xilinx_axienet_main.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c index 0ab607732bb4..a4840952d372 100644 --- a/drivers/net/ethernet/xilinx/xilinx_axienet_main.c +++ b/drivers/net/ethernet/xilinx/xilinx_axienet_main.c @@ -726,15 +726,15 @@ static void axienet_recv(struct net_device *ndev) u32 csumstatus; u32 size = 0; u32 packets = 0; - dma_addr_t tail_p; + dma_addr_t tail_p = 0; struct axienet_local *lp = netdev_priv(ndev); struct sk_buff *skb, *new_skb; struct axidma_bd *cur_p; - tail_p = lp->rx_bd_p + sizeof(*lp->rx_bd_v) * lp->rx_bd_ci; cur_p = &lp->rx_bd_v[lp->rx_bd_ci]; while ((cur_p->status & XAXIDMA_BD_STS_COMPLETE_MASK)) { + tail_p = lp->rx_bd_p + sizeof(*lp->rx_bd_v) * lp->rx_bd_ci; skb = (struct sk_buff *) (cur_p->sw_id_offset); length = cur_p->app4 & 0x0000FFFF; @@ -786,7 +786,8 @@ static void axienet_recv(struct net_device *ndev) ndev->stats.rx_packets += packets; ndev->stats.rx_bytes += size; - axienet_dma_out32(lp, XAXIDMA_RX_TDESC_OFFSET, tail_p); + if (tail_p) + axienet_dma_out32(lp, XAXIDMA_RX_TDESC_OFFSET, tail_p); } /** -- 2.3.5 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/