From mboxrd@z Thu Jan 1 00:00:00 1970 From: Rob Herring Subject: [PATCH v2 RESEND 6/6] net: calxedaxgmac: ip align receive buffers Date: Thu, 1 Nov 2012 05:41:04 -0500 Message-ID: <1351766464-27354-7-git-send-email-robherring2@gmail.com> References: <1351766464-27354-1-git-send-email-robherring2@gmail.com> Cc: eric.dumazet@gmail.com, Rob Herring To: netdev@vger.kernel.org, davem@davemloft.net Return-path: Received: from mail-bk0-f46.google.com ([209.85.214.46]:64228 "EHLO mail-bk0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756585Ab2KAKl0 (ORCPT ); Thu, 1 Nov 2012 06:41:26 -0400 Received: by mail-bk0-f46.google.com with SMTP id jk13so1000955bkc.19 for ; Thu, 01 Nov 2012 03:41:26 -0700 (PDT) In-Reply-To: <1351766464-27354-1-git-send-email-robherring2@gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Rob Herring On gcc 4.7, we will get alignment traps in the ip stack if we don't align the ip headers on receive. The h/w can support this, so use ip aligned allocations. Cut down the unnecessary padding on the allocation. The buffer can start on any byte alignment, but the size including the begining offset must be 8 byte aligned. So the h/w buffer size must include the NET_IP_ALIGN offset. Thanks to Eric Dumazet for the initial patch highlighting the padding issues. Signed-off-by: Rob Herring --- drivers/net/ethernet/calxeda/xgmac.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) diff --git a/drivers/net/ethernet/calxeda/xgmac.c b/drivers/net/ethernet/calxeda/xgmac.c index 38eb1b2..855bf2b 100644 --- a/drivers/net/ethernet/calxeda/xgmac.c +++ b/drivers/net/ethernet/calxeda/xgmac.c @@ -665,6 +665,7 @@ static void xgmac_rx_refill(struct xgmac_priv *priv) { struct xgmac_dma_desc *p; dma_addr_t paddr; + int bufsz = priv->dev->mtu + ETH_HLEN + ETH_FCS_LEN; while (dma_ring_space(priv->rx_head, priv->rx_tail, DMA_RX_RING_SZ) > 1) { int entry = priv->rx_head; @@ -673,13 +674,13 @@ static void xgmac_rx_refill(struct xgmac_priv *priv) p = priv->dma_rx + entry; if (priv->rx_skbuff[entry] == NULL) { - skb = netdev_alloc_skb(priv->dev, priv->dma_buf_sz); + skb = netdev_alloc_skb_ip_align(priv->dev, bufsz); if (unlikely(skb == NULL)) break; priv->rx_skbuff[entry] = skb; paddr = dma_map_single(priv->device, skb->data, - priv->dma_buf_sz, DMA_FROM_DEVICE); + bufsz, DMA_FROM_DEVICE); desc_set_buf_addr(p, paddr, priv->dma_buf_sz); } @@ -703,10 +704,10 @@ static int xgmac_dma_desc_rings_init(struct net_device *dev) unsigned int bfsize; /* Set the Buffer size according to the MTU; - * indeed, in case of jumbo we need to bump-up the buffer sizes. + * The total buffer size including any IP offset must be a multiple + * of 8 bytes. */ - bfsize = ALIGN(dev->mtu + ETH_HLEN + ETH_FCS_LEN + NET_IP_ALIGN + 64, - 64); + bfsize = ALIGN(dev->mtu + ETH_HLEN + ETH_FCS_LEN + NET_IP_ALIGN, 8); netdev_dbg(priv->dev, "mtu [%d] bfsize [%d]\n", dev->mtu, bfsize); -- 1.7.10.4