All of lore.kernel.org
 help / color / mirror / Atom feed
From: Felix Fietkau <nbd@openwrt.org>
To: netdev@vger.kernel.org
Cc: zajec5@gmail.com, hauke@hauke-m.de
Subject: [PATCH v7 3/9] bgmac: set received skb headroom to NET_SKB_PAD
Date: Tue, 14 Apr 2015 12:07:56 +0200	[thread overview]
Message-ID: <1429006082-23489-3-git-send-email-nbd@openwrt.org> (raw)
In-Reply-To: <1429006082-23489-1-git-send-email-nbd@openwrt.org>

A packet buffer offset of 30 bytes is inefficient, because the first 2
bytes end up in a different cacheline.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
---
 drivers/net/ethernet/broadcom/bgmac.c | 14 ++++++++------
 drivers/net/ethernet/broadcom/bgmac.h |  4 +++-
 2 files changed, 11 insertions(+), 7 deletions(-)

diff --git a/drivers/net/ethernet/broadcom/bgmac.c b/drivers/net/ethernet/broadcom/bgmac.c
index 9c3b2ff..424a68c 100644
--- a/drivers/net/ethernet/broadcom/bgmac.c
+++ b/drivers/net/ethernet/broadcom/bgmac.c
@@ -346,13 +346,13 @@ static int bgmac_dma_rx_skb_for_slot(struct bgmac *bgmac,
 		return -ENOMEM;
 
 	/* Poison - if everything goes fine, hardware will overwrite it */
-	rx = buf;
+	rx = buf + BGMAC_RX_BUF_OFFSET;
 	rx->len = cpu_to_le16(0xdead);
 	rx->flags = cpu_to_le16(0xbeef);
 
 	/* Map skb for the DMA */
-	dma_addr = dma_map_single(dma_dev, buf, BGMAC_RX_BUF_SIZE,
-				  DMA_FROM_DEVICE);
+	dma_addr = dma_map_single(dma_dev, buf + BGMAC_RX_BUF_OFFSET,
+				  BGMAC_RX_BUF_SIZE, DMA_FROM_DEVICE);
 	if (dma_mapping_error(dma_dev, dma_addr)) {
 		bgmac_err(bgmac, "DMA mapping error\n");
 		put_page(virt_to_head_page(buf));
@@ -403,7 +403,7 @@ static int bgmac_dma_rx_read(struct bgmac *bgmac, struct bgmac_dma_ring *ring,
 	while (ring->start != ring->end) {
 		struct device *dma_dev = bgmac->core->dma_dev;
 		struct bgmac_slot_info *slot = &ring->slots[ring->start];
-		struct bgmac_rx_header *rx = slot->buf;
+		struct bgmac_rx_header *rx = slot->buf + BGMAC_RX_BUF_OFFSET;
 		struct sk_buff *skb;
 		void *buf = slot->buf;
 		u16 len, flags;
@@ -454,8 +454,10 @@ static int bgmac_dma_rx_read(struct bgmac *bgmac, struct bgmac_dma_ring *ring,
 					 BGMAC_RX_BUF_SIZE, DMA_FROM_DEVICE);
 
 			skb = build_skb(buf, BGMAC_RX_ALLOC_SIZE);
-			skb_put(skb, BGMAC_RX_FRAME_OFFSET + len);
-			skb_pull(skb, BGMAC_RX_FRAME_OFFSET);
+			skb_put(skb, BGMAC_RX_FRAME_OFFSET +
+				BGMAC_RX_BUF_OFFSET + len);
+			skb_pull(skb, BGMAC_RX_FRAME_OFFSET +
+				 BGMAC_RX_BUF_OFFSET);
 
 			skb_checksum_none_assert(skb);
 			skb->protocol = eth_type_trans(skb, bgmac->net_dev);
diff --git a/drivers/net/ethernet/broadcom/bgmac.h b/drivers/net/ethernet/broadcom/bgmac.h
index abd50d1..3db0d52 100644
--- a/drivers/net/ethernet/broadcom/bgmac.h
+++ b/drivers/net/ethernet/broadcom/bgmac.h
@@ -360,9 +360,11 @@
 
 #define BGMAC_RX_HEADER_LEN			28		/* Last 24 bytes are unused. Well... */
 #define BGMAC_RX_FRAME_OFFSET			30		/* There are 2 unused bytes between header and real data */
+#define BGMAC_RX_BUF_OFFSET			(NET_SKB_PAD + NET_IP_ALIGN - \
+						 BGMAC_RX_FRAME_OFFSET)
 #define BGMAC_RX_MAX_FRAME_SIZE			1536		/* Copied from b44/tg3 */
 #define BGMAC_RX_BUF_SIZE			(BGMAC_RX_FRAME_OFFSET + BGMAC_RX_MAX_FRAME_SIZE)
-#define BGMAC_RX_ALLOC_SIZE			(SKB_DATA_ALIGN(BGMAC_RX_BUF_SIZE) + \
+#define BGMAC_RX_ALLOC_SIZE			(SKB_DATA_ALIGN(BGMAC_RX_BUF_SIZE + BGMAC_RX_BUF_OFFSET) + \
 						 SKB_DATA_ALIGN(sizeof(struct skb_shared_info)))
 
 #define BGMAC_BFL_ENETROBO			0x0010		/* has ephy roboswitch spi */
-- 
2.2.2

  parent reply	other threads:[~2015-04-14 10:08 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-04-14 10:07 [PATCH v7 1/9] bgmac: simplify tx ring index handling Felix Fietkau
2015-04-14 10:07 ` [PATCH v7 2/9] bgmac: leave interrupts disabled as long as there is work to do Felix Fietkau
2015-04-14 10:07 ` Felix Fietkau [this message]
2015-04-14 10:07 ` [PATCH v7 4/9] bgmac: simplify/optimize rx DMA error handling Felix Fietkau
2015-04-14 10:36   ` Rafał Miłecki
2015-04-14 10:07 ` [PATCH v7 5/9] bgmac: add check for oversized packets Felix Fietkau
2015-04-14 10:07 ` [PATCH v7 6/9] bgmac: increase rx ring size from 511 to 512 Felix Fietkau
2015-04-14 10:08 ` [PATCH v7 7/9] bgmac: simplify dma init/cleanup Felix Fietkau
2015-04-14 10:38   ` Rafał Miłecki
2015-04-14 10:08 ` [PATCH v7 8/9] bgmac: fix DMA rx corruption Felix Fietkau
2015-04-14 10:08 ` [PATCH v7 9/9] bgmac: drop ring->num_slots Felix Fietkau
2015-04-14 10:29 ` [PATCH v7 1/9] bgmac: simplify tx ring index handling Rafał Miłecki
2015-04-14 18:55 ` David Miller

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1429006082-23489-3-git-send-email-nbd@openwrt.org \
    --to=nbd@openwrt.org \
    --cc=hauke@hauke-m.de \
    --cc=netdev@vger.kernel.org \
    --cc=zajec5@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.