From: Andrew Lunn <andrew@lunn.ch>
To: David Miller <davem@davemloft.net>,
Vivien Didelot <vivien.didelot@savoirfairelinux.com>,
Florian Fainelli <f.fainelli@gmail.com>
Cc: netdev <netdev@vger.kernel.org>, Shawn Guo <shawnguo@kernel.org>,
Andrew Lunn <andrew@lunn.ch>
Subject: [PATCHv2 net-next] net: igb: Only dma sync frame length
Date: Sat, 4 Jun 2016 21:16:53 +0200 [thread overview]
Message-ID: <1465067829-1875-4-git-send-email-andrew@lunn.ch> (raw)
In-Reply-To: <1465067829-1875-1-git-send-email-andrew@lunn.ch>
On some platforms, syncing a buffer for DMA is expensive. Rather than
sync the whole 2K receive buffer, only synchronise the length of the
frame, which will typically be the MTU, or a much smaller TCP ACK.
For an IMX6Q, this gives around 6% increased TCP receive performance,
which is cache operations bound and reduces CPU load for TCP transmit.
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
---
v2:
Christmas tree the local variables
Pass size into igb_add_rx_frag() rather than repeating the endiness swap.
---
drivers/net/ethernet/intel/igb/igb_main.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/intel/igb/igb_main.c b/drivers/net/ethernet/intel/igb/igb_main.c
index 0a289dda604a..8fa9e6e8c3b0 100644
--- a/drivers/net/ethernet/intel/igb/igb_main.c
+++ b/drivers/net/ethernet/intel/igb/igb_main.c
@@ -6856,12 +6856,12 @@ static bool igb_can_reuse_rx_page(struct igb_rx_buffer *rx_buffer,
**/
static bool igb_add_rx_frag(struct igb_ring *rx_ring,
struct igb_rx_buffer *rx_buffer,
+ unsigned int size,
union e1000_adv_rx_desc *rx_desc,
struct sk_buff *skb)
{
struct page *page = rx_buffer->page;
unsigned char *va = page_address(page) + rx_buffer->page_offset;
- unsigned int size = le16_to_cpu(rx_desc->wb.upper.length);
#if (PAGE_SIZE < 8192)
unsigned int truesize = IGB_RX_BUFSZ;
#else
@@ -6913,6 +6913,7 @@ static struct sk_buff *igb_fetch_rx_buffer(struct igb_ring *rx_ring,
union e1000_adv_rx_desc *rx_desc,
struct sk_buff *skb)
{
+ unsigned int size = le16_to_cpu(rx_desc->wb.upper.length);
struct igb_rx_buffer *rx_buffer;
struct page *page;
@@ -6948,11 +6949,11 @@ static struct sk_buff *igb_fetch_rx_buffer(struct igb_ring *rx_ring,
dma_sync_single_range_for_cpu(rx_ring->dev,
rx_buffer->dma,
rx_buffer->page_offset,
- IGB_RX_BUFSZ,
+ size,
DMA_FROM_DEVICE);
/* pull page into skb */
- if (igb_add_rx_frag(rx_ring, rx_buffer, rx_desc, skb)) {
+ if (igb_add_rx_frag(rx_ring, rx_buffer, size, rx_desc, skb)) {
/* hand second half of page back to the ring */
igb_reuse_rx_page(rx_ring, rx_buffer);
} else {
--
2.8.1
next prev parent reply other threads:[~2016-06-04 19:24 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-06-04 19:16 [PATCHv2 net-next 00/17] New DSA bind, switches as devices Andrew Lunn
2016-06-04 19:16 ` [PATCH] dma: Add a benchmark for using SRAM memory as receiver buffers Andrew Lunn
2016-06-04 19:16 ` [PATCHv2 net-next 01/17] net: dsa: slave: chip data is optional, don't dereference NULL Andrew Lunn
2016-06-04 19:16 ` Andrew Lunn [this message]
2016-06-04 19:16 ` [PATCHv2 net-next 02/17] net: dsa: mv88e6xxx: fix circular lock in PPU work Andrew Lunn
2016-06-04 19:16 ` [PATCHv2 net-next 03/17] net: dsa: slave: Remove MDIO address from switch MDIO bus name Andrew Lunn
2016-06-04 19:16 ` [PATCHv2 net-next 04/17] net: dsa: tag_{e}dsa.c: Remove dependency on platform data Andrew Lunn
2016-06-04 19:16 ` [PATCHv2 net-next 05/17] net: dsa: Add a ports structure and use it in the switch structure Andrew Lunn
2016-06-04 19:16 ` [PATCHv2 net-next 06/17] net: dsa: Move port device node into port structure Andrew Lunn
2016-06-04 19:16 ` [PATCHv2 net-next 07/17] net: dsa: Remove dynamic allocate of routing table Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 08/17] net: dsa: Copy the routing table into the switch structure Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 09/17] net: dsa: Split up creating/destroying of DSA and CPU ports Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 10/17] net: dsa: mv88e6xxx: Only support EDSA tagging Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 11/17] net: dsa: Refactor selection of tag ops into a function Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 12/17] net: dsa: Make mdio bus optional Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 13/17] net: dsa: mv88e6xxx: Rename _phy_ to _mdio_ Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 14/17] net: dsa: mv88e6xxx: Refactor MDIO so driver registers mdio bus Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 15/17] net: dsa: Add new binding implementation Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 16/17] arm: dt: vf610-zii-devel-b: Make use of new DSA binding Andrew Lunn
2016-06-04 19:17 ` [PATCHv2 net-next 17/17] net: dsa: Document new binding Andrew Lunn
2016-06-04 19:28 ` [PATCHv2 net-next 00/17] New DSA bind, switches as devices Andrew Lunn
2016-06-04 21:30 ` David Miller
-- strict thread matches above, loose matches on Subject: below --
2016-06-03 21:03 [PATCHv2 net-next] net: igb: Only dma sync frame length Andrew Lunn
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=1465067829-1875-4-git-send-email-andrew@lunn.ch \
--to=andrew@lunn.ch \
--cc=davem@davemloft.net \
--cc=f.fainelli@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=shawnguo@kernel.org \
--cc=vivien.didelot@savoirfairelinux.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).