From: Madalin Bucur <madalin.bucur@nxp.com>
To: <netdev@vger.kernel.org>
Cc: <davem@davemloft.net>, <camelia.groza@nxp.com>,
<dan.carpenter@oracle.com>, <linux-kernel@vger.kernel.org>,
<linuxppc-dev@lists.ozlabs.org>,
Madalin Bucur <madalin.bucur@nxp.com>
Subject: [net-next v2 06/10] dpaa_eth: enable Rx checksum offload
Date: Thu, 9 Mar 2017 16:37:01 +0200 [thread overview]
Message-ID: <1489070225-7891-7-git-send-email-madalin.bucur@nxp.com> (raw)
In-Reply-To: <1489070225-7891-1-git-send-email-madalin.bucur@nxp.com>
Use the FMan HW parser L4CV flag to offload Rx checksumming.
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
---
drivers/net/ethernet/freescale/dpaa/dpaa_eth.c | 29 ++++++++++++++++++++++++--
1 file changed, 27 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c b/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
index e19181f..a7a595c 100644
--- a/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
+++ b/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
@@ -137,6 +137,13 @@ MODULE_PARM_DESC(tx_timeout, "The Tx timeout in ms");
/* L4 Type field: TCP */
#define FM_L4_PARSE_RESULT_TCP 0x20
+/* FD status field indicating whether the FM Parser has attempted to validate
+ * the L4 csum of the frame.
+ * Note that having this bit set doesn't necessarily imply that the checksum
+ * is valid. One would have to check the parse results to find that out.
+ */
+#define FM_FD_STAT_L4CV 0x00000004
+
#define DPAA_SGT_MAX_ENTRIES 16 /* maximum number of entries in SG Table */
#define DPAA_BUFF_RELEASE_MAX 8 /* maximum number of buffers released at once */
@@ -235,6 +242,7 @@ static int dpaa_netdev_init(struct net_device *net_dev,
* For conformity, we'll still declare GSO explicitly.
*/
net_dev->features |= NETIF_F_GSO;
+ net_dev->features |= NETIF_F_RXCSUM;
net_dev->priv_flags |= IFF_LIVE_ADDR_CHANGE;
/* we do not want shared skbs on TX */
@@ -1526,6 +1534,23 @@ static struct sk_buff *dpaa_cleanup_tx_fd(const struct dpaa_priv *priv,
return skb;
}
+static u8 rx_csum_offload(const struct dpaa_priv *priv, const struct qm_fd *fd)
+{
+ /* The parser has run and performed L4 checksum validation.
+ * We know there were no parser errors (and implicitly no
+ * L4 csum error), otherwise we wouldn't be here.
+ */
+ if ((priv->net_dev->features & NETIF_F_RXCSUM) &&
+ (be32_to_cpu(fd->status) & FM_FD_STAT_L4CV))
+ return CHECKSUM_UNNECESSARY;
+
+ /* We're here because either the parser didn't run or the L4 checksum
+ * was not verified. This may include the case of a UDP frame with
+ * checksum zero or an L4 proto other than TCP/UDP
+ */
+ return CHECKSUM_NONE;
+}
+
/* Build a linear skb around the received buffer.
* We are guaranteed there is enough room at the end of the data buffer to
* accommodate the shared info area of the skb.
@@ -1556,7 +1581,7 @@ static struct sk_buff *contig_fd_to_skb(const struct dpaa_priv *priv,
skb_reserve(skb, fd_off);
skb_put(skb, qm_fd_get_length(fd));
- skb->ip_summed = CHECKSUM_NONE;
+ skb->ip_summed = rx_csum_offload(priv, fd);
return skb;
@@ -1616,7 +1641,7 @@ static struct sk_buff *sg_fd_to_skb(const struct dpaa_priv *priv,
if (WARN_ON(unlikely(!skb)))
goto free_buffers;
- skb->ip_summed = CHECKSUM_NONE;
+ skb->ip_summed = rx_csum_offload(priv, fd);
/* Make sure forwarded skbs will have enough space
* on Tx, if extra headers are added.
--
2.1.0
next prev parent reply other threads:[~2017-03-09 14:37 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-09 14:36 [net-next v2 00/10] QorIQ DPAA 1 updates Madalin Bucur
2017-03-09 14:36 ` [net-next v2 01/10] fsl/fman: parse result data is big endian Madalin Bucur
2017-03-09 14:36 ` [net-next v2 02/10] fsl/fman: set HW parser as BMI next engine Madalin Bucur
2017-03-09 14:36 ` [net-next v2 03/10] fsl/fman: remove wrong free Madalin Bucur
2017-03-09 14:36 ` [net-next v2 04/10] fsl/fman: enlarge FIFO to allow for the 5th port Madalin Bucur
2017-03-09 14:37 ` [net-next v2 05/10] dpaa_eth: remove redundant initialization Madalin Bucur
2017-03-09 14:37 ` Madalin Bucur [this message]
2017-03-09 14:37 ` [net-next v2 07/10] dpaa_eth: do not ignore port api return value Madalin Bucur
2017-03-09 14:37 ` [net-next v2 08/10] dpaa_eth: add four prioritised Tx traffic classes Madalin Bucur
2017-03-09 14:37 ` [net-next v2 09/10] dpaa_eth: enable multiple " Madalin Bucur
2017-03-09 14:37 ` [net-next v2 10/10] dpaa_eth: enable context-A stashing Madalin Bucur
2017-03-13 6:01 ` [net-next v2 00/10] QorIQ DPAA 1 updates 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=1489070225-7891-7-git-send-email-madalin.bucur@nxp.com \
--to=madalin.bucur@nxp.com \
--cc=camelia.groza@nxp.com \
--cc=dan.carpenter@oracle.com \
--cc=davem@davemloft.net \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=netdev@vger.kernel.org \
/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).