diff mbox

[v2] IB/ipoib: Report SG feature regardless of HW UD CSUM capability

Message ID 1469003406-1540-1-git-send-email-yuval.shaia@oracle.com (mailing list archive)
State Accepted
Headers show

Commit Message

Yuval Shaia July 20, 2016, 8:30 a.m. UTC
Decouple SG support from HW ability to do UD checksum.
This coupling is for historical reasons and removed with 'commit
ec5f06156423 ("net: Kill link between CSUM and SG features.")'

During driver load it is assumed that device does not supports SG. The
final decision is taken after creating UD QP based on device capability.

Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com>
---
v0 -> v1:
	* Fix patch title
	* Move settings to ipoib_setup instead of ipoib_set_dev_feature
v1 -> v2:
	* Move settings to ipoib_transport_dev_init
---
 drivers/infiniband/ulp/ipoib/ipoib_main.c  |    3 +--
 drivers/infiniband/ulp/ipoib/ipoib_verbs.c |   10 +++++-----
 2 files changed, 6 insertions(+), 7 deletions(-)

Comments

Jason Gunthorpe July 20, 2016, 5:44 p.m. UTC | #1
On Wed, Jul 20, 2016 at 01:30:06AM -0700, Yuval Shaia wrote:
> Decouple SG support from HW ability to do UD checksum.
> This coupling is for historical reasons and removed with 'commit
> ec5f06156423 ("net: Kill link between CSUM and SG features.")'
> 
> During driver load it is assumed that device does not supports SG. The
> final decision is taken after creating UD QP based on device capability.
> 
> Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com>

Yes, this looks nice now.

Reviewed-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Doug Ledford Aug. 2, 2016, 7:11 p.m. UTC | #2
On Wed, 2016-07-20 at 01:30 -0700, Yuval Shaia wrote:
> Decouple SG support from HW ability to do UD checksum.
> This coupling is for historical reasons and removed with 'commit
> ec5f06156423 ("net: Kill link between CSUM and SG features.")'
> 
> During driver load it is assumed that device does not supports SG.
> The
> final decision is taken after creating UD QP based on device
> capability.
> 
> Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com>
> ---
> v0 -> v1:
> 	* Fix patch title
> 	* Move settings to ipoib_setup instead of ipoib_set_dev_feature
> v1 -> v2:
> 	* Move settings to ipoib_transport_dev_init
> ---
>  drivers/infiniband/ulp/ipoib/ipoib_main.c  |    3 +--
>  drivers/infiniband/ulp/ipoib/ipoib_verbs.c |   10 +++++-----
>  2 files changed, 6 insertions(+), 7 deletions(-)


Thanks, applied.
diff mbox

Patch

diff --git a/drivers/infiniband/ulp/ipoib/ipoib_main.c b/drivers/infiniband/ulp/ipoib/ipoib_main.c
index 5f58c41..74bcaa0 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_main.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_main.c
@@ -1967,8 +1967,7 @@  int ipoib_set_dev_features(struct ipoib_dev_priv *priv, struct ib_device *hca)
 	priv->hca_caps = hca->attrs.device_cap_flags;
 
 	if (priv->hca_caps & IB_DEVICE_UD_IP_CSUM) {
-		priv->dev->hw_features = NETIF_F_SG |
-			NETIF_F_IP_CSUM | NETIF_F_RXCSUM;
+		priv->dev->hw_features = NETIF_F_IP_CSUM | NETIF_F_RXCSUM;
 
 		if (priv->hca_caps & IB_DEVICE_UD_TSO)
 			priv->dev->hw_features |= NETIF_F_TSO;
diff --git a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
index 1e7cbba..c55ecb2 100644
--- a/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
+++ b/drivers/infiniband/ulp/ipoib/ipoib_verbs.c
@@ -135,7 +135,8 @@  int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca)
 		.cap = {
 			.max_send_wr  = ipoib_sendq_size,
 			.max_recv_wr  = ipoib_recvq_size,
-			.max_send_sge = 1,
+			.max_send_sge = min_t(u32, priv->ca->attrs.max_sge,
+					      MAX_SKB_FRAGS + 1),
 			.max_recv_sge = IPOIB_UD_RX_SG
 		},
 		.sq_sig_type = IB_SIGNAL_ALL_WR,
@@ -205,10 +206,6 @@  int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca)
 	if (priv->hca_caps & IB_DEVICE_MANAGED_FLOW_STEERING)
 		init_attr.create_flags |= IB_QP_CREATE_NETIF_QP;
 
-	if (dev->features & NETIF_F_SG)
-		init_attr.cap.max_send_sge =
-			min_t(u32, priv->ca->attrs.max_sge, MAX_SKB_FRAGS + 1);
-
 	priv->qp = ib_create_qp(priv->pd, &init_attr);
 	if (IS_ERR(priv->qp)) {
 		printk(KERN_WARNING "%s: failed to create QP\n", ca->name);
@@ -234,6 +231,9 @@  int ipoib_transport_dev_init(struct net_device *dev, struct ib_device *ca)
 	priv->rx_wr.next = NULL;
 	priv->rx_wr.sg_list = priv->rx_sge;
 
+	if (init_attr.cap.max_send_sge > 1)
+		dev->features |= NETIF_F_SG;
+
 	priv->max_send_sge = init_attr.cap.max_send_sge;
 
 	return 0;