[PATCH net-next] net: marvell: return csum computation result from mvneta_rx_csum/mvpp2_rx_csum

Subsystems: marvell mvneta ethernet driver, marvell mvpp2 ethernet driver, networking drivers, the rest

STALE1875d REVIEWED: 1 (0M)

1 review trailer.

4 messages, 3 authors, 2021-06-22 · open the first message on its own page

[PATCH net-next] net: marvell: return csum computation result from mvneta_rx_csum/mvpp2_rx_csum

From: Lorenzo Bianconi <lorenzo@kernel.org>
Date: 2021-06-17 15:34:18 

This is a preliminary patch to add hw csum hint support to
mvneta/mvpp2 xdp implementation

Tested-by: Matteo Croce <redacted>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
 drivers/net/ethernet/marvell/mvneta.c         | 19 +++++++------------
 .../net/ethernet/marvell/mvpp2/mvpp2_main.c   | 14 +++++---------
 2 files changed, 12 insertions(+), 21 deletions(-)

diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index c15ce06427d0..88a755034c39 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -1805,18 +1805,14 @@ static void mvneta_rx_error(struct mvneta_port *pp,
 }
 
 /* Handle RX checksum offload based on the descriptor's status */
-static void mvneta_rx_csum(struct mvneta_port *pp, u32 status,
-			   struct sk_buff *skb)
+static int mvneta_rx_csum(struct mvneta_port *pp, u32 status)
 {
 	if ((pp->dev->features & NETIF_F_RXCSUM) &&
 	    (status & MVNETA_RXD_L3_IP4) &&
-	    (status & MVNETA_RXD_L4_CSUM_OK)) {
-		skb->csum = 0;
-		skb->ip_summed = CHECKSUM_UNNECESSARY;
-		return;
-	}
+	    (status & MVNETA_RXD_L4_CSUM_OK))
+		return CHECKSUM_UNNECESSARY;
 
-	skb->ip_summed = CHECKSUM_NONE;
+	return CHECKSUM_NONE;
 }
 
 /* Return tx queue pointer (find last set bit) according to <cause> returned
@@ -2335,7 +2331,7 @@ mvneta_swbm_build_skb(struct mvneta_port *pp, struct page_pool *pool,
 
 	skb_reserve(skb, xdp->data - xdp->data_hard_start);
 	skb_put(skb, xdp->data_end - xdp->data);
-	mvneta_rx_csum(pp, desc_status, skb);
+	skb->ip_summed = mvneta_rx_csum(pp, desc_status);
 
 	for (i = 0; i < num_frags; i++) {
 		skb_frag_t *frag = &sinfo->frags[i];
@@ -2535,7 +2531,7 @@ static int mvneta_rx_hwbm(struct napi_struct *napi,
 				     rx_bytes);
 
 			skb->protocol = eth_type_trans(skb, dev);
-			mvneta_rx_csum(pp, rx_status, skb);
+			skb->ip_summed = mvneta_rx_csum(pp, rx_status);
 			napi_gro_receive(napi, skb);
 
 			rcvd_pkts++;
@@ -2584,8 +2580,7 @@ static int mvneta_rx_hwbm(struct napi_struct *napi,
 		skb_put(skb, rx_bytes);
 
 		skb->protocol = eth_type_trans(skb, dev);
-
-		mvneta_rx_csum(pp, rx_status, skb);
+		skb->ip_summed = mvneta_rx_csum(pp, rx_status);
 
 		napi_gro_receive(napi, skb);
 	}
diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
index 9bca8c8f9f8d..01f6078bc859 100644
--- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
+++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
@@ -3543,21 +3543,17 @@ static void mvpp2_rx_error(struct mvpp2_port *port,
 }
 
 /* Handle RX checksum offload */
-static void mvpp2_rx_csum(struct mvpp2_port *port, u32 status,
-			  struct sk_buff *skb)
+static int mvpp2_rx_csum(struct mvpp2_port *port, u32 status)
 {
 	if (((status & MVPP2_RXD_L3_IP4) &&
 	     !(status & MVPP2_RXD_IP4_HEADER_ERR)) ||
 	    (status & MVPP2_RXD_L3_IP6))
 		if (((status & MVPP2_RXD_L4_UDP) ||
 		     (status & MVPP2_RXD_L4_TCP)) &&
-		     (status & MVPP2_RXD_L4_CSUM_OK)) {
-			skb->csum = 0;
-			skb->ip_summed = CHECKSUM_UNNECESSARY;
-			return;
-		}
+		     (status & MVPP2_RXD_L4_CSUM_OK))
+			return CHECKSUM_UNNECESSARY;
 
-	skb->ip_summed = CHECKSUM_NONE;
+	return CHECKSUM_NONE;
 }
 
 /* Allocate a new skb and add it to BM pool */
@@ -4012,7 +4008,7 @@ static int mvpp2_rx(struct mvpp2_port *port, struct napi_struct *napi,
 
 		skb_reserve(skb, MVPP2_MH_SIZE + MVPP2_SKB_HEADROOM);
 		skb_put(skb, rx_bytes);
-		mvpp2_rx_csum(port, rx_status, skb);
+		skb->ip_summed = mvpp2_rx_csum(port, rx_status);
 		skb->protocol = eth_type_trans(skb, dev);
 
 		napi_gro_receive(napi, skb);
-- 
2.31.1

Re: [PATCH net-next] net: marvell: return csum computation result from mvneta_rx_csum/mvpp2_rx_csum

From: Lorenzo Bianconi <hidden>
Date: 2021-06-22 08:22:37 

This is a preliminary patch to add hw csum hint support to
mvneta/mvpp2 xdp implementation

Tested-by: Matteo Croce <redacted>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Hi Dave and Jakub,

I have just noticed this patch is marked as "Not Applicable" in patchwork. I
tried to rebase it on top of net-next and it applies and compiles so I am
wondering why it is "not applicable". Am I missing something?

Regards,
Lorenzo
quoted hunk
---
 drivers/net/ethernet/marvell/mvneta.c         | 19 +++++++------------
 .../net/ethernet/marvell/mvpp2/mvpp2_main.c   | 14 +++++---------
 2 files changed, 12 insertions(+), 21 deletions(-)

diff --git a/drivers/net/ethernet/marvell/mvneta.c b/drivers/net/ethernet/marvell/mvneta.c
index c15ce06427d0..88a755034c39 100644
--- a/drivers/net/ethernet/marvell/mvneta.c
+++ b/drivers/net/ethernet/marvell/mvneta.c
@@ -1805,18 +1805,14 @@ static void mvneta_rx_error(struct mvneta_port *pp,
 }
 
 /* Handle RX checksum offload based on the descriptor's status */
-static void mvneta_rx_csum(struct mvneta_port *pp, u32 status,
-			   struct sk_buff *skb)
+static int mvneta_rx_csum(struct mvneta_port *pp, u32 status)
 {
 	if ((pp->dev->features & NETIF_F_RXCSUM) &&
 	    (status & MVNETA_RXD_L3_IP4) &&
-	    (status & MVNETA_RXD_L4_CSUM_OK)) {
-		skb->csum = 0;
-		skb->ip_summed = CHECKSUM_UNNECESSARY;
-		return;
-	}
+	    (status & MVNETA_RXD_L4_CSUM_OK))
+		return CHECKSUM_UNNECESSARY;
 
-	skb->ip_summed = CHECKSUM_NONE;
+	return CHECKSUM_NONE;
 }
 
 /* Return tx queue pointer (find last set bit) according to <cause> returned
@@ -2335,7 +2331,7 @@ mvneta_swbm_build_skb(struct mvneta_port *pp, struct page_pool *pool,
 
 	skb_reserve(skb, xdp->data - xdp->data_hard_start);
 	skb_put(skb, xdp->data_end - xdp->data);
-	mvneta_rx_csum(pp, desc_status, skb);
+	skb->ip_summed = mvneta_rx_csum(pp, desc_status);
 
 	for (i = 0; i < num_frags; i++) {
 		skb_frag_t *frag = &sinfo->frags[i];
@@ -2535,7 +2531,7 @@ static int mvneta_rx_hwbm(struct napi_struct *napi,
 				     rx_bytes);
 
 			skb->protocol = eth_type_trans(skb, dev);
-			mvneta_rx_csum(pp, rx_status, skb);
+			skb->ip_summed = mvneta_rx_csum(pp, rx_status);
 			napi_gro_receive(napi, skb);
 
 			rcvd_pkts++;
@@ -2584,8 +2580,7 @@ static int mvneta_rx_hwbm(struct napi_struct *napi,
 		skb_put(skb, rx_bytes);
 
 		skb->protocol = eth_type_trans(skb, dev);
-
-		mvneta_rx_csum(pp, rx_status, skb);
+		skb->ip_summed = mvneta_rx_csum(pp, rx_status);
 
 		napi_gro_receive(napi, skb);
 	}
diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
index 9bca8c8f9f8d..01f6078bc859 100644
--- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
+++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
@@ -3543,21 +3543,17 @@ static void mvpp2_rx_error(struct mvpp2_port *port,
 }
 
 /* Handle RX checksum offload */
-static void mvpp2_rx_csum(struct mvpp2_port *port, u32 status,
-			  struct sk_buff *skb)
+static int mvpp2_rx_csum(struct mvpp2_port *port, u32 status)
 {
 	if (((status & MVPP2_RXD_L3_IP4) &&
 	     !(status & MVPP2_RXD_IP4_HEADER_ERR)) ||
 	    (status & MVPP2_RXD_L3_IP6))
 		if (((status & MVPP2_RXD_L4_UDP) ||
 		     (status & MVPP2_RXD_L4_TCP)) &&
-		     (status & MVPP2_RXD_L4_CSUM_OK)) {
-			skb->csum = 0;
-			skb->ip_summed = CHECKSUM_UNNECESSARY;
-			return;
-		}
+		     (status & MVPP2_RXD_L4_CSUM_OK))
+			return CHECKSUM_UNNECESSARY;
 
-	skb->ip_summed = CHECKSUM_NONE;
+	return CHECKSUM_NONE;
 }
 
 /* Allocate a new skb and add it to BM pool */
@@ -4012,7 +4008,7 @@ static int mvpp2_rx(struct mvpp2_port *port, struct napi_struct *napi,
 
 		skb_reserve(skb, MVPP2_MH_SIZE + MVPP2_SKB_HEADROOM);
 		skb_put(skb, rx_bytes);
-		mvpp2_rx_csum(port, rx_status, skb);
+		skb->ip_summed = mvpp2_rx_csum(port, rx_status);
 		skb->protocol = eth_type_trans(skb, dev);
 
 		napi_gro_receive(napi, skb);
-- 
2.31.1

Re: [PATCH net-next] net: marvell: return csum computation result from mvneta_rx_csum/mvpp2_rx_csum

From: David Miller <davem@davemloft.net>
Date: 2021-06-22 17:06:02 

From: Lorenzo Bianconi <redacted>
Date: Tue, 22 Jun 2021 10:22:27 +0200
quoted
This is a preliminary patch to add hw csum hint support to
mvneta/mvpp2 xdp implementation

Tested-by: Matteo Croce <redacted>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Hi Dave and Jakub,

I have just noticed this patch is marked as "Not Applicable" in patchwork. I
tried to rebase it on top of net-next and it applies and compiles so I am
wondering why it is "not applicable". Am I missing something?
It did not apply cleanly for me, please resend.

Thank you.

Re: [PATCH net-next] net: marvell: return csum computation result from mvneta_rx_csum/mvpp2_rx_csum

From: Lorenzo Bianconi <hidden>
Date: 2021-06-22 17:19:41 

On Jun 22, David Miller wrote:
From: Lorenzo Bianconi <redacted>
Date: Tue, 22 Jun 2021 10:22:27 +0200
quoted
quoted
This is a preliminary patch to add hw csum hint support to
mvneta/mvpp2 xdp implementation

Tested-by: Matteo Croce <redacted>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Hi Dave and Jakub,

I have just noticed this patch is marked as "Not Applicable" in patchwork. I
tried to rebase it on top of net-next and it applies and compiles so I am
wondering why it is "not applicable". Am I missing something?
It did not apply cleanly for me, please resend.

Thank you.
ack, I will post v2 soon.

Regards,
Lorenzo
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help