* [PATCH net-next 0/3] net: core: add helper tcp_v6_gso_csum_prep
@ 2020-02-17 21:39 Heiner Kallweit
2020-02-17 21:40 ` [PATCH net-next 1/3] " Heiner Kallweit
` (2 more replies)
0 siblings, 3 replies; 10+ messages in thread
From: Heiner Kallweit @ 2020-02-17 21:39 UTC (permalink / raw)
To: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.
Cc: netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
Several network drivers for chips that support TSO6 share the same code
for preparing the TCP header. A difference is that some reset the
payload_len whilst others don't do this. Let's factor out this common
code to a new helper.
Heiner Kallweit (3):
net: core: add helper tcp_v6_gso_csum_prep
r8169: use new helper tcp_v6_gso_csum_prep
net: use new helper tcp_v6_gso_csum_prep
drivers/net/ethernet/atheros/alx/main.c | 5 +---
.../net/ethernet/atheros/atl1c/atl1c_main.c | 6 ++---
drivers/net/ethernet/brocade/bna/bnad.c | 7 +----
drivers/net/ethernet/cisco/enic/enic_main.c | 3 +--
drivers/net/ethernet/intel/e1000/e1000_main.c | 6 +----
drivers/net/ethernet/intel/e1000e/netdev.c | 5 +---
drivers/net/ethernet/jme.c | 7 +----
.../net/ethernet/pensando/ionic/ionic_txrx.c | 5 +---
drivers/net/ethernet/qualcomm/emac/emac-mac.c | 7 ++---
drivers/net/ethernet/realtek/r8169_main.c | 26 ++-----------------
drivers/net/ethernet/socionext/netsec.c | 6 +----
drivers/net/hyperv/netvsc_drv.c | 5 +---
drivers/net/usb/r8152.c | 26 ++-----------------
drivers/net/vmxnet3/vmxnet3_drv.c | 5 +---
include/net/ip6_checksum.h | 12 +++++++++
15 files changed, 30 insertions(+), 101 deletions(-)
--
2.25.0
^ permalink raw reply [flat|nested] 10+ messages in thread
* [PATCH net-next 1/3] net: core: add helper tcp_v6_gso_csum_prep
2020-02-17 21:39 [PATCH net-next 0/3] net: core: add helper tcp_v6_gso_csum_prep Heiner Kallweit
@ 2020-02-17 21:40 ` Heiner Kallweit
2020-02-18 18:25 ` Alexander Duyck
2020-02-17 21:40 ` [PATCH net-next 2/3] r8169: use new " Heiner Kallweit
2020-02-17 21:42 ` [PATCH net-next 3/3] net: " Heiner Kallweit
2 siblings, 1 reply; 10+ messages in thread
From: Heiner Kallweit @ 2020-02-17 21:40 UTC (permalink / raw)
To: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.
Cc: netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
Several network drivers for chips that support TSO6 share the same code
for preparing the TCP header. A difference is that some reset the
payload_len whilst others don't do this. Let's factor out this common
code to a new helper.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
---
include/net/ip6_checksum.h | 12 ++++++++++++
1 file changed, 12 insertions(+)
diff --git a/include/net/ip6_checksum.h b/include/net/ip6_checksum.h
index 7bec95df4..ef0130023 100644
--- a/include/net/ip6_checksum.h
+++ b/include/net/ip6_checksum.h
@@ -76,6 +76,18 @@ static inline void __tcp_v6_send_check(struct sk_buff *skb,
}
}
+static inline void tcp_v6_gso_csum_prep(struct sk_buff *skb,
+ bool clear_payload_len)
+{
+ struct ipv6hdr *ipv6h = ipv6_hdr(skb);
+ struct tcphdr *th = tcp_hdr(skb);
+
+ if (clear_payload_len)
+ ipv6h->payload_len = 0;
+
+ th->check = ~tcp_v6_check(0, &ipv6h->saddr, &ipv6h->daddr, 0);
+}
+
#if IS_ENABLED(CONFIG_IPV6)
static inline void tcp_v6_send_check(struct sock *sk, struct sk_buff *skb)
{
--
2.25.0
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH net-next 2/3] r8169: use new helper tcp_v6_gso_csum_prep
2020-02-17 21:39 [PATCH net-next 0/3] net: core: add helper tcp_v6_gso_csum_prep Heiner Kallweit
2020-02-17 21:40 ` [PATCH net-next 1/3] " Heiner Kallweit
@ 2020-02-17 21:40 ` Heiner Kallweit
2020-02-18 18:37 ` Alexander Duyck
2020-02-17 21:42 ` [PATCH net-next 3/3] net: " Heiner Kallweit
2 siblings, 1 reply; 10+ messages in thread
From: Heiner Kallweit @ 2020-02-17 21:40 UTC (permalink / raw)
To: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.
Cc: netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
Simplify the code by using new helper tcp_v6_gso_csum_prep.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
---
drivers/net/ethernet/realtek/r8169_main.c | 26 ++---------------------
1 file changed, 2 insertions(+), 24 deletions(-)
diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
index 5a9143b50..75ba10069 100644
--- a/drivers/net/ethernet/realtek/r8169_main.c
+++ b/drivers/net/ethernet/realtek/r8169_main.c
@@ -4108,29 +4108,6 @@ static bool rtl_test_hw_pad_bug(struct rtl8169_private *tp, struct sk_buff *skb)
return skb->len < ETH_ZLEN && tp->mac_version == RTL_GIGA_MAC_VER_34;
}
-/* msdn_giant_send_check()
- * According to the document of microsoft, the TCP Pseudo Header excludes the
- * packet length for IPv6 TCP large packets.
- */
-static int msdn_giant_send_check(struct sk_buff *skb)
-{
- const struct ipv6hdr *ipv6h;
- struct tcphdr *th;
- int ret;
-
- ret = skb_cow_head(skb, 0);
- if (ret)
- return ret;
-
- ipv6h = ipv6_hdr(skb);
- th = tcp_hdr(skb);
-
- th->check = 0;
- th->check = ~tcp_v6_check(0, &ipv6h->saddr, &ipv6h->daddr, 0);
-
- return ret;
-}
-
static void rtl8169_tso_csum_v1(struct sk_buff *skb, u32 *opts)
{
u32 mss = skb_shinfo(skb)->gso_size;
@@ -4163,9 +4140,10 @@ static bool rtl8169_tso_csum_v2(struct rtl8169_private *tp,
break;
case htons(ETH_P_IPV6):
- if (msdn_giant_send_check(skb))
+ if (skb_cow_head(skb, 0))
return false;
+ tcp_v6_gso_csum_prep(skb, false);
opts[0] |= TD1_GTSENV6;
break;
--
2.25.0
^ permalink raw reply related [flat|nested] 10+ messages in thread
* [PATCH net-next 3/3] net: use new helper tcp_v6_gso_csum_prep
2020-02-17 21:39 [PATCH net-next 0/3] net: core: add helper tcp_v6_gso_csum_prep Heiner Kallweit
2020-02-17 21:40 ` [PATCH net-next 1/3] " Heiner Kallweit
2020-02-17 21:40 ` [PATCH net-next 2/3] r8169: use new " Heiner Kallweit
@ 2020-02-17 21:42 ` Heiner Kallweit
2020-02-18 18:34 ` Alexander Duyck
2 siblings, 1 reply; 10+ messages in thread
From: Heiner Kallweit @ 2020-02-17 21:42 UTC (permalink / raw)
To: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.
Cc: netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
Use new helper tcp_v6_gso_csum_prep in additional network drivers.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
---
drivers/net/ethernet/atheros/alx/main.c | 5 +---
.../net/ethernet/atheros/atl1c/atl1c_main.c | 6 ++---
drivers/net/ethernet/brocade/bna/bnad.c | 7 +----
drivers/net/ethernet/cisco/enic/enic_main.c | 3 +--
drivers/net/ethernet/intel/e1000/e1000_main.c | 6 +----
drivers/net/ethernet/intel/e1000e/netdev.c | 5 +---
drivers/net/ethernet/jme.c | 7 +----
.../net/ethernet/pensando/ionic/ionic_txrx.c | 5 +---
drivers/net/ethernet/qualcomm/emac/emac-mac.c | 7 ++---
drivers/net/ethernet/socionext/netsec.c | 6 +----
drivers/net/hyperv/netvsc_drv.c | 5 +---
drivers/net/usb/r8152.c | 26 ++-----------------
drivers/net/vmxnet3/vmxnet3_drv.c | 5 +---
13 files changed, 16 insertions(+), 77 deletions(-)
diff --git a/drivers/net/ethernet/atheros/alx/main.c b/drivers/net/ethernet/atheros/alx/main.c
index 1dcbc486e..3e0215887 100644
--- a/drivers/net/ethernet/atheros/alx/main.c
+++ b/drivers/net/ethernet/atheros/alx/main.c
@@ -1416,10 +1416,7 @@ static int alx_tso(struct sk_buff *skb, struct alx_txd *first)
0, IPPROTO_TCP, 0);
first->word1 |= 1 << TPD_IPV4_SHIFT;
} else if (skb_is_gso_v6(skb)) {
- ipv6_hdr(skb)->payload_len = 0;
- tcp_hdr(skb)->check = ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr,
- 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, true);
/* LSOv2: the first TPD only provides the packet length */
first->adrl.l.pkt_len = skb->len;
first->word1 |= 1 << TPD_LSO_V2_SHIFT;
diff --git a/drivers/net/ethernet/atheros/atl1c/atl1c_main.c b/drivers/net/ethernet/atheros/atl1c/atl1c_main.c
index 4c0b1f855..482e18d0d 100644
--- a/drivers/net/ethernet/atheros/atl1c/atl1c_main.c
+++ b/drivers/net/ethernet/atheros/atl1c/atl1c_main.c
@@ -2025,10 +2025,8 @@ static int atl1c_tso_csum(struct atl1c_adapter *adapter,
"IPV6 tso with zero data??\n");
goto check_sum;
} else
- tcp_hdr(skb)->check = ~csum_ipv6_magic(
- &ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr,
- 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, false);
+
etpd->word1 |= 1 << TPD_LSO_EN_SHIFT;
etpd->word1 |= 1 << TPD_LSO_VER_SHIFT;
etpd->pkt_len = cpu_to_le32(skb->len);
diff --git a/drivers/net/ethernet/brocade/bna/bnad.c b/drivers/net/ethernet/brocade/bna/bnad.c
index 01a50a4b2..c301ad736 100644
--- a/drivers/net/ethernet/brocade/bna/bnad.c
+++ b/drivers/net/ethernet/brocade/bna/bnad.c
@@ -2504,12 +2504,7 @@ bnad_tso_prepare(struct bnad *bnad, struct sk_buff *skb)
IPPROTO_TCP, 0);
BNAD_UPDATE_CTR(bnad, tso4);
} else {
- struct ipv6hdr *ipv6h = ipv6_hdr(skb);
-
- ipv6h->payload_len = 0;
- tcp_hdr(skb)->check =
- ~csum_ipv6_magic(&ipv6h->saddr, &ipv6h->daddr, 0,
- IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, true);
BNAD_UPDATE_CTR(bnad, tso6);
}
diff --git a/drivers/net/ethernet/cisco/enic/enic_main.c b/drivers/net/ethernet/cisco/enic/enic_main.c
index ddf60dc9a..683c628ef 100644
--- a/drivers/net/ethernet/cisco/enic/enic_main.c
+++ b/drivers/net/ethernet/cisco/enic/enic_main.c
@@ -696,8 +696,7 @@ static void enic_preload_tcp_csum(struct sk_buff *skb)
tcp_hdr(skb)->check = ~csum_tcpudp_magic(ip_hdr(skb)->saddr,
ip_hdr(skb)->daddr, 0, IPPROTO_TCP, 0);
} else if (skb->protocol == cpu_to_be16(ETH_P_IPV6)) {
- tcp_hdr(skb)->check = ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr, 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, false);
}
}
diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c
index 2bced34c1..0664985e8 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_main.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_main.c
@@ -2715,11 +2715,7 @@ static int e1000_tso(struct e1000_adapter *adapter,
cmd_length = E1000_TXD_CMD_IP;
ipcse = skb_transport_offset(skb) - 1;
} else if (skb_is_gso_v6(skb)) {
- ipv6_hdr(skb)->payload_len = 0;
- tcp_hdr(skb)->check =
- ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr,
- 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, true);
ipcse = 0;
}
ipcss = skb_network_offset(skb);
diff --git a/drivers/net/ethernet/intel/e1000e/netdev.c b/drivers/net/ethernet/intel/e1000e/netdev.c
index db4ea58ba..7dda7d407 100644
--- a/drivers/net/ethernet/intel/e1000e/netdev.c
+++ b/drivers/net/ethernet/intel/e1000e/netdev.c
@@ -5462,10 +5462,7 @@ static int e1000_tso(struct e1000_ring *tx_ring, struct sk_buff *skb,
cmd_length = E1000_TXD_CMD_IP;
ipcse = skb_transport_offset(skb) - 1;
} else if (skb_is_gso_v6(skb)) {
- ipv6_hdr(skb)->payload_len = 0;
- tcp_hdr(skb)->check = ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr,
- 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, true);
ipcse = 0;
}
ipcss = skb_network_offset(skb);
diff --git a/drivers/net/ethernet/jme.c b/drivers/net/ethernet/jme.c
index 2e4975572..cde9be497 100644
--- a/drivers/net/ethernet/jme.c
+++ b/drivers/net/ethernet/jme.c
@@ -2077,12 +2077,7 @@ jme_tx_tso(struct sk_buff *skb, __le16 *mss, u8 *flags)
IPPROTO_TCP,
0);
} else {
- struct ipv6hdr *ip6h = ipv6_hdr(skb);
-
- tcp_hdr(skb)->check = ~csum_ipv6_magic(&ip6h->saddr,
- &ip6h->daddr, 0,
- IPPROTO_TCP,
- 0);
+ tcp_v6_gso_csum_prep(skb, false);
}
return 0;
diff --git a/drivers/net/ethernet/pensando/ionic/ionic_txrx.c b/drivers/net/ethernet/pensando/ionic/ionic_txrx.c
index e452f4242..3d8469d97 100644
--- a/drivers/net/ethernet/pensando/ionic/ionic_txrx.c
+++ b/drivers/net/ethernet/pensando/ionic/ionic_txrx.c
@@ -632,10 +632,7 @@ static int ionic_tx_tcp_pseudo_csum(struct sk_buff *skb)
ip_hdr(skb)->daddr,
0, IPPROTO_TCP, 0);
} else if (skb->protocol == cpu_to_be16(ETH_P_IPV6)) {
- tcp_hdr(skb)->check =
- ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr,
- 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, false);
}
return 0;
diff --git a/drivers/net/ethernet/qualcomm/emac/emac-mac.c b/drivers/net/ethernet/qualcomm/emac/emac-mac.c
index bebe38d74..01bcc5e68 100644
--- a/drivers/net/ethernet/qualcomm/emac/emac-mac.c
+++ b/drivers/net/ethernet/qualcomm/emac/emac-mac.c
@@ -1288,11 +1288,8 @@ static int emac_tso_csum(struct emac_adapter *adpt,
memset(tpd, 0, sizeof(*tpd));
memset(&extra_tpd, 0, sizeof(extra_tpd));
- ipv6_hdr(skb)->payload_len = 0;
- tcp_hdr(skb)->check =
- ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr,
- 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, true);
+
TPD_PKT_LEN_SET(&extra_tpd, skb->len);
TPD_LSO_SET(&extra_tpd, 1);
TPD_LSOV_SET(&extra_tpd, 1);
diff --git a/drivers/net/ethernet/socionext/netsec.c b/drivers/net/ethernet/socionext/netsec.c
index e8224b543..d7a033053 100644
--- a/drivers/net/ethernet/socionext/netsec.c
+++ b/drivers/net/ethernet/socionext/netsec.c
@@ -1148,11 +1148,7 @@ static netdev_tx_t netsec_netdev_start_xmit(struct sk_buff *skb,
~tcp_v4_check(0, ip_hdr(skb)->saddr,
ip_hdr(skb)->daddr, 0);
} else {
- ipv6_hdr(skb)->payload_len = 0;
- tcp_hdr(skb)->check =
- ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr,
- 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, true);
}
tx_ctrl.tcp_seg_offload_flag = true;
diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c
index 65e12cb07..f41e48634 100644
--- a/drivers/net/hyperv/netvsc_drv.c
+++ b/drivers/net/hyperv/netvsc_drv.c
@@ -638,10 +638,7 @@ static int netvsc_xmit(struct sk_buff *skb, struct net_device *net, bool xdp_tx)
} else {
lso_info->lso_v2_transmit.ip_version =
NDIS_TCP_LARGE_SEND_OFFLOAD_IPV6;
- ipv6_hdr(skb)->payload_len = 0;
- tcp_hdr(skb)->check =
- ~csum_ipv6_magic(&ipv6_hdr(skb)->saddr,
- &ipv6_hdr(skb)->daddr, 0, IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, true);
}
lso_info->lso_v2_transmit.tcp_header_offset = skb_transport_offset(skb);
lso_info->lso_v2_transmit.mss = skb_shinfo(skb)->gso_size;
diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c
index 78ddbaf64..4ad2a1d42 100644
--- a/drivers/net/usb/r8152.c
+++ b/drivers/net/usb/r8152.c
@@ -1948,29 +1948,6 @@ static void r8152_csum_workaround(struct r8152 *tp, struct sk_buff *skb,
}
}
-/* msdn_giant_send_check()
- * According to the document of microsoft, the TCP Pseudo Header excludes the
- * packet length for IPv6 TCP large packets.
- */
-static int msdn_giant_send_check(struct sk_buff *skb)
-{
- const struct ipv6hdr *ipv6h;
- struct tcphdr *th;
- int ret;
-
- ret = skb_cow_head(skb, 0);
- if (ret)
- return ret;
-
- ipv6h = ipv6_hdr(skb);
- th = tcp_hdr(skb);
-
- th->check = 0;
- th->check = ~tcp_v6_check(0, &ipv6h->saddr, &ipv6h->daddr, 0);
-
- return ret;
-}
-
static inline void rtl_tx_vlan_tag(struct tx_desc *desc, struct sk_buff *skb)
{
if (skb_vlan_tag_present(skb)) {
@@ -2016,10 +1993,11 @@ static int r8152_tx_csum(struct r8152 *tp, struct tx_desc *desc,
break;
case htons(ETH_P_IPV6):
- if (msdn_giant_send_check(skb)) {
+ if (skb_cow_head(skb, 0)) {
ret = TX_CSUM_TSO;
goto unavailable;
}
+ tcp_v6_gso_csum_prep(skb, false);
opts1 |= GTSENDV6;
break;
diff --git a/drivers/net/vmxnet3/vmxnet3_drv.c b/drivers/net/vmxnet3/vmxnet3_drv.c
index 18f152fa0..92c2ecf3f 100644
--- a/drivers/net/vmxnet3/vmxnet3_drv.c
+++ b/drivers/net/vmxnet3/vmxnet3_drv.c
@@ -942,10 +942,7 @@ vmxnet3_prepare_tso(struct sk_buff *skb,
tcph->check = ~csum_tcpudp_magic(iph->saddr, iph->daddr, 0,
IPPROTO_TCP, 0);
} else if (ctx->ipv6) {
- struct ipv6hdr *iph = ipv6_hdr(skb);
-
- tcph->check = ~csum_ipv6_magic(&iph->saddr, &iph->daddr, 0,
- IPPROTO_TCP, 0);
+ tcp_v6_gso_csum_prep(skb, false);
}
}
--
2.25.0
^ permalink raw reply related [flat|nested] 10+ messages in thread
* Re: [PATCH net-next 1/3] net: core: add helper tcp_v6_gso_csum_prep
2020-02-17 21:40 ` [PATCH net-next 1/3] " Heiner Kallweit
@ 2020-02-18 18:25 ` Alexander Duyck
2020-02-18 18:53 ` Heiner Kallweit
0 siblings, 1 reply; 10+ messages in thread
From: Alexander Duyck @ 2020-02-18 18:25 UTC (permalink / raw)
To: Heiner Kallweit
Cc: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.,
netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
On Mon, Feb 17, 2020 at 1:41 PM Heiner Kallweit <hkallweit1@gmail.com> wrote:
>
> Several network drivers for chips that support TSO6 share the same code
> for preparing the TCP header. A difference is that some reset the
> payload_len whilst others don't do this. Let's factor out this common
> code to a new helper.
>
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> ---
> include/net/ip6_checksum.h | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/include/net/ip6_checksum.h b/include/net/ip6_checksum.h
> index 7bec95df4..ef0130023 100644
> --- a/include/net/ip6_checksum.h
> +++ b/include/net/ip6_checksum.h
> @@ -76,6 +76,18 @@ static inline void __tcp_v6_send_check(struct sk_buff *skb,
> }
> }
>
> +static inline void tcp_v6_gso_csum_prep(struct sk_buff *skb,
> + bool clear_payload_len)
> +{
> + struct ipv6hdr *ipv6h = ipv6_hdr(skb);
> + struct tcphdr *th = tcp_hdr(skb);
> +
> + if (clear_payload_len)
> + ipv6h->payload_len = 0;
> +
> + th->check = ~tcp_v6_check(0, &ipv6h->saddr, &ipv6h->daddr, 0);
> +}
> +
> #if IS_ENABLED(CONFIG_IPV6)
> static inline void tcp_v6_send_check(struct sock *sk, struct sk_buff *skb)
> {
So functionally I believe this is correct. The only piece I have a
question about is if we should just force the clear_payload_len as
always being the case since the value should either be
ignored/overwritten in any GSO case anyway.
Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH net-next 3/3] net: use new helper tcp_v6_gso_csum_prep
2020-02-17 21:42 ` [PATCH net-next 3/3] net: " Heiner Kallweit
@ 2020-02-18 18:34 ` Alexander Duyck
2020-02-18 18:54 ` Heiner Kallweit
0 siblings, 1 reply; 10+ messages in thread
From: Alexander Duyck @ 2020-02-18 18:34 UTC (permalink / raw)
To: Heiner Kallweit
Cc: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.,
netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
On Mon, Feb 17, 2020 at 1:43 PM Heiner Kallweit <hkallweit1@gmail.com> wrote:
>
> Use new helper tcp_v6_gso_csum_prep in additional network drivers.
>
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> ---
> drivers/net/ethernet/atheros/alx/main.c | 5 +---
> .../net/ethernet/atheros/atl1c/atl1c_main.c | 6 ++---
> drivers/net/ethernet/brocade/bna/bnad.c | 7 +----
> drivers/net/ethernet/cisco/enic/enic_main.c | 3 +--
> drivers/net/ethernet/intel/e1000/e1000_main.c | 6 +----
> drivers/net/ethernet/intel/e1000e/netdev.c | 5 +---
> drivers/net/ethernet/jme.c | 7 +----
> .../net/ethernet/pensando/ionic/ionic_txrx.c | 5 +---
> drivers/net/ethernet/qualcomm/emac/emac-mac.c | 7 ++---
> drivers/net/ethernet/socionext/netsec.c | 6 +----
> drivers/net/hyperv/netvsc_drv.c | 5 +---
> drivers/net/usb/r8152.c | 26 ++-----------------
> drivers/net/vmxnet3/vmxnet3_drv.c | 5 +---
> 13 files changed, 16 insertions(+), 77 deletions(-)
>
It might make sense to break this up into several smaller patches
based on the maintainers for the various driver bits.
So the changes all look fine to me, but I am not that familiar with
the non-Intel drivers.
Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH net-next 2/3] r8169: use new helper tcp_v6_gso_csum_prep
2020-02-17 21:40 ` [PATCH net-next 2/3] r8169: use new " Heiner Kallweit
@ 2020-02-18 18:37 ` Alexander Duyck
2020-02-18 18:45 ` Heiner Kallweit
0 siblings, 1 reply; 10+ messages in thread
From: Alexander Duyck @ 2020-02-18 18:37 UTC (permalink / raw)
To: Heiner Kallweit
Cc: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.,
netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
On Mon, Feb 17, 2020 at 1:42 PM Heiner Kallweit <hkallweit1@gmail.com> wrote:
>
> Simplify the code by using new helper tcp_v6_gso_csum_prep.
>
> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
> ---
> drivers/net/ethernet/realtek/r8169_main.c | 26 ++---------------------
> 1 file changed, 2 insertions(+), 24 deletions(-)
>
> diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
> index 5a9143b50..75ba10069 100644
> --- a/drivers/net/ethernet/realtek/r8169_main.c
> +++ b/drivers/net/ethernet/realtek/r8169_main.c
> @@ -4108,29 +4108,6 @@ static bool rtl_test_hw_pad_bug(struct rtl8169_private *tp, struct sk_buff *skb)
> return skb->len < ETH_ZLEN && tp->mac_version == RTL_GIGA_MAC_VER_34;
> }
>
> -/* msdn_giant_send_check()
> - * According to the document of microsoft, the TCP Pseudo Header excludes the
> - * packet length for IPv6 TCP large packets.
> - */
> -static int msdn_giant_send_check(struct sk_buff *skb)
> -{
> - const struct ipv6hdr *ipv6h;
> - struct tcphdr *th;
> - int ret;
> -
> - ret = skb_cow_head(skb, 0);
> - if (ret)
> - return ret;
> -
> - ipv6h = ipv6_hdr(skb);
> - th = tcp_hdr(skb);
> -
> - th->check = 0;
> - th->check = ~tcp_v6_check(0, &ipv6h->saddr, &ipv6h->daddr, 0);
> -
> - return ret;
> -}
> -
> static void rtl8169_tso_csum_v1(struct sk_buff *skb, u32 *opts)
> {
> u32 mss = skb_shinfo(skb)->gso_size;
> @@ -4163,9 +4140,10 @@ static bool rtl8169_tso_csum_v2(struct rtl8169_private *tp,
> break;
>
> case htons(ETH_P_IPV6):
> - if (msdn_giant_send_check(skb))
> + if (skb_cow_head(skb, 0))
> return false;
>
> + tcp_v6_gso_csum_prep(skb, false);
> opts[0] |= TD1_GTSENV6;
> break;
>
This change looks more or less identical to the one you made in
"drivers/net/usb/r8152.c" for patch 3. If you have to resubmit it
might make sense to pull that change out and include it here since
they are both essentially the same change.
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH net-next 2/3] r8169: use new helper tcp_v6_gso_csum_prep
2020-02-18 18:37 ` Alexander Duyck
@ 2020-02-18 18:45 ` Heiner Kallweit
0 siblings, 0 replies; 10+ messages in thread
From: Heiner Kallweit @ 2020-02-18 18:45 UTC (permalink / raw)
To: Alexander Duyck
Cc: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.,
netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
On 18.02.2020 19:37, Alexander Duyck wrote:
> On Mon, Feb 17, 2020 at 1:42 PM Heiner Kallweit <hkallweit1@gmail.com> wrote:
>>
>> Simplify the code by using new helper tcp_v6_gso_csum_prep.
>>
>> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
>> ---
>> drivers/net/ethernet/realtek/r8169_main.c | 26 ++---------------------
>> 1 file changed, 2 insertions(+), 24 deletions(-)
>>
>> diff --git a/drivers/net/ethernet/realtek/r8169_main.c b/drivers/net/ethernet/realtek/r8169_main.c
>> index 5a9143b50..75ba10069 100644
>> --- a/drivers/net/ethernet/realtek/r8169_main.c
>> +++ b/drivers/net/ethernet/realtek/r8169_main.c
>> @@ -4108,29 +4108,6 @@ static bool rtl_test_hw_pad_bug(struct rtl8169_private *tp, struct sk_buff *skb)
>> return skb->len < ETH_ZLEN && tp->mac_version == RTL_GIGA_MAC_VER_34;
>> }
>>
>> -/* msdn_giant_send_check()
>> - * According to the document of microsoft, the TCP Pseudo Header excludes the
>> - * packet length for IPv6 TCP large packets.
>> - */
>> -static int msdn_giant_send_check(struct sk_buff *skb)
>> -{
>> - const struct ipv6hdr *ipv6h;
>> - struct tcphdr *th;
>> - int ret;
>> -
>> - ret = skb_cow_head(skb, 0);
>> - if (ret)
>> - return ret;
>> -
>> - ipv6h = ipv6_hdr(skb);
>> - th = tcp_hdr(skb);
>> -
>> - th->check = 0;
>> - th->check = ~tcp_v6_check(0, &ipv6h->saddr, &ipv6h->daddr, 0);
>> -
>> - return ret;
>> -}
>> -
>> static void rtl8169_tso_csum_v1(struct sk_buff *skb, u32 *opts)
>> {
>> u32 mss = skb_shinfo(skb)->gso_size;
>> @@ -4163,9 +4140,10 @@ static bool rtl8169_tso_csum_v2(struct rtl8169_private *tp,
>> break;
>>
>> case htons(ETH_P_IPV6):
>> - if (msdn_giant_send_check(skb))
>> + if (skb_cow_head(skb, 0))
>> return false;
>>
>> + tcp_v6_gso_csum_prep(skb, false);
>> opts[0] |= TD1_GTSENV6;
>> break;
>>
>
> This change looks more or less identical to the one you made in
> "drivers/net/usb/r8152.c" for patch 3. If you have to resubmit it
> might make sense to pull that change out and include it here since
> they are both essentially the same change.
>
Right, it's the same change. I just treated r8169 separately because
I happen to be maintainer of it.
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH net-next 1/3] net: core: add helper tcp_v6_gso_csum_prep
2020-02-18 18:25 ` Alexander Duyck
@ 2020-02-18 18:53 ` Heiner Kallweit
0 siblings, 0 replies; 10+ messages in thread
From: Heiner Kallweit @ 2020-02-18 18:53 UTC (permalink / raw)
To: Alexander Duyck
Cc: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.,
netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
On 18.02.2020 19:25, Alexander Duyck wrote:
> On Mon, Feb 17, 2020 at 1:41 PM Heiner Kallweit <hkallweit1@gmail.com> wrote:
>>
>> Several network drivers for chips that support TSO6 share the same code
>> for preparing the TCP header. A difference is that some reset the
>> payload_len whilst others don't do this. Let's factor out this common
>> code to a new helper.
>>
>> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
>> ---
>> include/net/ip6_checksum.h | 12 ++++++++++++
>> 1 file changed, 12 insertions(+)
>>
>> diff --git a/include/net/ip6_checksum.h b/include/net/ip6_checksum.h
>> index 7bec95df4..ef0130023 100644
>> --- a/include/net/ip6_checksum.h
>> +++ b/include/net/ip6_checksum.h
>> @@ -76,6 +76,18 @@ static inline void __tcp_v6_send_check(struct sk_buff *skb,
>> }
>> }
>>
>> +static inline void tcp_v6_gso_csum_prep(struct sk_buff *skb,
>> + bool clear_payload_len)
>> +{
>> + struct ipv6hdr *ipv6h = ipv6_hdr(skb);
>> + struct tcphdr *th = tcp_hdr(skb);
>> +
>> + if (clear_payload_len)
>> + ipv6h->payload_len = 0;
>> +
>> + th->check = ~tcp_v6_check(0, &ipv6h->saddr, &ipv6h->daddr, 0);
>> +}
>> +
>> #if IS_ENABLED(CONFIG_IPV6)
>> static inline void tcp_v6_send_check(struct sock *sk, struct sk_buff *skb)
>> {
>
> So functionally I believe this is correct. The only piece I have a
> question about is if we should just force the clear_payload_len as
> always being the case since the value should either be
> ignored/overwritten in any GSO case anyway.
>
I also thought about this and just wasn't sure whether this functional
change may break any driver. But yes, then let's change it this way.
Breaking down the series into smaller patches makes it easier to
revert an individual patch in case of a problem.
> Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
>
Thanks for the review.
^ permalink raw reply [flat|nested] 10+ messages in thread
* Re: [PATCH net-next 3/3] net: use new helper tcp_v6_gso_csum_prep
2020-02-18 18:34 ` Alexander Duyck
@ 2020-02-18 18:54 ` Heiner Kallweit
0 siblings, 0 replies; 10+ messages in thread
From: Heiner Kallweit @ 2020-02-18 18:54 UTC (permalink / raw)
To: Alexander Duyck
Cc: David Miller, Realtek linux nic maintainers, Jay Cliburn,
Chris Snook, Rasesh Mody, Sudarsana Kalluru, GR-Linux-NIC-Dev,
Christian Benvenuti, Govindarajulu Varadarajan, Parvi Kaustubhi,
Jeff Kirsher, Guo-Fu Tseng, Shannon Nelson, Pensando Drivers,
Timur Tabi, Jassi Brar, Ilias Apalodimas, K. Y. Srinivasan,
Haiyang Zhang, Stephen Hemminger, Sasha Levin, Ronak Doshi,
VMware, Inc.,
netdev, Linux Kernel Mailing List, intel-wired-lan, linux-hyperv,
Linux USB Mailing List
On 18.02.2020 19:34, Alexander Duyck wrote:
> On Mon, Feb 17, 2020 at 1:43 PM Heiner Kallweit <hkallweit1@gmail.com> wrote:
>>
>> Use new helper tcp_v6_gso_csum_prep in additional network drivers.
>>
>> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
>> ---
>> drivers/net/ethernet/atheros/alx/main.c | 5 +---
>> .../net/ethernet/atheros/atl1c/atl1c_main.c | 6 ++---
>> drivers/net/ethernet/brocade/bna/bnad.c | 7 +----
>> drivers/net/ethernet/cisco/enic/enic_main.c | 3 +--
>> drivers/net/ethernet/intel/e1000/e1000_main.c | 6 +----
>> drivers/net/ethernet/intel/e1000e/netdev.c | 5 +---
>> drivers/net/ethernet/jme.c | 7 +----
>> .../net/ethernet/pensando/ionic/ionic_txrx.c | 5 +---
>> drivers/net/ethernet/qualcomm/emac/emac-mac.c | 7 ++---
>> drivers/net/ethernet/socionext/netsec.c | 6 +----
>> drivers/net/hyperv/netvsc_drv.c | 5 +---
>> drivers/net/usb/r8152.c | 26 ++-----------------
>> drivers/net/vmxnet3/vmxnet3_drv.c | 5 +---
>> 13 files changed, 16 insertions(+), 77 deletions(-)
>>
>
> It might make sense to break this up into several smaller patches
> based on the maintainers for the various driver bits.
>
OK
> So the changes all look fine to me, but I am not that familiar with
> the non-Intel drivers.
>
> Reviewed-by: Alexander Duyck <alexander.h.duyck@linux.intel.com>
>
^ permalink raw reply [flat|nested] 10+ messages in thread
end of thread, other threads:[~2020-02-18 18:54 UTC | newest]
Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-02-17 21:39 [PATCH net-next 0/3] net: core: add helper tcp_v6_gso_csum_prep Heiner Kallweit
2020-02-17 21:40 ` [PATCH net-next 1/3] " Heiner Kallweit
2020-02-18 18:25 ` Alexander Duyck
2020-02-18 18:53 ` Heiner Kallweit
2020-02-17 21:40 ` [PATCH net-next 2/3] r8169: use new " Heiner Kallweit
2020-02-18 18:37 ` Alexander Duyck
2020-02-18 18:45 ` Heiner Kallweit
2020-02-17 21:42 ` [PATCH net-next 3/3] net: " Heiner Kallweit
2020-02-18 18:34 ` Alexander Duyck
2020-02-18 18:54 ` Heiner Kallweit
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).