[PATCH net-next v2 1/2] IPv6/GRO: generic helper to remove temporary HBH/jumbo header in driver

[PATCH net-next v2 1/2] IPv6/GRO: generic helper to remove temporary HBH/jumbo header in driver

[Coco Li]

IPv6/TCP and GRO stacks can build big TCP packets with an added
temporary Hop By Hop header.

Is GSO is not involved, then the temporary header needs to be removed in
the driver. This patch provides a generic helper for drivers that need
to modify their headers in place.

Signed-off-by: Coco Li <lixiaoyan@google.com>
---
 include/net/ipv6.h     | 36 ++++++++++++++++++++++++++++++++++++
 net/ipv6/ip6_offload.c | 26 ++++----------------------
 2 files changed, 40 insertions(+), 22 deletions(-)

diff --git a/include/net/ipv6.h b/include/net/ipv6.h
index d383c895592a..c5a1daaf5056 100644
--- a/include/net/ipv6.h
+++ b/include/net/ipv6.h
@@ -500,6 +500,42 @@ static inline int ipv6_has_hopopt_jumbo(const struct sk_buff *skb)
 	return jhdr->nexthdr;
 }
 
+/* Return 0 if HBH header is successfully removed
+ * Or if HBH removal is unnecessary (packet is not big TCP)
+ * Return error to indicate dropping the packet
+ */
+static inline int ipv6_hopopt_jumbo_remove(struct sk_buff *skb)
+{
+	const int hophdr_len = sizeof(struct hop_jumbo_hdr);
+	int nexthdr = ipv6_has_hopopt_jumbo(skb);
+	struct ipv6hdr *h6;
+	int err = 0;
+
+	if (!nexthdr)
+		return err;
+
+	err = skb_cow_head(skb, 0);
+	if (err)
+		return err;
+
+	/* Remove the HBH header.
+	 * Layout: [Ethernet header][IPv6 header][HBH][L4 Header]
+	 */
+	memmove(skb_mac_header(skb) + hophdr_len, skb_mac_header(skb),
+		skb_network_header(skb) - skb_mac_header(skb) +
+		sizeof(struct ipv6hdr));
+
+	skb->data += hophdr_len;
+	skb->len -= hophdr_len;
+	skb->network_header += hophdr_len;
+	skb->mac_header += hophdr_len;
+
+	h6 = ipv6_hdr(skb);
+	h6->nexthdr = nexthdr;
+
+	return err;
+}
+
 static inline bool ipv6_accept_ra(struct inet6_dev *idev)
 {
 	/* If forwarding is enabled, RA are not accepted unless the special
diff --git a/net/ipv6/ip6_offload.c b/net/ipv6/ip6_offload.c
index 3ee345672849..60a82342fe57 100644
--- a/net/ipv6/ip6_offload.c
+++ b/net/ipv6/ip6_offload.c
@@ -77,7 +77,7 @@ static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb,
 	struct sk_buff *segs = ERR_PTR(-EINVAL);
 	struct ipv6hdr *ipv6h;
 	const struct net_offload *ops;
-	int proto, nexthdr;
+	int proto, err;
 	struct frag_hdr *fptr;
 	unsigned int payload_len;
 	u8 *prevhdr;
@@ -87,28 +87,10 @@ static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb,
 	bool gso_partial;
 
 	skb_reset_network_header(skb);
-	nexthdr = ipv6_has_hopopt_jumbo(skb);
-	if (nexthdr) {
-		const int hophdr_len = sizeof(struct hop_jumbo_hdr);
-		int err;
+	err = ipv6_hopopt_jumbo_remove(skb);
+	if (err)
+		return ERR_PTR(err);
 
-		err = skb_cow_head(skb, 0);
-		if (err < 0)
-			return ERR_PTR(err);
-
-		/* remove the HBH header.
-		 * Layout: [Ethernet header][IPv6 header][HBH][TCP header]
-		 */
-		memmove(skb_mac_header(skb) + hophdr_len,
-			skb_mac_header(skb),
-			ETH_HLEN + sizeof(struct ipv6hdr));
-		skb->data += hophdr_len;
-		skb->len -= hophdr_len;
-		skb->network_header += hophdr_len;
-		skb->mac_header += hophdr_len;
-		ipv6h = (struct ipv6hdr *)skb->data;
-		ipv6h->nexthdr = nexthdr;
-	}
 	nhoff = skb_network_header(skb) - skb_mac_header(skb);
 	if (unlikely(!pskb_may_pull(skb, sizeof(*ipv6h))))
 		goto out;
-- 
2.38.1.584.g0f3c55d4c2-goog

[RFC net-next v2 2/2] bnxt: Use generic HBH removal helper in tx path

[Coco Li]

Eric Dumazet implemented Big TCP that allowed bigger TSO/GRO packet sizes
for IPv6 traffic. See patch series:
'commit 89527be8d8d6 ("net: add IFLA_TSO_{MAX_SIZE|SEGS} attributes")'

This reduces the number of packets traversing the networking stack and
should usually improves performance. However, it also inserts a
temporary Hop-by-hop IPv6 extension header.

Using the HBH header removal method in the previous path, the extra header
be removed in bnxt drivers to allow it to send big TCP packets (bigger
TSO packets) as well.

Tested:
Compiled locally

To further test functional correctness, update the GSO/GRO limit on the
physical NIC:

ip link set eth0 gso_max_size 181000
ip link set eth0 gro_max_size 181000

Note that if there are bonding or ipvan devices on top of the physical
NIC, their GSO sizes need to be updated as well.

Then, IPv6/TCP packets with sizes larger than 64k can be observed.

Signed-off-by: Coco Li <lixiaoyan@google.com>
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
index 0fe164b42c5d..2bfa5e9fb179 100644
--- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
+++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
@@ -389,6 +389,9 @@ static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, struct net_device *dev)
 			return NETDEV_TX_BUSY;
 	}
 
+	if (unlikely(ipv6_hopopt_jumbo_remove(skb)))
+		goto tx_free;
+
 	length = skb->len;
 	len = skb_headlen(skb);
 	last_frag = skb_shinfo(skb)->nr_frags;
@@ -13657,6 +13660,7 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 		dev->features &= ~NETIF_F_LRO;
 	dev->priv_flags |= IFF_UNICAST_FLT;
 
+	netif_set_tso_max_size(dev, GSO_MAX_SIZE);
 #ifdef CONFIG_BNXT_SRIOV
 	init_waitqueue_head(&bp->sriov_cfg_wait);
 #endif
-- 
2.38.1.584.g0f3c55d4c2-goog

Re: [RFC net-next v2 2/2] bnxt: Use generic HBH removal helper in tx path

[Michael Chan]

[-- Attachment #1: Type: text/plain, Size: 2123 bytes --]

On Wed, Nov 23, 2022 at 11:16 AM Coco Li <lixiaoyan@google.com> wrote:
>
> Eric Dumazet implemented Big TCP that allowed bigger TSO/GRO packet sizes
> for IPv6 traffic. See patch series:
> 'commit 89527be8d8d6 ("net: add IFLA_TSO_{MAX_SIZE|SEGS} attributes")'
>
> This reduces the number of packets traversing the networking stack and
> should usually improves performance. However, it also inserts a
> temporary Hop-by-hop IPv6 extension header.
>
> Using the HBH header removal method in the previous path, the extra header
> be removed in bnxt drivers to allow it to send big TCP packets (bigger
> TSO packets) as well.
>
> Tested:
> Compiled locally
>
> To further test functional correctness, update the GSO/GRO limit on the
> physical NIC:
>
> ip link set eth0 gso_max_size 181000
> ip link set eth0 gro_max_size 181000
>
> Note that if there are bonding or ipvan devices on top of the physical
> NIC, their GSO sizes need to be updated as well.
>
> Then, IPv6/TCP packets with sizes larger than 64k can be observed.
>
> Signed-off-by: Coco Li <lixiaoyan@google.com>
> ---
>  drivers/net/ethernet/broadcom/bnxt/bnxt.c | 4 ++++
>  1 file changed, 4 insertions(+)
>
> diff --git a/drivers/net/ethernet/broadcom/bnxt/bnxt.c b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
> index 0fe164b42c5d..2bfa5e9fb179 100644
> --- a/drivers/net/ethernet/broadcom/bnxt/bnxt.c
> +++ b/drivers/net/ethernet/broadcom/bnxt/bnxt.c
> @@ -389,6 +389,9 @@ static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, struct net_device *dev)
>                         return NETDEV_TX_BUSY;
>         }
>
> +       if (unlikely(ipv6_hopopt_jumbo_remove(skb)))
> +               goto tx_free;
> +
>         length = skb->len;
>         len = skb_headlen(skb);
>         last_frag = skb_shinfo(skb)->nr_frags;
> @@ -13657,6 +13660,7 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
>                 dev->features &= ~NETIF_F_LRO;
>         dev->priv_flags |= IFF_UNICAST_FLT;
>
> +       netif_set_tso_max_size(dev, GSO_MAX_SIZE);

Our chips can only transmit TSO packets up to 64K bytes, so I think
this won't work.

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]

Re: [RFC net-next v2 2/2] bnxt: Use generic HBH removal helper in tx path

[Michael Chan]

[-- Attachment #1: Type: text/plain, Size: 648 bytes --]

On Wed, Nov 23, 2022 at 11:42 AM Michael Chan <michael.chan@broadcom.com> wrote:
>
> On Wed, Nov 23, 2022 at 11:16 AM Coco Li <lixiaoyan@google.com> wrote:
> >
> > @@ -13657,6 +13660,7 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
> >                 dev->features &= ~NETIF_F_LRO;
> >         dev->priv_flags |= IFF_UNICAST_FLT;
> >
> > +       netif_set_tso_max_size(dev, GSO_MAX_SIZE);
>
> Our chips can only transmit TSO packets up to 64K bytes, so I think
> this won't work.

I wanted to double check with the hardware team but there's no one in
the office today.  I will confirm early next week.  Thanks.

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]

Re: [RFC net-next v2 2/2] bnxt: Use generic HBH removal helper in tx path

[Michael Chan]

[-- Attachment #1: Type: text/plain, Size: 1455 bytes --]

On Wed, Nov 23, 2022 at 11:42 AM Michael Chan <michael.chan@broadcom.com> wrote:
>
> On Wed, Nov 23, 2022 at 11:16 AM Coco Li <lixiaoyan@google.com> wrote:
> >
> > Eric Dumazet implemented Big TCP that allowed bigger TSO/GRO packet sizes
> > for IPv6 traffic. See patch series:
> > 'commit 89527be8d8d6 ("net: add IFLA_TSO_{MAX_SIZE|SEGS} attributes")'
> >
> > This reduces the number of packets traversing the networking stack and
> > should usually improves performance. However, it also inserts a
> > temporary Hop-by-hop IPv6 extension header.
> >
> > Using the HBH header removal method in the previous path, the extra header
> > be removed in bnxt drivers to allow it to send big TCP packets (bigger
> > TSO packets) as well.
> >
> > Tested:
> > Compiled locally
> >
> > To further test functional correctness, update the GSO/GRO limit on the
> > physical NIC:
> >
> > ip link set eth0 gso_max_size 181000
> > ip link set eth0 gro_max_size 181000
> >
> > Note that if there are bonding or ipvan devices on top of the physical
> > NIC, their GSO sizes need to be updated as well.
> >
> > Then, IPv6/TCP packets with sizes larger than 64k can be observed.
> >
> > Signed-off-by: Coco Li <lixiaoyan@google.com>

I've confirmed with our hardware team that this is supported by our
chips, and I've tested it up to gso_max_size of 524280.  Thanks.

Tested-by: Michael Chan <michael.chan@broadcom.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]

Re: [RFC net-next v2 2/2] bnxt: Use generic HBH removal helper in tx path

[Coco Li]

Great, thanks Michael! Will send a new version of the patch.

Best,
Coco Li


On Tue, Nov 29, 2022 at 12:59 AM Michael Chan <michael.chan@broadcom.com> wrote:
>
> On Wed, Nov 23, 2022 at 11:42 AM Michael Chan <michael.chan@broadcom.com> wrote:
> >
> > On Wed, Nov 23, 2022 at 11:16 AM Coco Li <lixiaoyan@google.com> wrote:
> > >
> > > Eric Dumazet implemented Big TCP that allowed bigger TSO/GRO packet sizes
> > > for IPv6 traffic. See patch series:
> > > 'commit 89527be8d8d6 ("net: add IFLA_TSO_{MAX_SIZE|SEGS} attributes")'
> > >
> > > This reduces the number of packets traversing the networking stack and
> > > should usually improves performance. However, it also inserts a
> > > temporary Hop-by-hop IPv6 extension header.
> > >
> > > Using the HBH header removal method in the previous path, the extra header
> > > be removed in bnxt drivers to allow it to send big TCP packets (bigger
> > > TSO packets) as well.
> > >
> > > Tested:
> > > Compiled locally
> > >
> > > To further test functional correctness, update the GSO/GRO limit on the
> > > physical NIC:
> > >
> > > ip link set eth0 gso_max_size 181000
> > > ip link set eth0 gro_max_size 181000
> > >
> > > Note that if there are bonding or ipvan devices on top of the physical
> > > NIC, their GSO sizes need to be updated as well.
> > >
> > > Then, IPv6/TCP packets with sizes larger than 64k can be observed.
> > >
> > > Signed-off-by: Coco Li <lixiaoyan@google.com>
>
> I've confirmed with our hardware team that this is supported by our
> chips, and I've tested it up to gso_max_size of 524280.  Thanks.
>
> Tested-by: Michael Chan <michael.chan@broadcom.com>
> Reviewed-by: Michael Chan <michael.chan@broadcom.com>

Re: [RFC net-next v2 2/2] bnxt: Use generic HBH removal helper in tx path

[Michael Chan]

[-- Attachment #1: Type: text/plain, Size: 431 bytes --]

On Tue, Nov 29, 2022 at 9:29 AM Coco Li <lixiaoyan@google.com> wrote:
>
> Great, thanks Michael! Will send a new version of the patch.
>

I just realized one thing.  In bnxt_features_check() ->
bnxt_exthdr_check(), we check to make sure every extension header is
supported by the chip and the number of headers does not exceed what
the chip can support.  So to be more complete, I think we need to
exclude this from being counted.

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4209 bytes --]

Re: [PATCH net-next v2 1/2] IPv6/GRO: generic helper to remove temporary HBH/jumbo header in driver

[Paolo Abeni]

Hello,

Only a couple of minor things below, reporting them as this is still a
RFC, right ? ;)

On Wed, 2022-11-23 at 11:16 -0800, Coco Li wrote:
> IPv6/TCP and GRO stacks can build big TCP packets with an added
> temporary Hop By Hop header.
> 
> Is GSO is not involved, then the temporary header needs to be removed in
> the driver. This patch provides a generic helper for drivers that need
> to modify their headers in place.
> 
> Signed-off-by: Coco Li <lixiaoyan@google.com>
> ---
>  include/net/ipv6.h     | 36 ++++++++++++++++++++++++++++++++++++
>  net/ipv6/ip6_offload.c | 26 ++++----------------------
>  2 files changed, 40 insertions(+), 22 deletions(-)
> 
> diff --git a/include/net/ipv6.h b/include/net/ipv6.h
> index d383c895592a..c5a1daaf5056 100644
> --- a/include/net/ipv6.h
> +++ b/include/net/ipv6.h
> @@ -500,6 +500,42 @@ static inline int ipv6_has_hopopt_jumbo(const struct sk_buff *skb)
>  	return jhdr->nexthdr;
>  }
>  
> +/* Return 0 if HBH header is successfully removed
> + * Or if HBH removal is unnecessary (packet is not big TCP)
> + * Return error to indicate dropping the packet
> + */
> +static inline int ipv6_hopopt_jumbo_remove(struct sk_buff *skb)
> +{
> +	const int hophdr_len = sizeof(struct hop_jumbo_hdr);
> +	int nexthdr = ipv6_has_hopopt_jumbo(skb);
> +	struct ipv6hdr *h6;
> +	int err = 0;
> +
> +	if (!nexthdr)
> +		return err;

You can help a bit the compiler avoiding err initialization:

	int err;

	if (!nexthdr)
		return 0;

> +
> +	err = skb_cow_head(skb, 0);
> +	if (err)
> +		return err;
> +
> +	/* Remove the HBH header.
> +	 * Layout: [Ethernet header][IPv6 header][HBH][L4 Header]
> +	 */
> +	memmove(skb_mac_header(skb) + hophdr_len, skb_mac_header(skb),
> +		skb_network_header(skb) - skb_mac_header(skb) +

The have could be:

		skb_mac_header_len(skb)

which is IMHO a little more clear.

Thanks!

Paolo

Re: [PATCH net-next v2 1/2] IPv6/GRO: generic helper to remove temporary HBH/jumbo header in driver

[Alexander Lobakin]

From: Paolo Abeni <pabeni@redhat.com>
Date: Tue, 29 Nov 2022 10:33:59 +0100

> Hello,
> 
> Only a couple of minor things below, reporting them as this is still a
> RFC, right ? ;)
> 
> On Wed, 2022-11-23 at 11:16 -0800, Coco Li wrote:

[...]

> > +static inline int ipv6_hopopt_jumbo_remove(struct sk_buff *skb)

I thinks it's relatively small and inlineable enough to not make it
an external function, right? I'd keep it inline just how the author
does it, the compiler then will decide.

> > +{
> > +	const int hophdr_len = sizeof(struct hop_jumbo_hdr);
> > +	int nexthdr = ipv6_has_hopopt_jumbo(skb);
> > +	struct ipv6hdr *h6;
> > +	int err = 0;
> > +
> > +	if (!nexthdr)
> > +		return err;
> 
> You can help a bit the compiler avoiding err initialization:
> 
> 	int err;
> 
> 	if (!nexthdr)
> 		return 0;

Same with the end of the function, @err is unused after
skb_cow_head() and always equal 0, so the end return could be just
`return 0`.

> 
> > +
> > +	err = skb_cow_head(skb, 0);
> > +	if (err)
> > +		return err;
> > +
> > +	/* Remove the HBH header.
> > +	 * Layout: [Ethernet header][IPv6 header][HBH][L4 Header]
> > +	 */
> > +	memmove(skb_mac_header(skb) + hophdr_len, skb_mac_header(skb),
> > +		skb_network_header(skb) - skb_mac_header(skb) +
> 
> The have could be:
> 
> 		skb_mac_header_len(skb)
> 
> which is IMHO a little more clear.
> 
> Thanks!
> 
> Paolo

Thanks,
Olek