Message ID | 1574338995-14657-1-git-send-email-martinvarghesenokia@gmail.com |
---|---|
State | Changes Requested |
Delegated to: | David Miller |
Headers | show |
Series | [net-next] Enhanced skb_mpls_pop to update ethertype of the packet in all the cases when an ethernet header is present is the packet. | expand |
On Thu, Nov 21, 2019 at 4:23 AM Martin Varghese <martinvarghesenokia@gmail.com> wrote: > > From: Martin Varghese <martin.varghese@nokia.com> > > The skb_mpls_pop was not updating ethertype of an ethernet packet if the > packet was originally received from a non ARPHRD_ETHER device. > > In the below OVS data path flow, since the device corresponding to port 7 > is an l3 device (ARPHRD_NONE) the skb_mpls_pop function does not update > the ethertype of the packet even though the previous push_eth action had > added an ethernet header to the packet. > > recirc_id(0),in_port(7),eth_type(0x8847), > mpls(label=12/0xfffff,tc=0/0,ttl=0/0x0,bos=1/1), > actions:push_eth(src=00:00:00:00:00:00,dst=00:00:00:00:00:00), > pop_mpls(eth_type=0x800),4 > > Signed-off-by: Martin Varghese <martin.varghese@nokia.com> > --- > include/linux/skbuff.h | 3 ++- > net/core/skbuff.c | 8 +++++--- > net/openvswitch/actions.c | 2 +- > net/sched/act_mpls.c | 2 +- > 4 files changed, 9 insertions(+), 6 deletions(-) > ... > diff --git a/net/core/skbuff.c b/net/core/skbuff.c > index 867e61d..8ac377d 100644 > --- a/net/core/skbuff.c > +++ b/net/core/skbuff.c > @@ -5529,12 +5529,14 @@ int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto, > * @skb: buffer > * @next_proto: ethertype of header after popped MPLS header > * @mac_len: length of the MAC header > - * > + * @ethernet: flag to indicate if ethernet header is present in packet > + * ignored for device type ARPHRD_ETHER > * Expects skb->data at mac header. > * > * Returns 0 on success, -errno otherwise. > */ > -int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) > +int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len, > + bool ethernet) > { > int err; > > @@ -5553,7 +5555,7 @@ int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) > skb_reset_mac_header(skb); > skb_set_network_header(skb, mac_len); > > - if (skb->dev && skb->dev->type == ARPHRD_ETHER) { > + if ((skb->dev && skb->dev->type == ARPHRD_ETHER) || ethernet) { > struct ethhdr *hdr; Lets move the dev-type check to caller. That would also avoid one more argument to this function. > > /* use mpls_hdr() to get ethertype to account for VLANs. */ > diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c > index 12936c1..9e5d274 100644 > --- a/net/openvswitch/actions.c > +++ b/net/openvswitch/actions.c > @@ -179,7 +179,7 @@ static int pop_mpls(struct sk_buff *skb, struct sw_flow_key *key, > { > int err; > > - err = skb_mpls_pop(skb, ethertype, skb->mac_len); > + err = skb_mpls_pop(skb, ethertype, skb->mac_len, true); > if (err) OVS supports L3 packets, you need to check flow key for type of packet (ovs_key_mac_proto()) under process.
On Thu, Nov 21, 2019 at 06:22:29PM -0800, Pravin Shelar wrote: > On Thu, Nov 21, 2019 at 4:23 AM Martin Varghese > <martinvarghesenokia@gmail.com> wrote: > > > > From: Martin Varghese <martin.varghese@nokia.com> > > > > The skb_mpls_pop was not updating ethertype of an ethernet packet if the > > packet was originally received from a non ARPHRD_ETHER device. > > > > In the below OVS data path flow, since the device corresponding to port 7 > > is an l3 device (ARPHRD_NONE) the skb_mpls_pop function does not update > > the ethertype of the packet even though the previous push_eth action had > > added an ethernet header to the packet. > > > > recirc_id(0),in_port(7),eth_type(0x8847), > > mpls(label=12/0xfffff,tc=0/0,ttl=0/0x0,bos=1/1), > > actions:push_eth(src=00:00:00:00:00:00,dst=00:00:00:00:00:00), > > pop_mpls(eth_type=0x800),4 > > > > Signed-off-by: Martin Varghese <martin.varghese@nokia.com> > > --- > > include/linux/skbuff.h | 3 ++- > > net/core/skbuff.c | 8 +++++--- > > net/openvswitch/actions.c | 2 +- > > net/sched/act_mpls.c | 2 +- > > 4 files changed, 9 insertions(+), 6 deletions(-) > > > ... > > diff --git a/net/core/skbuff.c b/net/core/skbuff.c > > index 867e61d..8ac377d 100644 > > --- a/net/core/skbuff.c > > +++ b/net/core/skbuff.c > > @@ -5529,12 +5529,14 @@ int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto, > > * @skb: buffer > > * @next_proto: ethertype of header after popped MPLS header > > * @mac_len: length of the MAC header > > - * > > + * @ethernet: flag to indicate if ethernet header is present in packet > > + * ignored for device type ARPHRD_ETHER > > * Expects skb->data at mac header. > > * > > * Returns 0 on success, -errno otherwise. > > */ > > -int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) > > +int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len, > > + bool ethernet) > > { > > int err; > > > > @@ -5553,7 +5555,7 @@ int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) > > skb_reset_mac_header(skb); > > skb_set_network_header(skb, mac_len); > > > > - if (skb->dev && skb->dev->type == ARPHRD_ETHER) { > > + if ((skb->dev && skb->dev->type == ARPHRD_ETHER) || ethernet) { > > struct ethhdr *hdr; > Lets move the dev-type check to caller. That would also avoid one more > argument to this function. To have only the ethernet flag check in the function like below ? If (ethernet) { /*pseudo*/ Update ethertype } And pass the flag to the function considering the device type Fo example in case of tc. if (skb_mpls_pop(skb, p->tcfm_proto, mac_len, (skb->dev && skb->dev->type == ARPHRD_ETHER))). But how do we avoid an argument here ? I am missing anything ? > > > > > /* use mpls_hdr() to get ethertype to account for VLANs. */ > > diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c > > index 12936c1..9e5d274 100644 > > --- a/net/openvswitch/actions.c > > +++ b/net/openvswitch/actions.c > > @@ -179,7 +179,7 @@ static int pop_mpls(struct sk_buff *skb, struct sw_flow_key *key, > > { > > int err; > > > > - err = skb_mpls_pop(skb, ethertype, skb->mac_len); > > + err = skb_mpls_pop(skb, ethertype, skb->mac_len, true); > > if (err) > OVS supports L3 packets, you need to check flow key for type of packet > (ovs_key_mac_proto()) under process. Yes I missed that. err = skb_mpls_pop(skb, ethertype, skb->mac_len, ovs_key_mac_proto())); ? or err = skb_mpls_pop(skb, ethertype, skb->mac_len, key->mac_proto = MAC_PROTO_ETHERNET) I assume both acheives the same Thanks for reviewing Martin
On Thu, Nov 21, 2019 at 9:13 PM Martin Varghese <martinvarghesenokia@gmail.com> wrote: > > On Thu, Nov 21, 2019 at 06:22:29PM -0800, Pravin Shelar wrote: > > On Thu, Nov 21, 2019 at 4:23 AM Martin Varghese > > <martinvarghesenokia@gmail.com> wrote: > > > > > > From: Martin Varghese <martin.varghese@nokia.com> > > > > > > The skb_mpls_pop was not updating ethertype of an ethernet packet if the > > > packet was originally received from a non ARPHRD_ETHER device. > > > > > > In the below OVS data path flow, since the device corresponding to port 7 > > > is an l3 device (ARPHRD_NONE) the skb_mpls_pop function does not update > > > the ethertype of the packet even though the previous push_eth action had > > > added an ethernet header to the packet. > > > > > > recirc_id(0),in_port(7),eth_type(0x8847), > > > mpls(label=12/0xfffff,tc=0/0,ttl=0/0x0,bos=1/1), > > > actions:push_eth(src=00:00:00:00:00:00,dst=00:00:00:00:00:00), > > > pop_mpls(eth_type=0x800),4 > > > > > > Signed-off-by: Martin Varghese <martin.varghese@nokia.com> > > > --- > > > include/linux/skbuff.h | 3 ++- > > > net/core/skbuff.c | 8 +++++--- > > > net/openvswitch/actions.c | 2 +- > > > net/sched/act_mpls.c | 2 +- > > > 4 files changed, 9 insertions(+), 6 deletions(-) > > > > > ... > > > diff --git a/net/core/skbuff.c b/net/core/skbuff.c > > > index 867e61d..8ac377d 100644 > > > --- a/net/core/skbuff.c > > > +++ b/net/core/skbuff.c > > > @@ -5529,12 +5529,14 @@ int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto, > > > * @skb: buffer > > > * @next_proto: ethertype of header after popped MPLS header > > > * @mac_len: length of the MAC header > > > - * > > > + * @ethernet: flag to indicate if ethernet header is present in packet > > > + * ignored for device type ARPHRD_ETHER > > > * Expects skb->data at mac header. > > > * > > > * Returns 0 on success, -errno otherwise. > > > */ > > > -int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) > > > +int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len, > > > + bool ethernet) > > > { > > > int err; > > > > > > @@ -5553,7 +5555,7 @@ int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) > > > skb_reset_mac_header(skb); > > > skb_set_network_header(skb, mac_len); > > > > > > - if (skb->dev && skb->dev->type == ARPHRD_ETHER) { > > > + if ((skb->dev && skb->dev->type == ARPHRD_ETHER) || ethernet) { > > > struct ethhdr *hdr; > > Lets move the dev-type check to caller. That would also avoid one more > > argument to this function. > > > To have only the ethernet flag check in the function like below ? > If (ethernet) { > /*pseudo*/ Update ethertype > } > And pass the flag to the function considering the device type > Fo example in case of tc. > > if (skb_mpls_pop(skb, p->tcfm_proto, mac_len, > (skb->dev && skb->dev->type == ARPHRD_ETHER))). > > > But how do we avoid an argument here ? I am missing anything ? Right, I wanted to say new predicate can be eliminated. > > > > > > > > /* use mpls_hdr() to get ethertype to account for VLANs. */ > > > diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c > > > index 12936c1..9e5d274 100644 > > > --- a/net/openvswitch/actions.c > > > +++ b/net/openvswitch/actions.c > > > @@ -179,7 +179,7 @@ static int pop_mpls(struct sk_buff *skb, struct sw_flow_key *key, > > > { > > > int err; > > > > > > - err = skb_mpls_pop(skb, ethertype, skb->mac_len); > > > + err = skb_mpls_pop(skb, ethertype, skb->mac_len, true); > > > if (err) > > OVS supports L3 packets, you need to check flow key for type of packet > > (ovs_key_mac_proto()) under process. > > > Yes I missed that. > err = skb_mpls_pop(skb, ethertype, skb->mac_len, ovs_key_mac_proto())); ? > or > err = skb_mpls_pop(skb, ethertype, skb->mac_len, key->mac_proto = MAC_PROTO_ETHERNET) I like the second which is more explicit.
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index dfe02b6..70204b9 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -3530,7 +3530,8 @@ int skb_zerocopy(struct sk_buff *to, struct sk_buff *from, int skb_vlan_push(struct sk_buff *skb, __be16 vlan_proto, u16 vlan_tci); int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto, int mac_len); -int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len); +int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len, + bool ethernet); int skb_mpls_update_lse(struct sk_buff *skb, __be32 mpls_lse); int skb_mpls_dec_ttl(struct sk_buff *skb); struct sk_buff *pskb_extract(struct sk_buff *skb, int off, int to_copy, diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 867e61d..8ac377d 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -5529,12 +5529,14 @@ int skb_mpls_push(struct sk_buff *skb, __be32 mpls_lse, __be16 mpls_proto, * @skb: buffer * @next_proto: ethertype of header after popped MPLS header * @mac_len: length of the MAC header - * + * @ethernet: flag to indicate if ethernet header is present in packet + * ignored for device type ARPHRD_ETHER * Expects skb->data at mac header. * * Returns 0 on success, -errno otherwise. */ -int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) +int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len, + bool ethernet) { int err; @@ -5553,7 +5555,7 @@ int skb_mpls_pop(struct sk_buff *skb, __be16 next_proto, int mac_len) skb_reset_mac_header(skb); skb_set_network_header(skb, mac_len); - if (skb->dev && skb->dev->type == ARPHRD_ETHER) { + if ((skb->dev && skb->dev->type == ARPHRD_ETHER) || ethernet) { struct ethhdr *hdr; /* use mpls_hdr() to get ethertype to account for VLANs. */ diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c index 12936c1..9e5d274 100644 --- a/net/openvswitch/actions.c +++ b/net/openvswitch/actions.c @@ -179,7 +179,7 @@ static int pop_mpls(struct sk_buff *skb, struct sw_flow_key *key, { int err; - err = skb_mpls_pop(skb, ethertype, skb->mac_len); + err = skb_mpls_pop(skb, ethertype, skb->mac_len, true); if (err) return err; diff --git a/net/sched/act_mpls.c b/net/sched/act_mpls.c index 4d8c822..5fa39fc 100644 --- a/net/sched/act_mpls.c +++ b/net/sched/act_mpls.c @@ -76,7 +76,7 @@ static int tcf_mpls_act(struct sk_buff *skb, const struct tc_action *a, switch (p->tcfm_action) { case TCA_MPLS_ACT_POP: - if (skb_mpls_pop(skb, p->tcfm_proto, mac_len)) + if (skb_mpls_pop(skb, p->tcfm_proto, mac_len, false)) goto drop; break; case TCA_MPLS_ACT_PUSH: