[ovs-dev] [PATCH v2 1/2] nsh: rework NSH netlink keys and actions
Yang, Yi
yi.y.yang at intel.com
Sat Aug 19 02:00:03 UTC 2017
On Fri, Aug 18, 2017 at 11:02:35PM +0800, Jan Scheurich wrote:
> > diff --git a/datapath/linux/compat/include/linux/openvswitch.h
> > b/datapath/linux/compat/include/linux/openvswitch.h
> > index bc6c94b..d7f9029 100644
> > --- a/datapath/linux/compat/include/linux/openvswitch.h
> > +++ b/datapath/linux/compat/include/linux/openvswitch.h
>
> > +#define NSH_MD1_CONTEXT_SIZE 4
> [Jan] This shouldn’t be needed here and replaced by a macro defined in nsh.h
Jan, this is indeed needed here, context[4] won't work because ovs
script will convert array[4] in
datapath/linux/compat/include/linux/openvswitch.h
To a special type, so we have to hack this. In addition,
datapath/linux/compat/include/linux/openvswitch.h shouldn't include
nsh.h anyway.
>
> > +#define OVS_PUSH_NSH_MAX_MD_LEN 248
> [Jan] This shouldn’t be needed here and replaced by a macro defined in nsh.h
>
ok, I'll remove it from here.
> > diff --git a/include/openvswitch/nsh.h b/include/openvswitch/nsh.h
> > index f4ccadc..ed9054e 100644
> > --- a/include/openvswitch/nsh.h
> > +++ b/include/openvswitch/nsh.h
>
>
> > struct nsh_md2_tlv {
> > @@ -72,6 +72,8 @@ struct nsh_hdr {
> > };
> > };
> >
> > +#define NSH_M_TYPE2_MAX_LEN 256
> [Jan] This is a duplicate definition (see below)
>
Noted, will remove it.
>
> > +/* NSH MD Type 2 header maximum Length. */
> > +#define NSH_M_TYPE2_MAX_LEN 256
> [Jan] I suggest to rename this to NSH_HEADER_MAX_LEN as it is generic limit for
> any MD type, not only MD2, resulting from the fact that the length in 4 byte
> words is encoded in 6 bits.
>
> Add a macro NSH_CONTEXT_HDRS_MAX_LEN 248 and use that instead of
> OVS_PUSH_NSH_MAX_MD_LEN.
>
Ok, will use macros you suggest to replace them.
> > static inline uint16_t
> > nsh_hdr_len(const struct nsh_hdr *nsh)
> > {
> > - return 4 * (ntohs(nsh->ver_flags_len) & NSH_LEN_MASK) >>
> > NSH_LEN_SHIFT;
> > + return ((ntohs(nsh->ver_flags_len) & NSH_LEN_MASK) >>
> > NSH_LEN_SHIFT) << 2;
> [Jan] I personally find 4 * X clearer than X << 2 when the purpose really is to
> multiply by 4 rather than to shift left by two bits. The compiler should
> generate similar code with -O2 anyhow.
>
The issue is it is incorrect, Eric Garver told me operator ">>" will be
done before operator "*", it can run correctly just because
NSH_LEN_SHIFT is zero, so the new one is right.
"* 4" and "<< 2" have same readability, we mustn't argue such issue
here.
> > diff --git a/include/openvswitch/packets.h b/include/openvswitch/packets.h
> > index be91e02..5ee3099 100644
> > --- a/include/openvswitch/packets.h
> > +++ b/include/openvswitch/packets.h
> > @@ -84,7 +84,16 @@ struct flow_nsh {
> > uint8_t np;
> > uint8_t si;
> > ovs_be32 spi;
> > - ovs_be32 c[4];
> > + ovs_be32 context[4];
> > +};
> > +
> > +struct ovs_key_nsh {
> > + uint8_t flags;
> > + uint8_t mdtype;
> > + uint8_t np;
> > + uint8_t pad;
> > + ovs_be32 path_hdr;
> > + ovs_be32 context[4];
> > };
>
> [Jan] I see in the second patch that you use this new struct as replacement for
> struct flow_nsh in struct flow. The main difference being that the new struct
> combines spi and si fields into the path_hdr. That is of course possible
> provided that the necessary conversions are made where needed, but I doubt that
> it improves readability in the code overall.
>
> Was the main motivation to use the same data type both in struct flow as well
> as in the datapath for set(nsh) operations? Or did you want to prevent struct
> flow_nsh spilling over into 25 bytes as the spare octet used for the si field
> is now needed for the new ttl field?
>
> I believe in the datapath the NSH key attributes could be parsed into an NSH
> header struct directly, no need to have an struct ovs_key_nsh.
Jan, struct flow_nsh must be 64 bits aligned, new key ttl broke this.
struct ovs_key_nsh is the stuff we transfer by netlink, kernel data path
struct sw_flow_key also uses it. struct ovs_key_nsh is very appropriate
to do this in struct flow, it is very natural way to do right thing.
A field corresponding to multiple keys is noraml if you check ovs code,
mpls_lse is a typical case.
> > @@ -572,11 +574,12 @@ parse_nsh(const void **datap, size_t *sizep,
> > struct flow_nsh *key)
> > switch (key->mdtype) {
> > case NSH_M_TYPE1:
> > for (size_t i = 0; i < 4; i++) {
> > - key->c[i] = get_16aligned_be32(&nsh->md1.c[i]);
> > + key->context[i] = get_16aligned_be32(&nsh->md1.context[i]);
> > }
> > break;
> > case NSH_M_TYPE2:
> > - /* Don't support MD type 2 yet, so return false */
> > + /* Don't support MD type 2 metedata parsing yet */
> > + break;
> > default:
> > return false;
>
> [Jan] There is no need to return false here. If the md_type is unknown we don’t
> parse the context headers, but we can still continue processing the NSH packet
> as the total length is all we need.
But if md_type is unknow, it is wrong, why do we still continue it?
>
> > }
> > @@ -885,9 +888,7 @@ miniflow_extract(struct dp_packet *packet, struct
> > miniflow *dst)
> > sizeof(uint64_t));
> > }
> > else if (nsh.mdtype == NSH_M_TYPE2) {
> > - /* parse_nsh has stopped it from arriving here for
> > - * MD type 2, will add MD type 2 support code here later
> > - */
> > + /* Can't parse MD type 2 metedata yet */
> > }
>
> [Jan] If parse_nsh returns true, we always need to push the flow_nsh to
> miniflow. Please remove the extra check on MD1.
Ok, will remove extra check.
>
> > diff --git a/lib/odp-execute.c b/lib/odp-execute.c
> > index 5f4d23a..0a9a535 100644
> > --- a/lib/odp-execute.c
> > +++ b/lib/odp-execute.c
> > @@ -273,19 +273,22 @@ odp_set_nd(struct dp_packet *packet, const
> > struct ovs_key_nd *key,
> > /* Set the NSH header. Assumes the NSH header is present and matches
> > the
> > * MD format of the key. The slow path must take case of that. */
> > static void
> > -odp_set_nsh(struct dp_packet *packet, const struct ovs_key_nsh *key,
> > - const struct ovs_key_nsh *mask)
> > +odp_set_nsh(struct dp_packet *packet, const struct flow_nsh *key,
> > + const struct flow_nsh *mask)
> > {
> > struct nsh_hdr *nsh = dp_packet_l3(packet);
> > + ovs_be32 path_hdr;
> >
> > if (!mask) {
> > nsh->ver_flags_len = htons(key->flags << NSH_FLAGS_SHIFT) |
> > (nsh->ver_flags_len & ~htons(NSH_FLAGS_MASK));
> > - put_16aligned_be32(&nsh->path_hdr, key->path_hdr);
> > + path_hdr = htonl((ntohl(key->spi) << NSH_SPI_SHIFT) |
> > + key->si);
> > + put_16aligned_be32(&nsh->path_hdr, path_hdr);
> > switch (nsh->md_type) {
> > case NSH_M_TYPE1:
> > for (int i = 0; i < 4; i++) {
> > - put_16aligned_be32(&nsh->md1.c[i], key->c[i]);
> > + put_16aligned_be32(&nsh->md1.context[i], key->context
> [i]);
> > }
> > break;
> > case NSH_M_TYPE2:
> > @@ -300,16 +303,24 @@ odp_set_nsh(struct dp_packet *packet, const
> > struct ovs_key_nsh *key,
> > nsh->ver_flags_len = htons(flags << NSH_FLAGS_SHIFT) |
> > (nsh->ver_flags_len & ~htons(NSH_FLAGS_MASK));
> >
> > - ovs_be32 path_hdr = get_16aligned_be32(&nsh->path_hdr);
> > - path_hdr = key->path_hdr | (path_hdr & ~mask->path_hdr);
> > + path_hdr = get_16aligned_be32(&nsh->path_hdr);
> > + uint32_t spi = (ntohl(path_hdr) & NSH_SPI_MASK) >> NSH_SPI_SHIFT;
> > + uint8_t si = (ntohl(path_hdr) & NSH_SI_MASK) >> NSH_SI_SHIFT;
> > + uint32_t spi_mask = ntohl(mask->spi);
> > + if (spi_mask == 0x00ffffff) {
> > + spi_mask = UINT32_MAX;
> > + }
> > + spi = ntohl(key->spi) | (spi & ~spi_mask);
> > + si = key->si | (si & ~mask->si);
> > + path_hdr = htonl((spi << NSH_SPI_SHIFT) | si);
> > put_16aligned_be32(&nsh->path_hdr, path_hdr);
> > switch (nsh->md_type) {
> > case NSH_M_TYPE1:
> > for (int i = 0; i < 4; i++) {
> > - ovs_be32 p = get_16aligned_be32(&nsh->md1.c[i]);
> > - ovs_be32 k = key->c[i];
> > - ovs_be32 m = mask->c[i];
> > - put_16aligned_be32(&nsh->md1.c[i], k | (p & ~m));
> > + ovs_be32 p = get_16aligned_be32(&nsh->md1.context[i]);
> > + ovs_be32 k = key->context[i];
> > + ovs_be32 m = mask->context[i];
> > + put_16aligned_be32(&nsh->md1.context[i], k | (p & ~m));
> > }
> > break;
> > case NSH_M_TYPE2:
>
> [Jan] The whole function could be simplified a lot when passing in a struct
> nsh_hdr instead of struct ovs_key_nsh
Jan, but we need to handle mask use case. struct nsh_hdr can't provide
mask, isn't it?
>
> > @@ -345,9 +356,12 @@ odp_execute_set_action(struct dp_packet
> > *packet, const struct nlattr *a)
> > odp_eth_set_addrs(packet, nl_attr_get(a), NULL);
> > break;
> >
> > - case OVS_KEY_ATTR_NSH:
> > - odp_set_nsh(packet, nl_attr_get(a), NULL);
> > + case OVS_KEY_ATTR_NSH: {
> > + struct flow_nsh nsh;
> > + odp_nsh_key_from_attr(a, &nsh);
>
> [Jan] Replace this with an odp_nsh_hdr_from_attr that directly translates into
> struct nsh_hdr
Same as above.
>
> > + odp_set_nsh(packet, &nsh, NULL);
> > break;
> > + }
> >
> > case OVS_KEY_ATTR_IPV4:
> > ipv4_key = nl_attr_get_unspec(a, sizeof(struct ovs_key_ipv4));
> > @@ -473,10 +487,24 @@ odp_execute_masked_set_action(struct
> > dp_packet *packet,
> > get_mask(a, struct ovs_key_ethernet));
> > break;
> >
> > - case OVS_KEY_ATTR_NSH:
> > - odp_set_nsh(packet, nl_attr_get(a),
> > - get_mask(a, struct ovs_key_nsh));
> > + case OVS_KEY_ATTR_NSH: {
> > + struct flow_nsh nsh, nsh_mask;
> > + size_t size = nl_attr_get_size(a) / 2;
> > +
> > + struct nlattr attr[1 + size / sizeof(struct nlattr) + 1];
> > + struct nlattr mask[1 + size / sizeof(struct nlattr) + 1];
> > +
> > + mask->nla_type = attr->nla_type = nl_attr_type(a);
> > + mask->nla_len = attr->nla_len = NLA_HDRLEN + size;
> > + memcpy(attr + 1, (char *)(a + 1), size);
> > + memcpy(mask + 1, (char *)(a + 1) + size, size);
>
> [Jan] This code looks very obscure. Can you replace this with something more
> readable and safe?
Please provide a readable and safe code piece if you have, I'm not sure
where is not readable and not safe.
>
> > + odp_nsh_key_from_attr(attr, &nsh);
> > + odp_nsh_key_from_attr(mask, &nsh_mask);
> > + odp_set_nsh(packet, &nsh, &nsh_mask);
> > +
> > break;
> > + }
>
>
> > @@ -818,18 +846,21 @@ odp_execute_actions(void *dp, struct
>
> > + case OVS_ACTION_ATTR_PUSH_NSH: {
> > + uint8_t buffer[256];
> > + struct nsh_hdr *nsh_hdr = ALIGNED_CAST(struct nsh_hdr *,
> > buffer);
> > + const struct nsh_hdr *nsh_hdr_src = nsh_hdr;
> > + odp_nsh_hdr_from_attr(nl_attr_get(a), nsh_hdr);
>
> [Jan] Calling this function with nl_attr_get(a) as first argument seems to
> conflict with the implementation of function odp_nsh_hdr_from_attr(), which
> uses NL_NESTED_FOR_EACH() to iterate over the nested attributes. In my tests
> that caused crashes. Please double-check.
nl_attr_get(a) just returns nested OVS_KEY_ATTR_NSH attribute,
odp_nsh_hdr_from_attr NL_NESTED_FOR_EACH nested OVS_KEY_ATTR_NSH, isn't
it?
Please remember, here the attribute is OVS_ACTION_ATTR_PUSH_NSH, not
OVS_KEY_ATTR_NSH, nl_attr_get(a) will get data of
OVS_ACTION_ATTR_PUSH_NSH.
I have verified it. Maybe you're using a wrong ovs version.
>
> I have to stop here for today. I will hopefully continue review next week.
>
> BR, Jan
>
> > diff --git a/lib/odp-util.c b/lib/odp-util.c
> > index 4f1499e..3e4403a 100644
> > --- a/lib/odp-util.c
> > +++ b/lib/odp-util.c
> > @@ -129,8 +129,8 @@ odp_action_len(uint16_t type)
> > case OVS_ACTION_ATTR_PUSH_ETH: return sizeof(struct
> > ovs_action_push_eth);
> > case OVS_ACTION_ATTR_POP_ETH: return 0;
> > case OVS_ACTION_ATTR_CLONE: return ATTR_LEN_VARIABLE;
> > - case OVS_ACTION_ATTR_ENCAP_NSH: return ATTR_LEN_VARIABLE;
> > - case OVS_ACTION_ATTR_DECAP_NSH: return 0;
> > + case OVS_ACTION_ATTR_PUSH_NSH: return ATTR_LEN_VARIABLE;
> > + case OVS_ACTION_ATTR_POP_NSH: return 0;
> >
> > case OVS_ACTION_ATTR_UNSPEC:
> > case __OVS_ACTION_ATTR_MAX:
> > @@ -264,7 +264,7 @@ format_nsh_key(struct ds *ds, const struct
> > ovs_key_nsh *key)
> > switch (key->mdtype) {
> > case NSH_M_TYPE1:
> > for (int i = 0; i < 4; i++) {
> > - ds_put_format(ds, ",c%d=0x%x", i + 1, ntohl(key->c[i]));
> > + ds_put_format(ds, ",c%d=0x%x", i + 1, ntohl(key->context
> [i]));
> > }
> > break;
> > case NSH_M_TYPE2:
> > @@ -334,41 +334,50 @@ format_nsh_key_mask(struct ds *ds, const struct
> > ovs_key_nsh *key,
> > format_uint8_masked(ds, &first, "np", key->np, mask->np);
> > format_be32_masked(ds, &first, "spi", htonl(spi), htonl(spi_mask));
> > format_uint8_masked(ds, &first, "si", si, si_mask);
> > - format_be32_masked(ds, &first, "c1", key->c[0], mask->c[0]);
> > - format_be32_masked(ds, &first, "c2", key->c[1], mask->c[1]);
> > - format_be32_masked(ds, &first, "c3", key->c[2], mask->c[2]);
> > - format_be32_masked(ds, &first, "c4", key->c[3], mask->c[3]);
> > + format_be32_masked(ds, &first, "c1", key->context[0],
> > + mask->context[0]);
> > + format_be32_masked(ds, &first, "c2", key->context[1],
> > + mask->context[1]);
> > + format_be32_masked(ds, &first, "c3", key->context[2],
> > + mask->context[2]);
> > + format_be32_masked(ds, &first, "c4", key->context[3],
> > + mask->context[3]);
> > }
> > }
> >
> > static void
> > -format_odp_encap_nsh_action(struct ds *ds,
> > - const struct ovs_action_encap_nsh *encap_nsh)
> > +format_odp_push_nsh_action(struct ds *ds,
> > + const struct nsh_hdr *nsh_hdr)
> > {
> > - uint32_t path_hdr = ntohl(encap_nsh->path_hdr);
> > + size_t mdlen = (((ntohs(nsh_hdr->ver_flags_len) & NSH_LEN_MASK)
> > + >> NSH_LEN_SHIFT) << 2) - NSH_BASE_HDR_LEN;
> > + uint32_t path_hdr = ntohl(get_16aligned_be32(&nsh_hdr->path_hdr));
> > uint32_t spi = (path_hdr & NSH_SPI_MASK) >> NSH_SPI_SHIFT;
> > uint8_t si = (path_hdr & NSH_SI_MASK) >> NSH_SI_SHIFT;
> > + uint8_t flags = (ntohs(nsh_hdr->ver_flags_len) & NSH_FLAGS_MASK)
> > + >> NSH_FLAGS_SHIFT;
> >
> > - ds_put_cstr(ds, "encap_nsh(");
> > - ds_put_format(ds, "flags=%d", encap_nsh->flags);
> > - ds_put_format(ds, ",mdtype=%d", encap_nsh->mdtype);
> > - ds_put_format(ds, ",np=%d", encap_nsh->np);
> > + ds_put_cstr(ds, "push_nsh(");
> > + ds_put_format(ds, "flags=%d", flags);
> > + ds_put_format(ds, ",mdtype=%d", nsh_hdr->md_type);
> > + ds_put_format(ds, ",np=%d", nsh_hdr->next_proto);
> > ds_put_format(ds, ",spi=0x%x", spi);
> > ds_put_format(ds, ",si=%d", si);
> > - switch (encap_nsh->mdtype) {
> > + switch (nsh_hdr->md_type) {
> > case NSH_M_TYPE1: {
> > - struct nsh_md1_ctx *md1_ctx =
> > - ALIGNED_CAST(struct nsh_md1_ctx *, encap_nsh->metadata);
> > + const struct nsh_md1_ctx *md1_ctx = NSH_MD1_CTX(nsh_hdr);
> > for (int i = 0; i < 4; i++) {
> > ds_put_format(ds, ",c%d=0x%x", i + 1,
> > - ntohl(get_16aligned_be32(&md1_ctx->c[i])));
> > + ntohl(get_16aligned_be32(&md1_ctx->context[i])));
> > }
> > break;
> > }
> > - case NSH_M_TYPE2:
> > + case NSH_M_TYPE2: {
> > + const struct nsh_md2_tlv *md2_ctx = NSH_MD2_CTX(nsh_hdr);
> > ds_put_cstr(ds, ",md2=");
> > - ds_put_hex(ds, encap_nsh->metadata, encap_nsh->mdlen);
> > + ds_put_hex(ds, md2_ctx, mdlen);
> > break;
> > + }
> > default:
> > OVS_NOT_REACHED();
> > }
> > @@ -1057,11 +1066,16 @@ format_odp_action(struct ds *ds, const struct
> > nlattr *a,
> > case OVS_ACTION_ATTR_CLONE:
> > format_odp_clone_action(ds, a, portno_names);
> > break;
> > - case OVS_ACTION_ATTR_ENCAP_NSH:
> > - format_odp_encap_nsh_action(ds, nl_attr_get(a));
> > + case OVS_ACTION_ATTR_PUSH_NSH: {
> > + uint8_t buffer[256];
> > + struct nsh_hdr *nsh_hdr = ALIGNED_CAST(struct nsh_hdr *, buffer);
> > + const struct nsh_hdr *nsh_hdr_src = nsh_hdr;
> > + odp_nsh_hdr_from_attr(nl_attr_get(a), nsh_hdr);
> > + format_odp_push_nsh_action(ds, nsh_hdr_src);
> > break;
> > - case OVS_ACTION_ATTR_DECAP_NSH:
> > - ds_put_cstr(ds, "decap_nsh()");
> > + }
> > + case OVS_ACTION_ATTR_POP_NSH:
> > + ds_put_cstr(ds, "pop_nsh()");
> > break;
> > case OVS_ACTION_ATTR_UNSPEC:
> > case __OVS_ACTION_ATTR_MAX:
> > @@ -1780,27 +1794,74 @@ find_end:
> > return s - s_;
> > }
> >
> > +static void
> > +nsh_key_to_attr(struct ofpbuf *buf, const struct flow_nsh *nsh,
> > + uint8_t * metadata, size_t md_size,
> > + bool is_mask)
> > +{
> > + size_t nsh_key_ofs;
> > + struct ovs_nsh_key_base base;
> > + struct ovs_nsh_key_md1 md1;
> > +
> > + base.flags = nsh->flags;
> > + base.mdtype = nsh->mdtype;
> > + base.np = nsh->np;
> > + base.path_hdr = htonl((ntohl(nsh->spi) << NSH_SPI_SHIFT) |
> > + nsh->si);
> > +
> > + nsh_key_ofs = nl_msg_start_nested(buf, OVS_KEY_ATTR_NSH);
> > + nl_msg_put_unspec(buf, OVS_NSH_KEY_ATTR_BASE, &base, sizeof
> > base);
> > +
> > + if (is_mask) {
> > + for (int i = 0; i < 4; i++) {
> > + md1.context[i] = nsh->context[i];
> > + }
> > + nl_msg_put_unspec(buf, OVS_NSH_KEY_ATTR_MD1, &md1, sizeof
> > md1);
> > + } else {
> > + switch (nsh->mdtype) {
> > + case NSH_M_TYPE1:
> > + for (int i = 0; i < 4; i++) {
> > + md1.context[i] = nsh->context[i];
> > + }
> > + nl_msg_put_unspec(buf, OVS_NSH_KEY_ATTR_MD1, &md1, sizeof
> > md1);
> > + break;
> > + case NSH_M_TYPE2:
> > + if (metadata && md_size > 0) {
> > + nl_msg_put_unspec(buf, OVS_NSH_KEY_ATTR_MD2, metadata,
> > + md_size);
> > + }
> > + break;
> > + default:
> > + /* No match support for other MD formats yet. */
> > + break;
> > + }
> > + }
> > + nl_msg_end_nested(buf, nsh_key_ofs);
> > +}
> > +
> > +
> > static int
> > -parse_odp_encap_nsh_action(const char *s, struct ofpbuf *actions)
> > +parse_odp_push_nsh_action(const char *s, struct ofpbuf *actions)
> > {
> > int n = 0;
> > int ret = 0;
> > - struct ovs_action_encap_nsh encap_nsh;
> > - uint32_t spi;
> > - uint8_t si;
> > uint32_t cd;
> > + struct flow_nsh nsh;
> > + uint8_t *metadata = NULL;
> > + uint8_t md_size = 0;
> >
> > - if (!ovs_scan_len(s, &n, "encap_nsh(")) {
> > + if (!ovs_scan_len(s, &n, "push_nsh(")) {
> > ret = -EINVAL;
> > goto out;
> > }
> >
> > /* The default is NSH_M_TYPE1 */
> > - encap_nsh.flags = 0;
> > - encap_nsh.mdtype = NSH_M_TYPE1;
> > - encap_nsh.mdlen = NSH_M_TYPE1_MDLEN;
> > - encap_nsh.path_hdr = htonl(255);
> > - memset(encap_nsh.metadata, 0, NSH_M_TYPE1_MDLEN);
> > + nsh.flags = 0;
> > + nsh.mdtype = NSH_M_TYPE1;
> > + nsh.np = NSH_P_ETHERNET;
> > + nsh.spi = 0;
> > + nsh.si = 255;
> > + memset(nsh.context, 0, NSH_M_TYPE1_MDLEN);
> >
> > for (;;) {
> > n += strspn(s + n, delimiters);
> > @@ -1808,17 +1869,17 @@ parse_odp_encap_nsh_action(const char *s,
> > struct ofpbuf *actions)
> > break;
> > }
> >
> > - if (ovs_scan_len(s, &n, "flags=%"SCNi8, &encap_nsh.flags)) {
> > + if (ovs_scan_len(s, &n, "flags=%"SCNi8, &nsh.flags)) {
> > continue;
> > }
> > - if (ovs_scan_len(s, &n, "mdtype=%"SCNi8, &encap_nsh.mdtype)) {
> > - switch (encap_nsh.mdtype) {
> > + if (ovs_scan_len(s, &n, "mdtype=%"SCNi8, &nsh.mdtype)) {
> > + switch (nsh.mdtype) {
> > case NSH_M_TYPE1:
> > /* This is the default format. */;
> > break;
> > case NSH_M_TYPE2:
> > /* Length will be updated later. */
> > - encap_nsh.mdlen = 0;
> > + md_size = 0;
> > break;
> > default:
> > ret = -EINVAL;
> > @@ -1826,65 +1887,60 @@ parse_odp_encap_nsh_action(const char *s,
> > struct ofpbuf *actions)
> > }
> > continue;
> > }
> > - if (ovs_scan_len(s, &n, "np=%"SCNi8, &encap_nsh.np)) {
> > + if (ovs_scan_len(s, &n, "np=%"SCNi8, &nsh.np)) {
> > continue;
> > }
> > - if (ovs_scan_len(s, &n, "spi=0x%"SCNx32, &spi)) {
> > - encap_nsh.path_hdr =
> > - htonl(((spi << NSH_SPI_SHIFT) & NSH_SPI_MASK) |
> > - (ntohl(encap_nsh.path_hdr) & ~NSH_SPI_MASK));
> > + if (ovs_scan_len(s, &n, "spi=0x%"SCNx32, &nsh.spi)) {
> > + nsh.spi = htonl(nsh.spi);
> > continue;
> > }
> > - if (ovs_scan_len(s, &n, "si=%"SCNi8, &si)) {
> > - encap_nsh.path_hdr =
> > - htonl((si << NSH_SI_SHIFT) |
> > - (ntohl(encap_nsh.path_hdr) & ~NSH_SI_MASK));
> > + if (ovs_scan_len(s, &n, "si=%"SCNi8, &nsh.si)) {
> > continue;
> > }
> > - if (encap_nsh.mdtype == NSH_M_TYPE1) {
> > - struct nsh_md1_ctx *md1 =
> > - ALIGNED_CAST(struct nsh_md1_ctx *, encap_nsh.metadata);
> > + if (nsh.mdtype == NSH_M_TYPE1) {
> > if (ovs_scan_len(s, &n, "c1=0x%"SCNx32, &cd)) {
> > - put_16aligned_be32(&md1->c[0], htonl(cd));
> > + nsh.context[0] = htonl(cd);
> > continue;
> > }
> > if (ovs_scan_len(s, &n, "c2=0x%"SCNx32, &cd)) {
> > - put_16aligned_be32(&md1->c[1], htonl(cd));
> > + nsh.context[1] = htonl(cd);
> > continue;
> > }
> > if (ovs_scan_len(s, &n, "c3=0x%"SCNx32, &cd)) {
> > - put_16aligned_be32(&md1->c[2], htonl(cd));
> > + nsh.context[2] = htonl(cd);
> > continue;
> > }
> > if (ovs_scan_len(s, &n, "c4=0x%"SCNx32, &cd)) {
> > - put_16aligned_be32(&md1->c[3], htonl(cd));
> > + nsh.context[3] = htonl(cd);
> > continue;
> > }
> > }
> > - else if (encap_nsh.mdtype == NSH_M_TYPE2) {
> > + else if (nsh.mdtype == NSH_M_TYPE2) {
> > struct ofpbuf b;
> > char buf[512];
> > size_t mdlen;
> > if (ovs_scan_len(s, &n, "md2=0x%511[0-9a-fA-F]", buf)) {
> > - ofpbuf_use_stub(&b, encap_nsh.metadata,
> > - OVS_ENCAP_NSH_MAX_MD_LEN);
> > + metadata = xmalloc(NSH_M_TYPE2_MAX_LEN - 8);
> > + ofpbuf_use_stub(&b, metadata,
> > + NSH_M_TYPE2_MAX_LEN - 8);
> > ofpbuf_put_hex(&b, buf, &mdlen);
> > - encap_nsh.mdlen = mdlen;
> > + md_size = mdlen;
> > ofpbuf_uninit(&b);
> > }
> > continue;
> > }
> > }
> > out:
> > - if (ret < 0) {
> > - return ret;
> > - } else {
> > - size_t size = offsetof(struct ovs_action_encap_nsh, metadata)
> > - + ROUND_UP(encap_nsh.mdlen, 4);
> > - nl_msg_put_unspec(actions, OVS_ACTION_ATTR_ENCAP_NSH,
> > - &encap_nsh, size);
> > - return n;
> > + if (ret >= 0) {
> > + size_t offset = nl_msg_start_nested(actions,
> > OVS_ACTION_ATTR_PUSH_NSH);
> > + nsh_key_to_attr(actions, &nsh, metadata, md_size, false);
> > + nl_msg_end_nested(actions, offset);
> > + ret = n;
> > }
> > + if (metadata != NULL) {
> > + free(metadata);
> > + }
> > + return ret;
> > }
> >
> > static int
> > @@ -2089,8 +2145,8 @@ parse_odp_action(const char *s, const struct
> > simap *port_names,
> > }
> >
> > {
> > - if (!strncmp(s, "encap_nsh(", 10)) {
> > - int retval = parse_odp_encap_nsh_action(s, actions);
> > + if (!strncmp(s, "push_nsh(", 9)) {
> > + int retval = parse_odp_push_nsh_action(s, actions);
> > if (retval < 0) {
> > return retval;
> > }
> > @@ -2100,8 +2156,8 @@ parse_odp_action(const char *s, const struct
> > simap *port_names,
> >
> > {
> > int n;
> > - if (ovs_scan(s, "decap_nsh()%n", &n)) {
> > - nl_msg_put_flag(actions, OVS_ACTION_ATTR_DECAP_NSH);
> > + if (ovs_scan(s, "pop_nsh()%n", &n)) {
> > + nl_msg_put_flag(actions, OVS_ACTION_ATTR_POP_NSH);
> > return n;
> > }
> > }
> > @@ -2198,6 +2254,13 @@ static const struct attr_len_tbl
> > ovs_tun_key_attr_lens[OVS_TUNNEL_KEY_ATTR_MAX +
> > [OVS_TUNNEL_KEY_ATTR_IPV6_DST] = { .len = 16 },
> > };
> >
> > +static const struct attr_len_tbl
> > +ovs_nsh_key_attr_lens[OVS_NSH_KEY_ATTR_MAX + 1] = {
> > + [OVS_NSH_KEY_ATTR_BASE] = { .len = 8 },
> > + [OVS_NSH_KEY_ATTR_MD1] = { .len = 16 },
> > + [OVS_NSH_KEY_ATTR_MD2] = { .len = ATTR_LEN_VARIABLE },
> > +};
> > +
> > static const struct attr_len_tbl
> > ovs_flow_key_attr_lens[OVS_KEY_ATTR_MAX + 1] = {
> > [OVS_KEY_ATTR_ENCAP] = { .len = ATTR_LEN_NESTED },
> > [OVS_KEY_ATTR_PRIORITY] = { .len = 4 },
> > @@ -2229,7 +2292,9 @@ static const struct attr_len_tbl
> > ovs_flow_key_attr_lens[OVS_KEY_ATTR_MAX + 1] =
> > [OVS_KEY_ATTR_CT_ORIG_TUPLE_IPV4] = { .len = sizeof(struct
> > ovs_key_ct_tuple_ipv4) },
> > [OVS_KEY_ATTR_CT_ORIG_TUPLE_IPV6] = { .len = sizeof(struct
> > ovs_key_ct_tuple_ipv6) },
> > [OVS_KEY_ATTR_PACKET_TYPE] = { .len = 4 },
> > - [OVS_KEY_ATTR_NSH] = { .len = sizeof(struct ovs_key_nsh) },
> > + [OVS_KEY_ATTR_NSH] = { .len = ATTR_LEN_NESTED,
> > + .next = ovs_nsh_key_attr_lens,
> > + .next_max = OVS_NSH_KEY_ATTR_MAX },
> > };
> >
> > /* Returns the correct length of the payload for a flow key attribute of the
> > @@ -2280,6 +2345,135 @@ ovs_frag_type_to_string(enum ovs_frag_type
> > type)
> > }
> > }
> >
> > +enum odp_key_fitness
> > +odp_nsh_hdr_from_attr(const struct nlattr *attr,
> > + struct nsh_hdr *nsh_hdr)
> > +{
> > + unsigned int left;
> > + const struct nlattr *a;
> > + bool unknown = false;
> > + uint8_t flags = 0;
> > + size_t mdlen = 0;
> > + bool has_md1 = false;
> > + bool has_md2 = false;
> > +
> > + NL_NESTED_FOR_EACH (a, left, attr) {
> > + uint16_t type = nl_attr_type(a);
> > + size_t len = nl_attr_get_size(a);
> > + int expected_len = odp_key_attr_len(ovs_nsh_key_attr_lens,
> > + OVS_NSH_KEY_ATTR_MAX, type);
> > +
> > + if (len != expected_len && expected_len >= 0) {
> > + return ODP_FIT_ERROR;
> > + }
> > +
> > + switch (type) {
> > + case OVS_NSH_KEY_ATTR_BASE: {
> > + const struct ovs_nsh_key_base *base = nl_attr_get(a);
> > + nsh_hdr->next_proto = base->np;
> > + nsh_hdr->md_type = base->mdtype;
> > + put_16aligned_be32(&nsh_hdr->path_hdr, base->path_hdr);
> > + flags = base->flags;
> > + break;
> > + }
> > + case OVS_NSH_KEY_ATTR_MD1: {
> > + const struct ovs_nsh_key_md1 *md1 = nl_attr_get(a);
> > + struct nsh_md1_ctx *md1_dst = nsh_md1_ctx(nsh_hdr);
> > + has_md1 = true;
> > + mdlen = sizeof *md1;
> > + memcpy(md1_dst, md1, mdlen);
> > + break;
> > + }
> > + case OVS_NSH_KEY_ATTR_MD2: {
> > + struct nsh_md2_tlv *md2_dst = nsh_md2_ctx(nsh_hdr);
> > + const uint8_t *md2 = nl_attr_get(a);
> > + has_md2 = true;
> > + mdlen = nl_attr_get_size(a);
> > + memcpy(md2_dst, md2, mdlen);
> > + break;
> > + }
> > + default:
> > + /* Allow this to show up as unexpected, if there are unknown
> > + * tunnel attribute, eventually resulting in ODP_FIT_TOO_MUCH.
> > */
> > + unknown = true;
> > + break;
> > + }
> > + }
> > +
> > + if (unknown) {
> > + return ODP_FIT_TOO_MUCH;
> > + }
> > +
> > + if ((has_md1 && nsh_hdr->md_type != NSH_M_TYPE1)
> > + || (has_md2 && nsh_hdr->md_type != NSH_M_TYPE2)) {
> > + return ODP_FIT_ERROR;
> > + }
> > +
> > + /* nsh header length = NSH_BASE_HDR_LEN + mdlen */
> > + nsh_hdr->ver_flags_len = htons(flags << NSH_FLAGS_SHIFT |
> > + (NSH_BASE_HDR_LEN + mdlen) >> 2);
> > +
> > + return ODP_FIT_PERFECT;
> > +}
> > +
> > +enum odp_key_fitness
> > +odp_nsh_key_from_attr(const struct nlattr *attr, struct flow_nsh *nsh)
> > +{
> > + unsigned int left;
> > + const struct nlattr *a;
> > + bool unknown = false;
> > + bool has_md1 = false;
> > +
> > + NL_NESTED_FOR_EACH (a, left, attr) {
> > + uint16_t type = nl_attr_type(a);
> > + size_t len = nl_attr_get_size(a);
> > + int expected_len = odp_key_attr_len(ovs_nsh_key_attr_lens,
> > + OVS_NSH_KEY_ATTR_MAX, type);
> > +
> > + if (len != expected_len && expected_len >= 0) {
> > + return ODP_FIT_ERROR;
> > + }
> > +
> > + switch (type) {
> > + case OVS_NSH_KEY_ATTR_BASE: {
> > + const struct ovs_nsh_key_base *base = nl_attr_get(a);
> > + nsh->flags = base->flags;
> > + nsh->mdtype = base->mdtype;
> > + nsh->np = base->np;
> > + nsh->spi = htonl((ntohl(base->path_hdr) & NSH_SPI_MASK) >>
> > + NSH_SPI_SHIFT);
> > + nsh->si = (ntohl(base->path_hdr) & NSH_SI_MASK) >>
> > NSH_SI_SHIFT;
> > + break;
> > + }
> > + case OVS_NSH_KEY_ATTR_MD1: {
> > + const struct ovs_nsh_key_md1 *md1 = nl_attr_get(a);
> > + has_md1 = true;
> > + nsh->context[0] = md1->context[0];
> > + nsh->context[1] = md1->context[1];
> > + nsh->context[2] = md1->context[2];
> > + nsh->context[3] = md1->context[3];
> > + break;
> > + }
> > + case OVS_NSH_KEY_ATTR_MD2:
> > + default:
> > + /* Allow this to show up as unexpected, if there are unknown
> > + * tunnel attribute, eventually resulting in ODP_FIT_TOO_MUCH.
> > */
> > + unknown = true;
> > + break;
> > + }
> > + }
> > +
> > + if (unknown) {
> > + return ODP_FIT_TOO_MUCH;
> > + }
> > +
> > + if (has_md1 && nsh->mdtype != NSH_M_TYPE1) {
> > + return ODP_FIT_ERROR;
> > + }
> > +
> > + return ODP_FIT_PERFECT;
> > +}
> > +
> > static enum odp_key_fitness
> > odp_tun_key_from_attr__(const struct nlattr *attr, bool is_mask,
> > struct flow_tnl *tun)
> > @@ -2971,6 +3165,80 @@ format_odp_tun_geneve(const struct nlattr
> > *attr,
> > }
> >
> > static void
> > +format_odp_nsh_attr(const struct nlattr *attr, const struct nlattr
> > *mask_attr,
> > + struct ds *ds)
> > +{
> > + unsigned int left;
> > + const struct nlattr *a;
> > + struct ovs_key_nsh nsh;
> > + struct ovs_key_nsh nsh_mask;
> > +
> > + memset(&nsh, 0, sizeof nsh);
> > + memset(&nsh_mask, 0xff, sizeof nsh_mask);
> > +
> > + NL_NESTED_FOR_EACH (a, left, attr) {
> > + enum ovs_nsh_key_attr type = nl_attr_type(a);
> > + const struct nlattr *ma = NULL;
> > +
> > + if (mask_attr) {
> > + ma = nl_attr_find__(nl_attr_get(mask_attr),
> > + nl_attr_get_size(mask_attr), type);
> > + }
> > +
> > + if (!check_attr_len(ds, a, ma, ovs_nsh_key_attr_lens,
> > + OVS_NSH_KEY_ATTR_MAX, true)) {
> > + continue;
> > + }
> > +
> > + switch (type) {
> > + case OVS_NSH_KEY_ATTR_BASE: {
> > + const struct ovs_nsh_key_base * base = nl_attr_get(a);
> > + const struct ovs_nsh_key_base * base_mask
> > + = ma ? nl_attr_get(ma) : NULL;
> > + nsh.flags = base->flags;
> > + nsh.mdtype = base->mdtype;
> > + nsh.np = base->np;
> > + nsh.path_hdr = base->path_hdr;
> > + if (base_mask) {
> > + nsh_mask.flags = base_mask->flags;
> > + nsh_mask.mdtype = base_mask->mdtype;
> > + nsh_mask.np = base_mask->np;
> > + nsh_mask.path_hdr = base_mask->path_hdr;
> > + }
> > + break;
> > + }
> > + case OVS_NSH_KEY_ATTR_MD1: {
> > + const struct ovs_nsh_key_md1 * md1 = nl_attr_get(a);
> > + const struct ovs_nsh_key_md1 * md1_mask
> > + = ma ? nl_attr_get(ma) : NULL;
> > + nsh.context[0] = md1->context[0];
> > + nsh.context[1] = md1->context[1];
> > + nsh.context[2] = md1->context[2];
> > + nsh.context[3] = md1->context[3];
> > + if (md1_mask) {
> > + nsh_mask.context[0] = md1_mask->context[0];
> > + nsh_mask.context[1] = md1_mask->context[1];
> > + nsh_mask.context[2] = md1_mask->context[2];
> > + nsh_mask.context[3] = md1_mask->context[3];
> > + }
> > + break;
> > + }
> > + case OVS_NSH_KEY_ATTR_MD2:
> > + case __OVS_NSH_KEY_ATTR_MAX:
> > + default:
> > + /* No support for matching other metadata formats yet. */
> > + break;
> > + }
> > + }
> > +
> > + if (mask_attr) {
> > + format_nsh_key_mask(ds, &nsh, &nsh_mask);
> > + } else {
> > + format_nsh_key(ds, &nsh);
> > + }
> > +}
> > +
> > +static void
> > format_odp_tun_attr(const struct nlattr *attr, const struct nlattr
> > *mask_attr,
> > struct ds *ds, bool verbose)
> > {
> > @@ -3448,9 +3716,7 @@ format_odp_key_attr__(const struct nlattr *a,
> > const struct nlattr *ma,
> > break;
> > }
> > case OVS_KEY_ATTR_NSH: {
> > - const struct ovs_key_nsh *mask = ma ? nl_attr_get(ma) : NULL;
> > - const struct ovs_key_nsh *key = nl_attr_get(a);
> > - format_nsh_key_mask(ds, key, mask);
> > + format_odp_nsh_attr(a, ma, ds);
> > break;
> > }
> > case OVS_KEY_ATTR_UNSPEC:
> > @@ -4549,6 +4815,129 @@ geneve_to_attr(struct ofpbuf *a, const void
> > *data_)
> > } SCAN_END_SINGLE(ATTR)
> >
> > static int
> > +parse_odp_nsh_key_mask_attr(const char *s, struct ofpbuf *key,
> > + struct ofpbuf *mask)
> > +{
> > + if (strncmp(s, "nsh(", 4) == 0) {
> > + const char *start = s;
> > + int len;
> > + struct flow_nsh skey, smask;
> > +
> > + s += 4;
> > +
> > + memset(&skey, 0, sizeof skey);
> > + memset(&smask, 0, sizeof smask);
> > + do {
> > + len = 0;
> > +
> > + if (strncmp(s, "flags=", 6) == 0) {
> > + s += 6;
> > + len = scan_u8(s, &skey.flags, mask ? &smask.flags : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > +
> > + if (strncmp(s, "mdtype=", 7) == 0) {
> > + s += 7;
> > + len = scan_u8(s, &skey.mdtype, mask ? &smask.mdtype : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > +
> > + if (strncmp(s, "np=", 3) == 0) {
> > + s += 3;
> > + len = scan_u8(s, &skey.np, mask ? &smask.np : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > +
> > + if (strncmp(s, "spi=", 4) == 0) {
> > + s += 4;
> > + len = scan_be32(s, &skey.spi, mask ? &smask.spi : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > +
> > + if (strncmp(s, "si=", 3) == 0) {
> > + s += 3;
> > + len = scan_u8(s, &skey.si, mask ? &smask.si : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > +
> > + if (strncmp(s, "c1=", 3) == 0) {
> > + s += 3;
> > + len = scan_be32(s, &skey.context[0],
> > + mask ? &smask.context[0] : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > +
> > + if (strncmp(s, "c2=", 3) == 0) {
> > + s += 3;
> > + len = scan_be32(s, &skey.context[1],
> > + mask ? &smask.context[1] : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > +
> > + if (strncmp(s, "c3=", 3) == 0) {
> > + s += 3;
> > + len = scan_be32(s, &skey.context[2],
> > + mask ? &smask.context[2] : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > +
> > + if (strncmp(s, "c4=", 3) == 0) {
> > + s += 3;
> > + len = scan_be32(s, &skey.context[3],
> > + mask ? &smask.context[3] : NULL);
> > + if (len == 0) {
> > + return -EINVAL;
> > + }
> > + s += len;
> > + continue;
> > + }
> > + } while (*s++ == ',' && len != 0);
> > + if (s[-1] != ')') {
> > + return -EINVAL;
> > + }
> > +
> > + nsh_key_to_attr(key, &skey, NULL, 0, false);
> > + if (mask) {
> > + nsh_key_to_attr(mask, &smask, NULL, 0, true);
> > + }
> > + return s - start;
> > + }
> > + return 0;
> > +}
> > +
> > +static int
> > parse_odp_key_mask_attr(const char *s, const struct simap *port_names,
> > struct ofpbuf *key, struct ofpbuf *mask)
> > {
> > @@ -4694,16 +5083,13 @@ parse_odp_key_mask_attr(const char *s,
> > const struct simap *port_names,
> > SCAN_FIELD("id=", be16, id);
> > } SCAN_END(OVS_KEY_ATTR_PACKET_TYPE);
> >
> > - SCAN_BEGIN("nsh(", struct ovs_key_nsh) {
> > - SCAN_FIELD("flags=", u8, flags);
> > - SCAN_FIELD("mdtype=", u8, mdtype);
> > - SCAN_FIELD("np=", u8, np);
> > - SCAN_FIELD("path_hdr=", be32, path_hdr);
> > - SCAN_FIELD("c1=", be32, c[0]);
> > - SCAN_FIELD("c2=", be32, c[1]);
> > - SCAN_FIELD("c3=", be32, c[2]);
> > - SCAN_FIELD("c4=", be32, c[2]);
> > - } SCAN_END(OVS_KEY_ATTR_NSH);
> > + /* nsh is nested, it needs special process */
> > + int ret = parse_odp_nsh_key_mask_attr(s, key, mask);
> > + if (ret < 0) {
> > + return ret;
> > + } else {
> > + s += ret;
> > + }
> >
> > /* Encap open-coded. */
> > if (!strncmp(s, "encap(", 6)) {
> > @@ -4994,11 +5380,7 @@ odp_flow_key_from_flow__(const struct
> > odp_flow_key_parms *parms,
> > mpls_key[i].mpls_lse = data->mpls_lse[i];
> > }
> > } else if (flow->dl_type == htons(ETH_TYPE_NSH)) {
> > - struct ovs_key_nsh *nsh_key;
> > -
> > - nsh_key = nl_msg_put_unspec_uninit(buf, OVS_KEY_ATTR_NSH,
> > - sizeof *nsh_key);
> > - get_nsh_key(data, nsh_key, export_mask);
> > + nsh_key_to_attr(buf, &data->nsh, NULL, 0, export_mask);
> > }
> >
> > if (is_ip_any(flow) && !(flow->nw_frag & FLOW_NW_FRAG_LATER)) {
> > @@ -5558,13 +5940,10 @@ parse_l2_5_onward(const struct nlattr
> > *attrs[OVS_KEY_ATTR_MAX + 1],
> > expected_attrs |= UINT64_C(1) << OVS_KEY_ATTR_NSH;
> > }
> > if (present_attrs & (UINT64_C(1) << OVS_KEY_ATTR_NSH)) {
> > - const struct ovs_key_nsh *nsh_key;
> > -
> > - nsh_key = nl_attr_get(attrs[OVS_KEY_ATTR_NSH]);
> > - put_nsh_key(nsh_key, flow, false);
> > + odp_nsh_key_from_attr(attrs[OVS_KEY_ATTR_NSH], &flow->nsh);
> > if (is_mask) {
> > - check_start = nsh_key;
> > - check_len = sizeof *nsh_key;
> > + check_start = nl_attr_get(attrs[OVS_KEY_ATTR_NSH]);
> > + check_len = nl_attr_get_size(attrs[OVS_KEY_ATTR_NSH]);
> > expected_bit = OVS_KEY_ATTR_NSH;
> > }
> > }
> > @@ -6620,13 +6999,13 @@ get_nsh_key(const struct flow *flow, struct
> > ovs_key_nsh *nsh, bool is_mask)
> > flow->nsh.si);
> > if (is_mask) {
> > for (int i = 0; i < 4; i++) {
> > - nsh->c[i] = flow->nsh.c[i];
> > + nsh->context[i] = flow->nsh.context[i];
> > }
> > } else {
> > switch (nsh->mdtype) {
> > case NSH_M_TYPE1:
> > for (int i = 0; i < 4; i++) {
> > - nsh->c[i] = flow->nsh.c[i];
> > + nsh->context[i] = flow->nsh.context[i];
> > }
> > break;
> > case NSH_M_TYPE2:
> > @@ -6650,17 +7029,125 @@ put_nsh_key(const struct ovs_key_nsh *nsh,
> > struct flow *flow,
> > switch (nsh->mdtype) {
> > case NSH_M_TYPE1:
> > for (int i = 0; i < 4; i++) {
> > - flow->nsh.c[i] = nsh->c[i];
> > + flow->nsh.context[i] = nsh->context[i];
> > }
> > break;
> > case NSH_M_TYPE2:
> > default:
> > /* No match support for other MD formats yet. */
> > - memset(flow->nsh.c, 0, sizeof flow->nsh.c);
> > + memset(flow->nsh.context, 0, sizeof flow->nsh.context);
> > break;
> > }
> > }
> >
> > +static bool
> > +commit_nsh(const struct flow_nsh * flow_nsh, bool use_masked_set,
> > + const struct ovs_key_nsh *key, struct ovs_key_nsh *base,
> > + struct ovs_key_nsh *mask, size_t size,
> > + struct ofpbuf *odp_actions)
> > +{
> > + enum ovs_key_attr attr = OVS_KEY_ATTR_NSH;
> > +
> > + if (memcmp(key, base, size) == 0) {
> > + /* Mask bits are set when we have either read or set the
> > corresponding
> > + * values. Masked bits will be exact-matched, no need to set them
> > + * if the value did not actually change. */
> > + return false;
> > + }
> > +
> > + bool fully_masked = odp_mask_is_exact(attr, mask, size);
> > +
> > + if (use_masked_set && !fully_masked) {
> > + size_t nsh_key_ofs;
> > + struct ovs_nsh_key_base nsh_base;
> > + struct ovs_nsh_key_base nsh_base_mask;
> > + struct ovs_nsh_key_md1 md1;
> > + struct ovs_nsh_key_md1 md1_mask;
> > + size_t offset = nl_msg_start_nested(odp_actions,
> > + OVS_ACTION_ATTR_SET_MASKED);
> > +
> > + nsh_base.flags = key->flags;
> > + nsh_base.mdtype = key->mdtype;
> > + nsh_base.np = key->np;
> > + nsh_base.path_hdr = key->path_hdr;
> > +
> > + nsh_base_mask.flags = mask->flags;
> > + nsh_base_mask.mdtype = mask->mdtype;
> > + nsh_base_mask.np = mask->np;
> > + nsh_base_mask.path_hdr = mask->path_hdr;
> > +
> > + /* OVS_KEY_ATTR_NSH keys */
> > + nsh_key_ofs = nl_msg_start_nested(odp_actions,
> > OVS_KEY_ATTR_NSH);
> > +
> > + char *data = nl_msg_put_unspec_uninit(odp_actions,
> > + OVS_NSH_KEY_ATTR_BASE,
> > + sizeof(nsh_base));
> > + const char *lkey = (char *)&nsh_base, *lmask = (char
> > *)&nsh_base_mask;
> > + size_t lkey_size = sizeof(nsh_base);
> > +
> > + while (lkey_size--) {
> > + *data++ = *lkey++ & *lmask++;
> > + }
> > +
> > + switch (key->mdtype) {
> > + case NSH_M_TYPE1:
> > + for (int i = 0; i < 4; i++) {
> > + md1.context[i] = key->context[i];
> > + md1_mask.context[i] = mask->context[i];
> > + }
> > + data = nl_msg_put_unspec_uninit(odp_actions,
> > + OVS_NSH_KEY_ATTR_MD1,
> > + sizeof(md1));
> > + lkey = (char *)&md1;
> > + lmask = (char *)&md1_mask;
> > + lkey_size = sizeof(md1);
> > +
> > + while (lkey_size--) {
> > + *data++ = *lkey++ & *lmask++;
> > + }
> > + break;
> > + case NSH_M_TYPE2:
> > + default:
> > + /* No match support for other MD formats yet. */
> > + break;
> > + }
> > +
> > + /* OVS_KEY_ATTR_NSH masks */
> > + data = nl_msg_put_unspec_uninit(odp_actions,
> > + OVS_NSH_KEY_ATTR_BASE,
> > + sizeof(nsh_base_mask));
> > + lmask = (char *)&nsh_base_mask;
> > +
> > + memcpy(data, lmask, sizeof(nsh_base_mask));
> > +
> > + switch (key->mdtype) {
> > + case NSH_M_TYPE1:
> > + data = nl_msg_put_unspec_uninit(odp_actions,
> > + OVS_NSH_KEY_ATTR_MD1,
> > + sizeof(md1_mask));
> > + lmask = (char *)&md1_mask;
> > + memcpy(data, lmask, sizeof(md1_mask));
> > + break;
> > + case NSH_M_TYPE2:
> > + default:
> > + /* No match support for other MD formats yet. */
> > + break;
> > + }
> > + nl_msg_end_nested(odp_actions, nsh_key_ofs);
> > +
> > + nl_msg_end_nested(odp_actions, offset);
> > + } else {
> > + if (!fully_masked) {
> > + memset(mask, 0xff, size);
> > + }
> > + size_t offset = nl_msg_start_nested(odp_actions,
> > OVS_ACTION_ATTR_SET);
> > + nsh_key_to_attr(odp_actions, flow_nsh, NULL, 0, false);
> > + nl_msg_end_nested(odp_actions, offset);
> > + }
> > + memcpy(base, key, size);
> > + return true;
> > +}
> > +
> > static void
> > commit_set_nsh_action(const struct flow *flow, struct flow *base_flow,
> > struct ofpbuf *odp_actions,
> > @@ -6684,8 +7171,8 @@ commit_set_nsh_action(const struct flow *flow,
> > struct flow *base_flow,
> > mask.mdtype = 0; /* Not writable. */
> > mask.np = 0; /* Not writable. */
> >
> > - if (commit(OVS_KEY_ATTR_NSH, use_masked, &key, &base, &mask,
> > sizeof key,
> > - odp_actions)) {
> > + if (commit_nsh(&base_flow->nsh, use_masked, &key, &base, &mask,
> > + sizeof key, odp_actions)) {
> > put_nsh_key(&base, base_flow, false);
> > if (mask.mdtype != 0) { /* Mask was changed by commit(). */
> > put_nsh_key(&mask, &wc->masks, true);
> > @@ -6788,49 +7275,36 @@ commit_set_pkt_mark_action(const struct
> > flow *flow, struct flow *base_flow,
> > }
> >
> > static void
> > -odp_put_decap_nsh_action(struct ofpbuf *odp_actions)
> > +odp_put_pop_nsh_action(struct ofpbuf *odp_actions)
> > {
> > - nl_msg_put_flag(odp_actions, OVS_ACTION_ATTR_DECAP_NSH);
> > + nl_msg_put_flag(odp_actions, OVS_ACTION_ATTR_POP_NSH);
> > }
> >
> > static void
> > -odp_put_encap_nsh_action(struct ofpbuf *odp_actions,
> > +odp_put_push_nsh_action(struct ofpbuf *odp_actions,
> > const struct flow *flow,
> > struct ofpbuf *encap_data)
> > {
> > - struct ovs_action_encap_nsh encap_nsh;
> > -
> > - encap_nsh.flags = flow->nsh.flags;
> > - encap_nsh.mdtype = flow->nsh.mdtype;
> > - encap_nsh.np = flow->nsh.np;
> > - encap_nsh.path_hdr = htonl((ntohl(flow->nsh.spi) << NSH_SPI_SHIFT) |
> > - flow->nsh.si);
> > + uint8_t * metadata = NULL;
> > + uint8_t md_size = 0;
> >
> > - switch (encap_nsh.mdtype) {
> > - case NSH_M_TYPE1: {
> > - struct nsh_md1_ctx *md1 =
> > - ALIGNED_CAST(struct nsh_md1_ctx *, encap_nsh.metadata);
> > - encap_nsh.mdlen = NSH_M_TYPE1_MDLEN;
> > - for (int i = 0; i < 4; i++) {
> > - put_16aligned_be32(&md1->c[i], flow->nsh.c[i]);
> > - }
> > - break;
> > - }
> > + switch (flow->nsh.mdtype) {
> > case NSH_M_TYPE2:
> > if (encap_data) {
> > - ovs_assert(encap_data->size < OVS_ENCAP_NSH_MAX_MD_LEN);
> > - encap_nsh.mdlen = encap_data->size;
> > - memcpy(encap_nsh.metadata, encap_data->data, encap_data-
> > >size);
> > + ovs_assert(encap_data->size < OVS_PUSH_NSH_MAX_MD_LEN);
> > + metadata = encap_data->data;
> > + md_size = encap_data->size;
> > } else {
> > - encap_nsh.mdlen = 0;
> > + md_size = 0;
> > }
> > break;
> > default:
> > - encap_nsh.mdlen = 0;
> > + md_size = 0;
> > break;
> > }
> > - nl_msg_put_unspec(odp_actions, OVS_ACTION_ATTR_ENCAP_NSH,
> > - &encap_nsh, sizeof(encap_nsh));
> > + size_t offset = nl_msg_start_nested(odp_actions,
> > OVS_ACTION_ATTR_PUSH_NSH);
> > + nsh_key_to_attr(odp_actions, &flow->nsh, metadata, md_size, false);
> > + nl_msg_end_nested(odp_actions, offset);
> > }
> >
> > static void
> > @@ -6857,8 +7331,8 @@ commit_packet_type_change(const struct flow
> > *flow,
> > break;
> > }
> > case PT_NSH:
> > - /* encap_nsh */
> > - odp_put_encap_nsh_action(odp_actions, flow, encap_data);
> > + /* push_nsh */
> > + odp_put_push_nsh_action(odp_actions, flow, encap_data);
> > base_flow->packet_type = flow->packet_type;
> > /* Update all packet headers in base_flow. */
> > memcpy(&base_flow->dl_dst, &flow->dl_dst,
> > @@ -6883,8 +7357,8 @@ commit_packet_type_change(const struct flow
> > *flow,
> > * No need to update the base flow here. */
> > switch (ntohl(base_flow->packet_type)) {
> > case PT_NSH:
> > - /* decap_nsh. */
> > - odp_put_decap_nsh_action(odp_actions);
> > + /* pop_nsh. */
> > + odp_put_pop_nsh_action(odp_actions);
> > break;
> > default:
> > /* Checks are done during translation. */
> > diff --git a/lib/odp-util.h b/lib/odp-util.h
> > index 27c2ab4..bd6be60 100644
> > --- a/lib/odp-util.h
> > +++ b/lib/odp-util.h
> > @@ -158,6 +158,10 @@ struct odputil_keybuf {
> >
> > enum odp_key_fitness odp_tun_key_from_attr(const struct nlattr *,
> > struct flow_tnl *);
> > +enum odp_key_fitness odp_nsh_key_from_attr(const struct nlattr *,
> > + struct flow_nsh *);
> > +enum odp_key_fitness odp_nsh_hdr_from_attr(const struct nlattr *,
> > + struct nsh_hdr *);
> >
> > int odp_ufid_from_string(const char *s_, ovs_u128 *ufid);
> > void odp_format_ufid(const ovs_u128 *ufid, struct ds *);
> > diff --git a/lib/packets.c b/lib/packets.c
> > index 74d87ed..db98ab3 100644
> > --- a/lib/packets.c
> > +++ b/lib/packets.c
> > @@ -403,10 +403,10 @@ pop_mpls(struct dp_packet *packet, ovs_be16
> > ethtype)
> > }
> >
> > void
> > -encap_nsh(struct dp_packet *packet, const struct ovs_action_encap_nsh
> > *encap)
> > +push_nsh(struct dp_packet *packet, const struct nsh_hdr *nsh_hdr_src)
> > {
> > struct nsh_hdr *nsh;
> > - size_t length = NSH_BASE_HDR_LEN + encap->mdlen;
> > + size_t length = nsh_hdr_len(nsh_hdr_src);
> > uint8_t next_proto;
> >
> > switch (ntohl(packet->packet_type)) {
> > @@ -427,23 +427,8 @@ encap_nsh(struct dp_packet *packet, const struct
> > ovs_action_encap_nsh *encap)
> > }
> >
> > nsh = (struct nsh_hdr *) dp_packet_push_uninit(packet, length);
> > - nsh->ver_flags_len = htons(encap->flags << NSH_FLAGS_SHIFT | length
> > >> 2);
> > + memcpy(nsh, nsh_hdr_src, length);
> > nsh->next_proto = next_proto;
> > - put_16aligned_be32(&nsh->path_hdr, encap->path_hdr);
> > - nsh->md_type = encap->mdtype;
> > - switch (nsh->md_type) {
> > - case NSH_M_TYPE1:
> > - nsh->md1 = *ALIGNED_CAST(struct nsh_md1_ctx *, encap-
> > >metadata);
> > - break;
> > - case NSH_M_TYPE2: {
> > - /* The MD2 metadata in encap is already padded to 4 bytes. */
> > - size_t len = ROUND_UP(encap->mdlen, 4);
> > - memcpy(&nsh->md2, encap->metadata, len);
> > - break;
> > - }
> > - default:
> > - OVS_NOT_REACHED();
> > - }
> >
> > packet->packet_type = htonl(PT_NSH);
> > dp_packet_reset_offsets(packet);
> > @@ -451,7 +436,7 @@ encap_nsh(struct dp_packet *packet, const struct
> > ovs_action_encap_nsh *encap)
> > }
> >
> > bool
> > -decap_nsh(struct dp_packet *packet)
> > +pop_nsh(struct dp_packet *packet)
> > {
> > struct nsh_hdr *nsh = (struct nsh_hdr *) dp_packet_l3(packet);
> > size_t length;
> > diff --git a/lib/packets.h b/lib/packets.h
> > index 705d0b2..e7832ba 100644
> > --- a/lib/packets.h
> > +++ b/lib/packets.h
> > @@ -434,9 +434,8 @@ void push_eth(struct dp_packet *packet, const
> > struct eth_addr *dst,
> > const struct eth_addr *src);
> > void pop_eth(struct dp_packet *packet);
> >
> > -void encap_nsh(struct dp_packet *packet,
> > - const struct ovs_action_encap_nsh *encap_nsh);
> > -bool decap_nsh(struct dp_packet *packet);
> > +void push_nsh(struct dp_packet *packet, const struct nsh_hdr
> > *nsh_hdr_src);
> > +bool pop_nsh(struct dp_packet *packet);
> >
> > #define LLC_DSAP_SNAP 0xaa
> > #define LLC_SSAP_SNAP 0xaa
> > diff --git a/ofproto/ofproto-dpif-ipfix.c b/ofproto/ofproto-dpif-ipfix.c
> > index 472c272..16976c4 100644
> > --- a/ofproto/ofproto-dpif-ipfix.c
> > +++ b/ofproto/ofproto-dpif-ipfix.c
> > @@ -2823,8 +2823,8 @@ dpif_ipfix_read_actions(const struct flow *flow,
> > case OVS_ACTION_ATTR_POP_MPLS:
> > case OVS_ACTION_ATTR_PUSH_ETH:
> > case OVS_ACTION_ATTR_POP_ETH:
> > - case OVS_ACTION_ATTR_ENCAP_NSH:
> > - case OVS_ACTION_ATTR_DECAP_NSH:
> > + case OVS_ACTION_ATTR_PUSH_NSH:
> > + case OVS_ACTION_ATTR_POP_NSH:
> > case OVS_ACTION_ATTR_UNSPEC:
> > case __OVS_ACTION_ATTR_MAX:
> > default:
> > diff --git a/ofproto/ofproto-dpif-sflow.c b/ofproto/ofproto-dpif-sflow.c
> > index 65a2003..1af1569 100644
> > --- a/ofproto/ofproto-dpif-sflow.c
> > +++ b/ofproto/ofproto-dpif-sflow.c
> > @@ -1199,8 +1199,8 @@ dpif_sflow_read_actions(const struct flow *flow,
> > break;
> > case OVS_ACTION_ATTR_SAMPLE:
> > case OVS_ACTION_ATTR_CLONE:
> > - case OVS_ACTION_ATTR_ENCAP_NSH:
> > - case OVS_ACTION_ATTR_DECAP_NSH:
> > + case OVS_ACTION_ATTR_PUSH_NSH:
> > + case OVS_ACTION_ATTR_POP_NSH:
> > case OVS_ACTION_ATTR_UNSPEC:
> > case __OVS_ACTION_ATTR_MAX:
> > default:
> > diff --git a/ofproto/ofproto-dpif-xlate.c b/ofproto/ofproto-dpif-xlate.c
> > index 973e760..437c519 100644
> > --- a/ofproto/ofproto-dpif-xlate.c
> > +++ b/ofproto/ofproto-dpif-xlate.c
> > @@ -4392,8 +4392,8 @@ xlate_fixup_actions(struct ofpbuf *b, const struct
> > nlattr *actions,
> > case OVS_ACTION_ATTR_CT:
> > case OVS_ACTION_ATTR_PUSH_ETH:
> > case OVS_ACTION_ATTR_POP_ETH:
> > - case OVS_ACTION_ATTR_ENCAP_NSH:
> > - case OVS_ACTION_ATTR_DECAP_NSH:
> > + case OVS_ACTION_ATTR_PUSH_NSH:
> > + case OVS_ACTION_ATTR_POP_NSH:
> > case OVS_ACTION_ATTR_METER:
> > ofpbuf_put(b, a, nl_attr_len_pad(a, left));
> > break;
> > @@ -5804,17 +5804,17 @@ rewrite_flow_encap_ethernet(struct xlate_ctx
> > *ctx,
> >
> > /* For an MD2 NSH header returns a pointer to an ofpbuf with the
> > encoded
> > * MD2 TLVs provided as encap properties to the encap operation. This
> > - * will be stored as encap_data in the ctx and copied into the encap_nsh
> > + * will be stored as encap_data in the ctx and copied into the push_nsh
> > * action at the next commit. */
> > static struct ofpbuf *
> > -rewrite_flow_encap_nsh(struct xlate_ctx *ctx,
> > - const struct ofpact_encap *encap,
> > - struct flow *flow,
> > - struct flow_wildcards *wc)
> > +rewrite_flow_push_nsh(struct xlate_ctx *ctx,
> > + const struct ofpact_encap *encap,
> > + struct flow *flow,
> > + struct flow_wildcards *wc)
> > {
> > ovs_be32 packet_type = flow->packet_type;
> > const char *ptr = (char *) encap->props;
> > - struct ofpbuf *buf = ofpbuf_new(OVS_ENCAP_NSH_MAX_MD_LEN);
> > + struct ofpbuf *buf = ofpbuf_new(OVS_PUSH_NSH_MAX_MD_LEN);
> > uint8_t md_type = NSH_M_TYPE1;
> > uint8_t np = 0;
> > int i;
> > @@ -5854,7 +5854,7 @@ rewrite_flow_encap_nsh(struct xlate_ctx *ctx,
> > }
> > ptr += ROUND_UP(prop_ptr->len, 8);
> > }
> > - if (buf->size == 0 || buf->size > OVS_ENCAP_NSH_MAX_MD_LEN) {
> > + if (buf->size == 0 || buf->size > OVS_PUSH_NSH_MAX_MD_LEN) {
> > ofpbuf_delete(buf);
> > buf = NULL;
> > }
> > @@ -5899,7 +5899,7 @@ rewrite_flow_encap_nsh(struct xlate_ctx *ctx,
> >
> > if (md_type == NSH_M_TYPE1) {
> > flow->nsh.mdtype = NSH_M_TYPE1;
> > - memset(flow->nsh.c, 0, sizeof flow->nsh.c);
> > + memset(flow->nsh.context, 0, sizeof flow->nsh.context);
> > if (buf) {
> > /* Drop any MD2 context TLVs. */
> > ofpbuf_delete(buf);
> > @@ -5930,7 +5930,7 @@ xlate_generic_encap_action(struct xlate_ctx
> > *ctx,
> > rewrite_flow_encap_ethernet(ctx, flow, wc);
> > break;
> > case PT_NSH:
> > - encap_data = rewrite_flow_encap_nsh(ctx, encap, flow, wc);
> > + encap_data = rewrite_flow_push_nsh(ctx, encap, flow, wc);
> > break;
> > default:
> > /* New packet type was checked during decoding. */
> > @@ -5973,7 +5973,7 @@ xlate_generic_decap_action(struct xlate_ctx
> > *ctx,
> > }
> > return false;
> > case PT_NSH:
> > - /* The decap_nsh action is generated at the commit executed as
> > + /* The pop_nsh action is generated at the commit executed as
> > * part of freezing the ctx for recirculation. Here we just set
> > * the new packet type based on the NSH next protocol field. */
> > switch (flow->nsh.np) {
> > diff --git a/tests/nsh.at b/tests/nsh.at
> > index aa80a2a..562f3da 100644
> > --- a/tests/nsh.at
> > +++ b/tests/nsh.at
> > @@ -105,7 +105,7 @@ bridge("br0")
> >
> > Final flow:
> > in_port=1,vlan_tci=0x0000,dl_src=00:00:00:00:00:00,dl_dst=11:22:33:44:55:
> > 66,dl_type=0x894f,nsh_flags=0,nsh_mdtype=1,nsh_np=3,nsh_spi=0x1234,ns
> > h_si=255,nsh_c1=0x11223344,nsh_c2=0x0,nsh_c3=0x0,nsh_c4=0x0,nw_pro
> > to=0,nw_tos=0,nw_ecn=0,nw_ttl=0
> > Megaflow:
> > recirc_id=0,eth,ip,in_port=1,dl_dst=66:77:88:99:aa:bb,nw_frag=no
> > -Datapath actions:
> > encap_nsh(flags=0,mdtype=1,np=3,spi=0x1234,si=255,c1=0x11223344,c2=0
> > x0,c3=0x0,c4=0x0),push_eth(src=00:00:00:00:00:00,dst=11:22:33:44:55:66),
> > pop_eth,decap_nsh(),set(eth(dst=11:22:33:44:55:66)),recirc(0x1)
> > +Datapath actions:
> > push_nsh(flags=0,mdtype=1,np=3,spi=0x1234,si=255,c1=0x11223344,c2=0x
> > 0,c3=0x0,c4=0x0),push_eth(src=00:00:00:00:00:00,dst=11:22:33:44:55:66),p
> > op_eth,pop_nsh(),set(eth(dst=11:22:33:44:55:66)),recirc(0x1)
> > ])
> >
> > AT_CHECK([
> > @@ -121,7 +121,7 @@ bridge("br0")
> >
> > Final flow: unchanged
> > Megaflow:
> > recirc_id=0,eth,in_port=4,dl_type=0x894f,nsh_mdtype=1,nsh_np=3,nsh_spi
> > =0x1234,nsh_c1=0x11223344
> > -Datapath actions: pop_eth,decap_nsh(),recirc(0x2)
> > +Datapath actions: pop_eth,pop_nsh(),recirc(0x2)
> > ])
> >
> > # Now send two real ICMP echo request packets in on port p1
> > @@ -139,7 +139,7 @@ ovs-appctl time/warp 1000
> > AT_CHECK([
> > ovs-appctl dpctl/dump-flows dummy at ovs-dummy | strip_used | grep -v
> > ipv6 | sort
> > ], [0], [flow-dump from non-dpdk interfaces:
> > -
> > recirc_id(0),in_port(1),packet_type(ns=0,id=0),eth(dst=1e:2c:e9:2a:66:9e),e
> > th_type(0x0800),ipv4(frag=no), packets:1, bytes:98, used:0.0s,
> > actions:encap_nsh(flags=0,mdtype=1,np=3,spi=0x1234,si=255,c1=0x112233
> > 44,c2=0x0,c3=0x0,c4=0x0),push_eth(src=00:00:00:00:00:00,dst=11:22:33:44
> > :55:66),pop_eth,decap_nsh(),set(eth(dst=11:22:33:44:55:66)),recirc(0x3)
> > +recirc_id(0),in_port(1),packet_type(ns=0,id=0),eth(dst=1e:2c:e9:2a:66:9e),
> > eth_type(0x0800),ipv4(frag=no), packets:1, bytes:98, used:0.0s,
> > actions:push_nsh(flags=0,mdtype=1,np=3,spi=0x1234,si=255,c1=0x1122334
> > 4,c2=0x0,c3=0x0,c4=0x0),push_eth(src=00:00:00:00:00:00,dst=11:22:33:44:
> > 55:66),pop_eth,pop_nsh(),set(eth(dst=11:22:33:44:55:66)),recirc(0x3)
> >
> > recirc_id(0x3),in_port(1),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(fra
> > g=no), packets:1, bytes:98, used:0.0s, actions:2
> > ])
> >
> > @@ -170,7 +170,7 @@ ovs-appctl time/warp 1000
> > AT_CHECK([
> > ovs-appctl dpctl/dump-flows dummy at ovs-dummy | strip_used | grep -v
> > ipv6 | sort
> > ], [0], [flow-dump from non-dpdk interfaces:
> > -
> > recirc_id(0),in_port(1),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(frag=
> > no), packets:1, bytes:98, used:0.0s,
> > actions:push_vlan(vid=100,pcp=0),encap_nsh(flags=0,mdtype=1,np=3,spi=0
> > x0,si=255,c1=0x0,c2=0x0,c3=0x0,c4=0x0),decap_nsh(),recirc(0x4)
> > +recirc_id(0),in_port(1),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(frag
> > =no), packets:1, bytes:98, used:0.0s,
> > actions:push_vlan(vid=100,pcp=0),push_nsh(flags=0,mdtype=1,np=3,spi=0x
> > 0,si=255,c1=0x0,c2=0x0,c3=0x0,c4=0x0),pop_nsh(),recirc(0x4)
> >
> > recirc_id(0x4),in_port(1),packet_type(ns=0,id=0),eth_type(0x8100),vlan(vid
> > =100,pcp=0),encap(eth_type(0x0800),ipv4(frag=no)), packets:1, bytes:102,
> > used:0.0s, actions:2
> > ])
> >
> > @@ -195,7 +195,7 @@ ovs-vsctl set bridge br0 datapath_type=dummy \
> > add-port br0 v4 -- set Interface v4 type=patch options:peer=v3
> > ofport_request=4])
> >
> > AT_DATA([flows.txt], [dnl
> > -
> > table=0,in_port=1,ip,actions=encap(nsh(md_type=2,tlv(0x1000,10,0x12345
> > 678))),set_field:0x1234-
> > >nsh_spi,encap(ethernet),set_field:11:22:33:44:55:66->dl_dst,3
> > +
> > table=0,in_port=1,ip,actions=encap(nsh(md_type=2,tlv(0x1000,10,0x12345
> > 678),tlv(0x2000,20,0xfedcba9876543210))),set_field:0x1234-
> > >nsh_spi,encap(ethernet),set_field:11:22:33:44:55:66->dl_dst,3
> >
> > table=0,in_port=4,dl_type=0x894f,nsh_mdtype=2,nsh_spi=0x1234,actions=
> > decap(),decap(),2
> > ])
> >
> > @@ -205,7 +205,7 @@ AT_CHECK([
> > ovs-ofctl -Oopenflow13 dump-flows br0 | ofctl_strip | sort | grep
> > actions
> > ], [0], [dnl
> > in_port=4,dl_type=0x894f,nsh_mdtype=2,nsh_spi=0x1234
> > actions=decap(),decap(),output:2
> > - ip,in_port=1
> > actions=encap(nsh(md_type=2,tlv(0x1000,10,0x12345678))),set_field:0x123
> > 4->nsh_spi,encap(ethernet),set_field:11:22:33:44:55:66->eth_dst,output:3
> > + ip,in_port=1
> > actions=encap(nsh(md_type=2,tlv(0x1000,10,0x12345678),tlv(0x2000,20,0x
> > fedcba9876543210))),set_field:0x1234-
> > >nsh_spi,encap(ethernet),set_field:11:22:33:44:55:66->eth_dst,output:3
> > ])
> >
> > AT_CHECK([
> > @@ -216,7 +216,7 @@ Flow:
> > icmp,in_port=1,vlan_tci=0x0000,dl_src=00:11:22:33:44:55,dl_dst=66:77:88:
> > 99
> > bridge("br0")
> > -------------
> > 0. ip,in_port=1, priority 32768
> > - encap(nsh(md_type=2,tlv(0x1000,10,0x12345678)))
> > +
> > encap(nsh(md_type=2,tlv(0x1000,10,0x12345678),tlv(0x2000,20,0xfedcba9
> > 876543210)))
> > set_field:0x1234->nsh_spi
> > encap(ethernet)
> > set_field:11:22:33:44:55:66->eth_dst
> > @@ -230,7 +230,7 @@ bridge("br0")
> >
> > Final flow:
> > in_port=1,vlan_tci=0x0000,dl_src=00:00:00:00:00:00,dl_dst=11:22:33:44:55:
> > 66,dl_type=0x894f,nsh_flags=0,nsh_mdtype=2,nsh_np=3,nsh_spi=0x1234,ns
> > h_si=255,nw_proto=0,nw_tos=0,nw_ecn=0,nw_ttl=0
> > Megaflow:
> > recirc_id=0,eth,ip,in_port=1,dl_dst=66:77:88:99:aa:bb,nw_frag=no
> > -Datapath actions:
> > encap_nsh(flags=0,mdtype=2,np=3,spi=0x1234,si=255,md2=0x10000a0412
> > 345678),push_eth(src=00:00:00:00:00:00,dst=11:22:33:44:55:66),pop_eth,d
> > ecap_nsh(),set(eth(dst=11:22:33:44:55:66)),recirc(0x1)
> > +Datapath actions:
> > push_nsh(flags=0,mdtype=2,np=3,spi=0x1234,si=255,md2=0x10000a04123
> > 4567820001408fedcba9876543210),push_eth(src=00:00:00:00:00:00,dst=11
> > :22:33:44:55:66),pop_eth,pop_nsh(),set(eth(dst=11:22:33:44:55:66)),recirc(
> > 0x1)
> > ])
> >
> > AT_CHECK([
> > @@ -246,7 +246,7 @@ bridge("br0")
> >
> > Final flow: unchanged
> > Megaflow:
> > recirc_id=0,eth,in_port=4,dl_type=0x894f,nsh_mdtype=2,nsh_np=3,nsh_spi
> > =0x1234
> > -Datapath actions: pop_eth,decap_nsh(),recirc(0x2)
> > +Datapath actions: pop_eth,pop_nsh(),recirc(0x2)
> > ])
> >
> > # Now send two real ICMP echo request packets in on port p1
> > @@ -264,7 +264,7 @@ ovs-appctl time/warp 1000
> > AT_CHECK([
> > ovs-appctl dpctl/dump-flows dummy at ovs-dummy | strip_used | grep -v
> > ipv6 | sort
> > ], [0], [flow-dump from non-dpdk interfaces:
> > -
> > recirc_id(0),in_port(1),packet_type(ns=0,id=0),eth(dst=1e:2c:e9:2a:66:9e),e
> > th_type(0x0800),ipv4(frag=no), packets:1, bytes:98, used:0.0s,
> > actions:encap_nsh(flags=0,mdtype=2,np=3,spi=0x1234,si=255,md2=0x1000
> > 0a0412345678),push_eth(src=00:00:00:00:00:00,dst=11:22:33:44:55:66),po
> > p_eth,decap_nsh(),set(eth(dst=11:22:33:44:55:66)),recirc(0x3)
> > +recirc_id(0),in_port(1),packet_type(ns=0,id=0),eth(dst=1e:2c:e9:2a:66:9e),
> > eth_type(0x0800),ipv4(frag=no), packets:1, bytes:98, used:0.0s,
> > actions:push_nsh(flags=0,mdtype=2,np=3,spi=0x1234,si=255,md2=0x10000
> > a041234567820001408fedcba9876543210),push_eth(src=00:00:00:00:00:0
> > 0,dst=11:22:33:44:55:66),pop_eth,pop_nsh(),set(eth(dst=11:22:33:44:55:66)
> > ),recirc(0x3)
> >
> > recirc_id(0x3),in_port(1),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(fra
> > g=no), packets:1, bytes:98, used:0.0s, actions:2
> > ])
> >
> > @@ -577,8 +577,8 @@ ovs-appctl time/warp 1000
> > AT_CHECK([
> > ovs-appctl dpctl/dump-flows dummy at ovs-dummy | strip_used | grep -v
> > ipv6 | sort
> > ], [0], [flow-dump from non-dpdk interfaces:
> > -
> > recirc_id(0),in_port(4),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(dst=1
> > 92.168.10.30,frag=no), packets:1, bytes:98, used:0.0s,
> > actions:pop_eth,encap_nsh(flags=0,mdtype=1,np=1,spi=0x3000,si=255,c1=0
> > x0,c2=0x0,c3=0x0,c4=0x0),clone(tnl_push(tnl_port(4789),header(size=50,ty
> > pe=4,eth(dst=aa:55:00:00:00:03,src=aa:55:00:00:00:01,dl_type=0x0800),ipv
> > 4(src=10.0.0.1,dst=10.0.0.3,proto=17,tos=0,ttl=64,frag=0x4000),udp(src=0,d
> > st=4789,csum=0x0),vxlan(flags=0xc000004,vni=0x0)),out_port(1)),set(ipv4(s
> > rc=30.0.0.1,dst=30.0.0.3)),tnl_pop(4789))
> > -tunnel(tun_id=0x0,src=30.0.0.1,dst=30.0.0.3,flags(-df-
> > csum+key)),recirc_id(0),in_port(4789),packet_type(ns=1,id=0x894f),nsh(np
> > =1,spi=0x3000,si=255), packets:1, bytes:108, used:0.0s,
> > actions:decap_nsh(),recirc(0x1)
> > +recirc_id(0),in_port(4),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(dst=
> > 192.168.10.30,frag=no), packets:1, bytes:98, used:0.0s,
> > actions:pop_eth,push_nsh(flags=0,mdtype=1,np=1,spi=0x3000,si=255,c1=0x
> > 0,c2=0x0,c3=0x0,c4=0x0),clone(tnl_push(tnl_port(4789),header(size=50,typ
> > e=4,eth(dst=aa:55:00:00:00:03,src=aa:55:00:00:00:01,dl_type=0x0800),ipv4
> > (src=10.0.0.1,dst=10.0.0.3,proto=17,tos=0,ttl=64,frag=0x4000),udp(src=0,ds
> > t=4789,csum=0x0),vxlan(flags=0xc000004,vni=0x0)),out_port(1)),set(ipv4(sr
> > c=30.0.0.1,dst=30.0.0.3)),tnl_pop(4789))
> > +tunnel(tun_id=0x0,src=30.0.0.1,dst=30.0.0.3,flags(-df-
> > csum+key)),recirc_id(0),in_port(4789),packet_type(ns=1,id=0x894f),nsh(np
> > =1,spi=0x3000,si=255), packets:1, bytes:108, used:0.0s,
> > actions:pop_nsh(),recirc(0x1)
> > tunnel(tun_id=0x0,src=30.0.0.1,dst=30.0.0.3,flags(-df-
> > csum+key)),recirc_id(0x1),in_port(4789),packet_type(ns=1,id=0x800),ipv4(f
> > rag=no), packets:1, bytes:84, used:0.0s,
> > actions:push_eth(src=00:00:00:00:00:00,dst=aa:55:aa:55:00:03),6
> > ])
> >
> > @@ -631,9 +631,9 @@ ovs-appctl time/warp 1000
> > AT_CHECK([
> > ovs-appctl dpctl/dump-flows dummy at ovs-dummy | strip_used | grep -v
> > ipv6 | sort
> > ], [0], [flow-dump from non-dpdk interfaces:
> > -
> > recirc_id(0),in_port(4),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(dst=1
> > 92.168.10.20/255.255.255.248,frag=no), packets:1, bytes:98, used:0.0s,
> > actions:pop_eth,encap_nsh(flags=0,mdtype=1,np=1,spi=0x3020,si=255,c1=0
> > x0,c2=0x0,c3=0x0,c4=0x0),clone(tnl_push(tnl_port(4789),header(size=50,ty
> > pe=4,eth(dst=aa:55:00:00:00:02,src=aa:55:00:00:00:01,dl_type=0x0800),ipv
> > 4(src=10.0.0.1,dst=10.0.0.2,proto=17,tos=0,ttl=64,frag=0x4000),udp(src=0,d
> > st=4789,csum=0x0),vxlan(flags=0xc000004,vni=0x0)),out_port(1)),set(ipv4(s
> > rc=20.0.0.1,dst=20.0.0.2)),tnl_pop(4789))
> > +recirc_id(0),in_port(4),packet_type(ns=0,id=0),eth_type(0x0800),ipv4(dst=
> > 192.168.10.20/255.255.255.248,frag=no), packets:1, bytes:98, used:0.0s,
> > actions:pop_eth,push_nsh(flags=0,mdtype=1,np=1,spi=0x3020,si=255,c1=0x
> > 0,c2=0x0,c3=0x0,c4=0x0),clone(tnl_push(tnl_port(4789),header(size=50,typ
> > e=4,eth(dst=aa:55:00:00:00:02,src=aa:55:00:00:00:01,dl_type=0x0800),ipv4
> > (src=10.0.0.1,dst=10.0.0.2,proto=17,tos=0,ttl=64,frag=0x4000),udp(src=0,ds
> > t=4789,csum=0x0),vxlan(flags=0xc000004,vni=0x0)),out_port(1)),set(ipv4(sr
> > c=20.0.0.1,dst=20.0.0.2)),tnl_pop(4789))
> > tunnel(tun_id=0x0,src=20.0.0.1,dst=20.0.0.2,flags(-df-
> > csum+key)),recirc_id(0),in_port(4789),packet_type(ns=1,id=0x894f),nsh(spi
> > =0x3020,si=255), packets:1, bytes:108, used:0.0s,
> > actions:push_eth(src=00:00:00:00:00:00,dst=11:22:33:44:55:66),set(nsh(spi=
> > 0x3020,si=254)),pop_eth,clone(tnl_push(tnl_port(4789),header(size=50,typ
> > e=4,eth(dst=aa:55:00:00:00:03,src=aa:55:00:00:00:02,dl_type=0x0800),ipv4
> > (src=20.0.0.2,dst=20.0.0.3,proto=17,tos=0,ttl=64,frag=0x4000),udp(src=0,ds
> > t=4789,csum=0x0),vxlan(flags=0xc000004,vni=0x0)),out_port(2)),set(ipv4(sr
> > c=30.0.0.2,dst=30.0.0.3)),tnl_pop(4789))
> > -tunnel(tun_id=0x0,src=30.0.0.2,dst=30.0.0.3,flags(-df-
> > csum+key)),recirc_id(0),in_port(4789),packet_type(ns=1,id=0x894f),nsh(np
> > =1,spi=0x3020,si=254), packets:1, bytes:108, used:0.0s,
> > actions:decap_nsh(),recirc(0x2)
> > +tunnel(tun_id=0x0,src=30.0.0.2,dst=30.0.0.3,flags(-df-
> > csum+key)),recirc_id(0),in_port(4789),packet_type(ns=1,id=0x894f),nsh(np
> > =1,spi=0x3020,si=254), packets:1, bytes:108, used:0.0s,
> > actions:pop_nsh(),recirc(0x2)
> > tunnel(tun_id=0x0,src=30.0.0.2,dst=30.0.0.3,flags(-df-
> > csum+key)),recirc_id(0x2),in_port(4789),packet_type(ns=1,id=0x800),ipv4(f
> > rag=no), packets:1, bytes:84, used:0.0s,
> > actions:push_eth(src=00:00:00:00:00:00,dst=aa:55:aa:55:00:03),6
> > ])
> >
> > --
> > 2.1.0
>
>
More information about the dev
mailing list