From: David Marchand <david.marchand@redhat.com>
To: Pavan Nikhilesh <pbhagavatula@marvell.com>
Cc: Jerin Jacob Kollanukkaran <jerinj@marvell.com>,
dev <dev@dpdk.org>,
"Ananyev, Konstantin" <konstantin.ananyev@intel.com>
Subject: Re: [dpdk-dev] [PATCH v4 3/3] examples/l3fwd: use reserved addresses for EM mode
Date: Wed, 3 Nov 2021 21:53:12 +0100 [thread overview]
Message-ID: <CAJFAV8ytxybikS51z9+0E+EGZjQc5xZgo1yHTFtPZGUAy7qw8w@mail.gmail.com> (raw)
In-Reply-To: <20211003202448.10908-3-pbhagavatula@marvell.com>
On Sun, Oct 3, 2021 at 10:25 PM <pbhagavatula@marvell.com> wrote:
>
> From: Pavan Nikhilesh <pbhagavatula@marvell.com>
>
> The l3fwd example should use the reserved IPv4/v6 reserved address
> ranges defined in RFC5735, RFC5180 and RFC863 discard protocol for
> the port number in the exact match mode of L3 forwarding.
>
> Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
> ---
> doc/guides/rel_notes/release_21_11.rst | 6 ++
> examples/l3fwd/l3fwd.h | 2 +-
> examples/l3fwd/l3fwd_em.c | 142 ++++++++++++-------------
> 3 files changed, 73 insertions(+), 77 deletions(-)
>
> diff --git a/doc/guides/rel_notes/release_21_11.rst b/doc/guides/rel_notes/release_21_11.rst
> index 37dc1a7786..ccdfa47baa 100644
> --- a/doc/guides/rel_notes/release_21_11.rst
> +++ b/doc/guides/rel_notes/release_21_11.rst
> @@ -125,6 +125,12 @@ New Features
> * Added tests to validate packets hard expiry.
> * Added tests to verify tunnel header verification in IPsec inbound.
>
> +* **Increase number of routes configured in l3fwd lpm, em and fib lookups.**
> +
> + * Increase number of routes from 8 to 16 for all lookup modes of l3fwd, this
> + helps in validating SoC with more than 8 ethernet devices using l3fwd.
It should be in patch1.
> + * Update EM mode to use RFC2544 reserved IP address space.
> +
>
> Removed Items
> -------------
> diff --git a/examples/l3fwd/l3fwd.h b/examples/l3fwd/l3fwd.h
> index a808d60247..853a31d4c2 100644
> --- a/examples/l3fwd/l3fwd.h
> +++ b/examples/l3fwd/l3fwd.h
> @@ -54,7 +54,7 @@
> /* 32-bit has less address-space for hugepage memory, limit to 1M entries */
> #define L3FWD_HASH_ENTRIES (1024*1024*1)
> #endif
> -#define HASH_ENTRY_NUMBER_DEFAULT 4
> +#define HASH_ENTRY_NUMBER_DEFAULT 16
>
> struct mbuf_table {
> uint16_t len;
> diff --git a/examples/l3fwd/l3fwd_em.c b/examples/l3fwd/l3fwd_em.c
> index 2a8ab6aab5..92998e26b2 100644
> --- a/examples/l3fwd/l3fwd_em.c
> +++ b/examples/l3fwd/l3fwd_em.c
> @@ -99,33 +99,64 @@ struct ipv6_l3fwd_em_route {
> uint8_t if_out;
> };
>
> -static struct ipv4_l3fwd_em_route ipv4_l3fwd_em_route_array[] = {
> - {{RTE_IPV4(101, 0, 0, 0), RTE_IPV4(100, 10, 0, 1), 101, 11, IPPROTO_TCP}, 0},
> - {{RTE_IPV4(201, 0, 0, 0), RTE_IPV4(200, 20, 0, 1), 102, 12, IPPROTO_TCP}, 1},
> - {{RTE_IPV4(111, 0, 0, 0), RTE_IPV4(100, 30, 0, 1), 101, 11, IPPROTO_TCP}, 2},
> - {{RTE_IPV4(211, 0, 0, 0), RTE_IPV4(200, 40, 0, 1), 102, 12, IPPROTO_TCP}, 3},
> +/* 198.18.0.0/16 are set aside for RFC2544 benchmarking (RFC5735).
> + * Use RFC863 Discard Protocol.
> + */
> +static const struct ipv4_l3fwd_em_route ipv4_l3fwd_em_route_array[] = {
> + {{RTE_IPV4(198, 18, 0, 0), RTE_IPV4(198, 18, 0, 1), 9, 9, IPPROTO_UDP}, 0},
> + {{RTE_IPV4(198, 18, 1, 0), RTE_IPV4(198, 18, 1, 1), 9, 9, IPPROTO_UDP}, 1},
> + {{RTE_IPV4(198, 18, 2, 0), RTE_IPV4(198, 18, 2, 1), 9, 9, IPPROTO_UDP}, 2},
> + {{RTE_IPV4(198, 18, 3, 0), RTE_IPV4(198, 18, 3, 1), 9, 9, IPPROTO_UDP}, 3},
> + {{RTE_IPV4(198, 18, 4, 0), RTE_IPV4(198, 18, 4, 1), 9, 9, IPPROTO_UDP}, 4},
> + {{RTE_IPV4(198, 18, 5, 0), RTE_IPV4(198, 18, 5, 1), 9, 9, IPPROTO_UDP}, 5},
> + {{RTE_IPV4(198, 18, 6, 0), RTE_IPV4(198, 18, 6, 1), 9, 9, IPPROTO_UDP}, 6},
> + {{RTE_IPV4(198, 18, 7, 0), RTE_IPV4(198, 18, 7, 1), 9, 9, IPPROTO_UDP}, 7},
> + {{RTE_IPV4(198, 18, 8, 0), RTE_IPV4(198, 18, 8, 1), 9, 9, IPPROTO_UDP}, 8},
> + {{RTE_IPV4(198, 18, 9, 0), RTE_IPV4(198, 18, 9, 1), 9, 9, IPPROTO_UDP}, 9},
> + {{RTE_IPV4(198, 18, 10, 0), RTE_IPV4(198, 18, 10, 1), 9, 9, IPPROTO_UDP}, 10},
> + {{RTE_IPV4(198, 18, 11, 0), RTE_IPV4(198, 18, 11, 1), 9, 9, IPPROTO_UDP}, 11},
> + {{RTE_IPV4(198, 18, 12, 0), RTE_IPV4(198, 18, 12, 1), 9, 9, IPPROTO_UDP}, 12},
> + {{RTE_IPV4(198, 18, 13, 0), RTE_IPV4(198, 18, 13, 1), 9, 9, IPPROTO_UDP}, 13},
> + {{RTE_IPV4(198, 18, 14, 0), RTE_IPV4(198, 18, 14, 1), 9, 9, IPPROTO_UDP}, 14},
> + {{RTE_IPV4(198, 18, 15, 0), RTE_IPV4(198, 18, 15, 1), 9, 9, IPPROTO_UDP}, 15},
> };
>
> -static struct ipv6_l3fwd_em_route ipv6_l3fwd_em_route_array[] = {
> - {{
> - {0xfe, 0x80, 0, 0, 0, 0, 0, 0, 0x02, 0x1e, 0x67, 0xff, 0xfe, 0, 0, 0},
> - {0xfe, 0x80, 0, 0, 0, 0, 0, 0, 0x02, 0x1b, 0x21, 0xff, 0xfe, 0x91, 0x38, 0x05},
> - 101, 11, IPPROTO_TCP}, 0},
> -
> - {{
> - {0xfe, 0x90, 0, 0, 0, 0, 0, 0, 0x02, 0x1e, 0x67, 0xff, 0xfe, 0, 0, 0},
> - {0xfe, 0x90, 0, 0, 0, 0, 0, 0, 0x02, 0x1b, 0x21, 0xff, 0xfe, 0x91, 0x38, 0x05},
> - 102, 12, IPPROTO_TCP}, 1},
> -
> - {{
> - {0xfe, 0xa0, 0, 0, 0, 0, 0, 0, 0x02, 0x1e, 0x67, 0xff, 0xfe, 0, 0, 0},
> - {0xfe, 0xa0, 0, 0, 0, 0, 0, 0, 0x02, 0x1b, 0x21, 0xff, 0xfe, 0x91, 0x38, 0x05},
> - 101, 11, IPPROTO_TCP}, 2},
> -
> - {{
> - {0xfe, 0xb0, 0, 0, 0, 0, 0, 0, 0x02, 0x1e, 0x67, 0xff, 0xfe, 0, 0, 0},
> - {0xfe, 0xb0, 0, 0, 0, 0, 0, 0, 0x02, 0x1b, 0x21, 0xff, 0xfe, 0x91, 0x38, 0x05},
> - 102, 12, IPPROTO_TCP}, 3},
> +/* 2001:0200::/48 is IANA reserved range for IPv6 benchmarking (RFC5180).
> + * Use RFC863 Discard Protocol.
> + */
> +static const struct ipv6_l3fwd_em_route ipv6_l3fwd_em_route_array[] = {
> + {{{32, 1, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 0},
> + {{{32, 1, 2, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 1},
> + {{{32, 1, 2, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 2},
> + {{{32, 1, 2, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 3},
> + {{{32, 1, 2, 0, 0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 4},
> + {{{32, 1, 2, 0, 0, 0, 0, 5, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 5, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 5},
> + {{{32, 1, 2, 0, 0, 0, 0, 6, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 6, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 6},
> + {{{32, 1, 2, 0, 0, 0, 0, 7, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 7, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 7},
> + {{{32, 1, 2, 0, 0, 0, 0, 8, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 8, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 8},
> + {{{32, 1, 2, 0, 0, 0, 0, 9, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 9, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 9},
> + {{{32, 1, 2, 0, 0, 0, 0, 10, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 10, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 10},
> + {{{32, 1, 2, 0, 0, 0, 0, 11, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 11, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 11},
> + {{{32, 1, 2, 0, 0, 0, 0, 12, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 12, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 12},
> + {{{32, 1, 2, 0, 0, 0, 0, 13, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 13, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 13},
> + {{{32, 1, 2, 0, 0, 0, 0, 14, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 14, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 14},
> + {{{32, 1, 2, 0, 0, 0, 0, 15, 0, 0, 0, 0, 0, 0, 0, 0},
> + {32, 1, 2, 0, 0, 0, 0, 15, 0, 0, 0, 0, 0, 0, 0, 1}, 9, 9, IPPROTO_UDP}, 15},
> };
>
> struct rte_hash *ipv4_l3fwd_em_lookup_struct[NB_SOCKETS];
> @@ -405,7 +436,7 @@ populate_ipv6_few_flow_into_table(const struct rte_hash *h)
> (uint64_t)IPV6_L3FWD_EM_NUM_ROUTES);
> }
>
> -#define NUMBER_PORT_USED 4
> +#define NUMBER_PORT_USED 16
> static inline void
> populate_ipv4_many_flow_into_table(const struct rte_hash *h,
> unsigned int nr_flow)
> @@ -416,36 +447,16 @@ populate_ipv4_many_flow_into_table(const struct rte_hash *h,
> ALL_32_BITS, ALL_32_BITS} };
>
> for (i = 0; i < nr_flow; i++) {
> + uint8_t port = i % NUMBER_PORT_USED;
> struct ipv4_l3fwd_em_route entry;
> union ipv4_5tuple_host newkey;
>
> - uint8_t a = (uint8_t)
> - ((i/NUMBER_PORT_USED)%BYTE_VALUE_MAX);
> - uint8_t b = (uint8_t)
> - (((i/NUMBER_PORT_USED)/BYTE_VALUE_MAX)%BYTE_VALUE_MAX);
> - uint8_t c = (uint8_t)
> - ((i/NUMBER_PORT_USED)/(BYTE_VALUE_MAX*BYTE_VALUE_MAX));
> + uint8_t a = (uint8_t)((port + 1) % BYTE_VALUE_MAX);
>
> /* Create the ipv4 exact match flow */
> memset(&entry, 0, sizeof(entry));
> - switch (i & (NUMBER_PORT_USED - 1)) {
> - case 0:
> - entry = ipv4_l3fwd_em_route_array[0];
> - entry.key.ip_dst = RTE_IPV4(101, c, b, a);
> - break;
> - case 1:
> - entry = ipv4_l3fwd_em_route_array[1];
> - entry.key.ip_dst = RTE_IPV4(201, c, b, a);
> - break;
> - case 2:
> - entry = ipv4_l3fwd_em_route_array[2];
> - entry.key.ip_dst = RTE_IPV4(111, c, b, a);
> - break;
> - case 3:
> - entry = ipv4_l3fwd_em_route_array[3];
> - entry.key.ip_dst = RTE_IPV4(211, c, b, a);
> - break;
> - };
> + entry = ipv4_l3fwd_em_route_array[port];
> + entry.key.ip_dst = RTE_IPV4(198, 18, port, a);
> convert_ipv4_5tuple(&entry.key, &newkey);
> int32_t ret = rte_hash_add_key(h, (void *) &newkey);
>
> @@ -469,35 +480,14 @@ populate_ipv6_many_flow_into_table(const struct rte_hash *h,
> mask2 = (rte_xmm_t){.u32 = {ALL_32_BITS, ALL_32_BITS, 0, 0} };
>
> for (i = 0; i < nr_flow; i++) {
> + uint8_t port = i % NUMBER_PORT_USED;
> struct ipv6_l3fwd_em_route entry;
> union ipv6_5tuple_host newkey;
>
> - uint8_t a = (uint8_t)
> - ((i/NUMBER_PORT_USED)%BYTE_VALUE_MAX);
> - uint8_t b = (uint8_t)
> - (((i/NUMBER_PORT_USED)/BYTE_VALUE_MAX)%BYTE_VALUE_MAX);
> - uint8_t c = (uint8_t)
> - ((i/NUMBER_PORT_USED)/(BYTE_VALUE_MAX*BYTE_VALUE_MAX));
> -
> /* Create the ipv6 exact match flow */
> memset(&entry, 0, sizeof(entry));
> - switch (i & (NUMBER_PORT_USED - 1)) {
> - case 0:
> - entry = ipv6_l3fwd_em_route_array[0];
> - break;
> - case 1:
> - entry = ipv6_l3fwd_em_route_array[1];
> - break;
> - case 2:
> - entry = ipv6_l3fwd_em_route_array[2];
> - break;
> - case 3:
> - entry = ipv6_l3fwd_em_route_array[3];
> - break;
> - };
> - entry.key.ip_dst[13] = c;
> - entry.key.ip_dst[14] = b;
> - entry.key.ip_dst[15] = a;
> + entry = ipv6_l3fwd_em_route_array[port];
> + entry.key.ip_dst[15] = (port + 1) % BYTE_VALUE_MAX;
> convert_ipv6_5tuple(&entry.key, &newkey);
> int32_t ret = rte_hash_add_key(h, (void *) &newkey);
>
> @@ -584,7 +574,7 @@ em_parse_ptype(struct rte_mbuf *m)
> hdr_len = rte_ipv4_hdr_len(ipv4_hdr);
> if (hdr_len == sizeof(struct rte_ipv4_hdr)) {
> packet_type |= RTE_PTYPE_L3_IPV4;
> - if (ipv4_hdr->next_proto_id == IPPROTO_TCP)
> + if (ipv4_hdr->next_proto_id == IPPROTO_UDP)
I don't understand the intention here.
> packet_type |= RTE_PTYPE_L4_TCP;
> else if (ipv4_hdr->next_proto_id == IPPROTO_UDP)
> packet_type |= RTE_PTYPE_L4_UDP;
> @@ -592,7 +582,7 @@ em_parse_ptype(struct rte_mbuf *m)
> packet_type |= RTE_PTYPE_L3_IPV4_EXT;
> } else if (ether_type == rte_cpu_to_be_16(RTE_ETHER_TYPE_IPV6)) {
> ipv6_hdr = (struct rte_ipv6_hdr *)l3;
> - if (ipv6_hdr->proto == IPPROTO_TCP)
> + if (ipv6_hdr->proto == IPPROTO_UDP)
> packet_type |= RTE_PTYPE_L3_IPV6 | RTE_PTYPE_L4_TCP;
Idem.
> else if (ipv6_hdr->proto == IPPROTO_UDP)
> packet_type |= RTE_PTYPE_L3_IPV6 | RTE_PTYPE_L4_UDP;
> --
> 2.17.1
>
--
David Marchand
next prev parent reply other threads:[~2021-11-03 20:53 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-30 5:23 [dpdk-dev] [PATCH 1/2] examples/l3fwd: increase number of routes pbhagavatula
2019-10-30 5:23 ` [dpdk-dev] [PATCH 2/2] examples/l3fwd: print port bdf when initializing routes pbhagavatula
2019-11-08 8:51 ` [dpdk-dev] [PATCH 1/2] examples/l3fwd: increase number of routes David Marchand
2020-08-02 18:16 ` [dpdk-dev] [PATCH v2 1/3] " pbhagavatula
2020-08-02 18:16 ` [dpdk-dev] [PATCH v2 2/3] examples/l3fwd: use IP reserved addresses for EM mode pbhagavatula
2020-10-23 14:11 ` David Marchand
2020-08-02 18:16 ` [dpdk-dev] [PATCH v2 3/3] examples/l3fwd: print port bdf when initializing routes pbhagavatula
2021-10-03 20:11 ` [dpdk-dev] [PATCH 1/3] examples/l3fwd: increase number of routes pbhagavatula
2021-10-03 20:11 ` [dpdk-dev] [PATCH 2/3] examples/l3fwd: print port bdf when adding routes pbhagavatula
2021-11-04 10:28 ` Walsh, Conor
2021-11-05 7:27 ` Pavan Nikhilesh Bhagavatula
2021-10-03 20:11 ` [dpdk-dev] [PATCH 3/3] examples/l3fwd: use reserved addresses for EM mode pbhagavatula
2021-10-03 20:24 ` [dpdk-dev] [PATCH v4 1/3] examples/l3fwd: increase number of routes pbhagavatula
2021-10-03 20:24 ` [dpdk-dev] [PATCH v4 2/3] examples/l3fwd: print port bdf when adding routes pbhagavatula
2021-10-03 20:24 ` [dpdk-dev] [PATCH v4 3/3] examples/l3fwd: use reserved addresses for EM mode pbhagavatula
2021-11-03 20:53 ` David Marchand [this message]
2021-11-05 7:30 ` [dpdk-dev] [EXT] " Pavan Nikhilesh Bhagavatula
2021-11-04 10:28 ` [dpdk-dev] [PATCH v4 1/3] examples/l3fwd: increase number of routes Walsh, Conor
2021-11-05 7:27 ` Pavan Nikhilesh Bhagavatula
2021-11-05 8:38 ` [dpdk-dev] [PATCH v5 " pbhagavatula
2021-11-05 8:38 ` [dpdk-dev] [PATCH v5 2/3] examples/l3fwd: print port bdf when adding routes pbhagavatula
2021-11-05 9:34 ` Walsh, Conor
2021-11-05 8:38 ` [dpdk-dev] [PATCH v5 3/3] examples/l3fwd: use reserved addresses for EM mode pbhagavatula
2021-11-16 10:03 ` David Marchand
2021-11-05 9:28 ` [dpdk-dev] [PATCH v5 1/3] examples/l3fwd: increase number of routes Walsh, Conor
2021-10-04 15:46 ` [dpdk-dev] [PATCH " Stephen Hemminger
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAJFAV8ytxybikS51z9+0E+EGZjQc5xZgo1yHTFtPZGUAy7qw8w@mail.gmail.com \
--to=david.marchand@redhat.com \
--cc=dev@dpdk.org \
--cc=jerinj@marvell.com \
--cc=konstantin.ananyev@intel.com \
--cc=pbhagavatula@marvell.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).