d/numam-dpdk
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

examples/l3fwd: move packet group function in common

Browse Source 
This will make the packet grouping function common, so
that other examples can utilize as per need.

For each architecture sse/neon/altivec, port group
headers will be created under examples/common/<arch>.

Signed-off-by: Rahul Bhansali <rbhansali@marvell.com>
Tested-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
Acked-by: Konstantin Ananyev <konstantin.v.ananyev@yandex.ru>
This commit is contained in:
Rahul Bhansali 2022-06-23 15:08:15 +05:30 committed by Akhil Goyal
parent 65bd9c7abc
commit 732115ce38
12 changed files with 295 additions and 239 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
MAINTAINERS
View File

@ -293,6 +293,7 @@ F: lib/*/*_neon.*
F: drivers/*/*/*_neon.*
F: app/*/*_neon.*
F: examples/*/*_neon.*
F: examples/common/neon/
IBM POWER (alpha)
M: David Christensen <drc@linux.vnet.ibm.com>
@ -302,6 +303,7 @@ F: lib/*/*_altivec*
F: drivers/*/*/*_altivec.*
F: app/*/*_altivec.*
F: examples/*/*_altivec.*
F: examples/common/altivec/
RISC-V
M: Stanislaw Kardach <kda@semihalf.com>
@ -325,6 +327,7 @@ F: app/*/*_sse*
F: app/*/*_avx*
F: examples/*/*_sse*
F: examples/*/*_avx*
F: examples/common/sse/
Linux EAL (with overlaps)
F: lib/eal/linux/

48
examples/common/altivec/port_group.h Normal file
View File

@ -0,0 +1,48 @@
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright(c) 2016 Intel Corporation.
* Copyright(c) 2017 IBM Corporation.
* Copyright(C) 2022 Marvell.
*/
#ifndef PORT_GROUP_H
#define PORT_GROUP_H
#include "pkt_group.h"
/*
* Group consecutive packets with the same destination port in bursts of 4.
* Suppose we have array of destination ports:
* dst_port[] = {a, b, c, d,, e, ... }
* dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
* We doing 4 comparisons at once and the result is 4 bit mask.
* This mask is used as an index into prebuild array of pnum values.
*/
static inline uint16_t *
port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp,
__vector unsigned short dp1,
__vector unsigned short dp2)
{
union {
uint16_t u16[FWDSTEP + 1];
uint64_t u64;
} *pnum = (void *)pn;
int32_t v;
v = vec_any_eq(dp1, dp2);
/* update last port counter. */
lp[0] += gptbl[v].lpv;
/* if dest port value has changed. */
if (v != GRPMSK) {
pnum->u64 = gptbl[v].pnum;
pnum->u16[FWDSTEP] = 1;
lp = pnum->u16 + gptbl[v].idx;
}
return lp;
}
#endif /* PORT_GROUP_H */

50
examples/common/neon/port_group.h Normal file
View File

@ -0,0 +1,50 @@
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright(c) 2016-2018 Intel Corporation.
* Copyright(c) 2017-2018 Linaro Limited.
* Copyright(C) 2022 Marvell.
*/
#ifndef PORT_GROUP_H
#define PORT_GROUP_H
#include "pkt_group.h"
/*
* Group consecutive packets with the same destination port in bursts of 4.
* Suppose we have array of destination ports:
* dst_port[] = {a, b, c, d,, e, ... }
* dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
* We doing 4 comparisons at once and the result is 4 bit mask.
* This mask is used as an index into prebuild array of pnum values.
*/
static inline uint16_t *
port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp, uint16x8_t dp1,
uint16x8_t dp2)
{
union {
uint16_t u16[FWDSTEP + 1];
uint64_t u64;
} *pnum = (void *)pn;
uint16x8_t mask = {1, 2, 4, 8, 0, 0, 0, 0};
int32_t v;
dp1 = vceqq_u16(dp1, dp2);
dp1 = vandq_u16(dp1, mask);
v = vaddvq_u16(dp1);
/* update last port counter. */
lp[0] += gptbl[v].lpv;
rte_compiler_barrier();
/* if dest port value has changed. */
if (v != GRPMSK) {
pnum->u64 = gptbl[v].pnum;
pnum->u16[FWDSTEP] = 1;
lp = pnum->u16 + gptbl[v].idx;
}
return lp;
}
#endif /* PORT_GROUP_H */

139
examples/common/pkt_group.h Normal file
View File

@ -0,0 +1,139 @@
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright(c) 2016-2018 Intel Corporation.
* Copyright(c) 2017-2018 Linaro Limited.
* Copyright(C) 2022 Marvell.
*/
#ifndef PKT_GROUP_H
#define PKT_GROUP_H
#define FWDSTEP 4
/*
* Group consecutive packets with the same destination port into one burst.
* To avoid extra latency this is done together with some other packet
* processing, but after we made a final decision about packet's destination.
* To do this we maintain:
* pnum - array of number of consecutive packets with the same dest port for
* each packet in the input burst.
* lp - pointer to the last updated element in the pnum.
* dlp - dest port value lp corresponds to.
*/
#define GRPSZ (1 << FWDSTEP)
#define GRPMSK (GRPSZ - 1)
#define GROUP_PORT_STEP(dlp, dcp, lp, pn, idx) do { \
if (likely((dlp) == (dcp)[(idx)])) { \
(lp)[0]++; \
} else { \
(dlp) = (dcp)[idx]; \
(lp) = (pn) + (idx); \
(lp)[0] = 1; \
} \
} while (0)
static const struct {
uint64_t pnum; /* prebuild 4 values for pnum[]. */
int32_t idx; /* index for new last updated elemnet. */
uint16_t lpv; /* add value to the last updated element. */
} gptbl[GRPSZ] = {
{
/* 0: a != b, b != c, c != d, d != e */
.pnum = UINT64_C(0x0001000100010001),
.idx = 4,
.lpv = 0,
},
{
/* 1: a == b, b != c, c != d, d != e */
.pnum = UINT64_C(0x0001000100010002),
.idx = 4,
.lpv = 1,
},
{
/* 2: a != b, b == c, c != d, d != e */
.pnum = UINT64_C(0x0001000100020001),
.idx = 4,
.lpv = 0,
},
{
/* 3: a == b, b == c, c != d, d != e */
.pnum = UINT64_C(0x0001000100020003),
.idx = 4,
.lpv = 2,
},
{
/* 4: a != b, b != c, c == d, d != e */
.pnum = UINT64_C(0x0001000200010001),
.idx = 4,
.lpv = 0,
},
{
/* 5: a == b, b != c, c == d, d != e */
.pnum = UINT64_C(0x0001000200010002),
.idx = 4,
.lpv = 1,
},
{
/* 6: a != b, b == c, c == d, d != e */
.pnum = UINT64_C(0x0001000200030001),
.idx = 4,
.lpv = 0,
},
{
/* 7: a == b, b == c, c == d, d != e */
.pnum = UINT64_C(0x0001000200030004),
.idx = 4,
.lpv = 3,
},
{
/* 8: a != b, b != c, c != d, d == e */
.pnum = UINT64_C(0x0002000100010001),
.idx = 3,
.lpv = 0,
},
{
/* 9: a == b, b != c, c != d, d == e */
.pnum = UINT64_C(0x0002000100010002),
.idx = 3,
.lpv = 1,
},
{
/* 0xa: a != b, b == c, c != d, d == e */
.pnum = UINT64_C(0x0002000100020001),
.idx = 3,
.lpv = 0,
},
{
/* 0xb: a == b, b == c, c != d, d == e */
.pnum = UINT64_C(0x0002000100020003),
.idx = 3,
.lpv = 2,
},
{
/* 0xc: a != b, b != c, c == d, d == e */
.pnum = UINT64_C(0x0002000300010001),
.idx = 2,
.lpv = 0,
},
{
/* 0xd: a == b, b != c, c == d, d == e */
.pnum = UINT64_C(0x0002000300010002),
.idx = 2,
.lpv = 1,
},
{
/* 0xe: a != b, b == c, c == d, d == e */
.pnum = UINT64_C(0x0002000300040001),
.idx = 1,
.lpv = 0,
},
{
/* 0xf: a == b, b == c, c == d, d == e */
.pnum = UINT64_C(0x0002000300040005),
.idx = 0,
.lpv = 4,
},
};
#endif /* PKT_GROUP_H */

47
examples/common/sse/port_group.h Normal file
View File

@ -0,0 +1,47 @@
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright(c) 2016 Intel Corporation.
* Copyright(C) 2022 Marvell.
*/
#ifndef PORT_GROUP_H
#define PORT_GROUP_H
#include "pkt_group.h"
/*
* Group consecutive packets with the same destination port in bursts of 4.
* Suppose we have array of destination ports:
* dst_port[] = {a, b, c, d,, e, ... }
* dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
* We doing 4 comparisons at once and the result is 4 bit mask.
* This mask is used as an index into prebuild array of pnum values.
*/
static inline uint16_t *
port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp, __m128i dp1,
__m128i dp2)
{
union {
uint16_t u16[FWDSTEP + 1];
uint64_t u64;
} *pnum = (void *)pn;
int32_t v;
dp1 = _mm_cmpeq_epi16(dp1, dp2);
dp1 = _mm_unpacklo_epi16(dp1, dp1);
v = _mm_movemask_ps((__m128)dp1);
/* update last port counter. */
lp[0] += gptbl[v].lpv;
/* if dest port value has changed. */
if (v != GRPMSK) {
pnum->u64 = gptbl[v].pnum;
pnum->u16[FWDSTEP] = 1;
lp = pnum->u16 + gptbl[v].idx;
}
return lp;
}
#endif /* PORT_GROUP_H */

2
examples/l3fwd/Makefile
View File

@ -37,6 +37,8 @@ $(error "Cannot generate statically-linked binaries with this version of pkg-con
endif
endif
CFLAGS += -I../common
build/$(APP)-shared: $(SRCS-y) Makefile $(PC_FILE) | build
$(CC) $(CFLAGS) $(SRCS-y) -o $@ $(LDFLAGS) $(LDFLAGS_SHARED)

2
examples/l3fwd/l3fwd.h
View File

@ -44,8 +44,6 @@
/* Used to mark destination port as 'invalid'. */
#define BAD_PORT ((uint16_t)-1)
#define FWDSTEP 4
/* replace first 12B of the ethernet header. */
#define MASK_ETH 0x3f

37
examples/l3fwd/l3fwd_altivec.h
View File

@ -8,6 +8,7 @@
#define _L3FWD_ALTIVEC_H_
#include "l3fwd.h"
#include "altivec/port_group.h"
#include "l3fwd_common.h"
/*
@ -82,42 +83,6 @@ processx4_step3(struct rte_mbuf *pkt[FWDSTEP], uint16_t dst_port[FWDSTEP])
&dst_port[3], pkt[3]->packet_type);
}
/*
* Group consecutive packets with the same destination port in bursts of 4.
* Suppose we have array of destination ports:
* dst_port[] = {a, b, c, d,, e, ... }
* dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
* We doing 4 comparisons at once and the result is 4 bit mask.
* This mask is used as an index into prebuild array of pnum values.
*/
static inline uint16_t *
port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp,
__vector unsigned short dp1,
__vector unsigned short dp2)
{
union {
uint16_t u16[FWDSTEP + 1];
uint64_t u64;
} *pnum = (void *)pn;
int32_t v;
v = vec_any_eq(dp1, dp2);
/* update last port counter. */
lp[0] += gptbl[v].lpv;
/* if dest port value has changed. */
if (v != GRPMSK) {
pnum->u64 = gptbl[v].pnum;
pnum->u16[FWDSTEP] = 1;
lp = pnum->u16 + gptbl[v].idx;
}
return lp;
}
/**
* Process one packet:
* Update source and destination MAC addresses in the ethernet header.

129
examples/l3fwd/l3fwd_common.h
View File

@ -7,6 +7,8 @@
#ifndef _L3FWD_COMMON_H_
#define _L3FWD_COMMON_H_
#include "pkt_group.h"
#ifdef DO_RFC_1812_CHECKS
#define IPV4_MIN_VER_IHL 0x45
@ -50,133 +52,6 @@ rfc1812_process(struct rte_ipv4_hdr *ipv4_hdr, uint16_t *dp, uint32_t ptype)
#define rfc1812_process(mb, dp, ptype) do { } while (0)
#endif /* DO_RFC_1812_CHECKS */
/*
* We group consecutive packets with the same destination port into one burst.
* To avoid extra latency this is done together with some other packet
* processing, but after we made a final decision about packet's destination.
* To do this we maintain:
* pnum - array of number of consecutive packets with the same dest port for
* each packet in the input burst.
* lp - pointer to the last updated element in the pnum.
* dlp - dest port value lp corresponds to.
*/
#define GRPSZ (1 << FWDSTEP)
#define GRPMSK (GRPSZ - 1)
#define GROUP_PORT_STEP(dlp, dcp, lp, pn, idx) do { \
if (likely((dlp) == (dcp)[(idx)])) { \
(lp)[0]++; \
} else { \
(dlp) = (dcp)[idx]; \
(lp) = (pn) + (idx); \
(lp)[0] = 1; \
} \
} while (0)
static const struct {
uint64_t pnum; /* prebuild 4 values for pnum[]. */
int32_t idx; /* index for new last updated element. */
uint16_t lpv; /* add value to the last updated element. */
} gptbl[GRPSZ] = {
{
/* 0: a != b, b != c, c != d, d != e */
.pnum = UINT64_C(0x0001000100010001),
.idx = 4,
.lpv = 0,
},
{
/* 1: a == b, b != c, c != d, d != e */
.pnum = UINT64_C(0x0001000100010002),
.idx = 4,
.lpv = 1,
},
{
/* 2: a != b, b == c, c != d, d != e */
.pnum = UINT64_C(0x0001000100020001),
.idx = 4,
.lpv = 0,
},
{
/* 3: a == b, b == c, c != d, d != e */
.pnum = UINT64_C(0x0001000100020003),
.idx = 4,
.lpv = 2,
},
{
/* 4: a != b, b != c, c == d, d != e */
.pnum = UINT64_C(0x0001000200010001),
.idx = 4,
.lpv = 0,
},
{
/* 5: a == b, b != c, c == d, d != e */
.pnum = UINT64_C(0x0001000200010002),
.idx = 4,
.lpv = 1,
},
{
/* 6: a != b, b == c, c == d, d != e */
.pnum = UINT64_C(0x0001000200030001),
.idx = 4,
.lpv = 0,
},
{
/* 7: a == b, b == c, c == d, d != e */
.pnum = UINT64_C(0x0001000200030004),
.idx = 4,
.lpv = 3,
},
{
/* 8: a != b, b != c, c != d, d == e */
.pnum = UINT64_C(0x0002000100010001),
.idx = 3,
.lpv = 0,
},
{
/* 9: a == b, b != c, c != d, d == e */
.pnum = UINT64_C(0x0002000100010002),
.idx = 3,
.lpv = 1,
},
{
/* 0xa: a != b, b == c, c != d, d == e */
.pnum = UINT64_C(0x0002000100020001),
.idx = 3,
.lpv = 0,
},
{
/* 0xb: a == b, b == c, c != d, d == e */
.pnum = UINT64_C(0x0002000100020003),
.idx = 3,
.lpv = 2,
},
{
/* 0xc: a != b, b != c, c == d, d == e */
.pnum = UINT64_C(0x0002000300010001),
.idx = 2,
.lpv = 0,
},
{
/* 0xd: a == b, b != c, c == d, d == e */
.pnum = UINT64_C(0x0002000300010002),
.idx = 2,
.lpv = 1,
},
{
/* 0xe: a != b, b == c, c == d, d == e */
.pnum = UINT64_C(0x0002000300040001),
.idx = 1,
.lpv = 0,
},
{
/* 0xf: a == b, b == c, c == d, d == e */
.pnum = UINT64_C(0x0002000300040005),
.idx = 0,
.lpv = 4,
},
};
static __rte_always_inline void
send_packetsx4(struct lcore_conf *qconf, uint16_t port, struct rte_mbuf *m[],
uint32_t num)

39
examples/l3fwd/l3fwd_neon.h
View File

@ -7,6 +7,7 @@
#define _L3FWD_NEON_H_
#include "l3fwd.h"
#include "neon/port_group.h"
#include "l3fwd_common.h"
/*
@ -62,44 +63,6 @@ processx4_step3(struct rte_mbuf *pkt[FWDSTEP], uint16_t dst_port[FWDSTEP])
&dst_port[3], pkt[3]->packet_type);
}
/*
* Group consecutive packets with the same destination port in bursts of 4.
* Suppose we have array of destination ports:
* dst_port[] = {a, b, c, d,, e, ... }
* dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
* We doing 4 comparisons at once and the result is 4 bit mask.
* This mask is used as an index into prebuild array of pnum values.
*/
static inline uint16_t *
port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp, uint16x8_t dp1,
uint16x8_t dp2)
{
union {
uint16_t u16[FWDSTEP + 1];
uint64_t u64;
} *pnum = (void *)pn;
int32_t v;
uint16x8_t mask = {1, 2, 4, 8, 0, 0, 0, 0};
dp1 = vceqq_u16(dp1, dp2);
dp1 = vandq_u16(dp1, mask);
v = vaddvq_u16(dp1);
/* update last port counter. */
lp[0] += gptbl[v].lpv;
rte_compiler_barrier();
/* if dest port value has changed. */
if (v != GRPMSK) {
pnum->u64 = gptbl[v].pnum;
pnum->u16[FWDSTEP] = 1;
lp = pnum->u16 + gptbl[v].idx;
}
return lp;
}
/**
* Process one packet:
* Update source and destination MAC addresses in the ethernet header.

36
examples/l3fwd/l3fwd_sse.h
View File

@ -7,6 +7,7 @@
#define _L3FWD_SSE_H_
#include "l3fwd.h"
#include "sse/port_group.h"
#include "l3fwd_common.h"
/*
@ -62,41 +63,6 @@ processx4_step3(struct rte_mbuf *pkt[FWDSTEP], uint16_t dst_port[FWDSTEP])
&dst_port[3], pkt[3]->packet_type);
}
/*
* Group consecutive packets with the same destination port in bursts of 4.
* Suppose we have array of destination ports:
* dst_port[] = {a, b, c, d,, e, ... }
* dp1 should contain: <a, b, c, d>, dp2: <b, c, d, e>.
* We doing 4 comparisons at once and the result is 4 bit mask.
* This mask is used as an index into prebuild array of pnum values.
*/
static inline uint16_t *
port_groupx4(uint16_t pn[FWDSTEP + 1], uint16_t *lp, __m128i dp1, __m128i dp2)
{
union {
uint16_t u16[FWDSTEP + 1];
uint64_t u64;
} *pnum = (void *)pn;
int32_t v;
dp1 = _mm_cmpeq_epi16(dp1, dp2);
dp1 = _mm_unpacklo_epi16(dp1, dp1);
v = _mm_movemask_ps((__m128)dp1);
/* update last port counter. */
lp[0] += gptbl[v].lpv;
/* if dest port value has changed. */
if (v != GRPMSK) {
pnum->u64 = gptbl[v].pnum;
pnum->u16[FWDSTEP] = 1;
lp = pnum->u16 + gptbl[v].idx;
}
return lp;
}
/**
* Process one packet:
* Update source and destination MAC addresses in the ethernet header.

2
examples/meson.build
View File

@ -97,7 +97,7 @@ foreach example: examples
ldflags = default_ldflags
ext_deps = []
includes = [include_directories(example)]
includes = [include_directories(example, 'common')]
deps = ['eal', 'mempool', 'net', 'mbuf', 'ethdev', 'cmdline']
subdir(example)