From 732115ce38c63184cb706b9179c02ed04b961afa Mon Sep 17 00:00:00 2001 From: Rahul Bhansali Date: Thu, 23 Jun 2022 15:08:15 +0530 Subject: [PATCH] examples/l3fwd: move packet group function in common 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/. Signed-off-by: Rahul Bhansali Tested-by: Konstantin Ananyev Acked-by: Konstantin Ananyev --- MAINTAINERS | 3 + examples/common/altivec/port_group.h | 48 +++++++++ examples/common/neon/port_group.h | 50 ++++++++++ examples/common/pkt_group.h | 139 +++++++++++++++++++++++++++ examples/common/sse/port_group.h | 47 +++++++++ examples/l3fwd/Makefile | 2 + examples/l3fwd/l3fwd.h | 2 - examples/l3fwd/l3fwd_altivec.h | 37 +------ examples/l3fwd/l3fwd_common.h | 129 +------------------------ examples/l3fwd/l3fwd_neon.h | 39 +------- examples/l3fwd/l3fwd_sse.h | 36 +------ examples/meson.build | 2 +- 12 files changed, 295 insertions(+), 239 deletions(-) create mode 100644 examples/common/altivec/port_group.h create mode 100644 examples/common/neon/port_group.h create mode 100644 examples/common/pkt_group.h create mode 100644 examples/common/sse/port_group.h diff --git a/MAINTAINERS b/MAINTAINERS index bfeeb7d1b4..c923712946 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -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 @@ -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 @@ -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/ diff --git a/examples/common/altivec/port_group.h b/examples/common/altivec/port_group.h new file mode 100644 index 0000000000..5e209b02fa --- /dev/null +++ b/examples/common/altivec/port_group.h @@ -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: , dp2: . + * 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 */ diff --git a/examples/common/neon/port_group.h b/examples/common/neon/port_group.h new file mode 100644 index 0000000000..c4861d1b7e --- /dev/null +++ b/examples/common/neon/port_group.h @@ -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: , dp2: . + * 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 */ diff --git a/examples/common/pkt_group.h b/examples/common/pkt_group.h new file mode 100644 index 0000000000..12dfe48250 --- /dev/null +++ b/examples/common/pkt_group.h @@ -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 */ diff --git a/examples/common/sse/port_group.h b/examples/common/sse/port_group.h new file mode 100644 index 0000000000..cf9ec0913b --- /dev/null +++ b/examples/common/sse/port_group.h @@ -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: , dp2: . + * 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 */ diff --git a/examples/l3fwd/Makefile b/examples/l3fwd/Makefile index 8efe6378e2..e802f037be 100644 --- a/examples/l3fwd/Makefile +++ b/examples/l3fwd/Makefile @@ -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) diff --git a/examples/l3fwd/l3fwd.h b/examples/l3fwd/l3fwd.h index 8a52c90755..40b5f32a9e 100644 --- a/examples/l3fwd/l3fwd.h +++ b/examples/l3fwd/l3fwd.h @@ -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 diff --git a/examples/l3fwd/l3fwd_altivec.h b/examples/l3fwd/l3fwd_altivec.h index 88fb41843b..87018f5dbe 100644 --- a/examples/l3fwd/l3fwd_altivec.h +++ b/examples/l3fwd/l3fwd_altivec.h @@ -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: , dp2: . - * 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. diff --git a/examples/l3fwd/l3fwd_common.h b/examples/l3fwd/l3fwd_common.h index 8e4c27218f..224b1c08e8 100644 --- a/examples/l3fwd/l3fwd_common.h +++ b/examples/l3fwd/l3fwd_common.h @@ -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) diff --git a/examples/l3fwd/l3fwd_neon.h b/examples/l3fwd/l3fwd_neon.h index e3d33a5229..ce515e0bc4 100644 --- a/examples/l3fwd/l3fwd_neon.h +++ b/examples/l3fwd/l3fwd_neon.h @@ -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: , dp2: . - * 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. diff --git a/examples/l3fwd/l3fwd_sse.h b/examples/l3fwd/l3fwd_sse.h index d5a717e18c..0f0d0323a2 100644 --- a/examples/l3fwd/l3fwd_sse.h +++ b/examples/l3fwd/l3fwd_sse.h @@ -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: , dp2: . - * 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. diff --git a/examples/meson.build b/examples/meson.build index 78de0e1f37..81e93799f2 100644 --- a/examples/meson.build +++ b/examples/meson.build @@ -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) -- 2.39.5