From e6ba4731c0f3a0aa27873ec301d5e9305f1986fd Mon Sep 17 00:00:00 2001 From: Konstantin Ananyev Date: Mon, 20 Apr 2020 13:28:24 +0100 Subject: [PATCH] ring: introduce RTS ring mode Introduce relaxed tail sync (RTS) mode for MT ring synchronization. Aim to reduce stall times in case when ring is used on overcommited cpus (multiple active threads on the same cpu). The main difference from original MP/MC algorithm is that tail value is increased not by every thread that finished enqueue/dequeue, but only by the last one. That allows threads to avoid spinning on ring tail value, leaving actual tail value change to the last thread in the update queue. Signed-off-by: Konstantin Ananyev Acked-by: Honnappa Nagarahalli --- devtools/libabigail.abignore | 7 + doc/guides/prog_guide/ring_lib.rst | 16 + doc/guides/rel_notes/release_20_05.rst | 7 + lib/librte_ring/Makefile | 4 +- lib/librte_ring/meson.build | 4 +- lib/librte_ring/rte_ring.c | 100 +++++- lib/librte_ring/rte_ring.h | 118 +++++-- lib/librte_ring/rte_ring_core.h | 36 +- lib/librte_ring/rte_ring_elem.h | 114 ++++++- lib/librte_ring/rte_ring_rts.h | 439 +++++++++++++++++++++++++ lib/librte_ring/rte_ring_rts_c11_mem.h | 179 ++++++++++ 11 files changed, 976 insertions(+), 48 deletions(-) create mode 100644 lib/librte_ring/rte_ring_rts.h create mode 100644 lib/librte_ring/rte_ring_rts_c11_mem.h diff --git a/devtools/libabigail.abignore b/devtools/libabigail.abignore index a59df8f135..cd86d89cad 100644 --- a/devtools/libabigail.abignore +++ b/devtools/libabigail.abignore @@ -11,3 +11,10 @@ type_kind = enum name = rte_crypto_asym_xform_type changed_enumerators = RTE_CRYPTO_ASYM_XFORM_TYPE_LIST_END +; Ignore updates of ring prod/cons +[suppress_type] + type_kind = struct + name = rte_ring +[suppress_type] + type_kind = struct + name = rte_event_ring diff --git a/doc/guides/prog_guide/ring_lib.rst b/doc/guides/prog_guide/ring_lib.rst index d1ac475979..de4661aafe 100644 --- a/doc/guides/prog_guide/ring_lib.rst +++ b/doc/guides/prog_guide/ring_lib.rst @@ -374,6 +374,22 @@ SP/SC Single-producer (/single-consumer) mode. In this mode only one thread at a time is allowed to enqueue (/dequeue) objects to (/from) the ring. +MP_RTS/MC_RTS +~~~~~~~~~~~~~ + +Multi-producer (/multi-consumer) with Relaxed Tail Sync (RTS) mode. +The main difference from the original MP/MC algorithm is that +tail value is increased not by every thread that finished enqueue/dequeue, +but only by the last one. +That allows threads to avoid spinning on ring tail value, +leaving actual tail value change to the last thread at a given instance. +That technique helps to avoid the Lock-Waiter-Preemption (LWP) problem on tail +update and improves average enqueue/dequeue times on overcommitted systems. +To achieve that RTS requires 2 64-bit CAS for each enqueue(/dequeue) operation: +one for head update, second for tail update. +In comparison the original MP/MC algorithm requires one 32-bit CAS +for head update and waiting/spinning on tail value. + References ---------- diff --git a/doc/guides/rel_notes/release_20_05.rst b/doc/guides/rel_notes/release_20_05.rst index 1849678444..60bb4eb637 100644 --- a/doc/guides/rel_notes/release_20_05.rst +++ b/doc/guides/rel_notes/release_20_05.rst @@ -56,6 +56,13 @@ New Features Also, make sure to start the actual text at the margin. ========================================================= +* **New synchronization modes for rte_ring.** + + Introduced new optional MT synchronization mode for rte_ring: + Relaxed Tail Sync (RTS). With this mode selected, rte_ring shows + significant improvements for average enqueue/dequeue times on + overcommitted systems. + * **Updated Mellanox mlx5 driver.** Updated Mellanox mlx5 driver with new features and improvements, including: diff --git a/lib/librte_ring/Makefile b/lib/librte_ring/Makefile index 6572768c9e..04e446e379 100644 --- a/lib/librte_ring/Makefile +++ b/lib/librte_ring/Makefile @@ -19,6 +19,8 @@ SYMLINK-$(CONFIG_RTE_LIBRTE_RING)-include := rte_ring.h \ rte_ring_core.h \ rte_ring_elem.h \ rte_ring_generic.h \ - rte_ring_c11_mem.h + rte_ring_c11_mem.h \ + rte_ring_rts.h \ + rte_ring_rts_c11_mem.h include $(RTE_SDK)/mk/rte.lib.mk diff --git a/lib/librte_ring/meson.build b/lib/librte_ring/meson.build index c656781dad..42b98c6818 100644 --- a/lib/librte_ring/meson.build +++ b/lib/librte_ring/meson.build @@ -6,4 +6,6 @@ headers = files('rte_ring.h', 'rte_ring_core.h', 'rte_ring_elem.h', 'rte_ring_c11_mem.h', - 'rte_ring_generic.h') + 'rte_ring_generic.h', + 'rte_ring_rts.h', + 'rte_ring_rts_c11_mem.h') diff --git a/lib/librte_ring/rte_ring.c b/lib/librte_ring/rte_ring.c index fa57339074..222eec0fb6 100644 --- a/lib/librte_ring/rte_ring.c +++ b/lib/librte_ring/rte_ring.c @@ -45,6 +45,9 @@ EAL_REGISTER_TAILQ(rte_ring_tailq) /* true if x is a power of 2 */ #define POWEROF2(x) ((((x)-1) & (x)) == 0) +/* by default set head/tail distance as 1/8 of ring capacity */ +#define HTD_MAX_DEF 8 + /* return the size of memory occupied by a ring */ ssize_t rte_ring_get_memsize_elem(unsigned int esize, unsigned int count) @@ -79,11 +82,84 @@ rte_ring_get_memsize(unsigned int count) return rte_ring_get_memsize_elem(sizeof(void *), count); } +/* + * internal helper function to reset prod/cons head-tail values. + */ +static void +reset_headtail(void *p) +{ + struct rte_ring_headtail *ht; + struct rte_ring_rts_headtail *ht_rts; + + ht = p; + ht_rts = p; + + switch (ht->sync_type) { + case RTE_RING_SYNC_MT: + case RTE_RING_SYNC_ST: + ht->head = 0; + ht->tail = 0; + break; + case RTE_RING_SYNC_MT_RTS: + ht_rts->head.raw = 0; + ht_rts->tail.raw = 0; + break; + default: + /* unknown sync mode */ + RTE_ASSERT(0); + } +} + void rte_ring_reset(struct rte_ring *r) { - r->prod.head = r->cons.head = 0; - r->prod.tail = r->cons.tail = 0; + reset_headtail(&r->prod); + reset_headtail(&r->cons); +} + +/* + * helper function, calculates sync_type values for prod and cons + * based on input flags. Returns zero at success or negative + * errno value otherwise. + */ +static int +get_sync_type(uint32_t flags, enum rte_ring_sync_type *prod_st, + enum rte_ring_sync_type *cons_st) +{ + static const uint32_t prod_st_flags = + (RING_F_SP_ENQ | RING_F_MP_RTS_ENQ); + static const uint32_t cons_st_flags = + (RING_F_SC_DEQ | RING_F_MC_RTS_DEQ); + + switch (flags & prod_st_flags) { + case 0: + *prod_st = RTE_RING_SYNC_MT; + break; + case RING_F_SP_ENQ: + *prod_st = RTE_RING_SYNC_ST; + break; + case RING_F_MP_RTS_ENQ: + *prod_st = RTE_RING_SYNC_MT_RTS; + break; + default: + return -EINVAL; + } + + switch (flags & cons_st_flags) { + case 0: + *cons_st = RTE_RING_SYNC_MT; + break; + case RING_F_SC_DEQ: + *cons_st = RTE_RING_SYNC_ST; + break; + case RING_F_MC_RTS_DEQ: + *cons_st = RTE_RING_SYNC_MT_RTS; + break; + default: + return -EINVAL; + } + + return 0; } int @@ -100,16 +176,20 @@ rte_ring_init(struct rte_ring *r, const char *name, unsigned count, RTE_BUILD_BUG_ON((offsetof(struct rte_ring, prod) & RTE_CACHE_LINE_MASK) != 0); + RTE_BUILD_BUG_ON(offsetof(struct rte_ring_headtail, sync_type) != + offsetof(struct rte_ring_rts_headtail, sync_type)); + RTE_BUILD_BUG_ON(offsetof(struct rte_ring_headtail, tail) != + offsetof(struct rte_ring_rts_headtail, tail.val.pos)); + /* init the ring structure */ memset(r, 0, sizeof(*r)); ret = strlcpy(r->name, name, sizeof(r->name)); if (ret < 0 || ret >= (int)sizeof(r->name)) return -ENAMETOOLONG; r->flags = flags; - r->prod.sync_type = (flags & RING_F_SP_ENQ) ? - RTE_RING_SYNC_ST : RTE_RING_SYNC_MT; - r->cons.sync_type = (flags & RING_F_SC_DEQ) ? - RTE_RING_SYNC_ST : RTE_RING_SYNC_MT; + ret = get_sync_type(flags, &r->prod.sync_type, &r->cons.sync_type); + if (ret != 0) + return ret; if (flags & RING_F_EXACT_SZ) { r->size = rte_align32pow2(count + 1); @@ -126,8 +206,12 @@ rte_ring_init(struct rte_ring *r, const char *name, unsigned count, r->mask = count - 1; r->capacity = r->mask; } - r->prod.head = r->cons.head = 0; - r->prod.tail = r->cons.tail = 0; + + /* set default values for head-tail distance */ + if (flags & RING_F_MP_RTS_ENQ) + rte_ring_set_prod_htd_max(r, r->capacity / HTD_MAX_DEF); + if (flags & RING_F_MC_RTS_DEQ) + rte_ring_set_cons_htd_max(r, r->capacity / HTD_MAX_DEF); return 0; } diff --git a/lib/librte_ring/rte_ring.h b/lib/librte_ring/rte_ring.h index 35ee4491c6..c42e1cfc46 100644 --- a/lib/librte_ring/rte_ring.h +++ b/lib/librte_ring/rte_ring.h @@ -1,6 +1,6 @@ /* SPDX-License-Identifier: BSD-3-Clause * - * Copyright (c) 2010-2017 Intel Corporation + * Copyright (c) 2010-2020 Intel Corporation * Copyright (c) 2007-2009 Kip Macy kmacy@freebsd.org * All rights reserved. * Derived from FreeBSD's bufring.h @@ -79,12 +79,24 @@ ssize_t rte_ring_get_memsize(unsigned count); * The number of elements in the ring (must be a power of 2). * @param flags * An OR of the following: - * - RING_F_SP_ENQ: If this flag is set, the default behavior when - * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` - * is "single-producer". Otherwise, it is "multi-producers". - * - RING_F_SC_DEQ: If this flag is set, the default behavior when - * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` - * is "single-consumer". Otherwise, it is "multi-consumers". + * - One of mutually exclusive flags that define producer behavior: + * - RING_F_SP_ENQ: If this flag is set, the default behavior when + * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` + * is "single-producer". + * - RING_F_MP_RTS_ENQ: If this flag is set, the default behavior when + * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` + * is "multi-producer RTS mode". + * If none of these flags is set, then default "multi-producer" + * behavior is selected. + * - One of mutually exclusive flags that define consumer behavior: + * - RING_F_SC_DEQ: If this flag is set, the default behavior when + * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` + * is "single-consumer". Otherwise, it is "multi-consumers". + * - RING_F_MC_RTS_DEQ: If this flag is set, the default behavior when + * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` + * is "multi-consumer RTS mode". + * If none of these flags is set, then default "multi-consumer" + * behavior is selected. * @return * 0 on success, or a negative value on error. */ @@ -114,12 +126,24 @@ int rte_ring_init(struct rte_ring *r, const char *name, unsigned count, * constraint for the reserved zone. * @param flags * An OR of the following: - * - RING_F_SP_ENQ: If this flag is set, the default behavior when - * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` - * is "single-producer". Otherwise, it is "multi-producers". - * - RING_F_SC_DEQ: If this flag is set, the default behavior when - * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` - * is "single-consumer". Otherwise, it is "multi-consumers". + * - One of mutually exclusive flags that define producer behavior: + * - RING_F_SP_ENQ: If this flag is set, the default behavior when + * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` + * is "single-producer". + * - RING_F_MP_RTS_ENQ: If this flag is set, the default behavior when + * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` + * is "multi-producer RTS mode". + * If none of these flags is set, then default "multi-producer" + * behavior is selected. + * - One of mutually exclusive flags that define consumer behavior: + * - RING_F_SC_DEQ: If this flag is set, the default behavior when + * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` + * is "single-consumer". Otherwise, it is "multi-consumers". + * - RING_F_MC_RTS_DEQ: If this flag is set, the default behavior when + * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` + * is "multi-consumer RTS mode". + * If none of these flags is set, then default "multi-consumer" + * behavior is selected. * @return * On success, the pointer to the new allocated ring. NULL on error with * rte_errno set appropriately. Possible errno values include: @@ -389,8 +413,21 @@ static __rte_always_inline unsigned int rte_ring_enqueue_bulk(struct rte_ring *r, void * const *obj_table, unsigned int n, unsigned int *free_space) { - return __rte_ring_do_enqueue(r, obj_table, n, RTE_RING_QUEUE_FIXED, - r->prod.sync_type, free_space); + switch (r->prod.sync_type) { + case RTE_RING_SYNC_MT: + return rte_ring_mp_enqueue_bulk(r, obj_table, n, free_space); + case RTE_RING_SYNC_ST: + return rte_ring_sp_enqueue_bulk(r, obj_table, n, free_space); +#ifdef ALLOW_EXPERIMENTAL_API + case RTE_RING_SYNC_MT_RTS: + return rte_ring_mp_rts_enqueue_bulk(r, obj_table, n, + free_space); +#endif + } + + /* valid ring should never reach this point */ + RTE_ASSERT(0); + return 0; } /** @@ -524,8 +561,20 @@ static __rte_always_inline unsigned int rte_ring_dequeue_bulk(struct rte_ring *r, void **obj_table, unsigned int n, unsigned int *available) { - return __rte_ring_do_dequeue(r, obj_table, n, RTE_RING_QUEUE_FIXED, - r->cons.sync_type, available); + switch (r->cons.sync_type) { + case RTE_RING_SYNC_MT: + return rte_ring_mc_dequeue_bulk(r, obj_table, n, available); + case RTE_RING_SYNC_ST: + return rte_ring_sc_dequeue_bulk(r, obj_table, n, available); +#ifdef ALLOW_EXPERIMENTAL_API + case RTE_RING_SYNC_MT_RTS: + return rte_ring_mc_rts_dequeue_bulk(r, obj_table, n, available); +#endif + } + + /* valid ring should never reach this point */ + RTE_ASSERT(0); + return 0; } /** @@ -845,8 +894,21 @@ static __rte_always_inline unsigned rte_ring_enqueue_burst(struct rte_ring *r, void * const *obj_table, unsigned int n, unsigned int *free_space) { - return __rte_ring_do_enqueue(r, obj_table, n, RTE_RING_QUEUE_VARIABLE, - r->prod.sync_type, free_space); + switch (r->prod.sync_type) { + case RTE_RING_SYNC_MT: + return rte_ring_mp_enqueue_burst(r, obj_table, n, free_space); + case RTE_RING_SYNC_ST: + return rte_ring_sp_enqueue_burst(r, obj_table, n, free_space); +#ifdef ALLOW_EXPERIMENTAL_API + case RTE_RING_SYNC_MT_RTS: + return rte_ring_mp_rts_enqueue_burst(r, obj_table, n, + free_space); +#endif + } + + /* valid ring should never reach this point */ + RTE_ASSERT(0); + return 0; } /** @@ -925,9 +987,21 @@ static __rte_always_inline unsigned rte_ring_dequeue_burst(struct rte_ring *r, void **obj_table, unsigned int n, unsigned int *available) { - return __rte_ring_do_dequeue(r, obj_table, n, - RTE_RING_QUEUE_VARIABLE, - r->cons.sync_type, available); + switch (r->cons.sync_type) { + case RTE_RING_SYNC_MT: + return rte_ring_mc_dequeue_burst(r, obj_table, n, available); + case RTE_RING_SYNC_ST: + return rte_ring_sc_dequeue_burst(r, obj_table, n, available); +#ifdef ALLOW_EXPERIMENTAL_API + case RTE_RING_SYNC_MT_RTS: + return rte_ring_mc_rts_dequeue_burst(r, obj_table, n, + available); +#endif + } + + /* valid ring should never reach this point */ + RTE_ASSERT(0); + return 0; } #ifdef __cplusplus diff --git a/lib/librte_ring/rte_ring_core.h b/lib/librte_ring/rte_ring_core.h index d9cef763fd..bd21fa5355 100644 --- a/lib/librte_ring/rte_ring_core.h +++ b/lib/librte_ring/rte_ring_core.h @@ -57,6 +57,9 @@ enum rte_ring_queue_behavior { enum rte_ring_sync_type { RTE_RING_SYNC_MT, /**< multi-thread safe (default mode) */ RTE_RING_SYNC_ST, /**< single thread only */ +#ifdef ALLOW_EXPERIMENTAL_API + RTE_RING_SYNC_MT_RTS, /**< multi-thread relaxed tail sync */ +#endif }; /** @@ -76,6 +79,22 @@ struct rte_ring_headtail { }; }; +union __rte_ring_rts_poscnt { + /** raw 8B value to read/write *cnt* and *pos* as one atomic op */ + uint64_t raw __rte_aligned(8); + struct { + uint32_t cnt; /**< head/tail reference counter */ + uint32_t pos; /**< head/tail position */ + } val; +}; + +struct rte_ring_rts_headtail { + volatile union __rte_ring_rts_poscnt tail; + enum rte_ring_sync_type sync_type; /**< sync type of prod/cons */ + uint32_t htd_max; /**< max allowed distance between head/tail */ + volatile union __rte_ring_rts_poscnt head; +}; + /** * An RTE ring structure. * @@ -104,11 +123,21 @@ struct rte_ring { char pad0 __rte_cache_aligned; /**< empty cache line */ /** Ring producer status. */ - struct rte_ring_headtail prod __rte_cache_aligned; + RTE_STD_C11 + union { + struct rte_ring_headtail prod; + struct rte_ring_rts_headtail rts_prod; + } __rte_cache_aligned; + char pad1 __rte_cache_aligned; /**< empty cache line */ /** Ring consumer status. */ - struct rte_ring_headtail cons __rte_cache_aligned; + RTE_STD_C11 + union { + struct rte_ring_headtail cons; + struct rte_ring_rts_headtail rts_cons; + } __rte_cache_aligned; + char pad2 __rte_cache_aligned; /**< empty cache line */ }; @@ -125,6 +154,9 @@ struct rte_ring { #define RING_F_EXACT_SZ 0x0004 #define RTE_RING_SZ_MASK (0x7fffffffU) /**< Ring size mask */ +#define RING_F_MP_RTS_ENQ 0x0008 /**< The default enqueue is "MP RTS". */ +#define RING_F_MC_RTS_DEQ 0x0010 /**< The default dequeue is "MC RTS". */ + #ifdef __cplusplus } #endif diff --git a/lib/librte_ring/rte_ring_elem.h b/lib/librte_ring/rte_ring_elem.h index 7406c0b0f3..4030753b60 100644 --- a/lib/librte_ring/rte_ring_elem.h +++ b/lib/librte_ring/rte_ring_elem.h @@ -74,12 +74,24 @@ ssize_t rte_ring_get_memsize_elem(unsigned int esize, unsigned int count); * constraint for the reserved zone. * @param flags * An OR of the following: - * - RING_F_SP_ENQ: If this flag is set, the default behavior when - * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` - * is "single-producer". Otherwise, it is "multi-producers". - * - RING_F_SC_DEQ: If this flag is set, the default behavior when - * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` - * is "single-consumer". Otherwise, it is "multi-consumers". + * - One of mutually exclusive flags that define producer behavior: + * - RING_F_SP_ENQ: If this flag is set, the default behavior when + * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` + * is "single-producer". + * - RING_F_MP_RTS_ENQ: If this flag is set, the default behavior when + * using ``rte_ring_enqueue()`` or ``rte_ring_enqueue_bulk()`` + * is "multi-producer RTS mode". + * If none of these flags is set, then default "multi-producer" + * behavior is selected. + * - One of mutually exclusive flags that define consumer behavior: + * - RING_F_SC_DEQ: If this flag is set, the default behavior when + * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` + * is "single-consumer". Otherwise, it is "multi-consumers". + * - RING_F_MC_RTS_DEQ: If this flag is set, the default behavior when + * using ``rte_ring_dequeue()`` or ``rte_ring_dequeue_bulk()`` + * is "multi-consumer RTS mode". + * If none of these flags is set, then default "multi-consumer" + * behavior is selected. * @return * On success, the pointer to the new allocated ring. NULL on error with * rte_errno set appropriately. Possible errno values include: @@ -528,6 +540,10 @@ rte_ring_sp_enqueue_bulk_elem(struct rte_ring *r, const void *obj_table, RTE_RING_QUEUE_FIXED, RTE_RING_SYNC_ST, free_space); } +#ifdef ALLOW_EXPERIMENTAL_API +#include +#endif + /** * Enqueue several objects on a ring. * @@ -557,6 +573,26 @@ rte_ring_enqueue_bulk_elem(struct rte_ring *r, const void *obj_table, { return __rte_ring_do_enqueue_elem(r, obj_table, esize, n, RTE_RING_QUEUE_FIXED, r->prod.sync_type, free_space); + + switch (r->prod.sync_type) { + case RTE_RING_SYNC_MT: + return rte_ring_mp_enqueue_bulk_elem(r, obj_table, esize, n, + free_space); + case RTE_RING_SYNC_ST: + return rte_ring_sp_enqueue_bulk_elem(r, obj_table, esize, n, + free_space); +#ifdef ALLOW_EXPERIMENTAL_API + case RTE_RING_SYNC_MT_RTS: + return rte_ring_mp_rts_enqueue_bulk_elem(r, obj_table, esize, n, + free_space); +#endif + } + + /* valid ring should never reach this point */ + RTE_ASSERT(0); + if (free_space != NULL) + *free_space = 0; + return 0; } /** @@ -661,7 +697,7 @@ rte_ring_mc_dequeue_bulk_elem(struct rte_ring *r, void *obj_table, unsigned int esize, unsigned int n, unsigned int *available) { return __rte_ring_do_dequeue_elem(r, obj_table, esize, n, - RTE_RING_QUEUE_FIXED, RTE_RING_SYNC_MT, available); + RTE_RING_QUEUE_FIXED, RTE_RING_SYNC_MT, available); } /** @@ -719,8 +755,25 @@ static __rte_always_inline unsigned int rte_ring_dequeue_bulk_elem(struct rte_ring *r, void *obj_table, unsigned int esize, unsigned int n, unsigned int *available) { - return __rte_ring_do_dequeue_elem(r, obj_table, esize, n, - RTE_RING_QUEUE_FIXED, r->cons.sync_type, available); + switch (r->cons.sync_type) { + case RTE_RING_SYNC_MT: + return rte_ring_mc_dequeue_bulk_elem(r, obj_table, esize, n, + available); + case RTE_RING_SYNC_ST: + return rte_ring_sc_dequeue_bulk_elem(r, obj_table, esize, n, + available); +#ifdef ALLOW_EXPERIMENTAL_API + case RTE_RING_SYNC_MT_RTS: + return rte_ring_mc_rts_dequeue_bulk_elem(r, obj_table, esize, + n, available); +#endif + } + + /* valid ring should never reach this point */ + RTE_ASSERT(0); + if (available != NULL) + *available = 0; + return 0; } /** @@ -887,8 +940,25 @@ static __rte_always_inline unsigned rte_ring_enqueue_burst_elem(struct rte_ring *r, const void *obj_table, unsigned int esize, unsigned int n, unsigned int *free_space) { - return __rte_ring_do_enqueue_elem(r, obj_table, esize, n, - RTE_RING_QUEUE_VARIABLE, r->prod.sync_type, free_space); + switch (r->prod.sync_type) { + case RTE_RING_SYNC_MT: + return rte_ring_mp_enqueue_burst_elem(r, obj_table, esize, n, + free_space); + case RTE_RING_SYNC_ST: + return rte_ring_sp_enqueue_burst_elem(r, obj_table, esize, n, + free_space); +#ifdef ALLOW_EXPERIMENTAL_API + case RTE_RING_SYNC_MT_RTS: + return rte_ring_mp_rts_enqueue_burst_elem(r, obj_table, esize, + n, free_space); +#endif + } + + /* valid ring should never reach this point */ + RTE_ASSERT(0); + if (free_space != NULL) + *free_space = 0; + return 0; } /** @@ -979,9 +1049,25 @@ static __rte_always_inline unsigned int rte_ring_dequeue_burst_elem(struct rte_ring *r, void *obj_table, unsigned int esize, unsigned int n, unsigned int *available) { - return __rte_ring_do_dequeue_elem(r, obj_table, esize, n, - RTE_RING_QUEUE_VARIABLE, - r->cons.sync_type, available); + switch (r->cons.sync_type) { + case RTE_RING_SYNC_MT: + return rte_ring_mc_dequeue_burst_elem(r, obj_table, esize, n, + available); + case RTE_RING_SYNC_ST: + return rte_ring_sc_dequeue_burst_elem(r, obj_table, esize, n, + available); +#ifdef ALLOW_EXPERIMENTAL_API + case RTE_RING_SYNC_MT_RTS: + return rte_ring_mc_rts_dequeue_burst_elem(r, obj_table, esize, + n, available); +#endif + } + + /* valid ring should never reach this point */ + RTE_ASSERT(0); + if (available != NULL) + *available = 0; + return 0; } #include diff --git a/lib/librte_ring/rte_ring_rts.h b/lib/librte_ring/rte_ring_rts.h new file mode 100644 index 0000000000..d40e9994fb --- /dev/null +++ b/lib/librte_ring/rte_ring_rts.h @@ -0,0 +1,439 @@ +/* SPDX-License-Identifier: BSD-3-Clause + * + * Copyright (c) 2010-2020 Intel Corporation + * Copyright (c) 2007-2009 Kip Macy kmacy@freebsd.org + * All rights reserved. + * Derived from FreeBSD's bufring.h + * Used as BSD-3 Licensed with permission from Kip Macy. + */ + +#ifndef _RTE_RING_RTS_H_ +#define _RTE_RING_RTS_H_ + +/** + * @file rte_ring_rts.h + * @b EXPERIMENTAL: this API may change without prior notice + * It is not recommended to include this file directly. + * Please include instead. + * + * Contains functions for Relaxed Tail Sync (RTS) ring mode. + * The main idea remains the same as for our original MP/MC synchronization + * mechanism. + * The main difference is that tail value is increased not + * by every thread that finished enqueue/dequeue, + * but only by the current last one doing enqueue/dequeue. + * That allows threads to skip spinning on tail value, + * leaving actual tail value change to last thread at a given instance. + * RTS requires 2 64-bit CAS for each enqueue(/dequeue) operation: + * one for head update, second for tail update. + * As a gain it allows thread to avoid spinning/waiting on tail value. + * In comparison original MP/MC algorithm requires one 32-bit CAS + * for head update and waiting/spinning on tail value. + * + * Brief outline: + * - introduce update counter (cnt) for both head and tail. + * - increment head.cnt for each head.value update + * - write head.value and head.cnt atomically (64-bit CAS) + * - move tail.value ahead only when tail.cnt + 1 == head.cnt + * (indicating that this is the last thread updating the tail) + * - increment tail.cnt when each enqueue/dequeue op finishes + * (no matter if tail.value going to change or not) + * - write tail.value and tail.cnt atomically (64-bit CAS) + * + * To avoid producer/consumer starvation: + * - limit max allowed distance between head and tail value (HTD_MAX). + * I.E. thread is allowed to proceed with changing head.value, + * only when: head.value - tail.value <= HTD_MAX + * HTD_MAX is an optional parameter. + * With HTD_MAX == 0 we'll have fully serialized ring - + * i.e. only one thread at a time will be able to enqueue/dequeue + * to/from the ring. + * With HTD_MAX >= ring.capacity - no limitation. + * By default HTD_MAX == ring.capacity / 8. + */ + +#ifdef __cplusplus +extern "C" { +#endif + +#include + +/** + * @internal Enqueue several objects on the RTS ring. + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of objects. + * @param esize + * The size of ring element, in bytes. It must be a multiple of 4. + * This must be the same value used while creating the ring. Otherwise + * the results are undefined. + * @param n + * The number of objects to add in the ring from the obj_table. + * @param behavior + * RTE_RING_QUEUE_FIXED: Enqueue a fixed number of items from a ring + * RTE_RING_QUEUE_VARIABLE: Enqueue as many items as possible from ring + * @param free_space + * returns the amount of space after the enqueue operation has finished + * @return + * Actual number of objects enqueued. + * If behavior == RTE_RING_QUEUE_FIXED, this will be 0 or n only. + */ +static __rte_always_inline unsigned int +__rte_ring_do_rts_enqueue_elem(struct rte_ring *r, const void *obj_table, + uint32_t esize, uint32_t n, enum rte_ring_queue_behavior behavior, + uint32_t *free_space) +{ + uint32_t free, head; + + n = __rte_ring_rts_move_prod_head(r, n, behavior, &head, &free); + + if (n != 0) { + __rte_ring_enqueue_elems(r, head, obj_table, esize, n); + __rte_ring_rts_update_tail(&r->rts_prod); + } + + if (free_space != NULL) + *free_space = free - n; + return n; +} + +/** + * @internal Dequeue several objects from the RTS ring. + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of objects. + * @param esize + * The size of ring element, in bytes. It must be a multiple of 4. + * This must be the same value used while creating the ring. Otherwise + * the results are undefined. + * @param n + * The number of objects to pull from the ring. + * @param behavior + * RTE_RING_QUEUE_FIXED: Dequeue a fixed number of items from a ring + * RTE_RING_QUEUE_VARIABLE: Dequeue as many items as possible from ring + * @param available + * returns the number of remaining ring entries after the dequeue has finished + * @return + * - Actual number of objects dequeued. + * If behavior == RTE_RING_QUEUE_FIXED, this will be 0 or n only. + */ +static __rte_always_inline unsigned int +__rte_ring_do_rts_dequeue_elem(struct rte_ring *r, void *obj_table, + uint32_t esize, uint32_t n, enum rte_ring_queue_behavior behavior, + uint32_t *available) +{ + uint32_t entries, head; + + n = __rte_ring_rts_move_cons_head(r, n, behavior, &head, &entries); + + if (n != 0) { + __rte_ring_dequeue_elems(r, head, obj_table, esize, n); + __rte_ring_rts_update_tail(&r->rts_cons); + } + + if (available != NULL) + *available = entries - n; + return n; +} + +/** + * Enqueue several objects on the RTS ring (multi-producers safe). + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of objects. + * @param esize + * The size of ring element, in bytes. It must be a multiple of 4. + * This must be the same value used while creating the ring. Otherwise + * the results are undefined. + * @param n + * The number of objects to add in the ring from the obj_table. + * @param free_space + * if non-NULL, returns the amount of space in the ring after the + * enqueue operation has finished. + * @return + * The number of objects enqueued, either 0 or n + */ +__rte_experimental +static __rte_always_inline unsigned int +rte_ring_mp_rts_enqueue_bulk_elem(struct rte_ring *r, const void *obj_table, + unsigned int esize, unsigned int n, unsigned int *free_space) +{ + return __rte_ring_do_rts_enqueue_elem(r, obj_table, esize, n, + RTE_RING_QUEUE_FIXED, free_space); +} + +/** + * Dequeue several objects from an RTS ring (multi-consumers safe). + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of objects that will be filled. + * @param esize + * The size of ring element, in bytes. It must be a multiple of 4. + * This must be the same value used while creating the ring. Otherwise + * the results are undefined. + * @param n + * The number of objects to dequeue from the ring to the obj_table. + * @param available + * If non-NULL, returns the number of remaining ring entries after the + * dequeue has finished. + * @return + * The number of objects dequeued, either 0 or n + */ +__rte_experimental +static __rte_always_inline unsigned int +rte_ring_mc_rts_dequeue_bulk_elem(struct rte_ring *r, void *obj_table, + unsigned int esize, unsigned int n, unsigned int *available) +{ + return __rte_ring_do_rts_dequeue_elem(r, obj_table, esize, n, + RTE_RING_QUEUE_FIXED, available); +} + +/** + * Enqueue several objects on the RTS ring (multi-producers safe). + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of objects. + * @param esize + * The size of ring element, in bytes. It must be a multiple of 4. + * This must be the same value used while creating the ring. Otherwise + * the results are undefined. + * @param n + * The number of objects to add in the ring from the obj_table. + * @param free_space + * if non-NULL, returns the amount of space in the ring after the + * enqueue operation has finished. + * @return + * - n: Actual number of objects enqueued. + */ +__rte_experimental +static __rte_always_inline unsigned +rte_ring_mp_rts_enqueue_burst_elem(struct rte_ring *r, const void *obj_table, + unsigned int esize, unsigned int n, unsigned int *free_space) +{ + return __rte_ring_do_rts_enqueue_elem(r, obj_table, esize, n, + RTE_RING_QUEUE_VARIABLE, free_space); +} + +/** + * Dequeue several objects from an RTS ring (multi-consumers safe). + * When the requested objects are more than the available objects, + * only dequeue the actual number of objects. + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of objects that will be filled. + * @param esize + * The size of ring element, in bytes. It must be a multiple of 4. + * This must be the same value used while creating the ring. Otherwise + * the results are undefined. + * @param n + * The number of objects to dequeue from the ring to the obj_table. + * @param available + * If non-NULL, returns the number of remaining ring entries after the + * dequeue has finished. + * @return + * - n: Actual number of objects dequeued, 0 if ring is empty + */ +__rte_experimental +static __rte_always_inline unsigned +rte_ring_mc_rts_dequeue_burst_elem(struct rte_ring *r, void *obj_table, + unsigned int esize, unsigned int n, unsigned int *available) +{ + return __rte_ring_do_rts_dequeue_elem(r, obj_table, esize, n, + RTE_RING_QUEUE_VARIABLE, available); +} + +/** + * Enqueue several objects on the RTS ring (multi-producers safe). + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of void * pointers (objects). + * @param n + * The number of objects to add in the ring from the obj_table. + * @param free_space + * if non-NULL, returns the amount of space in the ring after the + * enqueue operation has finished. + * @return + * The number of objects enqueued, either 0 or n + */ +__rte_experimental +static __rte_always_inline unsigned int +rte_ring_mp_rts_enqueue_bulk(struct rte_ring *r, void * const *obj_table, + unsigned int n, unsigned int *free_space) +{ + return rte_ring_mp_rts_enqueue_bulk_elem(r, obj_table, + sizeof(uintptr_t), n, free_space); +} + +/** + * Dequeue several objects from an RTS ring (multi-consumers safe). + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of void * pointers (objects) that will be filled. + * @param n + * The number of objects to dequeue from the ring to the obj_table. + * @param available + * If non-NULL, returns the number of remaining ring entries after the + * dequeue has finished. + * @return + * The number of objects dequeued, either 0 or n + */ +__rte_experimental +static __rte_always_inline unsigned int +rte_ring_mc_rts_dequeue_bulk(struct rte_ring *r, void **obj_table, + unsigned int n, unsigned int *available) +{ + return rte_ring_mc_rts_dequeue_bulk_elem(r, obj_table, + sizeof(uintptr_t), n, available); +} + +/** + * Enqueue several objects on the RTS ring (multi-producers safe). + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of void * pointers (objects). + * @param n + * The number of objects to add in the ring from the obj_table. + * @param free_space + * if non-NULL, returns the amount of space in the ring after the + * enqueue operation has finished. + * @return + * - n: Actual number of objects enqueued. + */ +__rte_experimental +static __rte_always_inline unsigned +rte_ring_mp_rts_enqueue_burst(struct rte_ring *r, void * const *obj_table, + unsigned int n, unsigned int *free_space) +{ + return rte_ring_mp_rts_enqueue_burst_elem(r, obj_table, + sizeof(uintptr_t), n, free_space); +} + +/** + * Dequeue several objects from an RTS ring (multi-consumers safe). + * When the requested objects are more than the available objects, + * only dequeue the actual number of objects. + * + * @param r + * A pointer to the ring structure. + * @param obj_table + * A pointer to a table of void * pointers (objects) that will be filled. + * @param n + * The number of objects to dequeue from the ring to the obj_table. + * @param available + * If non-NULL, returns the number of remaining ring entries after the + * dequeue has finished. + * @return + * - n: Actual number of objects dequeued, 0 if ring is empty + */ +__rte_experimental +static __rte_always_inline unsigned +rte_ring_mc_rts_dequeue_burst(struct rte_ring *r, void **obj_table, + unsigned int n, unsigned int *available) +{ + return rte_ring_mc_rts_dequeue_burst_elem(r, obj_table, + sizeof(uintptr_t), n, available); +} + +/** + * Return producer max Head-Tail-Distance (HTD). + * + * @param r + * A pointer to the ring structure. + * @return + * Producer HTD value, if producer is set in appropriate sync mode, + * or UINT32_MAX otherwise. + */ +__rte_experimental +static inline uint32_t +rte_ring_get_prod_htd_max(const struct rte_ring *r) +{ + if (r->prod.sync_type == RTE_RING_SYNC_MT_RTS) + return r->rts_prod.htd_max; + return UINT32_MAX; +} + +/** + * Set producer max Head-Tail-Distance (HTD). + * Note that producer has to use appropriate sync mode (RTS). + * + * @param r + * A pointer to the ring structure. + * @param v + * new HTD value to setup. + * @return + * Zero on success, or negative error code otherwise. + */ +__rte_experimental +static inline int +rte_ring_set_prod_htd_max(struct rte_ring *r, uint32_t v) +{ + if (r->prod.sync_type != RTE_RING_SYNC_MT_RTS) + return -ENOTSUP; + + r->rts_prod.htd_max = v; + return 0; +} + +/** + * Return consumer max Head-Tail-Distance (HTD). + * + * @param r + * A pointer to the ring structure. + * @return + * Consumer HTD value, if consumer is set in appropriate sync mode, + * or UINT32_MAX otherwise. + */ +__rte_experimental +static inline uint32_t +rte_ring_get_cons_htd_max(const struct rte_ring *r) +{ + if (r->cons.sync_type == RTE_RING_SYNC_MT_RTS) + return r->rts_cons.htd_max; + return UINT32_MAX; +} + +/** + * Set consumer max Head-Tail-Distance (HTD). + * Note that consumer has to use appropriate sync mode (RTS). + * + * @param r + * A pointer to the ring structure. + * @param v + * new HTD value to setup. + * @return + * Zero on success, or negative error code otherwise. + */ +__rte_experimental +static inline int +rte_ring_set_cons_htd_max(struct rte_ring *r, uint32_t v) +{ + if (r->cons.sync_type != RTE_RING_SYNC_MT_RTS) + return -ENOTSUP; + + r->rts_cons.htd_max = v; + return 0; +} + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_RING_RTS_H_ */ diff --git a/lib/librte_ring/rte_ring_rts_c11_mem.h b/lib/librte_ring/rte_ring_rts_c11_mem.h new file mode 100644 index 0000000000..327f227964 --- /dev/null +++ b/lib/librte_ring/rte_ring_rts_c11_mem.h @@ -0,0 +1,179 @@ +/* SPDX-License-Identifier: BSD-3-Clause + * + * Copyright (c) 2010-2020 Intel Corporation + * Copyright (c) 2007-2009 Kip Macy kmacy@freebsd.org + * All rights reserved. + * Derived from FreeBSD's bufring.h + * Used as BSD-3 Licensed with permission from Kip Macy. + */ + +#ifndef _RTE_RING_RTS_C11_MEM_H_ +#define _RTE_RING_RTS_C11_MEM_H_ + +/** + * @file rte_ring_rts_c11_mem.h + * It is not recommended to include this file directly, + * include instead. + * Contains internal helper functions for Relaxed Tail Sync (RTS) ring mode. + * For more information please refer to . + */ + +/** + * @internal This function updates tail values. + */ +static __rte_always_inline void +__rte_ring_rts_update_tail(struct rte_ring_rts_headtail *ht) +{ + union __rte_ring_rts_poscnt h, ot, nt; + + /* + * If there are other enqueues/dequeues in progress that + * might preceded us, then don't update tail with new value. + */ + + ot.raw = __atomic_load_n(&ht->tail.raw, __ATOMIC_ACQUIRE); + + do { + /* on 32-bit systems we have to do atomic read here */ + h.raw = __atomic_load_n(&ht->head.raw, __ATOMIC_RELAXED); + + nt.raw = ot.raw; + if (++nt.val.cnt == h.val.cnt) + nt.val.pos = h.val.pos; + + } while (__atomic_compare_exchange_n(&ht->tail.raw, &ot.raw, nt.raw, + 0, __ATOMIC_RELEASE, __ATOMIC_ACQUIRE) == 0); +} + +/** + * @internal This function waits till head/tail distance wouldn't + * exceed pre-defined max value. + */ +static __rte_always_inline void +__rte_ring_rts_head_wait(const struct rte_ring_rts_headtail *ht, + union __rte_ring_rts_poscnt *h) +{ + uint32_t max; + + max = ht->htd_max; + + while (h->val.pos - ht->tail.val.pos > max) { + rte_pause(); + h->raw = __atomic_load_n(&ht->head.raw, __ATOMIC_ACQUIRE); + } +} + +/** + * @internal This function updates the producer head for enqueue. + */ +static __rte_always_inline uint32_t +__rte_ring_rts_move_prod_head(struct rte_ring *r, uint32_t num, + enum rte_ring_queue_behavior behavior, uint32_t *old_head, + uint32_t *free_entries) +{ + uint32_t n; + union __rte_ring_rts_poscnt nh, oh; + + const uint32_t capacity = r->capacity; + + oh.raw = __atomic_load_n(&r->rts_prod.head.raw, __ATOMIC_ACQUIRE); + + do { + /* Reset n to the initial burst count */ + n = num; + + /* + * wait for prod head/tail distance, + * make sure that we read prod head *before* + * reading cons tail. + */ + __rte_ring_rts_head_wait(&r->rts_prod, &oh); + + /* + * The subtraction is done between two unsigned 32bits value + * (the result is always modulo 32 bits even if we have + * *old_head > cons_tail). So 'free_entries' is always between 0 + * and capacity (which is < size). + */ + *free_entries = capacity + r->cons.tail - oh.val.pos; + + /* check that we have enough room in ring */ + if (unlikely(n > *free_entries)) + n = (behavior == RTE_RING_QUEUE_FIXED) ? + 0 : *free_entries; + + if (n == 0) + break; + + nh.val.pos = oh.val.pos + n; + nh.val.cnt = oh.val.cnt + 1; + + /* + * this CAS(ACQUIRE, ACQUIRE) serves as a hoist barrier to prevent: + * - OOO reads of cons tail value + * - OOO copy of elems to the ring + */ + } while (__atomic_compare_exchange_n(&r->rts_prod.head.raw, + &oh.raw, nh.raw, + 0, __ATOMIC_ACQUIRE, __ATOMIC_ACQUIRE) == 0); + + *old_head = oh.val.pos; + return n; +} + +/** + * @internal This function updates the consumer head for dequeue + */ +static __rte_always_inline unsigned int +__rte_ring_rts_move_cons_head(struct rte_ring *r, uint32_t num, + enum rte_ring_queue_behavior behavior, uint32_t *old_head, + uint32_t *entries) +{ + uint32_t n; + union __rte_ring_rts_poscnt nh, oh; + + oh.raw = __atomic_load_n(&r->rts_cons.head.raw, __ATOMIC_ACQUIRE); + + /* move cons.head atomically */ + do { + /* Restore n as it may change every loop */ + n = num; + + /* + * wait for cons head/tail distance, + * make sure that we read cons head *before* + * reading prod tail. + */ + __rte_ring_rts_head_wait(&r->rts_cons, &oh); + + /* The subtraction is done between two unsigned 32bits value + * (the result is always modulo 32 bits even if we have + * cons_head > prod_tail). So 'entries' is always between 0 + * and size(ring)-1. + */ + *entries = r->prod.tail - oh.val.pos; + + /* Set the actual entries for dequeue */ + if (n > *entries) + n = (behavior == RTE_RING_QUEUE_FIXED) ? 0 : *entries; + + if (unlikely(n == 0)) + break; + + nh.val.pos = oh.val.pos + n; + nh.val.cnt = oh.val.cnt + 1; + + /* + * this CAS(ACQUIRE, ACQUIRE) serves as a hoist barrier to prevent: + * - OOO reads of prod tail value + * - OOO copy of elems from the ring + */ + } while (__atomic_compare_exchange_n(&r->rts_cons.head.raw, + &oh.raw, nh.raw, + 0, __ATOMIC_ACQUIRE, __ATOMIC_ACQUIRE) == 0); + + *old_head = oh.val.pos; + return n; +} + +#endif /* _RTE_RING_RTS_C11_MEM_H_ */ -- 2.20.1