From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id 28B9445920; Fri, 6 Sep 2024 19:09:13 +0200 (CEST) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 124F142E82; Fri, 6 Sep 2024 19:09:13 +0200 (CEST) Received: from mgamail.intel.com (mgamail.intel.com [192.198.163.19]) by mails.dpdk.org (Postfix) with ESMTP id B46C3400D5 for ; Fri, 6 Sep 2024 19:09:10 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1725642551; x=1757178551; h=from:to:cc:subject:date:message-id:mime-version: content-transfer-encoding; bh=T5uuFEb4LeQ17sRrEeSDWS8SfyVm9VAhx6qTN4QJXPo=; b=Xwx0CH5TmMMoI37T+f3vr0ayLAC5MB7q9X8iTl1FNCFaGKZ3fW+CFoHL 5EvTb5KVtdXTpVwScA0lzPf5oeEjKUyG7dThZqY0yM1jDj6vDm/xM/E+v Sh9+icxwtQ9/NsNAJLKmDpmN9ihpW50KjFOUc7R531USjq7P7jxaOfB12 EY0RfSHDNUIMUnIh62tFZRciGvGnW9/jpvf6sHHbhit4RhiZL8fAePXz2 1uIO9snEKYC9z7kMtnpH+9yikHd9K0nOk4PD1Fpe/kYmeMgElcKWNnf+t /ugQQDPhIVNKJvXntmUd2+Wjeekfwvpw82QOYCP6twzH7hyqSbt39hQB6 A==; X-CSE-ConnectionGUID: 559s/0CvRQyLBtUmozCVYg== X-CSE-MsgGUID: kv2pfmSZRE691ViWOrSzfw== X-IronPort-AV: E=McAfee;i="6700,10204,11187"; a="23969523" X-IronPort-AV: E=Sophos;i="6.10,208,1719903600"; d="scan'208";a="23969523" Received: from fmviesa007.fm.intel.com ([10.60.135.147]) by fmvoesa113.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 06 Sep 2024 10:09:10 -0700 X-CSE-ConnectionGUID: NqsiF5HkQWegzWjo7BIF+Q== X-CSE-MsgGUID: iMGI9PrLQdqIZMIgZ6NIFQ== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.10,208,1719903600"; d="scan'208";a="65698308" Received: from silpixa00401176.ir.intel.com ([10.243.22.170]) by fmviesa007.fm.intel.com with ESMTP; 06 Sep 2024 10:09:08 -0700 From: Vladimir Medvedkin To: dev@dpdk.org Cc: rjarry@redhat.com, ruifeng.wang@arm.com, honnappa.nagarahalli@arm.com Subject: [PATCH] fib: implement RCU rule reclamation Date: Fri, 6 Sep 2024 17:09:07 +0000 Message-Id: <20240906170907.1325808-1-vladimir.medvedkin@intel.com> X-Mailer: git-send-email 2.34.1 MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Currently, for DIR24-8 algorithm, the tbl8 group is freed even though the readers might be using the tbl8 group entries. The freed tbl8 group can be reallocated quickly. As a result, lookup may be performed incorrectly. To address that, RCU QSBR is integrated for safe tbl8 group reclamation. Signed-off-by: Vladimir Medvedkin --- lib/fib/dir24_8.c | 104 +++++++++++++++++++++++++++++++++++++++----- lib/fib/dir24_8.h | 9 ++++ lib/fib/meson.build | 1 + lib/fib/rte_fib.c | 11 +++++ lib/fib/rte_fib.h | 50 ++++++++++++++++++++- lib/fib/version.map | 7 +++ 6 files changed, 171 insertions(+), 11 deletions(-) diff --git a/lib/fib/dir24_8.c b/lib/fib/dir24_8.c index c739e92304..f884b02d2c 100644 --- a/lib/fib/dir24_8.c +++ b/lib/fib/dir24_8.c @@ -14,6 +14,7 @@ #include #include #include "dir24_8.h" +#include "fib_log.h" #ifdef CC_DIR24_8_AVX512_SUPPORT @@ -176,6 +177,13 @@ tbl8_alloc(struct dir24_8_tbl *dp, uint64_t nh) uint8_t *tbl8_ptr; tbl8_idx = tbl8_get_idx(dp); + if ((tbl8_idx == -ENOSPC) && dp->dq != NULL) { + /* If there are no tbl8 groups try to reclaim one. */ + if (rte_rcu_qsbr_dq_reclaim(dp->dq, 1, + NULL, NULL, NULL) == 0) + tbl8_idx = tbl8_get_idx(dp); + } + if (tbl8_idx < 0) return tbl8_idx; tbl8_ptr = (uint8_t *)dp->tbl8 + @@ -189,6 +197,27 @@ tbl8_alloc(struct dir24_8_tbl *dp, uint64_t nh) return tbl8_idx; } +static void +tbl8_cleanup_and_free(struct dir24_8_tbl *dp, uint64_t tbl8_idx) +{ + uint8_t *ptr = (uint8_t *)dp->tbl8 + + (tbl8_idx * DIR24_8_TBL8_GRP_NUM_ENT << dp->nh_sz); + + memset(ptr, 0, DIR24_8_TBL8_GRP_NUM_ENT << dp->nh_sz); + tbl8_free_idx(dp, tbl8_idx); + dp->cur_tbl8s--; +} + +static void +__rcu_qsbr_free_resource(void *p, void *data, unsigned int n) +{ + struct dir24_8_tbl *dp = p; + uint64_t tbl8_idx = *(uint64_t *)data; + RTE_SET_USED(n); + + tbl8_cleanup_and_free(dp, tbl8_idx); +} + static void tbl8_recycle(struct dir24_8_tbl *dp, uint32_t ip, uint64_t tbl8_idx) { @@ -210,8 +239,6 @@ tbl8_recycle(struct dir24_8_tbl *dp, uint32_t ip, uint64_t tbl8_idx) } ((uint8_t *)dp->tbl24)[ip >> 8] = nh & ~DIR24_8_EXT_ENT; - for (i = 0; i < DIR24_8_TBL8_GRP_NUM_ENT; i++) - ptr8[i] = 0; break; case RTE_FIB_DIR24_8_2B: ptr16 = &((uint16_t *)dp->tbl8)[tbl8_idx * @@ -223,8 +250,6 @@ tbl8_recycle(struct dir24_8_tbl *dp, uint32_t ip, uint64_t tbl8_idx) } ((uint16_t *)dp->tbl24)[ip >> 8] = nh & ~DIR24_8_EXT_ENT; - for (i = 0; i < DIR24_8_TBL8_GRP_NUM_ENT; i++) - ptr16[i] = 0; break; case RTE_FIB_DIR24_8_4B: ptr32 = &((uint32_t *)dp->tbl8)[tbl8_idx * @@ -236,8 +261,6 @@ tbl8_recycle(struct dir24_8_tbl *dp, uint32_t ip, uint64_t tbl8_idx) } ((uint32_t *)dp->tbl24)[ip >> 8] = nh & ~DIR24_8_EXT_ENT; - for (i = 0; i < DIR24_8_TBL8_GRP_NUM_ENT; i++) - ptr32[i] = 0; break; case RTE_FIB_DIR24_8_8B: ptr64 = &((uint64_t *)dp->tbl8)[tbl8_idx * @@ -249,12 +272,20 @@ tbl8_recycle(struct dir24_8_tbl *dp, uint32_t ip, uint64_t tbl8_idx) } ((uint64_t *)dp->tbl24)[ip >> 8] = nh & ~DIR24_8_EXT_ENT; - for (i = 0; i < DIR24_8_TBL8_GRP_NUM_ENT; i++) - ptr64[i] = 0; break; } - tbl8_free_idx(dp, tbl8_idx); - dp->cur_tbl8s--; + + if (dp->v == NULL) + tbl8_cleanup_and_free(dp, tbl8_idx); + else if (dp->rcu_mode == RTE_FIB_QSBR_MODE_SYNC) { + rte_rcu_qsbr_synchronize(dp->v, + RTE_QSBR_THRID_INVALID); + tbl8_cleanup_and_free(dp, tbl8_idx); + } else { /* RTE_FIB_QSBR_MODE_DQ */ + if (rte_rcu_qsbr_dq_enqueue(dp->dq, + (void *)&tbl8_idx)) + FIB_LOG(ERR, "Failed to push QSBR FIFO"); + } } static int @@ -569,7 +600,60 @@ dir24_8_free(void *p) { struct dir24_8_tbl *dp = (struct dir24_8_tbl *)p; + if (dp->dq != NULL) + rte_rcu_qsbr_dq_delete(dp->dq); + rte_free(dp->tbl8_idxes); rte_free(dp->tbl8); rte_free(dp); } + +int +dir24_8_rcu_qsbr_add(struct dir24_8_tbl *dp, struct rte_fib_rcu_config *cfg, + const char *name) +{ + struct rte_rcu_qsbr_dq_parameters params = {0}; + char rcu_dq_name[RTE_RCU_QSBR_DQ_NAMESIZE]; + + if (dp == NULL || cfg == NULL) { + rte_errno = EINVAL; + return 1; + } + + if (dp->v != NULL) { + rte_errno = EEXIST; + return 1; + } + + if (cfg->mode == RTE_FIB_QSBR_MODE_SYNC) { + /* No other things to do. */ + } else if (cfg->mode == RTE_FIB_QSBR_MODE_DQ) { + /* Init QSBR defer queue. */ + snprintf(rcu_dq_name, sizeof(rcu_dq_name), + "FIB_RCU_%s", name); + params.name = rcu_dq_name; + params.size = cfg->dq_size; + if (params.size == 0) + params.size = RTE_FIB_RCU_DQ_RECLAIM_SZ; + params.trigger_reclaim_limit = cfg->reclaim_thd; + params.max_reclaim_size = cfg->reclaim_max; + if (params.max_reclaim_size == 0) + params.max_reclaim_size = RTE_FIB_RCU_DQ_RECLAIM_MAX; + params.esize = sizeof(uint64_t); + params.free_fn = __rcu_qsbr_free_resource; + params.p = dp; + params.v = cfg->v; + dp->dq = rte_rcu_qsbr_dq_create(¶ms); + if (dp->dq == NULL) { + FIB_LOG(ERR, "LPM defer queue creation failed"); + return 1; + } + } else { + rte_errno = EINVAL; + return 1; + } + dp->rcu_mode = cfg->mode; + dp->v = cfg->v; + + return 0; +} \ No newline at end of file diff --git a/lib/fib/dir24_8.h b/lib/fib/dir24_8.h index 7125049f15..08fd818ce4 100644 --- a/lib/fib/dir24_8.h +++ b/lib/fib/dir24_8.h @@ -10,6 +10,7 @@ #include #include +#include /** * @file @@ -30,6 +31,10 @@ struct dir24_8_tbl { uint32_t rsvd_tbl8s; /**< Number of reserved tbl8s */ uint32_t cur_tbl8s; /**< Current number of tbl8s */ enum rte_fib_dir24_8_nh_sz nh_sz; /**< Size of nexthop entry */ + /* RCU config. */ + enum rte_fib_qsbr_mode rcu_mode;/* Blocking, defer queue. */ + struct rte_rcu_qsbr *v; /* RCU QSBR variable. */ + struct rte_rcu_qsbr_dq *dq; /* RCU QSBR defer queue. */ uint64_t def_nh; /**< Default next hop */ uint64_t *tbl8; /**< tbl8 table. */ uint64_t *tbl8_idxes; /**< bitmap containing free tbl8 idxes*/ @@ -250,4 +255,8 @@ int dir24_8_modify(struct rte_fib *fib, uint32_t ip, uint8_t depth, uint64_t next_hop, int op); +int +dir24_8_rcu_qsbr_add(struct dir24_8_tbl *dp, struct rte_fib_rcu_config *cfg, + const char *name); + #endif /* _DIR24_8_H_ */ diff --git a/lib/fib/meson.build b/lib/fib/meson.build index 6795f41a0a..1895f37050 100644 --- a/lib/fib/meson.build +++ b/lib/fib/meson.build @@ -11,6 +11,7 @@ endif sources = files('rte_fib.c', 'rte_fib6.c', 'dir24_8.c', 'trie.c') headers = files('rte_fib.h', 'rte_fib6.h') deps += ['rib'] +deps += ['rcu'] # compile AVX512 version if: # we are building 64-bit binary AND binutils can generate proper code diff --git a/lib/fib/rte_fib.c b/lib/fib/rte_fib.c index 4f9fba5a4f..f1b73d64cb 100644 --- a/lib/fib/rte_fib.c +++ b/lib/fib/rte_fib.c @@ -338,3 +338,14 @@ rte_fib_select_lookup(struct rte_fib *fib, return -EINVAL; } } + +int +rte_fib_rcu_qsbr_add(struct rte_fib *fib, struct rte_fib_rcu_config *cfg) +{ + switch (fib->type) { + case RTE_FIB_DIR24_8: + return dir24_8_rcu_qsbr_add(fib->dp, cfg, fib->name); + default: + return -ENOTSUP; + } +} \ No newline at end of file diff --git a/lib/fib/rte_fib.h b/lib/fib/rte_fib.h index d7a5aafe53..346eb7f149 100644 --- a/lib/fib/rte_fib.h +++ b/lib/fib/rte_fib.h @@ -16,7 +16,7 @@ */ #include - +#include #ifdef __cplusplus extern "C" { @@ -28,6 +28,19 @@ struct rte_rib; /** Maximum depth value possible for IPv4 FIB. */ #define RTE_FIB_MAXDEPTH 32 +/** @internal Default RCU defer queue entries to reclaim in one go. */ +#define RTE_FIB_RCU_DQ_RECLAIM_MAX 16 +/** @internal Default RCU defer queue size. */ +#define RTE_FIB_RCU_DQ_RECLAIM_SZ 128 + +/** RCU reclamation modes */ +enum rte_fib_qsbr_mode { + /** Create defer queue for reclaim. */ + RTE_FIB_QSBR_MODE_DQ = 0, + /** Use blocking mode reclaim. No defer queue created. */ + RTE_FIB_QSBR_MODE_SYNC +}; + /** Type of FIB struct */ enum rte_fib_type { RTE_FIB_DUMMY, /**< RIB tree based FIB */ @@ -89,6 +102,22 @@ struct rte_fib_conf { }; }; +/** FIB RCU QSBR configuration structure. */ +struct rte_fib_rcu_config { + struct rte_rcu_qsbr *v; /* RCU QSBR variable. */ + /* Mode of RCU QSBR. RTE_FIB_QSBR_MODE_xxx + * '0' for default: create defer queue for reclaim. + */ + enum rte_fib_qsbr_mode mode; + uint32_t dq_size; /* RCU defer queue size. + * default: RTE_FIB_RCU_DQ_RECLAIM_SZ. + */ + uint32_t reclaim_thd; /* Threshold to trigger auto reclaim. */ + uint32_t reclaim_max; /* Max entries to reclaim in one go. + * default: RTE_FIB_RCU_DQ_RECLAIM_MAX. + */ +}; + /** * Create FIB * @@ -219,6 +248,25 @@ rte_fib_get_rib(struct rte_fib *fib); int rte_fib_select_lookup(struct rte_fib *fib, enum rte_fib_lookup_type type); +/** + * Associate RCU QSBR variable with a FIB object. + * + * @param fib + * the fib object to add RCU QSBR + * @param cfg + * RCU QSBR configuration + * @return + * On success - 0 + * On error - 1 with error code set in rte_errno. + * Possible rte_errno codes are: + * - EINVAL - invalid pointer + * - EEXIST - already added QSBR + * - ENOMEM - memory allocation failure + * - ENOTSUP - not supported by configured dataplane algorithm + */ +__rte_experimental +int rte_fib_rcu_qsbr_add(struct rte_fib *fib, struct rte_fib_rcu_config *cfg); + #ifdef __cplusplus } #endif diff --git a/lib/fib/version.map b/lib/fib/version.map index c6d2769611..df8f113df3 100644 --- a/lib/fib/version.map +++ b/lib/fib/version.map @@ -22,3 +22,10 @@ DPDK_25 { local: *; }; + +EXPERIMENTAL { + global: + + # added in 24.11 + rte_fib_rcu_qsbr_add; +}; -- 2.34.1