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 7A92E43F19; Fri, 26 Apr 2024 17:52:20 +0200 (CEST) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id C8A6E43F1E; Fri, 26 Apr 2024 17:51:46 +0200 (CEST) Received: from mail-pg1-f170.google.com (mail-pg1-f170.google.com [209.85.215.170]) by mails.dpdk.org (Postfix) with ESMTP id AB0B143EFB for ; Fri, 26 Apr 2024 17:51:41 +0200 (CEST) Received: by mail-pg1-f170.google.com with SMTP id 41be03b00d2f7-5e42b4bbfa4so1460818a12.1 for ; Fri, 26 Apr 2024 08:51:41 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=networkplumber-org.20230601.gappssmtp.com; s=20230601; t=1714146701; x=1714751501; darn=dpdk.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=aYJCmNFOVedHApS/WzQzrBs6k52ZDo9m1VFd7MFhABU=; b=OF9K/k0gk38YHwPUrkLgVr9JuyBKtOlUxV7YOf78qiB3pBS0KH3NR0bu4k2+83cALV 3LDxZXXSroaXWcIL37pDolA8g2SCdSpx2flbIBGG+bClfzGKx8/prE+qso63mbmrf37L xDOKsj4PMH54Srx34ZNldtiO6uspXGgDCInQ3MrFn5lWtQizLprejYytjqYGaWeukR3U HnWJm12/ych18+tuu99Ec6IedYgbSB4ZlgB0F/J8u+WGKrwSSmenGjFLc6JyzwJfOtr4 CxZVxm9k//c9GmrmGGzlu+G8tAfi7WVaq+MFFE8XO8Akqhb6UexlgtkNI2nWY55FPMEM dPTw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1714146701; x=1714751501; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=aYJCmNFOVedHApS/WzQzrBs6k52ZDo9m1VFd7MFhABU=; b=pIWcryZAJT3r8pj/SIpA17wRcth7STVIHa+a0+qFmPcCBrFGD0xy2y0tIs9XIxURXY q3fE13c3mAa1pI1rRmvqnIbJTSwgrXkqCXgmMkHGN7uMN+ooAyFHNCfTm5o+xsunfAY0 Vy2nzJVGcO9Uy0K5ZbsywMOs0dtKq6Nnyy4L1wHQa9aLmDaib6WKV9WkLBFxvVKF5wnL lWq2wMk15pSY0cBMCKk52iFe7YSYQxauM7tub6E08WZWgsyndwPKjcHm1a+pR/W0freR 2FqzGN5BgD2YOQs3UJpwA/U5N+Xy3TRDW1a27hLIH1U6uz0HrFcpN0hmp5EMwRg+G7ea 5Zvg== X-Gm-Message-State: AOJu0YyTLINQHkjTUzyDs1f37x1zgJEAACzkS/zdCo343WyheQgorRjE 2IN18ksXR3SL2syEQF8YlhIYP6ObL5bSGWwgPGQp2zu20ZnJDUDwKq53sY6TW+Gnhp+Qwpd2SQD k774= X-Google-Smtp-Source: AGHT+IFXzbAXX/yrpmJwQdl5LAf0x5aoMqLRWMUf1RoQIgbrDTUAoACOKhH0r4Ww6YwpCdYxgixV2w== X-Received: by 2002:a17:90a:c292:b0:2ae:814f:43fa with SMTP id f18-20020a17090ac29200b002ae814f43famr2957130pjt.46.1714146700789; Fri, 26 Apr 2024 08:51:40 -0700 (PDT) Received: from hermes.local (204-195-96-226.wavecable.com. [204.195.96.226]) by smtp.gmail.com with ESMTPSA id ns5-20020a17090b250500b002a5f44353d2sm16770978pjb.7.2024.04.26.08.51.39 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 26 Apr 2024 08:51:40 -0700 (PDT) From: Stephen Hemminger To: dev@dpdk.org Cc: Stephen Hemminger Subject: [PATCH v9 6/9] net/tap: use libbpf to load new BPF program Date: Fri, 26 Apr 2024 08:48:56 -0700 Message-ID: <20240426155125.125910-7-stephen@networkplumber.org> X-Mailer: git-send-email 2.43.0 In-Reply-To: <20240426155125.125910-1-stephen@networkplumber.org> References: <20240130034925.44869-1-stephen@networkplumber.org> <20240426155125.125910-1-stephen@networkplumber.org> 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 There were multiple issues in the RSS queue support in the TAP driver. This required extensive rework of the BPF support. Change the BPF loading to use bpftool to create a skeleton header file, and load with libbpf. The BPF is always compiled from source so less chance that source and instructions diverge. Also resolves issue where libbpf and source get out of sync. The program is only loaded once, so if multiple rules are created only one BPF program is loaded in kernel. The new BPF program only needs a single action. No need for action and re-classification step. It also fixes the missing bits from the original. - supports setting RSS key per flow - level of hash can be L3 or L3/L4. Signed-off-by: Stephen Hemminger --- drivers/net/tap/bpf/meson.build | 36 ++- drivers/net/tap/bpf/tap_rss.c | 45 ++-- drivers/net/tap/meson.build | 35 ++- drivers/net/tap/rte_eth_tap.c | 14 +- drivers/net/tap/rte_eth_tap.h | 6 +- drivers/net/tap/tap_flow.c | 418 ++++++++------------------------ drivers/net/tap/tap_flow.h | 17 +- drivers/net/tap/tap_rss.h | 10 +- drivers/net/tap/tap_tcmsgs.h | 4 +- 9 files changed, 189 insertions(+), 396 deletions(-) diff --git a/drivers/net/tap/bpf/meson.build b/drivers/net/tap/bpf/meson.build index f2c03a19fd..45e8c3b498 100644 --- a/drivers/net/tap/bpf/meson.build +++ b/drivers/net/tap/bpf/meson.build @@ -1,17 +1,26 @@ # SPDX-License-Identifier: BSD-3-Clause # Copyright 2024 Stephen Hemminger -enable_tap_rss = false - -libbpf = dependency('libbpf', required: false, method: 'pkg-config') +# Loading BPF requires libbpf +# and the bpf_map__XXX API's were introduced in 0.8.0 +libbpf = dependency('libbpf', version: '>= 1.0', + required: false, method: 'pkg-config') if not libbpf.found() message('net/tap: no RSS support missing libbpf') subdir_done() endif +# Making skeleton needs bpftool # Debian install this in /usr/sbin which is not in $PATH -bpftool = find_program('bpftool', '/usr/sbin/bpftool', required: false, version: '>= 5.6.0') -if not bpftool.found() +bpftool_supports_skel = false +bpftool = find_program('bpftool', '/usr/sbin/bpftool', required: false) +if bpftool.found() + # Some Ubuntu versions have non-functional bpftool + bpftool_supports_skel = run_command(bpftool, 'gen', 'help', + check:false).returncode() == 0 +endif + +if not bpftool_supports_skel message('net/tap: no RSS support missing bpftool') subdir_done() endif @@ -42,6 +51,7 @@ clang_flags = [ '-O2', '-Wall', '-Wextra', + max_queues, '-target', 'bpf', '-g', @@ -67,6 +77,22 @@ skel_h_cmd = [ '@INPUT@' ] +vmlinux_h_cmd = [ + bpftool, + 'btf', + 'dump', + 'file', + '/sys/kernel/btf/vmlinux', + 'format', + 'c' +] + +vmlinux_h = custom_target( + 'vmlinux.h', + output: 'vmlinux.h', + command: vmlinux_h_cmd, + capture: true) + tap_rss_o = custom_target( 'tap_rss.bpf.o', input: 'tap_rss.c', diff --git a/drivers/net/tap/bpf/tap_rss.c b/drivers/net/tap/bpf/tap_rss.c index 888b3bdc24..1d71941c53 100644 --- a/drivers/net/tap/bpf/tap_rss.c +++ b/drivers/net/tap/bpf/tap_rss.c @@ -203,23 +203,6 @@ parse_ipv6(const struct __sk_buff *skb, __u32 hash_type, const __u32 *key) return 0; } -/* - * Compute RSS hash for packets. - * Returns 0 if no hash is possible. - */ -static __u32 __attribute__((always_inline)) -calculate_rss_hash(const struct __sk_buff *skb, const struct rss_key *rsskey) -{ - const __u32 *key = (const __u32 *)rsskey->key; - - if (skb->protocol == bpf_htons(ETH_P_IP)) - return parse_ipv4(skb, rsskey->hash_fields, key); - else if (skb->protocol == bpf_htons(ETH_P_IPV6)) - return parse_ipv6(skb, rsskey->hash_fields, key); - else - return 0; -} - /* * Scale value to be into range [0, n) * Assumes val is large (ie hash covers whole u32 range) @@ -244,20 +227,40 @@ SEC("action") int rss_flow_action(struct __sk_buff *skb) { const struct rss_key *rsskey; - __u32 mark = skb->mark; + const __u32 *key; + __be16 proto; + __u32 mark; __u32 hash; + __u16 queue; + + __builtin_preserve_access_index(({ + mark = skb->mark; + proto = skb->protocol; + })); /* Lookup RSS configuration for that BPF class */ rsskey = bpf_map_lookup_elem(&rss_map, &mark); if (rsskey == NULL) return TC_ACT_OK; - hash = calculate_rss_hash(skb, rsskey); - if (!hash) + key = (const __u32 *)rsskey->key; + + if (proto == bpf_htons(ETH_P_IP)) + hash = parse_ipv4(skb, rsskey->hash_fields, key); + else if (proto == bpf_htons(ETH_P_IPV6)) + hash = parse_ipv6(skb, rsskey->hash_fields, key); + else + hash = 0; + + if (hash == 0) return TC_ACT_OK; /* Fold hash to the number of queues configured */ - skb->queue_mapping = reciprocal_scale(hash, rsskey->nb_queues); + queue = reciprocal_scale(hash, rsskey->nb_queues); + + __builtin_preserve_access_index(({ + skb->queue_mapping = queue; + })); return TC_ACT_PIPE; } diff --git a/drivers/net/tap/meson.build b/drivers/net/tap/meson.build index 66647a1c62..26074059fa 100644 --- a/drivers/net/tap/meson.build +++ b/drivers/net/tap/meson.build @@ -5,36 +5,33 @@ if not is_linux build = false reason = 'only supported on Linux' endif + sources = files( 'rte_eth_tap.c', 'tap_intr.c', 'tap_netlink.c', ) +deps = ['bus_vdev', 'gso', 'hash'] + +max_queues = '-DTAP_MAX_QUEUES=16' +cflags += max_queues + +require_iova_in_mbuf = false + if cc.has_header_symbol('linux/pkt_cls.h', 'TCA_FLOWER_ACT') cflags += '-DHAVE_TCA_FLOWER' sources += files( - 'tap_bpf_api.c', 'tap_flow.c', 'tap_tcmsgs.c', ) -endif - -deps = ['bus_vdev', 'gso', 'hash'] -cflags += '-DTAP_MAX_QUEUES=16' - -# input array for meson symbol search: -# [ "MACRO to define if found", "header for the search", -# "enum/define", "symbol to search" ] -# -args = [ - [ 'HAVE_TC_ACT_BPF', 'linux/tc_act/tc_bpf.h', 'TCA_ACT_BPF_UNSPEC' ], -] -config = configuration_data() -foreach arg:args - config.set(arg[0], cc.has_header_symbol(arg[1], arg[2])) -endforeach -configure_file(output : 'tap_autoconf.h', configuration : config) + enable_tap_rss = false -require_iova_in_mbuf = false + subdir('bpf') + if enable_tap_rss + cflags += '-DHAVE_BPF_RSS' + ext_deps += libbpf + sources += tap_rss_skel_h + endif +endif diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c index 3f4fe95a09..793730de08 100644 --- a/drivers/net/tap/rte_eth_tap.c +++ b/drivers/net/tap/rte_eth_tap.c @@ -1131,6 +1131,7 @@ tap_dev_close(struct rte_eth_dev *dev) tap_flow_implicit_flush(internals, NULL); tap_nl_final(internals->nlsk_fd); internals->nlsk_fd = -1; + tap_flow_bpf_destroy(internals); } #endif @@ -1956,6 +1957,7 @@ eth_dev_tap_create(struct rte_vdev_device *vdev, const char *tap_name, strlcpy(pmd->name, tap_name, sizeof(pmd->name)); pmd->type = type; pmd->ka_fd = -1; + #ifdef HAVE_TCA_FLOWER pmd->nlsk_fd = -1; #endif @@ -2038,13 +2040,6 @@ eth_dev_tap_create(struct rte_vdev_device *vdev, const char *tap_name, /* Make network device persist after application exit */ pmd->persist = persist; - pmd->if_index = if_nametoindex(pmd->name); - if (!pmd->if_index) { - TAP_LOG(ERR, "%s: failed to get if_index.", pmd->name); - goto disable_rte_flow; - } - - #ifdef HAVE_TCA_FLOWER /* * Set up everything related to rte_flow: @@ -2060,6 +2055,11 @@ eth_dev_tap_create(struct rte_vdev_device *vdev, const char *tap_name, pmd->name); goto disable_rte_flow; } + pmd->if_index = if_nametoindex(pmd->name); + if (!pmd->if_index) { + TAP_LOG(ERR, "%s: failed to get if_index.", pmd->name); + goto disable_rte_flow; + } if (qdisc_create_multiq(pmd->nlsk_fd, pmd->if_index) < 0) { TAP_LOG(ERR, "%s: failed to create multiq qdisc.", pmd->name); diff --git a/drivers/net/tap/rte_eth_tap.h b/drivers/net/tap/rte_eth_tap.h index af18b29090..ce4322ad04 100644 --- a/drivers/net/tap/rte_eth_tap.h +++ b/drivers/net/tap/rte_eth_tap.h @@ -81,10 +81,8 @@ struct pmd_internals { #ifdef HAVE_TCA_FLOWER int nlsk_fd; /* Netlink socket fd */ int flow_isolate; /* 1 if flow isolation is enabled */ - int rss_enabled; /* 1 if RSS is enabled, else 0 */ - /* implicit rules set when RSS is enabled */ - int map_fd; /* BPF RSS map fd */ - int bpf_fd[RTE_PMD_TAP_MAX_QUEUES];/* List of bpf fds per queue */ + + struct tap_rss *rss; /* BPF program */ LIST_HEAD(tap_flows, rte_flow) flows; /* rte_flow rules */ /* implicit rte_flow rules set when a remote device is active */ diff --git a/drivers/net/tap/tap_flow.c b/drivers/net/tap/tap_flow.c index 5a5a4ea048..322153d806 100644 --- a/drivers/net/tap/tap_flow.c +++ b/drivers/net/tap/tap_flow.c @@ -15,25 +15,19 @@ #include #include #include +#include #include -#include #include #include -/* RSS key management */ -enum bpf_rss_key_e { - KEY_CMD_GET = 1, - KEY_CMD_RELEASE, - KEY_CMD_INIT, - KEY_CMD_DEINIT, -}; - -enum key_status_e { - KEY_STAT_UNSPEC, - KEY_STAT_USED, - KEY_STAT_AVAILABLE, -}; +#ifdef HAVE_BPF_RSS +/* Workaround for warning in bpftool generated skeleton code */ +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wcast-qual" +#include "tap_rss.skel.h" +#pragma GCC diagnostic pop +#endif #define ISOLATE_HANDLE 1 #define REMOTE_PROMISCUOUS_HANDLE 2 @@ -41,8 +35,6 @@ enum key_status_e { struct rte_flow { LIST_ENTRY(rte_flow) next; /* Pointer to the next rte_flow structure */ struct rte_flow *remote_flow; /* associated remote flow */ - int bpf_fd[SEC_MAX]; /* list of bfs fds per ELF section */ - uint32_t key_idx; /* RSS rule key index into BPF map */ struct nlmsg msg; }; @@ -69,12 +61,16 @@ struct action_data { struct skbedit { struct tc_skbedit skbedit; uint16_t queue; + uint32_t mark; } skbedit; +#ifdef HAVE_BPF_RSS struct bpf { struct tc_act_bpf bpf; + uint32_t map_key; int bpf_fd; const char *annotation; } bpf; +#endif }; }; @@ -112,13 +108,12 @@ tap_flow_isolate(struct rte_eth_dev *dev, int set, struct rte_flow_error *error); -static int bpf_rss_key(enum bpf_rss_key_e cmd, __u32 *key_idx); -static int rss_enable(struct pmd_internals *pmd, - const struct rte_flow_attr *attr, - struct rte_flow_error *error); +#ifdef HAVE_BPF_RSS +static int rss_enable(struct pmd_internals *pmd, struct rte_flow_error *error); static int rss_add_actions(struct rte_flow *flow, struct pmd_internals *pmd, const struct rte_flow_action_rss *rss, struct rte_flow_error *error); +#endif static const struct rte_flow_ops tap_flow_ops = { .validate = tap_flow_validate, @@ -853,11 +848,14 @@ add_action(struct rte_flow *flow, size_t *act_index, struct action_data *adata) &adata->mirred); } else if (strcmp("skbedit", adata->id) == 0) { tap_nlattr_add(&msg->nh, TCA_SKBEDIT_PARMS, - sizeof(adata->skbedit.skbedit), - &adata->skbedit.skbedit); - tap_nlattr_add16(&msg->nh, TCA_SKBEDIT_QUEUE_MAPPING, - adata->skbedit.queue); - } else if (strcmp("bpf", adata->id) == 0) { + sizeof(adata->skbedit.skbedit), &adata->skbedit.skbedit); + if (adata->skbedit.mark) + tap_nlattr_add32(&msg->nh, TCA_SKBEDIT_MARK, adata->skbedit.mark); + else + tap_nlattr_add16(&msg->nh, TCA_SKBEDIT_QUEUE_MAPPING, adata->skbedit.queue); + } +#ifdef HAVE_BPF_RSS + else if (strcmp("bpf", adata->id) == 0) { tap_nlattr_add32(&msg->nh, TCA_ACT_BPF_FD, adata->bpf.bpf_fd); tap_nlattr_add(&msg->nh, TCA_ACT_BPF_NAME, strlen(adata->bpf.annotation) + 1, @@ -865,7 +863,9 @@ add_action(struct rte_flow *flow, size_t *act_index, struct action_data *adata) tap_nlattr_add(&msg->nh, TCA_ACT_BPF_PARMS, sizeof(adata->bpf.bpf), &adata->bpf.bpf); - } else { + } +#endif + else { return -1; } tap_nlattr_nested_finish(msg); /* nested TCA_ACT_OPTIONS */ @@ -1104,8 +1104,7 @@ priv_flow_process(struct pmd_internals *pmd, }, }; - err = add_actions(flow, 1, &adata, - TCA_FLOWER_ACT); + err = add_actions(flow, 1, &adata, TCA_FLOWER_ACT); } } else if (actions->type == RTE_FLOW_ACTION_TYPE_QUEUE) { const struct rte_flow_action_queue *queue = @@ -1135,6 +1134,7 @@ priv_flow_process(struct pmd_internals *pmd, err = add_actions(flow, 1, &adata, TCA_FLOWER_ACT); } +#ifdef HAVE_BPF_RSS } else if (actions->type == RTE_FLOW_ACTION_TYPE_RSS) { const struct rte_flow_action_rss *rss = (const struct rte_flow_action_rss *) @@ -1143,13 +1143,14 @@ priv_flow_process(struct pmd_internals *pmd, if (action++) goto exit_action_not_supported; - if (!pmd->rss_enabled) { - err = rss_enable(pmd, attr, error); + if (pmd->rss == NULL) { + err = rss_enable(pmd, error); if (err) goto exit_return_error; } if (flow) err = rss_add_actions(flow, pmd, rss, error); +#endif } else { goto exit_action_not_supported; } @@ -1246,26 +1247,17 @@ tap_flow_set_handle(struct rte_flow *flow) * */ static void -tap_flow_free(struct pmd_internals *pmd, struct rte_flow *flow) +tap_flow_free(struct pmd_internals *pmd __rte_unused, struct rte_flow *flow) { - int i; - if (!flow) return; - if (pmd->rss_enabled) { - /* Close flow BPF file descriptors */ - for (i = 0; i < SEC_MAX; i++) - if (flow->bpf_fd[i] != 0) { - close(flow->bpf_fd[i]); - flow->bpf_fd[i] = 0; - } - - /* Release the map key for this RSS rule */ - bpf_rss_key(KEY_CMD_RELEASE, &flow->key_idx); - flow->key_idx = 0; - } - +#ifdef HAVE_BPF_RSS + struct tap_rss *rss = pmd->rss; + if (rss) + bpf_map__delete_elem(rss->maps.rss_map, + &flow->msg.t.tcm_handle, sizeof(uint32_t), 0); +#endif /* Free flow allocated memory */ rte_free(flow); } @@ -1733,14 +1725,18 @@ tap_flow_implicit_flush(struct pmd_internals *pmd, struct rte_flow_error *error) return 0; } -#define MAX_RSS_KEYS 256 -#define KEY_IDX_OFFSET (3 * MAX_RSS_KEYS) -#define SEC_NAME_CLS_Q "cls_q" - -static const char *sec_name[SEC_MAX] = { - [SEC_L3_L4] = "l3_l4", -}; +/** + * Cleanup when device is closed + */ +void tap_flow_bpf_destroy(struct pmd_internals *pmd __rte_unused) +{ +#ifdef HAVE_BPF_RSS + tap_rss__destroy(pmd->rss); + pmd->rss = NULL; +#endif +} +#ifdef HAVE_BPF_RSS /** * Enable RSS on tap: create TC rules for queuing. * @@ -1755,225 +1751,32 @@ static const char *sec_name[SEC_MAX] = { * * @return 0 on success, negative value on failure. */ -static int rss_enable(struct pmd_internals *pmd, - const struct rte_flow_attr *attr, - struct rte_flow_error *error) +static int rss_enable(struct pmd_internals *pmd, struct rte_flow_error *error) { - struct rte_flow *rss_flow = NULL; - struct nlmsg *msg = NULL; - /* 4096 is the maximum number of instructions for a BPF program */ - char annotation[64]; - int i; - int err = 0; - - /* unlimit locked memory */ - struct rlimit memlock_limit = { - .rlim_cur = RLIM_INFINITY, - .rlim_max = RLIM_INFINITY, - }; - setrlimit(RLIMIT_MEMLOCK, &memlock_limit); - - /* Get a new map key for a new RSS rule */ - err = bpf_rss_key(KEY_CMD_INIT, NULL); - if (err < 0) { - rte_flow_error_set( - error, EINVAL, RTE_FLOW_ERROR_TYPE_HANDLE, NULL, - "Failed to initialize BPF RSS keys"); - - return -1; - } - - /* - * Create BPF RSS MAP - */ - pmd->map_fd = tap_flow_bpf_rss_map_create(sizeof(__u32), /* key size */ - sizeof(struct rss_key), - MAX_RSS_KEYS); - if (pmd->map_fd < 0) { - TAP_LOG(ERR, - "Failed to create BPF map (%d): %s", - errno, strerror(errno)); - rte_flow_error_set( - error, ENOTSUP, RTE_FLOW_ERROR_TYPE_HANDLE, NULL, - "Kernel too old or not configured " - "to support BPF maps"); + int err; - return -ENOTSUP; + /* Load the BPF program (defined in tap_bpf.h from skeleton) */ + pmd->rss = tap_rss__open_and_load(); + if (pmd->rss == NULL) { + TAP_LOG(ERR, "Failed to load BPF object: %s", strerror(errno)); + rte_flow_error_set(error, errno, RTE_FLOW_ERROR_TYPE_HANDLE, NULL, + "BPF object could not be loaded"); + return -errno; } - /* - * Add a rule per queue to match reclassified packets and direct them to - * the correct queue. - */ - for (i = 0; i < pmd->dev->data->nb_rx_queues; i++) { - pmd->bpf_fd[i] = tap_flow_bpf_cls_q(i); - if (pmd->bpf_fd[i] < 0) { - TAP_LOG(ERR, - "Failed to load BPF section %s for queue %d", - SEC_NAME_CLS_Q, i); - rte_flow_error_set( - error, ENOTSUP, RTE_FLOW_ERROR_TYPE_HANDLE, - NULL, - "Kernel too old or not configured " - "to support BPF programs loading"); - - return -ENOTSUP; - } - - rss_flow = rte_zmalloc(__func__, sizeof(struct rte_flow), 0); - if (!rss_flow) { - TAP_LOG(ERR, - "Cannot allocate memory for rte_flow"); - return -1; - } - msg = &rss_flow->msg; - tc_init_msg(msg, pmd->if_index, RTM_NEWTFILTER, NLM_F_REQUEST | - NLM_F_ACK | NLM_F_EXCL | NLM_F_CREATE); - msg->t.tcm_info = TC_H_MAKE(0, htons(ETH_P_ALL)); - tap_flow_set_handle(rss_flow); - uint16_t group = attr->group << GROUP_SHIFT; - uint16_t prio = group | (i + PRIORITY_OFFSET); - msg->t.tcm_info = TC_H_MAKE(prio << 16, msg->t.tcm_info); - msg->t.tcm_parent = TC_H_MAKE(MULTIQ_MAJOR_HANDLE, 0); - - tap_nlattr_add(&msg->nh, TCA_KIND, sizeof("bpf"), "bpf"); - if (tap_nlattr_nested_start(msg, TCA_OPTIONS) < 0) - return -1; - tap_nlattr_add32(&msg->nh, TCA_BPF_FD, pmd->bpf_fd[i]); - snprintf(annotation, sizeof(annotation), "[%s%d]", - SEC_NAME_CLS_Q, i); - tap_nlattr_add(&msg->nh, TCA_BPF_NAME, strlen(annotation) + 1, - annotation); - /* Actions */ - { - struct action_data adata = { - .id = "skbedit", - .skbedit = { - .skbedit = { - .action = TC_ACT_PIPE, - }, - .queue = i, - }, - }; - if (add_actions(rss_flow, 1, &adata, TCA_BPF_ACT) < 0) - return -1; - } - tap_nlattr_nested_finish(msg); /* nested TCA_OPTIONS */ - - /* Netlink message is now ready to be sent */ - if (tap_nl_send(pmd->nlsk_fd, &msg->nh) < 0) - return -1; - err = tap_nl_recv_ack(pmd->nlsk_fd); - if (err < 0) { - TAP_LOG(ERR, - "Kernel refused TC filter rule creation (%d): %s", - errno, strerror(errno)); - return err; - } - } - - pmd->rss_enabled = 1; - return err; -} - -/** - * Manage bpf RSS keys repository with operations: init, get, release - * - * @param[in] cmd - * Command on RSS keys: init, get, release - * - * @param[in, out] key_idx - * Pointer to RSS Key index (out for get command, in for release command) - * - * @return -1 if couldn't get, release or init the RSS keys, 0 otherwise. - */ -static int bpf_rss_key(enum bpf_rss_key_e cmd, __u32 *key_idx) -{ - __u32 i; - int err = 0; - static __u32 num_used_keys; - static __u32 rss_keys[MAX_RSS_KEYS] = {KEY_STAT_UNSPEC}; - static __u32 rss_keys_initialized; - __u32 key; - - switch (cmd) { - case KEY_CMD_GET: - if (!rss_keys_initialized) { - err = -1; - break; - } - - if (num_used_keys == RTE_DIM(rss_keys)) { - err = -1; - break; - } - - *key_idx = num_used_keys % RTE_DIM(rss_keys); - while (rss_keys[*key_idx] == KEY_STAT_USED) - *key_idx = (*key_idx + 1) % RTE_DIM(rss_keys); - - rss_keys[*key_idx] = KEY_STAT_USED; - - /* - * Add an offset to key_idx in order to handle a case of - * RSS and non RSS flows mixture. - * If a non RSS flow is destroyed it has an eBPF map - * index 0 (initialized on flow creation) and might - * unintentionally remove RSS entry 0 from eBPF map. - * To avoid this issue, add an offset to the real index - * during a KEY_CMD_GET operation and subtract this offset - * during a KEY_CMD_RELEASE operation in order to restore - * the real index. - */ - *key_idx += KEY_IDX_OFFSET; - num_used_keys++; - break; - - case KEY_CMD_RELEASE: - if (!rss_keys_initialized) - break; - - /* - * Subtract offset to restore real key index - * If a non RSS flow is falsely trying to release map - * entry 0 - the offset subtraction will calculate the real - * map index as an out-of-range value and the release operation - * will be silently ignored. - */ - key = *key_idx - KEY_IDX_OFFSET; - if (key >= RTE_DIM(rss_keys)) - break; - - if (rss_keys[key] == KEY_STAT_USED) { - rss_keys[key] = KEY_STAT_AVAILABLE; - num_used_keys--; - } - break; - - case KEY_CMD_INIT: - for (i = 0; i < RTE_DIM(rss_keys); i++) - rss_keys[i] = KEY_STAT_AVAILABLE; - - rss_keys_initialized = 1; - num_used_keys = 0; - break; - - case KEY_CMD_DEINIT: - for (i = 0; i < RTE_DIM(rss_keys); i++) - rss_keys[i] = KEY_STAT_UNSPEC; - - rss_keys_initialized = 0; - num_used_keys = 0; - break; - - default: - break; + /* Attach the maps defined in BPF program */ + err = tap_rss__attach(pmd->rss); + if (err < 0) { + TAP_LOG(ERR, "Failed to attach BPF object: %d", err); + rte_flow_error_set(error, -err, RTE_FLOW_ERROR_TYPE_HANDLE, NULL, + "BPF object could not be attached"); + tap_flow_bpf_destroy(pmd); + return err; } - return err; + return 0; } - /* Default RSS hash key also used by mlx devices */ static const uint8_t rss_hash_default_key[] = { 0x2c, 0xc6, 0x81, 0xd1, @@ -2006,9 +1809,11 @@ static int rss_add_actions(struct rte_flow *flow, struct pmd_internals *pmd, const struct rte_flow_action_rss *rss, struct rte_flow_error *error) { + const struct bpf_program *rss_prog = pmd->rss->progs.rss_flow_action; struct rss_key rss_entry = { }; const uint8_t *key_in; uint32_t hash_type = 0; + uint32_t handle = flow->msg.t.tcm_handle; unsigned int i; int err; @@ -2057,34 +1862,24 @@ static int rss_add_actions(struct rte_flow *flow, struct pmd_internals *pmd, else if (rss->types & (RTE_ETH_RSS_IPV6 | RTE_ETH_RSS_FRAG_IPV6 | RTE_ETH_RSS_IPV6_EX)) hash_type |= RTE_BIT32(HASH_FIELD_IPV6_L3); - /* Get a new map key for a new RSS rule */ - err = bpf_rss_key(KEY_CMD_GET, &flow->key_idx); - if (err < 0) { - rte_flow_error_set( - error, EINVAL, RTE_FLOW_ERROR_TYPE_HANDLE, NULL, - "Failed to get BPF RSS key"); - - return -1; - } + rss_entry.hash_fields = hash_type; + rte_convert_rss_key((const uint32_t *)key_in, (uint32_t *)rss_entry.key, + TAP_RSS_HASH_KEY_SIZE); /* Update RSS map entry with queues */ rss_entry.nb_queues = rss->queue_num; for (i = 0; i < rss->queue_num; i++) rss_entry.queues[i] = rss->queue[i]; - rss_entry.hash_fields = hash_type; - rte_convert_rss_key((const uint32_t *)key_in, (uint32_t *)rss_entry.key, - TAP_RSS_HASH_KEY_SIZE); - - - /* Add this RSS entry to map */ - err = tap_flow_bpf_update_rss_elem(pmd->map_fd, - &flow->key_idx, &rss_entry); + /* Add this way for BPF to find entry in map */ + err = bpf_map__update_elem(pmd->rss->maps.rss_map, + &handle, sizeof(handle), + &rss_entry, sizeof(rss_entry), 0); if (err) { TAP_LOG(ERR, - "Failed to update BPF map entry #%u (%d): %s", - flow->key_idx, errno, strerror(errno)); + "Failed to update BPF map entry %#x (%d): %s", + handle, errno, strerror(errno)); rte_flow_error_set( error, ENOTSUP, RTE_FLOW_ERROR_TYPE_HANDLE, NULL, "Kernel too old or not configured " @@ -2093,47 +1888,28 @@ static int rss_add_actions(struct rte_flow *flow, struct pmd_internals *pmd, return -ENOTSUP; } - - /* - * Load bpf rules to calculate hash for this key_idx - */ - - flow->bpf_fd[SEC_L3_L4] = - tap_flow_bpf_calc_l3_l4_hash(flow->key_idx, pmd->map_fd); - if (flow->bpf_fd[SEC_L3_L4] < 0) { - TAP_LOG(ERR, - "Failed to load BPF section %s (%d): %s", - sec_name[SEC_L3_L4], errno, strerror(errno)); - rte_flow_error_set( - error, ENOTSUP, RTE_FLOW_ERROR_TYPE_HANDLE, NULL, - "Kernel too old or not configured " - "to support BPF program loading"); - - return -ENOTSUP; - } - - /* Actions */ - { - struct action_data adata[] = { - { - .id = "bpf", - .bpf = { - .bpf_fd = flow->bpf_fd[SEC_L3_L4], - .annotation = sec_name[SEC_L3_L4], - .bpf = { - .action = TC_ACT_PIPE, - }, - }, + /* Add actions to mark packet then run the RSS BPF program */ + struct action_data adata[] = { + { + .id = "skbedit", + .skbedit = { + .skbedit.action = TC_ACT_PIPE, + .mark = handle, }, - }; - - if (add_actions(flow, RTE_DIM(adata), adata, - TCA_FLOWER_ACT) < 0) - return -1; - } + }, + { + .id = "bpf", + .bpf = { + .bpf.action = TC_ACT_PIPE, + .annotation = "tap_rss", + .bpf_fd = bpf_program__fd(rss_prog), + }, + }, + }; - return 0; + return add_actions(flow, RTE_DIM(adata), adata, TCA_FLOWER_ACT); } +#endif /** * Get rte_flow operations. diff --git a/drivers/net/tap/tap_flow.h b/drivers/net/tap/tap_flow.h index 240fbc3dfa..8b19347a93 100644 --- a/drivers/net/tap/tap_flow.h +++ b/drivers/net/tap/tap_flow.h @@ -9,7 +9,11 @@ #include #include #include -#include + +/** + * Mask of unsupported RSS types + */ +#define TAP_RSS_HF_MASK (~(RTE_ETH_RSS_IP | RTE_ETH_RSS_UDP | RTE_ETH_RSS_TCP)) /** * In TC, priority 0 means we require the kernel to allocate one for us. @@ -41,11 +45,6 @@ enum implicit_rule_index { TAP_REMOTE_MAX_IDX, }; -enum bpf_fd_idx { - SEC_L3_L4, - SEC_MAX, -}; - int tap_dev_flow_ops_get(struct rte_eth_dev *dev, const struct rte_flow_ops **ops); int tap_flow_flush(struct rte_eth_dev *dev, struct rte_flow_error *error); @@ -57,10 +56,6 @@ int tap_flow_implicit_destroy(struct pmd_internals *pmd, int tap_flow_implicit_flush(struct pmd_internals *pmd, struct rte_flow_error *error); -int tap_flow_bpf_cls_q(__u32 queue_idx); -int tap_flow_bpf_calc_l3_l4_hash(__u32 key_idx, int map_fd); -int tap_flow_bpf_rss_map_create(unsigned int key_size, unsigned int value_size, - unsigned int max_entries); -int tap_flow_bpf_update_rss_elem(int fd, void *key, void *value); +void tap_flow_bpf_destroy(struct pmd_internals *pmd); #endif /* _TAP_FLOW_H_ */ diff --git a/drivers/net/tap/tap_rss.h b/drivers/net/tap/tap_rss.h index 6009be7031..65bd8991b1 100644 --- a/drivers/net/tap/tap_rss.h +++ b/drivers/net/tap/tap_rss.h @@ -5,16 +5,14 @@ #ifndef _TAP_RSS_H_ #define _TAP_RSS_H_ -#ifndef TAP_MAX_QUEUES -#define TAP_MAX_QUEUES 16 +/* Size of the map from BPF classid to queue table */ +#ifndef TAP_RSS_MAX +#define TAP_RSS_MAX 32 #endif -/* Fixed RSS hash key size in bytes. */ +/* Standard Toeplitz hash key size */ #define TAP_RSS_HASH_KEY_SIZE 40 -/* Supported RSS */ -#define TAP_RSS_HF_MASK (~(RTE_ETH_RSS_IP | RTE_ETH_RSS_UDP | RTE_ETH_RSS_TCP)) - /* hashed fields for RSS */ enum hash_field { HASH_FIELD_IPV4_L3, /* IPv4 src/dst addr */ diff --git a/drivers/net/tap/tap_tcmsgs.h b/drivers/net/tap/tap_tcmsgs.h index a64cb29d6f..9411626661 100644 --- a/drivers/net/tap/tap_tcmsgs.h +++ b/drivers/net/tap/tap_tcmsgs.h @@ -6,7 +6,6 @@ #ifndef _TAP_TCMSGS_H_ #define _TAP_TCMSGS_H_ -#include #include #include #include @@ -14,9 +13,10 @@ #include #include #include -#ifdef HAVE_TC_ACT_BPF +#ifdef HAVE_BPF_RSS #include #endif + #include #include -- 2.43.0