From patchwork Thu Aug 24 13:21:50 2017 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Saeed Mahameed X-Patchwork-Id: 805485 X-Patchwork-Delegate: davem@davemloft.net Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Authentication-Results: ozlabs.org; spf=none (mailfrom) smtp.mailfrom=vger.kernel.org (client-ip=209.132.180.67; helo=vger.kernel.org; envelope-from=netdev-owner@vger.kernel.org; receiver=) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id 3xdQ3M1ldXz9s7m for ; Thu, 24 Aug 2017 23:23:27 +1000 (AEST) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753149AbdHXNW6 (ORCPT ); Thu, 24 Aug 2017 09:22:58 -0400 Received: from mail-il-dmz.mellanox.com ([193.47.165.129]:34102 "EHLO mellanox.co.il" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1751116AbdHXNV6 (ORCPT ); Thu, 24 Aug 2017 09:21:58 -0400 Received: from Internal Mail-Server by MTLPINE1 (envelope-from saeedm@mellanox.com) with ESMTPS (AES256-SHA encrypted); 24 Aug 2017 16:21:55 +0300 Received: from sws.mtl.labs.mlnx (reg-l-vrt-045-015.mtl.labs.mlnx [10.135.45.15]) by labmailer.mlnx (8.13.8/8.13.8) with ESMTP id v7ODLsQv020626; Thu, 24 Aug 2017 16:21:55 +0300 From: Saeed Mahameed To: "David S. Miller" Cc: netdev@vger.kernel.org, Leon Romanovsky , Matan Barak , Saeed Mahameed Subject: [net-next 6/8] net/mlx5: Add hash table to search FTEs in a flow-group Date: Thu, 24 Aug 2017 16:21:50 +0300 Message-Id: <20170824132152.29538-7-saeedm@mellanox.com> X-Mailer: git-send-email 2.13.0 In-Reply-To: <20170824132152.29538-1-saeedm@mellanox.com> References: <20170824132152.29538-1-saeedm@mellanox.com> Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org From: Matan Barak When adding a flow table entry (fte) to a flow group (fg), we first need to check whether this fte exist. In such a case we just merge the destinations (if possible). Currently, this is done by traversing the fte list available in a fg. This could take a lot of time when using large flow groups. Speeding this up by using rhashtable, which is much faster. Signed-off-by: Matan Barak Reviewed-by: Maor Gottlieb Signed-off-by: Saeed Mahameed --- drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 228 +++++++++++++++------- drivers/net/ethernet/mellanox/mlx5/core/fs_core.h | 3 + 2 files changed, 160 insertions(+), 71 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c index 986c5e50e872..d8d45b006996 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c @@ -150,6 +150,14 @@ enum fs_i_mutex_lock_class { FS_MUTEX_CHILD }; +static const struct rhashtable_params rhash_fte = { + .key_len = FIELD_SIZEOF(struct fs_fte, val), + .key_offset = offsetof(struct fs_fte, val), + .head_offset = offsetof(struct fs_fte, hash), + .automatic_shrinking = true, + .min_size = 1, +}; + static void del_rule(struct fs_node *node); static void del_flow_table(struct fs_node *node); static void del_flow_group(struct fs_node *node); @@ -255,63 +263,59 @@ static struct fs_prio *find_prio(struct mlx5_flow_namespace *ns, return NULL; } -static bool masked_memcmp(void *mask, void *val1, void *val2, size_t size) +static bool check_last_reserved(const u32 *match_criteria) { - unsigned int i; - - for (i = 0; i < size; i++, mask++, val1++, val2++) - if ((*((u8 *)val1) & (*(u8 *)mask)) != - ((*(u8 *)val2) & (*(u8 *)mask))) - return false; + char *match_criteria_reserved = + MLX5_ADDR_OF(fte_match_param, match_criteria, MLX5_FTE_MATCH_PARAM_RESERVED); - return true; + return !match_criteria_reserved[0] && + !memcmp(match_criteria_reserved, match_criteria_reserved + 1, + MLX5_FLD_SZ_BYTES(fte_match_param, + MLX5_FTE_MATCH_PARAM_RESERVED) - 1); } -static bool compare_match_value(struct mlx5_flow_group_mask *mask, - void *fte_param1, void *fte_param2) +static bool check_valid_mask(u8 match_criteria_enable, const u32 *match_criteria) { - if (mask->match_criteria_enable & - 1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_OUTER_HEADERS) { - void *fte_match1 = MLX5_ADDR_OF(fte_match_param, - fte_param1, outer_headers); - void *fte_match2 = MLX5_ADDR_OF(fte_match_param, - fte_param2, outer_headers); - void *fte_mask = MLX5_ADDR_OF(fte_match_param, - mask->match_criteria, outer_headers); + if (match_criteria_enable & ~( + (1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_OUTER_HEADERS) | + (1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_MISC_PARAMETERS) | + (1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_INNER_HEADERS))) + return false; - if (!masked_memcmp(fte_mask, fte_match1, fte_match2, - MLX5_ST_SZ_BYTES(fte_match_set_lyr_2_4))) + if (!(match_criteria_enable & + 1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_OUTER_HEADERS)) { + char *fg_type_mask = MLX5_ADDR_OF(fte_match_param, + match_criteria, outer_headers); + + if (fg_type_mask[0] || + memcmp(fg_type_mask, fg_type_mask + 1, + MLX5_ST_SZ_BYTES(fte_match_set_lyr_2_4) - 1)) return false; } - if (mask->match_criteria_enable & - 1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_MISC_PARAMETERS) { - void *fte_match1 = MLX5_ADDR_OF(fte_match_param, - fte_param1, misc_parameters); - void *fte_match2 = MLX5_ADDR_OF(fte_match_param, - fte_param2, misc_parameters); - void *fte_mask = MLX5_ADDR_OF(fte_match_param, - mask->match_criteria, misc_parameters); + if (!(match_criteria_enable & + 1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_MISC_PARAMETERS)) { + char *fg_type_mask = MLX5_ADDR_OF(fte_match_param, + match_criteria, misc_parameters); - if (!masked_memcmp(fte_mask, fte_match1, fte_match2, - MLX5_ST_SZ_BYTES(fte_match_set_misc))) + if (fg_type_mask[0] || + memcmp(fg_type_mask, fg_type_mask + 1, + MLX5_ST_SZ_BYTES(fte_match_set_misc) - 1)) return false; } - if (mask->match_criteria_enable & - 1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_INNER_HEADERS) { - void *fte_match1 = MLX5_ADDR_OF(fte_match_param, - fte_param1, inner_headers); - void *fte_match2 = MLX5_ADDR_OF(fte_match_param, - fte_param2, inner_headers); - void *fte_mask = MLX5_ADDR_OF(fte_match_param, - mask->match_criteria, inner_headers); + if (!(match_criteria_enable & + 1 << MLX5_CREATE_FLOW_GROUP_IN_MATCH_CRITERIA_ENABLE_INNER_HEADERS)) { + char *fg_type_mask = MLX5_ADDR_OF(fte_match_param, + match_criteria, inner_headers); - if (!masked_memcmp(fte_mask, fte_match1, fte_match2, - MLX5_ST_SZ_BYTES(fte_match_set_lyr_2_4))) + if (fg_type_mask[0] || + memcmp(fg_type_mask, fg_type_mask + 1, + MLX5_ST_SZ_BYTES(fte_match_set_lyr_2_4) - 1)) return false; } - return true; + + return check_last_reserved(match_criteria); } static bool compare_match_criteria(u8 match_criteria_enable1, @@ -410,6 +414,18 @@ static void del_rule(struct fs_node *node) } } +static void destroy_fte(struct fs_fte *fte, struct mlx5_flow_group *fg) +{ + struct mlx5_flow_table *ft; + int ret; + + ret = rhashtable_remove_fast(&fg->ftes_hash, &fte->hash, rhash_fte); + WARN_ON(ret); + fte->status = 0; + fs_get_obj(ft, fg->node.parent); + ida_simple_remove(&ft->fte_allocator, fte->index); +} + static void del_fte(struct fs_node *node) { struct mlx5_flow_table *ft; @@ -430,8 +446,7 @@ static void del_fte(struct fs_node *node) "flow steering can't delete fte in index %d of flow group id %d\n", fte->index, fg->id); - ida_simple_remove(&ft->fte_allocator, fte->index); - fte->status = 0; + destroy_fte(fte, fg); } static void del_flow_group(struct fs_node *node) @@ -447,6 +462,7 @@ static void del_flow_group(struct fs_node *node) if (ft->autogroup.active) ft->autogroup.num_groups--; + rhashtable_destroy(&fg->ftes_hash); if (mlx5_cmd_destroy_flow_group(dev, ft, fg->id)) mlx5_core_warn(dev, "flow steering can't destroy fg %d of ft %d\n", fg->id, ft->id); @@ -481,10 +497,17 @@ static struct mlx5_flow_group *alloc_flow_group(u32 *create_fg_in) u8 match_criteria_enable = MLX5_GET(create_flow_group_in, create_fg_in, match_criteria_enable); + int ret; + fg = kzalloc(sizeof(*fg), GFP_KERNEL); if (!fg) return ERR_PTR(-ENOMEM); + ret = rhashtable_init(&fg->ftes_hash, &rhash_fte); + if (ret) { + kfree(fg); + return ERR_PTR(ret); + } fg->mask.match_criteria_enable = match_criteria_enable; memcpy(&fg->mask.match_criteria, match_criteria, sizeof(fg->mask.match_criteria)); @@ -920,10 +943,8 @@ static struct mlx5_flow_group *create_flow_group_common(struct mlx5_flow_table * return fg; err = mlx5_cmd_create_flow_group(dev, ft, fg_in, &fg->id); - if (err) { - kfree(fg); - return ERR_PTR(err); - } + if (err) + goto err_free_fg; if (ft->autogroup.active) ft->autogroup.num_groups++; @@ -934,13 +955,27 @@ static struct mlx5_flow_group *create_flow_group_common(struct mlx5_flow_table * list_add(&fg->node.list, prev_fg); return fg; + +err_free_fg: + rhashtable_destroy(&fg->ftes_hash); + kfree(fg); + + return ERR_PTR(err); } struct mlx5_flow_group *mlx5_create_flow_group(struct mlx5_flow_table *ft, u32 *fg_in) { + void *match_criteria = MLX5_ADDR_OF(create_flow_group_in, + fg_in, match_criteria); + u8 match_criteria_enable = MLX5_GET(create_flow_group_in, + fg_in, + match_criteria_enable); struct mlx5_flow_group *fg; + if (!check_valid_mask(match_criteria_enable, match_criteria)) + return ERR_PTR(-EINVAL); + if (ft->autogroup.active) return ERR_PTR(-EPERM); @@ -1104,6 +1139,7 @@ static struct fs_fte *create_fte(struct mlx5_flow_group *fg, struct mlx5_flow_table *ft; struct fs_fte *fte; int index; + int ret; fs_get_obj(ft, fg->node.parent); index = ida_simple_get(&ft->fte_allocator, fg->start_index, @@ -1114,11 +1150,20 @@ static struct fs_fte *create_fte(struct mlx5_flow_group *fg, fte = alloc_fte(flow_act, match_value, index); if (IS_ERR(fte)) { - ida_simple_remove(&ft->fte_allocator, index); - return fte; + ret = PTR_ERR(fte); + goto err_alloc; } + ret = rhashtable_insert_fast(&fg->ftes_hash, &fte->hash, rhash_fte); + if (ret) + goto err_hash; return fte; + +err_hash: + kfree(fte); +err_alloc: + ida_simple_remove(&ft->fte_allocator, index); + return ERR_PTR(ret); } static struct mlx5_flow_group *create_autogroup(struct mlx5_flow_table *ft, @@ -1206,42 +1251,78 @@ static struct mlx5_flow_rule *find_flow_rule(struct fs_fte *fte, return NULL; } +static bool check_conflicting_actions(u32 action1, u32 action2) +{ + u32 xored_actions = action1 ^ action2; + + /* if one rule only wants to count, it's ok */ + if (action1 == MLX5_FLOW_CONTEXT_ACTION_COUNT || + action2 == MLX5_FLOW_CONTEXT_ACTION_COUNT) + return false; + + if (xored_actions & (MLX5_FLOW_CONTEXT_ACTION_DROP | + MLX5_FLOW_CONTEXT_ACTION_ENCAP | + MLX5_FLOW_CONTEXT_ACTION_DECAP)) + return true; + + return false; +} + +static int check_conflicting_ftes(struct fs_fte *fte, const struct mlx5_flow_act *flow_act) +{ + if (check_conflicting_actions(flow_act->action, fte->action)) { + mlx5_core_warn(get_dev(&fte->node), + "Found two FTEs with conflicting actions\n"); + return -EEXIST; + } + + if (fte->flow_tag != flow_act->flow_tag) { + mlx5_core_warn(get_dev(&fte->node), + "FTE flow tag %u already exists with different flow tag %u\n", + fte->flow_tag, + flow_act->flow_tag); + return -EEXIST; + } + + return 0; +} + static struct mlx5_flow_handle *add_rule_fg(struct mlx5_flow_group *fg, u32 *match_value, struct mlx5_flow_act *flow_act, struct mlx5_flow_destination *dest, int dest_num) { + u32 masked_val[sizeof(fg->mask.match_criteria)]; struct mlx5_flow_handle *handle; struct mlx5_flow_table *ft; struct fs_fte *fte; int i; nested_lock_ref_node(&fg->node, FS_MUTEX_PARENT); - fs_for_each_fte(fte, fg) { + for (i = 0; i < sizeof(masked_val); i++) + masked_val[i] = match_value[i] & fg->mask.match_criteria[i]; + fte = rhashtable_lookup_fast(&fg->ftes_hash, masked_val, rhash_fte); + if (fte) { + int old_action; + int ret; + nested_lock_ref_node(&fte->node, FS_MUTEX_CHILD); - if (compare_match_value(&fg->mask, match_value, &fte->val) && - (flow_act->action & fte->action)) { - int old_action = fte->action; - - if (fte->flow_tag != flow_act->flow_tag) { - mlx5_core_warn(get_dev(&fte->node), - "FTE flow tag %u already exists with different flow tag %u\n", - fte->flow_tag, - flow_act->flow_tag); - handle = ERR_PTR(-EEXIST); - goto unlock_fte; - } + ret = check_conflicting_ftes(fte, flow_act); + if (ret) { + handle = ERR_PTR(ret); + goto unlock_fte; + } - fte->action |= flow_act->action; - handle = add_rule_fte(fte, fg, dest, dest_num, - old_action != flow_act->action); - if (IS_ERR(handle)) { - fte->action = old_action; - goto unlock_fte; - } else { - goto add_rules; - } + old_action = fte->action; + fte->action |= flow_act->action; + handle = add_rule_fte(fte, fg, dest, dest_num, + old_action != flow_act->action); + if (IS_ERR(handle)) { + fte->action = old_action; + goto unlock_fte; + } else { + goto add_rules; } unlock_ref_node(&fte->node); } @@ -1257,6 +1338,7 @@ static struct mlx5_flow_handle *add_rule_fg(struct mlx5_flow_group *fg, handle = add_rule_fte(fte, fg, dest, dest_num, false); if (IS_ERR(handle)) { unlock_ref_node(&fte->node); + destroy_fte(fte, fg); kfree(fte); goto unlock_fg; } @@ -1332,6 +1414,10 @@ _mlx5_add_flow_rules(struct mlx5_flow_table *ft, struct mlx5_flow_handle *rule; int i; + if (!check_valid_mask(spec->match_criteria_enable, + spec->match_criteria)) + return ERR_PTR(-EINVAL); + for (i = 0; i < dest_num; i++) { if (!dest_is_valid(&dest[i], flow_act->action, ft)) return ERR_PTR(-EINVAL); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h index bfbc081d17dc..62709a3865d2 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.h @@ -34,6 +34,7 @@ #define _MLX5_FS_CORE_ #include +#include enum fs_node_type { FS_TYPE_NAMESPACE, @@ -168,6 +169,7 @@ struct fs_fte { u32 modify_id; enum fs_fte_status status; struct mlx5_fc *counter; + struct rhash_head hash; }; /* Type of children is mlx5_flow_table/namespace */ @@ -197,6 +199,7 @@ struct mlx5_flow_group { u32 start_index; u32 max_ftes; u32 id; + struct rhashtable ftes_hash; }; struct mlx5_flow_root_namespace {