diff mbox series

[net-next,09/15] net/mlx5: HWS, num_of_rules counter on matcher should be atomic

Message ID 20250102181415.1477316-10-tariqt@nvidia.com (mailing list archive)
State New
Delegated to: Netdev Maintainers
Headers show
Series mlx5 Hardware Steering part 2 | expand

Checks

Context Check Description
netdev/series_format success Posting correctly formatted
netdev/tree_selection success Clearly marked for net-next
netdev/ynl success Generated files up to date; no warnings/errors; no diff in generated;
netdev/fixes_present success Fixes tag not required for -next series
netdev/header_inline success No static functions without inline keyword in header files
netdev/build_32bit success Errors and warnings before: 1 this patch: 1
netdev/build_tools success No tools touched, skip
netdev/cc_maintainers warning 1 maintainers not CCed: linux-rdma@vger.kernel.org
netdev/build_clang success Errors and warnings before: 16 this patch: 16
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/deprecated_api success None detected
netdev/check_selftest success No net selftest shell script
netdev/verify_fixes success No Fixes tag
netdev/build_allmodconfig_warn success Errors and warnings before: 1 this patch: 1
netdev/checkpatch warning WARNING: line length of 89 exceeds 80 columns
netdev/build_clang_rust success No Rust files in patch. Skipping build
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/source_inline success Was 0 now: 0
netdev/contest success net-next-2025-01-04--18-00 (tests: 887)

Commit Message

Tariq Toukan Jan. 2, 2025, 6:14 p.m. UTC 
From: Yevgeny Kliteynik <kliteyn@nvidia.com>

Rule counter in matcher's struct is used in two places:

1. As heuristics to decide when the number of rules have crossed a
certain percentage threshold and the matcher should be resized.
We don't mind here if the number will be off by 1-2 due to concurrency.

2. When destroying matcher, the counter value is checked and the
user is warned if it is not 0. Here we lock all the queues, so the
counter will be correct.

We don't need to always have *exact* number, but we do need this
number to not be corrupted, which is what is happening when the
counter isn't atomic, due to update by different threads.

Signed-off-by: Yevgeny Kliteynik <kliteyn@nvidia.com>
Reviewed-by: Erez Shitrit <erezsh@nvidia.com>
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 .../mellanox/mlx5/core/steering/hws/bwc.c       | 17 +++++++++++------
 .../mellanox/mlx5/core/steering/hws/bwc.h       |  2 +-
 2 files changed, 12 insertions(+), 7 deletions(-)
diff mbox series

Patch

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.c
index af8ab8750c70..40d688ed6153 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.c
@@ -152,6 +152,8 @@  mlx5hws_bwc_matcher_create(struct mlx5hws_table *table,
 	if (!bwc_matcher)
 		return NULL;
 
+	atomic_set(&bwc_matcher->num_of_rules, 0);
+
 	/* Check if the required match params can be all matched
 	 * in single STE, otherwise complex matcher is needed.
 	 */
@@ -199,10 +201,12 @@  int mlx5hws_bwc_matcher_destroy_simple(struct mlx5hws_bwc_matcher *bwc_matcher)
 
 int mlx5hws_bwc_matcher_destroy(struct mlx5hws_bwc_matcher *bwc_matcher)
 {
-	if (bwc_matcher->num_of_rules)
+	u32 num_of_rules = atomic_read(&bwc_matcher->num_of_rules);
+
+	if (num_of_rules)
 		mlx5hws_err(bwc_matcher->matcher->tbl->ctx,
 			    "BWC matcher destroy: matcher still has %d rules\n",
-			    bwc_matcher->num_of_rules);
+			    num_of_rules);
 
 	mlx5hws_bwc_matcher_destroy_simple(bwc_matcher);
 
@@ -309,7 +313,7 @@  static void hws_bwc_rule_list_add(struct mlx5hws_bwc_rule *bwc_rule, u16 idx)
 {
 	struct mlx5hws_bwc_matcher *bwc_matcher = bwc_rule->bwc_matcher;
 
-	bwc_matcher->num_of_rules++;
+	atomic_inc(&bwc_matcher->num_of_rules);
 	bwc_rule->bwc_queue_idx = idx;
 	list_add(&bwc_rule->list_node, &bwc_matcher->rules[idx]);
 }
@@ -318,7 +322,7 @@  static void hws_bwc_rule_list_remove(struct mlx5hws_bwc_rule *bwc_rule)
 {
 	struct mlx5hws_bwc_matcher *bwc_matcher = bwc_rule->bwc_matcher;
 
-	bwc_matcher->num_of_rules--;
+	atomic_dec(&bwc_matcher->num_of_rules);
 	list_del_init(&bwc_rule->list_node);
 }
 
@@ -711,7 +715,8 @@  hws_bwc_matcher_rehash_size(struct mlx5hws_bwc_matcher *bwc_matcher)
 	 * Need to check again if we really need rehash.
 	 * If the reason for rehash was size, but not any more - skip rehash.
 	 */
-	if (!hws_bwc_matcher_rehash_size_needed(bwc_matcher, bwc_matcher->num_of_rules))
+	if (!hws_bwc_matcher_rehash_size_needed(bwc_matcher,
+						atomic_read(&bwc_matcher->num_of_rules)))
 		return 0;
 
 	/* Now we're done all the checking - do the rehash:
@@ -804,7 +809,7 @@  int mlx5hws_bwc_rule_create_simple(struct mlx5hws_bwc_rule *bwc_rule,
 	}
 
 	/* check if number of rules require rehash */
-	num_of_rules = bwc_matcher->num_of_rules;
+	num_of_rules = atomic_read(&bwc_matcher->num_of_rules);
 
 	if (unlikely(hws_bwc_matcher_rehash_size_needed(bwc_matcher, num_of_rules))) {
 		mutex_unlock(queue_lock);
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.h b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.h
index 1d27638fa171..06c2a30c0d4e 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.h
+++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/hws/bwc.h
@@ -25,7 +25,7 @@  struct mlx5hws_bwc_matcher {
 	u8 num_of_at;
 	u16 priority;
 	u8 size_log;
-	u32 num_of_rules; /* atomically accessed */
+	atomic_t num_of_rules;
 	struct list_head *rules;
 };