[PATCH v4 net-next] net: flow_offload: simplify hw stats check handling

[PATCH v4 net-next] net: flow_offload: simplify hw stats check handling

[Edward Cree]

Make FLOW_ACTION_HW_STATS_DONT_CARE be all bits, rather than none, so that
 drivers and __flow_action_hw_stats_check can use simple bitwise checks.

Pre-fill all actions with DONT_CARE in flow_rule_alloc(), rather than
 relying on implicit semantics of zero from kzalloc, so that callers which
 don't configure action stats themselves (i.e. netfilter) get the correct
 behaviour by default.

Only the kernel's internal API semantics change; the TC uAPI is unaffected.

v4: move DONT_CARE setting to flow_rule_alloc() for robustness and simplicity.

v3: set DONT_CARE in nft and ct offload.

v2: rebased on net-next, removed RFC tags.

Signed-off-by: Edward Cree <ecree@solarflare.com>
---
Again, I've tested that conntrack entry actions have hw_stats=7, but can't
 test other netfilter offloads.

 drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c |  8 ++++----
 include/net/flow_offload.h                            | 11 +++++++----
 net/core/flow_offload.c                               |  6 ++++++
 3 files changed, 17 insertions(+), 8 deletions(-)

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c
index b286fe158820..51e1b3930c56 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_flower.c
@@ -30,14 +30,14 @@ static int mlxsw_sp_flower_parse_actions(struct mlxsw_sp *mlxsw_sp,
 		return -EOPNOTSUPP;
 
 	act = flow_action_first_entry_get(flow_action);
-	if (act->hw_stats == FLOW_ACTION_HW_STATS_ANY ||
-	    act->hw_stats == FLOW_ACTION_HW_STATS_IMMEDIATE) {
+	if (act->hw_stats & FLOW_ACTION_HW_STATS_DISABLED) {
+		/* Nothing to do */
+	} else if (act->hw_stats & FLOW_ACTION_HW_STATS_IMMEDIATE) {
 		/* Count action is inserted first */
 		err = mlxsw_sp_acl_rulei_act_count(mlxsw_sp, rulei, extack);
 		if (err)
 			return err;
-	} else if (act->hw_stats != FLOW_ACTION_HW_STATS_DISABLED &&
-		   act->hw_stats != FLOW_ACTION_HW_STATS_DONT_CARE) {
+	} else {
 		NL_SET_ERR_MSG_MOD(extack, "Unsupported action HW stats type");
 		return -EOPNOTSUPP;
 	}
diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index 4001ffb04f0d..95d633785ef9 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h
@@ -168,10 +168,11 @@ enum flow_action_hw_stats_bit {
 	FLOW_ACTION_HW_STATS_IMMEDIATE_BIT,
 	FLOW_ACTION_HW_STATS_DELAYED_BIT,
 	FLOW_ACTION_HW_STATS_DISABLED_BIT,
+
+	FLOW_ACTION_HW_STATS_NUM_BITS
 };
 
 enum flow_action_hw_stats {
-	FLOW_ACTION_HW_STATS_DONT_CARE = 0,
 	FLOW_ACTION_HW_STATS_IMMEDIATE =
 		BIT(FLOW_ACTION_HW_STATS_IMMEDIATE_BIT),
 	FLOW_ACTION_HW_STATS_DELAYED = BIT(FLOW_ACTION_HW_STATS_DELAYED_BIT),
@@ -179,6 +180,7 @@ enum flow_action_hw_stats {
 				   FLOW_ACTION_HW_STATS_DELAYED,
 	FLOW_ACTION_HW_STATS_DISABLED =
 		BIT(FLOW_ACTION_HW_STATS_DISABLED_BIT),
+	FLOW_ACTION_HW_STATS_DONT_CARE = BIT(FLOW_ACTION_HW_STATS_NUM_BITS) - 1,
 };
 
 typedef void (*action_destr)(void *priv);
@@ -340,11 +342,12 @@ __flow_action_hw_stats_check(const struct flow_action *action,
 		return false;
 
 	action_entry = flow_action_first_entry_get(action);
-	if (action_entry->hw_stats == FLOW_ACTION_HW_STATS_DONT_CARE)
-		return true;
+
+	/* Zero is not a legal value for hw_stats, catch anyone passing it */
+	WARN_ON_ONCE(!action_entry->hw_stats);
 
 	if (!check_allow_bit &&
-	    action_entry->hw_stats != FLOW_ACTION_HW_STATS_ANY) {
+	    ~action_entry->hw_stats & FLOW_ACTION_HW_STATS_ANY) {
 		NL_SET_ERR_MSG_MOD(extack, "Driver supports only default HW stats type \"any\"");
 		return false;
 	} else if (check_allow_bit &&
diff --git a/net/core/flow_offload.c b/net/core/flow_offload.c
index e951b743bed3..e64941c526b1 100644
--- a/net/core/flow_offload.c
+++ b/net/core/flow_offload.c
@@ -8,6 +8,7 @@
 struct flow_rule *flow_rule_alloc(unsigned int num_actions)
 {
 	struct flow_rule *rule;
+	int i;
 
 	rule = kzalloc(struct_size(rule, action.entries, num_actions),
 		       GFP_KERNEL);
@@ -15,6 +16,11 @@ struct flow_rule *flow_rule_alloc(unsigned int num_actions)
 		return NULL;
 
 	rule->action.num_entries = num_actions;
+	/* Pre-fill each action hw_stats with DONT_CARE.
+	 * Caller can override this if it wants stats for a given action.
+	 */
+	for (i = 0; i < num_actions; i++)
+		rule->action.entries[i].hw_stats = FLOW_ACTION_HW_STATS_DONT_CARE;
 
 	return rule;
 }

Re: [PATCH v4 net-next] net: flow_offload: simplify hw stats check handling

[David Miller]

From: Edward Cree <ecree@solarflare.com>
Date: Wed, 20 May 2020 19:18:10 +0100

> Make FLOW_ACTION_HW_STATS_DONT_CARE be all bits, rather than none, so that
>  drivers and __flow_action_hw_stats_check can use simple bitwise checks.
> 
> Pre-fill all actions with DONT_CARE in flow_rule_alloc(), rather than
>  relying on implicit semantics of zero from kzalloc, so that callers which
>  don't configure action stats themselves (i.e. netfilter) get the correct
>  behaviour by default.
> 
> Only the kernel's internal API semantics change; the TC uAPI is unaffected.
> 
> v4: move DONT_CARE setting to flow_rule_alloc() for robustness and simplicity.
> 
> v3: set DONT_CARE in nft and ct offload.
> 
> v2: rebased on net-next, removed RFC tags.
> 
> Signed-off-by: Edward Cree <ecree@solarflare.com>

Applied, thanks Edward.