* [PATCH] net/mlx5: fix async flow create error handling
@ 2024-03-05 18:05 Dariusz Sosnowski
2024-03-13 7:43 ` Raslan Darawsheh
0 siblings, 1 reply; 2+ messages in thread
From: Dariusz Sosnowski @ 2024-03-05 18:05 UTC (permalink / raw)
To: Matan Azrad, Viacheslav Ovsiienko, Ori Kam, Suanming Mou; +Cc: dev, stable
Whenever processing of asynchronous flow rule create operation failed,
but after some dynamic flow actions had already been allocated,
these actions were not freed during error handling flow.
That behavior lead to leaks e.g., RSS/QUEUE action objects were leaked
which triggered assertions during device cleanup.
This patch adds flow rule cleanup handling in case of an error
during async flow rule creation.
Fixes: 3a2f674b6aa8 ("net/mlx5: add queue and RSS HW steering action")
Cc: suanmingm@nvidia.com
Cc: stable@dpdk.org
Signed-off-by: Dariusz Sosnowski <dsosnowski@nvidia.com>
Acked-by: Ori Kam <orika@nvidia.com>
---
drivers/net/mlx5/mlx5_flow_hw.c | 78 +++++++++++++++++++++++----------
1 file changed, 55 insertions(+), 23 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_flow_hw.c b/drivers/net/mlx5/mlx5_flow_hw.c
index 4216433c6e..5a407d592c 100644
--- a/drivers/net/mlx5/mlx5_flow_hw.c
+++ b/drivers/net/mlx5/mlx5_flow_hw.c
@@ -174,7 +174,7 @@ mlx5_flow_hw_aux_set_mtr_id(struct rte_flow_hw *flow,
aux->orig.mtr_id = mtr_id;
}
-static __rte_always_inline uint32_t __rte_unused
+static __rte_always_inline uint32_t
mlx5_flow_hw_aux_get_mtr_id(struct rte_flow_hw *flow, struct rte_flow_hw_aux *aux)
{
if (unlikely(flow->operation_type == MLX5_FLOW_HW_FLOW_OP_TYPE_UPDATE))
@@ -183,6 +183,10 @@ mlx5_flow_hw_aux_get_mtr_id(struct rte_flow_hw *flow, struct rte_flow_hw_aux *au
return aux->orig.mtr_id;
}
+static void
+flow_hw_age_count_release(struct mlx5_priv *priv, uint32_t queue, struct rte_flow_hw *flow,
+ struct rte_flow_error *error);
+
static int
mlx5_tbl_multi_pattern_process(struct rte_eth_dev *dev,
struct rte_flow_template_table *tbl,
@@ -3034,6 +3038,31 @@ flow_hw_modify_field_construct(struct mlx5_modification_cmd *mhdr_cmd,
return 0;
}
+/**
+ * Release any actions allocated for the flow rule during actions construction.
+ *
+ * @param[in] flow
+ * Pointer to flow structure.
+ */
+static void
+flow_hw_release_actions(struct rte_eth_dev *dev,
+ uint32_t queue,
+ struct rte_flow_hw *flow)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ struct mlx5_aso_mtr_pool *pool = priv->hws_mpool;
+ struct rte_flow_hw_aux *aux = mlx5_flow_hw_aux(dev->data->port_id, flow);
+
+ if (flow->flags & MLX5_FLOW_HW_FLOW_FLAG_FATE_JUMP)
+ flow_hw_jump_release(dev, flow->jump);
+ else if (flow->flags & MLX5_FLOW_HW_FLOW_FLAG_FATE_HRXQ)
+ mlx5_hrxq_obj_release(dev, flow->hrxq);
+ if (flow->flags & MLX5_FLOW_HW_FLOW_FLAG_CNT_ID)
+ flow_hw_age_count_release(priv, queue, flow, NULL);
+ if (flow->flags & MLX5_FLOW_HW_FLOW_FLAG_MTR_ID)
+ mlx5_ipool_free(pool->idx_pool, mlx5_flow_hw_aux_get_mtr_id(flow, aux));
+}
+
/**
* Construct flow action array.
*
@@ -3156,7 +3185,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
(dev, queue, action, table, it_idx,
at->action_flags, flow,
&rule_acts[act_data->action_dst]))
- return -1;
+ goto error;
break;
case RTE_FLOW_ACTION_TYPE_VOID:
break;
@@ -3176,7 +3205,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
jump = flow_hw_jump_action_register
(dev, &table->cfg, jump_group, NULL);
if (!jump)
- return -1;
+ goto error;
rule_acts[act_data->action_dst].action =
(!!attr.group) ? jump->hws_action : jump->root_action;
flow->jump = jump;
@@ -3188,7 +3217,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
ft_flag,
action);
if (!hrxq)
- return -1;
+ goto error;
rule_acts[act_data->action_dst].action = hrxq->action;
flow->hrxq = hrxq;
flow->flags |= MLX5_FLOW_HW_FLOW_FLAG_FATE_HRXQ;
@@ -3198,19 +3227,19 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
if (flow_hw_shared_action_get
(dev, act_data, item_flags,
&rule_acts[act_data->action_dst]))
- return -1;
+ goto error;
break;
case RTE_FLOW_ACTION_TYPE_VXLAN_ENCAP:
enc_item = ((const struct rte_flow_action_vxlan_encap *)
action->conf)->definition;
if (flow_dv_convert_encap_data(enc_item, ap->encap_data, &encap_len, NULL))
- return -1;
+ goto error;
break;
case RTE_FLOW_ACTION_TYPE_NVGRE_ENCAP:
enc_item = ((const struct rte_flow_action_nvgre_encap *)
action->conf)->definition;
if (flow_dv_convert_encap_data(enc_item, ap->encap_data, &encap_len, NULL))
- return -1;
+ goto error;
break;
case RTE_FLOW_ACTION_TYPE_RAW_ENCAP:
raw_encap_data =
@@ -3238,12 +3267,12 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
hw_acts,
action);
if (ret)
- return -1;
+ goto error;
break;
case RTE_FLOW_ACTION_TYPE_REPRESENTED_PORT:
port_action = action->conf;
if (!priv->hw_vport[port_action->port_id])
- return -1;
+ goto error;
rule_acts[act_data->action_dst].action =
priv->hw_vport[port_action->port_id];
break;
@@ -3263,7 +3292,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
jump = flow_hw_jump_action_register
(dev, &table->cfg, aso_mtr->fm.group, NULL);
if (!jump)
- return -1;
+ goto error;
MLX5_ASSERT
(!rule_acts[act_data->action_dst + 1].action);
rule_acts[act_data->action_dst + 1].action =
@@ -3272,7 +3301,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
flow->jump = jump;
flow->flags |= MLX5_FLOW_HW_FLOW_FLAG_FATE_JUMP;
if (mlx5_aso_mtr_wait(priv->sh, MLX5_HW_INV_QUEUE, aso_mtr))
- return -1;
+ goto error;
break;
case RTE_FLOW_ACTION_TYPE_AGE:
aux = mlx5_flow_hw_aux(dev->data->port_id, flow);
@@ -3288,7 +3317,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
flow->res_idx,
error);
if (age_idx == 0)
- return -rte_errno;
+ goto error;
mlx5_flow_hw_aux_set_age_idx(flow, aux, age_idx);
flow->flags |= MLX5_FLOW_HW_FLOW_FLAG_AGE_IDX;
if (at->action_flags & MLX5_FLOW_ACTION_INDIRECT_COUNT)
@@ -3303,7 +3332,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
cnt_queue = mlx5_hws_cnt_get_queue(priv, &queue);
ret = mlx5_hws_cnt_pool_get(priv->hws_cpool, cnt_queue, &cnt_id, age_idx);
if (ret != 0)
- return ret;
+ goto error;
ret = mlx5_hws_cnt_pool_get_action_offset
(priv->hws_cpool,
cnt_id,
@@ -3311,7 +3340,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
&rule_acts[act_data->action_dst].counter.offset
);
if (ret != 0)
- return ret;
+ goto error;
flow->flags |= MLX5_FLOW_HW_FLOW_FLAG_CNT_ID;
flow->cnt_id = cnt_id;
break;
@@ -3323,7 +3352,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
&rule_acts[act_data->action_dst].counter.offset
);
if (ret != 0)
- return ret;
+ goto error;
flow->flags |= MLX5_FLOW_HW_FLOW_FLAG_CNT_ID;
flow->cnt_id = act_data->shared_counter.id;
break;
@@ -3331,7 +3360,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
ct_idx = MLX5_INDIRECT_ACTION_IDX_GET(action->conf);
if (flow_hw_ct_compile(dev, queue, ct_idx,
&rule_acts[act_data->action_dst]))
- return -1;
+ goto error;
break;
case MLX5_RTE_FLOW_ACTION_TYPE_METER_MARK:
mtr_id = act_data->shared_meter.id &
@@ -3339,7 +3368,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
/* Find ASO object. */
aso_mtr = mlx5_ipool_get(pool->idx_pool, mtr_id);
if (!aso_mtr)
- return -1;
+ goto error;
rule_acts[act_data->action_dst].action =
pool->action;
rule_acts[act_data->action_dst].aso_meter.offset =
@@ -3354,7 +3383,7 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
act_data->action_dst, action,
rule_acts, &mtr_idx, MLX5_HW_INV_QUEUE, error);
if (ret != 0)
- return ret;
+ goto error;
aux = mlx5_flow_hw_aux(dev->data->port_id, flow);
mlx5_flow_hw_aux_set_mtr_id(flow, aux, mtr_idx);
flow->flags |= MLX5_FLOW_HW_FLOW_FLAG_MTR_ID;
@@ -3396,11 +3425,11 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
struct mlx5dr_rule_action *ra = &rule_acts[hw_acts->encap_decap_pos];
if (ix < 0)
- return -1;
+ goto error;
if (!mp_segment)
mp_segment = mlx5_multi_pattern_segment_find(table, flow->res_idx);
if (!mp_segment || !mp_segment->reformat_action[ix])
- return -1;
+ goto error;
ra->action = mp_segment->reformat_action[ix];
/* reformat offset is relative to selected DR action */
ra->reformat.offset = flow->res_idx - mp_segment->head_index;
@@ -3416,6 +3445,11 @@ flow_hw_actions_construct(struct rte_eth_dev *dev,
flow->cnt_id = hw_acts->cnt_id;
}
return 0;
+
+error:
+ flow_hw_release_actions(dev, queue, flow);
+ rte_errno = EINVAL;
+ return -rte_errno;
}
static const struct rte_flow_item *
@@ -3565,10 +3599,8 @@ flow_hw_async_flow_create(struct rte_eth_dev *dev,
if (flow_hw_actions_construct(dev, flow, &ap,
&table->ats[action_template_index],
pattern_template_index, actions,
- rule_acts, queue, error)) {
- rte_errno = EINVAL;
+ rule_acts, queue, error))
goto error;
- }
rule_items = flow_hw_get_rule_items(dev, table, items,
pattern_template_index, &pp);
if (!rule_items)
--
2.39.2
^ permalink raw reply [flat|nested] 2+ messages in thread
* RE: [PATCH] net/mlx5: fix async flow create error handling
2024-03-05 18:05 [PATCH] net/mlx5: fix async flow create error handling Dariusz Sosnowski
@ 2024-03-13 7:43 ` Raslan Darawsheh
0 siblings, 0 replies; 2+ messages in thread
From: Raslan Darawsheh @ 2024-03-13 7:43 UTC (permalink / raw)
To: Dariusz Sosnowski, Matan Azrad, Slava Ovsiienko, Ori Kam, Suanming Mou
Cc: dev, stable
Hi,
> -----Original Message-----
> From: Dariusz Sosnowski <dsosnowski@nvidia.com>
> Sent: Tuesday, March 5, 2024 8:05 PM
> To: Matan Azrad <matan@nvidia.com>; Slava Ovsiienko
> <viacheslavo@nvidia.com>; Ori Kam <orika@nvidia.com>; Suanming Mou
> <suanmingm@nvidia.com>
> Cc: dev@dpdk.org; stable@dpdk.org
> Subject: [PATCH] net/mlx5: fix async flow create error handling
>
> Whenever processing of asynchronous flow rule create operation failed, but
> after some dynamic flow actions had already been allocated, these actions
> were not freed during error handling flow.
> That behavior lead to leaks e.g., RSS/QUEUE action objects were leaked which
> triggered assertions during device cleanup.
>
> This patch adds flow rule cleanup handling in case of an error during async
> flow rule creation.
>
> Fixes: 3a2f674b6aa8 ("net/mlx5: add queue and RSS HW steering action")
> Cc: suanmingm@nvidia.com
> Cc: stable@dpdk.org
>
> Signed-off-by: Dariusz Sosnowski <dsosnowski@nvidia.com>
> Acked-by: Ori Kam <orika@nvidia.com>
Patch applied to next-net-mlx,
Kindest regards,
Raslan Darawsheh
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2024-03-13 7:43 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-03-05 18:05 [PATCH] net/mlx5: fix async flow create error handling Dariusz Sosnowski
2024-03-13 7:43 ` Raslan Darawsheh
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).