From: Rongwei Liu <rongweil@nvidia.com>
To: <dev@dpdk.org>, <matan@nvidia.com>, <viacheslavo@nvidia.com>,
<orika@nvidia.com>, <suanmingm@nvidia.com>, <thomas@monjalon.net>
Cc: <rongweil@nvidia.com>, <stable@dpdk.org>,
Raslan Darawsheh <rasland@nvidia.com>
Subject: [21.11 PATCH v1] net/mlx5: fix matcher layout size calculation
Date: Fri, 3 Nov 2023 05:07:21 +0200 [thread overview]
Message-ID: <20231103030721.930738-1-rongweil@nvidia.com> (raw)
Initially, the rdma-core library only supported misc0 to misc3 fields in
matching resources, misc4 and misc5 fields were added to handle new
features. The matcher layout, passing from DPDK to rdma-core, shouldn't
exceed the size of the engaged library version capabilities.
For now, there is no way to know what is the maximum capability of
rdma-core, and DPDK should limit the matcher layout to misc3 if
possible (no matching on fields misc4 and misc5 are requested by the
application).
The check if misc4 and misc5 features were requested was based on checking
the values against zeroes. The matching mask should be checked instead.
Fixes: 630a587bfb37 ("net/mlx5: support matching on VXLAN reserved field")
Cc: rongweil@nvidia.com
Cc: stable@dpdk.org
Signed-off-by: Rongwei Liu <rongweil@nvidia.com>
Acked-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
---
drivers/net/mlx5/mlx5_flow_dv.c | 18 +++++++++---------
1 file changed, 9 insertions(+), 9 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_flow_dv.c b/drivers/net/mlx5/mlx5_flow_dv.c
index 2896bed5dc..535e0ea618 100644
--- a/drivers/net/mlx5/mlx5_flow_dv.c
+++ b/drivers/net/mlx5/mlx5_flow_dv.c
@@ -14067,7 +14067,7 @@ flow_dv_apply(struct rte_eth_dev *dev, struct rte_flow *flow,
}
dv->actions[n++] = priv->sh->default_miss_action;
}
- misc_mask = flow_dv_matcher_enable(dv->value.buf);
+ misc_mask = flow_dv_matcher_enable(dv_h->matcher->mask.buf);
__flow_dv_adjust_buf_size(&dv->value.size, misc_mask);
err = mlx5_flow_os_create_flow(dv_h->matcher->matcher_object,
(void *)&dv->value, n,
@@ -16236,7 +16236,7 @@ flow_dv_destroy_def_policy(struct rte_eth_dev *dev)
static int
__flow_dv_create_policy_flow(struct rte_eth_dev *dev,
uint32_t color_reg_c_idx,
- enum rte_color color, void *matcher_object,
+ enum rte_color color, struct mlx5_flow_dv_matcher *dv_matcher,
int actions_n, void *actions,
bool match_src_port, const struct rte_flow_item *item,
void **rule, const struct rte_flow_attr *attr)
@@ -16262,9 +16262,9 @@ __flow_dv_create_policy_flow(struct rte_eth_dev *dev,
flow_dv_match_meta_reg(matcher.buf, value.buf,
(enum modify_reg)color_reg_c_idx,
rte_col_2_mlx5_col(color), UINT32_MAX);
- misc_mask = flow_dv_matcher_enable(value.buf);
+ misc_mask = flow_dv_matcher_enable(dv_matcher->mask.buf);
__flow_dv_adjust_buf_size(&value.size, misc_mask);
- ret = mlx5_flow_os_create_flow(matcher_object, (void *)&value,
+ ret = mlx5_flow_os_create_flow(dv_matcher->matcher_object, (void *)&value,
actions_n, actions, rule);
if (ret) {
DRV_LOG(ERR, "Failed to create meter policy%d flow.", color);
@@ -16412,7 +16412,7 @@ __flow_dv_create_domain_policy_rules(struct rte_eth_dev *dev,
/* Create flow, matching color. */
if (__flow_dv_create_policy_flow(dev,
color_reg_c_idx, (enum rte_color)i,
- color_rule->matcher->matcher_object,
+ color_rule->matcher,
acts[i].actions_n, acts[i].dv_actions,
svport_match, NULL, &color_rule->rule,
&attr)) {
@@ -16872,7 +16872,7 @@ flow_dv_create_mtr_tbls(struct rte_eth_dev *dev,
actions[i++] = priv->sh->dr_drop_action;
flow_dv_match_meta_reg(matcher_para.buf, value.buf,
(enum modify_reg)mtr_id_reg_c, 0, 0);
- misc_mask = flow_dv_matcher_enable(value.buf);
+ misc_mask = flow_dv_matcher_enable(mtrmng->def_matcher[domain]->mask.buf);
__flow_dv_adjust_buf_size(&value.size, misc_mask);
ret = mlx5_flow_os_create_flow
(mtrmng->def_matcher[domain]->matcher_object,
@@ -16917,7 +16917,7 @@ flow_dv_create_mtr_tbls(struct rte_eth_dev *dev,
fm->drop_cnt, NULL);
actions[i++] = cnt->action;
actions[i++] = priv->sh->dr_drop_action;
- misc_mask = flow_dv_matcher_enable(value.buf);
+ misc_mask = flow_dv_matcher_enable(drop_matcher->mask.buf);
__flow_dv_adjust_buf_size(&value.size, misc_mask);
ret = mlx5_flow_os_create_flow(drop_matcher->matcher_object,
(void *)&value, i, actions,
@@ -17292,7 +17292,7 @@ flow_dv_meter_hierarchy_rule_create(struct rte_eth_dev *dev,
}
if (__flow_dv_create_policy_flow(dev, color_reg_c_idx,
(enum rte_color)i,
- color_rule->matcher->matcher_object,
+ color_rule->matcher,
acts.actions_n, acts.dv_actions,
true, item,
&color_rule->rule, &attr)) {
@@ -18309,7 +18309,7 @@ flow_dv_discover_priorities(struct rte_eth_dev *dev,
break;
}
/* Try to apply the flow to HW. */
- misc_mask = flow_dv_matcher_enable(flow.dv.value.buf);
+ misc_mask = flow_dv_matcher_enable(flow.handle->dvh.matcher->mask.buf);
__flow_dv_adjust_buf_size(&flow.dv.value.size, misc_mask);
err = mlx5_flow_os_create_flow
(flow.handle->dvh.matcher->matcher_object,
--
2.27.0
next reply other threads:[~2023-11-03 3:08 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-11-03 3:07 Rongwei Liu [this message]
2023-11-08 19:24 ` Kevin Traynor
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20231103030721.930738-1-rongweil@nvidia.com \
--to=rongweil@nvidia.com \
--cc=dev@dpdk.org \
--cc=matan@nvidia.com \
--cc=orika@nvidia.com \
--cc=rasland@nvidia.com \
--cc=stable@dpdk.org \
--cc=suanmingm@nvidia.com \
--cc=thomas@monjalon.net \
--cc=viacheslavo@nvidia.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).