From: "Xueming(Steven) Li" <xuemingl@nvidia.com>
To: Lior Margalit <lmargalit@nvidia.com>, Matan Azrad <matan@nvidia.com>
Cc: "stable@dpdk.org" <stable@dpdk.org>
Subject: Re: [PATCH 20.11 1/5] net/mlx5: fix RSS expansion for inner tunnel VLAN
Date: Tue, 30 Nov 2021 13:55:39 +0000 [thread overview]
Message-ID: <96abe15bb427b7e8e03ccf83817ed02b852e832d.camel@nvidia.com> (raw)
In-Reply-To: <20211130121714.4170879-1-lmargalit@nvidia.com>
Series applied to 20.11.4 list, thanks!
On Tue, 2021-11-30 at 14:17 +0200, Lior Margalit wrote:
> [ upstream commit 3f02c7ff6815286dfe81efbf2db196bdafc752b3 ]
>
> The RSS expansion algorithm is using a graph to find the possible
> expansion paths. The VLAN item in the flow pattern requires special
> treatment, because it should not be added implicitly by the expansion
> algorithm. If the flow pattern ends with ETH item, the pattern will be
> expanded with IPv4 and IPv6.
> For example:
> testpmd> flow create ... eth / end actions rss / end
> ETH END
> ETH IPV4 END
> ETH IPV6 END
> If a VLAN item follows the ETH item in the flow pattern, the pattern
> will be expanded with IPv4 and IPv6 following the VLAN item.
> For example:
> testpmd> flow create ... eth / vlan / end actions rss level 1 / end
> ETH VLAN END
> ETH VLAN IPV4 END
> ETH VLAN IPV6 END
>
> The case of inner tunnel VLAN item was not taken care of so the flow
> pattern did not expand with IPv6 and IPv4 as expected.
> Example with inner VLAN:
> testpmd> flow create ... / vxlan / eth / vlan / end actions rss level 2
> / end
> The current result of the expansion alg:
> ETH IPV6 UDP VXLAN ETH VLAN END
> The expected result of the expansion alg:
> ETH IPV6 UDP VXLAN ETH VLAN END
> ETH IPV6 UDP VXLAN ETH VLAN IPV4 END
> ETH IPV6 UDP VXLAN ETH VLAN IPV6 END
>
> The fix is to introduce a new flag to set on a graph expansion node
> to apply the 'explicit' behavior, meaning the node is not added to
> the expanded pattern, if it is not found in the flow pattern, but the
> expansion alg can go deeper to its next nodes.
>
> Fixes: c7870bfe09dc ("ethdev: move RSS expansion code to mlx5 driver")
> Cc: stable@dpdk.org
>
> Signed-off-by: Lior Margalit <lmargalit@nvidia.com>
> Acked-by: Matan Azrad <matan@nvidia.com>
> ---
> drivers/net/mlx5/mlx5_flow.c | 88 ++++++++++++++++++------------------
> 1 file changed, 44 insertions(+), 44 deletions(-)
>
> diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
> index 3c35b87964..2b916899e8 100644
> --- a/drivers/net/mlx5/mlx5_flow.c
> +++ b/drivers/net/mlx5/mlx5_flow.c
> @@ -100,10 +100,25 @@ struct mlx5_flow_expand_node {
> * RSS types bit-field associated with this node
> * (see ETH_RSS_* definitions).
> */
> - uint8_t optional;
> - /**< optional expand field. Default 0 to expand, 1 not go deeper. */
> + uint64_t node_flags;
> + /**<
> + * Bit-fields that define how the node is used in the expansion.
> + * (see MLX5_EXPANSION_NODE_* definitions).
> + */
> };
>
> +/* Optional expand field. The expansion alg will not go deeper. */
> +#define MLX5_EXPANSION_NODE_OPTIONAL (UINT64_C(1) << 0)
> +
> +/* The node is not added implicitly as expansion to the flow pattern.
> + * If the node type does not match the flow pattern item type, the
> + * expansion alg will go deeper to its next items.
> + * In the current implementation, the list of next nodes indexes can
> + * have up to one node with this flag set and it has to be the last
> + * node index (before the list terminator).
> + */
> +#define MLX5_EXPANSION_NODE_EXPLICIT (UINT64_C(1) << 1)
> +
> /** Object returned by mlx5_flow_expand_rss(). */
> struct mlx5_flow_expand_rss {
> uint32_t entries;
> @@ -331,10 +346,17 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size,
> continue;
> }
> last_item = item;
> - for (i = 0; node->next && node->next[i]; ++i) {
> + i = 0;
> + while (node->next && node->next[i]) {
> next = &graph[node->next[i]];
> if (next->type == item->type)
> break;
> + if (next->node_flags & MLX5_EXPANSION_NODE_EXPLICIT) {
> + node = next;
> + i = 0;
> + } else {
> + ++i;
> + }
> }
> if (next)
> node = next;
> @@ -393,6 +415,16 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size,
> }
> memset(flow_items, 0, sizeof(flow_items));
> next_node = node->next;
> + while (next_node) {
> + /*
> + * Skip the nodes with the MLX5_EXPANSION_NODE_EXPLICIT
> + * flag set, because they were not found in the flow pattern.
> + */
> + node = &graph[*next_node];
> + if (!(node->node_flags & MLX5_EXPANSION_NODE_EXPLICIT))
> + break;
> + next_node = node->next;
> + }
> stack[stack_pos] = next_node;
> node = next_node ? &graph[*next_node] : NULL;
> while (node) {
> @@ -427,7 +459,8 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size,
> addr = (void *)(((uintptr_t)addr) + n);
> }
> /* Go deeper. */
> - if (!node->optional && node->next) {
> + if (!(node->node_flags & MLX5_EXPANSION_NODE_OPTIONAL) &&
> + node->next) {
> next_node = node->next;
> if (stack_pos++ == MLX5_RSS_EXP_ELT_N) {
> rte_errno = E2BIG;
> @@ -452,10 +485,7 @@ mlx5_flow_expand_rss(struct mlx5_flow_expand_rss *buf, size_t size,
> enum mlx5_expansion {
> MLX5_EXPANSION_ROOT,
> MLX5_EXPANSION_ROOT_OUTER,
> - MLX5_EXPANSION_ROOT_ETH_VLAN,
> - MLX5_EXPANSION_ROOT_OUTER_ETH_VLAN,
> MLX5_EXPANSION_OUTER_ETH,
> - MLX5_EXPANSION_OUTER_ETH_VLAN,
> MLX5_EXPANSION_OUTER_VLAN,
> MLX5_EXPANSION_OUTER_IPV4,
> MLX5_EXPANSION_OUTER_IPV4_UDP,
> @@ -470,7 +500,6 @@ enum mlx5_expansion {
> MLX5_EXPANSION_GRE_KEY,
> MLX5_EXPANSION_MPLS,
> MLX5_EXPANSION_ETH,
> - MLX5_EXPANSION_ETH_VLAN,
> MLX5_EXPANSION_VLAN,
> MLX5_EXPANSION_IPV4,
> MLX5_EXPANSION_IPV4_UDP,
> @@ -497,22 +526,7 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = {
> MLX5_EXPANSION_OUTER_IPV6),
> .type = RTE_FLOW_ITEM_TYPE_END,
> },
> - [MLX5_EXPANSION_ROOT_ETH_VLAN] = {
> - .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_ETH_VLAN),
> - .type = RTE_FLOW_ITEM_TYPE_END,
> - },
> - [MLX5_EXPANSION_ROOT_OUTER_ETH_VLAN] = {
> - .next = MLX5_FLOW_EXPAND_RSS_NEXT
> - (MLX5_EXPANSION_OUTER_ETH_VLAN),
> - .type = RTE_FLOW_ITEM_TYPE_END,
> - },
> [MLX5_EXPANSION_OUTER_ETH] = {
> - .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_OUTER_IPV4,
> - MLX5_EXPANSION_OUTER_IPV6),
> - .type = RTE_FLOW_ITEM_TYPE_ETH,
> - .rss_types = 0,
> - },
> - [MLX5_EXPANSION_OUTER_ETH_VLAN] = {
> .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_OUTER_VLAN),
> .type = RTE_FLOW_ITEM_TYPE_ETH,
> .rss_types = 0,
> @@ -521,6 +535,7 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = {
> .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_OUTER_IPV4,
> MLX5_EXPANSION_OUTER_IPV6),
> .type = RTE_FLOW_ITEM_TYPE_VLAN,
> + .node_flags = MLX5_EXPANSION_NODE_EXPLICIT,
> },
> [MLX5_EXPANSION_OUTER_IPV4] = {
> .next = MLX5_FLOW_EXPAND_RSS_NEXT
> @@ -597,7 +612,7 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = {
> MLX5_EXPANSION_IPV6,
> MLX5_EXPANSION_MPLS),
> .type = RTE_FLOW_ITEM_TYPE_GRE_KEY,
> - .optional = 1,
> + .node_flags = MLX5_EXPANSION_NODE_OPTIONAL,
> },
> [MLX5_EXPANSION_NVGRE] = {
> .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_ETH),
> @@ -608,13 +623,9 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = {
> MLX5_EXPANSION_IPV6,
> MLX5_EXPANSION_ETH),
> .type = RTE_FLOW_ITEM_TYPE_MPLS,
> + .node_flags = MLX5_EXPANSION_NODE_OPTIONAL,
> },
> [MLX5_EXPANSION_ETH] = {
> - .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4,
> - MLX5_EXPANSION_IPV6),
> - .type = RTE_FLOW_ITEM_TYPE_ETH,
> - },
> - [MLX5_EXPANSION_ETH_VLAN] = {
> .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_VLAN),
> .type = RTE_FLOW_ITEM_TYPE_ETH,
> },
> @@ -622,6 +633,7 @@ static const struct mlx5_flow_expand_node mlx5_support_expansion[] = {
> .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4,
> MLX5_EXPANSION_IPV6),
> .type = RTE_FLOW_ITEM_TYPE_VLAN,
> + .node_flags = MLX5_EXPANSION_NODE_EXPLICIT,
> },
> [MLX5_EXPANSION_IPV4] = {
> .next = MLX5_FLOW_EXPAND_RSS_NEXT(MLX5_EXPANSION_IPV4_UDP,
> @@ -3458,20 +3470,8 @@ flow_get_shared_rss_action(struct rte_eth_dev *dev,
> }
>
> static unsigned int
> -find_graph_root(const struct rte_flow_item pattern[], uint32_t rss_level)
> +find_graph_root(uint32_t rss_level)
> {
> - const struct rte_flow_item *item;
> - unsigned int has_vlan = 0;
> -
> - for (item = pattern; item->type != RTE_FLOW_ITEM_TYPE_END; item++) {
> - if (item->type == RTE_FLOW_ITEM_TYPE_VLAN) {
> - has_vlan = 1;
> - break;
> - }
> - }
> - if (has_vlan)
> - return rss_level < 2 ? MLX5_EXPANSION_ROOT_ETH_VLAN :
> - MLX5_EXPANSION_ROOT_OUTER_ETH_VLAN;
> return rss_level < 2 ? MLX5_EXPANSION_ROOT :
> MLX5_EXPANSION_ROOT_OUTER;
> }
> @@ -5361,7 +5361,7 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list,
> int shared_actions_n = MLX5_MAX_SHARED_ACTIONS;
> union {
> struct mlx5_flow_expand_rss buf;
> - uint8_t buffer[2048];
> + uint8_t buffer[4096];
> } expand_buffer;
> union {
> struct rte_flow_action actions[MLX5_MAX_SPLIT_ACTIONS];
> @@ -5451,7 +5451,7 @@ flow_list_create(struct rte_eth_dev *dev, uint32_t *list,
> if (rss && rss->types) {
> unsigned int graph_root;
>
> - graph_root = find_graph_root(items, rss->level);
> + graph_root = find_graph_root(rss->level);
> ret = mlx5_flow_expand_rss(buf, sizeof(expand_buffer.buffer),
> items, rss->types,
> mlx5_support_expansion, graph_root);
prev parent reply other threads:[~2021-11-30 13:55 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-30 12:17 Lior Margalit
2021-11-30 12:17 ` [PATCH 20.11 2/5] net/mlx5: fix RSS expansion for explicit graph node Lior Margalit
2021-11-30 12:17 ` [PATCH 20.11 3/5] net/mlx5: fix RSS expansion traversal over next nodes Lior Margalit
2021-11-30 12:17 ` [PATCH 20.11 4/5] net/mlx5: fix RSS expansion for L2/L3 VXLAN Lior Margalit
2021-11-30 12:17 ` [PATCH 20.11 5/5] net/mlx5: fix RSS expansion with EtherType Lior Margalit
2021-11-30 13:55 ` Xueming(Steven) Li [this message]
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=96abe15bb427b7e8e03ccf83817ed02b852e832d.camel@nvidia.com \
--to=xuemingl@nvidia.com \
--cc=lmargalit@nvidia.com \
--cc=matan@nvidia.com \
--cc=stable@dpdk.org \
/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).