From: Mark Bloch <markb@mellanox.com>
To: Leon Romanovsky <leon@kernel.org>,
Doug Ledford <dledford@redhat.com>,
Jason Gunthorpe <jgg@mellanox.com>
Cc: Yevgeny Kliteynik <kliteyn@mellanox.com>,
RDMA mailing list <linux-rdma@vger.kernel.org>,
Leon Romanovsky <leonro@mellanox.com>
Subject: Re: [PATCH rdma-next v2] IB/mlx5: Support flow counters offset for bulk counters
Date: Mon, 4 Nov 2019 08:42:10 +0000 [thread overview]
Message-ID: <06ad7417-7427-3402-0b8c-772ae2ea83dd@mellanox.com> (raw)
In-Reply-To: <20191103140723.77411-1-leon@kernel.org>
On 11/3/2019 06:07, Leon Romanovsky wrote:
> From: Yevgeny Kliteynik <kliteyn@mellanox.com>
>
> Add support for flow steering counters action with
> a non-base counter ID (offset) for bulk counters.
>
> When creating a flow counter object, save the bulk value.
> This value is used when a flow action with a non-base
> counter ID is requested - to validate that the required
> offset is in the range of the allocated bulk.
>
> Signed-off-by: Yevgeny Kliteynik <kliteyn@mellanox.com>
> Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
> ---
> Changelog
> v1->v2: https://lore.kernel.org/linux-rdma/20191029155020.20792-1-leon@kernel.org
> * Fixed mlx5_ib_devx_is_flow_counter() logic
> v0 -> v1: https://lore.kernel.org/linux-rdma/20191029055916.7322-1-leon@kernel.org
> * Change ffs to multiply bitmap
> * Changed uint32_t to be u32
> * Added offset to mlx5_ib_devx_is_flow_counter()
> ---
> drivers/infiniband/hw/mlx5/devx.c | 15 +++++++++++-
> drivers/infiniband/hw/mlx5/flow.c | 29 ++++++++++++++++++++++--
> drivers/infiniband/hw/mlx5/mlx5_ib.h | 2 +-
> include/uapi/rdma/mlx5_user_ioctl_cmds.h | 1 +
> 4 files changed, 43 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/infiniband/hw/mlx5/devx.c b/drivers/infiniband/hw/mlx5/devx.c
> index 6b1fca91d7d3..2a5812a4c3bb 100644
> --- a/drivers/infiniband/hw/mlx5/devx.c
> +++ b/drivers/infiniband/hw/mlx5/devx.c
> @@ -100,6 +100,7 @@ struct devx_obj {
> struct mlx5_ib_devx_mr devx_mr;
> struct mlx5_core_dct core_dct;
> struct mlx5_core_cq core_cq;
> + u32 flow_counter_bulk_size;
> };
> struct list_head event_sub; /* holds devx_event_subscription entries */
> };
> @@ -192,15 +193,20 @@ bool mlx5_ib_devx_is_flow_dest(void *obj, int *dest_id, int *dest_type)
> }
> }
>
> -bool mlx5_ib_devx_is_flow_counter(void *obj, u32 *counter_id)
> +bool mlx5_ib_devx_is_flow_counter(void *obj, u32 offset, u32 *counter_id)
> {
> struct devx_obj *devx_obj = obj;
> u16 opcode = MLX5_GET(general_obj_in_cmd_hdr, devx_obj->dinbox, opcode);
>
> if (opcode == MLX5_CMD_OP_DEALLOC_FLOW_COUNTER) {
> +
> + if (offset && offset >= devx_obj->flow_counter_bulk_size)
> + return false;
> +
> *counter_id = MLX5_GET(dealloc_flow_counter_in,
> devx_obj->dinbox,
> flow_counter_id);
> + *counter_id += offset;
> return true;
> }
>
> @@ -1463,6 +1469,13 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_DEVX_OBJ_CREATE)(
> if (err)
> goto obj_free;
>
> + if (opcode == MLX5_CMD_OP_ALLOC_FLOW_COUNTER) {
> + u8 bulk = MLX5_GET(alloc_flow_counter_in,
> + cmd_in,
> + flow_counter_bulk);
> + obj->flow_counter_bulk_size = 128UL * bulk;
> + }
> +
> uobj->object = obj;
> INIT_LIST_HEAD(&obj->event_sub);
> obj->ib_dev = dev;
> diff --git a/drivers/infiniband/hw/mlx5/flow.c b/drivers/infiniband/hw/mlx5/flow.c
> index b198ff10cde9..dbee17d22d50 100644
> --- a/drivers/infiniband/hw/mlx5/flow.c
> +++ b/drivers/infiniband/hw/mlx5/flow.c
> @@ -85,6 +85,8 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_CREATE_FLOW)(
> struct mlx5_ib_dev *dev = mlx5_udata_to_mdev(&attrs->driver_udata);
> int len, ret, i;
> u32 counter_id = 0;
> + u32 *offset_attr;
> + u32 offset = 0;
>
> if (!capable(CAP_NET_RAW))
> return -EPERM;
> @@ -151,8 +153,27 @@ static int UVERBS_HANDLER(MLX5_IB_METHOD_CREATE_FLOW)(
> if (len) {
> devx_obj = arr_flow_actions[0]->object;
>
> - if (!mlx5_ib_devx_is_flow_counter(devx_obj, &counter_id))
> + if (uverbs_attr_is_valid(attrs,
> + MLX5_IB_ATTR_CREATE_FLOW_ARR_COUNTERS_DEVX_OFFSET)) {
> +
> + int num_offsets = uverbs_attr_ptr_get_array_size(
> + attrs,
> + MLX5_IB_ATTR_CREATE_FLOW_ARR_COUNTERS_DEVX_OFFSET,
> + sizeof(u32));
> +
> + if (num_offsets != 1)
> + return -EINVAL;
> +
> + offset_attr = uverbs_attr_get_alloced_ptr(
> + attrs,
> + MLX5_IB_ATTR_CREATE_FLOW_ARR_COUNTERS_DEVX_OFFSET);
> + offset = *offset_attr;
I would have used offset_arr and would have done:
offset = offset_attr[0] seems more appropriate but I don't mind either way.
Reviewed-by: Mark Bloch <markb@mellanox.com>
> + }
> +
> + if (!mlx5_ib_devx_is_flow_counter(devx_obj, offset,
> + &counter_id))
> return -EINVAL;
> +
> flow_act.action |= MLX5_FLOW_CONTEXT_ACTION_COUNT;
> }
>
> @@ -598,7 +619,11 @@ DECLARE_UVERBS_NAMED_METHOD(
> UVERBS_ATTR_IDRS_ARR(MLX5_IB_ATTR_CREATE_FLOW_ARR_COUNTERS_DEVX,
> MLX5_IB_OBJECT_DEVX_OBJ,
> UVERBS_ACCESS_READ, 1, 1,
> - UA_OPTIONAL));
> + UA_OPTIONAL),
> + UVERBS_ATTR_PTR_IN(MLX5_IB_ATTR_CREATE_FLOW_ARR_COUNTERS_DEVX_OFFSET,
> + UVERBS_ATTR_MIN_SIZE(sizeof(u32)),
> + UA_OPTIONAL,
> + UA_ALLOC_AND_COPY));
>
> DECLARE_UVERBS_NAMED_METHOD_DESTROY(
> MLX5_IB_METHOD_DESTROY_FLOW,
> diff --git a/drivers/infiniband/hw/mlx5/mlx5_ib.h b/drivers/infiniband/hw/mlx5/mlx5_ib.h
> index 0bdb8b45ea15..8ca72cc8797e 100644
> --- a/drivers/infiniband/hw/mlx5/mlx5_ib.h
> +++ b/drivers/infiniband/hw/mlx5/mlx5_ib.h
> @@ -1367,7 +1367,7 @@ struct mlx5_ib_flow_handler *mlx5_ib_raw_fs_rule_add(
> struct mlx5_flow_act *flow_act, u32 counter_id,
> void *cmd_in, int inlen, int dest_id, int dest_type);
> bool mlx5_ib_devx_is_flow_dest(void *obj, int *dest_id, int *dest_type);
> -bool mlx5_ib_devx_is_flow_counter(void *obj, u32 *counter_id);
> +bool mlx5_ib_devx_is_flow_counter(void *obj, u32 offset, u32 *counter_id);
> int mlx5_ib_get_flow_trees(const struct uverbs_object_tree_def **root);
> void mlx5_ib_destroy_flow_action_raw(struct mlx5_ib_flow_action *maction);
> #else
> diff --git a/include/uapi/rdma/mlx5_user_ioctl_cmds.h b/include/uapi/rdma/mlx5_user_ioctl_cmds.h
> index d0da070cf0ab..20d88307f75f 100644
> --- a/include/uapi/rdma/mlx5_user_ioctl_cmds.h
> +++ b/include/uapi/rdma/mlx5_user_ioctl_cmds.h
> @@ -198,6 +198,7 @@ enum mlx5_ib_create_flow_attrs {
> MLX5_IB_ATTR_CREATE_FLOW_ARR_FLOW_ACTIONS,
> MLX5_IB_ATTR_CREATE_FLOW_TAG,
> MLX5_IB_ATTR_CREATE_FLOW_ARR_COUNTERS_DEVX,
> + MLX5_IB_ATTR_CREATE_FLOW_ARR_COUNTERS_DEVX_OFFSET,
> };
>
> enum mlx5_ib_destoy_flow_attrs {
> --
> 2.20.1
>
next prev parent reply other threads:[~2019-11-04 8:42 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-11-03 14:07 [PATCH rdma-next v2] IB/mlx5: Support flow counters offset for bulk counters Leon Romanovsky
2019-11-04 8:42 ` Mark Bloch [this message]
2019-11-06 20:21 ` Jason Gunthorpe
2019-11-07 10:11 ` Yishai Hadas
2019-11-13 19:49 ` Jason Gunthorpe
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=06ad7417-7427-3402-0b8c-772ae2ea83dd@mellanox.com \
--to=markb@mellanox.com \
--cc=dledford@redhat.com \
--cc=jgg@mellanox.com \
--cc=kliteyn@mellanox.com \
--cc=leon@kernel.org \
--cc=leonro@mellanox.com \
--cc=linux-rdma@vger.kernel.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).