From: Tal Shnaiderman <talshn@nvidia.com>
To: <dev@dpdk.org>
Cc: <thomas@monjalon.net>, <matan@nvidia.com>
Subject: [RFC PATCH 2/3] net/mlx5: support CQE compression on Windows
Date: Tue, 12 Apr 2022 17:59:59 +0300 [thread overview]
Message-ID: <20220412150000.3412-3-talshn@nvidia.com> (raw)
In-Reply-To: <20220412150000.3412-1-talshn@nvidia.com>
CQE Compression reduces PCI overhead by coalescing and compressing
multiple CQEs into a single merged CQE.
Add supported for the CQE compression feature on Windows.
feature is enabled by default unless not supported by the HW
or if the rxq_cqe_comp_en PMD argument is explicitly disabled.
Signed-off-by: Tal Shnaiderman <talshn@nvidia.com>
---
drivers/common/mlx5/mlx5_devx_cmds.c | 2 ++
drivers/common/mlx5/mlx5_devx_cmds.h | 1 +
drivers/net/mlx5/windows/mlx5_os.c | 8 ++++++++
3 files changed, 11 insertions(+)
diff --git a/drivers/common/mlx5/mlx5_devx_cmds.c b/drivers/common/mlx5/mlx5_devx_cmds.c
index 7900254287..a109341d02 100644
--- a/drivers/common/mlx5/mlx5_devx_cmds.c
+++ b/drivers/common/mlx5/mlx5_devx_cmds.c
@@ -956,6 +956,8 @@ mlx5_devx_cmd_query_hca_attr(void *ctx,
attr->cqe_compression = MLX5_GET(cmd_hca_cap, hcattr, cqe_compression);
attr->mini_cqe_resp_flow_tag = MLX5_GET(cmd_hca_cap, hcattr,
mini_cqe_resp_flow_tag);
+ attr->cqe_compression_128 = MLX5_GET(cmd_hca_cap, hcattr,
+ cqe_compression_128);
attr->mini_cqe_resp_l3_l4_tag = MLX5_GET(cmd_hca_cap, hcattr,
mini_cqe_resp_l3_l4_tag);
attr->umr_indirect_mkey_disabled =
diff --git a/drivers/common/mlx5/mlx5_devx_cmds.h b/drivers/common/mlx5/mlx5_devx_cmds.h
index 2d813c0fdc..6413176f2e 100644
--- a/drivers/common/mlx5/mlx5_devx_cmds.h
+++ b/drivers/common/mlx5/mlx5_devx_cmds.h
@@ -258,6 +258,7 @@ struct mlx5_hca_attr {
uint16_t max_wqe_sz_sq;
uint32_t striding_rq:1;
uint32_t ext_stride_num_range:1;
+ uint32_t cqe_compression_128:1;
};
/* LAG Context. */
diff --git a/drivers/net/mlx5/windows/mlx5_os.c b/drivers/net/mlx5/windows/mlx5_os.c
index 02438ece56..232f22232b 100644
--- a/drivers/net/mlx5/windows/mlx5_os.c
+++ b/drivers/net/mlx5/windows/mlx5_os.c
@@ -237,6 +237,14 @@ mlx5_os_capabilities_prepare(struct mlx5_dev_ctx_shared *sh)
} else {
DRV_LOG(DEBUG, "Tunnel offloading is not supported.");
}
+#if (RTE_CACHE_LINE_SIZE == 128)
+ if (hca_attr->hca_attr.cqe_compression_128)
+ sh->dev_cap.cqe_comp = 1;
+ DRV_LOG(DEBUG, "Rx CQE 128B compression is %ssupported.",
+ sh->dev_cap.cqe_comp ? "" : "not ");
+#else
+ sh->dev_cap.cqe_comp = 1;
+#endif
snprintf(sh->dev_cap.fw_ver, 64, "%x.%x.%04x",
MLX5_GET(initial_seg, pv_iseg, fw_rev_major),
MLX5_GET(initial_seg, pv_iseg, fw_rev_minor),
--
2.16.1.windows.4
next prev parent reply other threads:[~2022-04-12 15:01 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-12 14:59 [RFC PATCH 0/3] Windows performance enhancements Tal Shnaiderman
2022-04-12 14:59 ` [RFC PATCH 1/3] net/mlx5: support multi-packet RQ on Windows Tal Shnaiderman
2022-04-12 14:59 ` Tal Shnaiderman [this message]
2022-04-12 15:00 ` [RFC PATCH 3/3] net/mlx5: support enhanced multi-packet write " Tal Shnaiderman
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=20220412150000.3412-3-talshn@nvidia.com \
--to=talshn@nvidia.com \
--cc=dev@dpdk.org \
--cc=matan@nvidia.com \
--cc=thomas@monjalon.net \
/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).