[dpdk-dev] [PATCH] net/mlx5: document workaround for ConnectX-4 with L2 encap

[Dmitry Kozlyuk <dkozlyuk@nvidia.com>]

ConnectX-4 and ConnectX-4 Lx NICs require all L2 headers of transmitted
packets to be inlined. By default only first 18 bytes are inlined,
which is insufficient if additional encapsulation is used, like Q-in-Q.
Thus, default settings caused such traffic to be dropepd on Tx.
Document a workaround to increase inlined data size in such cases.

Fixes: 505f1fe426d3 ("net/mlx5: add Tx devargs")
Cc: stable@dpdk.org

Signed-off-by: Dmitry Kozlyuk <dkozlyuk@nvidia.com>
Reviewed-by: Viacheslav Ovsiienko <viacheslavo@nvidia.com>
---
 doc/guides/nics/mlx5.rst | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/doc/guides/nics/mlx5.rst b/doc/guides/nics/mlx5.rst
index ebefbe607e..05a89d08f2 100644
--- a/doc/guides/nics/mlx5.rst
+++ b/doc/guides/nics/mlx5.rst
@@ -705,6 +705,13 @@ Driver options
   it is not recommended and may prevent NIC from sending packets over
   some configurations.
 
+  For ConnectX-4 and ConnectX-4 Lx NICs, automatically configured value
+  is insufficient for some traffic, because they require at least all L2 headers
+  to be inlined. For example, Q-in-Q adds 4 bytes to default 18 bytes
+  of Ethernet and VLAN, thus ``txq_inline_min`` must be set to 22.
+  MPLS would add 4 bytes per label. Final value must account for all possible
+  L2 encapsulation headers used in particular environment.
+
   Please, note, this minimal data inlining disengages eMPW feature (Enhanced
   Multi-Packet Write), because last one does not support partial packet inlining.
   This is not very critical due to minimal data inlining is mostly required
-- 
2.25.1