The integrity item uses the DW "oks1" in header layout. It includes the
all supported bits for both inner and outer. When item is integrity
type, the relevant bits are turned on and all DW is submitted.
When user provides more then single integrity item in same pattern, the
last one overrides the values were submitted before. This is problematic
when user wants to match integrity bits for both inner and outer in same
pattern, he cannot merge them into single item since rte_flow API
provides encapsulation level field to match either inner or outer.
This patch avoids the overriding values, when "oks1" is submitted,
operator "or" is used instead of regular set.
Fixes: c55c2bf35333 ("net/mlx5/hws: add definer layer")
Cc: va...@nvidia.com
Cc: sta...@dpdk.org
Signed-off-by: Michael Baum <michae...@nvidia.com>
Reviewed-by: Erez Shitrit <ere...@nvidia.com>
Acked-by: Matan Azrad <ma...@nvidia.com>
---
drivers/net/mlx5/hws/mlx5dr_definer.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/net/mlx5/hws/mlx5dr_definer.c
b/drivers/net/mlx5/hws/mlx5dr_definer.c
index e036aca781..0e15aafb8a 100644
--- a/drivers/net/mlx5/hws/mlx5dr_definer.c
+++ b/drivers/net/mlx5/hws/mlx5dr_definer.c
@@ -44,6 +44,10 @@
(bit_off))); \
} while (0)
+/* Getter function based on bit offset and mask, for 32bit DW*/
+#define DR_GET_32(p, byte_off, bit_off, mask) \
+ ((rte_be_to_cpu_32(*((const rte_be32_t *)(p) + ((byte_off) / 4))) >>
(bit_off)) & (mask))
+
/* Setter function based on bit offset and mask */
#define DR_SET(p, v, byte_off, bit_off, mask) \
do { \
@@ -509,7 +513,7 @@ mlx5dr_definer_integrity_set(struct mlx5dr_definer_fc *fc,
{
bool inner = (fc->fname == MLX5DR_DEFINER_FNAME_INTEGRITY_I);
const struct rte_flow_item_integrity *v = item_spec;
- uint32_t ok1_bits = 0;
+ uint32_t ok1_bits = DR_GET_32(tag, fc->byte_off, fc->bit_off,
fc->bit_mask);
if (v->l3_ok)
ok1_bits |= inner ? BIT(MLX5DR_DEFINER_OKS1_SECOND_L3_OK) :
--
2.25.1
