The changeset that introduced common flow API thread safety
in fact introduced double locking to this particular PMD as
RTE flow API implementation in the PMD has been thread-safe
since the day zero. State this by setting the corresponding
device flag to skip locking imposed by generic RTE flow API.
Fixes: 80d1a9aff7f6 ("ethdev: make flow API thread safe")
Signed-off-by: Ivan Malov <ivan.ma...@oktetlabs.ru>
---
doc/guides/nics/sfc_efx.rst | 2 ++
doc/guides/rel_notes/release_20_11.rst | 1 +
drivers/net/sfc/sfc_ethdev.c | 1 +
3 files changed, 4 insertions(+)
diff --git a/doc/guides/nics/sfc_efx.rst b/doc/guides/nics/sfc_efx.rst
index 962e54389..f5e9f9495 100644
--- a/doc/guides/nics/sfc_efx.rst
+++ b/doc/guides/nics/sfc_efx.rst
@@ -246,6 +246,8 @@ Validating flow rules depends on the firmware variant.
The :ref:`flow_isolated_mode` is supported.
+The implementation is natively thread-safe.
+
Ethernet destination individual/group match
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
diff --git a/doc/guides/rel_notes/release_20_11.rst
b/doc/guides/rel_notes/release_20_11.rst
index 24cedba16..1c262d39a 100644
--- a/doc/guides/rel_notes/release_20_11.rst
+++ b/doc/guides/rel_notes/release_20_11.rst
@@ -220,6 +220,7 @@ New Features
* Added Alveo SN1000 SmartNICs (EF100 architecture) support including
flow API transfer rules for switch HW offload
* Added ARMv8 support
+ * Claimed RTE flow API native thread safety
* **Added Wangxun txgbe PMD.**
diff --git a/drivers/net/sfc/sfc_ethdev.c b/drivers/net/sfc/sfc_ethdev.c
index db81e705f..93fc7baa0 100644
--- a/drivers/net/sfc/sfc_ethdev.c
+++ b/drivers/net/sfc/sfc_ethdev.c
@@ -2198,6 +2198,7 @@ sfc_eth_dev_init(struct rte_eth_dev *dev)
/* Copy PCI device info to the dev->data */
rte_eth_copy_pci_info(dev, pci_dev);
dev->data->dev_flags |= RTE_ETH_DEV_AUTOFILL_QUEUE_XSTATS;
+ dev->data->dev_flags |= RTE_ETH_DEV_FLOW_OPS_THREAD_SAFE;
rc = sfc_kvargs_parse(sa);
if (rc != 0)
--
2.20.1
