[v2,03/29] net/dpaa2: enable timestamp for Rx offload case as well
Checks
Commit Message
From: Gagandeep Singh <g.singh@nxp.com>
This patch enables the packet timestamping
conditionally when Rx offload is enabled for timestamp.
Signed-off-by: Gagandeep Singh <g.singh@nxp.com>
---
drivers/net/dpaa2/dpaa2_ethdev.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
Comments
07/07/2020 11:22, Hemant Agrawal:
> From: Gagandeep Singh <g.singh@nxp.com>
>
> This patch enables the packet timestamping
> conditionally when Rx offload is enabled for timestamp.
>
> Signed-off-by: Gagandeep Singh <g.singh@nxp.com>
> ---
> drivers/net/dpaa2/dpaa2_ethdev.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
> index a1f19194d..8edd4b3cd 100644
> --- a/drivers/net/dpaa2/dpaa2_ethdev.c
> +++ b/drivers/net/dpaa2/dpaa2_ethdev.c
> @@ -524,8 +524,10 @@ dpaa2_eth_dev_configure(struct rte_eth_dev *dev)
> return ret;
> }
>
> +#if !defined(RTE_LIBRTE_IEEE1588)
> if (rx_offloads & DEV_RX_OFFLOAD_TIMESTAMP)
> - dpaa2_enable_ts = true;
> +#endif
> + dpaa2_enable_ts = true;
I don't understand this patch at all.
There is no comment in the code, and the commit log
is not very explanatory.
You are lucky Ferruh is less strict than me.
I remember I already said I was bored of the lack of explanations
in NXP drivers.
-----Original Message-----
From: Thomas Monjalon <thomas@monjalon.net>
Sent: Saturday, July 11, 2020 7:16 PM
To: Gagandeep Singh <G.Singh@nxp.com>; Hemant Agrawal <hemant.agrawal@nxp.com>
Cc: dev@dpdk.org; ferruh.yigit@intel.com
Subject: Re: [dpdk-dev] [PATCH v2 03/29] net/dpaa2: enable timestamp for Rx offload case as well
Importance: High
07/07/2020 11:22, Hemant Agrawal:
> From: Gagandeep Singh <g.singh@nxp.com>
>
> This patch enables the packet timestamping conditionally when Rx
> offload is enabled for timestamp.
>
> Signed-off-by: Gagandeep Singh <g.singh@nxp.com>
> ---
> drivers/net/dpaa2/dpaa2_ethdev.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c
> b/drivers/net/dpaa2/dpaa2_ethdev.c
> index a1f19194d..8edd4b3cd 100644
> --- a/drivers/net/dpaa2/dpaa2_ethdev.c
> +++ b/drivers/net/dpaa2/dpaa2_ethdev.c
> @@ -524,8 +524,10 @@ dpaa2_eth_dev_configure(struct rte_eth_dev *dev)
> return ret;
> }
>
> +#if !defined(RTE_LIBRTE_IEEE1588)
> if (rx_offloads & DEV_RX_OFFLOAD_TIMESTAMP)
> - dpaa2_enable_ts = true;
> +#endif
> + dpaa2_enable_ts = true;
I don't understand this patch at all.
There is no comment in the code, and the commit log is not very explanatory.
You are lucky Ferruh is less strict than me.
I remember I already said I was bored of the lack of explanations in NXP drivers.
[Hemant] We will improve next time.
The patch description says: "> This patch enables the packet timestamping conditionally when Rx
> offload is enabled for timestamp."
It should be improved with - Enable the timestamping by default when IEEE1588 is enabled irrespective of offload flag.
@@ -524,8 +524,10 @@ dpaa2_eth_dev_configure(struct rte_eth_dev *dev)
return ret;
}
+#if !defined(RTE_LIBRTE_IEEE1588)
if (rx_offloads & DEV_RX_OFFLOAD_TIMESTAMP)
- dpaa2_enable_ts = true;
+#endif
+ dpaa2_enable_ts = true;
if (tx_offloads & DEV_TX_OFFLOAD_IPV4_CKSUM)
tx_l3_csum_offload = true;