On Tue, Jul 30, 2019 at 12:18:59AM +0200, Thomas Monjalon wrote:
> External Email
>
> ----------------------------------------------------------------------
> 25/07/2019 17:51, Harman Kalra:
> > Offloafing Rx timestamp is a device capability than queue capability.
>
> Why is it a device capability and not a queue capability?
1. Since all PMDs doesn't implements per queue offload capabilities but
supports RX timestamping and also since rx_offload_capa includes all
rx_queue_offload_capa's. So we moved the DEV_RX_OFFLOAD_TIMESTAMP
configuration setting to device from queue, as its a test application
and should work with all PMDs.
2. Or we can have a test in this application, i.e. if PMD has
rx_queue_offload_capa implemented go with per queue configuration else
make it a device configuration.
If OK with this first approach, we will send V2 with reworded commit message
explaining the reason as above.
>
> > Hence the logic to enable HW timestamp via DEV_RX_OFFLOAD_TIMESTAMP
> > flag should be before device configuration.
> >
> > Fixes: cd1dadeb9b2a ("examples/rxtx_callbacks: support HW timestamp")
> > Cc: barbe...@kth.se
> >
> > Signed-off-by: Harman Kalra <hka...@marvell.com>
>
>
>
