Re: [RFC net-next 4/7] net/mlx5e: Drain old TX SQs before activating new channels

[Tariq Toukan]

On 10/11/2026 0:19, Rishikesh Jethwani wrote:
> The regular/QoS drain and the timeout ordering look right to me. The
> PTP SQs can carry TLS data, though. Their indices are in
> real_num_tx_queues and txq2sq, and egress skbedit can select them,
> bypassing mlx5e_select_queue().

[...]

> I'd prefer to include the PTP SQs in the drain/close ordering.

Agreed, v2 drains the PTP SQs with the others, so a timeout closes the
old PTP channel first too, along with the old channels.

Thanks,
Tariq