Re: [PATCH v4 0/2] net/af_xdp: add Rx timestamping and read_clock support

[Stephen Hemminger <stephen@networkplumber.org>]

On Sat,  1 Aug 2026 03:25:31 +0000
Mark Blasko <blasko@google.com> wrote:

> This patch series introduces support for dynamic RX timestamping and
> clock querying in the AF_XDP Poll Mode Driver.
> 
> The first patch introduces three new vdev devargs to specify
> layout-agnostic metadata offsets and bitmasks for extracting hardware
> RX timestamps from XDP metadata into the mbuf dynamic timestamp field.
> 
> The second patch implements the read_clock ethdev operation, querying
> ethtool for the interface's PTP Hardware Clock index at start and using
> clock_gettime to query the NIC hardware clock time.
> ---

AI review with Claude Opus sees some valid issues:

The significant one is in patch 1: rx_timestamp_enabled is computed in
eth_rx_queue_setup() and includes timestamp_dynfield_offset >= 0, but
the dynfield is only registered in eth_dev_start(), which runs
afterwards. Since af_xdp doesn't advertise runtime queue setup, the
flag is latched false and timestamps are never written. af_packet has
the same start-time registration but keeps the offset out of the enable
condition — setting the flag from rxmode.offloads in
eth_dev_configure() is the fix.

Second error: the release notes hunks target release_26_07.rst, but
main is 26.11.0-rc0 and 26.07 has shipped — that's also why the series
no longer applies.

Beyond that, mostly documentation/code mismatches: parse_hex_arg() is
base-16 only while the doc promises hex or decimal, the offset
direction is undocumented (they're measured backwards from mtod), and
the doc's own example puts the validity byte inside the timestamp bytes.