Hi

On 06/03/2026 14:58, Simon Horman wrote:

This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.

For local reproduction steps see https://netdev-ai.bots.linux.dev/ai-local.html
---
net: ethtool: Introduce ethtool command to list ports

This commit introduces a new ethtool netlink command to expose phy_port
information to userspace, allowing enumeration of available ports and
their capabilities on a given interface.

quoted

+static int port_dump_one(struct sk_buff *skb, struct net_device *dev,
+			 struct netlink_callback *cb)
+{
+	struct port_dump_ctx *ctx = port_dump_ctx_get(cb);
+	void *ehdr;
+	int ret;
+
+	ehdr = ethnl_dump_put(skb, cb, ETHTOOL_A_PORT_HEADER);
+	if (!ehdr)
+		return -EMSGSIZE;

Is ETHTOOL_A_PORT_HEADER the correct constant here?

The third parameter to ethnl_dump_put() should be the genetlink command
ID that gets stored in the message header's cmd field. Looking at other
dump implementations in net/ethtool/:

net/ethtool/tunnels.c:
    ehdr = ethnl_dump_put(skb, cb, ETHTOOL_MSG_TUNNEL_INFO_GET_REPLY);

net/ethtool/tsinfo.c:
    ehdr = ethnl_dump_put(skb, cb, ETHTOOL_MSG_TSINFO_GET_REPLY);

net/ethtool/rss.c:
    ehdr = ethnl_dump_put(skb, cb, ETHTOOL_MSG_RSS_GET_REPLY);

These all pass ETHTOOL_MSG_*_GET_REPLY command constants, not
ETHTOOL_A_* attribute constants.

ETHTOOL_A_PORT_HEADER is a netlink attribute with value 1 (from
include/uapi/linux/ethtool_netlink.h). Meanwhile, this file's
ethnl_port_request_ops structure correctly uses
ETHTOOL_MSG_PORT_GET_REPLY for .reply_cmd, suggesting the same
constant should be used here. Can this cause the wrong command ID
to be sent in every dump response message?

hmmm true indeed, however ethtool did show me the replies. I'll fix that
though.

Thanks !

Maxime