On 1/15/26 10:18 AM, Oliver Hartkopp wrote:

On 15.01.26 09:59, Marc Kleine-Budde wrote:

quoted

On 15.01.2026 08:55:33, Oliver Hartkopp wrote:

quoted

Hello Marc,

On 14.01.26 11:45, Marc Kleine-Budde wrote:

quoted

From: Oliver Hartkopp <socketcan@hartkopp.net>

quoted

@@ -944,6 +945,10 @@ static int raw_sendmsg(struct socket *sock, struct msghdr *msg, size_t size)
   	if (!dev)
   		return -ENXIO;
+	/* no sending on a CAN device in read-only mode */
+	if (can_cap_enabled(dev, CAN_CAP_RO))
+		return -EACCES;
+
   	skb = sock_alloc_send_skb(sk, size + sizeof(struct can_skb_priv),
   				  msg->msg_flags & MSG_DONTWAIT, &err);
   	if (!skb)

At midnight the AI review from the netdev patchwork correctly identified a
problem with the above code:

https://netdev-ai.bots.linux.dev/ai-review.html?id=fb201338-eed0-488f-bb32-5240af254cf4

Is the review sent exclusively in a direct email or available in a
mailing list?

No. I have checked the status of our PR in patchwork yesterday:

https://patchwork.kernel.org/project/netdevbpf/list/?series=1042268

And I was wondering why my patch was marked "yellow"

https://patchwork.kernel.org/project/netdevbpf/patch/20260114105212.1034554-4-mkl@pengutronix.de/

The AI review marked the patch as "yellow" but the review result was not 
accessible until midnight.

A direct feedback to the authors would be helpful.

The AI review is intentionally "revealed" in PW after a grace period to
avoid random people sending unreviewed/half-finished patches to the ML
just to get the AI review.

I insisted to raise such grace period to 24h to align with the maximum
re-submit rate, but I did not consider carefully the trusted PR cases.

/P