On Wed, Oct 15, 2025 at 12:00:33PM +0200, Markus Elfring wrote:

quoted

To: Fan Gong …

Please reconsider the distribution of recipient information between message fields
once more.


…

quoted

+++ b/drivers/net/ethernet/huawei/hinic3/hinic3_mgmt.c

@@ -3,19 +3,325 @@

…

quoted

+static void mgmt_resp_msg_handler(struct hinic3_msg_pf_to_mgmt *pf_to_mgmt,
+				  struct hinic3_recv_msg *recv_msg)
+{

…

quoted

+	spin_lock(&pf_to_mgmt->sync_event_lock);
+	if (recv_msg->msg_id != pf_to_mgmt->sync_msg_id) {

…

quoted

+	}
+	spin_unlock(&pf_to_mgmt->sync_event_lock);
+}

…

Will development interests grow to apply a call like “scoped_guard(spinlock, &pf_to_mgmt->sync_event_lock)”?
https://elixir.bootlin.com/linux/v6.17.1/source/include/linux/spinlock.h#L565-L567

Hi Markus, all.

It's up to the developer. But there is a still a weak preference
for open-coding, as is done above, in Networking code. So in the
absence of some other motivation (I have not reviewed this code)
I would suggest leaving this as-is.

https://docs.kernel.org/process/maintainer-netdev.html#using-device-managed-and-cleanup-h-constructs