Thread (20 messages) 20 messages, 3 authors, 2026-02-26

Re: [PATCH net-next 2/7] bng_en: query PHY capabilities and report link status

From: Bhargava Chenna Marreddy <hidden>
Date: 2026-02-26 14:24:32
Also in: lkml

On Wed, Feb 25, 2026 at 8:05 AM Jakub Kicinski [off-list ref] wrote:
On Mon, 23 Feb 2026 22:06:36 +0530 Bhargava Marreddy wrote:
quoted
+     /* To protect link related settings during link changes and
+      * ethtool settings changes.
+      */
+     struct mutex            link_lock;
+     struct bnge_link_info   link_info;
Why does link deserve its own lock? Are the operations slow?

bnxt was written before the netdev instance lock existed, if you're
just trying to avoid taking rtnl_lock the instance lock would do.
Hi Jakub, thanks for the feedback.

Agreed. I will replace link_lock with the netdev instance lock in the
next spin.

These link operations are relatively slow as they involve firmware HWRM
mailbox round-trips (PHY query/config). Since the firmware latency is
the main factor in hold time, the instance lock is sufficient and
avoids extra complexity.

Thanks,
Bhargava Marreddy

Attachments

  • smime.p7s [application/pkcs7-signature] 5496 bytes
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help