Re: [PATCH net v4 06/14] net: pse-pd: check the PI state under the lock when setting a power limit
From: Kory Maincent
Date: Wed Nov 18 2026 - 05:08:14 EST
On 11/16/26 19:20, Carlo Szelinsky wrote:
On Sun, 15 Nov 2026 21:00:55 +0000, netdev-bot+sashiko@xxxxxxxxxx wrote:
Does this description understate what the old race did? It looks like
the budget was leaked permanently, not just reserved for a while.
Right, the budget is lost until the controller is bound again.
v5 says that, and fixes the old "powered" comment.
Is admin_state_enabled && isr_pd_detected enough to show that the port is
actually powered?
Does this still leave part of the race with pse_isr() open?
Both were there before this patch, which only moves the old check
under the lock. They need their own fixes, so I keep them out of
this series. The cover letter lists both.
Kory, are you OK with fixing the two older problems in a follow-up?
I am happy to work on them next, if you have a preferred approach.
Yes, I am ok with this.
Also, are you fine with the rest of v4? I kept your Reviewed-by on 8
after the small change there, so please tell me if that is not OK.
Except for comment change or commit message change, I would prefer to have my R-by removed if there is code change. Thanks for asking.
For patch 8, I am ok with the changes, you can keep my R-by.
Regards,
--
Köry Maincent, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com