Re: [PATCH net-next 4/4] net: pse-pd: check the PI vpwr supply before registering the controller

[Carlo Szelinsky <github@szelinsky.de>]

On Mon, 19 Oct 2026 08:35:39 +0000, netdev-bot+sashiko@kernel.org wrote:
> [Severity: Medium]
> Can this check pass even though resolving the PI's supply later still
> defers?

Yes, as the commit message says. The bound check is inside the
regulator core, and a PSE driver cannot do it from outside. fw_devlink
does not help either: the vpwr-supply is in the PI child nodes, and
for those fw_devlink adds no link that delays the controller's probe.
So this window is real while the supplier has registered its
regulator but has not finished probing.

Once the supplier has bound, PI lookups work again. But the notifier
series only retries a failed lookup on the next PSE_REGISTERED, so
until then the PHY gets no PSE control.

> Also, pse_register_pw_ds() only runs at registration. Would that PI then
> never get a power domain?

Yes, until the controller probes again. The commit message says that
too.

> [Severity: Medium]
> Does this new exit, and the pse_controller_check_supply() exit just
> below it, leak the vmain power budget that pd692x0 reserved in
> setup_pi_matrix()?

Yes, an older bug, and this patch makes it happen more often. It is
fixed in net by "net: pse-pd: pd692x0: release the manager power
budget through devres", which this patch will wait for.

> [Severity: Low]
> This isn't a bug, but is this comment accurate?

Right, the fallback also walks the controller's child nodes. I will
fix the comment in the next version.

Carlo