Re: [PATCH V3 1/2] net: ethernet: ti: cpsw_new: enable mac_managed_pm to fix mdio

[Paolo Abeni <pabeni@redhat.com>]

On Tue, 2024-02-06 at 12:49 +0530, Sinthu Raja M wrote:
> On Tue, Feb 6, 2024 at 11:31 AM Ravi Gunasekaran <r-gunasekaran@ti.com> wrote:
> > In v1, you received a comment to add the fixes tag. The reference stmmac patch also points
> > to this commit as Fixes tag. But as Paolo pointed out in v2, this is not the right
> > fixes tag for your patch series.
> > 
> > I did a git blame on few drivers where PHY is managed by MAC. These have
> > Fixes: fba863b81604 ("net: phy: make PHY PM ops a no-op if MAC driver manages PHY PM")
> 
> Thanks, Ravi for pointing this out.
> But the warning message was caused only after the below commit had been added.
> 744d23c71af3 ("net: phy: Warn about incorrect mdio_bus_phy_resume() state")
> 
> With the below commit the warning didn't pop up.
> fba863b81604 ("net: phy: make PHY PM ops a no-op if MAC driver manages PHY PM")
> That is the reason I have not changed the Fixes tag.
> 
> Let's wait for Paolo's comment on this Fixes: fba863b81604 ("net: phy:
> make PHY PM ops a no-op if MAC driver manages PHY PM")

The main point of fixes tag is helping stable teams to backport the
fixes if and where relevant.

In this case the issue addressed is arguably an inconsistent state,
which may lead to later functional issue. The warning emitted by the
phy core is just an head-up, but it does not look like the 'real'
problem.

An older kernel including fba863b81604 but not including 744d23c71af3
will likely be prone to the issue.

TL;DR: I think it's better to include both fixes tag.

I'll add the 2nd one while applying the patch, no need to repost.

Cheers,

Paolo