Re: [RFC PATCH] net: phy: integrate reset-after-clock quirk into phy_init_hw

[Michael Nazzareno Trimarchi]

Hi

On Fri, Jan 30, 2026 at 3:25 AM Wei Fang <wei.fang@xxxxxxx> wrote:
>
> > The current implementation of phy_reset_after_clk_enable requires MAC
> > drivers
> > (like fec_main.c) to manually track PHY probing states and trigger resets
> > at specific points in their clock management flow and was created just for a SoC
> > vendor. This leads to fragile code in the Ethernet driver, involving complex
> > checks to see if the PHY device is bound.
> >
> > This patch proposes moving the handling of the PHY_RST_AFTER_CLK_EN flag
> > directly
> > into the generic phy_init_hw() function.
> >
> > Signed-off-by: Michael Trimarchi <michael@xxxxxxxxxxxxxxxxxxxx>
> > ---
> >
> > I need to test better some boards that need this reset, but
> > I need to confirm if the general approach is ok
>
> This patch looks good to me, it is more generic than the previous
> solution. I do not have a board that uses SMSC LAN8710/LAN8720
> or other PHYs, so I can not help verify the solution, but I think it
> should fix the issue that reported on the i.MX platform. Thanks.
>
> Acked-by: Wei Fang <wei.fang@xxxxxxx>
>

Let me re-test and try to get rid of this flag totally. According to
the discussion seems
that should not be needed

Michael