It doesn't matter whether "mux-states" is not present or there is some
other issue parsing it causing an error. Drop the presence check and
rework the error handling to ignore anything other than deferred probe.

Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
---
Now a warning in v6.14-rc1, so please apply for 6.14.

v2:
 - Use brackets on else clause
---
 drivers/phy/phy-can-transceiver.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/drivers/phy/phy-can-transceiver.c b/drivers/phy/phy-can-transceiver.c
index ee4ce4249698..2bec70615449 100644
--- a/drivers/phy/phy-can-transceiver.c
+++ b/drivers/phy/phy-can-transceiver.c
@@ -103,6 +103,7 @@ static int can_transceiver_phy_probe(struct platform_device *pdev)
 	struct phy *phy;
 	struct gpio_desc *standby_gpio;
 	struct gpio_desc *enable_gpio;
+	struct mux_state *mux_state;
 	u32 max_bitrate = 0;
 	int err;
 
@@ -113,13 +114,11 @@ static int can_transceiver_phy_probe(struct platform_device *pdev)
 	match = of_match_node(can_transceiver_phy_ids, pdev->dev.of_node);
 	drvdata = match->data;
 
-	if (of_property_read_bool(dev->of_node, "mux-states")) {
-		struct mux_state *mux_state;
-
-		mux_state = devm_mux_state_get(dev, NULL);
-		if (IS_ERR(mux_state))
-			return dev_err_probe(&pdev->dev, PTR_ERR(mux_state),
-					     "failed to get mux\n");
+	mux_state = devm_mux_state_get(dev, NULL);
+	if (IS_ERR(mux_state)) {
+		if (PTR_ERR(mux_state) == -EPROBE_DEFER)
+			return PTR_ERR(mux_state);
+	} else {
 		can_transceiver_phy->mux_state = mux_state;
 	}
 
-- 
2.47.2

Hi Rob,

On Mon, 3 Feb 2025 at 19:55, Rob Herring (Arm) <robh@kernel.org> wrote:
> It doesn't matter whether "mux-states" is not present or there is some
> other issue parsing it causing an error. Drop the presence check and
> rework the error handling to ignore anything other than deferred probe.
>
> Acked-by: Marc Kleine-Budde <mkl@pengutronix.de>
> Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
> ---
> Now a warning in v6.14-rc1, so please apply for 6.14.
>
> v2:
>  - Use brackets on else clause

Thanks for your patch, which is now commit d02dfd4ceb2e9f34 ("phy:
can-transceiver: Drop unnecessary "mux-states" property presence check")
in phy/next (next-20250212 and later).

I have bisected the following error during boot on the Gray Hawk Single
and White Hawk development boards:

    can-transceiver-phy can-phy0: /can-phy0: failed to get mux-state (0)

> --- a/drivers/phy/phy-can-transceiver.c
> +++ b/drivers/phy/phy-can-transceiver.c
> @@ -113,13 +114,11 @@ static int can_transceiver_phy_probe(struct platform_device *pdev)
>         match = of_match_node(can_transceiver_phy_ids, pdev->dev.of_node);
>         drvdata = match->data;
>
> -       if (of_property_read_bool(dev->of_node, "mux-states")) {
> -               struct mux_state *mux_state;
> -
> -               mux_state = devm_mux_state_get(dev, NULL);
> -               if (IS_ERR(mux_state))
> -                       return dev_err_probe(&pdev->dev, PTR_ERR(mux_state),
> -                                            "failed to get mux\n");
> +       mux_state = devm_mux_state_get(dev, NULL);

If the (optional) "mux-states" property is not present, mux_get()
prints an error message, and returns -ENOENT....

> +       if (IS_ERR(mux_state)) {
> +               if (PTR_ERR(mux_state) == -EPROBE_DEFER)
> +                       return PTR_ERR(mux_state);

... which is ignored here, so in the end all is good.
Still, it is confusing to erroneously print an error message.

> +       } else {
>                 can_transceiver_phy->mux_state = mux_state;
>         }

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

On Mon, 03 Feb 2025 12:54:21 -0600, Rob Herring (Arm) wrote:
> It doesn't matter whether "mux-states" is not present or there is some
> other issue parsing it causing an error. Drop the presence check and
> rework the error handling to ignore anything other than deferred probe.
> 
> 

Applied, thanks!

[1/1] phy: can-transceiver: Drop unnecessary "mux-states" property presence check
      commit: d02dfd4ceb2e9f34caaaaf87105267e2f722f443

Best regards,
-- 
~Vinod