2021-04-12 19:07:12

by Dinghao Liu

Subject: [PATCH] clk: renesas: rcar-usb2-clock-sel: Fix error handling in rcar_usb2_clock_sel_probe

When clk_get_rate() fails, a pairing PM usage counter decrement
and disable is required to prevent refcount leak. It's the same
for the subsequent error paths. When of_clk_add_hw_provider()
fails, we need to unregister clk_hw.

Signed-off-by: Dinghao Liu <[email protected]>
---
drivers/clk/renesas/rcar-usb2-clock-sel.c | 22 ++++++++++++++++++----
1 file changed, 18 insertions(+), 4 deletions(-)

diff --git a/drivers/clk/renesas/rcar-usb2-clock-sel.c b/drivers/clk/renesas/rcar-usb2-clock-sel.c
index 3abafd78f7c8..ad7bd50b9d1b 100644
--- a/drivers/clk/renesas/rcar-usb2-clock-sel.c
+++ b/drivers/clk/renesas/rcar-usb2-clock-sel.c
@@ -180,7 +180,8 @@ static int rcar_usb2_clock_sel_probe(struct platform_device *pdev)

if (!priv->extal && !priv->xtal) {
dev_err(dev, "This driver needs usb_extal or usb_xtal\n");
- return -ENOENT;
+ ret = -ENOENT;
+ goto pm_put;
}

platform_set_drvdata(pdev, priv);
@@ -194,10 +195,23 @@ static int rcar_usb2_clock_sel_probe(struct platform_device *pdev)
priv->hw.init = &init;

clk = clk_register(NULL, &priv->hw);
- if (IS_ERR(clk))
- return PTR_ERR(clk);
+ if (IS_ERR(clk)) {
+ ret = PTR_ERR(clk);
+ goto pm_put;
+ }
+
+ ret = of_clk_add_hw_provider(np, of_clk_hw_simple_get, &priv->hw);
+ if (ret)
+ goto clk_unregister;
+
+ return 0;

- return of_clk_add_hw_provider(np, of_clk_hw_simple_get, &priv->hw);
+clk_unregister:
+ clk_hw_unregister(&priv->hw);
+pm_put:
+ pm_runtime_put(dev);
+ pm_runtime_disable(dev);
+ return ret;
}

static const struct dev_pm_ops rcar_usb2_clock_sel_pm_ops = {
--
2.17.1

2021-04-12 22:19:33

by Geert Uytterhoeven

Subject: Re: [PATCH] clk: renesas: rcar-usb2-clock-sel: Fix error handling in rcar_usb2_clock_sel_probe

Hi Dinghao,

On Mon, Apr 12, 2021 at 9:51 AM Dinghao Liu <[email protected]> wrote:
> When clk_get_rate() fails, a pairing PM usage counter decrement
> and disable is required to prevent refcount leak. It's the same
> for the subsequent error paths. When of_clk_add_hw_provider()
> fails, we need to unregister clk_hw.
>
> Signed-off-by: Dinghao Liu <[email protected]>

Thanks for your patch, which looks correct to me.

> --- a/drivers/clk/renesas/rcar-usb2-clock-sel.c
> +++ b/drivers/clk/renesas/rcar-usb2-clock-sel.c
> @@ -180,7 +180,8 @@ static int rcar_usb2_clock_sel_probe(struct platform_device *pdev)
>
> if (!priv->extal && !priv->xtal) {
> dev_err(dev, "This driver needs usb_extal or usb_xtal\n");
> - return -ENOENT;
> + ret = -ENOENT;
> + goto pm_put;
> }

As the code above doesn't rely on the device being powered yet, you
could move the pm_runtime_{enable,get_sync}() calls below the clock
checks instead.

>
> platform_set_drvdata(pdev, priv);
> @@ -194,10 +195,23 @@ static int rcar_usb2_clock_sel_probe(struct platform_device *pdev)
> priv->hw.init = &init;
>
> clk = clk_register(NULL, &priv->hw);
> - if (IS_ERR(clk))
> - return PTR_ERR(clk);
> + if (IS_ERR(clk)) {
> + ret = PTR_ERR(clk);
> + goto pm_put;
> + }
> +
> + ret = of_clk_add_hw_provider(np, of_clk_hw_simple_get, &priv->hw);
> + if (ret)
> + goto clk_unregister;
> +
> + return 0;
>
> - return of_clk_add_hw_provider(np, of_clk_hw_simple_get, &priv->hw);
> +clk_unregister:
> + clk_hw_unregister(&priv->hw);

The error path can be simplified by replacing the call to clk_register()
by a call to devm_clk_register(), to match the style of the other
initialization steps.

> +pm_put:
> + pm_runtime_put(dev);
> + pm_runtime_disable(dev);
> + return ret;

This part has to stay, of course.

> }

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

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