Re: [PATCH] usb: misc: usb3503: call clk_disable_unprepare in the error handling

From: Dongliang Mu
Date: Thu Sep 08 2022 - 02:02:37 EST


On Wed, Sep 7, 2022 at 10:25 PM Greg Kroah-Hartman
<gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Sat, Sep 03, 2022 at 03:15:40PM +0800, Dongliang Mu wrote:
> > From: Dongliang Mu <mudongliangabcd@xxxxxxxxx>
> >
> > Smatch reports the following warning:
> >
> > vers/usb/misc/usb3503.c:267 usb3503_probe() warn: 'hub->clk'
> > from clk_prepare_enable() not released on lines: 240,246,252
> >
> > Fix this by adding a flag to indicate if hub->clk is prepared or not and
> > invoke clk_disable_unprepare in the error handling.
> >
> > Signed-off-by: Dongliang Mu <mudongliangabcd@xxxxxxxxx>
> > ---
> > drivers/usb/misc/usb3503.c | 18 +++++++++++++++---
> > 1 file changed, 15 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/usb/misc/usb3503.c b/drivers/usb/misc/usb3503.c
> > index 330f494cd158..add47dd964b2 100644
> > --- a/drivers/usb/misc/usb3503.c
> > +++ b/drivers/usb/misc/usb3503.c
> > @@ -160,6 +160,7 @@ static int usb3503_probe(struct usb3503 *hub)
> > struct usb3503_platform_data *pdata = dev_get_platdata(dev);
> > struct device_node *np = dev->of_node;
> > int err;
> > + int is_clk_enable = 0;
>
> bool?
>
> > u32 mode = USB3503_MODE_HUB;
> > const u32 *property;
> > enum gpiod_flags flags;
> > @@ -217,6 +218,8 @@ static int usb3503_probe(struct usb3503 *hub)
> > return err;
> > }
> >
> > + // set a flag for successful clk_prepare_enable
>
> Comment isn't needed.
>
> > + is_clk_enable = 1;
> > property = of_get_property(np, "disabled-ports", &len);
> > if (property && (len / sizeof(u32)) > 0) {
> > int i;
> > @@ -236,20 +239,29 @@ static int usb3503_probe(struct usb3503 *hub)
> > else
> > flags = GPIOD_OUT_HIGH;
> > hub->intn = devm_gpiod_get_optional(dev, "intn", flags);
> > - if (IS_ERR(hub->intn))
> > + if (IS_ERR(hub->intn)) {
> > + if (is_clk_enable)
> > + clk_disable_unprepare(hub->clk);
> > return PTR_ERR(hub->intn);
>
> This is getting messy, any way to make a common error handler at the
> bottom of the function instead? That's the common kernel coding style
> for this type of thing.

I have addressed all the issues mentioned above:
1. change is_clk_enable to is_clk_enabled, and change its type to bool.
2. remove the comment and move the error handling code to the end of
probe function.
>
> thanks,
>
> greg k-h