[PATCH] mtd: nand: vf610: Check the return value from clk_prepare_enable()

[PATCH] mtd: nand: vf610: Check the return value from clk_prepare_enable()

[Fabio Estevam]

From: Fabio Estevam <fabio.estevam@nxp.com>

clk_prepare_enable() may fail, so we should better check its return value
and propagate it in the case of error.

Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
---
 drivers/mtd/nand/vf610_nfc.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/mtd/nand/vf610_nfc.c b/drivers/mtd/nand/vf610_nfc.c
index 744ab10..9e49672 100644
--- a/drivers/mtd/nand/vf610_nfc.c
+++ b/drivers/mtd/nand/vf610_nfc.c
@@ -814,12 +814,16 @@ static int vf610_nfc_suspend(struct device *dev)
 
 static int vf610_nfc_resume(struct device *dev)
 {
+	int err;
+
 	struct mtd_info *mtd = dev_get_drvdata(dev);
 	struct vf610_nfc *nfc = mtd_to_nfc(mtd);
 
 	pinctrl_pm_select_default_state(dev);
 
-	clk_prepare_enable(nfc->clk);
+	err = clk_prepare_enable(nfc->clk);
+	if (err)
+		return err;
 
 	vf610_nfc_preinit_controller(nfc);
 	vf610_nfc_init_controller(nfc);
-- 
2.7.4

Re: [PATCH] mtd: nand: vf610: Check the return value from clk_prepare_enable()

[Boris Brezillon]

Le Mon, 17 Jul 2017 17:55:59 -0300,
Fabio Estevam <festevam@gmail.com> a écrit :

> From: Fabio Estevam <fabio.estevam@nxp.com>
> 
> clk_prepare_enable() may fail, so we should better check its return value
> and propagate it in the case of error.
> 
> Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
> ---
>  drivers/mtd/nand/vf610_nfc.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mtd/nand/vf610_nfc.c b/drivers/mtd/nand/vf610_nfc.c
> index 744ab10..9e49672 100644
> --- a/drivers/mtd/nand/vf610_nfc.c
> +++ b/drivers/mtd/nand/vf610_nfc.c
> @@ -814,12 +814,16 @@ static int vf610_nfc_suspend(struct device *dev)
>  
>  static int vf610_nfc_resume(struct device *dev)
>  {
> +	int err;
> +
>  	struct mtd_info *mtd = dev_get_drvdata(dev);
>  	struct vf610_nfc *nfc = mtd_to_nfc(mtd);
>  
>  	pinctrl_pm_select_default_state(dev);

AFAICT pinctrl_pm_select_default_state() can also fail. What makes it
different from clk_prepare_enable() failures?

>  
> -	clk_prepare_enable(nfc->clk);
> +	err = clk_prepare_enable(nfc->clk);
> +	if (err)
> +		return err;
>

I guess the idea was that clk_prepare_enabled already succeeded at probe
time and there was not real reason for it to fail at resume time. But
extra checks are never bad. BTW, I don't remember what happens if
->resume() returns an error. Does it prevent the whole system from
resuming or just make this specific device unusable?

>  	vf610_nfc_preinit_controller(nfc);
>  	vf610_nfc_init_controller(nfc);

Re: [PATCH] mtd: nand: vf610: Check the return value from clk_prepare_enable()

[Boris Brezillon]

Le Mon, 17 Jul 2017 23:08:28 +0200,
Boris Brezillon <boris.brezillon@free-electrons.com> a écrit :

> Le Mon, 17 Jul 2017 17:55:59 -0300,
> Fabio Estevam <festevam@gmail.com> a écrit :
> 
> > From: Fabio Estevam <fabio.estevam@nxp.com>
> > 
> > clk_prepare_enable() may fail, so we should better check its return value
> > and propagate it in the case of error.
> > 
> > Signed-off-by: Fabio Estevam <fabio.estevam@nxp.com>
> > ---
> >  drivers/mtd/nand/vf610_nfc.c | 6 +++++-
> >  1 file changed, 5 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/mtd/nand/vf610_nfc.c b/drivers/mtd/nand/vf610_nfc.c
> > index 744ab10..9e49672 100644
> > --- a/drivers/mtd/nand/vf610_nfc.c
> > +++ b/drivers/mtd/nand/vf610_nfc.c
> > @@ -814,12 +814,16 @@ static int vf610_nfc_suspend(struct device *dev)
> >  
> >  static int vf610_nfc_resume(struct device *dev)
> >  {
> > +	int err;
> > +
> >  	struct mtd_info *mtd = dev_get_drvdata(dev);
> >  	struct vf610_nfc *nfc = mtd_to_nfc(mtd);
> >  
> >  	pinctrl_pm_select_default_state(dev);  
> 
> AFAICT pinctrl_pm_select_default_state() can also fail. What makes it
> different from clk_prepare_enable() failures?
> 
> >  
> > -	clk_prepare_enable(nfc->clk);
> > +	err = clk_prepare_enable(nfc->clk);
> > +	if (err)
> > +		return err;
> >  
> 
> I guess the idea was that clk_prepare_enabled already succeeded at probe
> time and there was not real reason for it to fail at resume time. But
> extra checks are never bad. BTW, I don't remember what happens if
> ->resume() returns an error. Does it prevent the whole system from  
> resuming or just make this specific device unusable?

Just found the answer [1]: it's harmless.

[1]http://elixir.free-electrons.com/linux/latest/source/include/linux/pm.h#L262

Re: [PATCH] mtd: nand: vf610: Check the return value from clk_prepare_enable()

[Fabio Estevam]

Hi Boris,

On Mon, Jul 17, 2017 at 6:08 PM, Boris Brezillon
<boris.brezillon@free-electrons.com> wrote:

> AFAICT pinctrl_pm_select_default_state() can also fail. What makes it
> different from clk_prepare_enable() failures?

IMHO we could even remove pinctrl_pm_select_default_state() from this
driver since pinctrl_pm_select_sleep_state() is never used.

This is the only driver that uses pinctrl_pm_select_default_state()
inside drivers/mtd/.