2018-11-21 20:20:07

by Nicholas Mc Guire

Subject: [PATCH 1/2] gpio: mt7621: report failure of devm_kasprintf()

kasprintf() may return NULL on failure of internal allocation thus the
assigned label is not safe if not explicitly checked. On error
mediatek_gpio_bank_probe() returns negative values so -ENOMEM in the
(unlikely) failure case should be fine here.

Signed-off-by: Nicholas Mc Guire <[email protected]>
Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
---

V2: The dev_err() for the unlikely case of the devm_kasprintf()
failing was removed on request from Bartosz Golaszewski
<[email protected]>

Problem located with experimental coccinelle script

Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
GPIOLIB=y, GPIO_MT7621=y

Patch is against 4.20-rc3 (localversion-next is next-20181121)

drivers/gpio/gpio-mt7621.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
index d72af6f..1ec95bc 100644
--- a/drivers/gpio/gpio-mt7621.c
+++ b/drivers/gpio/gpio-mt7621.c
@@ -244,6 +244,8 @@ mediatek_gpio_bank_probe(struct device *dev,
rg->chip.of_xlate = mediatek_gpio_xlate;
rg->chip.label = devm_kasprintf(dev, GFP_KERNEL, "%s-bank%d",
dev_name(dev), bank);
+ if (!rg->chip.label)
+ return -ENOMEM;

ret = devm_gpiochip_add_data(dev, &rg->chip, mtk);
if (ret < 0) {
--
2.1.4

2018-11-21 20:29:35

by Nicholas Mc Guire

Subject: [PATCH 2/2] gpio: mt7621: pass mediatek_gpio_bank_probe() failure up the stack

The error cases of mediatek_gpio_bank_probe() would go unnoticed (except
for the dev_err() messages). The probe function should return an error
if one of the banks failed to initialize properly.

Signed-off-by: Nicholas Mc Guire <[email protected]>
Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
---

Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
GPIOLIB=y, GPIO_MT7621=y

Patch is against 4.20-rc3 (localversion-next is next-20181121)

drivers/gpio/gpio-mt7621.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
index 1ec95bc..68fca8b 100644
--- a/drivers/gpio/gpio-mt7621.c
+++ b/drivers/gpio/gpio-mt7621.c
@@ -297,6 +297,7 @@ mediatek_gpio_probe(struct platform_device *pdev)
struct device_node *np = dev->of_node;
struct mtk *mtk;
int i;
+ int ret;

mtk = devm_kzalloc(dev, sizeof(*mtk), GFP_KERNEL);
if (!mtk)
@@ -311,8 +312,11 @@ mediatek_gpio_probe(struct platform_device *pdev)
platform_set_drvdata(pdev, mtk);
mediatek_gpio_irq_chip.name = dev_name(dev);

- for (i = 0; i < MTK_BANK_CNT; i++)
- mediatek_gpio_bank_probe(dev, np, i);
+ for (i = 0; i < MTK_BANK_CNT; i++) {
+ ret = mediatek_gpio_bank_probe(dev, np, i);
+ if (!ret)
+ return ret;
+ }

return 0;
}
--
2.1.4

2018-11-24 03:14:27

by Bartosz Golaszewski

Subject: Re: [PATCH 2/2] gpio: mt7621: pass mediatek_gpio_bank_probe() failure up the stack

śr., 21 lis 2018 o 19:13 Nicholas Mc Guire <[email protected]> napisał(a):
>
> The error cases of mediatek_gpio_bank_probe() would go unnoticed (except
> for the dev_err() messages). The probe function should return an error
> if one of the banks failed to initialize properly.
>
> Signed-off-by: Nicholas Mc Guire <[email protected]>
> Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
> ---
>
> Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
> GPIOLIB=y, GPIO_MT7621=y
>
> Patch is against 4.20-rc3 (localversion-next is next-20181121)
>
> drivers/gpio/gpio-mt7621.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
> index 1ec95bc..68fca8b 100644
> --- a/drivers/gpio/gpio-mt7621.c
> +++ b/drivers/gpio/gpio-mt7621.c
> @@ -297,6 +297,7 @@ mediatek_gpio_probe(struct platform_device *pdev)
> struct device_node *np = dev->of_node;
> struct mtk *mtk;
> int i;
> + int ret;
>
> mtk = devm_kzalloc(dev, sizeof(*mtk), GFP_KERNEL);
> if (!mtk)
> @@ -311,8 +312,11 @@ mediatek_gpio_probe(struct platform_device *pdev)
> platform_set_drvdata(pdev, mtk);
> mediatek_gpio_irq_chip.name = dev_name(dev);
>
> - for (i = 0; i < MTK_BANK_CNT; i++)
> - mediatek_gpio_bank_probe(dev, np, i);
> + for (i = 0; i < MTK_BANK_CNT; i++) {
> + ret = mediatek_gpio_bank_probe(dev, np, i);
> + if (!ret)
> + return ret;
> + }

Looks like are resources allocated in mediatek_gpio_bank_probe() will
be freed automatically so:

Reviewed-by: Bartosz Golaszewski <[email protected]>

2018-11-24 03:14:56

by Bartosz Golaszewski

Subject: Re: [PATCH 1/2] gpio: mt7621: report failure of devm_kasprintf()

śr., 21 lis 2018 o 19:13 Nicholas Mc Guire <[email protected]> napisał(a):
>
> kasprintf() may return NULL on failure of internal allocation thus the
> assigned label is not safe if not explicitly checked. On error
> mediatek_gpio_bank_probe() returns negative values so -ENOMEM in the
> (unlikely) failure case should be fine here.
>
> Signed-off-by: Nicholas Mc Guire <[email protected]>
> Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
> ---
>
> V2: The dev_err() for the unlikely case of the devm_kasprintf()
> failing was removed on request from Bartosz Golaszewski
> <[email protected]>
>
> Problem located with experimental coccinelle script
>
> Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
> GPIOLIB=y, GPIO_MT7621=y
>
> Patch is against 4.20-rc3 (localversion-next is next-20181121)
>
> drivers/gpio/gpio-mt7621.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
> index d72af6f..1ec95bc 100644
> --- a/drivers/gpio/gpio-mt7621.c
> +++ b/drivers/gpio/gpio-mt7621.c
> @@ -244,6 +244,8 @@ mediatek_gpio_bank_probe(struct device *dev,
> rg->chip.of_xlate = mediatek_gpio_xlate;
> rg->chip.label = devm_kasprintf(dev, GFP_KERNEL, "%s-bank%d",
> dev_name(dev), bank);
> + if (!rg->chip.label)
> + return -ENOMEM;
>
> ret = devm_gpiochip_add_data(dev, &rg->chip, mtk);
> if (ret < 0) {
> --
> 2.1.4
>

Reviewed-by: Bartosz Golaszewski <[email protected]>

2018-11-27 07:51:06

by Sean Wang

Subject: Re: [PATCH 2/2] gpio: mt7621: pass mediatek_gpio_bank_probe() failure up the stack

Nicholas Mc Guire <[email protected]> 於 2018年11月21日 週三 上午10:13寫道：
>
> The error cases of mediatek_gpio_bank_probe() would go unnoticed (except
> for the dev_err() messages). The probe function should return an error
> if one of the banks failed to initialize properly.
>
> Signed-off-by: Nicholas Mc Guire <[email protected]>
> Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
> ---
>
> Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
> GPIOLIB=y, GPIO_MT7621=y
>
> Patch is against 4.20-rc3 (localversion-next is next-20181121)
>
> drivers/gpio/gpio-mt7621.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
> index 1ec95bc..68fca8b 100644
> --- a/drivers/gpio/gpio-mt7621.c
> +++ b/drivers/gpio/gpio-mt7621.c
> @@ -297,6 +297,7 @@ mediatek_gpio_probe(struct platform_device *pdev)
> struct device_node *np = dev->of_node;
> struct mtk *mtk;
> int i;
> + int ret;
>
> mtk = devm_kzalloc(dev, sizeof(*mtk), GFP_KERNEL);
> if (!mtk)
> @@ -311,8 +312,11 @@ mediatek_gpio_probe(struct platform_device *pdev)
> platform_set_drvdata(pdev, mtk);
> mediatek_gpio_irq_chip.name = dev_name(dev);
>
> - for (i = 0; i < MTK_BANK_CNT; i++)
> - mediatek_gpio_bank_probe(dev, np, i);
> + for (i = 0; i < MTK_BANK_CNT; i++) {
> + ret = mediatek_gpio_bank_probe(dev, np, i);
> + if (!ret)

it should be if (ret < 0) ?

> + return ret;
> + }
>
> return 0;
> }
> --
> 2.1.4
>
>
> _______________________________________________
> Linux-mediatek mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-mediatek

2018-11-27 08:02:45

by Nicholas Mc Guire

Subject: Re: [PATCH 2/2] gpio: mt7621: pass mediatek_gpio_bank_probe() failure up the stack

On Mon, Nov 26, 2018 at 11:49:26PM -0800, Sean Wang wrote:
> Nicholas Mc Guire <[email protected]> ??? 2018???11???21??? ?????? ??????10:13?????????
> >
> > The error cases of mediatek_gpio_bank_probe() would go unnoticed (except
> > for the dev_err() messages). The probe function should return an error
> > if one of the banks failed to initialize properly.
> >
> > Signed-off-by: Nicholas Mc Guire <[email protected]>
> > Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
> > ---
> >
> > Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
> > GPIOLIB=y, GPIO_MT7621=y
> >
> > Patch is against 4.20-rc3 (localversion-next is next-20181121)
> >
> > drivers/gpio/gpio-mt7621.c | 8 ++++++--
> > 1 file changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
> > index 1ec95bc..68fca8b 100644
> > --- a/drivers/gpio/gpio-mt7621.c
> > +++ b/drivers/gpio/gpio-mt7621.c
> > @@ -297,6 +297,7 @@ mediatek_gpio_probe(struct platform_device *pdev)
> > struct device_node *np = dev->of_node;
> > struct mtk *mtk;
> > int i;
> > + int ret;
> >
> > mtk = devm_kzalloc(dev, sizeof(*mtk), GFP_KERNEL);
> > if (!mtk)
> > @@ -311,8 +312,11 @@ mediatek_gpio_probe(struct platform_device *pdev)
> > platform_set_drvdata(pdev, mtk);
> > mediatek_gpio_irq_chip.name = dev_name(dev);
> >
> > - for (i = 0; i < MTK_BANK_CNT; i++)
> > - mediatek_gpio_bank_probe(dev, np, i);
> > + for (i = 0; i < MTK_BANK_CNT; i++) {
> > + ret = mediatek_gpio_bank_probe(dev, np, i);
> > + if (!ret)
>
> it should be if (ret < 0) ?

I don?t think so mediatek_gpio_bank_probe() returns 0 on success
and all other returns are error paths - while the current code
only returns negative values I do thik that any non 0 would be
an error indication so !ret should be fine here.

thx!
hofrat

>
> > + return ret;
> > + }
> >
> > return 0;
> > }
> > --
> > 2.1.4
> >
> >
> > _______________________________________________
> > Linux-mediatek mailing list
> > [email protected]
> > http://lists.infradead.org/mailman/listinfo/linux-mediatek

2018-11-27 08:11:38

by Sean Wang

Subject: Re: [PATCH 1/2] gpio: mt7621: report failure of devm_kasprintf()

Nicholas Mc Guire <[email protected]> 於 2018年11月21日 週三 上午10:13寫道：
>
> kasprintf() may return NULL on failure of internal allocation thus the
> assigned label is not safe if not explicitly checked. On error
> mediatek_gpio_bank_probe() returns negative values so -ENOMEM in the
> (unlikely) failure case should be fine here.
>
> Signed-off-by: Nicholas Mc Guire <[email protected]>

Acked-by: Sean Wang <[email protected]>

> Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
> ---
>
> V2: The dev_err() for the unlikely case of the devm_kasprintf()
> failing was removed on request from Bartosz Golaszewski
> <[email protected]>
>
> Problem located with experimental coccinelle script
>
> Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
> GPIOLIB=y, GPIO_MT7621=y
>
> Patch is against 4.20-rc3 (localversion-next is next-20181121)
>
> drivers/gpio/gpio-mt7621.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
> index d72af6f..1ec95bc 100644
> --- a/drivers/gpio/gpio-mt7621.c
> +++ b/drivers/gpio/gpio-mt7621.c
> @@ -244,6 +244,8 @@ mediatek_gpio_bank_probe(struct device *dev,
> rg->chip.of_xlate = mediatek_gpio_xlate;
> rg->chip.label = devm_kasprintf(dev, GFP_KERNEL, "%s-bank%d",
> dev_name(dev), bank);
> + if (!rg->chip.label)
> + return -ENOMEM;
>
> ret = devm_gpiochip_add_data(dev, &rg->chip, mtk);
> if (ret < 0) {
> --
> 2.1.4
>
>
> _______________________________________________
> Linux-mediatek mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-mediatek

2018-11-27 08:34:52

by Sean Wang

Subject: Re: [PATCH 2/2] gpio: mt7621: pass mediatek_gpio_bank_probe() failure up the stack

>
> On Mon, Nov 26, 2018 at 11:49:26PM -0800, Sean Wang wrote:
> > Nicholas Mc Guire <[email protected]> ??? 2018???11???21??? ?????? ??????10:13?????????
> > >
> > > The error cases of mediatek_gpio_bank_probe() would go unnoticed (except
> > > for the dev_err() messages). The probe function should return an error
> > > if one of the banks failed to initialize properly.
> > >
> > > Signed-off-by: Nicholas Mc Guire <[email protected]>
> > > Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
> > > ---
> > >
> > > Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
> > > GPIOLIB=y, GPIO_MT7621=y
> > >
> > > Patch is against 4.20-rc3 (localversion-next is next-20181121)
> > >
> > > drivers/gpio/gpio-mt7621.c | 8 ++++++--
> > > 1 file changed, 6 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
> > > index 1ec95bc..68fca8b 100644
> > > --- a/drivers/gpio/gpio-mt7621.c
> > > +++ b/drivers/gpio/gpio-mt7621.c
> > > @@ -297,6 +297,7 @@ mediatek_gpio_probe(struct platform_device *pdev)
> > > struct device_node *np = dev->of_node;
> > > struct mtk *mtk;
> > > int i;
> > > + int ret;
> > >
> > > mtk = devm_kzalloc(dev, sizeof(*mtk), GFP_KERNEL);
> > > if (!mtk)
> > > @@ -311,8 +312,11 @@ mediatek_gpio_probe(struct platform_device *pdev)
> > > platform_set_drvdata(pdev, mtk);
> > > mediatek_gpio_irq_chip.name = dev_name(dev);
> > >
> > > - for (i = 0; i < MTK_BANK_CNT; i++)
> > > - mediatek_gpio_bank_probe(dev, np, i);
> > > + for (i = 0; i < MTK_BANK_CNT; i++) {
> > > + ret = mediatek_gpio_bank_probe(dev, np, i);
> > > + if (!ret)
> >
> > it should be if (ret < 0) ?
>
> I don´t think so mediatek_gpio_bank_probe() returns 0 on success
> and all other returns are error paths - while the current code
> only returns negative values I do thik that any non 0 would be
> an error indication so !ret should be fine here.
>
!0 would be true

> thx!
> hofrat
>
> >
> > > + return ret;
> > > + }
> > >
> > > return 0;
> > > }
> > > --
> > > 2.1.4
> > >
> > >
> > > _______________________________________________
> > > Linux-mediatek mailing list
> > > [email protected]
> > > http://lists.infradead.org/mailman/listinfo/linux-mediatek
>
> _______________________________________________
> Linux-mediatek mailing list
> [email protected]
> http://lists.infradead.org/mailman/listinfo/linux-mediatek

2018-11-27 08:40:30

by Nicholas Mc Guire

Subject: Re: [PATCH 2/2] gpio: mt7621: pass mediatek_gpio_bank_probe() failure up the stack

On Tue, Nov 27, 2018 at 12:32:59AM -0800, Sean Wang wrote:
> >
> > On Mon, Nov 26, 2018 at 11:49:26PM -0800, Sean Wang wrote:
> > > Nicholas Mc Guire <[email protected]> ??? 2018???11???21??? ?????? ??????10:13?????????
> > > >
> > > > The error cases of mediatek_gpio_bank_probe() would go unnoticed (except
> > > > for the dev_err() messages). The probe function should return an error
> > > > if one of the banks failed to initialize properly.
> > > >
> > > > Signed-off-by: Nicholas Mc Guire <[email protected]>
> > > > Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")
> > > > ---
> > > >
> > > > Patch was compile tested with: omega2p_defconfig, SOC_MT7621=y,
> > > > GPIOLIB=y, GPIO_MT7621=y
> > > >
> > > > Patch is against 4.20-rc3 (localversion-next is next-20181121)
> > > >
> > > > drivers/gpio/gpio-mt7621.c | 8 ++++++--
> > > > 1 file changed, 6 insertions(+), 2 deletions(-)
> > > >
> > > > diff --git a/drivers/gpio/gpio-mt7621.c b/drivers/gpio/gpio-mt7621.c
> > > > index 1ec95bc..68fca8b 100644
> > > > --- a/drivers/gpio/gpio-mt7621.c
> > > > +++ b/drivers/gpio/gpio-mt7621.c
> > > > @@ -297,6 +297,7 @@ mediatek_gpio_probe(struct platform_device *pdev)
> > > > struct device_node *np = dev->of_node;
> > > > struct mtk *mtk;
> > > > int i;
> > > > + int ret;
> > > >
> > > > mtk = devm_kzalloc(dev, sizeof(*mtk), GFP_KERNEL);
> > > > if (!mtk)
> > > > @@ -311,8 +312,11 @@ mediatek_gpio_probe(struct platform_device *pdev)
> > > > platform_set_drvdata(pdev, mtk);
> > > > mediatek_gpio_irq_chip.name = dev_name(dev);
> > > >
> > > > - for (i = 0; i < MTK_BANK_CNT; i++)
> > > > - mediatek_gpio_bank_probe(dev, np, i);
> > > > + for (i = 0; i < MTK_BANK_CNT; i++) {
> > > > + ret = mediatek_gpio_bank_probe(dev, np, i);
> > > > + if (!ret)
> > >
> > > it should be if (ret < 0) ?
> >
> > I don?t think so mediatek_gpio_bank_probe() returns 0 on success
> > and all other returns are error paths - while the current code
> > only returns negative values I do thik that any non 0 would be
> > an error indication so !ret should be fine here.
> >
> !0 would be true
>
...sorry - stupid me - thanks for catching that !

thx!
hofrat

2018-12-07 09:47:39

by Linus Walleij

Subject: Re: [PATCH 1/2] gpio: mt7621: report failure of devm_kasprintf()

On Wed, Nov 21, 2018 at 7:12 PM Nicholas Mc Guire <[email protected]> wrote:

> kasprintf() may return NULL on failure of internal allocation thus the
> assigned label is not safe if not explicitly checked. On error
> mediatek_gpio_bank_probe() returns negative values so -ENOMEM in the
> (unlikely) failure case should be fine here.
>
> Signed-off-by: Nicholas Mc Guire <[email protected]>
> Fixes: 4ba9c3afda41 ("gpio: mt7621: Add a driver for MT7621")

Patch applied with Bartosz and Sean's tags.

Yours,
Linus Walleij