| Message ID | 1573231008-24239-1-git-send-email-fabrizio.castro@bp.renesas.com (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | [RFT/RFC,linux-4.4.y-cip] gpiolib: Fix bad of_node pointer | expand |
Hi Fab, This patch is good for my system (ls1021a), so many thanks! Thanks, Johnson -----Original Message----- From: Fabrizio Castro <fabrizio.castro@bp.renesas.com> Sent: Saturday, November 9, 2019 12:37 AM To: cip-dev@lists.cip-project.org Cc: nobuhiro1.iwamatsu@toshiba.co.jp; pavel@denx.de; Chris Paterson <Chris.Paterson2@renesas.com>; Biju Das <biju.das@bp.renesas.com>; Fabrizio Castro <fabrizio.castro@bp.renesas.com>; Johnson CH Chen (陳昭勳) <JohnsonCH.Chen@moxa.com> Subject: [cip-dev][RFT/RFC linux-4.4.y-cip] gpiolib: Fix bad of_node pointer Not every driver initialises of_node from struct gpio_chip, therefore the replacement of of_node from struct gpio_chip with dev->of_node in the below commit won't work on every platform: baff4777cdb8 ("gpiolib: Support 'gpio-reserved-ranges' property") The final result is that on some platforms the kernel will try to dereference a NULL pointer, with obvious consequences. This patch makes sure the pointer gets initialised before its first usage. Fixes: baff4777cdb8 ("gpiolib: Support 'gpio-reserved-ranges' property") Reported-by: Johnson CH Chen <JohnsonCH.Chen@moxa.com> Signed-off-by: Fabrizio Castro <fabrizio.castro@bp.renesas.com> --- Hi Johnson, could you please test this patch on your system? Thanks, Fab drivers/gpio/gpiolib-of.c | 4 +++- drivers/gpio/gpiolib.c | 6 +++++- 2 files changed, 8 insertions(+), 2 deletions(-) diff --git a/drivers/gpio/gpiolib-of.c b/drivers/gpio/gpiolib-of.c index ec642bf..36b60dd 100644 --- a/drivers/gpio/gpiolib-of.c +++ b/drivers/gpio/gpiolib-of.c @@ -338,7 +338,7 @@ static void of_gpiochip_init_valid_mask(struct gpio_chip *chip) { int len, i; u32 start, count; - struct device_node *np = chip->dev->of_node; + struct device_node *np = chip->of_node; len = of_property_count_u32_elems(np, "gpio-reserved-ranges"); if (len < 0 || len % 2 != 0) @@ -445,8 +445,10 @@ int of_gpiochip_add(struct gpio_chip *chip) { int status; + /* if ((!chip->of_node) && (chip->dev)) chip->of_node = chip->dev->of_node; + */ if (!chip->of_node) return 0; diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c index d72218f..0fd443a 100644 --- a/drivers/gpio/gpiolib.c +++ b/drivers/gpio/gpiolib.c @@ -296,7 +296,7 @@ static int gpiochip_init_valid_mask(struct gpio_chip *gpiochip) { #ifdef CONFIG_OF_GPIO int size; - struct device_node *np = gpiochip->dev->of_node; + struct device_node *np = gpiochip->of_node; size = of_property_count_u32_elems(np, "gpio-reserved-ranges"); if (size > 0 && size % 2 == 0) @@ -354,12 +354,16 @@ int gpiochip_add_data(struct gpio_chip *chip, void *data) int base = chip->base; struct gpio_desc *descs; + descs = kcalloc(chip->ngpio, sizeof(descs[0]), GFP_KERNEL); if (!descs) return -ENOMEM; chip->data = data; + if ((!chip->of_node) && (chip->dev)) + chip->of_node = chip->dev->of_node; + spin_lock_irqsave(&gpio_lock, flags); if (base < 0) { -- 2.7.4
Thank you for testing! Cheers, Fab > -----Original Message----- > From: Johnson CH Chen (陳昭勳) <JohnsonCH.Chen@moxa.com> > Sent: 11 November 2019 03:36 > To: Fabrizio Castro <fabrizio.castro@bp.renesas.com>; cip-dev@lists.cip-project.org > Cc: nobuhiro1.iwamatsu@toshiba.co.jp; pavel@denx.de; Chris Paterson <Chris.Paterson2@renesas.com>; Biju Das > <biju.das@bp.renesas.com> > Subject: RE: [cip-dev][RFT/RFC linux-4.4.y-cip] gpiolib: Fix bad of_node pointer > > Hi Fab, > > This patch is good for my system (ls1021a), so many thanks! > > Thanks, > Johnson > > -----Original Message----- > From: Fabrizio Castro <fabrizio.castro@bp.renesas.com> > Sent: Saturday, November 9, 2019 12:37 AM > To: cip-dev@lists.cip-project.org > Cc: nobuhiro1.iwamatsu@toshiba.co.jp; pavel@denx.de; Chris Paterson <Chris.Paterson2@renesas.com>; Biju Das > <biju.das@bp.renesas.com>; Fabrizio Castro <fabrizio.castro@bp.renesas.com>; Johnson CH Chen (陳昭勳) > <JohnsonCH.Chen@moxa.com> > Subject: [cip-dev][RFT/RFC linux-4.4.y-cip] gpiolib: Fix bad of_node pointer > > Not every driver initialises of_node from struct gpio_chip, therefore the replacement of of_node from struct gpio_chip with dev- > >of_node in the below commit won't work on every > platform: > baff4777cdb8 ("gpiolib: Support 'gpio-reserved-ranges' property") The final result is that on some platforms the kernel will try to > dereference a NULL pointer, with obvious consequences. > > This patch makes sure the pointer gets initialised before its first usage. > > Fixes: baff4777cdb8 ("gpiolib: Support 'gpio-reserved-ranges' property") > Reported-by: Johnson CH Chen <JohnsonCH.Chen@moxa.com> > Signed-off-by: Fabrizio Castro <fabrizio.castro@bp.renesas.com> > --- > > Hi Johnson, > > could you please test this patch on your system? > > Thanks, > Fab > > > drivers/gpio/gpiolib-of.c | 4 +++- > drivers/gpio/gpiolib.c | 6 +++++- > 2 files changed, 8 insertions(+), 2 deletions(-) > > diff --git a/drivers/gpio/gpiolib-of.c b/drivers/gpio/gpiolib-of.c index ec642bf..36b60dd 100644 > --- a/drivers/gpio/gpiolib-of.c > +++ b/drivers/gpio/gpiolib-of.c > @@ -338,7 +338,7 @@ static void of_gpiochip_init_valid_mask(struct gpio_chip *chip) { > int len, i; > u32 start, count; > - struct device_node *np = chip->dev->of_node; > + struct device_node *np = chip->of_node; > > len = of_property_count_u32_elems(np, "gpio-reserved-ranges"); > if (len < 0 || len % 2 != 0) > @@ -445,8 +445,10 @@ int of_gpiochip_add(struct gpio_chip *chip) { > int status; > > + /* > if ((!chip->of_node) && (chip->dev)) > chip->of_node = chip->dev->of_node; > + */ > > if (!chip->of_node) > return 0; > diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c index d72218f..0fd443a 100644 > --- a/drivers/gpio/gpiolib.c > +++ b/drivers/gpio/gpiolib.c > @@ -296,7 +296,7 @@ static int gpiochip_init_valid_mask(struct gpio_chip *gpiochip) { #ifdef CONFIG_OF_GPIO > int size; > - struct device_node *np = gpiochip->dev->of_node; > + struct device_node *np = gpiochip->of_node; > > size = of_property_count_u32_elems(np, "gpio-reserved-ranges"); > if (size > 0 && size % 2 == 0) > @@ -354,12 +354,16 @@ int gpiochip_add_data(struct gpio_chip *chip, void *data) > int base = chip->base; > struct gpio_desc *descs; > > + > descs = kcalloc(chip->ngpio, sizeof(descs[0]), GFP_KERNEL); > if (!descs) > return -ENOMEM; > > chip->data = data; > > + if ((!chip->of_node) && (chip->dev)) > + chip->of_node = chip->dev->of_node; > + > spin_lock_irqsave(&gpio_lock, flags); > > if (base < 0) { > -- > 2.7.4
diff --git a/drivers/gpio/gpiolib-of.c b/drivers/gpio/gpiolib-of.c index ec642bf..36b60dd 100644 --- a/drivers/gpio/gpiolib-of.c +++ b/drivers/gpio/gpiolib-of.c @@ -338,7 +338,7 @@ static void of_gpiochip_init_valid_mask(struct gpio_chip *chip) { int len, i; u32 start, count; - struct device_node *np = chip->dev->of_node; + struct device_node *np = chip->of_node; len = of_property_count_u32_elems(np, "gpio-reserved-ranges"); if (len < 0 || len % 2 != 0) @@ -445,8 +445,10 @@ int of_gpiochip_add(struct gpio_chip *chip) { int status; + /* if ((!chip->of_node) && (chip->dev)) chip->of_node = chip->dev->of_node; + */ if (!chip->of_node) return 0; diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c index d72218f..0fd443a 100644 --- a/drivers/gpio/gpiolib.c +++ b/drivers/gpio/gpiolib.c @@ -296,7 +296,7 @@ static int gpiochip_init_valid_mask(struct gpio_chip *gpiochip) { #ifdef CONFIG_OF_GPIO int size; - struct device_node *np = gpiochip->dev->of_node; + struct device_node *np = gpiochip->of_node; size = of_property_count_u32_elems(np, "gpio-reserved-ranges"); if (size > 0 && size % 2 == 0) @@ -354,12 +354,16 @@ int gpiochip_add_data(struct gpio_chip *chip, void *data) int base = chip->base; struct gpio_desc *descs; + descs = kcalloc(chip->ngpio, sizeof(descs[0]), GFP_KERNEL); if (!descs) return -ENOMEM; chip->data = data; + if ((!chip->of_node) && (chip->dev)) + chip->of_node = chip->dev->of_node; + spin_lock_irqsave(&gpio_lock, flags); if (base < 0) {
Not every driver initialises of_node from struct gpio_chip, therefore the replacement of of_node from struct gpio_chip with dev->of_node in the below commit won't work on every platform: baff4777cdb8 ("gpiolib: Support 'gpio-reserved-ranges' property") The final result is that on some platforms the kernel will try to dereference a NULL pointer, with obvious consequences. This patch makes sure the pointer gets initialised before its first usage. Fixes: baff4777cdb8 ("gpiolib: Support 'gpio-reserved-ranges' property") Reported-by: Johnson CH Chen <JohnsonCH.Chen@moxa.com> Signed-off-by: Fabrizio Castro <fabrizio.castro@bp.renesas.com> --- Hi Johnson, could you please test this patch on your system? Thanks, Fab drivers/gpio/gpiolib-of.c | 4 +++- drivers/gpio/gpiolib.c | 6 +++++- 2 files changed, 8 insertions(+), 2 deletions(-)