Skip to contentSkip to conversation

gpio: Remove VLA from stmpe driver

PATCH In review Laura Abbott 4 participants 4 messages Mar 28, 2018 – May 24, 2018

Subsystems (2)
GPIO Arm STM32
lore
Share & tools
lore

Laura Abbott

author
#0 lore raw

PATCH v3 1/1 gpio: Remove VLA from stmpe driver

The new challenge is to remove VLAs from the kernel (see https://lkml.org/lkml/2018/3/7/621)

The number of GPIOs on the supported chips is fairly small so stack allocate to a known upper bound and spit out a warning if any new chips have more gpios.

Signed-off-by Laura Abbott <[email protected]>

---
v3: Split this off from the rest of the series since some of the patches had been picked up. Switched to just hardcoding an upper bound for the stack array since it's only a few extra bytes of stack space.

---
 drivers/gpio/gpio-stmpe.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/gpio/gpio-stmpe.c +8 -1
index f8d7d1cd8488..8d6a5a7e612d 100644
@@ -363,13 +363,15 @@ static struct irq_chip stmpe_gpio_irq_chip = {
 	.irq_set_type		= stmpe_gpio_irq_set_type,
 };
 
+#define MAX_GPIOS 24
+
 static irqreturn_t stmpe_gpio_irq(int irq, void *dev)
 {
 	struct stmpe_gpio *stmpe_gpio = dev;
 	struct stmpe *stmpe = stmpe_gpio->stmpe;
 	u8 statmsbreg;
 	int num_banks = DIV_ROUND_UP(stmpe->num_gpios, 8);
-	u8 status[num_banks];
+	u8 status[DIV_ROUND_UP(MAX_GPIOS, 8)];
 	int ret;
 	int i;
 

@@ -434,6 +436,11 @@ static int stmpe_gpio_probe(struct platform_device *pdev)
 	struct stmpe_gpio *stmpe_gpio;
 	int ret, irq;
 
+	if (stmpe->num_gpios > MAX_GPIOS) {
+		dev_err(&pdev->dev, "Need to increase maximum GPIO number\n");
+		return -EINVAL;
+	}
+
 	stmpe_gpio = kzalloc(sizeof(*stmpe_gpio), GFP_KERNEL);
 	if (!stmpe_gpio)
 		return -ENOMEM;
Signature · 1 line
2.14.3
Reviewed-by on v3 1/1
#1 lore raw
Quoted 41 lines · Laura Abbott

The new challenge is to remove VLAs from the kernel (see https://lkml.org/lkml/2018/3/7/621)

The number of GPIOs on the supported chips is fairly small so stack allocate to a known upper bound and spit out a warning if any new chips have more gpios.

Signed-off-by Laura Abbott <[email protected]>

---
v3: Split this off from the rest of the series since some of the patches had been picked up. Switched to just hardcoding an upper bound for the stack array since it's only a few extra bytes of stack space.

---
  drivers/gpio/gpio-stmpe.c | 9 ++++++++-
  1 file changed, 8 insertions(+), 1 deletion(-)
drivers/gpio/gpio-stmpe.c +3 -1
index f8d7d1cd8488..8d6a5a7e612d 100644
@@ -363,13 +363,15 @@ static struct irq_chip stmpe_gpio_irq_chip = {
  	.irq_set_type		= stmpe_gpio_irq_set_type,
  };
 
+#define MAX_GPIOS 24
+
  static irqreturn_t stmpe_gpio_irq(int irq, void *dev)
  {
  	struct stmpe_gpio *stmpe_gpio = dev;
  	struct stmpe *stmpe = stmpe_gpio->stmpe;
  	u8 statmsbreg;
  	int num_banks = DIV_ROUND_UP(stmpe->num_gpios, 8);
-	u8 status[num_banks];
+	u8 status[DIV_ROUND_UP(MAX_GPIOS, 8)];
  	int ret;
  	int i;
 

@@ -434,6 +436,11 @@ static int stmpe_gpio_probe(struct platform_device *pdev)
  	struct stmpe_gpio *stmpe_gpio;

drivers/gpio/gpio-stmpe.c static int stmpe_gpio_probe(struct platform_device *pdev) · line 446

Quoted 10 lines
drivers/gpio/gpio-stmpe.c +5 -0
  	int ret, irq;
 
+	if (stmpe->num_gpios > MAX_GPIOS) {
+		dev_err(&pdev->dev, "Need to increase maximum GPIO number\n");
+		return -EINVAL;
+	}
+
  	stmpe_gpio = kzalloc(sizeof(*stmpe_gpio), GFP_KERNEL);
  	if (!stmpe_gpio)
  		return -ENOMEM;

FWIW

Reviewed-by Phil Reid <[email protected]>
Signature · 2 lines
Regards
Phil Reid
Reviewed-by on v3 1/1
#2 lore raw
In reply to Laura Abbott
Quoted 8 lines · Laura Abbott

The new challenge is to remove VLAs from the kernel (see https://lkml.org/lkml/2018/3/7/621)

The number of GPIOs on the supported chips is fairly small so stack allocate to a known upper bound and spit out a warning if any new chips have more gpios.

Signed-off-by Laura Abbott <[email protected]>
Reviewed-by Kees Cook <[email protected]>

Linus, I think this patch is still needed and got missed? Can you take it as well?

Thanks!

-Kees

Quoted 45 lines

---
v3: Split this off from the rest of the series since some of the patches had been picked up. Switched to just hardcoding an upper bound for the stack array since it's only a few extra bytes of stack space.

---
 drivers/gpio/gpio-stmpe.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
drivers/gpio/gpio-stmpe.c +8 -2
index f8d7d1cd8488..8d6a5a7e612d 100644
@@ -363,13 +363,15 @@ static struct irq_chip stmpe_gpio_irq_chip = {
        .irq_set_type           = stmpe_gpio_irq_set_type,
 };
 
+#define MAX_GPIOS 24
+
 static irqreturn_t stmpe_gpio_irq(int irq, void *dev)
 {
        struct stmpe_gpio *stmpe_gpio = dev;
        struct stmpe *stmpe = stmpe_gpio->stmpe;
        u8 statmsbreg;
        int num_banks = DIV_ROUND_UP(stmpe->num_gpios, 8);
-       u8 status[num_banks];
+       u8 status[DIV_ROUND_UP(MAX_GPIOS, 8)];
        int ret;
        int i;
 

@@ -434,6 +436,11 @@ static int stmpe_gpio_probe(struct platform_device *pdev)
        struct stmpe_gpio *stmpe_gpio;
        int ret, irq;
 
+       if (stmpe->num_gpios > MAX_GPIOS) {
+               dev_err(&pdev->dev, "Need to increase maximum GPIO number\n");
+               return -EINVAL;
+       }
+
        stmpe_gpio = kzalloc(sizeof(*stmpe_gpio), GFP_KERNEL);
        if (!stmpe_gpio)
                return -ENOMEM;
--

2.14.3

Signature · 2 lines
Kees Cook
Pixel Security