gpio-timberdale: fix a potential wrapping issue

PATCH Applied Dan Carpenter 2 participants 3 messages Oct 11, 2012 – Oct 20, 2012

Subsystems (2)

Kernel janitors GPIO

lore

Share & tools

lore

Dan Carpenter

author

2012-10-11 06:56 UTC #0 lore raw

PATCH 1/1 gpio-timberdale: fix a potential wrapping issue

->last_ier is an unsigned long but the high bits can't be used int the original code because the shift wraps.

Signed-off-by Dan Carpenter <[email protected]>

drivers/gpio/gpio-timberdale.c +2 -2

index 031c6ad..1a3e2b9 100644
@@ -116,7 +116,7 @@ static void timbgpio_irq_disable(struct irq_data *d)
 	unsigned long flags;
 
 	spin_lock_irqsave(&tgpio->lock, flags);
-	tgpio->last_ier &= ~(1 << offset);
+	tgpio->last_ier &= ~(1UL << offset);
 	iowrite32(tgpio->last_ier, tgpio->membase + TGPIO_IER);
 	spin_unlock_irqrestore(&tgpio->lock, flags);
 }

@@ -128,7 +128,7 @@ static void timbgpio_irq_enable(struct irq_data *d)
 	unsigned long flags;
 
 	spin_lock_irqsave(&tgpio->lock, flags);
-	tgpio->last_ier |= 1 << offset;
+	tgpio->last_ier |= 1UL << offset;
 	iowrite32(tgpio->last_ier, tgpio->membase + TGPIO_IER);
 	spin_unlock_irqrestore(&tgpio->lock, flags);
 }

Linus Walleij

Applied

2012-10-12 21:01 UTC #1 lore raw

Quoted 4 lines · Dan Carpenter

->last_ier is an unsigned long but the high bits can't be used int the original code because the shift wraps.
Signed-off-by Dan Carpenter <[email protected]>

Applied, thanks!
Should we send this to stable? (I assume so ...)

Yours,
Linus Walleij

Dan Carpenter

author

2012-10-20 03:46 UTC #2 lore raw

Quoted 9 lines · Linus Walleij

Quoted 4 lines · Dan Carpenter
->last_ier is an unsigned long but the high bits can't be used int the original code because the shift wraps.
Signed-off-by Dan Carpenter <[email protected]>
Applied, thanks!
Should we send this to stable? (I assume so ...)

This was a static checker fix. I guess the answer would depend on if there are over 30 IRQs.

regards,
dan carpenter