| Message ID | 20220715205203.82591-1-andriy.shevchenko@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v1,1/1] irqchip/stm32-exti: Use INVALID_HWIRQ definition | expand |
On Fri, 15 Jul 2022 21:52:03 +0100, Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote: > > Use specific definition for invalid IRQ. It makes the > code uniform in respect to the constant used for that. > No functional change intended. > > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> > --- > drivers/irqchip/irq-stm32-exti.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/irqchip/irq-stm32-exti.c b/drivers/irqchip/irq-stm32-exti.c > index a73763d475f0..a6ae9f38aaf0 100644 > --- a/drivers/irqchip/irq-stm32-exti.c > +++ b/drivers/irqchip/irq-stm32-exti.c > @@ -170,7 +170,7 @@ static const struct stm32_exti_bank *stm32mp1_exti_banks[] = { > static struct irq_chip stm32_exti_h_chip; > static struct irq_chip stm32_exti_h_chip_direct; > > -#define EXTI_INVALID_IRQ U8_MAX > +#define EXTI_INVALID_IRQ ((u8)INVALID_HWIRQ) This looks like a terrible idea. It gives the impression that you can now rely on comparing the internal data structure field to INVALID_HWIRQ. Which is of course bound to fail. To be honest, I'd rather *kill* INVALID_HWIRQ, because apart from cherryview, nobody even *checks* for this value by that name. So much for the "code uniformity"... M.
On Sat, Jul 16, 2022 at 10:41:59AM +0100, Marc Zyngier wrote: > On Fri, 15 Jul 2022 21:52:03 +0100, > Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote: > > > > Use specific definition for invalid IRQ. It makes the > > code uniform in respect to the constant used for that. > > No functional change intended. ... > > -#define EXTI_INVALID_IRQ U8_MAX > > +#define EXTI_INVALID_IRQ ((u8)INVALID_HWIRQ) > > This looks like a terrible idea. It gives the impression that you can > now rely on comparing the internal data structure field to > INVALID_HWIRQ. Which is of course bound to fail. I don't know how one can go to that conclusion, but okay, I understood you. > To be honest, I'd rather *kill* INVALID_HWIRQ, because apart from > cherryview, nobody even *checks* for this value by that name. So much > for the "code uniformity"... It's used by two or three, I don't remember by heart.
diff --git a/drivers/irqchip/irq-stm32-exti.c b/drivers/irqchip/irq-stm32-exti.c index a73763d475f0..a6ae9f38aaf0 100644 --- a/drivers/irqchip/irq-stm32-exti.c +++ b/drivers/irqchip/irq-stm32-exti.c @@ -170,7 +170,7 @@ static const struct stm32_exti_bank *stm32mp1_exti_banks[] = { static struct irq_chip stm32_exti_h_chip; static struct irq_chip stm32_exti_h_chip_direct; -#define EXTI_INVALID_IRQ U8_MAX +#define EXTI_INVALID_IRQ ((u8)INVALID_HWIRQ) #define STM32MP1_DESC_IRQ_SIZE (ARRAY_SIZE(stm32mp1_exti_banks) * IRQS_PER_BANK) static const u8 stm32mp1_desc_irq[] = {
Use specific definition for invalid IRQ. It makes the code uniform in respect to the constant used for that. No functional change intended. Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> --- drivers/irqchip/irq-stm32-exti.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)