it's definitely not always that we will have all 16 interrupts fired at the same time, so in order to avoid looping so many times, we are now using ffs() which is implemented (on ARM) using the far better clz instruction.
This will save us quite some loops and could improve IRQ latency on Retu significantly (no actual measurements were made, though) Signed-off-by: Felipe Balbi <[email protected]> --- drivers/cbus/retu.c | 14 ++++++++------ 1 files changed, 8 insertions(+), 6 deletions(-) diff --git a/drivers/cbus/retu.c b/drivers/cbus/retu.c index b7fbd18..e749c0e 100644 --- a/drivers/cbus/retu.c +++ b/drivers/cbus/retu.c @@ -36,6 +36,8 @@ #include <linux/platform_device.h> #include <linux/platform_data/cbus.h> +#include <asm/bitops.h> + #include "cbus.h" #include "retu.h" @@ -183,8 +185,6 @@ static irqreturn_t retu_irq_handler(int irq, void *_retu) { struct retu *retu = _retu; - int i; - u16 idr; u16 imr; @@ -199,11 +199,13 @@ static irqreturn_t retu_irq_handler(int irq, void *_retu) return IRQ_NONE; } - for (i = retu->irq_base; idr != 0; i++, idr >>= 1) { - if (!(idr & 1)) - continue; + while (idr) { + unsigned long pending = __ffs(idr); + unsigned int irq; - handle_nested_irq(i); + idr &= ~BIT(pending); + irq = pending + retu->irq_base; + handle_nested_irq(irq); } return IRQ_HANDLED; -- 1.7.6.396.ge0613 -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
