[patch 08/75] genirq: Fixup poll handling

From: Thomas Gleixner
Date: Thu Feb 10 2011 - 18:52:35 EST


try_one_irq() contains redundant code and lots of useless checks for
shared interrupts. Check for shared before setting IRQ_INPROGRESS and
then call handle_IRQ_event() while pending. Shorter version with the
same functionality.

Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
---
kernel/irq/spurious.c | 50 +++++++++++++++++++-------------------------------
1 file changed, 19 insertions(+), 31 deletions(-)

Index: linux-2.6-tip/kernel/irq/spurious.c
===================================================================
--- linux-2.6-tip.orig/kernel/irq/spurious.c
+++ linux-2.6-tip/kernel/irq/spurious.c
@@ -42,48 +42,36 @@ static int try_one_irq(int irq, struct i
raw_spin_unlock(&desc->lock);
return ok;
}
- /* Honour the normal IRQ locking */
- desc->status |= IRQ_INPROGRESS;
- action = desc->action;
- raw_spin_unlock(&desc->lock);
-
- while (action) {
- /* Only shared IRQ handlers are safe to call */
- if (action->flags & IRQF_SHARED) {
- if (action->handler(irq, action->dev_id) ==
- IRQ_HANDLED)
- ok = 1;
- }
- action = action->next;
- }
- local_irq_disable();
- /* Now clean up the flags */
- raw_spin_lock(&desc->lock);
- action = desc->action;
-
/*
- * While we were looking for a fixup someone queued a real
- * IRQ clashing with our walk:
+ * All handlers must agree on IRQF_SHARED, so we test just the
+ * first. Check for action->next as well.
*/
- while ((desc->status & IRQ_PENDING) && action) {
- /*
- * Perform real IRQ processing for the IRQ we deferred
- */
- work = 1;
+ action = desc->action;
+ if (!action || !(action->flags & IRQF_SHARED) || !action->next)
+ goto out;
+
+ /* Honour the normal IRQ locking */
+ desc->status |= IRQ_INPROGRESS;
+ do {
+ work++;
+ desc->status &= ~IRQ_PENDING;
raw_spin_unlock(&desc->lock);
- handle_IRQ_event(irq, action);
+ if (handle_IRQ_event(irq, action) != IRQ_NONE)
+ ok = 1;
raw_spin_lock(&desc->lock);
- desc->status &= ~IRQ_PENDING;
- }
+ action = desc->action;
+ } while ((desc->status & IRQ_PENDING) && action);
+
desc->status &= ~IRQ_INPROGRESS;
/*
* If we did actual work for the real IRQ line we must let the
* IRQ controller clean up too
*/
- if (work)
+ if (work > 1)
irq_end(irq, desc);
- raw_spin_unlock(&desc->lock);

+out:
+ raw_spin_unlock(&desc->lock);
return ok;
}



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/