twl4030-core irq simplification

From: David Brownell <[email protected]>

Simplify twl4030 IRQ handling by removing a needless custom flow handler. The top level IRQs, from the PIH, are well suited for handle_simple_irq() ... they can't be acked or masked.

Switching resolves some issues with how IRQs were dispatched. Notably, abuse of desc->status, IRQ accounting, and handling of various faults.

In short, use standard genirq code.

Drivers that request_irq() to the PIH will need to pay more attention to things like setting IRQF_DISABLED (since it's no longer ignored), and making I2C calls from handlers (you'll need a lockdep workaround).

Signed-off-by David Brownell <[email protected]>

---
Tested with the updated TWL GPIO and pwrirq drivers (now using handle_edge_irq). It should behave without those updates, too.

 drivers/mfd/twl4030-core.c |   75 ++-----------------------------------------
 1 file changed, 4 insertions(+), 71 deletions(-)

drivers/mfd/twl4030-core.c +2 -67

@@ -535,68 +535,6 @@ EXPORT_SYMBOL(twl4030_i2c_read_u8);
 
 /*----------------------------------------------------------------------*/
 
-/*
- * do_twl4030_module_irq() is the desc->handle method for each of the twl4030
- * module interrupts that doesn't chain to another irq_chip (GPIO, power, etc).
- * It executes in kernel thread context.  On entry, cpu interrupts are disabled.
- */
-static void do_twl4030_module_irq(unsigned int irq, irq_desc_t *desc)
-{
-	struct irqaction *action;
-	const unsigned int cpu = smp_processor_id();
-
-	/*
-	 * Earlier this was desc->triggered = 1;
-	 */
-	desc->status |= IRQ_LEVEL;
-
-	/*
-	 * The desc->handle method would normally call the desc->chip->ack
-	 * method here, but we won't bother since our ack method is NULL.
-	 */
-
-	if (!desc->depth) {
-		kstat_cpu(cpu).irqs[irq]++;
-
-		action = desc->action;
-		if (action) {
-			int ret;
-			int status = 0;
-			int retval = 0;
-
-			local_irq_enable();
-
-			do {
-				/* Call the ISR with cpu interrupts enabled */
-				ret = action->handler(irq, action->dev_id);
-				if (ret == IRQ_HANDLED)
-					status |= action->flags;
-				retval |= ret;
-				action = action->next;
-			} while (action);
-
-			if (status & IRQF_SAMPLE_RANDOM)
-				add_interrupt_randomness(irq);
-
-			local_irq_disable();
-
-			if (retval != IRQ_HANDLED)
-				printk(KERN_ERR "ISR for TWL4030 module"
-					" irq %d can't handle interrupt\n",
-					irq);
-
-			/*
-			 * Here is where we should call the unmask method, but
-			 * again we won't bother since it is NULL.
-			 */
-		} else
-			printk(KERN_CRIT "TWL4030 module irq %d has no ISR"
-					" but can't be masked!\n", irq);
-	} else
-		printk(KERN_CRIT "TWL4030 module irq %d is disabled but can't"
-				" be masked!\n", irq);
-}
-
 static unsigned twl4030_irq_base;
 
 static struct completion irq_event;

@@ -611,7 +549,6 @@ static int twl4030_irq_thread(void *data
 	static unsigned i2c_errors;
 	const static unsigned max_i2c_errors = 100;
 
-	daemonize("twl4030-irq");
 	current->flags |= PF_NOFREEZE;
 
 	while (!kthread_should_stop()) {

@@ -691,8 +625,7 @@ static struct task_struct * __init start
 	struct task_struct *thread;
 
 	init_completion(&irq_event);
-	thread = kthread_run(twl4030_irq_thread, (void *)irq,
-			     "twl4030 irq %ld", irq);
+	thread = kthread_run(twl4030_irq_thread, (void *)irq, "twl4030-irq");
 	if (!thread)
 		pr_err("%s: could not create twl4030 irq %ld thread!\n",
 		       DRIVER_NAME, irq);

@@ -1126,7 +1059,7 @@ static void twl_init_irq(int irq_num, un
 	/* install an irq handler for each of the PIH modules */
 	for (i = irq_base; i < irq_end; i++) {
 		set_irq_chip_and_handler(i, &twl4030_irq_chip,
-				do_twl4030_module_irq);
+				handle_simple_irq);
 		activate_irq(i);
 	}
 
--

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

Pushed.

Quoted 1 line · Tony Lindgren

Pushed.

Cool. Now, once the pwrirq updates get sorted out, there will be only two main steps left in my plan:

- Rework the twl4030 core SIH init. Basically, use a different IRQ setup data structure, which will be kept around for later use by ...

- New generic SIH IRQ handling code using that data, and looking much like the current GPIO stuff (in current GIT, after you merged the patch updating it) -- but sharable between at least GPIO and pwrirq, and replacing pwrirq.

I'm going to hold back on sending that stuff out until at
least next week, to let things shake out a bit more. Find
the i2c-omap issue; sort the Overo puzzle; etc.