[PATCH] mfd: twl6030-irq: Make sure to call request_irq before using the irq_num

[PATCH] mfd: twl6030-irq: Make sure to call request_irq before using the irq_num

[Axel Lin]

I was trying to fix the error handling part because in the case of
request_irq failure, it should call kthread_stop instead of free_irq.
But it seems more reasonable to do request_irq before calling kthread_run.

Signed-off-by: Axel Lin <axel.lin@gmail.com>
---
 drivers/mfd/twl6030-irq.c |   18 ++++++++++--------
 1 files changed, 10 insertions(+), 8 deletions(-)

diff --git a/drivers/mfd/twl6030-irq.c b/drivers/mfd/twl6030-irq.c
index b57383b..e0baff8 100644
--- a/drivers/mfd/twl6030-irq.c
+++ b/drivers/mfd/twl6030-irq.c
@@ -332,12 +332,6 @@ int twl6030_init_irq(int irq_num, unsigned irq_base, unsigned irq_end)
 
 	/* install an irq handler to demultiplex the TWL6030 interrupt */
 	init_completion(&irq_event);
-	task = kthread_run(twl6030_irq_thread, (void *)irq_num, "twl6030-irq");
-	if (IS_ERR(task)) {
-		pr_err("twl6030: could not create irq %d thread!\n", irq_num);
-		status = PTR_ERR(task);
-		goto fail_kthread;
-	}
 
 	status = request_irq(irq_num, handle_twl6030_pih, IRQF_DISABLED,
 				"TWL6030-PIH", &irq_event);
@@ -345,11 +339,19 @@ int twl6030_init_irq(int irq_num, unsigned irq_base, unsigned irq_end)
 		pr_err("twl6030: could not claim irq%d: %d\n", irq_num, status);
 		goto fail_irq;
 	}
+
+	task = kthread_run(twl6030_irq_thread, (void *)irq_num, "twl6030-irq");
+	if (IS_ERR(task)) {
+		pr_err("twl6030: could not create irq %d thread!\n", irq_num);
+		status = PTR_ERR(task);
+		goto fail_kthread;
+	}
 	return status;
-fail_irq:
-	free_irq(irq_num, &irq_event);
 
 fail_kthread:
+	free_irq(irq_num, &irq_event);
+
+fail_irq:
 	for (i = irq_base; i < irq_end; i++)
 		irq_set_chip_and_handler(i, NULL, NULL);
 	return status;
-- 
1.7.4.1

Re: [PATCH] mfd: twl6030-irq: Make sure to call request_irq before using the irq_num

[Samuel Ortiz]

Hi Axel,

On Thu, Aug 11, 2011 at 03:21:00PM +0800, Axel Lin wrote:
> I was trying to fix the error handling part because in the case of
> request_irq failure, it should call kthread_stop instead of free_irq.
> But it seems more reasonable to do request_irq before calling kthread_run.
I don't have the hardware, but that makes sense to me.
Unless some from TI NACK this patch, I'm applying it.

Cheers,
Samuel.

-- 
Intel Open Source Technology Centre
http://oss.intel.com/