[PATCH] Input: wm97xx - use EBUSY instead of EINVAL

[PATCH] Input: wm97xx - use EBUSY instead of EINVAL

[Joonyoung Shim]

If request_irq() fails, we have to return -EBUSY instead of -EINVAL.

Signed-off-by: Joonyoung Shim <jy0922.shim@samsung.com>
---
 drivers/input/touchscreen/wm97xx-core.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/input/touchscreen/wm97xx-core.c b/drivers/input/touchscreen/wm97xx-core.c
index 2957d48..06cc8cf 100644
--- a/drivers/input/touchscreen/wm97xx-core.c
+++ b/drivers/input/touchscreen/wm97xx-core.c
@@ -375,7 +375,7 @@ static int wm97xx_init_pen_irq(struct wm97xx *wm)
 		dev_err(wm->dev,
 			"Failed to register pen down interrupt, polling");
 		wm->pen_irq = 0;
-		return -EINVAL;
+		return -EBUSY;
 	}
 
 	/* Configure GPIO as interrupt source on WM971x */
-- 
1.6.0.4

Re: [PATCH] Input: wm97xx - use EBUSY instead of EINVAL

[Mark Brown]

On Wed, Jun 03, 2009 at 02:45:15PM +0900, Joonyoung Shim wrote:
> If request_irq() fails, we have to return -EBUSY instead of -EINVAL.

Could you provide a bit more explanation for this one, please?  -EBUSY
seems to give entirely the wrong impression to applications since in the
systems where this driver will be deployed a failure to acquire the IRQ
is most likely to be due to incorrect platform data and it is highly
unlikely to resolve itself.

Re: [PATCH] Input: wm97xx - use EBUSY instead of EINVAL

[Dmitry Torokhov]

On Wed, Jun 03, 2009 at 11:01:21AM +0100, Mark Brown wrote:
> On Wed, Jun 03, 2009 at 02:45:15PM +0900, Joonyoung Shim wrote:
> > If request_irq() fails, we have to return -EBUSY instead of -EINVAL.
> 
> Could you provide a bit more explanation for this one, please?  -EBUSY
> seems to give entirely the wrong impression to applications since in the
> systems where this driver will be deployed a failure to acquire the IRQ
> is most likely to be due to incorrect platform data and it is highly
> unlikely to resolve itself.

The "most correct" thing would be to propagate the value returned by
request_irq() I think but since nobidy actually uses result of
wm97xx_init_pen_irq() I just would not bother at all.

-- 
Dmitry

Re: [PATCH] Input: wm97xx - use EBUSY instead of EINVAL

[Mark Brown]

On Wed, Jun 03, 2009 at 07:19:14AM -0700, Dmitry Torokhov wrote:

> The "most correct" thing would be to propagate the value returned by
> request_irq() I think but since nobidy actually uses result of
> wm97xx_init_pen_irq() I just would not bother at all.

Yeah, I was looking at that - the driver can run in polling mode so it's
fine as-is.  I could change the code to return void, I suppose.

Re: [PATCH] Input: wm97xx - use EBUSY instead of EINVAL

[Dmitry Torokhov]

On Wednesday 03 June 2009 07:40:24 Mark Brown wrote:
> On Wed, Jun 03, 2009 at 07:19:14AM -0700, Dmitry Torokhov wrote:
> > The "most correct" thing would be to propagate the value returned by
> > request_irq() I think but since nobidy actually uses result of
> > wm97xx_init_pen_irq() I just would not bother at all.
>
> Yeah, I was looking at that - the driver can run in polling mode so it's
> fine as-is.  I could change the code to return void, I suppose.

Yes, I think that would be the best.

-- 
Dmitry

Re: [PATCH] Input: wm97xx - use EBUSY instead of EINVAL

[Joonyoung Shim]

On 6/3/2009 7:01 PM, Mark Brown wrote:
> On Wed, Jun 03, 2009 at 02:45:15PM +0900, Joonyoung Shim wrote:
>> If request_irq() fails, we have to return -EBUSY instead of -EINVAL.
> 
> Could you provide a bit more explanation for this one, please?  -EBUSY
> seems to give entirely the wrong impression to applications since in the
> systems where this driver will be deployed a failure to acquire the IRQ
> is most likely to be due to incorrect platform data and it is highly
> unlikely to resolve itself.

-EBUSY is used in many touchscreen drivers currently, but i agree dmitry
opinion to propagate what request_irq() returned because request_irq() can
return different errnos.

Thanks.