[patch] input/tps6507x-ts: a couple work queue cleanups

[patch] input/tps6507x-ts: a couple work queue cleanups

[Dan Carpenter]

1) Use msecs_to_jiffies() instead of calculating by hand.
2) Call cancel_delayed_work_sync() instead of cancel_delayed_work() 
   followed by a separate flush_workqueue().
3) Remove the "tsc->wq = 0;"  Sparse complains about that because
   tsc->wq is a pointer, not an int.  It's not needed because we just
   free the pointer anyway.

Signed-off-by: Dan Carpenter <error27@gmail.com>

diff --git a/drivers/input/touchscreen/tps6507x-ts.c b/drivers/input/touchscreen/tps6507x-ts.c
index 5de80a1..5b70a14 100644
--- a/drivers/input/touchscreen/tps6507x-ts.c
+++ b/drivers/input/touchscreen/tps6507x-ts.c
@@ -221,7 +221,7 @@ done:
 
 	if (poll) {
 		schd = queue_delayed_work(tsc->wq, &tsc->work,
-					  tsc->poll_period * HZ / 1000);
+					  msecs_to_jiffies(tsc->poll_period));
 		if (schd)
 			tsc->polling = 1;
 		else {
@@ -326,7 +326,7 @@ static int tps6507x_ts_probe(struct platform_device *pdev)
 		goto err2;
 
 	schd = queue_delayed_work(tsc->wq, &tsc->work,
-				  tsc->poll_period * HZ / 1000);
+				  msecs_to_jiffies(tsc->poll_period));
 
 	if (schd)
 		tsc->polling = 1;
@@ -339,10 +339,8 @@ static int tps6507x_ts_probe(struct platform_device *pdev)
 	return 0;
 
 err2:
-	cancel_delayed_work(&tsc->work);
-	flush_workqueue(tsc->wq);
+	cancel_delayed_work_sync(&tsc->work);
 	destroy_workqueue(tsc->wq);
-	tsc->wq = 0;
 	input_free_device(input_dev);
 err1:
 	kfree(tsc);
@@ -360,10 +358,8 @@ static int __devexit tps6507x_ts_remove(struct platform_device *pdev)
 	if (!tsc)
 		return 0;
 
-	cancel_delayed_work(&tsc->work);
-	flush_workqueue(tsc->wq);
+	cancel_delayed_work_sync(&tsc->work);
 	destroy_workqueue(tsc->wq);
-	tsc->wq = 0;
 
 	input_free_device(input_dev);
 

Re: [patch] input/tps6507x-ts: a couple work queue cleanups

[Dmitry Torokhov]

Hi Dan,

On Tue, Jun 01, 2010 at 05:27:45PM +0200, Dan Carpenter wrote:
> 1) Use msecs_to_jiffies() instead of calculating by hand.
> 2) Call cancel_delayed_work_sync() instead of cancel_delayed_work() 
>    followed by a separate flush_workqueue().
> 3) Remove the "tsc->wq = 0;"  Sparse complains about that because
>    tsc->wq is a pointer, not an int.  It's not needed because we just
>    free the pointer anyway.
> 
> Signed-off-by: Dan Carpenter <error27@gmail.com>
> 

Applied the patch, thank you, but more cleanups needed.

> diff --git a/drivers/input/touchscreen/tps6507x-ts.c b/drivers/input/touchscreen/tps6507x-ts.c
> index 5de80a1..5b70a14 100644
> --- a/drivers/input/touchscreen/tps6507x-ts.c
> +++ b/drivers/input/touchscreen/tps6507x-ts.c
> @@ -221,7 +221,7 @@ done:
>  
>  	if (poll) {
>  		schd = queue_delayed_work(tsc->wq, &tsc->work,
> -					  tsc->poll_period * HZ / 1000);
> +					  msecs_to_jiffies(tsc->poll_period));
>  		if (schd)


I think the driver author misunderstood the return values of
queue_delayed_work().

>  			tsc->polling = 1;
>  		else {
> @@ -326,7 +326,7 @@ static int tps6507x_ts_probe(struct platform_device *pdev)
>  		goto err2;
>  
>  	schd = queue_delayed_work(tsc->wq, &tsc->work,
> -				  tsc->poll_period * HZ / 1000);
> +				  msecs_to_jiffies(tsc->poll_period));
>  
>  	if (schd)
>  		tsc->polling = 1;
> @@ -339,10 +339,8 @@ static int tps6507x_ts_probe(struct platform_device *pdev)
>  	return 0;
>  
>  err2:
> -	cancel_delayed_work(&tsc->work);
> -	flush_workqueue(tsc->wq);
> +	cancel_delayed_work_sync(&tsc->work);
>  	destroy_workqueue(tsc->wq);
> -	tsc->wq = 0;
>  	input_free_device(input_dev);
>  err1:
>  	kfree(tsc);
> @@ -360,10 +358,8 @@ static int __devexit tps6507x_ts_remove(struct platform_device *pdev)
>  	if (!tsc)
>  		return 0;
>  
> -	cancel_delayed_work(&tsc->work);
> -	flush_workqueue(tsc->wq);
> +	cancel_delayed_work_sync(&tsc->work);
>  	destroy_workqueue(tsc->wq);
> -	tsc->wq = 0;
>  
>  	input_free_device(input_dev);

This should be input_unregister_device().

I also question the production utility of a driver that continuously
poll device every 30 ms... I think that it was merged prematurely,
without adequate review ;(

-- 
Dmitry