[patch] [media] Input: synaptics-rmi4 - unlock on error

We should unlock before returning on this error path.

Fixes: 3a762dbd5347 ('[media] Input: synaptics-rmi4 - add support for F54 diagnostics')
Signed-off-by: Dan Carpenter <redacted>

diff --git a/drivers/input/rmi4/rmi_f54.c b/drivers/input/rmi4/rmi_f54.c
index cf805b9..2e934ae 100644
--- a/drivers/input/rmi4/rmi_f54.c
+++ b/drivers/input/rmi4/rmi_f54.c

@@ -200,7 +200,7 @@ static int rmi_f54_request_report(struct rmi_function *fn, u8 report_type)
 
 	error = rmi_write(rmi_dev, fn->fd.command_base_addr, F54_GET_REPORT);
 	if (error < 0)
-		return error;
+		goto unlock;
 
 	init_completion(&f54->cmd_done);
 

@@ -209,9 +209,10 @@ static int rmi_f54_request_report(struct rmi_function *fn, u8 report_type)
 
 	queue_delayed_work(f54->workqueue, &f54->work, 0);
 
+unlock:
 	mutex_unlock(&f54->data_mutex);
 
-	return 0;
+	return error;
 }
 
 static size_t rmi_f54_get_report_size(struct f54_data *f54)

On 10/14/2016 07:14 PM, Dan Carpenter wrote:

We should unlock before returning on this error path.

Fixes: 3a762dbd5347 ('[media] Input: synaptics-rmi4 - add support for F54 diagnostics')
Signed-off-by: Dan Carpenter <redacted>

Dmitry, will you pick this up? It probably should to in for 4.9.

Even though it was prefixed with [media] it wasn't send to linux-media, so it never
ended up in our patchwork.

If you don't want to process this patch, then let me know and I'll take it.

Regards,

	Hans

On Mon, Nov 14, 2016 at 01:24:12PM +0100, Hans Verkuil wrote:

On 10/14/2016 07:14 PM, Dan Carpenter wrote:

quoted

We should unlock before returning on this error path.

Fixes: 3a762dbd5347 ('[media] Input: synaptics-rmi4 - add support for F54 diagnostics')
Signed-off-by: Dan Carpenter <redacted>

Dmitry, will you pick this up? It probably should to in for 4.9.

Even though it was prefixed with [media] it wasn't send to linux-media, so it never
ended up in our patchwork.

If you don't want to process this patch, then let me know and I'll take it.

I picked it up. Given it is error handling that should not really fail I
think it can wait for next merge.

Thanks.

Regards,

	Hans

quoted

diff --git a/drivers/input/rmi4/rmi_f54.c b/drivers/input/rmi4/rmi_f54.c
index cf805b9..2e934ae 100644
--- a/drivers/input/rmi4/rmi_f54.c
+++ b/drivers/input/rmi4/rmi_f54.c

@@ -200,7 +200,7 @@ static int rmi_f54_request_report(struct rmi_function *fn, u8 report_type)

	error = rmi_write(rmi_dev, fn->fd.command_base_addr, F54_GET_REPORT);
	if (error < 0)
-		return error;
+		goto unlock;

	init_completion(&f54->cmd_done);

@@ -209,9 +209,10 @@ static int rmi_f54_request_report(struct rmi_function *fn, u8 report_type)

	queue_delayed_work(f54->workqueue, &f54->work, 0);

+unlock:
	mutex_unlock(&f54->data_mutex);

-	return 0;
+	return error;
}

static size_t rmi_f54_get_report_size(struct f54_data *f54)

-- 
Dmitry