[PATCH] Bluetooth: btqcomsmd: Fix channel open check

[PATCH] Bluetooth: btqcomsmd: Fix channel open check

[Bjorn Andersson]

rpmsg_create_ept() returns NULL on error and as such
qcom_wcnss_open_channel() does the same. Clarify this in the
kernel-doc and correct the error checks in btqcomsmd.

Fixes: 1511cc750c3d ("Bluetooth: Introduce Qualcomm WCNSS SMD based HCI driver")
Cc: stable@vger.kernel.org
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
---
 drivers/bluetooth/btqcomsmd.c | 8 ++++----
 drivers/soc/qcom/wcnss_ctrl.c | 2 ++
 2 files changed, 6 insertions(+), 4 deletions(-)

diff --git a/drivers/bluetooth/btqcomsmd.c b/drivers/bluetooth/btqcomsmd.c
index 2c9a5fc9137d..24f702927875 100644
--- a/drivers/bluetooth/btqcomsmd.c
+++ b/drivers/bluetooth/btqcomsmd.c
@@ -151,13 +151,13 @@ static int btqcomsmd_probe(struct platform_device *pdev)
 
 	btq->acl_channel = qcom_wcnss_open_channel(wcnss, "APPS_RIVA_BT_ACL",
 						   btqcomsmd_acl_callback, btq);
-	if (IS_ERR(btq->acl_channel))
-		return PTR_ERR(btq->acl_channel);
+	if (!btq->acl_channel)
+		return -ENOENT;
 
 	btq->cmd_channel = qcom_wcnss_open_channel(wcnss, "APPS_RIVA_BT_CMD",
 						   btqcomsmd_cmd_callback, btq);
-	if (IS_ERR(btq->cmd_channel))
-		return PTR_ERR(btq->cmd_channel);
+	if (!btq->cmd_channel)
+		return -ENOENT;
 
 	/* The local-bd-address property is usually injected by the
 	 * bootloader which has access to the allocated BD address.
diff --git a/drivers/soc/qcom/wcnss_ctrl.c b/drivers/soc/qcom/wcnss_ctrl.c
index d008e5b82db4..e968559b0fc4 100644
--- a/drivers/soc/qcom/wcnss_ctrl.c
+++ b/drivers/soc/qcom/wcnss_ctrl.c
@@ -275,6 +275,8 @@ static int wcnss_download_nv(struct wcnss_ctrl *wcnss, bool *expect_cbc)
  * @wcnss:	wcnss handle, retrieved from drvdata
  * @name:	SMD channel name
  * @cb:		callback to handle incoming data on the channel
+ *
+ * Return: rpmsg_endpoint object on success, NULL on error
  */
 struct rpmsg_endpoint *qcom_wcnss_open_channel(void *wcnss, const char *name, rpmsg_rx_cb_t cb, void *priv)
 {
-- 
2.16.2

Re: [PATCH] Bluetooth: btqcomsmd: Fix channel open check

[Marcel Holtmann]

Hi Bjorn,

> rpmsg_create_ept() returns NULL on error and as such
> qcom_wcnss_open_channel() does the same. Clarify this in the
> kernel-doc and correct the error checks in btqcomsmd.
> 
> Fixes: 1511cc750c3d ("Bluetooth: Introduce Qualcomm WCNSS SMD based HCI driver")
> Cc: stable@vger.kernel.org
> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
> ---
> drivers/bluetooth/btqcomsmd.c | 8 ++++----
> drivers/soc/qcom/wcnss_ctrl.c | 2 ++
> 2 files changed, 6 insertions(+), 4 deletions(-)

I think you better split this into two separate patches so they can go via its subsystems independently.

> 
> diff --git a/drivers/bluetooth/btqcomsmd.c b/drivers/bluetooth/btqcomsmd.c
> index 2c9a5fc9137d..24f702927875 100644
> --- a/drivers/bluetooth/btqcomsmd.c
> +++ b/drivers/bluetooth/btqcomsmd.c
> @@ -151,13 +151,13 @@ static int btqcomsmd_probe(struct platform_device *pdev)
> 
> 	btq->acl_channel = qcom_wcnss_open_channel(wcnss, "APPS_RIVA_BT_ACL",
> 						   btqcomsmd_acl_callback, btq);
> -	if (IS_ERR(btq->acl_channel))
> -		return PTR_ERR(btq->acl_channel);
> +	if (!btq->acl_channel)
> +		return -ENOENT;
> 
> 	btq->cmd_channel = qcom_wcnss_open_channel(wcnss, "APPS_RIVA_BT_CMD",
> 						   btqcomsmd_cmd_callback, btq);
> -	if (IS_ERR(btq->cmd_channel))
> -		return PTR_ERR(btq->cmd_channel);
> +	if (!btq->cmd_channel)
> +		return -ENOENT;

Any reason not to fix qcom_wcnss_open_channel to use PTR_ERR?

Regards

Marcel

Re: [PATCH] Bluetooth: btqcomsmd: Fix channel open check

[Bjorn Andersson]

On Tue 27 Feb 22:25 PST 2018, Marcel Holtmann wrote:
> > 	btq->cmd_channel = qcom_wcnss_open_channel(wcnss, "APPS_RIVA_BT_CMD",
> > 						   btqcomsmd_cmd_callback, btq);
> > -	if (IS_ERR(btq->cmd_channel))
> > -		return PTR_ERR(btq->cmd_channel);
> > +	if (!btq->cmd_channel)
> > +		return -ENOENT;
> 
> Any reason not to fix qcom_wcnss_open_channel to use PTR_ERR?
> 

Not that I can come up with, and obviously I did expect it to do so in
the past. I'll do that.

Thanks,
Bjorn