[PATCH] Bluetooth: Extra sanity check for AMP discover response

[PATCH] Bluetooth: Extra sanity check for AMP discover response

[Marcel Holtmann]

Just in case the Bluetooth standard introduces a third device type,
make sure that we only include AMP controllers in the AMP discover
response.

Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
---
 net/bluetooth/a2mp.c | 18 +++++++-----------
 1 file changed, 7 insertions(+), 11 deletions(-)

diff --git a/net/bluetooth/a2mp.c b/net/bluetooth/a2mp.c
index 529789d..5278a65 100644
--- a/net/bluetooth/a2mp.c
+++ b/net/bluetooth/a2mp.c
@@ -78,24 +78,20 @@ u8 __next_ident(struct amp_mgr *mgr)
 /* hci_dev_list shall be locked */
 static void __a2mp_add_cl(struct amp_mgr *mgr, struct a2mp_cl *cl)
 {
-	int i = 0;
 	struct hci_dev *hdev;
+	int i = 1;
 
 	cl[0].id = AMP_ID_BREDR;
 	cl[0].type = AMP_TYPE_BREDR;
 	cl[0].status = AMP_STATUS_BLUETOOTH_ONLY;
 
 	list_for_each_entry(hdev, &hci_dev_list, list) {
-		/* Iterate through AMP controllers */
-		if (hdev->dev_type != HCI_AMP)
-			continue;
-
-		/* Starting from second entry */
-		++i;
-
-		cl[i].id = hdev->id;
-		cl[i].type = hdev->amp_type;
-		cl[i].status = hdev->amp_status;
+		if (hdev->dev_type == HCI_AMP) {
+			cl[i].id = hdev->id;
+			cl[i].type = hdev->amp_type;
+			cl[i].status = hdev->amp_status;
+			i++;
+		}
 	}
 }
 
-- 
1.8.3.1

Re: [PATCH] Bluetooth: Extra sanity check for AMP discover response

[Andrei Emeltchenko]

Hi Marcel,

On Sat, Oct 05, 2013 at 11:56:16AM -0700, Marcel Holtmann wrote:
> Just in case the Bluetooth standard introduces a third device type,
> make sure that we only include AMP controllers in the AMP discover
> response.
> 
> Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
> ---
>  net/bluetooth/a2mp.c | 18 +++++++-----------
>  1 file changed, 7 insertions(+), 11 deletions(-)
> 
> diff --git a/net/bluetooth/a2mp.c b/net/bluetooth/a2mp.c
> index 529789d..5278a65 100644
> --- a/net/bluetooth/a2mp.c
> +++ b/net/bluetooth/a2mp.c
> @@ -78,24 +78,20 @@ u8 __next_ident(struct amp_mgr *mgr)
>  /* hci_dev_list shall be locked */
>  static void __a2mp_add_cl(struct amp_mgr *mgr, struct a2mp_cl *cl)
>  {
> -	int i = 0;
>  	struct hci_dev *hdev;
> +	int i = 1;
>  
>  	cl[0].id = AMP_ID_BREDR;
>  	cl[0].type = AMP_TYPE_BREDR;
>  	cl[0].status = AMP_STATUS_BLUETOOTH_ONLY;
>  
>  	list_for_each_entry(hdev, &hci_dev_list, list) {
> -		/* Iterate through AMP controllers */
> -		if (hdev->dev_type != HCI_AMP)
> -			continue;
> -
> -		/* Starting from second entry */
> -		++i;
> -
> -		cl[i].id = hdev->id;
> -		cl[i].type = hdev->amp_type;
> -		cl[i].status = hdev->amp_status;
> +		if (hdev->dev_type == HCI_AMP) {
> +			cl[i].id = hdev->id;
> +			cl[i].type = hdev->amp_type;
> +			cl[i].status = hdev->amp_status;
> +			i++;
> +		}

I wonder did it make sense to merge 3 patches which touches this small
function together?

Best regards 
Andrei Emeltchenko