[PATCH][next] USB: gadget: Fix return of -EBUSY

[PATCH][next] USB: gadget: Fix return of -EBUSY

[Colin Ian King]

Currently when driver->match_existing_only is true, the error return is
set to -EBUSY however ret is then set to 0 at the end of the if/else
statement. I believe the ret = 0 statement should be set in the else
part of the if statement and not at the end to ensure -EBUSY is being
returned correctly.

Detected by clang scan:
drivers/usb/gadget/udc/core.c:1558:4: warning: Value stored to 'ret' is
never read [deadcode.DeadStores]

Fixes: fc274c1e9973 ("USB: gadget: Add a new bus for gadgets")
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
---
 drivers/usb/gadget/udc/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/usb/gadget/udc/core.c b/drivers/usb/gadget/udc/core.c
index 61790592b2c8..3281d8a3dae7 100644
--- a/drivers/usb/gadget/udc/core.c
+++ b/drivers/usb/gadget/udc/core.c
@@ -1559,8 +1559,8 @@ int usb_gadget_register_driver_owner(struct usb_gadget_driver *driver,
 		} else {
 			pr_info("%s: couldn't find an available UDC\n",
 					driver->function);
+			ret = 0;
 		}
-		ret = 0;
 	}
 	mutex_unlock(&udc_lock);
 
-- 
2.35.1

Re: [PATCH][next] USB: gadget: Fix return of -EBUSY

[Alan Stern]

On Wed, May 04, 2022 at 02:58:40PM +0100, Colin Ian King wrote:
> Currently when driver->match_existing_only is true, the error return is
> set to -EBUSY however ret is then set to 0 at the end of the if/else
> statement. I believe the ret = 0 statement should be set in the else
> part of the if statement and not at the end to ensure -EBUSY is being
> returned correctly.
> 
> Detected by clang scan:
> drivers/usb/gadget/udc/core.c:1558:4: warning: Value stored to 'ret' is
> never read [deadcode.DeadStores]
> 
> Fixes: fc274c1e9973 ("USB: gadget: Add a new bus for gadgets")
> Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
> ---

Acked-by: Alan Stern <stern@rowland.harvard.edu>

Don't know how I missed that.  Thanks for fixing it.

In fact, since it's guaranteed that ret is already 0 before the 
driver->match_existing_only test, you could eliminate the assignment 
entirely.  But moving it into the second branch of the "if" statement is 
probably more clear.

Alan Stern

>  drivers/usb/gadget/udc/core.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/usb/gadget/udc/core.c b/drivers/usb/gadget/udc/core.c
> index 61790592b2c8..3281d8a3dae7 100644
> --- a/drivers/usb/gadget/udc/core.c
> +++ b/drivers/usb/gadget/udc/core.c
> @@ -1559,8 +1559,8 @@ int usb_gadget_register_driver_owner(struct usb_gadget_driver *driver,
>  		} else {
>  			pr_info("%s: couldn't find an available UDC\n",
>  					driver->function);
> +			ret = 0;
>  		}
> -		ret = 0;
>  	}
>  	mutex_unlock(&udc_lock);
>  
> -- 
> 2.35.1
>