[PATCH 1/1] usb:gadget:function:fix memory leak

[PATCH 1/1] usb:gadget:function:fix memory leak

[Xidong Wang]

In function f_audio_set_alt(), the memory allocated by
usb_ep_alloc_request() is not released on the error path
that req->buf, which holds the return value of kzalloc(),
is NULL. This will result in a memory leak bug.

Signed-off-by: Xidong Wang <wangxidong_97@163.com>
---
 drivers/usb/gadget/function/f_uac1_legacy.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
index 24c086b..2fcdade 100644
--- a/drivers/usb/gadget/function/f_uac1_legacy.c
+++ b/drivers/usb/gadget/function/f_uac1_legacy.c
@@ -630,8 +630,11 @@ static int f_audio_set_alt(struct usb_function *f, unsigned intf, unsigned alt)
 							ERROR(cdev,
 							"%s queue req: %d\n",
 							out_ep->name, err);
-					} else
+					} else {
+						usb_ep_free_request(
+							out_ep, req);
 						err = -ENOMEM;
+					}
 				} else
 					err = -ENOMEM;
 			}
-- 
2.7.4

Re: [PATCH 1/1] usb:gadget:function:fix memory leak

[Felipe Balbi]

[-- Attachment #1: Type: text/plain, Size: 1320 bytes --]


hi,

Xidong Wang <wangxidong_97@163.com> writes:
> In function f_audio_set_alt(), the memory allocated by
> usb_ep_alloc_request() is not released on the error path
> that req->buf, which holds the return value of kzalloc(),
> is NULL. This will result in a memory leak bug.
>
> Signed-off-by: Xidong Wang <wangxidong_97@163.com>
> ---
>  drivers/usb/gadget/function/f_uac1_legacy.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> index 24c086b..2fcdade 100644
> --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> @@ -630,8 +630,11 @@ static int f_audio_set_alt(struct usb_function *f, unsigned intf, unsigned alt)
>  							ERROR(cdev,
>  							"%s queue req: %d\n",
>  							out_ep->name, err);
> -					} else
> +					} else {
> +						usb_ep_free_request(
> +							out_ep, req);
>  						err = -ENOMEM;
> +					}

I feel like this hunk has been ping ponging between having
usb_ep_free_request() and not having it because completion callback will
call usb_ep_free_request() or something along those lines.

Can we get a final solution that solves all cases and doesn't introduce
other bugs?

-- 
balbi

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]