[PATCH] staging: r8188eu: fix memleak in rtw_wx_set_enc_ext

[PATCH] staging: r8188eu: fix memleak in rtw_wx_set_enc_ext

[Martin Kaiser]

Free the param struct if the caller sets an unsupported algorithm
and we return an error.

Fixes: 2b42bd58b321 ("staging: r8188eu: introduce new os_dep dir for RTL8188eu driver")
Signed-off-by: Martin Kaiser <martin@kaiser.cx>
---
 drivers/staging/r8188eu/os_dep/ioctl_linux.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/staging/r8188eu/os_dep/ioctl_linux.c b/drivers/staging/r8188eu/os_dep/ioctl_linux.c
index 4f0ae821d193..4e51d5a55985 100644
--- a/drivers/staging/r8188eu/os_dep/ioctl_linux.c
+++ b/drivers/staging/r8188eu/os_dep/ioctl_linux.c
@@ -1897,7 +1897,7 @@ static int rtw_wx_set_enc_ext(struct net_device *dev,
 	struct ieee_param *param = NULL;
 	struct iw_point *pencoding = &wrqu->encoding;
 	struct iw_encode_ext *pext = (struct iw_encode_ext *)extra;
-	int ret = 0;
+	int ret = -1;
 
 	param_len = sizeof(struct ieee_param) + pext->key_len;
 	param = kzalloc(param_len, GFP_KERNEL);
@@ -1923,7 +1923,7 @@ static int rtw_wx_set_enc_ext(struct net_device *dev,
 		alg_name = "CCMP";
 		break;
 	default:
-		return -1;
+		goto out;
 	}
 
 	strlcpy((char *)param->u.crypt.alg, alg_name, IEEE_CRYPT_ALG_NAME_LEN);
@@ -1950,6 +1950,7 @@ static int rtw_wx_set_enc_ext(struct net_device *dev,
 
 	ret =  wpa_set_encryption(dev, param, param_len);
 
+out:
 	kfree(param);
 	return ret;
 }
-- 
2.20.1

Re: [PATCH] staging: r8188eu: fix memleak in rtw_wx_set_enc_ext

[Dan Carpenter]

On Tue, Oct 19, 2021 at 10:23:56PM +0200, Martin Kaiser wrote:
> Free the param struct if the caller sets an unsupported algorithm
> and we return an error.
> 
> Fixes: 2b42bd58b321 ("staging: r8188eu: introduce new os_dep dir for RTL8188eu driver")
> Signed-off-by: Martin Kaiser <martin@kaiser.cx>
> ---
>  drivers/staging/r8188eu/os_dep/ioctl_linux.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/staging/r8188eu/os_dep/ioctl_linux.c b/drivers/staging/r8188eu/os_dep/ioctl_linux.c
> index 4f0ae821d193..4e51d5a55985 100644
> --- a/drivers/staging/r8188eu/os_dep/ioctl_linux.c
> +++ b/drivers/staging/r8188eu/os_dep/ioctl_linux.c
> @@ -1897,7 +1897,7 @@ static int rtw_wx_set_enc_ext(struct net_device *dev,
>  	struct ieee_param *param = NULL;
>  	struct iw_point *pencoding = &wrqu->encoding;
>  	struct iw_encode_ext *pext = (struct iw_encode_ext *)extra;
> -	int ret = 0;
> +	int ret = -1;
>  

This was in the original code, so it's fine to leave it as -1, but in
future patches could you change it to -EINVAL instead.  (Or don't.  I'm
not your boss.  I appretiate all the clean up you have done to this
driver).

regards,
dan carpenter

Re: [PATCH] staging: r8188eu: fix memleak in rtw_wx_set_enc_ext

[Martin Kaiser]

Thus wrote Dan Carpenter (dan.carpenter@oracle.com):

> On Tue, Oct 19, 2021 at 10:23:56PM +0200, Martin Kaiser wrote:
> > Free the param struct if the caller sets an unsupported algorithm
> > and we return an error.

> > Fixes: 2b42bd58b321 ("staging: r8188eu: introduce new os_dep dir for RTL8188eu driver")
> > Signed-off-by: Martin Kaiser <martin@kaiser.cx>
> > ---
> >  drivers/staging/r8188eu/os_dep/ioctl_linux.c | 5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)

> > diff --git a/drivers/staging/r8188eu/os_dep/ioctl_linux.c b/drivers/staging/r8188eu/os_dep/ioctl_linux.c
> > index 4f0ae821d193..4e51d5a55985 100644
> > --- a/drivers/staging/r8188eu/os_dep/ioctl_linux.c
> > +++ b/drivers/staging/r8188eu/os_dep/ioctl_linux.c
> > @@ -1897,7 +1897,7 @@ static int rtw_wx_set_enc_ext(struct net_device *dev,
> >  	struct ieee_param *param = NULL;
> >  	struct iw_point *pencoding = &wrqu->encoding;
> >  	struct iw_encode_ext *pext = (struct iw_encode_ext *)extra;
> > -	int ret = 0;
> > +	int ret = -1;


> This was in the original code, so it's fine to leave it as -1, but in
> future patches could you change it to -EINVAL instead.  (Or don't.  I'm
> not your boss.  I appretiate all the clean up you have done to this
> driver).

Thanks.

I tried to change as little as possible to allow for backporting to
stable.

Using a standard error code makes more sense. Some of the ioctl
functions do this already, others use -1. This is something we should
fix when cfg80211 support is in place and we know which of the ioctl
functions survive.

Best regards,
Martin