2022-03-11 20:36:48

by baihaowen

Subject: [PATCH] staging: r8188eu: simplify unnecessary NULL check

remove the optimisation of NULL checking it inline, kfree/rtw_free_netdev
will take care if that would ever be the case.

Signed-off-by: Haowen Bai <[email protected]>
---
drivers/staging/r8188eu/os_dep/usb_intf.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/drivers/staging/r8188eu/os_dep/usb_intf.c b/drivers/staging/r8188eu/os_dep/usb_intf.c
index 91792df..8d1ac48 100644
--- a/drivers/staging/r8188eu/os_dep/usb_intf.c
+++ b/drivers/staging/r8188eu/os_dep/usb_intf.c
@@ -425,10 +425,8 @@ static struct adapter *rtw_usb_if1_init(struct dvobj_priv *dvobj,
rtw_handle_dualmac(padapter, 0);
free_adapter:
if (status != _SUCCESS) {
- if (pnetdev)
- rtw_free_netdev(pnetdev);
- else if (padapter)
- vfree(padapter);
+ rtw_free_netdev(pnetdev);
+ vfree(padapter);
padapter = NULL;
}
exit:
--
2.7.4

2022-03-11 23:21:05

by Pavel Skripkin

Subject: Re: [PATCH] staging: r8188eu: simplify unnecessary NULL check

Hi Haowen,

On 3/11/22 13:33, Haowen Bai wrote:
> remove the optimisation of NULL checking it inline, kfree/rtw_free_netdev
> will take care if that would ever be the case.
>
> Signed-off-by: Haowen Bai <[email protected]>
> ---
> drivers/staging/r8188eu/os_dep/usb_intf.c | 6 ++----
> 1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/staging/r8188eu/os_dep/usb_intf.c b/drivers/staging/r8188eu/os_dep/usb_intf.c
> index 91792df..8d1ac48 100644
> --- a/drivers/staging/r8188eu/os_dep/usb_intf.c
> +++ b/drivers/staging/r8188eu/os_dep/usb_intf.c
> @@ -425,10 +425,8 @@ static struct adapter *rtw_usb_if1_init(struct dvobj_priv *dvobj,
> rtw_handle_dualmac(padapter, 0);
> free_adapter:
> if (status != _SUCCESS) {
> - if (pnetdev)
> - rtw_free_netdev(pnetdev);
> - else if (padapter)
> - vfree(padapter);
> + rtw_free_netdev(pnetdev);
> + vfree(padapter);
> padapter = NULL;
> }
> exit:

I don't see such code on staging-testing branch.

All error handling code was refactored in commit 13456b9. Please, always
base your code on top of newest branch but not upstream one



With regards,
Pavel Skripkin

2022-03-11 23:33:58

by Dan Carpenter

Subject: Re: [PATCH] staging: r8188eu: simplify unnecessary NULL check

On Fri, Mar 11, 2022 at 06:33:21PM +0800, Haowen Bai wrote:
> remove the optimisation of NULL checking it inline, kfree/rtw_free_netdev
> will take care if that would ever be the case.
>
> Signed-off-by: Haowen Bai <[email protected]>
> ---
> drivers/staging/r8188eu/os_dep/usb_intf.c | 6 ++----
> 1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/staging/r8188eu/os_dep/usb_intf.c b/drivers/staging/r8188eu/os_dep/usb_intf.c
> index 91792df..8d1ac48 100644
> --- a/drivers/staging/r8188eu/os_dep/usb_intf.c
> +++ b/drivers/staging/r8188eu/os_dep/usb_intf.c
> @@ -425,10 +425,8 @@ static struct adapter *rtw_usb_if1_init(struct dvobj_priv *dvobj,
> rtw_handle_dualmac(padapter, 0);
> free_adapter:
> if (status != _SUCCESS) {
> - if (pnetdev)
> - rtw_free_netdev(pnetdev);
> - else if (padapter)
> - vfree(padapter);
> + rtw_free_netdev(pnetdev);
> + vfree(padapter);

The rtw_free_netdev() frees padapter so this patch would introduce a
double free.

This driver is kind of garbage so I don't really fault you for making
this mistake. If the error handling and cleanup were written in the
correct way it would avoid accidentally creating bugs like this:

https://lore.kernel.org/all/20210831084735.GL12231@kadam/

The correct thing is to not call rtw_free_netdev() from rtw_usb_if1_init()
but to instead call:

free_netdev:
free_netdev(pnetdev);
handle_dualmac:
rtw_handle_dualmac(padapter, 0);
free_padapter:
vfree(padapter);

Also you need to work against the latest linux-next or staging-next
tree.

regards,
dan carpenter