2021-11-03 07:50:56

by Jiasheng Jiang

Subject: [PATCH] drm: mxsfb: Check NULL pointer

As we see in the drm_connector_list_iter_next(), it could return
NULL. In order to avoid the use of the NULL pointer, it may be
better to check the return value.

Fixes: c42001e ("drm: mxsfb: Use drm_panel_bridge")
Signed-off-by: Jiasheng Jiang <[email protected]>
---
drivers/gpu/drm/mxsfb/mxsfb_drv.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/mxsfb/mxsfb_drv.c b/drivers/gpu/drm/mxsfb/mxsfb_drv.c
index 6da9355..b875c11 100644
--- a/drivers/gpu/drm/mxsfb/mxsfb_drv.c
+++ b/drivers/gpu/drm/mxsfb/mxsfb_drv.c
@@ -145,6 +145,8 @@ static int mxsfb_attach_bridge(struct mxsfb_drm_private *mxsfb)
*/
drm_connector_list_iter_begin(drm, &iter);
mxsfb->connector = drm_connector_list_iter_next(&iter);
+ if (!mxsfb->connector)
+ return 1;
drm_connector_list_iter_end(&iter);

return 0;
--
2.7.4

2021-11-03 07:59:27

by Marek Vasut

Subject: Re: [PATCH] drm: mxsfb: Check NULL pointer

On 11/3/21 8:48 AM, Jiasheng Jiang wrote:
> As we see in the drm_connector_list_iter_next(), it could return
> NULL. In order to avoid the use of the NULL pointer, it may be
> better to check the return value.
>
> Fixes: c42001e ("drm: mxsfb: Use drm_panel_bridge")
> Signed-off-by: Jiasheng Jiang <[email protected]>
> ---
> drivers/gpu/drm/mxsfb/mxsfb_drv.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/gpu/drm/mxsfb/mxsfb_drv.c b/drivers/gpu/drm/mxsfb/mxsfb_drv.c
> index 6da9355..b875c11 100644
> --- a/drivers/gpu/drm/mxsfb/mxsfb_drv.c
> +++ b/drivers/gpu/drm/mxsfb/mxsfb_drv.c
> @@ -145,6 +145,8 @@ static int mxsfb_attach_bridge(struct mxsfb_drm_private *mxsfb)
> */
> drm_connector_list_iter_begin(drm, &iter);
> mxsfb->connector = drm_connector_list_iter_next(&iter);
> + if (!mxsfb->connector)
> + return 1;

In which case does this happen failure happen ?
What is the test case ?

2021-11-04 02:43:56

by Jiasheng Jiang

Subject: Re: Re: [PATCH] drm: mxsfb: Check NULL pointer

On 11/3/21 8:58 AM, Marek Vasut wrote:
>> As we see in the drm_connector_list_iter_next(), it could return
>> NULL. In order to avoid the use of the NULL pointer, it may be
>> better to check the return value.
>>
>> Fixes: c42001e ("drm: mxsfb: Use drm_panel_bridge")
>> Signed-off-by: Jiasheng Jiang <[email protected]>
>> ---
>> drivers/gpu/drm/mxsfb/mxsfb_drv.c | 2 ++
>> 1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/mxsfb/mxsfb_drv.c b/drivers/gpu/drm/mxsfb/mxsfb_drv.c
>> index 6da9355..b875c11 100644
>> --- a/drivers/gpu/drm/mxsfb/mxsfb_drv.c
>> +++ b/drivers/gpu/drm/mxsfb/mxsfb_drv.c
>> @@ -145,6 +145,8 @@ static int mxsfb_attach_bridge(struct mxsfb_drm_private *mxsfb)
>> */
>> drm_connector_list_iter_begin(drm, &iter);
>> mxsfb->connector = drm_connector_list_iter_next(&iter);
>> + if (!mxsfb->connector)
>> + return 1;

> In which case does this happen failure happen ?
> What is the test case ?

As fas as I am concerned, 'connector_list' in the
drm_connector_list_iter_next() is initialized in the
drm_mode_config_init(), which is 'connector_list->next = connector_list'.
And therefore, the check in drm_connector_list_iter_next() that
'if (lhead->next == &config->connector_list)' is directly satisfied and
returns NULL.
I am not sure wheter it is designed on purpose. If not, please fix it.