2022-04-02 15:40:25

by Xiaomeng Tong

Subject: [PATCH] drm/gma500: fix a missing break in psb_driver_load

Instead of exiting the loop as expected when an entry is found, the
list_for_each_entry() continues until the traversal is complete. To
avoid potential executing 'ret = gma_backlight_init(dev);' repeatly,
break the loop when the entry is found.

Signed-off-by: Xiaomeng Tong <[email protected]>
---
drivers/gpu/drm/gma500/psb_drv.c | 3 +++
1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
index 2aff54d505e2..b61a8b0eea38 100644
--- a/drivers/gpu/drm/gma500/psb_drv.c
+++ b/drivers/gpu/drm/gma500/psb_drv.c
@@ -402,6 +402,9 @@ static int psb_driver_load(struct drm_device *dev, unsigned long flags)
ret = gma_backlight_init(dev);
break;
}
+
+ if (!ret)
+ break;
}
drm_connector_list_iter_end(&conn_iter);

--
2.17.1

2022-04-06 14:25:27

by Patrik Jakobsson

Subject: Re: [PATCH] drm/gma500: fix a missing break in psb_driver_load

On Fri, Apr 1, 2022 at 1:58 PM Xiaomeng Tong <[email protected]> wrote:
>
> Instead of exiting the loop as expected when an entry is found, the
> list_for_each_entry() continues until the traversal is complete. To
> avoid potential executing 'ret = gma_backlight_init(dev);' repeatly,
> break the loop when the entry is found.
>
> Signed-off-by: Xiaomeng Tong <[email protected]>
> ---
> drivers/gpu/drm/gma500/psb_drv.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
> index 2aff54d505e2..b61a8b0eea38 100644
> --- a/drivers/gpu/drm/gma500/psb_drv.c
> +++ b/drivers/gpu/drm/gma500/psb_drv.c
> @@ -402,6 +402,9 @@ static int psb_driver_load(struct drm_device *dev, unsigned long flags)
> ret = gma_backlight_init(dev);
> break;
> }
> +
> + if (!ret)
> + break;

Hi,
We cannot do it like this either. If the first connector isn't LVDS or
MIPI we would just break out of the iteration because ret is
presumably 0 at start and gma_backlight_init() would never run.

-Patrik

> }
> drm_connector_list_iter_end(&conn_iter);
>
> --
> 2.17.1
>

2022-04-07 07:48:09

by Xiaomeng Tong

Subject: Re: [PATCH] drm/gma500: fix a missing break in psb_driver_load

On Wed, 6 Apr 2022 09:27:19 +0200, Patrik Jakobsson wrote:
> On Fri, Apr 1, 2022 at 1:58 PM Xiaomeng Tong <[email protected]> wrote:
> >
> > Instead of exiting the loop as expected when an entry is found, the
> > list_for_each_entry() continues until the traversal is complete. To
> > avoid potential executing 'ret = gma_backlight_init(dev);' repeatly,
> > break the loop when the entry is found.
> >
> > Signed-off-by: Xiaomeng Tong <[email protected]>
> > ---
> > drivers/gpu/drm/gma500/psb_drv.c | 3 +++
> > 1 file changed, 3 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c
> > index 2aff54d505e2..b61a8b0eea38 100644
> > --- a/drivers/gpu/drm/gma500/psb_drv.c
> > +++ b/drivers/gpu/drm/gma500/psb_drv.c
> > @@ -402,6 +402,9 @@ static int psb_driver_load(struct drm_device *dev, unsigned long flags)
> > ret = gma_backlight_init(dev);
> > break;
> > }
> > +
> > + if (!ret)
> > + break;
>
> Hi,
> We cannot do it like this either. If the first connector isn't LVDS or
> MIPI we would just break out of the iteration because ret is
> presumably 0 at start and gma_backlight_init() would never run.
>
> -Patrik

Yes, you are correct. I have sent a v2 patch[1]: just goto out; when found.
Please check it, thank you very much.

[1]: https://lore.kernel.org/all/[email protected]/

--
Xiaomeng Tong

>
> > }
> > drm_connector_list_iter_end(&conn_iter);
> >
> > --
> > 2.17.1