Re: [PATCH v2 1/4] drm/bridge: Add fwnode based helpers to get the next bridge

[Dmitry Baryshkov]

On Thu, 7 Mar 2024 at 21:20, Sui Jingfeng <sui.jingfeng@xxxxxxxxx> wrote:
>
> Hi,
>
>
> On 2024/3/8 02:43, Dmitry Baryshkov wrote:
> >> +
> >>   MODULE_AUTHOR("Ajay Kumar<ajaykumar.rs@xxxxxxxxxxx>");
> >>   MODULE_DESCRIPTION("DRM bridge infrastructure");
> >>   MODULE_LICENSE("GPL and additional rights");
> >> diff --git a/include/drm/drm_bridge.h b/include/drm/drm_bridge.h
> >> index 3606e1a7f965..d4c95afdd662 100644
> >> --- a/include/drm/drm_bridge.h
> >> +++ b/include/drm/drm_bridge.h
> >> @@ -26,6 +26,7 @@
> >>   #include <linux/ctype.h>
> >>   #include <linux/list.h>
> >>   #include <linux/mutex.h>
> >> +#include <linux/of.h>
> >>
> >>   #include <drm/drm_atomic.h>
> >>   #include <drm/drm_encoder.h>
> >> @@ -721,6 +722,8 @@ struct drm_bridge {
> >>          struct list_head chain_node;
> >>          /** @of_node: device node pointer to the bridge */
> >>          struct device_node *of_node;
> > In my opinion, if you are adding fwnode, we can drop of_node
> > completely. There is no need to keep both of them.
>
>
> But the 'struct device' have both of them contained, we should *follow the core*, right?
> They are two major firmware subsystems (DT and ACPT), both are great and large, right?
> Personally, I think the drm_bridge should embeds 'struct device', after all, drm bridge
> are mainly stand for display bridges device. And also to reflect what you said: "to
> reflect the hardware perfectly" and remove some boilerplate.

struct device contains both because it is at the root of the hierarchy
and it should support both API. drm_bridge is a consumer, so it's fine
to have just one.

>
> I think I'm not good enough to do such a big refactor, sorry. I believe that Maxime
> and Laurent are the advanced programmers who is good enough to do such things, maybe
> you can ask them for help?

Well, you picked up the task ;-)

But really, there is nothing so hard about it:
- Change of_node to fw_node, apply an automatic patch changing this in
bridge drivers.
- Make drm_of_bridge functions convert passed of_node and comp

After this we can start cleaning up bridge drivers to use fw_node API
natively as you did in your patches 2-4.

>
> Beside this, other reviews are acceptable and will be fixed at the next version,
> thanks a lot for review.
>
>
> Best Regards,
> Sui
>


-- 
With best wishes
Dmitry