Re: [PATCH v12 1/3] dt-bindings: display: mediatek: Add OF graph support for board path

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Il 16/10/24 16:00, Rob Herring ha scritto:
On Wed, Oct 16, 2024 at 4:23 AM AngeloGioacchino Del Regno
<angelogioacchino.delregno@xxxxxxxxxxxxx> wrote:

Il 15/10/24 15:48, Rob Herring ha scritto:
On Tue, Oct 15, 2024 at 10:32:22AM +0200, AngeloGioacchino Del Regno wrote:
Il 14/10/24 19:36, Rob Herring ha scritto:
On Mon, Oct 14, 2024 at 3:51 AM AngeloGioacchino Del Regno
<angelogioacchino.delregno@xxxxxxxxxxxxx> wrote:

The display IPs in MediaTek SoCs support being interconnected with
different instances of DDP IPs (for example, merge0 or merge1) and/or
with different DDP IPs (for example, rdma can be connected with either
color, dpi, dsi, merge, etc), forming a full Display Data Path that
ends with an actual display.

The final display pipeline is effectively board specific, as it does
depend on the display that is attached to it, and eventually on the
sensors supported by the board (for example, Adaptive Ambient Light
would need an Ambient Light Sensor, otherwise it's pointless!), other
than the output type.

Add support for OF graphs to most of the MediaTek DDP (display) bindings
to add flexibility to build custom hardware paths, hence enabling board
specific configuration of the display pipeline and allowing to finally
migrate away from using hardcoded paths.

Reviewed-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
Reviewed-by: Alexandre Mergnat <amergnat@xxxxxxxxxxxx>
Tested-by: Alexandre Mergnat <amergnat@xxxxxxxxxxxx>
Reviewed-by: CK Hu <ck.hu@xxxxxxxxxxxx>
Tested-by: Michael Walle <mwalle@xxxxxxxxxx> # on kontron-sbc-i1200
Signed-off-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@xxxxxxxxxxxxx>
---
    .../display/mediatek/mediatek,aal.yaml        | 40 +++++++++++++++++++
    .../display/mediatek/mediatek,ccorr.yaml      | 21 ++++++++++
    .../display/mediatek/mediatek,color.yaml      | 22 ++++++++++
    .../display/mediatek/mediatek,dither.yaml     | 22 ++++++++++
    .../display/mediatek/mediatek,dpi.yaml        | 25 +++++++++++-
    .../display/mediatek/mediatek,dsc.yaml        | 24 +++++++++++
    .../display/mediatek/mediatek,dsi.yaml        | 27 ++++++++++++-
    .../display/mediatek/mediatek,ethdr.yaml      | 22 ++++++++++
    .../display/mediatek/mediatek,gamma.yaml      | 19 +++++++++
    .../display/mediatek/mediatek,merge.yaml      | 23 +++++++++++
    .../display/mediatek/mediatek,od.yaml         | 22 ++++++++++
    .../display/mediatek/mediatek,ovl-2l.yaml     | 22 ++++++++++
    .../display/mediatek/mediatek,ovl.yaml        | 22 ++++++++++
    .../display/mediatek/mediatek,postmask.yaml   | 21 ++++++++++
    .../display/mediatek/mediatek,rdma.yaml       | 22 ++++++++++
    .../display/mediatek/mediatek,ufoe.yaml       | 21 ++++++++++
    16 files changed, 372 insertions(+), 3 deletions(-)

diff --git a/Documentation/devicetree/bindings/display/mediatek/mediatek,aal.yaml b/Documentation/devicetree/bindings/display/mediatek/mediatek,aal.yaml
index cf24434854ff..47ddba5c41af 100644
--- a/Documentation/devicetree/bindings/display/mediatek/mediatek,aal.yaml
+++ b/Documentation/devicetree/bindings/display/mediatek/mediatek,aal.yaml
@@ -62,6 +62,27 @@ properties:
        $ref: /schemas/types.yaml#/definitions/phandle-array
        maxItems: 1

+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+    description:
+      Input and output ports can have multiple endpoints, each of those
+      connects to either the primary, secondary, etc, display pipeline.
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: AAL input port
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description:
+          AAL output to the next component's input, for example could be one
+          of many gamma, overdrive or other blocks.
+
+    required:
+      - port@0
+      - port@1
+
    required:
      - compatible
      - reg
@@ -89,5 +110,24 @@ examples:
               power-domains = <&scpsys MT8173_POWER_DOMAIN_MM>;
               clocks = <&mmsys CLK_MM_DISP_AAL>;
               mediatek,gce-client-reg = <&gce SUBSYS_1401XXXX 0x5000 0x1000>;
+
+           ports {
+               #address-cells = <1>;
+               #size-cells = <0>;
+
+               port@0 {
+                   reg = <0>;
+                   aal0_in: endpoint {
+                       remote-endpoint = <&ccorr0_out>;
+                   };
+               };
+
+               port@1 {
+                   reg = <1>;
+                   aal0_out: endpoint {
+                       remote-endpoint = <&gamma0_in>;
+                   };
+               };
+           };
           };
        };
diff --git a/Documentation/devicetree/bindings/display/mediatek/mediatek,ccorr.yaml b/Documentation/devicetree/bindings/display/mediatek/mediatek,ccorr.yaml
index 9f8366763831..fca8e7bb0cbc 100644
--- a/Documentation/devicetree/bindings/display/mediatek/mediatek,ccorr.yaml
+++ b/Documentation/devicetree/bindings/display/mediatek/mediatek,ccorr.yaml
@@ -57,6 +57,27 @@ properties:
        $ref: /schemas/types.yaml#/definitions/phandle-array
        maxItems: 1

+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+    description:
+      Input and output ports can have multiple endpoints, each of those
+      connects to either the primary, secondary, etc, display pipeline.
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: CCORR input port
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description:
+          CCORR output to the input of the next desired component in the
+          display pipeline, usually only one of the available AAL blocks.
+
+    required:
+      - port@0
+      - port@1
+
    required:
      - compatible
      - reg
diff --git a/Documentation/devicetree/bindings/display/mediatek/mediatek,color.yaml b/Documentation/devicetree/bindings/display/mediatek/mediatek,color.yaml
index 7df786bbad20..6160439ce4d7 100644
--- a/Documentation/devicetree/bindings/display/mediatek/mediatek,color.yaml
+++ b/Documentation/devicetree/bindings/display/mediatek/mediatek,color.yaml
@@ -65,6 +65,28 @@ properties:
        $ref: /schemas/types.yaml#/definitions/phandle-array
        maxItems: 1

+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+    description:
+      Input and output ports can have multiple endpoints, each of those
+      connects to either the primary, secondary, etc, display pipeline.
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: COLOR input port
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description:
+          COLOR output to the input of the next desired component in the
+          display pipeline, for example one of the available CCORR or AAL
+          blocks.
+
+    required:
+      - port@0
+      - port@1
+
    required:
      - compatible
      - reg
diff --git a/Documentation/devicetree/bindings/display/mediatek/mediatek,dither.yaml b/Documentation/devicetree/bindings/display/mediatek/mediatek,dither.yaml
index 6fceb1f95d2a..abaf27916d13 100644
--- a/Documentation/devicetree/bindings/display/mediatek/mediatek,dither.yaml
+++ b/Documentation/devicetree/bindings/display/mediatek/mediatek,dither.yaml
@@ -56,6 +56,28 @@ properties:
        $ref: /schemas/types.yaml#/definitions/phandle-array
        maxItems: 1

+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+    description:
+      Input and output ports can have multiple endpoints, each of those
+      connects to either the primary, secondary, etc, display pipeline.
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: DITHER input, usually from a POSTMASK or GAMMA block.
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description:
+          DITHER output to the input of the next desired component in the
+          display pipeline, for example one of the available DSC compressors,
+          DP_INTF, DSI, LVDS or others.
+
+    required:
+      - port@0
+      - port@1
+
    required:
      - compatible
      - reg
diff --git a/Documentation/devicetree/bindings/display/mediatek/mediatek,dpi.yaml b/Documentation/devicetree/bindings/display/mediatek/mediatek,dpi.yaml
index 3a82aec9021c..b567e3d58aa1 100644
--- a/Documentation/devicetree/bindings/display/mediatek/mediatek,dpi.yaml
+++ b/Documentation/devicetree/bindings/display/mediatek/mediatek,dpi.yaml
@@ -71,13 +71,34 @@ properties:
          Output port node. This port should be connected to the input port of an
          attached HDMI, LVDS or DisplayPort encoder chip.

+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: DPI input port
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: DPI output to an HDMI, LVDS or DisplayPort encoder input

This is wrong. The existing 'port' is the output. 'port' and 'port@0'
are treated as the same thing. Since you are adding an input port, the
new port has to be 'port@1' (or any number but 0).

I haven't looked at the driver code, but it should request port 0 and
always get the output port. And requesting port 1 will return an error
or the input port.

Hello Rob,

I want to remind you that in v2 of this series you said that it'd be wrong for
port@0 to be an output, I replied that you misread that as I had modeled it indeed
as an input, and then you gave me your Reviewed-by tag.

I have not misread it. Then I guess I forgot about it and missed it the
next time on v3.


Okay, that was some misunderstanding then - it's fine, no problem.

Anyway - I get your concern about the previous behavior of `port`, but I chose to
model this that way purely for consistency.

First of all - the driver(s) will check if we're feeding a full graph, as it will
indeed first check if port@1 is present: if it is, then it follows this scheme with
port@0 as INPUT and port@1 as OUTPUT.
If the component in port@0 is an OUTPUT, the bridge attach will fail.

Getting to bindings themselves, then... it would be a mistake to model port@0 as an
output and port@1 as an input, because that would be not only inconsistent with the
DRM Bridge bindings, but would be highly confusing when reading the devicetree.

Somewhat confusing, yes. Highly, no. Put a comment in the DT.


"Somewhat or highly" boils down to personal opinion, so I still go for "highly"
but won't argue about that as wouldn't be productive and would bring us nowhere
anyway, so, whatever :-)

Putting a comment in DT is an option, yes, but that comment would need to be put
on all of the MediaTek SoC device trees - current and future - and IMO would scream
"inconsistency warning" (in different words, of course) all over, which honestly
doesn't really look clean... at least to my eyes...

What I find more confusing is my updated DT doesn't work with my
existing kernel...

Please note that the bridge bindings are always declaring port@0 as an INPUT and
other ports as OUTPUT(s).

There is no guarantee on that. Port numbering is local to the bridge and
opaque to anything outside that bridge. That is why you have to document
what each port represents.


I know and I agree that there's no guarantee on the numbering. I can see that in
other non-drm-bridge bindings, and that's fine.

Would we have followed that convention if all the ports were defined
from the start? Certainly. But that didn't happen and you are stuck with
the existing binding and ABI.


I thought about adding a new compatible for the new port scheme, but that looked
even worse to me as, after doing that (yeah, I actually went for it in the first
version that I have never sent upstream) during my own proof-read I started
screaming "HACK! HACK! NOOO!" all over, and rewritten the entire thing.

As an example, you can check display/bridge/analogix,anx7625.yaml or
display/bridge/samsung,mipi-dsim.yaml (and others) for bridges, otherwise
display/st,stm32mp25-lvds.yaml or display/allwinner,sun4i-a10-display-frontend.yaml
(and others) for display controllers, which do all conform to this logic, where
the input is always @0, and the output is @1.

Of course, doing this required me to do extra changes to the MTK DRM drivers to
actually be retro-compatible with the old devicetrees as I explained before.

You can't fix existing software...


That's true, but I don't see that as an "excuse" (grant me this term please) to
"carelessly" keep it in a suboptimal state.

This driver has been growing almost uncontrollably with (wrong, anyway!)
board-specific component vectors - and writing a new one would just add up
more code duplication to the mix and/or worsen the maintainability of older
MediaTek SoCs (as the "old" driver will get forgotten and never updated anymore).

If you want to break the ABI, then that's ultimately up to you and
Mediatek maintainers to decide0. This case is easy to avoid, why would
you knowingly break the ABI here.

Because if we don't do this, we condemn (forever) this driver, or part of it
to have an inverted port scheme compared to either:
   A. The other drm/mediatek components; or
   B. The other bridge drivers (of which, some are used with MTK as well)

...and we also condemn (forever, again) all MediaTek device trees to scream
"port inconsistency warning: A for drm/mediatek components, B for every other
thing", which would scream "drm/mediatek is somewhat broken", which can be
avoided.

Sure. The cost is just an ABI break to do that.

OTOH, this seems like a big enough
change I would imagine it is a matter of time before supporting a
missing OF graph for the components will be a problem.


Sorry I didn't understand this part, can you please-please-please reword?

I think keeping the kernel working with the old and new binding will
be a challenge. Partly because testing with the old binding won't
happen, but also if the binding and drivers continue to evolve. So
while maintaining old ABI might be possible with this change, it will
continue to be an issue with each change.

That's... exactly my point, so I am happy that we agree on that.

BTW, did you actually test
backwards compatibility with this? I can see you fallback to the old
binding, but there's a lot of other changes in there I can't really
tell by looking at it.


Short answer: Yes, largely

Long answer:

Yes, with this series applied, I have tested both *old* and *new* devicetrees on
7 boards with 4 different SoCs and different display pipelines (hence, graphs):
one smartphone (MT6795), a "bunch of" Chromebooks (MT8173, MT8186, MT8195), and
a SBC (MT8195).

Of course those had DSI or eDP displays, with or without DisplayPort external
display support.


What I haven't heard from you is "yes, we need to maintain the ABI" or
"no, we can break it". Decide that, then the path here is simple.

No, we have to break it.

Cheers,
Angelo





[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux