Re: [PATCH v2 0/8] Add display subsystem dt node on rk3576

["Heiko Stübner" <heiko@sntech.de>]

Hi Andy,

Am Sonntag, 29. Dezember 2024, 07:48:36 CET schrieb Andy Yan:
> 
> Hi Heiko,
> 
> At 2024-12-29 02:36:36, "Heiko Stübner" <heiko@sntech.de> wrote:
> >Hi Andy,
> >
> >Am Samstag, 28. Dezember 2024, 13:21:43 CET schrieb Andy Yan:
> >> From: Andy Yan <andy.yan@rock-chips.com>
> >> 
> >> 
> >> As the VOP[0] and HDMI[1] driver have already been submitted for review.
> >> This series send all display related device tree part together.
> >> [0] https://lore.kernel.org/linux-rockchip/20241219073931.3997788-1-andyshrk@163.com/T/#t
> >> [1] https://lore.kernel.org/linux-rockchip/20241225103741.364597-1-andyshrk@163.com/T/#t
> >
> >not to confuse you too much, I mainly meant that we were missing the
> >"arm64: dts: rockchip:" patches from this series.
> 
> Sorry, I misunderstood it in V2 . Before go to the next step, I would like to confirm it again:
> Does that mean I should:
> 1.  the dts patches(6/8, 7/8 8/8) should be sent as a separate series.
> 2. PATCH 5/8 should still go with the rk3576 vop driver[0]
> 3. PATCH 3/8 should still go with the rk3576 hdmi dirver[1]
> 4. PATCH 1/8, 2/8 4/8  should still be sent separate  ?

exactly. The reason behind this is to prevent confusion about who is
expected to apply a patch.

The dt-binding goes together with driver-changes into that subsystem-tree.
This normally also includes single bindings. So the "normal" expectation is
that the iommu-binding patch goes through the iommu tree, the phy-binding
patch goes through the phy tree and so on.

And of course the hdmi-binding patch goes together with the hdmi-driver
change.


Heiko


> >The dt-binding patches are (already) part of the driver changes and
> >should stay part of those series, to prevent confusion of what part goes
> >where :-) .
> >
> >
> >Heiko
> >
> >
> >> Changes in v2:
> >> - Wrap commit message according to Linux coding style
> >> - Make "rockchip,rk3588-hdptx-phy" const for "rockchip,rk3576-hdptx-phy"
> >> - Make declare phy/ropll/lcpll reset line are not exit on rk3576
> >> - describe constraint SOC by SOC for vop2 binding, as interrupts of rk3576
> >>   is very different from others
> >> - Drop Krzysztof's Reviewed-by for vopp2 bindings, as this version changed a lot.
> >> 
> >> Changes in v1:
> >> - ordered by soc name
> >> - Add description for newly added interrupt
> >> 
> >> Andy Yan (8):
> >>   dt-bindings: soc: rockchip: add rk3576 hdptxphy grf syscon
> >>   dt-bindings: phy: Add rk3576 hdptx phy
> >>   dt-bindings: display: rockchip: Add rk3576 hdmi controller
> >>   dt-bindings: iommu: rockchip: Add Rockchip RK3576
> >>   dt-bindings: display: vop2: Add rk3576 support
> >>   arm64: dts: rockchip: Add vop for rk3576
> >>   arm64: dts: rockchip: Add hdmi for rk3576
> >>   arm64: dts: rockchip: Enable hdmi display on sige5
> >> 
> >>  .../rockchip/rockchip,rk3588-dw-hdmi-qp.yaml  |   1 +
> >>  .../display/rockchip/rockchip-vop2.yaml       |  70 ++++++++--
> >>  .../bindings/iommu/rockchip,iommu.yaml        |   1 +
> >>  .../phy/rockchip,rk3588-hdptx-phy.yaml        |  62 ++++++---
> >>  .../devicetree/bindings/soc/rockchip/grf.yaml |   1 +
> >>  .../boot/dts/rockchip/rk3576-armsom-sige5.dts |  47 +++++++
> >>  arch/arm64/boot/dts/rockchip/rk3576.dtsi      | 126 ++++++++++++++++++
> >>  7 files changed, 276 insertions(+), 32 deletions(-)
> >> 
> >> 
> >
> >
> >
> >
>