DPUM (Display Processing Unit Main) is main dpu for Exynosautov9.
    
This patches enable clock for dpum, sysmmu(iommu).

I tested using sysmmu, DPP, DECON, DP on sdk board (modetest works).

Signed-off-by: Kwanghoon Son <k.son@samsung.com>
---
Changes in v3:
- Rename clock-names dout_clkcmu_dpum_bus to bus

Changes in v2:
- Squash patch 1,2 to make headers with schema.

---
Kwanghoon Son (3):
      dt-bindings: clock: exynosautov9: add dpum clock
      arm64: dts: exynosautov9: add dpum clock DT nodes
      clk: samsung: exynosautov9: add dpum clock support

 .../bindings/clock/samsung,exynosautov9-clock.yaml | 19 +++++
 arch/arm64/boot/dts/exynos/exynosautov9.dtsi       | 10 +++
 drivers/clk/samsung/clk-exynosautov9.c             | 83 ++++++++++++++++++++++
 include/dt-bindings/clock/samsung,exynosautov9.h   | 11 +++
 4 files changed, 123 insertions(+)
---
base-commit: de9c2c66ad8e787abec7c9d7eff4f8c3cdd28aed
change-id: 20240809-clk_dpum-66a8b28fd7f4

Best regards,
-- 
Kwanghoon Son <k.son@samsung.com>

Hi Kwanghoon,

On Fri, 2024-08-09 at 20:54 +0900, Kwanghoon Son wrote:
> DPUM (Display Processing Unit Main) is main dpu for Exynosautov9.
>     
> This patches enable clock for dpum, sysmmu(iommu).
> 
> I tested using sysmmu, DPP, DECON, DP on sdk board (modetest works).

Does this mean you are working on mainline drivers for these?

If you don't mind me asking, what state are these in?
Is DSIM being worked on?

I am asking these, because I am also working on some drivers for DPP,
DECON, and DSIM on the Exynos7885.

Now, my DPP and DECON drivers are not even close to complete, they just
have the bare minimum for output and vsync, using them as initialized
by the bootloader, but I do have a semi-functional DSIM driver (well,
existing mainline DSIM driver patched up to work on 7885) that is
enough for controlling the DSI panel attached to my phone.

If you are working on DPP and DECON drivers, I won't spend much time
with those for now, as your driver for ExynosAutoV9 may help a lot with
7885 as well.

If you are not working on a DSIM driver, I'll keep working on it for
7885. If you are, I'll have to think about how to move forward.

Best regards,
David

On Thu, 2024-08-22 at 20:59 +0200, David Virag wrote:
> Hi Kwanghoon,
> 
> On Fri, 2024-08-09 at 20:54 +0900, Kwanghoon Son wrote:
> > DPUM (Display Processing Unit Main) is main dpu for Exynosautov9.
> >     
> > This patches enable clock for dpum, sysmmu(iommu).
> > 
> > I tested using sysmmu, DPP, DECON, DP on sdk board (modetest works).
> 
> Does this mean you are working on mainline drivers for these?

Yes. I working on decon exynosautov9 now. (maybe exynosautov920 in
future..)

> If you don't mind me asking, what state are these in?

I'm glad to.
I shift vendor code to mainline. vendor code had split directory, now I
merged to proper mainline drm directory. And it work with modetest. But
a lot of todo.. It's 'somehow' works level.

> Is DSIM being worked on?

For now No. Exynosauto series support DP, MIPI both, but SDK board is
shape of easy to test with DP. For now only modetest works through DP
port interface.

> I am asking these, because I am also working on some drivers for DPP,
> DECON, and DSIM on the Exynos7885.

Ah, that's nice. But actually I'm not involved with exynosauto or the
hardware team. (But I get board and information from them to upstream.)
So I have no idea what the Exynos7885 drm hardware has.

> Now, my DPP and DECON drivers are not even close to complete, they just
> have the bare minimum for output and vsync, using them as initialized
> by the bootloader, but I do have a semi-functional DSIM driver (well,
> existing mainline DSIM driver patched up to work on 7885) that is
> enough for controlling the DSI panel attached to my phone.

For now, I don't understands how your display works because I really
don't know about 7885. But my DPP and decon also far to complete.
(really ugly code.)

> If you are working on DPP and DECON drivers, I won't spend much time
> with those for now, as your driver for ExynosAutoV9 may help a lot with
> 7885 as well.

Here is my suggestion. I will share my code after clean up with NO-
MERGE or WIP tag ASAP. But mind that I only code kernel when have spare
time..

Could you share your code also? even It's not ready.

> If you are not working on a DSIM driver, I'll keep working on it for
> 7885. If you are, I'll have to think about how to move forward.
> 
> Best regards,
> David

+ I added Sam Protsenko <semen.protsenko@linaro.org>
Because I also want to know maybe he has some process for drm, since he
worked on exynos850 sysmmu.

+ Alim Akhtar, Do you know some information around theses?
Especially between 7885, exynosauto?

On Fri, Aug 23, 2024 at 9:56 AM Kwanghoon Son <kwangson@yahoo.com> wrote:

[snip]

>
> > If you are not working on a DSIM driver, I'll keep working on it for
> > 7885. If you are, I'll have to think about how to move forward.
> >
> > Best regards,
> > David
>
> + I added Sam Protsenko <semen.protsenko@linaro.org>
> Because I also want to know maybe he has some process for drm, since he
> worked on exynos850 sysmmu.
>

I'm not working on DRM at the moment. After enabling SysMMU I've
switched to other Exynos tasks, not related to graphics. FWIW, I've
implemented the test driver for SysMMU, which you can use to verify
your SysMMU driver is functional. It's able to perform address
translations initiated from the test driver (using regular CPU). That
way you can test your SysMMU without DPU or other IP blocks initiating
the translation. You can find SysMMU test driver in GitHub repo [1],
and the commit adding the driver is [2].

All that said, I'd be excited to see DRM working on Exynos850 some
day! So please keep me in the loop, and let me know if I can help
somehow (by testing the code, etc).

Thanks!

[1] https://github.com/joe-skb7/linux/tree/e850-96-mainline-iommu
[2] https://github.com/joe-skb7/linux/commit/d712fb52a12ac0958118f9c23a4f1f0d7b9276a7

[snip]