[GIT PULL] Renesas ARM64 Based SoC DT Updates for v4.18

On Sat, Jun 02, 2018 at 01:32:20AM -0700, Olof Johansson wrote:

On Mon, May 28, 2018 at 09:44:41AM +0200, Simon Horman wrote:

quoted

Hi Olof,

On Sat, May 26, 2018 at 02:14:20PM -0700, Olof Johansson wrote:

quoted

Hi Simon,

On Fri, May 18, 2018 at 01:16:00PM +0200, Simon Horman wrote:

quoted

Hi Olof, Hi Kevin, Hi Arnd,

Please consider these Renesas ARM64 based SoC DT updates for v4.18.


The following changes since commit 60cc43fc888428bb2f18f08997432d426a243338:

  Linux 4.17-rc1 (2018-04-15 18:24:20 -0700)

are available in the git repository at:

  https://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas.git tags/renesas-arm64-dt-for-v4.18

for you to fetch changes up to 908001d778eba06ee1d832863d4e9a1e2cfd4746:

  arm64: dts: renesas: salvator-common: Add ADV7482 support (2018-05-18 11:52:03 +0200)

This pull request is really, really hard for us to digest. The tag
description is very large, and it repeats several SoCs several times,
making it hard to get an overview of what is in it. The verbosity of "<x>
says.." makes it harder on this size of a pull request as well.

I appologise that this pull-request has turned out to be hard to digest.

I think that one reason for this is that there are an unusally large number
of commits. It seems that the team has been a little more productive than
usual and I might have been better to send a smaller pull-request earlier
in the development cycle and then follow-up with a second batch. I'll try
to pay more attention to this aspect of things going forwards.

Regarding your specific comments:

* Repetition of SoCs. I do try to group things in a logical manner,
  but clearly I failed in this case. I'll try to make that a bit clearer
  in future.

I don't think you're helped by the fact that Renesas product names seem
to have little rhyme or reason to them, and mumbles families and product
names and numbers seemingly random. :(

quoted

* Verbosity: There as a request a few development cycles for me
  to include more information in the commit logs. It seems that
  I may have gone a bit too far. I'll try to find a better balance next
  time around.

There's a balance to be found here. You chose to describe some very small
changes with 3 lines of description, and as a result it's hard to get an
overview of what's in the pull request.

As a maintainer, you should strive to be the editor that makes sure that
things are easy to understand, and arranged in a manner that makes sense
for both your developers and the rest of the community (and the upstream
maintainers).

Understood. I'll work on making things easier to digest for all concerned.

quoted

* Patches for the same change split up for different SoCs/boards.
  Featurs broken out into incremental patches. And so on.

  This has been a long-standing practice for Renesas SoC development.
  We find that in general it aids review. It also works well
  with the way we develop patches. But I do see your point that
  it may be a little excessive - f.e. multiple patches for the same
  whitespace fixes. Again, we'll try to find a better balance.

I can understand where the policy comes from, but it seems to me that at least
in this case it's one that's no longer scaling to the number of SoCs you're
supporting, and the number of (each small) additions that you're doing.
Grouping in some of the simpler dtsi additions into one per SoC, or one per IP
block across all SoCs, seems like something to attempt here.

I think your point about scaling is well made. We'll start tweaking our
processes to see if we can find a better way.

quoted

quoted

For example:

quoted

* Condor board with R-Car V3H (r8a77980) SoC
  - Enable eMMC

    Sergei Shtylyov says "We're adding the R8A77980 MMC (SDHI)
    device nodes and then enable eMMC support on the Condor board."

The "Enable eMMC" line is just fine here.

quoted

----------------------------------------------------------------
Geert Uytterhoeven (11):
      arm64: dts: renesas: draak: Rename EtherAVB "mdc" pin group to "mdio"
      arm64: dts: renesas: salvator-common: Rename EtherAVB "mdc" pin group to "mdio"
      arm64: dts: renesas: ulcb: Rename EtherAVB "mdc" pin group to "mdio"

Why can't these be done in one commit?

quoted

      arm64: dts: renesas: r8a7795: Correct whitespace
      arm64: dts: renesas: r8a7796: Correct whitespace
      arm64: dts: renesas: r8a77965: Correct whitespace

Do these really need to be three commits to fix some whitespace?

quoted

      arm64: dts: renesas: ulcb: Add BD9571 PMIC
      arm64: dts: renesas: salvator-common: Add PMIC DDR Backup Power config
      arm64: dts: renesas: ulcb: Add PMIC DDR Backup Power config
      arm64: dts: renesas: r8a77970: Add secondary CA53 CPU core
      arm64: dts: renesas: r8a77970: Add Cortex-A53 PMU node

Why can't these be done in the same commit?

quoted

Kieran Bingham (7):
      arm64: dts: renesas: r8a77965: Add FCPF and FCPV instances
      arm64: dts: renesas: r8a77965: Add VSP instances
      arm64: dts: renesas: r8a77965: Populate the DU instance placeholder
      arm64: dts: renesas: r8a77965: Add HDMI encoder instance
      arm64: dts: renesas: r8a77965-salvator-x: Enable DU external clocks and HDMI
      arm64: dts: renesas: r8a77965-salvator-xs: Enable DU external clocks and HDMI

These two can probably be in one commit as well.

quoted

      arm64: dts: renesas: salvator-common: Add ADV7482 support

Kuninori Morimoto (8):
      arm64: dts: renesas: r8a7795: add HDMI sound support
      arm64: dts: renesas: r8a7796: add HDMI sound support

... starting to see a trend?

quoted

      arm64: dts: renesas: salvator-common: use audio-graph-card for Sound
      arm64: dts: renesas: r8a7795-es1-salvator-x: enable HDMI sound
      arm64: dts: renesas: r8a7795-salvator-xs: enable HDMI sound
      arm64: dts: renesas: r8a7796-salvator-xs: enable HDMI sound
      arm64: dts: renesas: r8a7795-salvator-x: enable HDMI sound
      arm64: dts: renesas: r8a7796-salvator-x: enable HDMI sound

... and more.

quoted

Magnus Damm (5):
      arm64: dts: renesas: r8a77970: Update IPMMU DS1 bit number
      arm64: dts: renesas: r8a7795: Enable IPMMU devices
      arm64: dts: renesas: r8a7796: Enable IPMMU devices
      arm64: dts: renesas: r8a77970: Enable IPMMU devices
      arm64: dts: renesas: r8a77995: Enable IPMMU devices

I think these 4 could be in one commit too.

quoted

Niklas S?derlund (11):
      arm64: dts: renesas: r8a7795: decrease temperature hysteresis
      arm64: dts: renesas: r8a7796: decrease temperature hysteresis
      arm64: dts: renesas: r8a77965: use r8a77965-sysc binding definitions
      arm64: dts: renesas: r8a77965: Add R-Car Gen3 thermal support
      arm64: dts: renesas: r8a77965: add I2C support
      arm64: dts: renesas: r8a7795: add VIN and CSI-2 nodes
      arm64: dts: renesas: r8a7795-es1: add CSI-2 node
      arm64: dts: renesas: r8a7796: add VIN and CSI-2 nodes
      arm64: dts: renesas: r8a77965: add VIN and CSI-2 nodes
      arm64: dts: renesas: r8a77970: add VIN and CSI-2 nodes

.... etc, etc. I'll stop here.


I haven't merged this branch yet, will need to set aside more time to review
the contents. I can't guarantee that it'll make v4.18 but I'll try.

Please let me know if there is anything I can do to help make this happen.
This work is very important to Renesas.

I've queued this and the SoC branch up in a next/late branch now. I'm planning
on sending it in unless it's a messy or hard merge window for some reason.

Thanks, much appreciated.