Re: [PATCH v5 00/39] ep93xx device tree conversion

On Wed, Nov 22, 2023 at 11:59:38AM +0300, Nikita Shubin via B4 Relay wrote:

This series aims to convert ep93xx from platform to full device tree support.

The main goal is to receive ACK's to take it via Arnd's arm-soc branch.

Bit thanks to Krzysztof Kozlowski for his zero-day review.

Krzysztof, Sergey, i've changed some files that you have already provided tag:
- dt-bindings: spi: Add Cirrus EP93xx
- ata: pata_ep93xx: add device tree support

Added DMA and renamed clock header file.

Sergey, Damien, i've changed some files that you have already provided tag:
- ata: pata_ep93xx: add device tree support

Added OF DMA setup and dropped platform file.

Mark, i've changed some files that you have already provided tag:
- spi: ep93xx: add DT support for Cirrus EP93xx

Added OF DMA setup and dropped platform file.

Major updates:

- reboot, pinctrl, clk are now auxiliary and instantiated from SoC driver as
  Stephen Boyd suggested
- i moved all clock code to clk-ep93xx.c, as it no longer has a separate dt node, 
  so XTALI is externalk for this driver and passed as index 0, and pll1, pll2 are
  internal and passed via pointer for parent_data
- reboot bindings dropped
- pinctrl and clk bindings moved to syscon YAML
- xlate added for DMA, so now all DMA users use it via device tree, otherwise probe 
  order messed up and we might end up probing before DMA with no possibility to defer probing
- DMA port bindings dropped, they are described in YAML file
- DMA platform code dropped
- i2s, spi, pata now use OF DMA
- YAML and dtsi/dts changed to reflect auxiliary conversion and DMA changes

Patches should be now formated with '--patience'

It seems we are a step behind on Torvalds' mind :-)
A few weeks ago he suggested to use --histogram is the best option.
But hey, --patience probably is good enough for this version, so
thank you for using it!

-- 
With Best Regards,
Andy Shevchenko