treewide: adjust CMakeLists.txt files indentation

[josuah]

The rules are interpolated from what there is on the docs:

https://docs.zephyrproject.org/latest/contribute/style/cmake.html#commands-and-syntax

# Good:
target_sources(my_target PRIVATE
  src/file1.cpp
  src/file2.cpp
)

And what seemed to be common on in-tree, in particular for if() and messages():

• Long lines have continuation lines use 2 spaces

  if(long line ...
    continuation line...
  

  function(arg1 arg2...
    continuation line...
  

• When the first line is filled with content, the closing parenthesis is
  at the end of the last continuation line

  if(long line filled until the end ...
    continuation line...)
  

  message(long line filled until the end...
    continuation line...)
  

• When the first line is empty or only containing a distinct label, such
  as an UPPERCASE label) the closing parenthesis is on its own line,
  with same indentation level as the first line.

  zephyr_library_sources(
    something_a.c
    something_b.c
  )
  

  zephyr_library_sources_ifdef(CONFIG_SOMETHING
    something_a.c
    something_b.c
  )
  

• Call to message() are converted from \ for escaping newlines to a list
  of "... " strings, with an empty space at the end if separation between
  words is needed.

  message("long line... Hello "
    "cmake! continuation line...")
  

If anything seems wrong choice, let me know of what to apply and I'd be glad to apply it again tree-wide.

If the scope of modifications is wrong (i.e. should do all CMakeLists.txt and .cmake), then let me know and I would change the scope for this PR.

I am keeping it as a draft for now to avoid triggering the bot.

[josuah]

Since this touches to every CMakeLists.txt, a second batch of commit also adds zephyr-keep-sorted-start/end comments for both CMakeLists.txt and Kconfig, making just one PR for all the "CMake style related changes".

If this is too much for a single PR, then I can easily split it.

[tejlmand]

approved.

Observations to be followed up.

[josuah]

force-push:

• grouped by top-level dirs instead of per-area granularity
• style: prefix
• applied review comments from @tejlmand

[sonarqubecloud]

Please retry analysis of this Pull-Request directly on SonarQube Cloud

[nashif]

@tejlmand already approved, merging this now to avoid any need to rebase again given the amount of files being changed and the fact this is considered "trivial"