Re: [PATCH v2 0/3] target/arm: Allow compilation without CONFIG_ARM_V7M

[Thomas Huth]

On 29/01/2024 09.18, Thomas Huth wrote:
> We've got a switch to disable v7m code since a long time - but it
> currently cannot be disabled since linking then fails due to missing
> functions. But thanks to the clean-ups that have been done during the
> past years, it's not that difficult anymore to finally make it possible
> to disable CONFIG_ARM_V7M: We just have to move some v7m-related code
> out of cpu32.c to a separate file (that we only compile if the switch
> CONFIG_ARM_V7M is enabled) and make sure to use the stub functions in
> m_helper.c if it is disabled. Then we can finally remove the hard-coded
> "select ARM_V7M" from the Kconfig file.
>
> v2:
> - Updated a comment
> - Avoid #ifdef in cpu-v7m.c, handle it via meson.build instead
>
> Thomas Huth (3):
>    target/arm: Move v7m-related code from cpu32.c into a separate file
>    target/arm/tcg/m_helper.c: Include the full helpers only with
>      CONFIG_ARM_V7M
>    target/arm/Kconfig: Stop requiring CONFIG_ARM_V7M
>
>   target/arm/tcg/cpu-v7m.c   | 290 +++++++++++++++++++++++++++++++++++++
>   target/arm/tcg/cpu32.c     | 261 ---------------------------------
>   target/arm/tcg/m_helper.c  |   3 +-
>   target/arm/Kconfig         |   4 -
>   target/arm/meson.build     |   3 +
>   target/arm/tcg/meson.build |   3 +
>   6 files changed, 298 insertions(+), 266 deletions(-)
>   create mode 100644 target/arm/tcg/cpu-v7m.c

*ping*

Any hints how to continue here? Or is the series fine as it is?

 Thomas