linux-asahi: kconfig

[Calandracas606]

Testing the changes

• I tested the changes in this PR: YES

Improvements to kconfig

[dkwo]

This pr turns off rust, drm_asahi, etc. Not sure what to do with it. Maybe could you rebase it?

[Calandracas606]

This pr turns off rust, drm_asahi, etc. Not sure what to do with it. Maybe could you rebase it?

Weird, looks like CONFIG_MODVERSIONS disables rust

I was running this kconfig exactly as it, but I guess I was not using DRM, and didn't notice

[dkwo]

Done with first pass. There were some things that did not make sense to me (e.g. stuff related to x86 or ecc), but when in doubt I turned on most of other stuff. Some things I did not bother changing from m to y or viceversa. If there's something left that you feel is important, can you address it directly in the main pr?

[Calandracas606]

Done with first pass. There were some things that did not make sense to me (e.g. stuff related to x86 or ecc), but when in doubt I turned on most of other stuff. Some things I did not bother changing from m to y or viceversa. If there's something left that you feel is important, can you address it directly in the main pr?

What stuff related to x86? The only thing I see is the XZ_DEC, which while I don't think is needed, is copied from void's generic arm64 kconfig.

The ECC bits refer to MTD NAND devices, not ECC memory, it's probably not needed, but when in doubt I've fallen back to the generic kconfig.

Changes from m to y are based on the generic arm64 kconfig

[Calandracas606]

Are there any other specific parts you disagree with? I think it would be easier to use this kconfig, and discuss any changes you would like to make, as I think it is closer to void's configuration than that of the main PR

[dkwo]

It would help if you could distinguish between (1) things that are (probably) not needed vs (2) things that are really needed for Apple silicon to work properly on Void. In my first pass I tried to enable all of (2), and ignored (1), in particular if their config has n as suggested answer, but if I missed something from (2), please tell me.

For example, CONFIG_RCU_TORTURE_TEST and/or CONFIG_RCU_REF_SCALE_TEST break things when set to y, so I'm not sure how you tested those (just like disabling rust before), and I'm not sure it's worth it to enable all of (1), just because they're in Void's arm64, and then fix whatever breaks.

The changing from m to y I don't care, unless it is advised to keep it y in asahi kconfig.

[dkwo]

Since the original pr has now been merged, perhaps we can discuss any further changes (if needed) directly in void linux pull requests.