Tailor config to non-surface laptops (mainly AMD)

[m4l352k4]

Hi!

I've been using Linux-NSPA for half a year now on my thinkpad p14s gen4 (ryzen 7840U machine, same as many frameworks, etc.) and it works really well. I'm running Ableton 12.5.1 under wine (soda-9.0.1 and wine-nspa 8.19) and it's all perfect, so thank you!

I had to change few things to make it work on my AMD platform:

• I added logic that disables certain patches when config_Surface7 is unset, e.g. 1116-kt-pm-suspend_patches.patch prevented my system from entering suspend state when closing the lid.
• fixed building issues with 1131-pmdomain_rt_hacks.patch (see Build fails with _config_Surface unset #5)

I have also made some ‘convenience’ changes:

• make -j $(nproc) instead of make -j 3
• removed auto-cpu-optimization as it was unused (it requires custom kbuild options that are in e.g. zen kernel and weren't patched here); we can use CONFIG_X86_NATIVE_CPU now
• answer automatically default options (usually 'no') to localmodconfig

Regards!

[m4l352k4]

Oh, patch selection logic is completely broken lol. All patches are included independently from options like _vmx_hack=. I added a fix for that: 6828979

Also a sidenote: 1130-sched-Introduce-Cache-aware-scheduling-V3.patch fails if _sched_feat_comp is unset. But _sched_feat_comp=y doesn't make my system unstable so it's a good default

[nine7nine]

Okay, so a few things:

1. I will NOT be pulling your patch to use all cpu cores, but i can add an option to customize how many cores to use. Using all cores makes my surface hot, and I don't need to use 8 cpu threads for kernel compilation.
2. I can drop that auo-cpu-optimization, now that NATIVE-CPU is merged. ~ However, it is incorrect to state that as a Zen-Kernel thing though. That was Graysky's cpu-arch patchwork, which i recently dropped.
3. Not sure why you put 1119-mm-compaction-Skip_all_non-migratable_pages_during_scan.patch as a surface only thing?!
4. I will pull the make localmodconfig patch. that makes sense
5. PM stuff will be dropped and/or changing on next update, so not relevant.
6. 1130-sched-Introduce-Cache-aware-scheduling-V3.patch very specifically says in the patch that it has been modified to work with sched_feat_comp > and I don't plan on maintaining 2 separate patches to support different configs that i don't use... I can just set that behind sched_feat_comp - or just make it all non-optional.
7. Patching logic was right to begin with: your first patch is broken, so it seems silly to pickup your broken patch, or the one that fixes it.
8. I'll update the generic config to use native cpu option

So I won't be merging this pull request (as is), but I will cherry-pick a couple of them, along with taking others into consideration/adapting my package slightly, when I push linux-6.16.1 (not moving to 6.17 for a bit > due to the proxy execution patchwork that landed && needing an updated version of SCHED_CACHE patchwork).