Arch Linux User Repository

The kernel config process is too unweildly for me to debug in detail. I rebuilt my config my hand in make xconfig. I no longer see the L1_CACHE_SHIFT error, but I also do not know what originally caused it to happen.

It may be related to the native-architecture options. I'm not sure the string the Xanmod script is using is the correct string expected by the kernel scripts.

Hi @saburouta

I'll check later. Didn't build the -git package in a while, but should work.

Part of the linux-kernel config scripts seem to break every couple of years or so. Is anyone else seeing this old issue return? I tried manually adding the symbol back to the config file, but apparently all make targets overwrite the file with a bad default file. (For what it's worth, I have a 7960X, which may have a weirdly-detected architecture.)

./arch/x86/include/asm/cache.h:8:26: error: 'CONFIG_X86_L1_CACHE_SHIFT' undeclared here (not in a function); did you mean 'CONFIG_X86_INTERNODE_CACHE_SHIFT'?
    8 | #define L1_CACHE_SHIFT  (CONFIG_X86_L1_CACHE_SHIFT)
      |                          ^~~~~~~~~~~~~~~~~~~~~~~~~

Does anyone know of a workaround? I tried several different make targets to see if I could get the build to just accept my file, but it seems they all perform a broken kind of over-writing.

Current build command: make ${_compiler_flags} KCONFIG_CONFIG=.config

Thanks for fixing it!

According to xanmod.org official page, sources now are pushed to https://gitlab.com/xanmod/linux.git

I'll change the source pointing there.

It works perfectly (thanks!) though I did update the source from git+https://github.com/xanmod/linux.git to git+https://gitlab.com/xanmod/linux.git in order to compile 6.12.1.

I have updated the PKGBUILD and the microarchitecture script.

    Available CPU microarchitectures:

    1) AMD Opteron/Athlon64/Hammer/K8
    2) AMD Opteron/Athlon64/Hammer/K8 with SSE3
    3) AMD 61xx/7x50/PhenomX3/X4/II/K10
    4) AMD Family 10h (Barcelona)
    5) AMD Family 14h (Bobcat)
    6) AMD Family 16h (Jaguar)
    7) AMD Family 15h (Bulldozer)
    8) AMD Family 15h (Piledriver)
    9) AMD Family 15h (Steamroller)
   10) AMD Family 15h (Excavator)
   11) AMD Family 17h (Zen)
   12) AMD Family 17h (Zen 2)
   13) AMD Family 19h (Zen 3)
   14) AMD Family 19h (Zen 4)
   15) Intel P4 / older Netburst based Xeon
   16) Intel Core 2 and newer Core 2 Xeons (Xeon 51xx and 53xx)
   17) Intel Atom
   18) Intel 1st Gen Core i3/i5/i7-family (Nehalem)
   19) Intel 1.5 Gen Core i3/i5/i7-family (Westmere)
   20) Intel Silvermont
   21) Intel Goldmont (Apollo Lake and Denverton)
   22) Intel Goldmont Plus (Gemini Lake)
   23) Intel 2nd Gen Core i3/i5/i7-family (Sandybridge)
   24) Intel 3rd Gen Core i3/i5/i7-family (Ivybridge)
   25) Intel 4th Gen Core i3/i5/i7-family (Haswell)
   26) Intel 5th Gen Core i3/i5/i7-family (Broadwell)
   27) Intel 6th Gen Core i3/i5/i7-family (Skylake)
   28) Intel 6th Gen Core i7/i9-family (Skylake X)
   29) Intel 8th Gen Core i3/i5/i7-family (Cannon Lake)
   30) Intel 8th Gen Core i7/i9-family (Ice Lake)
   31) Xeon processors in the Cascade Lake family
   32) Intel Xeon (Cooper Lake)
   33) Intel 3rd Gen 10nm++ i3/i5/i7/i9-family (Tiger Lake)
   34) Intel Sapphire Rapids
   35) Intel Rocket Lake
   36) Intel Alder Lake

   91) Generic-x86-64-v1 (LEGACY < 2003. Not supported by XanMod project)
   92) Generic-x86-64-v2 (Nehalem and newer)
   93) Generic-x86-64-v3 (Haswell and newer)
   94) Generic-x86-64-v4 (AVX512 CPUs)

   98) Intel-Native optimizations autodetected by GCC
   99) AMD-Native optimizations autodetected by GCC

    0) Generic x64-v2 (default)

I haven't time to test it, please do it if you can, or I'll try to test it next week. Thanks!

Other than the git origin & the weird x64-v2 localversion naming issue, it works great. Thanks for looking into it!

@CrimsonKnight13 I have not used this package since xanmod always make releases, but if I time this weekend I'll update it and check it.

After changing the remote origin to the gitlab version due to the github one being archived, I've hit a weird wall. It compiles fine & the gcc optimization script sees that I have it set to AMD native (99) but once the kernel installs, it shows up as x64-v2.

Did something change in the source to no longer recognize 99 as AMD native (currently set as an export in makepkg.conf)? It appears to be auto-selecting 96 or the code from PKGBUILD isn't working properly as follows:

``# Choose between the 4 main configs for stable branch. Default x86-64-v1 which use CONFIG_GENERIC_CPU2:

Possible values: config_x86-64-v1 / config_x86-64-v2 (default) / config_x86-64-v3 / config_x86-64-v4

This will be overwritten by selecting any option in microarchitecture script

Source files: https://github.com/xanmod/linux/tree/5.17/CONFIGS/xanmod/gcc

if [ -z ${_config+x} ]; then _config=config_x86-64-v2 fi``

Any help is greatly appreciated. Also, thanks for all of your hard work with the xanmod packages, figue!

Edit: I'm not super familiar with python-markdown & fenced code. Any pointers to get the code block right?

Edit 2: It appears that its just the naming scheme that is not reflecting it as native

zcat /proc/config.gz | grep CONFIG_MNATIVE_AMD

CONFIG_MNATIVE_AMD=y