| View previous topic :: View next topic |
| Author |
Message |
simonbcn
n00b
Joined: 01 Aug 2011
Posts: 69
Location: Denmark
|
 Posted: Wed Dec 28, 2022 11:15 am Post subject: [SOLVED]gentoo-sources with experimental omits grasky2 patch |
 |
|
I am installing Gentoo in a virtual machine.
Active profile is: /etc/portage/make.profile -> ../../var/db/repos/gentoo/profiles/default/linux/amd64/17.1/systemd/clang/
The "problem" is that when I install gentoo-sources with USE "experimental", the graysky2 kernel_compiler_patch patch to allow compiling the kernel with march=native should be applied and it is not doing it.
These are the options shown for "Processor family".
I have tested with both the stable kernel version (5.15.80) and the latest available (6.1.1).
Last edited by simonbcn on Wed Dec 28, 2022 12:45 pm; edited 1 time in total |
|
| Back to top |
|
 |
NeddySeagoon
Administrator
Joined: 05 Jul 2003
Posts: 54300
Location: 56N 3W
|
| Posted: Wed Dec 28, 2022 11:22 am Post subject: |
|
|
simonbcn,
Feel free to add the patch yourself.
If it breaks, you can keep all the pieces. Its your Gentoo your way.
_________________
Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
| Back to top |
|
|
simonbcn
n00b
Joined: 01 Aug 2011
Posts: 69
Location: Denmark
|
| Posted: Wed Dec 28, 2022 12:44 pm Post subject: |
|
|
Thanks. I just did it. In this way, which I didn't know, it is very easy to apply a patch like this and it works.
Anyway, I have created a bug report because this patch should be applied even if the compiler is Clang: https://bugs.gentoo.org/888727 |
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 21708
|
| Posted: Wed Dec 28, 2022 1:46 pm Post subject: |
|
|
It looks like the patch can be dropped for a few different cases. In my opinion, the most likely reason is that your gcc-major-version reports less than 9: | Code: | local GCC_MAJOR_VER=$(gcc-major-version)
# ...
if [[ ${GCC_MAJOR_VER} -lt 9 ]]; then
UNIPATCH_DROP+=" 5010_enable-cpu-optimizations-universal.patch" |
What value does GCC_MAJOR_VER receive on your system? |
|
| Back to top |
|
|
simonbcn
n00b
Joined: 01 Aug 2011
Posts: 69
Location: Denmark
|
| Posted: Wed Dec 28, 2022 1:51 pm Post subject: |
|
|
| Hu wrote: | It looks like the patch can be dropped for a few different cases. In my opinion, the most likely reason is that your gcc-major-version reports less than 9: | Code: | local GCC_MAJOR_VER=$(gcc-major-version)
# ...
if [[ ${GCC_MAJOR_VER} -lt 9 ]]; then
UNIPATCH_DROP+=" 5010_enable-cpu-optimizations-universal.patch" |
What value does GCC_MAJOR_VER receive on your system? |
| Code: | $ gcc --version
gcc (Gentoo 11.3.1_p20221209 p3) 11.3.1 20221209
... |
|
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 21708
|
| Posted: Wed Dec 28, 2022 2:20 pm Post subject: |
|
|
That is not what I asked. I want to know what value that variable has. Its value is a function of eclass/toolchain-funcs.eclass:gcc-major-version. That is a function of eclass/toolchain-funcs.eclass:_gcc_fullversion. That depends on the output of tc-getCPP processing __GNUC__. If you provide emerge --info for this package, I can try to extrapolate the value. If I had to guess, you set CC=clang somewhere, and clang has a known issue that it claims to be gcc-4: | Code: | $ echo __GNUC__ | clang -E -x c -
# 1 "<stdin>"
# 1 "<built-in>" 1
# 1 "<built-in>" 3
# 361 "<built-in>" 3
# 1 "<command line>" 1
# 1 "<built-in>" 2
# 1 "<stdin>" 2
4 |
4 is less than 9, so this check would trigger. Arguably, the toolchain-funcs test ought to verify that CC is actually gcc before trusting that __GNUC__ is sane. |
|
| Back to top |
|
|
|
Installing Gentoo
