Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-08-24 Thread Fangrui Song
> Hmm, I see and you are right. Well, I hope you can use 
> `__builtin_cpu_supports("x86-64-v3")` in the future once LLVM implements it. 
> It's very unfortunate we've added a useful builti-in, but projects are still 
> forced to parse cpuinfo itself :/

I created https://reviews.llvm.org/D158811 to implement 
`__builtin_cpu_supports("x86-64")` (and v2 v3 v4). It will hopefully be part of 
Clang 18.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1692747146
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-03-14 Thread Martin Liška
Hmm, I see and you are right. Well, I hope you can use 
`__builtin_cpu_supports("x86-64-v3")` in the future once LLVM implements it. 
It's very unfortunate we've added a useful builti-in, but projects are still 
forced to parse cpuinfo itself :/

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1467793381
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-03-10 Thread Fabian Vogt
> > FWICT, the auxiliary vector for HWCAPS is no longer really used and 
> > applications (including glibc, gcc runtime code) have to resort to methods 
> > like this instead. GCC's `__builtin_cpu_supports` does unfortunately not 
> > support all features needed to detect these levels properly.
> 
> What feature do you miss? Note that using `__builtin_cpu_supports` would be 
> much cleaner code and it's supported also by LLVM.

Of the one mentioned here:

> AVX + AVX2 + F16C + FMA + LZCNT + MOVBE + BMI + BMI2 + OSXSAVE

F16C, LZCNT, MOVBE and OSXSAVE aren't exposed through `__builtin_cpu_supports`, 
except by using `x86-64-v3`.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1464415351
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-03-10 Thread Martin Liška
> It's not implemented by clang.

https://github.com/llvm/llvm-project/issues/59961

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1463940613
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-03-10 Thread Martin Liška
> FWICT, the auxiliary vector for HWCAPS is no longer really used and 
> applications (including glibc, gcc runtime code) have to resort to methods 
> like this instead. GCC's `__builtin_cpu_supports` does unfortunately not 
> support all features needed to detect these levels properly.

What feature do you miss? Note that using `__builtin_cpu_supports` would be 
much cleaner code and it's supported also by LLVM.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1463939145
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-03-10 Thread Martin Liška
> It's just not documented... Is that an option for RPM? I guess not. It's not 
> implemented by clang.

It's documented here: 
https://gcc.gnu.org/onlinedocs/gcc/x86-Built-in-Functions.html#index-_005f_005fbuiltin_005fcpu_005fsupports-1
 (though we wrongly documented that under `__builtin_cpu_is` for GCC 12.x).

Well, that's a pity that clang does not support that. It would make the 
implementation much simpler!

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1463933551
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-03-07 Thread Fabian Vogt
> GCC's __builtin_cpu_supports does unfortunately not support all features 
> needed to detect these levels properly.

Looks like since GCC 12.2 it actually can do 
`__builtin_cpu_supports("x86-64-v3")`: 
https://github.com/gcc-mirror/gcc/commit/8ea292591e42aa4d52b4b7a00b86335bfd2e2e85#diff-64afd6a149a7486d332f3d75ac9479f58d0fb674be6920a6e7db0e54cd205e70R38

It's just not documented... Is that an option for RPM? I guess not. It's not 
implemented by clang.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1458293778
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-01-09 Thread Panu Matilainen
Merged manually (cd46c1704ccd8eeb9b600729a0a1c8738b66b847) to tweak the commit 
message a bit, the lines were overflowing in 'git log' view because of the 
indentation it adds and looked ugly.

Thanks for suggesting a way out of this particular arch madness, and the patch!

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1375300128
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2023-01-09 Thread Panu Matilainen
Closed #2315.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#event-8179573548
You are receiving this because you are subscribed to this thread.

Message ID: 
___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2022-12-22 Thread Fabian Vogt
@Vogtinator commented on this pull request.



> @@ -1063,6 +1063,10 @@ package or when debugging this package.\
 #
 %ix86   i386 i486 i586 i686 pentium3 pentium4 athlon geode
 
+#--
+# arch macro for all supported x86_64 processors
+%x86_64x86_64 x86_64_v2 x86_64_v3 x86_64_v4 amd64 em64t

IMO `x86_64v2` looks weird. If someone else also requests that I'll change it.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#discussion_r1055308541
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2022-12-22 Thread ニール・ゴンパ
@Conan-Kudo commented on this pull request.



> @@ -1063,6 +1063,10 @@ package or when debugging this package.\
 #
 %ix86   i386 i486 i586 i686 pentium3 pentium4 athlon geode
 
+#--
+# arch macro for all supported x86_64 processors
+%x86_64x86_64 x86_64_v2 x86_64_v3 x86_64_v4 amd64 em64t

I've slept on this and thought about it, and I think I like @dirkmueller's 
original suggestion of just dropping the `-` instead of converting it to an 
underscore. It's slightly more readable and less confusing. Can you do that, 
please?

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#discussion_r1055305010
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2022-12-22 Thread Michael Schroeder
What's the state of this pull request? Can this be merged now?
I would need to add support for the new architectures to libsolv as well.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1362649807
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint


Re: [Rpm-maint] [rpm-software-management/rpm] Add x86-64 architecture levels (v2-v4) as architectures (PR #2315)

2022-12-15 Thread Fabian Vogt
> See issue #2319

Perfect match - so indeed a mostly independent topic.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/2315#issuecomment-1352967373
You are receiving this because you are subscribed to this thread.

Message ID: ___
Rpm-maint mailing list
Rpm-maint@lists.rpm.org
http://lists.rpm.org/mailman/listinfo/rpm-maint