[llvm-dev] LLVM Assembly Instructions Generation Syntax

[mats petersson via llvm-dev]

There are (at least) two styles of x86 assembly: "Intel" and "AT&T" (aka
"att"). You can tell clang/llvm which you want to use - e.g in llc you can
use -x86-asm-syntax=intel to get intel syntax. When using clang, you can
use -mllvm -x86-asm-syntax=intel to achieve the same thing.

The gnu assembler (by default at least) does indeed use the AT&T syntax,
but it's the way most assemblers on any Unix-related toolchains work. I
think the reason behind this is to have the same syntax, whether you are
using a motorola 68k, a digital pdp-11 or an x86 processor, although I
wasn't there when this was decided some 40-50 years back when Unix was
first invented at AT&T (hence the name of the syntax).

--
Mats



On 22 June 2017 at 08:32, hameeza ahmed via llvm-dev <
llvm-dev at lists.llvm.org> wrote:

> Hello,
> I am generating some assembly code via llvm and using x86.
> in all the instructions source appears before destination like;
>
> subl $8, %esp
> movl 16(%esp), %eax
>
> vmovdqu32 c(%rip), %zmm0
> addl b+132(%rip), %eax
> movl %eax, a+132(%rip)
>
>
> but when i refered the intel x86 manual, it says instruction has 1st
> destination and then source;
>
> is it due to llvm follows gas syntax instead of intel? but my machine and
> normally machines are intel based then how does this gas syntax assembly of
> llvm execute on our machines?
>
> please correct me if i am wrong.
>
>
> Thank You
>
>
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20170622/60a01ff2/attachment.html>