GAS

From OSDev.wiki
Revision as of 13:02, 17 December 2016 by Jankowalski25 (talk | contribs) (typo)
Jump to navigation Jump to search
This page is a stub.
You can help the wiki by accurately adding more contents to it.

GNU Assembler

[[Image:{{{image}}}|300px|]]
Website:http://www.gnu.org/software/binutils/

The GNU Assembler (GAS, executable named 'as') is part of the binutils package. GCC generates assembly source code and automatically invokes GAS which assembles the code into machine code in an object file. GCC lets you forward assembly directly to the assembler using inline assembly feature.

GAS is not one assembler, but a collection of assemblers - one for each of the platforms supported by GCC. They are very similar in their available options, their macro language, and so on. GAS provides you with a very solid and well-supported assembler that is fully integrated with the other GNU tools (implicit make rules, inline assembly from C/C++ sources, the ability to be called as gcc -o myfile.o myfile.s, the same assembly syntax used by all the tools).

Usage

AT&T Syntax

The AT&T assembly syntax is traditional for Unix-like operating systems on the x86 and x86-64 platforms. The syntax is reminiscent of the original assembler format used in the original Unix operating system. The operand order for instructions is source and then destination. The assembly language is unambiguous as registers must be prefixed with a % character and constants must be prefixed with the $ character.

Intel Syntax Support

Up until v2.10 of binutils, GAS supported only the AT&T syntax for x86 and x86-64, which differs significantly from the Intel syntax used by virtually every other assembler. Today, GAS supports both syntax sets (.intel_syntax and the default .att_syntax), and even allows disabling the otherwise mandatory operand prefixes '%' or '$' (..._syntax noprefix). There are some pitfalls - several FP opcodes suffer from a reverse operand ordering that is bound to stay in there for compatibility reasons, .intel_syntax generates less optimized opcodes on occasion (try mov'ing to %si...).

It is generally discouraged to use the support for Intel Syntax because it can subtly and surprisingly different than the real Intel Syntax found in other assemblers. If you want that syntax, you should use an assembler that uses it natively.

See Also

Articles

External Links