What does % mean on register names in assembly language?

Question

I tried to convert and interpret C language code into assembly language with GCC -S option. What is the difference between push %rbp and push rbp?

Answer 1

Both statements you gave do the same thing, the main difference is in the syntax.
There are 2 major syntax conventions for X86, the Intel convention and the AT&T convention.
See syntax comparison for details.

[UPD:] Just in case, duplicating the syntax information here:

	AT&T	Intel
Parameter order	Source before the destination. movl $5, %eax	Destination before source. mov eax, 5
Parameter size	Mnemonics are suffixed with a letter indicating the size of the operands: q for qword, l for long (dword), w for word, and b for byte. addl $4, %esp	Derived from the name of the register that is used (e.g. rax, eax, ax, al imply q, l, w, b, respectively). add esp, 4
Sigils	Immediate values prefixed with a $, registers prefixed with a %.	The assembler automatically detects the type of symbols; i.e., whether they are registers, constants or something else.
Effective addresses	General syntax of DISP(BASE,INDEX,SCALE). movl mem_addr(%ebx,%ecx,4), %eax	Arithmetic expressions in square brackets; additionally, size keywords like byte, word, or dword have to be used if the size cannot be determined from the operands. mov eax, [ebx + ecx*4 + mem_addr]

Answer 2

In AT&T syntax, push rbp would be a push of a memory operand, with symbol name rbp, like if you compiled some C code that used long rbp; as a global variable.

rbp doesn't have a % so the assembler knows you aren't referring to the register of the same name; that's how AT&T syntax disambiguates this case. If you do want the register, you must use % like %rbp.

• Questions about AT&T x86 Syntax design
• What was the original reason for the design of AT&T assembly syntax?

rbp is an absolute addressing mode; normally you'd use pushq rbp(%rip) if you actually wanted to push the contents of a global. Push-memory with any other addressing mode looks like pushq (%rdi, %rsi, 8). In 32-bit code it is normal to access global variables with their bare symbol name, like mov foo, %eax to load from one.

---

But more likely you meant push rbp as the Intel-syntax version of push %rbp, pushing the register like a normal person.

gcc -masm=intel will use Intel syntax, the default is -masm=att. Pick whichever you find easier to read. See How to remove "noise" from GCC/clang assembly output?

For more details on syntax differences, see @hidefromkgb's answer, and:

• https://stackoverflow.com/tags/intel-syntax/info
• https://stackoverflow.com/tags/att/info
• Both are linked from https://stackoverflow.com/tags/x86/info which has lots of useful links to manuals and guides.