104

I'm a beginner in assembly language and have noticed that the x86 code emitted by compilers usually keeps the frame pointer around even in release/optimized mode when it could use the EBP register for something else.

I understand why the frame pointer might make code easier to debug, and might be necessary if alloca() is called within a function. However, x86 has very few registers and using two of them to hold the location of the stack frame when one would suffice just doesn't make sense to me. Why is omitting the frame pointer considered a bad idea even in optimized/release builds?

Improve this question
10

5 Answers 5

Reset to default
107

The frame pointer is a reference pointer allowing a debugger to know where a local variable or an argument is, with a single constant offset. Although ESP's value changes over the course of execution, EBP remains the same making it possible to reach the same variable at the same offset (such as the first parameter always being at EBP+8 while ESP offsets can change significantly since you'll be pushing/popping things).

Why don't compilers throw away the frame pointer? Because with the frame pointer, the debugger can figure out where local variables and arguments are using the symbol table since they are guaranteed to be at a constant offset to EBP. Otherwise there isn't an easy way to figure where a local variable is at any point in code.

As Greg mentioned, it also helps stack unwinding for a debugger since EBP provides a reverse linked list of stack frames therefore letting the debugger figure out the size of a stack frame (local variables + arguments) of the function.

Most compilers provide an option to omit frame pointers although it makes debugging really hard. That option should never be used globally, even in release code. You don't know when you'll need to debug a user's crash.

Improve this answer
8
  • 11
    The compiler probably knows what it does to ESP. The other points are valid, though, +1
    – erikkallen
    Commented Feb 24, 2009 at 21:52
  • 11
    Modern debuggers can do stack backtraces even in code compiled with -fomit-frame-pointer. That setting is the default in recent gcc.
    – Peter Cordes
    Commented Oct 25, 2015 at 2:31
  • 3
    @SedatKapanoglu: A data section records the necessary info: yosefk.com/blog/…
    – Peter Cordes
    Commented Oct 25, 2015 at 6:10
  • 4
    @SedatKapanoglu: the .eh_frame_hdr section is used for runtime exceptions, too. You'll find it (with objdump -h) in most binaries on a Linux system, It's about 16k for /bin/bash, vs. 572B for GNU /bin/true, 108k for ffmpeg. There is a gcc option to disable generating it, but it's a "normal" data section, not a debug section that strip removes by default. Otherwise you couldn't backtrace through a library function that didn't have debug symbols. That section may be bigger than the push/mov/pop instructions it replaces, but it has near zero runtime cost (e.g. uop cache).
    – Peter Cordes
    Commented Oct 25, 2015 at 6:32
  • 1
    @AydinK. yes, exactly. thank you for the correction.
    – Sedat Kapanoglu
    Commented May 8, 2022 at 18:32
33

Just adding my two cents to already good answers.

It's part of a good language architecture to have a chain of stack frames. The BP points to the current frame, where subroutine-local variables are stored. (Locals are at negative offsets, and arguments are at positive offsets.)

The idea that it is preventing a perfectly good register from being used in optimization raises the question: when and where is optimization actually worthwhile?

Optimization is only worthwhile in tight loops that 1) do not call functions, 2) where the program counter spends a significant fraction of its time, and 3) in code the compiler actually will ever see (i.e. non-library functions). This is usually a very small fraction of the overall code, especially in large systems.

Other code can be twisted and squeezed to get rid of cycles, and it simply won't matter, because the program counter is practically never there.

I know you didn't ask this, but in my experience, 99% of performance problems have nothing at all to do with compiler optimization. They have everything to do with over-design.

Improve this answer
3
  • 2
    Doing away with the frame pointer also saves you a couple of instructions every function call, which is a small optimization on its own. BTW, your use of "begs the question" is incorrect; you mean "raises the question".
    – augurar
    Commented Feb 14, 2014 at 3:03
  • @augurar: Fixed. Thanks. I'm a bit of a grammar grump myself :)
    – Mike Dunlavey
    Commented Feb 14, 2014 at 3:06
  • 4
    @augurar Language evolves: "Begs the question" now just means "raises the question". Being a prescriptivist nitpicker for outdated usage adds nothing.
    – user3364825
    Commented Jul 21, 2014 at 9:39
11

It depends on the compiler, certainly. I've seen optimized code emitted by x86 compilers that freely uses the EBP register as a general purpose register. (I don't recall which compiler I noticed that with, though.)

Compilers may also choose to maintain the EBP register to assist with stack unwinding during exception handling, but again this depends on the precise compiler implementation.

Improve this answer
1
  • Most compilers default to -fomit-frame-pointer when optimization is enabled. (when the ABI allows it). GCC, clang, ICC, and MSVC all do it, IIRC, even when targeting 32-bit Windows. Yup, my answer on Why is it better to use the ebp than the esp register to locate parameters on the stack? shows that even 32-bit Windows can omit the frame pointer. 32-bit x86 Linux definitely can and does. And of course 64-bit ABIs have allowed frame-pointer omission from the start.
    – Peter Cordes
    Commented Oct 4, 2019 at 11:07
4

However, x86 has very few registers

This is true only in the sense that opcodes can only address 8 registers. The processor itself will actually have many more registers than that and use register renaming, pipelining, speculative execution, and other processor buzzwords to get around that limit. Wikipedia has a good introductory paragraph as to what an x86 processor can do to overcome the register limit: http://en.wikipedia.org/wiki/X86#Current_implementations.

Improve this answer
3
  • 1
    The original question is about generated code, which is strictly limited to the registers referenceable by opcodes.
    – Darron
    Commented Feb 23, 2009 at 21:07
  • 1
    Yes, but this is why ommitting the frame pointer in optimized builds isn't as important nowadays.
    – Michael
    Commented Feb 23, 2009 at 21:10
  • 2
    Register renaming isn't quite the same thing as actually having a larger number of registers available though. There are still plenty of situations where register renaming won't help, but more "regular" registers would.
    – Stack Overflow is garbage
    Commented Feb 23, 2009 at 23:13
1

Using stack frames has gotten incredibly cheap in any hardware even remotely modern. If you have cheap stack frames then saving a couple of registers isn't as important. I'm sure fast stack frames vs. more registers was an engineering trade-off, and fast stack frames won.

How much are you saving going pure register? Is it worth it?

Improve this answer
1
  • More registers is limited by instruction encoding. x86-64 uses bits in the REX prefix byte to extend the register-specifying part of instructions from 3 to 4 bits for src and dest registers. If there was room, x86-64 probably would have gone to 32 architectural registers, although saving/restoring that many on context switches starts to add up. 15 is a huge step up from 7, but 31 is a much smaller improvement in most cases. (not counting the stack pointer as general-purpose.) Making push/pop fast is great for more than just stack frames. It's not a tradeoff with # of regs, though.
    – Peter Cordes
    Commented Oct 25, 2015 at 2:39

Not the answer you're looking for? Browse other questions tagged or ask your own question.