All Questions
Tagged with gcc compiler-construction
791
questions
-1
votes
0
answers
21
views
Why is GCC not used as Shader Compiler in Mesa? [closed]
Mesa GPU driver currently employs LLVM by default or Valve's ACO as a shader compiler. Why doesn't Mesa use GCC as default for shader compilation like GCC is default for building Linux/BSD?
Mesa is ...
- 966
0
votes
3
answers
115
views
When inlining a C function, can an optimizing compiler dereference a pointer more times than explicitly written in the source code?
Consider the following code:
int y = 1;
int* p = &y;
inline int f(int x) {
return x + x;
}
int g(void) {
return f(*p);
}
In this code, there is one explicit dereference.
Is a C compiler ...
0
votes
0
answers
19
views
why in calling a function by GCC convention , the passed parameters are stored in a reversed order in the stack
Can you help me answer this question. I have tried to figure this out by myself but didn't succeed. Is it because we don't know the amount of memory the function will use ?
- 25
1
vote
1
answer
125
views
What is the reasoning behind this boilerplate start-up code?
Here's a snippet from disassembled output of a simple program with gcc on Linux:
400478: ff 15 6a 0b 20 00 callq *0x200b6a(%rip) # 600fe8 <__libc_start_main@GLIBC_2.2.5>
...
- 1,183
0
votes
0
answers
89
views
How to compile to RISC-V?
I'm writing the final stage for my compiler to riscv; however, I have run into a minor issue of figuring out what the start point of the program is. Are we guaranteed that the PC starts execution with ...
- 1,183
1
vote
0
answers
32
views
Unknown type name ‘Node’ in the parser file. (Yacc program)
I've been working on a Yacc program (parser.y) to develop a small compiler that handles arithmetic operations. I'm using Yacc (Bison) and Lex tools to generate the parser and lexical analyzer, ...
- 41
0
votes
1
answer
28
views
Where can I find the body of the function pointed to by the function pointer of cfg_hooks in gcc source code？
In gcc cfghooks.cc
void
predict_edge (edge e, enum br_predictor predictor, int probability)
{
if (!cfg_hooks->predict_edge)
internal_error ("%s does not support predict_edge", ...
- 11
2
votes
0
answers
110
views
Write a gcc plugin that inserts macros at the beginning of each function
I understand that I can use __cyg_profile_func_enter and -finstrument-functions to insert code, but is it possible to insert macros by writing a gcc plugin?
Is there any documentation or possibly an ...
- 39
1
vote
0
answers
65
views
How to force GCC change the generated assembly order of a branch?
My problem is deterministically arranging the position of different branches' code in the compiled binary/assembly. Note that my primary goal is NOT to optimize performance, but to be sure to have a ...
0
votes
0
answers
20
views
modifying AST as constructed in gcc
I would like to add a few branches to the AST as being constructed by the parser of gcc, https://github.com/gcc-mirror/gcc/blob/master/gcc/cp/parser.cc
What is the variable that holds the AST as being ...
- 185
3
votes
0
answers
69
views
With gcc -O3 optimization, why does the loop index's type affect gcc's internal GIMPLE vector constants (but not the final asm)?
I created a struct array, and here is my source code.
#include <stdio.h>
#include <stdlib.h>
#define N 256
typedef struct arc {
int num;
} arc_t;
typedef double cost_t;
typedef ...
- 31
6
votes
3
answers
380
views
Why compilers doesn't inline everything, analyze it and then generate their own optimized functions?
Problem summary
I'm trying to do an in-memory database with transactions. I'm reaching bottlenecks on compiler level, specifically inlining limit. I don't have much knowledge about compilers and I ...
- 556
0
votes
0
answers
26
views
Which gcc O2 option do this constant folding？
I am not familiar with gcc,this code:
int a=145;int b=267+a;
printf("%d/n",b);
if comiple with no optimize option,will use the "addl" command,but if open O2 option,it will show me ...
- 1,735
0
votes
1
answer
204
views
DWARF output differs between compiler versions
I compile an ELF file including debug info and then read the debug-info back and parse it to get a list of all variables (global variables and class variables) in the RAM. This already works with the ...
- 11
1
vote
1
answer
86
views
Why doesn't the generated GCC code for array allocation and access follow the `movl (%rdx, %rcx, 4), %eax` formula?
I am thinking about how to compile arrays in x86-64 assembly.
I am reading "Computer Systems - A Programmer's Perspective" and the authors give the following formula:
E[i] -> `movl (%rdx,...
- 149