https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88686
Bug ID: 88686
Summary: gcc generates wrong debug information at -O1
Product: gcc
Version: unknown
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: debug
Assignee: unassigned at gcc dot gnu.org
Reporter: qrzhang at gatech dot edu
Target Milestone: ---
I have reported this bug at gdb
bugzilla(https://sourceware.org/bugzilla/show_bug.cgi?id=24032). However, it
appears to be a gcc bug. I can reproduce the bug with "gcc version 9.0.0
20190103 (experimental) [trunk revision 267553] (GCC)". The gdb version does
not matter. I can reproduce it with gdb 7.7.
Below is my original report:
=====8<=========
$ gdb-trunk --version
GNU gdb (GDB) 8.2.50.20181223-git
*It correctly prints the value without optimization*
$ gcc-trunk -g abc.c outer.c
$ gdb-trunk -x cmds -batch a.out
Breakpoint 1 at 0x400485: file abc.c, line 8.
Breakpoint 1, main () at abc.c:8
8 optimize_me_not();
9 c = 0;
$1 = 0
Kill the program being debugged? (y or n) [answered Y; input not from terminal]
[Inferior 1 (process 2315) killed]
*It incorrectly prints "j=1" at -O1*
$ gcc-trunk -g -O1 abc.c outer.c
$ gdb-trunk -x cmds -batch a.out
Breakpoint 1 at 0x400491: file abc.c, line 8.
Breakpoint 1, main () at abc.c:8
8 optimize_me_not();
10 for (; c < 6; c++)
$1 = 1
Kill the program being debugged? (y or n) [answered Y; input not from terminal]
[Inferior 1 (process 2331) killed]
*Files used to reproduce*
$ cat abc.c
volatile int a, b;
int c;
int main() {
int j;
for (; b < 10; b++) {
j = 0;
for (; j < 2; j++) {
optimize_me_not();
c = 0;
for (; c < 6; c++)
a = b + c;
}
}
}
$ cat outer.c
optimize_me_not() {}
$ cat cmds
b 8
r
n
p j
k
q
=====8<=========
