[Bug symtab/26061] New: Wrong debug info for argc at -O2

[Bug symtab/26061] New: Wrong debug info for argc at -O2

[yangyibiao at hust dot edu.cn]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

            Bug ID: 26061
           Summary: Wrong debug info for argc at -O2
           Product: gdb
           Version: HEAD
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: symtab
          Assignee: unassigned at sourceware dot org
          Reporter: yangyibiao at hust dot edu.cn
  Target Milestone: ---

Created attachment 12577
  --> https://sourceware.org/bugzilla/attachment.cgi?id=12577&action=edit
a.out

$ gcc --version
gcc (GCC) 10.0.1 20200419 (experimental)
Copyright (C) 2020 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

yibiao:~/Debugger$ gdb --version
GNU gdb (GDB) 10.0.50.20200517-git
Copyright (C) 2020 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law.


Consider test-case:
...
$ cat small.c
#include <stdio.h>

int main(int argc, char **argv) {
  char buf[6];
  char c[] = "abc";
  sprintf(buf, (char *)c, 1);

  return 0;
}
...

$ gcc -O2 -g small.c; gdb -q a.out
Reading symbols from a.out...
(gdb) b main
Breakpoint 1 at 0x401040: file small.c, line 5.
(gdb) r
Starting program: /home/yibiao/Debugger/a.out 

Breakpoint 1, main (argc=1, argv=0x7fffffffdff8) at small.c:5
5         char c[] = "abc";
(gdb) info args argc
argc = 1
(gdb) step
6         sprintf(buf, (char *)c, 1);
(gdb) info args argc
argc = -8454
(gdb) 


/**************************************
We can find that at line 5, the value of argc is 1. When stepping to line 6,
the value of argc is changed to -8454. When stepping with stepi, the value of
argc is correct.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[yangyibiao at hust dot edu.cn]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

Yibiao Yang <yangyibiao at hust dot edu.cn> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tankut.baris.aktemur@intel.
                   |                            |com

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[tankut.baris.aktemur at intel dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

--- Comment #1 from Baris Aktemur <tankut.baris.aktemur at intel dot com> ---
I used gcc-10 ((Ubuntu 10.1.0-2ubuntu1~18.04) 10.1.0) and gdb
(10.0.50.20200525-git), but the scenario did not reproduce.

My initial guess is that it's related to the debug info. I'd suggest printing
the value of argc in hex (e.g. `(gdb) p/x -8454` gives 0xffffdefa) and also
printing the address of local variables, in particular `buf`. Also, doing stepi
multiple times and checking the values each time an instruction is stepped may
help narrow down the problem.

Judging from the title, I conclude you also think that the problem is related
to the debug info. In that case, I believe it would fit better to report this
to gcc, rather than gdb.

PS: Just out of curiosity, I wonder why you CC'ed me to this bug report.
Thanks.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[yangyibiao at hust dot edu.cn]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

--- Comment #2 from Yibiao Yang <yangyibiao at hust dot edu.cn> ---
(In reply to Baris Aktemur from comment #1)
> I used gcc-10 ((Ubuntu 10.1.0-2ubuntu1~18.04) 10.1.0) and gdb
> (10.0.50.20200525-git), but the scenario did not reproduce.
> 
> My initial guess is that it's related to the debug info. I'd suggest
> printing the value of argc in hex (e.g. `(gdb) p/x -8454` gives 0xffffdefa)
> and also printing the address of local variables, in particular `buf`. Also,
> doing stepi multiple times and checking the values each time an instruction
> is stepped may help narrow down the problem.
> 
> Judging from the title, I conclude you also think that the problem is
> related to the debug info. In that case, I believe it would fit better to
> report this to gcc, rather than gdb.
> 
> PS: Just out of curiosity, I wonder why you CC'ed me to this bug report.
> Thanks.

Thank you very much for your suggestion.

I CC'ed it to you as I found you are mentioned in a prior bug relate to this
one.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[yangyibiao at outlook dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

Yibiao Yang <yangyibiao at outlook dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
                 CC|                            |yangyibiao at outlook dot com
             Status|UNCONFIRMED                 |RESOLVED

--- Comment #3 from Yibiao Yang <yangyibiao at outlook dot com> ---
This problem is no longer exists in the latest version of gdb.

$ gdb --version
GNU gdb (GDB) 11.0.50.20201224-git

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[yangyibiao at outlook dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

Yibiao Yang <yangyibiao at outlook dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|FIXED                       |---
             Status|RESOLVED                    |UNCONFIRMED

--- Comment #4 from Yibiao Yang <yangyibiao at outlook dot com> ---
Not sure whether this bug is fixed in the latest version.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[yangyibiao at nju dot edu.cn]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

Yibiao Yang <yangyibiao at nju dot edu.cn> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tromey at sourceware dot org

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[sam at gentoo dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

Sam James <sam at gentoo dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sam at gentoo dot org

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[tromey at sourceware dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

Tom Tromey <tromey at sourceware dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |NOTABUG

--- Comment #5 from Tom Tromey <tromey at sourceware dot org> ---
I can reproduce this with the attached a.out and with git master gdb.
Here's an informative transcript:

(gdb) info addr argc
Symbol "argc" is multi-location:
  Range 0x401040-0x401061: a variable in $rdi
  Range 0x401061-0x401069: a complex DWARF expression:
     0: DW_OP_GNU_entry_value
       2: DW_OP_reg5 [$rdi]
     3: DW_OP_stack_value

.
(gdb) p $pc
$4 = (void (*)()) 0x401040 <main>
(gdb) p argc
$5 = 1
(gdb) step
6       in small.c
(gdb) p $pc
$6 = (void (*)()) 0x40105d <main+29>
(gdb) p argc
$7 = -7062
(gdb) p $rdi
$8 = 140737488348266
(gdb) p/x $rdi
$9 = 0x7fffffffe46a
(gdb) p/x argc
$10 = 0xffffe46a


Here you can see that the debuginfo claims that 'argc' is
held in '$rdi' for this PC range.  However, this isn't true --
rdi changes to some other value even though PC is still in
that range.

Based on this I think this is a compiler bug and not a gdb bug.
There's nothing gdb can do to correct this kind of problem.

I'm going to close this.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug symtab/26061] Wrong debug info for argc at -O2

[yangyibiao at nju dot edu.cn]

https://sourceware.org/bugzilla/show_bug.cgi?id=26061

--- Comment #6 from Yibiao Yang <yangyibiao at nju dot edu.cn> ---
Thanks for the detailed explaination, Tom.

I update gcc to the latest trunk and found this problem have been fixed in the
gcc side.

-- 
You are receiving this mail because:
You are on the CC list for the bug.