[Bug debug/90717] New: wrong stmt location for breakpoint, XFAIL gcc.dg/guality/pr90716.c -flto -fuse-linker-plugin

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90717

            Bug ID: 90717
           Summary: wrong stmt location for breakpoint, XFAIL
                    gcc.dg/guality/pr90716.c -flto -fuse-linker-plugin
           Product: gcc
           Version: 10.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: rguenth at gcc dot gnu.org
  Target Milestone: ---

The gcc.dg/guality/pr90716.c testcase fails with -O2 -fwhole-program,
expanding from

main ()
{
  <bb 2> [local count: 17041817]:
  [t.c:12:3] # DEBUG BEGIN_STMT
  [t.c:13:3] # DEBUG BEGIN_STMT
  [t.c:13:5] # DEBUG b => 0
  [t.c:14:3] # DEBUG BEGIN_STMT
  # DEBUG b => 0
  [t.c:14:10] # DEBUG BEGIN_STMT
  # DEBUG j => 0
  [t.c:16:14] # DEBUG BEGIN_STMT
... unrolled loop ...
  [t.c:16:14] # DEBUG BEGIN_STMT
  [t.c:17:2] # DEBUG BEGIN_STMT
  [t.c:16:21] # DEBUG BEGIN_STMT
  # DEBUG j => 8
  [t.c:16:14] # DEBUG BEGIN_STMT
  [t.c:14:17] # DEBUG BEGIN_STMT
  # DEBUG b => 7
  [t.c:14:10] # DEBUG BEGIN_STMT
  [t.c:23:3] # DEBUG BEGIN_STMT
  [t.c:23:3] optimize_me_not ();
  [t.c:24:3] # DEBUG BEGIN_STMT
  return 0;

gdb puts the breakpoint at

Breakpoint 1, main ()
    at
/space/rguenther/src/svn/trunk2/gcc/testsuite/gcc.dg/guality/pr90716.c:23
23        optimize_me_not(); /* { dg-final { gdb-test . "j + 1" "9" } } */
(gdb) disassemble
Dump of assembler code for function main:
=> 0x00000000004003e0 <+0>:     xor    %eax,%eax
   0x00000000004003e2 <+2>:     callq  0x4004d0 <optimize_me_not>
   0x00000000004003e7 <+7>:     xor    %eax,%eax
   0x00000000004003e9 <+9>:     retq   

where

(gdb) p j
$1 = 0

but at the "correct" location the correct value is displayed.

(gdb) si
0x00000000004003e2      23        optimize_me_not(); /* { dg-final { gdb-test .
"j + 1" "9" } } */
(gdb) p j
$2 = 8

Somehow things must go wrong on the RTL / debuginfo creation level putting
the location views for 'j' on a assembler stmt associated with line 23.