[Bug debug/112372] New: GCC omits function location in DWARF at higher optimisation levels

[Bug debug/112372] New: GCC omits function location in DWARF at higher optimisation levels

[gprocida at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112372

            Bug ID: 112372
           Summary: GCC omits function location in DWARF at higher
                    optimisation levels
           Product: gcc
           Version: 13.2.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gprocida at google dot com
  Target Milestone: ---

Source:

void x(signed char c) { (void) c; }
void y(char c) { (void) c; }
void z(unsigned char c) { (void) c; } // not needed

GCC versions tested: 11.4.0, 12.3.0, 13.2.0
DWARF versions tested: 3, 4, 5
Optimisation levels tests: 0, 1, g (all good); 2, 3, fast, s (all bad)

At lower levels of optimisation, the 3 functions are described with
straightforward DIEs.

At higher levels of optimisation, they are... a mess.

x has abstract origin
y is missing location (PC)
z looks normal

The missing location means that binary analysis tools, like libabigail, cannot
associated the symbol y with its type.

FWIW, Clang does not have this issue.

Remove x: issue disappears.
Remove y: issue disappears.
Remove z: issue persists.

gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-linux-gnu/13/lto-wrapper
OFFLOAD_TARGET_NAMES=nvptx-none:amdgcn-amdhsa
OFFLOAD_TARGET_DEFAULT=1
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Debian 13.2.0-4'
--with-bugurl=file:///usr/share/doc/gcc-13/README.Bugs
--enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2 --prefix=/usr
--with-gcc-major-version-only --program-suffix=-13
--program-prefix=x86_64-linux-gnu- --enable-shared --enable-linker-build-id
--libexecdir=/usr/libexec --without-included-gettext --enable-threads=posix
--libdir=/usr/lib --enable-nls --enable-bootstrap --enable-clocale=gnu
--enable-libstdcxx-debug --enable-libstdcxx-time=yes
--with-default-libstdcxx-abi=new --enable-gnu-unique-object
--disable-vtable-verify --enable-plugin --enable-default-pie --with-system-zlib
--enable-libphobos-checking=release --with-target-system-zlib=auto
--enable-objc-gc=auto --enable-multiarch --disable-werror --enable-cet
--with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32
--enable-multilib --with-tune=generic
--enable-offload-targets=nvptx-none=/build/gcc-13-oyarai/gcc-13-13.2.0/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/gcc-13-oyarai/gcc-13-13.2.0/debian/tmp-gcn/usr
--enable-offload-defaulted --without-cuda-driver --enable-checking=release
--build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
--with-build-config=bootstrap-lto-lean --enable-link-serialization=28
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 13.2.0 (Debian 13.2.0-4) 

gcc -Wall -Wextra -o x.o -O2 -c -g x.c

[Bug debug/112372] GCC omits function location in DWARF at higher optimisation levels

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112372

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=65797,
                   |                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=63572

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Most likely a dup of bug 65797.

[Bug debug/112372] GCC omits function location in DWARF at higher optimisation levels

[gprocida at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112372

--- Comment #2 from Giuliano Procida <gprocida at google dot com> ---
The symbols are not aliased (which is what I thought might have happened with
very aggressive optimisations). If they had been aliased, it would be much
harder to give the symbols types.

0000000000000000 T x
0000000000000010 T y
0000000000000020 T z

Also:

Disassembly of section .text:

0000000000000000 <x>:
   0:   c3                      ret
   1:   66 66 2e 0f 1f 84 00    data16 cs nopw 0x0(%rax,%rax,1)
   8:   00 00 00 00 
   c:   0f 1f 40 00             nopl   0x0(%rax)

0000000000000010 <y>:
  10:   c3                      ret
  11:   66 66 2e 0f 1f 84 00    data16 cs nopw 0x0(%rax,%rax,1)
  18:   00 00 00 00 
  1c:   0f 1f 40 00             nopl   0x0(%rax)

0000000000000020 <z>:
  20:   c3                      ret

So it should have been entirely possible to assign the correct byte ranges
(low/high PC) to the 3 functions.

[Bug debug/112372] GCC omits function location in DWARF at higher optimisation levels

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112372

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Giuliano Procida from comment #2)
> The symbols are not aliased (which is what I thought might have happened
> with very aggressive optimisations). If they had been aliased, it would be
> much harder to give the symbols types.

They can't alias for language reasons that is `&y != &x` (and `&x != &z`) has
to be true. Anyways as mentioned in PR 65797, a thunk is being created and
going wrong. 

Yes maybe ICF should not touch empty functions ... But I doubt that was your
original testcase.

[Bug debug/112372] GCC omits function location in DWARF at higher optimisation levels

[gprocida at google dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112372

--- Comment #4 from Giuliano Procida <gprocida at google dot com> ---
The provided example was exactly my original test case.

It's part of our test suite for
https://android.googlesource.com/platform/external/stg/.

I'm currently running multiple versions of GCC and Clang over 100s of tests,
looking for discrepancies. There will probably be more bug reports.

[Bug ipa/112372] GCC omits function location in DWARF at higher optimisation levels

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112372

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|debug                       |ipa
                 CC|                            |hubicka at gcc dot gnu.org,
                   |                            |marxin at gcc dot gnu.org

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
Probably a duplicate of a recent similar bug also because of ICF.