[Bug c++/95516] New: [coroutines] incorrect warning "'coro.gro' is used uninitialized"

[Bug c++/95516] New: [coroutines] incorrect warning "'coro.gro' is used uninitialized"

[bruck.michael at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95516

            Bug ID: 95516
           Summary: [coroutines] incorrect warning "'coro.gro' is used
                    uninitialized"
           Product: gcc
           Version: 10.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: bruck.michael at gmail dot com
  Target Milestone: ---

#include <coroutine>

struct dummy
{
    struct promise_type
    {
        dummy get_return_object() const noexcept { return {}; }
        std::suspend_never initial_suspend() const noexcept { return {}; }
        std::suspend_never final_suspend() const noexcept { return {}; }
        void return_void() const noexcept {}
        void unhandled_exception() const noexcept {}
    };
    // int i; // this kills the warning
};

dummy foo()
{ 
    co_return;
}

int main()
{
    foo();
}
---

<source>: In function 'dummy foo()':
<source>:19:1: warning: 'coro.gro' is used uninitialized in this function
[-Wuninitialized]
   19 | }
      | ^

[Bug c++/95516] [coroutines] incorrect warning "'coro.gro' is used uninitialized"

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95516

Iain Sandoe <iains at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2020-06-04
             Status|UNCONFIRMED                 |NEW

--- Comment #1 from Iain Sandoe <iains at gcc dot gnu.org> ---
Thanks for the report.

this is a kinda tricky situation;
I have also seen this occur (in larger test cases)

1) the coro.gro variable is an internal one to the compiler (it is also marked
as such). So the user can't "fix" the problem...

2) coro.gro is always initialised unconditionally in the same fragment of code
that creates it.

so it seems a false positive (actually, the example code posted here does not
fail for me on either 10.x or master) - however, as noted, I have seen this
occur.

- so this is a problem with the subsequent diagnostics processing or
optimisation.

[Bug c++/95516] [coroutines] incorrect warning "'coro.gro' is used uninitialized"

[bruck.michael at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95516

--- Comment #2 from Michael Bruck <bruck.michael at gmail dot com> ---
I was using -Ofast -Wall:
https://gcc.godbolt.org/z/pAkVS8

[Bug c++/95516] [coroutines] incorrect warning "'coro.gro' is used uninitialized"

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95516

Iain Sandoe <iains at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |iains at gcc dot gnu.org
   Target Milestone|---                         |10.2

--- Comment #3 from Iain Sandoe <iains at gcc dot gnu.org> ---
(In reply to Michael Bruck from comment #2)
> I was using -Ofast -Wall:
> https://gcc.godbolt.org/z/pAkVS8

I can repeat it with -Ofast.
[this still might have nothing to do with coroutines, but rather the tracking
of var usage, since the complaint is in the ramp function (which is a regular
c++ function)]

[Bug c++/95516] [coroutines] incorrect warning "'coro.gro' is used uninitialized"

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95516

--- Comment #4 from Iain Sandoe <iains at gcc dot gnu.org> ---
(In reply to Iain Sandoe from comment #3)
> (In reply to Michael Bruck from comment #2)
> > I was using -Ofast -Wall:
> > https://gcc.godbolt.org/z/pAkVS8
> 
> I can repeat it with -Ofast.
> [this still might have nothing to do with coroutines, but rather the
> tracking of var usage, since the complaint is in the ramp function (which is
> a regular c++ function)]

... and now it's gone again ...

[Bug c++/95516] [coroutines] incorrect warning "'coro.gro' is used uninitialized"

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95516

Iain Sandoe <iains at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |WORKSFORME
             Status|NEW                         |RESOLVED

--- Comment #5 from Iain Sandoe <iains at gcc dot gnu.org> ---
I cannot reproduce this any more, it would seem that the reorganisation of the
g-r-o and return code has made this false-positive disappear.

This test case is in my regular "additional" testing and it has not fired for
weeks.

please reopen if there's a repeatable case

in my opinion, user should not be presented with such warnings (when they can't
fix them - but that's a separate discussion).