[Bug c/103492] New: 2 * new warnings in clang build

[Bug c/103492] New: 2 * new warnings in clang build

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

            Bug ID: 103492
           Summary: 2 * new warnings in clang build
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
  Target Milestone: ---

../../trunk.git/gcc/fold-const-call.c:432:1: warning: non-void function does
not return a value in all control paths [-Wreturn-type]
../../trunk.git/gcc/fold-const-call.c:465:1: warning: non-void function does
not return a value in all control paths [-Wreturn-type]

It looks like clang is being a bit keen. 

Suggest add a simple return false at the end of a couple of functions.
Either that or start decorating the switch statements in some way.

[Bug c/103492] 2 * new warnings in clang build

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Given that the switch uses arg->cl expression which is an unsigned int cl : 2
bitfield and the switch has all the 4 values covered, that clearly looks like a
clang bug.
The gcc_unreachable () has been removed because of the -Wunreachable-code
warnings...

[Bug c/103492] 2 * new warnings in clang build

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

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

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

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The gcc_unreachable called was explicitly removed in r12-5576-ga3b31fe3692894e8
. The bug is obvious in clang where the switch only can have 4 values as it is
a load from a bitfield of size of 2 bits.

[Bug c/103492] 2 * new warnings in clang build

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |marxin at gcc dot gnu.org

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 103575 has been marked as a duplicate of this bug. ***

[Bug c/103492] 2 * new warnings in clang build

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

--- Comment #4 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #2)
> The gcc_unreachable called was explicitly removed in
> r12-5576-ga3b31fe3692894e8 . The bug is obvious in clang where the switch
> only can have 4 values as it is a load from a bitfield of size of 2 bits.

All right! Is there any Clang bug report for this, or should I create it?

[Bug c/103492] 2 * new warnings in clang build

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

--- Comment #5 from David Binderman <dcb314 at hotmail dot com> ---
(In reply to Martin Liška from comment #4)
> All right! Is there any Clang bug report for this, or should I create it?

Option 2, I think. 

I have a 0.0 % success rate at getting any response to any clang bug I report
;-<

[Bug c/103492] 2 * new warnings in clang build

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

--- Comment #6 from Martin Liška <marxin at gcc dot gnu.org> ---
I'm going to do that, but right now they migrate bugzilla to GitHub issues..

[Bug c/103492] 2 * new warnings in clang build

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

--- Comment #7 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
reduced testcase for clang's warning:
struct f
{
    unsigned a:2;
    unsigned b:(32-2);
};

int g(struct f *b)
{
    switch (b->a)
    {
        case 0:
          return 2;
        case 1:
          return 3;
        case 2:
          return 5;
        case 3:
          return 1;
    }
}

[Bug c/103492] 2 * new warnings in clang build

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103492

--- Comment #8 from Martin Liška <marxin at gcc dot gnu.org> ---
Even simpler test-case:

struct A
{
  unsigned int val: 1;
};

int foo(struct A *a)
{
  switch (a->val)
  {
    case 0:
      return 123;
    case 1:
      return 456;
  }
}