[Bug debug/102442] New: Incorrect debug info for C89-style function parameter

[Bug debug/102442] New: Incorrect debug info for C89-style function parameter

[liyd2021 at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102442

            Bug ID: 102442
           Summary: Incorrect debug info for C89-style function parameter
           Product: gcc
           Version: 11.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: liyd2021 at gmail dot com
  Target Milestone: ---

Affected versions: gcc 11.1.0 with gdb (Ubuntu 20.04.2)

(terminal) $ cat simple.c && gcc -g -Og simple.c
#include <assert.h>

// float fx(float x)
float fx(x) float x;
{
    return x + 1.0;
}

float inita() { return 3.0; }

int main()
{
    float a = inita();
    assert(fx(a) == 4.0);
}
--------------------

(terminal) $ cat run.gdb
b fx
r

(terminal) $ gdb -x run.gdb a.out
Breakpoint 1, fx (x=0) at simple.c:5 <- **BUG** <- x should be 3.0

[Bug debug/102442] Incorrect debug info for C89-style function parameter

[josephcsible at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102442

Joseph C. Sible <josephcsible at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |josephcsible at gmail dot com

--- Comment #1 from Joseph C. Sible <josephcsible at gmail dot com> ---
I wonder if this is related to fx actually being a float(double) and not a
float(float) as it appears to be.

[Bug debug/102442] Incorrect debug info for C89-style function parameter

[egallager at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102442

Eric Gallager <egallager at gcc dot gnu.org> changed:

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

--- Comment #2 from Eric Gallager <egallager at gcc dot gnu.org> ---
That's not a C89 function parameter, that's a K&R-style parameter, which was
outdated even in C89. GCC warns about it:

$ /usr/local/bin/gcc -c -Wall -Wextra -Wstrict-prototypes -Wimplicit
-Wold-style-definition -Wold-style-declaration -Wmissing-prototypes
-Wmissing-declarations -pedantic -Wconversion -Wdouble-promotion
-Wtraditional-conversion 102442.c
102442.c:4:7: warning: function declaration isn't a prototype
[-Wstrict-prototypes]
    4 | float fx(x) float x;
      |       ^~
102442.c: In function 'fx':
102442.c:4:7: warning: old-style function definition [-Wold-style-definition]
102442.c:6:14: warning: implicit conversion from 'float' to 'double' to match
other operand of binary expression [-Wdouble-promotion]
    6 |     return x + 1.0;
      |              ^
102442.c:6:14: warning: conversion from 'double' to 'float' may change value
[-Wfloat-conversion]
    6 |     return x + 1.0;
      |            ~~^~~~~
102442.c: At top level:
102442.c:9:7: warning: function declaration isn't a prototype
[-Wstrict-prototypes]
    9 | float inita() { return 3.0; }
      |       ^~~~~
102442.c: In function 'inita':
102442.c:9:7: warning: old-style function definition [-Wold-style-definition]
102442.c: At top level:
102442.c:11:5: warning: function declaration isn't a prototype
[-Wstrict-prototypes]
   11 | int main()
      |     ^~~~
102442.c: In function 'main':
102442.c:11:5: warning: old-style function definition [-Wold-style-definition]
In file included from 102442.c:1:
102442.c:14:15: warning: implicit conversion from 'float' to 'double' when
passing argument to function [-Wdouble-promotion]
   14 |     assert(fx(a) == 4.0);
      |               ^
102442.c:14:18: warning: implicit conversion from 'float' to 'double' to match
other operand of binary expression [-Wdouble-promotion]
   14 |     assert(fx(a) == 4.0);
      |                  ^~
102442.c:14:5: warning: passing argument 1 of '__builtin_expect' with different
width due to prototype [-Wtraditional-conversion]
   14 |     assert(fx(a) == 4.0);
      |     ^~~~~~
102442.c:14:5: warning: passing argument 2 of '__builtin_expect' with different
width due to prototype [-Wtraditional-conversion]
$

As does clang:

$ clang -c -Weverything 102442.c
warning: include location '/usr/local/include' is unsafe for cross-compilation
[-Wpoison-system-directories]
102442.c:6:14: warning: implicit conversion loses floating-point precision:
'double' to 'float' [-Wimplicit-float-conversion]
    return x + 1.0;
    ~~~~~~ ~~^~~~~
102442.c:6:12: warning: implicit conversion increases floating-point precision:
'float' to 'double' [-Wdouble-promotion]
    return x + 1.0;
           ^ ~
102442.c:4:7: warning: no previous prototype for function 'fx'
[-Wmissing-prototypes]
float fx(x) float x;
      ^
102442.c:4:1: note: declare 'static' if the function is not intended to be used
outside of this translation unit
float fx(x) float x;
^
static 
102442.c:4:9: warning: this old-style function definition is not preceded by a
prototype [-Wstrict-prototypes]
float fx(x) float x;
        ^
102442.c:9:7: warning: no previous prototype for function 'inita'
[-Wmissing-prototypes]
float inita() { return 3.0; }
      ^
102442.c:9:1: note: declare 'static' if the function is not intended to be used
outside of this translation unit
float inita() { return 3.0; }
^
static 
102442.c:9:12: warning: this old-style function definition is not preceded by a
prototype [-Wstrict-prototypes]
float inita() { return 3.0; }
           ^
102442.c:14:12: warning: implicit conversion increases floating-point
precision: 'float' to 'double' [-Wdouble-promotion]
    assert(fx(a) == 4.0);
           ^~~~~ ~~
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/assert.h:93:25:
note: expanded from macro 'assert'
    (__builtin_expect(!(e), 0) ? __assert_rtn(__func__, __FILE__, __LINE__, #e)
: (void)0)
                        ^
102442.c:14:15: warning: implicit conversion increases floating-point
precision: 'float' to 'double' [-Wdouble-promotion]
    assert(fx(a) == 4.0);
           ~~ ^
/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/assert.h:93:25:
note: expanded from macro 'assert'
    (__builtin_expect(!(e), 0) ? __assert_rtn(__func__, __FILE__, __LINE__, #e)
: (void)0)
                        ^
9 warnings generated.
$

I think the warnings make it pretty clear that code like this needs to be
updated.