[Bug c/114657] New: Invalid type conversion from some _BitInt bit-fields

[Bug c/114657] New: Invalid type conversion from some _BitInt bit-fields

[juuso.alasuutari at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114657

            Bug ID: 114657
           Summary: Invalid type conversion from some _BitInt bit-fields
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: juuso.alasuutari at gmail dot com
  Target Milestone: ---

Created attachment 57911
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57911&action=edit
Preprocessed test program source

Summary
-------

GCC's type conversion from some bit-precise integer-typed bit-fields
does not match the description in the C23 standard.


Expected vs. observed behavior
------------------------------

Based on footnote on page 47 of ISO/IEC 9899/2024 6.3.1.1 (n3320.pdf),
the converted type of a _BitInt bit-field should match the original
bit-precise integer:

  "E.g. unsigned _BitInt(7): 2 is a bit-field that can hold the
   values 0, 1, 2, 3, and converts to unsigned _BitInt(7)."

My expectation would then be to see the following output, which does
in fact happen when compiling with clang-19:

  $ ./test
  unsigned _BitInt(7)

This is what GCC does:

  $ ./test
  unsigned _BitInt(2)


Output of gcc-14 -v
-------------------

Using built-in specs.
COLLECT_GCC=gcc-14
COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-linux-gnu/14/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 14-20240330-1'
--with-bugurl=file:///usr/share/doc/gcc-14/README.Bugs
--enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2,rust --prefix=/usr
--with-gcc-major-version-only --program-suffix=-14
--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-clocale=gnu --enable-libstdcxx-debug
--enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new
--enable-libstdcxx-backtrace --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/reproducible-path/gcc-14-14-20240330/debian/tmp-nvptx/usr,amdgcn-amdhsa=/build/reproducible-path/gcc-14-14-20240330/debian/tmp-gcn/usr
--enable-offload-defaulted --without-cuda-driver
--enable-checking=yes,extra,rtl --build=x86_64-linux-gnu
--host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 14.0.1 20240330 (experimental) [master r14-9728-g6fc84f680d0]
(Debian 14-20240330-1) 


Compilation command
-------------------

gcc-14 -std=gnu23 -Wall -Wextra -Wpedantic -save-temps -o test test.c

[Bug c/114657] Invalid type conversion from some _BitInt bit-fields

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114657

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org,
                   |                            |jsm28 at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
I believe Joseph said this isn't well specified in the C standard, see
https://gcc.gnu.org/pipermail/gcc-patches/2023-July/625765.html
If you use something like + 0uwb it should be the type after the integral
promotions and so what the standard specifies for integral promotions of the
_BitInt bit-fields.

[Bug c/114657] Invalid type conversion from some _BitInt bit-fields

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114657

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
In particular, 6.5.1.1/2 says
"The type of the controlling expression is the type of the expression as if it
had undergone an lvalue conversion, array to pointer conversion, or function to
pointer conversion."
but doesn't list integer promotions (if those were to occur, you'd e.g. never
be able to match a char, signed char, unsigned char, short, unsigned short
types with _Generic because everything would be promoted to int.

[Bug c/114657] Invalid type conversion from some _BitInt bit-fields

[jsm28 at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114657

--- Comment #3 from Joseph S. Myers <jsm28 at gcc dot gnu.org> ---
https://www.open-std.org/jtc1/sc22/wg14/www/docs/n2958.htm has my analysis of
the various notions of "type" used in relation to bit-fields and the questions
of what expressions are considered to have special properties associated with
referring to a bit-field.