[Bug target/85927] ud2 instruction generated starting with gcc 8

[Bug target/85927] ud2 instruction generated starting with gcc 8

[pskocik at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85927

Petr Skocik <pskocik at gmail dot com> changed:

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

--- Comment #5 from Petr Skocik <pskocik at gmail dot com> ---
I think it'd be more welcome if gcc just put nothing there like clang does.

[Bug target/85927] ud2 instruction generated starting with gcc 8

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85927

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

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

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
The ud2 is there since naked has been implemented for x86 in PR25967.
It is true it is a kind of debugging aid, so perhaps more suitable for
sanitization or if we add some option to decide preferred behavior on e.g.
__builtin_unreachable or for detected UB in the code (whether to use
__builtin_unreachable with its current behavior, just assume it doesn't happen
and don't emit anything in there, or whether to act as __builtin_trap etc.).
But without it e.g. the reporter wouldn't know so easily the function is
invalid and won't work correctly.