[Bug other/52646] New: Clang LLVM's __attribute__((naked)) for GCC when compiling for x86

[Bug other/52646] New: Clang LLVM's __attribute__((naked)) for GCC when compiling for x86

[th020394 at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52646

             Bug #: 52646
           Summary: Clang LLVM's __attribute__((naked)) for GCC when
                    compiling for x86
    Classification: Unclassified
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P3
         Component: other
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: th020394@gmail.com


It would be helpful to system developers if GCC had something similar to Clang
LLVM's __attribute__((naked)) for x86 targets. There are cases, such as when
interfacing with assembly or when handling an interrupt, when the code that
precedes and follows a normal function causes problems.

GCC already has __attribute__((interrupt)) for multiple targets, but doesn't
implement it for x86. Personally, if the feature were to be implemented, I
believe naked is more explicit.

I'd be glad to do this myself, but I'm not acquainted with GCC at all. What
files should I look in to get started?

[Bug other/52646] Clang LLVM's __attribute__((naked)) for GCC when compiling for x86

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52646

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

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

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> 2012-03-21 00:32:40 UTC ---
What is the reasoning for needing the naked attribute?  For me it seems like a
hack to add naked/interrupt attribute support for any target really.

*** This bug has been marked as a duplicate of bug 25967 ***

[Bug other/52646] Clang LLVM's __attribute__((naked)) for GCC when compiling for x86

[th020394 at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52646

--- Comment #2 from Tyler Hardin <th020394 at gmail dot com> 2012-03-21 00:44:39 UTC ---
While I admit my knowledge of interrupts on ARM, AVR, CR16, Epiphany, M32C,
M32R/D, m68k, MeP, MIPS, RL78, RX and Xstormy16 is limited; I don't see why, if
GCC supports interrupts on all of them, then it shouldn't have support for
others. Is there something special about them?

[Bug other/52646] Clang LLVM's __attribute__((naked)) for GCC when compiling for x86

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52646

Steven Bosscher <steven at gcc dot gnu.org> changed:

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

--- Comment #3 from Steven Bosscher <steven at gcc dot gnu.org> 2012-03-21 00:58:42 UTC ---
They are embedded targets, where naked functions are used more frequently, and
someone cared enough to implement it for those targets? Otherwise, nothing
special, I suppose.

[Bug other/52646] Clang LLVM's __attribute__((naked)) for GCC when compiling for x86

[th020394 at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52646

--- Comment #4 from Tyler Hardin <th020394 at gmail dot com> 2012-03-21 01:07:20 UTC ---
If I care enough to implement them for x86, do you believe it would be included
in GCC? If not, I won't waste my time with it (while it's worth my time to
implement it, it's not worth my time to maintain a fork that has my patch).

If you do, then I would really appreciate a hint about where to start. I
realize it may be obscure, so I understand if no one has a clue.