[Bug target/64008] New: [SH] sh4-linux configured compiler rejects -m4-nofpu

[Bug target/64008] New: [SH] sh4-linux configured compiler rejects -m4-nofpu

[olegendo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64008

            Bug ID: 64008
           Summary: [SH] sh4-linux configured compiler rejects -m4-nofpu
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: olegendo at gcc dot gnu.org
                CC: kkojima at gcc dot gnu.org, segher at gcc dot gnu.org
            Target: sh*-*-*

This issue has been mentioned by Segher while he was trying to build an SH4
linux kernel on a powerpc host.  Segher, could you please add the configure
options you have used for that?

Looking at the SUPPORT_SH* macros in sh.h starting at line 161, it's not
entirely clear to me what those are trying to accomplish.

For instance:

#if SUPPORT_SH1
#define SUPPORT_SH2 1
#endif

... to me reads that when SH1 support is enabled (-m1 option), then support for
SH2 (-m2 option) is also enabled.  This looks like it should be the other way
around.

Kaz, do you have any idea why these SUPPORT_SH* macros are needed?  Why isn't
just every CPU/FPU type marked as supported?

[Bug target/64008] [SH] sh4-linux configured compiler rejects -m4-nofpu

[kkojima at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64008

--- Comment #1 from Kazumoto Kojima <kkojima at gcc dot gnu.org> ---
(In reply to Oleg Endo from comment #0)
> Kaz, do you have any idea why these SUPPORT_SH* macros are needed?  Why
> isn't just every CPU/FPU type marked as supported?

I have no idea for those macros.  Some historical reason?

[Bug target/64008] [SH] sh4-linux configured compiler rejects -m4-nofpu

[olegendo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64008

Oleg Endo <olegendo at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |iwamatsu at nigauri dot org

--- Comment #2 from Oleg Endo <olegendo at gcc dot gnu.org> ---
I guess the SUPPORT_* macros are there to prevent scenarios like:

- SH4 machine with a GCC configured to produce SH4 code (with FPU, double
precision by default) without any -m options.

- No soft-fpu support code nor other multilib support code is present on the
SH4 machine

- Compiling a program with -m2 would produce SH2 code which might use soft-fpu
support code and various libgcc functions (e.g. dynamic shifts)

- Link error

Looking at config.gcc, it seems that specifying --enable-incomplete-targets
solves the problem, as it effectively enables all the target options.

Some related links:

https://lkml.org/lkml/2014/7/8/732
http://lists.openwall.net/linux-kernel/2014/07/08/1071

and a patch from somewhere else that seems related:

http://cgit.openembedded.org/openembedded/plain/recipes/gcc/gcc-4.5/sh4-multilib.patch

Maybe Nobuhiro has some comments regarding this issue...

[Bug target/64008] [SH] sh4-linux configured compiler rejects -m4-nofpu

[kkojima at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64008

--- Comment #3 from Kazumoto Kojima <kkojima at gcc dot gnu.org> ---
(In reply to Oleg Endo from comment #2)
> and a patch from somewhere else that seems related:
> 
> http://cgit.openembedded.org/openembedded/plain/recipes/gcc/gcc-4.5/sh4-
> multilib.patch

Perhaps I don't get the original problem.  The kernel folks used to
specify --with-multilib-list=m4,m4-nofpu during configuring sh4-linux
gcc.  Doesn't it work now?

[Bug target/64008] [SH] sh4-linux configured compiler rejects -m4-nofpu

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64008

--- Comment #4 from Segher Boessenkool <segher at gcc dot gnu.org> ---
I configured for sh4-linux, foolishly thinking that sh4-nofpu would
work with that as well.  Why not build all "regular" multilibs for
every "regular" config?  Maybe keep sh5, sh2a separate, I dunno (I
note that sh2a-nofpu is built for sh1 and sh2 configs, but not e.g.
sh3 or sh4, or sh2a itself!  Truly strange stuff.  I suspect it makes
sense in that it builds a minimal amount of multilibs and then supports
everything it can
with that, but that does not make terribly much sense from the viewpoint
of a user).

[Bug target/64008] [SH] sh4-linux configured compiler rejects -m4-nofpu

[kkojima at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64008

--- Comment #5 from Kazumoto Kojima <kkojima at gcc dot gnu.org> ---
(In reply to Segher Boessenkool from comment #4)

At least for sh4, it would have a historical reason.  In the old
time, -m4-nofpu confused many users (including me).  From its name,
those users expected that it'll give the same ABI for integer only
programs with -m4 and fallen into the pit.  Then no one complained
when the default configuration was -m4 only for sh4-linux gcc and
the kernel folks satisfied with --with-multilib-list=m4,m4-nofpu.

[Bug target/64008] [SH] sh4-linux configured compiler rejects -m4-nofpu

[olegendo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64008

--- Comment #6 from Oleg Endo <olegendo at gcc dot gnu.org> ---
Ah, ancient issue is ancient:
Here it starts: https://gcc.gnu.org/ml/gcc/2003-07/msg01851.html

Still, I think Segher has a point in comment #4.  Why not just enable all the
multilibs for any configuration?  At least it would be a friendly default
value.  If users don't want that (space limitations?), they can still disable
it with --disable-multilib.