[Bug target/13297] New: in x86 target -msoft_float should imply -mno-fp-ret-in-387

[Bug target/13297] New: in x86 target -msoft_float should imply -mno-fp-ret-in-387

[john32979 at hotmail dot com]

If -msoft_float is used to generate hooks for linking to FP emulation library, 
the results and arguments and results are still passed via the x87 stack if -
mno-fp-ret-in-387 is not defined as well.  This will generate invalid code for 
many embedded x86 platforms that do not have an FPU (e.g. AMD SC420).

-msoft_float should imply -mno-fp-ret-in-387 to avoid confusion.  At the bare 
minimum, a note regarding switching on -mno-fp-ret-in-387 in conjunction with -
msoft_float when developing on FPU-less platforms should be mentioned.

-- 
           Summary: in x86 target -msoft_float should imply -mno-fp-ret-in-
                    387
           Product: gcc
           Version: 3.2.1
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: P2
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: john32979 at hotmail dot com
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13297

[Bug target/13297] in x86 target -msoft_float should imply -mno-fp-ret-in-387

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2003-12-04 08:38 -------
The manual already warns about this:
-msoft-float 
.....
On machines where a function returns floating point results in the 80387 register stack, some 
floating point opcodes may be emitted even if -msoft-float is used. 

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=13297