[LLVMbugs] [Bug 7143] New: error: SSE register return with SSE disabled (with -mno-mmx)

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=7143

           Summary: error: SSE register return with SSE disabled (with
                    -mno-mmx)
           Product: new-bugs
           Version: trunk
          Platform: PC
        OS/Version: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: new bugs
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: nomura at netapp.com
                CC: llvmbugs at cs.uiuc.edu


Compiling floating-point code with -mno-mmx, clang aborts as follows:

[cyclsvl05]$ cat bug.i
double bar (void);
double foo()
{
  return bar()+1.0;
}

/u/nomura/dl/Installed/clang/bin/clang -c -O0 -mno-mmx -w bug.i
fatal error: error in backend: SSE register return with SSE disabled


gcc would issue a similar error if -mno-sse were specified.  

But gcc with -mno-mmx appears to proceed happily using SSE registers,
for example:

[cyclsvl05]$ gcc -v
Using built-in specs.
Target: x86_64-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--enable-checking=release --with-system-zlib --enable-__cxa_atexit
--disable-libunwind-exceptions --enable-libgcj-multifile
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada --enable-java-awt=gtk
--disable-dssi --enable-plugin
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre --with-cpu=generic
--host=x86_64-redhat-linux
Thread model: posix
gcc version 4.1.2 20080704 (Red Hat 4.1.2-46)

[cyclsvl05]$ gcc -O0 -mno-mmx -S bug.i
[cyclsvl05]$ 

.LCFI0:
        movq    %rsp, %rbp
.LCFI1:
        subq    $16, %rsp
.LCFI2:
        call    bar
        movapd  %xmm0, %xmm1
        movsd   .LC0(%rip), %xmm0
        addsd   %xmm1, %xmm0
        movsd   %xmm0, -8(%rbp)
        movq    -8(%rbp), %rax
        movq    %rax, -8(%rbp)
        movsd   -8(%rbp), %xmm0
        leave
        ret

I'm using a sync of clang+llvm from a couple days ago.

Is clang doing the right thing by aborting?  (If so, what clang options
would you suggest to obtain compatibility with gcc and -mno-mmx?)  Thanks.

-- 
Configure bugmail: http://llvm.org/bugs/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.