[LLVMbugs] [Bug 11860] New: ARM: floating-point arguments of variadic functions passed incorrectly with AAPCS-VFP ABI

[bugzilla-daemon at llvm.org]

http://llvm.org/bugs/show_bug.cgi?id=11860

             Bug #: 11860
           Summary: ARM: floating-point arguments of variadic functions
                    passed incorrectly with AAPCS-VFP ABI
           Product: libraries
           Version: trunk
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: Backend: ARM
        AssignedTo: unassignedbugs at nondot.org
        ReportedBy: mans at mansr.com
                CC: llvmbugs at cs.uiuc.edu
    Classification: Unclassified


Created attachment 7948
  --> http://llvm.org/bugs/attachment.cgi?id=7948
Fix vararg calls with float arguments under AAPCS-VFP ABI

Calls to variadic functions should always use the base AAPCS variant, even with
AAPCS-VFP is in use.  This is currently broken as demonstrated by this test:

float foo(float x, ...);
float bar(void)
{
    return foo(0.0);
}

Using -mfloat-abi hard, this compiles to the following:

00000000 <bar>:
   0:   e92d4800        push    {r11, lr}
   4:   ed9f0a01        vldr    s0, [pc, #4]    ; 10 <bar+0x10>
   8:   ebfffffe        bl      0 <foo>
   c:   e8bd8800        pop     {r11, pc}
  10:   00000000        andeq   r0, r0, r0

The attached patch fixes the problem.

-- 
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.