[llvm-commits] [llvm] r134018 - in /llvm/trunk: lib/Target/X86/X86FastISel.cpp lib/Target/X86/X86FloatingPoint.cpp lib/Target/X86/X86ISelLowering.cpp lib/Target/X86/X86InstrFPStack.td lib/Target/X86/X86RegisterInfo.cpp lib/Target/X86/X86RegisterInfo.td test/CodeGen/X86/inline-asm-fpstack.ll
Jakob Stoklund Olesen
stoklund at 2pi.dk
Fri Jul 1 14:29:42 PDT 2011
On Jul 1, 2011, at 1:53 PM, Chad Rosier wrote:
>
> On Jul 1, 2011, at 11:34 AM, Jakob Stoklund Olesen wrote:
>
>>
>> On Jun 30, 2011, at 1:33 PM, Chad Rosier wrote:
>>
>>> Hi Jakob,
>>> I'm seeing a failure for one of the clang llvm-gcc tests. Bisect brought me to this revision, which seems very relevant.
>>> gcc.target/i386/pr30848.c (test for excess errors)
>>> mcrosier$ /Users/mcrosier/llvm-clean/install/bin/clang /Users/mcrosier/gcc-4_2-testsuite/src/gcc.target/i386/pr30848.c -ansi -pedantic-errors -fno-show-column -S -o pr30848.s
>>> fatal error: error in backend: Inline asm fixed outputs must be last on the x87 stack
>>
>> Clang is failing correctly, but I fixed the message thusly:
>>
>> clang -c /d/g/clang-tests/gcc-4_2-testsuite/src/gcc.target/i386/pr30848.c
>> fatal error: error in backend: Inline asm output regs must be last on the x87 stack
>>
>> However, DejaGNU is still failing the test:
>>
>> $ cat src/gcc.target/i386/pr30848.c
>> /* { dg-do compile } */
>>
>> void foo(double d)
>> {
>> __asm__ ("" : "=u" (d)); /* { dg-error "output regs" } */
>> }
>>
>> Anyone know how to fix the test suite?
>>
>
> I tried a few different regular expressions that I expected would work, but it still failed miserably? Hopefully, someone with more DejaGNU experience can speak up.
>
> Chad
>
>> Here is gcc-4.2:
>>
>> $ gcc-4.2 -c /d/g/clang-tests/gcc-4_2-testsuite/src/gcc.target/i386/pr30848.c
>> /d/g/clang-tests/gcc-4_2-testsuite/src/gcc.target/i386/pr30848.c: In function ‘foo’:
>> /d/g/clang-tests/gcc-4_2-testsuite/src/gcc.target/i386/pr30848.c:5: error: output regs must be grouped at top of stack
Perhaps the problem is that clang doesn't print a line number. We can't currently do that.
/jakob
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20110701/a31e26c6/attachment.html>
More information about the llvm-commits
mailing list