[LLVMbugs] [Bug 7327] New: failed assertion while building SPEC CPU2006's 481.wrf in dragonegg-2.7 and llvm-gcc-4.2-2.7

bugzilla-daemon at llvm.org bugzilla-daemon at llvm.org
Wed Jun 9 03:03:46 PDT 2010


http://llvm.org/bugs/show_bug.cgi?id=7327

           Summary: failed assertion while building SPEC CPU2006's 481.wrf
                    in dragonegg-2.7 and llvm-gcc-4.2-2.7
           Product: dragonegg
           Version: 2.7
          Platform: PC
        OS/Version: Linux
            Status: NEW
          Severity: normal
          Priority: P
         Component: New Bugs
        AssignedTo: baldrick at free.fr
        ReportedBy: kenneth.hoste at ugent.be
                CC: llvmbugs at cs.uiuc.edu


Created an attachment (id=4993)
 --> (http://llvm.org/bugs/attachment.cgi?id=4993)
minimized testcase taken from 481.wrf (original was 10k lines, now ~1k)

While building the 481.wrf SPEC CPU2006 benchmark using dragonegg-2.7 (in
combination with gcc-4.5.0), I'm running into the following assertion failure:

gfortran -fplugin=dragonegg.so -c -o module_ra_gfdleta.fppized.o -I.
-I./netcdf/include -O2 module_ra_gfdleta.fppized.f90

f951: /path/to/LLVM/2.7/dragonegg-2.7/llvm-convert.cpp:8559: static
llvm::Constant* TreeConstantToLLVM::ConvertUnionCONSTRUCTOR(tree_node*):
Assertion `(VEC_constructor_elt_base_length(((elt) ? &(elt)->base : 0))) == 1
&& "Union CONSTRUCTOR with multiple elements?"' failed.

I ran into the same problem using llvm-gcc-4.2-2.7: 

gfortran -c -o module_ra_gfdleta.fppized.o -I. -I./netcdf/include -O2
module_ra_gfdleta.fppized.f90

f951: ../../llvm-gcc-4.2-2.7.source/gcc/llvm-convert.cpp:8207: static
llvm::Constant* TreeConstantToLLVM::ConvertUnionCONSTRUCTOR(tree_node*):
Assertion `(VEC_constructor_elt_base_length(((elt) ? &(elt)->base : 0))) == 1
&& "Union CONSTRUCTOR with multiple elements?"' failed.

A minimized testcase which should allow to reproduce is attached.

-- 
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.



More information about the llvm-bugs mailing list