[Bug tree-optimization/55415] New: Early SRA produces unaligned complex types

[rth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=55415



             Bug #: 55415

           Summary: Early SRA produces unaligned complex types

    Classification: Unclassified

           Product: gcc

           Version: unknown

            Status: UNCONFIRMED

          Severity: normal

          Priority: P3

         Component: tree-optimization

        AssignedTo: unassig...@gcc.gnu.org

        ReportedBy: r...@gcc.gnu.org





While looking at PR55403 one has to wonder where the unaligned

complex long double type came from in the first place:



#0  copy_node_stat (node=0x7ffff1aeb9d8) at ../../git-master/gcc/tree.c:961

#1  0x0000000000e33162 in build_distinct_type_copy (type=0x7ffff1aeb9d8)

    at ../../git-master/gcc/tree.c:5856

#2  0x0000000000e332d5 in build_variant_type_copy (type=0x7ffff1aeb9d8)

    at ../../git-master/gcc/tree.c:5890

#3  0x0000000000e33119 in build_aligned_type (type=0x7ffff1aeb9d8, align=8)

    at ../../git-master/gcc/tree.c:5842

#4  0x0000000000a39b7d in ipa_modify_call_arguments (cs=0x7ffff11493a8, 

    stmt=0x7ffff1147428, adjustments=...)

    at ../../git-master/gcc/ipa-prop.c:2967



I'm still not quite sure why ipa_modify_call_arguments *ever* wants to

create misaligned types as function interfaces?