[llvm-bugs] [Bug 35352] New: RTTI cxxabi pointers are 'external global i8*', but are referenced using 'getelementptr inbounds' with index '2'

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=35352

            Bug ID: 35352
           Summary: RTTI cxxabi pointers are 'external global i8*', but
                    are referenced using 'getelementptr inbounds' with
                    index '2'
           Product: clang
           Version: trunk
          Hardware: PC
                OS: All
            Status: NEW
          Severity: normal
          Priority: P
         Component: LLVM Codegen
          Assignee: unassignedclangbugs at nondot.org
          Reporter: ahmed.bougacha at gmail.com
                CC: llvm-bugs at lists.llvm.org

Take a simple class:

  class A { virtual void f1(); };
  void A::f1() { }

With RTTI enabled, we emit a reference to the libcxxabi RTTI base:

  @_ZTVN10__cxxabiv117__class_type_infoE = external global i8*

and later use it in a GEP:

  @_ZTI1A = constant ... { ... i8** getelementptr inbounds (i8*, i8**
@_ZTVN10__cxxabiv117__class_type_infoE, i64 2) to i8*), ... }

Is that correct?  As far as IR is concerned, the global is only 8-bytes wide. 
LangRef says:

  If the inbounds keyword is present, the result value of the getelementptr is
a poison value if the base pointer is not an in bounds address of an allocated
object

The object isn't allocated in this module, and irgen knows that the allocated
object is large enough.   But still, it seems fishy.  I haven't found anything
conclusive either way in the GEP/pointer aliasing rules.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20171117/57d43fbb/attachment.html>