Comment 7 for bug 910363

It's a bug in IPA-SRA that creates non-mode-size stores:

void llvm::Type::_ZN4llvm4Type15setSubclassDataEj.clone.1(unsigned int:24*, unsigned int) (<unnamed-unsigned:24> * ISRA.6, unsigned int val)
{
...
<bb 2>:
  D.87358_2 = (<unnamed-unsigned:24>) val_1(D);
  *ISRA.6_8(D) = D.87358_2;

I think this has been fixed in 4.6 (not on the 4.5 branch though) which
no longer performs this substitution. You can work around this using
-fno-ipa-sra.

The following is a simplified testcase:

extern "C" void abort (void);
struct S
{
  void __attribute__((noinline)) set(unsigned val)
    {
      data = val;
      if (data != val)
        abort ();
    }
  int pad0;
  unsigned pad1 : 8;
  unsigned data : 24;
  int pad2;
};
int main()
{
  S s;
  s.pad2 = -1;
  s.set(0);
  if (s.pad2 != -1)
    abort ();
}

Where 4.6 says:

Candidate (2069): this
! Disqualifying this - Encountered a bit-field access.

which hints at what needs backporting.

Martin?