It turns out that we emitted completely wrong common EH frame since the early beginni...
authorAnton Korobeynikov <asl@math.spbu.ru>
Mon, 15 Feb 2010 22:36:41 +0000 (22:36 +0000)
committerAnton Korobeynikov <asl@math.spbu.ru>
Mon, 15 Feb 2010 22:36:41 +0000 (22:36 +0000)
commitac8a3d0b5b464e4dd7c45ba111e6da1aeed39bc5
treec4ea542eabb85a355bbde7e6767142147658b13d
parent990a9fde39b7d9ff7af8beb51752c16f53c0b3d3
It turns out that we emitted completely wrong common EH frame since the early beginning! The "z" CIE augmentation mark indicates that there is an extra word containing the size of augmentation section. However, even for frames w/o any augmentation data we emitted size uleb! (this is the case of "unwinding through C code"). Thus frame moves information was emitterd completely wrong.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@96287 91177308-0d34-0410-b5e6-96231b3b80d8
lib/CodeGen/AsmPrinter/DwarfException.cpp