search for: eh_dwarf_cfa

Hi All, While compiling libgcc with LLVM for our target, it seems I don't get __builtin_dwarf_cfa() right. That causes uw_init_context_1 [1] initialize the wrong context, and the overall exception handling fail. I find Hal provided patch years ago adding ISD::EH_DWARF_CFA, which corresponding to __builtin_dwarf_cfa(), and let us have the opportunity to handle __builtin_dwarf_cfa() for our own [2]. I copy Mips' lowerEH_DWARF_CFA for our target, and it *just* works [3]. I have to admit that I don't understand what Hal explained in his patch, also don't...

...lvm-convert.cpp (revision 41260) +++ llvm-convert.cpp (working copy) @@ -4240,11 +4240,12 @@ return false; int cfa_offset = ARG_POINTER_CFA_OFFSET(0); - - Result = Builder.CreateCall(Intrinsic::getDeclaration(TheModule, - Intrinsic::eh_dwarf_cfa), - ConstantInt::get(Type::Int32Ty, cfa_offset)); - + Result = + Builder.CreateCall(Intrinsic::getDeclaration(TheModule, + Intrinsic::eh_dwarf_cfa), + Builder.CreateIntToPtr(ConstantInt::get(Type:...

...was adding an offset of type i32 to a pointer. This is > fine if pointers are 32-bits, but we get an assert if the types are > different. By converting the cfa_offset to a pointer, we get it to be > that size. Please ignore my prev. e-mail. The code breaks at least 32-bit stuff. The arg of eh_dwarf_cfa is i32, how it can be pointer? -- With best regards, Anton Korobeynikov. Faculty of Mathematics & Mechanics, Saint Petersburg State University.

Bill, > The built-in dwarf_cfa doesn't take an argument. It returns a pointer. Ah. Sorry for confusion: gcc's __builtin_draft_cfa doesn't have any argument LLVM's eh_dwarf_cfa intrinsic has extra "offset" argument, which is i32, because we have to propagate some information from GCC to LLVM. > But there's an offset associated with the CFA. So it looks like LLVM > is taking the pointer and adding the offset to it to synthesize what > GCC is doing....