kristina added inline comments.
================
Comment at: src/UnwindCursor.hpp:1801
+ if (pc != getLastPC()) {
+ UNWIND_INFO *xdata = reinterpret_cast<UNWIND_INFO *>(base +
unwindEntry->UnwindData);
+ if (xdata->Flags & (UNW_FLAG_EHANDLER|UNW_FLAG_UHANDLER)) {
----------------
mstorsjo wrote:
> I can't say I understand all of this yet, but I'm slowly getting there, and
> I'm trying to compare this to what libgcc does.
>
> libgcc doesn't use any definition of UNWIND_INFO and doesn't need to do the
> equivalent of `getInfoFromSEH`, used by `step()`, anywhere. `unw_step()` is
> used in `_Unwind_ForcedUnwind`, which in libgcc is implemented using
> `RaiseException (STATUS_GCC_FORCED, ...`.
>
> I guess if you happen to have all of the `unw_step` API available, it's
> easier to just do it like this, in custom code without relying on the NTDLL
> functions for it, while the libgcc version relies more on the NTDLL API.
This primarily deals with the SEH exceptions re-purposed as a C++ exception
mechanism on x86_64 (if I understood this right), it's possible to set a custom
filter using a runtime call so I suspect GCC does that or defines a translation
function (also via a runtime call) which acts as a filter for "true" SEH
exceptions behind the scenes deep within the runtime. Typially "true" SEH
exceptions don't, outside of runtime magic, play nicely with C++ exceptions,
with the `__C_specific_handler` ones being a completely different paradigm that
falls far outside the scope of libunwind (those ones being the "true"/explicit
SEH exceptions).
(Don't take my word for it, it's been a while and I only implemented the "true"
variation for 64-bit Linux by reserving some RT signals and using that to
invoke additional runtime glue that would then do the unwinding, completely
ignoring DWARF since CFI exceptions and SEH exceptions really don't mix
especially on platforms that are not Windows-like)
Repository:
rUNW libunwind
https://reviews.llvm.org/D50564
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
