How about this as a compromize:
- in 6.1
your original patch (but with a comment saying that the
+ if (IN_SOLIB_CALL_TRAMPOLINE (stop_pc, ecs->stop_func_name))
is a hack and shouldn't be included in the mainline)
- in mainline:
Assuming Mark's ok with the sparc changes, your patch without that part?
This way 6.1 is robust regardless of which SPARC architecture code is in
place.
it looks reasonable, but it'd be nice to have Mark's opinion on the
unwinder problem. How about we let it sit for another day or two before
we commit anything?
Also, do we need this change in 6.1? It seemed like a pretty minor bug
that I managed to produce on Tru64 only. Or were there other occurrences
of this bug?