This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: i386 nonlocal_goto_receiver patch



  In message <u9wwa995mi.fsf@yorick.cygnus.com>you write:
  > >>>>> John Wehle <john@feith.com> writes:
  > 
  > > This patch removes the nonlocal_goto_receiver pattern since all it
  > > does is restore the global table pointer which is in the PIC register.
  > > The documentation in md.texi indicates that a nonlocal goto only
  > > happens within a unit-of-translation and as a result doesn't need
  > > to restore the global table pointer since it's already correct.
  > 
  > Hmm...but exceptions also use the nonlocal_goto_receiver.
I believe rth refined the meaning of nonlocal_goto_receiver a while
back and that we have some other mechanism to deal with the PIC register
when doing a throw across module boundaries.

I believe exception_receiver is responsible for dealing with this
stuff now.


jeff


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]