This is the mail archive of the gcc@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]
Other format: [Raw text]

Re: birthpoints in rtl.


"Steven Bosscher" <stevenb.gcc@gmail.com> writes:

> On 28 Feb 2008 12:41:30 -0800, Ian Lance Taylor <iant@google.com> wrote:
> >  libcalls are still used for no-conflict blocks.  This may be what you
> >  mean by the scheduling thing.  No-conflict blocks are emitted by
> >  emit_no_conflict_block and checked in local-alloc.
> 
> I thought the no-conflict blocks were not as important anymore since
> your subreg lowering work, and maybe even since we started doing
> "real" liveness analysis (i.e. flow.c in ~1997 ;-))
> What do you think, is this a blocker issue?

I think it's probably a blocker issue on 32-bit x86 until we complete
the lower subreg work to track subreg lifetimes to detect no-conflict
sections manually.  I have an implementation of that written before
DF, and I haven't found the time to upgrade it to the DF framework.

Ian


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