This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Scope of work to fix sched REG_DEAD handling?
- To: John Wehle <john at feith dot com>
- Subject: Re: Scope of work to fix sched REG_DEAD handling?
- From: Jeffrey A Law <law at upchuck dot cygnus dot com>
- Date: Sat, 24 Apr 1999 16:34:31 -0600
- cc: egcs at egcs dot cygnus dot com
- Reply-To: law at cygnus dot com
In message <199904242041.QAA06962@jwlab.FEITH.COM>you write:
> Any idea what's involved in fixing the schedulers so that they
> don't misplace the REG_DEAD notes? I'm willing to take a look
> at fixing them so that other optimizations can be enabled (such
> as having delete_computation recursively delete insns).
It should be relatively easy. The scheduler already knows how to deal with
notes in the first pass, but that code is disabled in the second pass.
> Also, what's involved in getting additional constant propagation
> to work on cc0 machines? If I recall the web announcement
> said that the new work currently doesn't support cc0 machines.
I think Bernd posted a patch for this recently. It probably won't get into
egcs-1.2, but should probably be installed into the mainline sources after
we create the egcs-1.2 branch.
jeff