This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: What is this patch doing?
- To: Bernd Schmidt <bernds at pasanda dot cygnus dot co dot uk>
- Subject: Re: What is this patch doing?
- From: Joern Rennecke <amylaar at cygnus dot co dot uk>
- Date: Tue, 1 Aug 2000 22:04:14 +0100 (BST)
- CC: Joern Rennecke <amylaar at pasanda dot cygnus dot co dot uk>, gcc at gcc dot gnu dot org, rth at cygnus dot com
> Just -m3 is enough to reproduce it here.
I have updated my tree now and can reporduce the problem.
There are actually two problem with using the first post-reload
splitting pass: first, rebuild_notes_after_reload is not set
depending on whether the splitter created any notes or not.
But, more importantly, this splitter is only run when optimizing.
So a split pattern that is only supposed to be used at that point
would have to be guarded against being run too late. Do we want to
introduce a new global variable that says when a split pattern may
generate labels? Or is there already one?
Or should we rather make this post-reload splitter always executed?