This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Do alias analysis after SSA. Improvements to PR8361.
- From: Chris Lattner <sabre at nondot dot org>
- To: Jan Hubicka <hubicka at ucw dot cz>
- Cc: Daniel Berlin <dberlin at dberlin dot org>,Diego Novillo <dnovillo at redhat dot com>, Jeff Law <law at redhat dot com>,"gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
- Date: Thu, 8 Jan 2004 15:21:15 -0600 (CST)
- Subject: Re: [tree-ssa] Do alias analysis after SSA. Improvements to PR8361.
On Thu, 8 Jan 2004, Jan Hubicka wrote:
> >
> > On Jan 8, 2004, at 11:22 AM, Diego Novillo wrote:
> >
> > >On Thu, 2004-01-08 at 11:14, Chris Lattner wrote:
> > >
> > >>Just a clarification: you both mean flow sensitive analysis, not
> > >>context
> > >>sensitive, right?
> > >>
> > >Well, we do have call-graph info, so context sensitivity could be used.
> > Our call graph doesn't resolve function pointers, AFAIK, so you can't
> > make it context sensitive without being extra-conservative.
> I plan to add the function pointers bits in the future too, but first I
> need to be able to do analysis in early pass, so I would like to get
> into shape I can build CFG and do some trivial changes before inlining.
Besides, any good context sensitive pointer analysis should be able to
build the call graph on the fly, refining it with the information it is
gathering. e.g.:
http://llvm.cs.uiuc.edu/pubs/2003-04-29-DataStructureAnalysisTR.html
Note that the biggest limiting factor is going to be lack of whole-program
information. Without that, you have to treat _direct_ calls just as
pessimistically as indirect calls... but the program-at-a-time stuff in
GCC will probably help with this problem.
-Chris
--
http://llvm.cs.uiuc.edu/
http://www.nondot.org/~sabre/Projects/