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]
Other format:	[Raw text]

Re: [tree-ssa]: Loop LICM, take 2

From: Daniel Berlin <dberlin at dberlin dot org>
To: Diego Novillo <dnovillo at redhat dot com>
Cc: "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
Date: Wed, 27 Aug 2003 11:54:43 -0400
Subject: Re: [tree-ssa]: Loop LICM, take 2
References: <Pine.LNX.4.56.0308260928040.3711@dberlin.org> <1061986632.3255.1.camel@frodo.toronto.redhat.com> <D84602BC-D88F-11D7-A2B1-000A95A34564@dberlin.org> <1061991008.3255.38.camel@frodo.toronto.redhat.com>

Because we can't lift statements with unrenamed variables.

Why? Those unrenamed variables will have VUSE/VDEF operands. If they don't, the statement will be marked as having volatile operands. Do you have a test case that would fail without this?

I do now. :) I changed the test to use stmt_ann (stmt)->has_volatile_ops, and gengtype screws up.

Without the contains_unrenamed_variables test, we start hoisting calls, and doing so on one of them causes a problem.

First, are we supposed to be able to do this, or should i not hoist calls/should has_volatile_ops prevent us from doing it?

Follow-Ups:
- Re: [tree-ssa]: Loop LICM, take 2
  - From: Diego Novillo

References:
- [tree-ssa]: Loop LICM, take 2
  - From: Daniel Berlin
- Re: [tree-ssa]: Loop LICM, take 2
  - From: Diego Novillo
- Re: [tree-ssa]: Loop LICM, take 2
  - From: Daniel Berlin
- Re: [tree-ssa]: Loop LICM, take 2
  - From: Diego Novillo

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