From mboxrd@z Thu Jan  1 00:00:00 1970
From: Mark Mitchell <mark@codesourcery.com>
To: rth@cygnus.com
Cc: geoffk@cygnus.com, gcc@gcc.gnu.org
Subject: Re: ssa bootstrap problem on x86 (cmpstrsi_1 pattern)
Date: Thu, 27 Jul 2000 15:13:00 -0000
Message-id: <20000727151254U.mitchell@codesourcery.com>
References: <200007271958.MAA00620@localhost.cygnus.com> <20000727131011W.mitchell@codesourcery.com> <20000727144813.A625@cygnus.com>
X-SW-Source: 2000-07/msg00879.html

>>>>> "Richard" == Richard Henderson <rth@cygnus.com> writes:

    Richard> On Thu, Jul 27, 2000 at 01:10:11PM -0700, Mark Mitchell
    Richard> wrote:
    >> I think we need to bite the bullet and get the CLOBBERs out of
    >> there; otherwise, we are forced to do some kind of dataflow
    >> anlaysis in SSA...

    Richard> Not at all.  We have a register set, once, to an
    Richard> unspecified value.  That unspecified value is "valid"
    Richard> everywhere, just as with a register set to a useful
    Richard> value.

In Geoff's examples, I believe it was SET once, and then CLOBBERed.
Did I misread?

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com