From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jason Merrill <jason_merrill@redhat.com>
To: Jan Hubicka <jh@suse.cz>
Cc: Mark Mitchell <mark@codesourcery.com>, Richard Kenner <kenner@vlsi1.ultra.nyu.edu>, "gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Subject: Re: spec2000 regression
Date: Thu, 16 Aug 2001 07:46:00 -0000
Message-id: <m3hev8cbba.fsf@prospero.cambridge.redhat.com>
References: <m3bslhquqn.fsf@prospero.cambridge.redhat.com> <26940000.997887761@warlock.codesourcery.com> <20010816145745.J14857@atrey.karlin.mff.cuni.cz> <m37kw4due7.fsf@prospero.cambridge.redhat.com> <20010816163901.L14857@atrey.karlin.mff.cuni.cz>
X-SW-Source: 2001-08/msg00972.html

>>>>> "Jan" == Jan Hubicka <jh@suse.cz> writes:

>> >>>>> "Jan" == Jan Hubicka <jh@suse.cz> writes:
>> 
>> > +       /* set_mem_attributes uses DECL_RTL to avoid re-generating of
>> > +          already computed alias sets.  Here we want to re-generate.  */
>> > +       if (TREE_CODE (decl) != SAVE_EXPR)
>> > + 	SET_DECL_RTL (decl, NULL);
>> >         set_mem_attributes (reg, decl, 1);
>> > +       if (TREE_CODE (decl) != SAVE_EXPR)
>> > + 	SET_DECL_RTL (decl, reg);
>> 
>> This should check DECL_P.  Otherwise, looks good.

> May I interpret this as approval with this change?

Yes.  Please apply all the changes you have; I haven't checked in any of
the patches I've sent out.

Jason