This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: MEM_COPY_ATTRIBUTES
- To: kenner at vlsi1 dot ultra dot nyu dot edu
- Subject: Re: MEM_COPY_ATTRIBUTES
- From: Mark Mitchell <mark at codesourcery dot com>
- Date: Wed, 03 May 2000 12:09:19 -0700
- Cc: gcc at gcc dot gnu dot org
- Organization: CodeSourcery, LLC
- References: <10005031859.AA28233@vlsi1.ultra.nyu.edu>
>>>>> "Richard" == Richard Kenner <kenner@vlsi1.ultra.nyu.edu> writes:
Richard> What does "some" mean? That's the key point as I see it.
Richard> Right now, we've picked a peculiar subset of the
Richard> properties. I can't see why volatility is a property
Richard> that would be useful to copy in cases when readonly
Richard> isn't. Why make that distinction?
Reasonable argument. But, volatility is a conservative choice;
readonly-ness is an optimistic choice.
Richard> Note also that I think it is safe to always copy
Richard> MEM_ALIAS_SET because that alis can always safely be used
Richard> to reference a component: it's just that in some cases
Richard> you might be able to find a *better* alias set for that
Richard> component.
It's certainly not safe if the new MEM is bigger than the old one.
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com