This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix store motion, revised PRE memory handling
- To: law at redhat dot com
- Subject: Re: [PATCH] Fix store motion, revised PRE memory handling
- From: Daniel Berlin <dan at cgsoftware dot com>
- Date: Mon, 16 Jul 2001 13:16:08 -0400
- Cc: Daniel Berlin <dan at cgsoftware dot com>,Richard Henderson <rth at redhat dot com>, gcc-patches at gcc dot gnu dot org
- References: <29038.995303195@localhost.localdomain>
law@redhat.com writes:
> In message <87u20cesr3.fsf@cgsoftware.com>you write:
> > Richard Henderson <rth@redhat.com> writes:
> >
> > > On Mon, Jul 16, 2001 at 01:45:02AM -0400, Daniel Berlin wrote:
> > >> There is a consistent pattern of
> > >>
> > >> if (load_killed_in_block_p (...)) <which is the way you want to handle i
> > t>
> > >> if (mem_*set_* crap) <which is what we used to do>
> > >
> > > Wow, that's ... stunning. I can only guess that there was
> > > a botched cvs merge at some point in the past.
> > >
> > >> This patch won't be committed until a proper fix is committed to df.c
> > >> to fix the not marking all the right regs as def'd in call insns, but
> > >> i'd still like it reviewed anyway, since that's just a simple matter
> > >> to fix anyway.
> > >
> > > I'm working on that.
> > >
> > > This is a somewhat large-ish. Would you split out the stupidity
> > > killing into a separate patch? It would make it easier to examine
> > > the actual logic changes you've got in there.
> >
> > Sure.
> > I've attached the first patch (the stupidity killing).
> > Speaking of logic changes, if we use !rtx_addr_varies_p when picking
> > out candidate stores, do we still need to bother to see if the
> > register operands change in various blocks? If the address doesn't
> > vary, doesn't it mean they couldn't be possibly changing in a
> > meaningful way, because if they did, it would vary?
> >
> > (We used to use rtx_varies_p, which was overkill, and would disqualify
> > basically everything).
> Err, umm. I don't see any indication this patch was bootstrap tested.
>
I said the original, larger, patch, was bootstrapped on
powerpc-linux-gnu. When I split it, I bootstrapped it again on
powerpc-linux-gnu, again. One would assume that if i bootstrapped the
larger patch, i would have bootstrapped the smaller broken up patches
too, before submitting them.
However, i'll explicitly note it in the future.
> jeff
--
"I bought one of those little glass ball things with the snow in
it. You know, you turn it upside down then you turn it back and
it starts to snow. I bought one, except this has a snow plow
that does it in rows.
"-Steven Wright