This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] make has_gate and has_execute useless
- From: Trevor Saunders <tsaunders at mozilla dot com>
- To: Jeff Law <law at redhat dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 7 Nov 2013 13:24:57 -0500
- Subject: Re: [PATCH] make has_gate and has_execute useless
- Authentication-results: sourceware.org; auth=none
- References: <1383840050-26620-1-git-send-email-tsaunders at mozilla dot com> <527BCE28 dot 1070606 at redhat dot com>
On Thu, Nov 07, 2013 at 10:30:16AM -0700, Jeff Law wrote:
> On 11/07/13 09:00, tsaunders@mozilla.com wrote:
> >From: Trevor Saunders <tsaunders@mozilla.com>
> >
> >Hi,
> >
> > This is the result of seeing what it would take to get rid of the has_gate and
> >has_execute flags on pass_data. It turns out not much, but I wanted
> >confirmation this part is ok before I go deal with all the places that
> >initialize the fields.
> Most definitely, the has_gate/has_execute stuff is definitely a wart.
>
> >
> >I bootstrapped this part on x86_64-unknown-linux-gnu and saw no new test suite
> >regressions (ignoring the silk stuff because the full paths in its test names
> >break my test script for now) Any reason this patch with the actual removal of the flags wouldn't be ok?
> I don't see anything glaringly wrong.
great, I'll work on ripping them out then :)
> Any reason why the register_dump_files bits were mucking with
> properties. That just seems wrong.
no, it looks like its pretty old, the most recent change I can see is
r110742 in 2006, but I agree it seems crazy, but atleast its not
actually doing anything now.
> >+ todo_after = pass->execute ();
> >+ if (todo_after != TODO_absolutely_nothing)
> >+ do_per_function (clear_last_verified, NULL);
> >+ else
> >+ todo_after &= ~TODO_absolutely_nothing;
>
> Isn't the last assignment there just
> todo_after = 0;
yes
Trev
>
> ?
>
> Jeff