public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "rguenther at suse dot de" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug rtl-optimization/64164] [4.9/5 Regression] one more stack slot used due to one less inlining level
Date: Wed, 18 Mar 2015 10:16:00 -0000 [thread overview]
Message-ID: <bug-64164-4-1M89iahTYH@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-64164-4@http.gcc.gnu.org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64164
--- Comment #19 from rguenther at suse dot de <rguenther at suse dot de> ---
On Wed, 18 Mar 2015, aoliva at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64164
>
> --- Comment #18 from Alexandre Oliva <aoliva at gcc dot gnu.org> ---
> (In reply to Jeffrey A. Law from comment #14)
> > So, forgive me, but is -DOPT supposed to be the good or the bad code?
>
> It's the good one. As noted in comment 12, I had that backwards in comment 6.
>
> (In reply to Richard Biener from comment #15)
> > Because "coalescing" them doesn't do anything. copyrename coalescing
> > assigns the same underlying DECL to SSA names, thus makes SSA_NAME_VAR
> > the same. But when both SSA_NAME_VARs are NULL this won't do anything.
>
> This may be true when you look at only one pair of SSA names, but when you have
> 2 such candidate pairs involving one common SSA name, it can, and it does make
> a difference, as described in comment 6.
>
> (In reply to Richard Biener from comment #16)
> > And as I analyzed in comment #3 we chose the now different coalescing because
> > it is more profitable (to the cost analysis we perform in out-of-SSA
> > coalescing).
>
> No, we don't even *consider* the coalescing performed in the -DOPT case,
> because, as noted in comment 13, the SSA names ended up with different base
> names, because copyrename wouldn't give them the same base name.
>
> (In reply to Richard Biener from comment #17)
> > To add to all this - IMHO copyrename should go
>
> That's fine with me.
>
> > Yes, out-of-SSA coalescing could be changed to allow coalescing of SSA names
> > with a user-DECL and anonymous SSA names (or SSA names with a DECL_IGNORED_P
> > decl). But that will make the conflict graph much larger(?).
>
> Ok, I'll give that a shot tomorrow (I'll be away for the whole day today).
>
> > might happen that we end up coalescing things in a way
> > that there ends up being no DECL for a register and thus we lose in debug
> > quality (not sure if we could compensate by inserting debug stmts on the
> > edges to compensate for that - we couldn't do that if it requires splitting
> > the edge).
>
> I'm not concerned about that. Any debug stmts needed to make this work will
> already be in place, after the actual assignments, and after early PHI nodes.
> Sure, edge insertions might separate the copies from the debug stmts inserted
> after the PHI nodes, but ultimately the bindings should take care of it.
>
> > So to avoid wrong debug we'd have to always coalesce to an anonymous
> > entity
>
> With debug stmts, we don't care what the base names are any more. All the info
> we need (for tracked auto variables) is in the debug stmts.
But we do not always have debug stmts!
int bar (int b)
{
int i;
if (b > 7)
i = b / 3;
else
return b * 4;
return i;
}
is a testcase for what I am thinking of. Hmm. For some reason
into-SSA inserts a debug stmt, creating an extra copy _5 = i_4?!
<bb 2>:
if (b_2(D) > 7)
goto <bb 3>;
else
goto <bb 4>;
<bb 3>:
i_4 = b_2(D) / 3;
# DEBUG i => i_4
_5 = i_4;
goto <bb 5>;
<bb 4>:
_3 = b_2(D) * 4;
<bb 5>:
# _1 = PHI <_3(4), _5(3)>
so indeed we shouldn't lose anything here. But are you sure
that we are never using REG_EXPR for debug? With
-fno-var-tracking-assignments for sure we do (I suppose we don't care),
at least var-tracking.c does look at REG_EXPR.
next prev parent reply other threads:[~2015-03-18 10:16 UTC|newest]
Thread overview: 67+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-12-03 9:33 [Bug ipa/64164] New: " patrick.marlier at gmail dot com
2014-12-03 9:44 ` [Bug middle-end/64164] " pinskia at gcc dot gnu.org
2014-12-03 9:45 ` pinskia at gcc dot gnu.org
2014-12-03 10:03 ` [Bug rtl-optimization/64164] " rguenth at gcc dot gnu.org
2014-12-03 10:06 ` rguenth at gcc dot gnu.org
2014-12-11 11:46 ` rguenth at gcc dot gnu.org
2014-12-17 11:38 ` patrick.marlier at gmail dot com
2015-03-17 18:19 ` law at redhat dot com
2015-03-17 19:55 ` amacleod at redhat dot com
2015-03-17 22:15 ` aoliva at gcc dot gnu.org
2015-03-17 22:18 ` aoliva at gcc dot gnu.org
2015-03-17 22:19 ` aoliva at gcc dot gnu.org
2015-03-17 23:01 ` aoliva at gcc dot gnu.org
2015-03-18 0:20 ` aoliva at gcc dot gnu.org
2015-03-18 5:48 ` law at redhat dot com
2015-03-18 9:02 ` rguenth at gcc dot gnu.org
2015-03-18 9:03 ` rguenth at gcc dot gnu.org
2015-03-18 9:19 ` rguenth at gcc dot gnu.org
2015-03-18 10:05 ` aoliva at gcc dot gnu.org
2015-03-18 10:16 ` rguenther at suse dot de [this message]
2015-03-18 10:18 ` aoliva at gcc dot gnu.org
2015-03-19 5:17 ` aoliva at gcc dot gnu.org
2015-03-20 8:07 ` law at redhat dot com
2015-03-20 10:03 ` rguenther at suse dot de
2015-03-20 20:44 ` law at redhat dot com
2015-03-27 18:26 ` aoliva at gcc dot gnu.org
2015-03-29 20:58 ` aoliva at gcc dot gnu.org
2015-03-30 23:06 ` patrick.marlier at gmail dot com
2015-03-30 23:09 ` law at redhat dot com
2015-03-30 23:34 ` law at redhat dot com
2015-03-31 12:21 ` rguenth at gcc dot gnu.org
2015-03-31 12:21 ` [Bug rtl-optimization/64164] [4.9/5/6 " rguenth at gcc dot gnu.org
2015-03-31 12:56 ` jakub at gcc dot gnu.org
2015-04-01 7:51 ` rguenth at gcc dot gnu.org
2015-06-09 5:06 ` aoliva at gcc dot gnu.org
2015-06-09 5:34 ` [Bug rtl-optimization/64164] [4.9/5 " aoliva at gcc dot gnu.org
2015-06-09 9:10 ` ebotcazou at gcc dot gnu.org
2015-06-09 16:35 ` dje at gcc dot gnu.org
2015-06-10 0:15 ` [Bug rtl-optimization/64164] [4.9/5/6 " aoliva at gcc dot gnu.org
2015-06-10 13:16 ` clyon at gcc dot gnu.org
2015-06-10 14:43 ` dje at gcc dot gnu.org
2015-06-10 14:45 ` dje at gcc dot gnu.org
2015-07-23 15:35 ` aoliva at gcc dot gnu.org
2015-07-24 10:37 ` schwab@linux-m68k.org
2015-07-24 10:51 ` schwab@linux-m68k.org
2015-07-24 16:57 ` sje at gcc dot gnu.org
2015-07-27 9:07 ` ro at gcc dot gnu.org
2015-07-30 18:20 ` aoliva at gcc dot gnu.org
2015-08-02 21:11 ` gary at intrepid dot com
2015-08-14 18:52 ` aoliva at gcc dot gnu.org
2015-08-15 2:24 ` aoliva at gcc dot gnu.org
2015-08-15 2:25 ` aoliva at gcc dot gnu.org
2015-08-15 2:25 ` aoliva at gcc dot gnu.org
2015-08-15 2:26 ` aoliva at gcc dot gnu.org
2015-08-15 2:26 ` aoliva at gcc dot gnu.org
2015-08-15 2:27 ` aoliva at gcc dot gnu.org
2015-08-19 17:01 ` aoliva at gcc dot gnu.org
2015-08-21 20:04 ` aoliva at gcc dot gnu.org
2015-08-21 20:05 ` aoliva at gcc dot gnu.org
2015-09-23 21:13 ` aoliva at gcc dot gnu.org
2015-09-27 9:03 ` aoliva at gcc dot gnu.org
2015-09-27 12:13 ` trippels at gcc dot gnu.org
2015-09-28 1:15 ` aoliva at gcc dot gnu.org
2015-09-28 1:16 ` aoliva at gcc dot gnu.org
2015-09-28 1:16 ` aoliva at gcc dot gnu.org
2015-10-09 12:21 ` aoliva at gcc dot gnu.org
2015-10-09 12:22 ` aoliva at gcc dot gnu.org
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=bug-64164-4-1M89iahTYH@http.gcc.gnu.org/bugzilla/ \
--to=gcc-bugzilla@gcc.gnu.org \
--cc=gcc-bugs@gcc.gnu.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).