From: Segher Boessenkool <segher@kernel.crashing.org>
To: Jeff Law <law@redhat.com>
Cc: Bernd Schmidt <bschmidt@redhat.com>,
gcc-patches@gcc.gnu.org, dje.gcc@gmail.com
Subject: Re: [PATCH 5/9] regrename: Don't run if function was separately shrink-wrapped
Date: Wed, 14 Sep 2016 19:13:00 -0000 [thread overview]
Message-ID: <20160914190353.GA22273@gate.crashing.org> (raw)
In-Reply-To: <85cb1e59-f383-129e-bea7-e179f913c0c9@redhat.com>
On Wed, Sep 14, 2016 at 12:11:50PM -0600, Jeff Law wrote:
> On 09/14/2016 07:04 AM, Segher Boessenkool wrote:
> >>I'd
> >>probably start by fixing the dataflow issues and see if that fixes the
> >>regrename thing as a side effect.
> >
> >Have you seen https://gcc.gnu.org/ml/gcc-patches/2016-07/msg00091.html ?
> I missed it.
Yeah thought so, too much stuff in flight here.
> My interpretation....
>
> The uses at each "strange" exit fixing the first issue with
> shrink-wrapping definitely sounds like we've got a dataflow problem of
> some sort.
>
> If you think about it, conceptually we want the return insn to make the
> callee saved registers "used" so that DCE, regrename and friends don't
> muck with them. The fact that we don't is as much never having to care
> all that much until recently.
(There is no return insn at those exits; these are exits *without*
successor block, not the exit block).
It is puzzling to me why adding USEs for just the registers that *are*
separately shrink-wrapped does not work; only all those that *could* be
shrink-wrapped does. Do you have any idea about that?
> I continue to wonder if we could add something similar to
> CALL_INSN_FUNCTION_USAGE where we attach uses for all the call-saved
> registers onto a return insn. We would attach them at the end of
> prologue/epilogue generation and only attach those where were live
> somewhere in the code.
Maybe adding the new insns to the {pro,epi}logue_insn_hash will help
something. Or maybe it will blow up spectacularly. Will know in a
bit.
> By deferring that step until after prologue/epilogue generation we
> shouldn't cause unnecessary register saves/restores.
Hrm. I'll see about that as well.
> I pondered just doing it for the separately wrapped components on that
> particular path, but I've yet to convince myself that's actually correct.
If that is not correct, how is the status quo correct? That is what
puzzles me above, too.
> Bernd knows the regrename code better than anyone. Is there any way the
> two of you could work together to try and track down what's going on in
> the hash_map_rand case? Even throwing in some more debug stuff might
> help narrow things down since it's renaming something to a non-volatile,
> non-separately shrink wrapped register that's causing problems.
Okay with me, I could certainly use his help. I'll try the above things
first though, so not before friday.
> Can we agree that there's a set of targets that will improve and a set
> that are harmed? And that to enable regrename by default we need to
> either better describe the pipeline characteristics we're optimizing for
> or a well defined way for targets to turn it off?
There is a well-defined way to turn it off, via common/config/*/*-common.c ,
TARGET_OPTION_OPTIMIZATION_TABLE. We disagree on whether most targets will
want it enabled, i.e. whether it should (eventually) be enabled by default.
Segher
next prev parent reply other threads:[~2016-09-14 19:04 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-06-08 1:48 [PATCH 0/9] separate shrink-wrapping Segher Boessenkool
2016-06-08 1:48 ` [PATCH 2/9] cfgcleanup: Don't confuse CFI when -fshrink-wrap-separate Segher Boessenkool
2016-06-08 1:48 ` [PATCH 1/9] separate shrink-wrap: New command-line flag, status flag, hooks, and doc Segher Boessenkool
2016-06-08 1:48 ` [PATCH 3/9] dce: Don't dead-code delete separately wrapped restores Segher Boessenkool
2016-06-08 1:53 ` [PATCH 6/9] sel-sched: Don't mess with register restores Segher Boessenkool
2016-06-08 1:53 ` [PATCH 4/9] regrename: Don't rename restores Segher Boessenkool
2016-06-08 1:54 ` [PATCH 5/9] regrename: Don't run if function was separately shrink-wrapped Segher Boessenkool
2016-06-08 9:18 ` Bernd Schmidt
2016-09-09 18:41 ` Jeff Law
2016-09-09 20:56 ` Segher Boessenkool
2016-09-09 23:12 ` Jeff Law
2016-09-10 6:59 ` Segher Boessenkool
2016-09-12 16:36 ` Jeff Law
[not found] ` <CAGWvny=fHHZtKF4_D2098+3PTPPzxtg3EjKDWHyJwUxz8g_tEA@mail.gmail.com>
[not found] ` <CAGWvnymZVg_FR_PHqhwkgrAkHDntVMEiG4shfst_GA9OnZKvWg@mail.gmail.com>
[not found] ` <CAGWvnykQ3oz0UpcF6U1WYivbJww65h2EH5n3FocQ8JGY9hrOrA@mail.gmail.com>
2016-09-12 17:04 ` Jeff Law
2016-09-14 13:08 ` Segher Boessenkool
2016-09-14 13:18 ` Bernd Schmidt
2016-09-14 14:01 ` Segher Boessenkool
2016-09-14 14:54 ` Bernd Schmidt
2016-09-14 16:33 ` Segher Boessenkool
2016-09-14 19:10 ` Jeff Law
2016-09-14 17:55 ` Jeff Law
2016-09-14 19:13 ` Segher Boessenkool
2016-09-14 19:36 ` Jeff Law
2016-09-14 18:21 ` Jeff Law
2016-09-14 19:13 ` Segher Boessenkool [this message]
2016-09-14 19:38 ` Jeff Law
2016-09-14 22:34 ` Segher Boessenkool
2016-09-15 17:28 ` Jeff Law
2016-09-19 17:11 ` Segher Boessenkool
2016-09-14 20:04 ` Jeff Law
2016-09-14 22:51 ` Segher Boessenkool
2016-06-08 1:54 ` [PATCH 7/9] cprop: Leave RTX_FRAME_RELATED_P instructions alone Segher Boessenkool
2016-06-08 2:03 ` [PATCH 8/9] shrink-wrap: shrink-wrapping for separate concerns Segher Boessenkool
2016-07-15 12:42 ` Bernd Schmidt
2016-07-18 16:34 ` Segher Boessenkool
2016-07-18 17:03 ` Bernd Schmidt
2016-07-19 14:46 ` Segher Boessenkool
2016-07-19 14:49 ` Bernd Schmidt
2016-07-19 15:35 ` Segher Boessenkool
2016-07-20 11:23 ` Bernd Schmidt
2016-07-20 15:06 ` Segher Boessenkool
2016-06-08 2:04 ` [PATCH 9/9] rs6000: Separate shrink-wrapping Segher Boessenkool
2016-06-08 11:56 ` [PATCH 0/9] separate shrink-wrapping Bernd Schmidt
2016-06-08 12:45 ` Eric Botcazou
2016-06-08 15:16 ` Segher Boessenkool
2016-06-08 16:43 ` Bernd Schmidt
2016-06-08 17:26 ` Segher Boessenkool
2016-06-29 23:06 ` Bernd Schmidt
2016-06-29 23:24 ` Segher Boessenkool
2016-07-04 8:57 ` Segher Boessenkool
2016-06-14 21:24 ` Segher Boessenkool
2016-07-08 10:42 ` Bernd Schmidt
2016-07-08 12:11 ` Segher Boessenkool
2016-07-08 13:16 ` David Malcolm
2016-07-08 13:45 ` Segher Boessenkool
2016-07-08 14:35 ` Bill Schmidt
2016-06-09 16:12 ` Jeff Law
2016-06-09 19:57 ` Segher Boessenkool
2016-06-28 0:22 ` PING " Segher Boessenkool
2016-07-07 10:16 ` PING x2 " Segher Boessenkool
2016-08-01 1:43 [PATCH v2 0/9] Separate shrink-wrapping Segher Boessenkool
2016-08-01 2:12 ` [PATCH 5/9] regrename: Don't run if function was separately shrink-wrapped Segher Boessenkool
2016-09-08 17:54 ` Jeff Law
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=20160914190353.GA22273@gate.crashing.org \
--to=segher@kernel.crashing.org \
--cc=bschmidt@redhat.com \
--cc=dje.gcc@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=law@redhat.com \
/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).