From: Segher Boessenkool <segher@kernel.crashing.org>
To: Jeff Law <law@redhat.com>
Cc: gcc-patches@gcc.gnu.org, bschmidt@redhat.com
Subject: Re: [PATCH 8/9] shrink-wrap: shrink-wrapping for separate components
Date: Wed, 14 Sep 2016 13:45:00 -0000 [thread overview]
Message-ID: <20160914133812.GD4896@gate.crashing.org> (raw)
In-Reply-To: <0782459b-852f-f667-ca59-ade837ca2988@redhat.com>
On Mon, Sep 12, 2016 at 12:02:50PM -0600, Jeff Law wrote:
> >>>As a final optimisation, if a block needs a prologue and its immediate
> >>>dominator has the block as a post-dominator, the dominator gets the
> >>>prologue as well.
> >>So why not just put it in the idom and not in the dominated block?
> >
> >That's what it does :-)
> Then I must have mis-parsed. Thanks for clarifying.
"As a final optimisation, if a block needs a prologue and its immediate
dominator has the block as a post-dominator, ***that immediate dominator***
gets the prologue as well."
That is clearer I hope :-)
> Hmm, then explain again why DCE is mucking up? I don't immediately see
> how EPILOGUE_BEG notes come into play with DCE. It seems to rely on the
> DF data and AFAICT DF only cares about the EPILOGUE_BEG note in
> can_move_insns_across which shouldn't be used by DCE.
The register restore *is* dead code, but we need to have the same CFI
for all convergent paths.
> >>Consider using auto_sbitmap rather than manually managing
> >>allocation/releasing of the per-block structures. In fact, can't all of
> >>SW become a class and we lose the explicit init/fini routines in favor
> >>of a ctor/dtor?
> >
> >Yes, you can always add indirection. I do not think the code becomes
> >more readable that way (quite the opposite). Explicit is *good*.
> The GCC project is moving away from this kind of explicit
> allocation/deallocation and more towards a RAII. Unless there is a
> clear need for the explicit allocation/deallocation, please put this
> stuff into a class with an appropriate ctor/dtor.
>
> FWIW, I was a big opponent of how much stuff happens "behind your back"
> with some languages (including C++). But over the last few years my
> personal stance has softened considerably after seeing how cleanly RAII
> solves certain problems.
We then still cannot get rid of SW, which is a convenience macro to do
a nasty cast on bb->aux. If bb->aux was some pretty class hierarchy,
easy to use and all that, I would of course agree with your suggestion.
But as it is it is just a bare pointer, so the less we hide the safer
it is.
> >>For the PPC R0 vs LR is the only thing that causes disqualification
> >>right?
> >
> >Currently, yes.
> >
> >>Can't that be handled when we build the set of components we
> >>want to insert for each edge/block? Is there some advantage to handling
> >>disqualifications after all the potential insertion points have been
> >>handled?
> >
> >We do not know if an edge needs a prologue, epilogue, or neither, until
> >we have decided whether *both* ends of that edge want the component active
> >or not.
> Right. Hmm, maybe I'm not asking the question clearly.
>
> Whether or not an edge needs a prologue or epilogue is a function not
> just of the state at the head or tail of the edge, but instead is a
> function of global dataflow propagation? Thus we can't disqualify until
> after we've done the dataflow propagation? Right?
We can figure out before we decide what blocks need what components, what
edges can not get a prologue or epilogue for which components. This
complicates the selection algorithm a whole lot, for not much gain that
I have seen so far, so I just give up in the cases that end up "bad".
It is not easy at all to see what edges will need to get a *logue,
because not always both blocks that edge connects are in the same
dominator subtree (or tree even, for an epilogue-aware placement
algorithm, but this patch doesn't do that yet; it's a more minor
optimisation, only reduces code size a little).
Segher
next prev parent reply other threads:[~2016-09-14 13:38 UTC|newest]
Thread overview: 73+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-08-01 1:43 [PATCH v2 0/9] Separate shrink-wrapping Segher Boessenkool
2016-08-01 1:43 ` [PATCH 3/9] dce: Don't dead-code delete separately wrapped restores Segher Boessenkool
2016-09-08 17:52 ` Jeff Law
2016-09-09 15:59 ` Segher Boessenkool
2016-09-09 16:39 ` Jeff Law
2016-08-01 1:43 ` [PATCH 2/9] cfgcleanup: Don't confuse CFI when -fshrink-wrap-separate Segher Boessenkool
2016-09-08 17:51 ` Jeff Law
2016-08-01 1:43 ` [PATCH 1/9] separate shrink-wrap: New command-line flag, status flag, hooks, and doc Segher Boessenkool
2016-08-29 9:31 ` Bernd Schmidt
2016-08-29 14:30 ` Segher Boessenkool
2016-09-08 17:37 ` Jeff Law
2016-09-09 11:03 ` Bernd Schmidt
2016-09-09 15:13 ` Segher Boessenkool
2016-09-09 18:31 ` Jeff Law
2016-09-09 20:41 ` Segher Boessenkool
2016-09-12 16:49 ` Jeff Law
2016-09-09 15:51 ` Jeff Law
2016-09-09 15:40 ` Segher Boessenkool
2016-09-09 16:58 ` Jeff Law
2016-09-08 17:48 ` Jeff Law
2016-09-09 15:44 ` Segher Boessenkool
2016-08-01 1:57 ` [PATCH 4/9] regrename: Don't rename restores Segher Boessenkool
2016-09-08 17:54 ` Jeff Law
2016-09-09 21:05 ` Segher Boessenkool
2016-09-12 17:01 ` Jeff Law
2016-08-01 1:57 ` [PATCH 9/9] rs6000: Separate shrink-wrapping Segher Boessenkool
2016-08-01 1:57 ` [PATCH 8/9] shrink-wrap: shrink-wrapping for separate components Segher Boessenkool
2016-09-08 19:03 ` Jeff Law
2016-09-09 22:03 ` Segher Boessenkool
2016-09-12 18:21 ` Jeff Law
2016-09-14 13:45 ` Segher Boessenkool [this message]
2016-09-15 16:47 ` Jeff Law
2016-08-01 1:57 ` [PATCH 7/9] cprop: Leave RTX_FRAME_RELATED_P instructions alone Segher Boessenkool
2016-09-08 18:34 ` Jeff Law
2016-09-09 21:21 ` 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
2016-08-01 2:12 ` [PATCH 6/9] sel-sched: Don't mess with register restores Segher Boessenkool
2016-08-04 7:33 ` Andrey Belevantsev
2016-09-08 17:55 ` Jeff Law
2016-09-09 21:13 ` Segher Boessenkool
2016-09-12 17:39 ` Jeff Law
2016-09-14 13:28 ` Segher Boessenkool
2016-08-04 0:05 ` [PATCH v2 0/9] Separate shrink-wrapping Segher Boessenkool
2016-08-24 16:04 ` Segher Boessenkool
2016-08-26 13:03 ` Bernd Schmidt
2016-08-26 13:48 ` David Malcolm
2016-08-26 13:55 ` Bernd Schmidt
2016-08-26 14:50 ` Segher Boessenkool
2016-08-26 15:03 ` Bernd Schmidt
2016-08-26 16:27 ` Segher Boessenkool
2016-09-08 16:58 ` Jeff Law
2016-09-09 15:26 ` Segher Boessenkool
2016-09-09 16:26 ` Jeff Law
2016-09-09 16:51 ` Segher Boessenkool
2016-09-09 17:22 ` Jeff Law
2016-08-30 12:31 ` Michael Matz
2016-09-08 16:41 ` Jeff Law
2016-09-09 6:31 ` Segher Boessenkool
2016-09-09 15:28 ` Jeff Law
2016-09-09 15:43 ` Segher Boessenkool
2016-09-09 18:25 ` Jeff Law
2016-09-09 20:29 ` Segher Boessenkool
2016-09-08 17:20 ` Jeff Law
2016-09-09 15:33 ` Segher Boessenkool
2016-09-09 16:49 ` Jeff Law
2016-09-09 17:00 ` Segher Boessenkool
2016-09-09 17:44 ` Jeff Law
2016-09-09 19:36 ` Jeff Law
2016-09-09 21:00 ` Segher Boessenkool
2016-09-12 11:00 ` Bernd Schmidt
2016-09-12 16:59 ` Jeff Law
2016-09-14 13:22 ` Segher Boessenkool
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=20160914133812.GD4896@gate.crashing.org \
--to=segher@kernel.crashing.org \
--cc=bschmidt@redhat.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).