From: Richard Biener <richard.guenther@gmail.com>
To: Jan Hubicka <hubicka@ucw.cz>
Cc: Martin Liska <mliska@suse.cz>, GCC Patches <gcc-patches@gcc.gnu.org>
Subject: Re: [PATCH 3/9] operand_equal_p: add support for OBJ_TYPE_REF.
Date: Mon, 19 Aug 2019 14:03:00 -0000 [thread overview]
Message-ID: <CAFiYyc0GrdzuZp8zTweJSiKwCm747RyW--hPCoa973=OBOVQxA@mail.gmail.com> (raw)
In-Reply-To: <20190816120727.wttrwdhp4dphkksx@kam.mff.cuni.cz>
On Fri, Aug 16, 2019 at 2:07 PM Jan Hubicka <hubicka@ucw.cz> wrote:
>
> >
> > When we compare OBJ_TYPE_REF_TOKEN and OBJ_TYPE_REF_OBJECT
> > and they are equal, are there cases where types_same_for_odr returns false?
>
> OBJ_TYPE_REF_OBJECT is pointer to the instance, OBJ_TYPE_REF_TOKEN is
> index in the vtable or the type given by obj_ref_type_class. I guess
> one can do something like
>
> void *ptr;
> ...
> if (cond)
> ((class_a *)ptr)->firstvirtualmethod_of_class_a ();
> else
> ((class_b *)ptr)->firstvirtualmethod_of_class_b ();
>
> Here OBJECT will be always *ptr, TOKEN will be 0, but the actual virtual
> method is different. So merging this may lead to invalid
> devirtualization at least when the classes are anonymous namespace and
> we work out late in compilation that they are not derived.
But we also compare OBJ_TYPE_REF_EXPR and later we expand to a call
with exactly that address...
> >
> > I guess we went over this some time ago when talking about value-numbering
> > of them. I realize the devirt machinery use the class type to get at the
> > virtual table but since we later expand simply OBJ_TPE_REF_EXPR _that_
> > already has the result of the load from the vitual table.
> >
> > So - do we need to compare obj_type_ref_class at all? The worst thing
> > that could happen is that devirt no longer can devirtualize after the merging
> > but it should never result in "wrong" devirtualization? So, do we really
> > want to inhibit ICF of two equal functions that could be eventually
> > devirtualized both but into different direct calls? Can't this sort of thing
>
> I think it is like with the TBAA info. I would suggest first implement
> tests that preserve correctness of both TBAA and devirt annotations.
> Incrementally we can teach ICF to be agressive with -Os and merge anyway
> while dropping the metadata (removing OBJ_TYPE_REF or adjusting alias
> sets
> > happen anyway when you'd factor in IPA-CP and two identical functions
> > with indirect calls to a function argument but called with different constant
> > args? And should IPA-CP then not consider cloning after ICF merging
> > (not sure how ordering works here).
>
> I am not sure how those are realated. For sure IPA-CP can clone when it
> works out that different types are passed to the function and
> devirtualize to different calls. But this is bit different - it is about
> the type annotation we
>
> Honza
> >
> > Richard.
> >
> > > Honza
> > > >
> > > > + }
> > > >
> > > >
> > > > > 2019-07-24 Martin Liska <mliska@suse.cz>
> > > > >
> > > > > * fold-const.c (operand_equal_p): Support OBJ_TYPE_REF.
> > > > > * tree.c (add_expr): Hash parts of OBJ_TYPE_REF.
> > > > > ---
> > > > > gcc/fold-const.c | 21 +++++++++++++++++++++
> > > > > gcc/tree.c | 9 +++++++++
> > > > > 2 files changed, 30 insertions(+)
> > > > >
next prev parent reply other threads:[~2019-08-19 13:33 UTC|newest]
Thread overview: 59+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-22 12:50 [RFC] operand_equal_p with valueization Jan Hubicka
2015-05-22 13:40 ` Richard Biener
2015-05-22 14:12 ` Jan Hubicka
2015-05-26 8:18 ` Richard Biener
2015-05-26 19:09 ` Jan Hubicka
2015-05-27 8:49 ` Richard Biener
2019-06-18 11:10 ` [RFC] " Martin Liška
2019-08-06 15:44 ` [PATCH 0/9] IPA ICF overhaul Martin Liska
2019-08-06 15:43 ` [PATCH 7/9] IPA ICF: remove dead code Martin Liska
2019-08-08 14:44 ` Jeff Law
2019-08-06 15:43 ` [PATCH 6/9] Integrate that for IPA ICF Martin Liska
2019-08-16 11:10 ` Martin Liška
2019-08-06 15:43 ` [PATCH 5/9] Come up with an abstraction Martin Liska
2019-08-08 16:29 ` Michael Matz
2019-08-12 11:49 ` Martin Liška
2019-08-12 12:27 ` Richard Biener
2019-08-12 12:43 ` Martin Liška
2019-08-12 13:26 ` Richard Biener
2019-08-12 14:48 ` Martin Liška
2019-08-14 13:17 ` Richard Biener
2019-08-14 13:50 ` Martin Liška
2019-08-14 14:38 ` Richard Biener
2019-08-16 11:06 ` Martin Liška
2019-09-18 7:56 ` Martin Liška
2019-09-19 11:30 ` Richard Biener
2019-08-12 13:40 ` Michael Matz
2019-08-09 11:48 ` Richard Biener
2019-08-06 15:43 ` [PATCH 9/9] Remove alias set comparison Martin Liska
2019-08-07 15:58 ` Martin Sebor
2019-08-08 8:43 ` Martin Liška
2019-08-08 15:21 ` Martin Sebor
2019-08-08 14:44 ` Jeff Law
2019-08-06 15:43 ` [PATCH 2/9] operand_equal_p: add support for FIELD_DECL Martin Liska
2019-08-07 12:21 ` Richard Biener
2019-08-15 14:19 ` Jan Hubicka
2019-08-16 9:28 ` Richard Biener
2019-08-16 12:17 ` Jan Hubicka
2019-09-11 12:58 ` Martin Liška
2019-08-06 15:43 ` [PATCH 4/9] Strengthen alias_ptr_types_compatible_p in LTO mode Martin Liska
2019-08-07 12:05 ` Richard Biener
2019-08-08 12:09 ` Martin Liška
2019-08-09 11:20 ` Richard Biener
2019-08-06 15:43 ` [PATCH 8/9] Remove comparison for polymorphic types Martin Liska
2019-08-06 15:43 ` [PATCH 1/9] Replace int with boolean in predicate functions Martin Liska
2019-08-07 12:38 ` Richard Biener
2019-08-06 15:55 ` [PATCH 3/9] operand_equal_p: add support for OBJ_TYPE_REF Martin Liska
2019-08-07 12:09 ` Richard Biener
2019-08-15 15:44 ` Jan Hubicka
2019-08-16 9:25 ` Richard Biener
2019-08-16 12:11 ` Jan Hubicka
2019-08-19 14:03 ` Richard Biener [this message]
2019-08-19 15:12 ` Jan Hubicka
2019-08-20 14:29 ` Richard Biener
2019-08-20 14:42 ` Jan Hubicka
2019-09-13 12:30 ` Martin Liška
2019-09-16 6:45 ` Richard Biener
2019-08-16 11:53 ` [PATCH 10/N] Use const_tree more in IPA ICF Martin Liška
2019-08-19 13:57 ` Richard Biener
2019-10-30 11:54 ` [PATCH 0/9] IPA ICF overhaul Martin Liška
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='CAFiYyc0GrdzuZp8zTweJSiKwCm747RyW--hPCoa973=OBOVQxA@mail.gmail.com' \
--to=richard.guenther@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=hubicka@ucw.cz \
--cc=mliska@suse.cz \
/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).