From: Richard Biener <richard.guenther@gmail.com>
To: Qing Zhao <qing.zhao@oracle.com>
Cc: "Martin Jambor" <mjambor@suse.cz>,
"GCC Patches" <gcc-patches@gcc.gnu.org>,
"Martin Liška" <mliska@suse.cz>,
live-patching@vger.kernel.org, "Jan Hubicka" <hubicka@ucw.cz>
Subject: Re: GCC options for kernel live-patching (Was: Add a new option to control inlining only on static functions)
Date: Thu, 04 Oct 2018 13:20:00 -0000 [thread overview]
Message-ID: <CAFiYyc30SZd+HUGS26bbhx5UyEsSxBy=iwHng+eK0UN0Jb4Ekw@mail.gmail.com> (raw)
In-Reply-To: <F6E335BD-4ADE-4E1B-BC6B-6BDFF4A9E117@oracle.com>
On Wed, Oct 3, 2018 at 6:00 PM Qing Zhao <qing.zhao@oracle.com> wrote:
>
>
> > On Oct 3, 2018, at 5:19 AM, Martin Jambor <mjambor@suse.cz> wrote:
> >
> > Hi,
> >
> > On Tue, Oct 02 2018, Qing Zhao wrote:
> >> So, here comes the next question:
> >>
> >> what will be the better approach for helping live-patching?
> >>
> >> A. disable all the optimization/analyzes that changing a function based on other functions. this includes most of
> >> the IPA optimizations, inlining, cloning, ipa side effect analysis, etc.
> >
> > We came to the conclusion that we do not want to disable (any) inlining
> > - primarily for performance reasons but also because of always_inline.
> > When you have to deal with inlining, you can use the same mechanism for
> > IPA-CP clones, IPA-SRA and anything that creates clones and callers need
> > to be patched.
> >
> > Always_inline is also the reason why I would discourage anyone from
> > attempting this, it looks too much like an irresolvable conflict.
>
> Okay.
>
> However, there are also other users of live-patching who request only inlining static routines to control the code size explosion and debugging
> complexity as the start of the whole discussion.
>
> in addition to control inlining, I am not sure whether we want to control ipa-cp, ipa-sra, etc to further control the code size and debugging complexity?
>
> we should also include such request into this work.
>
> >
> >>
> >> B. enable all such optimizations, but compute the list of possibly affected functions from those IPA optimizations.
> >
> > Although certainly possible, we came to the conclusion that it would be
> > impractical to attempt to track all decisions affected by, say, a
> > function being pure, or a variable read-only, etc. Impractical to the
> > point when at least I personally would also discourage anyone from
> > attempting this too, it would be necessary to record this info deep
> > inside AA, VRP, who know what else, and keep it up to date.
>
> I think that theoretically the compiler should be able to do this. But the impacted routine list might be huge. that also lead problems of code size explosion and debugging complexity.
I think it boils down to have a "global" lattice of non-VARYING-X
"because of $FN" for each of the analyses X properly
propagated along the callgraph.
Something as "simple" as annotating each function output or inlined
with its lattice state "PURE, 1st param ~[0,0]", etc.
could be helpful to spot differences before/after the patch and raise a flag.
Richard.
> >
> >>
> >> C. enable part of them and disable the other part of them. for the part of enabled ones, compute the list of possibly affected functions
> >> from them.
> >
> > Which means we would like to go this route.
> Qing
> >
> > Martin
> >
>
next prev parent reply other threads:[~2018-10-04 12:47 UTC|newest]
Thread overview: 124+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-18 18:58 [PATCH][Middle-end][Version 3]Add a new option to control inlining only on static functions Qing Zhao
2018-09-18 23:18 ` Martin Sebor
2018-09-19 15:07 ` Qing Zhao
2018-09-21 15:19 ` [PATCH][Middle-end][Version 4]Add " Qing Zhao
2018-09-26 1:01 ` PING: " Qing Zhao
2018-09-26 11:20 ` Alexander Monakov
2018-09-26 12:49 ` Paolo Carlini
2018-09-26 14:43 ` Qing Zhao
2018-09-26 15:18 ` Alexander Monakov
2018-09-26 23:07 ` Qing Zhao
2018-09-26 23:54 ` Alexander Monakov
2018-09-27 5:16 ` Qing Zhao
2018-09-27 7:46 ` Richard Biener
2018-09-27 16:30 ` Qing Zhao
2018-09-27 8:55 ` Alexander Monakov
2018-09-27 8:58 ` Jan Hubicka
2018-09-27 11:12 ` Richard Biener
2018-09-26 13:24 ` Jason Merrill
2018-09-26 13:31 ` Richard Biener
2018-09-26 13:40 ` Jason Merrill
2018-09-26 14:46 ` Jeff Law
2018-09-26 14:58 ` Jason Merrill
2018-09-26 15:10 ` Jan Hubicka
2018-09-26 17:12 ` Qing Zhao
2018-09-26 17:22 ` Jan Hubicka
2018-09-26 21:36 ` Qing Zhao
2018-09-27 9:47 ` Jan Hubicka
2018-09-27 12:29 ` GCC options for kernel live-patching (Was: Add a new option to control inlining only on static functions) Martin Jambor
2018-09-27 16:40 ` Qing Zhao
2018-10-01 17:14 ` Qing Zhao
2018-10-02 8:38 ` Martin Jambor
2018-10-02 14:00 ` Qing Zhao
2018-10-02 14:11 ` Martin Liška
2018-10-02 14:55 ` Qing Zhao
2018-10-02 14:59 ` Martin Liška
2018-10-02 15:24 ` Qing Zhao
2018-10-02 17:16 ` Martin Liška
2018-10-02 18:34 ` Richard Biener
2018-10-02 21:25 ` Qing Zhao
2018-10-03 9:05 ` Martin Liška
2018-10-03 10:53 ` Martin Jambor
2018-10-03 16:11 ` Qing Zhao
2018-10-04 13:20 ` Richard Biener [this message]
2018-10-18 19:36 ` [RFC] GCC support for live-patching Qing Zhao
2018-10-19 8:53 ` Bernhard Reutner-Fischer
2018-10-19 18:33 ` Qing Zhao
2018-10-24 21:16 ` Alexandre Oliva
2018-10-19 14:25 ` Andi Kleen
2018-10-19 18:33 ` Qing Zhao
2018-10-20 6:47 ` Andi Kleen
2018-10-22 13:07 ` Martin Jambor
2018-10-22 17:59 ` Qing Zhao
2018-10-22 15:36 ` Miroslav Benes
2018-10-22 21:01 ` Qing Zhao
2018-10-23 9:37 ` Miroslav Benes
2018-10-23 19:54 ` Qing Zhao
2018-10-24 13:17 ` Miroslav Benes
2018-10-23 13:37 ` Nicolai Stange
2018-10-23 20:34 ` Qing Zhao
2018-10-23 21:18 ` Nicolai Stange
2018-10-24 17:19 ` Qing Zhao
2018-10-31 22:04 ` [RFC] [2nd version] " Qing Zhao
2018-10-31 22:08 ` [RFC] " Qing Zhao
2018-10-03 9:23 ` GCC options for kernel live-patching (Was: Add a new option to control inlining only on static functions) Jan Hubicka
2018-10-03 11:27 ` Martin Liška
2018-10-22 10:49 ` Martin Liška
[not found] ` <20181105095135.j3mnzox6rkktkoto@kam.mff.cuni.cz>
2018-11-05 22:26 ` Qing Zhao
2018-11-07 14:22 ` Martin Liška
2018-11-07 14:27 ` Jan Hubicka
2018-11-07 14:44 ` Martin Liška
2018-11-08 14:59 ` Jan Hubicka
2018-11-09 15:33 ` [PATCH][RFC] Come up with -flive-patching master option Martin Liška
2018-11-09 17:43 ` Qing Zhao
2018-11-10 8:51 ` Martin Liška
2018-11-12 2:28 ` Qing Zhao
2018-11-12 8:53 ` Martin Liška
2018-11-12 22:29 ` Qing Zhao
2018-11-13 17:49 ` Qing Zhao
2018-11-13 19:18 ` Miroslav Benes
2018-11-13 21:16 ` Qing Zhao
2018-11-14 15:03 ` Martin Liška
2018-11-14 17:54 ` Qing Zhao
2018-11-15 8:41 ` Martin Liška
2018-11-15 15:41 ` Qing Zhao
2018-11-16 1:36 ` [PATCH]Come " Qing Zhao
2018-11-16 15:26 ` Martin Liška
2018-11-16 15:51 ` Jan Hubicka
2018-11-16 16:25 ` Qing Zhao
2018-11-16 16:05 ` Qing Zhao
2018-11-19 8:12 ` Martin Liška
[not found] ` <F78B52B9-9F9A-4FA0-91BF-A33307D87AA8@oracle.c! ! om>
2018-11-19 15:53 ` Qing Zhao
2018-11-20 12:10 ` Martin Liška
2018-11-20 15:32 ` [PATCH][Version 3]Come " Qing Zhao
2018-11-26 15:54 ` PING: " Qing Zhao
2018-11-30 17:04 ` [wwwdocs] [PATCH]introduce new -flive-patching master option into gcc9 Qing Zhao
2018-11-30 18:29 ` Jeff Law
2018-11-30 20:44 ` Qing Zhao
2018-11-28 15:52 ` [PATCH][Version 3]Come up with -flive-patching master option Jan Hubicka
2018-11-28 20:25 ` Qing Zhao
2018-11-29 16:16 ` Qing Zhao
2018-12-05 23:16 ` Question on Disable no throw for " Qing Zhao
2019-01-02 11:47 ` Martin Liška
2019-01-09 11:28 ` Martin Liška
2018-12-07 13:07 ` [PATCH][Version 3]Come up with " Rainer Orth
2018-12-07 15:14 ` Qing Zhao
2019-04-10 14:46 ` Jonathan Wakely
2019-04-10 19:24 ` Qing Zhao
2019-04-10 20:01 ` Jonathan Wakely
2019-04-11 8:11 ` Richard Biener
2018-11-14 22:05 ` [PATCH][RFC] Come " Miroslav Benes
[not found] ` <20181110170343.g3k7j7rlydid3ahr@kam.mff.cuni.cz>
2018-11-12 9:29 ` Martin Liška
2018-11-09 18:38 ` Bernhard Reutner-Fischer
2018-11-10 8:48 ` Martin Liška
2018-10-03 12:18 ` GCC options for kernel live-patching (Was: Add a new option to control inlining only on static functions) Martin Jambor
2018-10-03 16:00 ` Qing Zhao
2018-10-23 12:34 ` Performance impact of disabling non-clone IPA optimizations for the Linux kernel (was: "GCC options for kernel live-patching") Nicolai Stange
2018-10-24 14:30 ` Jiri Kosina
2018-10-24 14:44 ` Miroslav Benes
2018-09-27 16:32 ` [PATCH][Middle-end][Version 4]Add a new option to control inlining only on static functions Qing Zhao
2018-09-26 22:42 ` Qing Zhao
2018-09-26 16:02 ` Qing Zhao
2018-09-26 15:52 ` Qing Zhao
2018-09-26 16:02 ` Jan Hubicka
2018-09-26 18:51 ` Qing Zhao
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='CAFiYyc30SZd+HUGS26bbhx5UyEsSxBy=iwHng+eK0UN0Jb4Ekw@mail.gmail.com' \
--to=richard.guenther@gmail.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=hubicka@ucw.cz \
--cc=live-patching@vger.kernel.org \
--cc=mjambor@suse.cz \
--cc=mliska@suse.cz \
--cc=qing.zhao@oracle.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).