From: Paolo Bonzini <bonzini@gnu.org>
To: Jeff Law <law@redhat.com>
Cc: gcc-patches@gcc.gnu.org
Subject: Re: RFA: New pass to delete unexecutable paths in the CFG
Date: Wed, 09 Nov 2011 08:37:00 -0000 [thread overview]
Message-ID: <4EBA3543.10708@gnu.org> (raw)
In-Reply-To: <4EB99006.4060501@redhat.com>
On 11/08/2011 09:24 PM, Jeff Law wrote:
> We don't have access to those assertions as they're removed well prior
> to this pass running. However, if we did, or if we had redundant PHIs
> in the stream which were propagated we'd be presented with something like
>
> BB0 if (p_1) goto BB1 else goto BB2
>
> BB1: ... goto BB3
> BB2:
> BB3: p_2 = phi (p_1 (BB1), 0(BB2))
> *p_2 = 2;
>
>
> We'd recognize that the edge bb2->bb3 is unexecutable as doing so
> leads to a NULL pointer dereference. Since the edge bb2->bb3 is not a
> critical edge, we know that bb2 as a whole is unexecutable. bb2 is
> control dependent on the edge bb0->bb2.
(Side note regarding critical edges: have you tried splitting them
before your pass?)
> We would remove the edge bb0->bb2 and the control statement if (p_1)
> .... That makes BB2 unreachable resulting in
>
> BB0 goto BB1
> BB1 ...
> BB3 p_2 = phi (p_1)
> *p_2 = 2;
>
> Which would then be optimized into
>
> BB0: ...
> *p_1 = 2;
>
> Which is exactly what I would expect the code to do with the knowledge
> that passing 0 to f results in undefined behaviour.
Ok, so that's exactly what I was thinking about. In this case the
optimization is obviously allowed by the C standard; you have
if (p)
something;
*p = 0;
and the "*p = 0" has been in some sense translated to
if (!p)
something;
*p = 0;
which is only different on undefined paths. But I'm not sure that more
complicated cases, where there are other instructions between the "if"
and "*p = 0", would also be allowed by the C standard. For example, I
think a function call in the "else" branch, or between the PHI and the
dereference should prevent the optimization, because the function call
might never return for what we know. Probably a volatile asm too. Does
your patch do that? (Testcases! :)).
In general, this is quite different from all other existing GCC
optimizations based on undefined behavior. Whenever you trigger
undefined behavior, right now the effects do not extend *before* the
undefined operation. The proposed pass would change that, so that its
effects are a bit more surprising when debugging. If your bug is that
you forgot a "return;" in the else branch, you surely wouldn't expect
the compiler to swallow the entire branch. Unfortunately debugging at
-O0 is not always an option.
Paolo
next prev parent reply other threads:[~2011-11-09 8:10 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-11-07 9:58 Jeff Law
2011-11-07 10:19 ` Jakub Jelinek
2011-11-07 10:21 ` Richard Guenther
2011-11-07 10:30 ` Richard Guenther
2011-11-07 19:20 ` Jeff Law
2011-11-07 16:14 ` Jeff Law
2011-11-07 16:30 ` Richard Guenther
2011-11-07 16:57 ` Kai Tietz
2011-11-07 19:03 ` Jeff Law
2011-11-08 11:50 ` Paolo Bonzini
2011-11-08 19:48 ` Jeff Law
2011-11-08 20:38 ` Paolo Bonzini
2011-11-08 20:59 ` Jeff Law
2011-11-09 8:37 ` Paolo Bonzini [this message]
2011-11-09 18:11 ` Jeff Law
2011-11-09 18:12 ` Jakub Jelinek
2011-11-09 22:45 ` Paolo Bonzini
2011-11-10 19:27 ` Jeff Law
2011-11-07 19:14 ` Jeff Law
2011-11-07 14:16 ` Tom Tromey
2011-11-07 15:54 ` Jeff Law
2011-11-07 15:54 ` Richard Guenther
2011-11-07 19:09 ` Jeff Law
2011-11-07 22:34 ` Richard Guenther
2011-11-08 20:02 ` Jeff Law
2011-11-09 9:50 ` Richard Guenther
2011-11-09 17:43 ` Jeff Law
2011-11-07 15:55 ` Tom Tromey
2011-11-07 17:01 ` Paolo Bonzini
2011-11-15 7:52 ` RFA: disable -fdelete-null-pointer-checks for Java Jeff Law
2011-11-07 19:05 ` RFA: New pass to delete unexecutable paths in the CFG 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=4EBA3543.10708@gnu.org \
--to=bonzini@gnu.org \
--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).