From: Arnd Bergmann <arnd@arndb.de>
To: libc-alpha@sourceware.org
Cc: Jeff Law <law@redhat.com>, Florian Weimer <fweimer@redhat.com>,
Carlos O'Donell <carlos@redhat.com>
Subject: Re: [PATCH] Fix -Os related -Werror failures.
Date: Fri, 28 Oct 2016 08:12:00 -0000 [thread overview]
Message-ID: <20863164.XNWC5rYB1g@wuerfel> (raw)
In-Reply-To: <9d58289e-07fb-4bae-d7d3-8055a6c96a3a@redhat.com>
On Friday, October 28, 2016 12:44:32 AM CEST Jeff Law wrote:
> On 10/28/2016 12:32 AM, Florian Weimer wrote:
> > On 10/28/2016 06:46 AM, Carlos O'Donell wrote:
> >> +/* With GCC 5.3 when compiling with -Os the compiler emits a warning
> >> + that buf[0] and buf[1] may be used uninitialized. This can only
> >> + happen in the case where tmpbuf[3] is used, and in that case the
> >> + write to the tmpbuf[1] and tmpbuf[2] was assured because
> >> + ucs4_to_cns11643 would have filled in those entries. The difficulty
> >> + is in getting the compiler to see this logic because tmpbuf[0] is
> >> + involved in determining the code page and is the indicator that
> >> + tmpbuf[2] is initialized. */
> >> +DIAG_PUSH_NEEDS_COMMENT;
> >> +DIAG_IGNORE_NEEDS_COMMENT (5.3, "-Wmaybe-uninitialized");
> >
> > This hides the warning for -O2 builds as well, so I don't think this is
> > a good idea.
> >
> > Those who want to build with -Os or other special compiler flags should
> > just configure with --disable-werror. We can't account for every
> > optimization someone might want to disable in their build.
> That'd be my recommendation.
>
> What often happens in these cases is the compiler in its default mode of
> operation is able to statically eliminate a conditional branch on a
> particular path. However, to do so the compiler has to duplicate code.
>
> Not surprisingly, there's a cost/benefit tradeoff here and the
> heuristics are largely driven by the real or estimated profile data as
> well as the coarser "optimize for code space". So changing flags
> changes the output of those heuristics and ultimately can result in
> leaving paths in the CFG that can not be executed -- and that often
> leads to false positive may-be-uninitialized warnings and such.
>
> Long term I would like to find a good way to mark paths that are not
> executable, but are not profitable to eliminate, then utilize that
> information to prune various "may" warnings. That would make those kind
> of warnings more stable across different optimization levels as well as
> more stable release-to-release. But that's definitely in the "future
> work" area.
I've spent a lot of time trying to eliminate -Wmaybe-uninitialized
warnings from the Linux kernel. Here are some data points that you
may find useful too:
- Building with -Os causes many false positives starting with gcc-4.9,
and I have disabled the warning for this specific flag. I believe
this is due to the lack of the "-fschedule-insns" optimization step
- Building with -O3 apparently also causes some false positives, but
we don't normally do that in the kernel, and the only architecture
port that does it also disables the warnings
- Two more gcc options that cause false positives are -fprofile-arcs
and some of the -fsanitize=... options
- overall, gcc-4.9 improved much over gcc-4.8 in these warnings,
but they have a different set of false-positives. As gcc-4.8 is
getting old, I'm pushing a patch to also disable the warning
for all 4.8 builds. Prior to v4.8, there was no option to disable
maybe-uninitialized warnings.
- gcc-5 and gcc-6 appear to be slightly better than gcc-4.9 but also
introduce a small number of additional false-positive warnings,
apparently this happens mostly because they make different
inlining decisions, not because something fundamentally changed.
Generally speaking, if any of 4.9, 4.x or 5.x produce a warning
in some configurations, it's likely that the other ones will
do the same, depending on a combination target architecture and
optimization flags that impact inlining.
- I found that most often when gcc is confused about whether a
variable is uninitialized or not, the source code tends to be
confusing to a human reader as well and rewriting it differently
results in better readability and better object code while
avoiding the warning. There are always other cases in which
this is not possible though.
Arnd
next prev parent reply other threads:[~2016-10-28 8:12 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-28 4:48 Carlos O'Donell
2016-10-28 6:25 ` Andreas Schwab
2016-10-28 6:32 ` Florian Weimer
2016-10-28 6:44 ` Jeff Law
2016-10-28 8:12 ` Arnd Bergmann [this message]
2016-10-28 8:17 ` Andrew Pinski
2016-10-28 13:28 ` Jeff Law
2016-10-28 20:10 ` Paul Eggert
2016-10-29 3:03 ` Jeff Law
2016-10-30 4:25 ` Paul Eggert
2016-10-28 12:09 ` Carlos O'Donell
2016-10-28 12:43 ` Florian Weimer
2016-10-28 13:04 ` Joseph Myers
2016-10-28 13:07 ` Carlos O'Donell
2016-10-28 12:49 ` Joseph Myers
2016-10-28 12:55 ` Florian Weimer
2016-10-28 13:18 ` Carlos O'Donell
2016-10-28 13:58 ` [PATCH v2] Fix -Os related build and test failures Carlos O'Donell
2016-10-28 14:17 ` Joseph Myers
2016-10-29 2:59 ` [PATCH v3] " Carlos O'Donell
2016-10-29 3:26 ` Carlos O'Donell
2016-10-29 17:35 ` Joseph Myers
2016-10-30 3:51 ` [PATCH v4] " Carlos O'Donell
2016-10-31 8:33 ` Andreas Schwab
2016-10-31 9:16 ` Carlos O'Donell
2016-10-31 9:22 ` Florian Weimer
2016-10-31 12:56 ` David Miller
2016-10-31 19:56 ` Carlos O'Donell
2016-11-01 22:59 ` Joseph Myers
2016-11-02 12:52 ` Carlos O'Donell
2016-11-01 9:17 ` Andreas Schwab
2016-11-01 11:13 ` Joseph Myers
2016-11-01 15:58 ` Tamar Christina
2016-11-01 16:06 ` David Miller
2016-11-01 16:15 ` Tamar Christina
2016-11-02 11:53 ` Carlos O'Donell
2016-11-02 17:03 ` Carlos O'Donell
2016-11-02 13:22 ` Carlos O'Donell
2016-10-31 18:38 ` [PATCH v3] " Steve Ellcey
2016-10-31 19:50 ` Carlos O'Donell
2016-10-31 19:57 ` Steve Ellcey
2016-10-31 20:50 ` Carlos O'Donell
2016-10-31 21:00 ` Steve Ellcey
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=20863164.XNWC5rYB1g@wuerfel \
--to=arnd@arndb.de \
--cc=carlos@redhat.com \
--cc=fweimer@redhat.com \
--cc=law@redhat.com \
--cc=libc-alpha@sourceware.org \
/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).