From: Jakub Jelinek <jakub@redhat.com>
To: Bernd Edlinger <bernd.edlinger@hotmail.de>
Cc: Mike Stump <mikestump@comcast.net>,
"H.J. Lu" <hjl.tools@gmail.com>,
"gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>,
Dmitry Vyukov <dvyukov@google.com>
Subject: Re: [PATCH] Fix sporadic failure in g++.dg/tsan/aligned_vs_unaligned_race.C
Date: Tue, 06 Jan 2015 09:16:00 -0000 [thread overview]
Message-ID: <20150106091633.GF1667@tucnak.redhat.com> (raw)
In-Reply-To: <DUB118-W468072AA78C7E7CD1DCF6BE4590@phx.gbl>
On Tue, Jan 06, 2015 at 10:08:22AM +0100, Bernd Edlinger wrote:
> Hi Mike,
>
> after some hours of sleep I realized that your step function can do something very interesting,
> (which you already requested previously):
>
> That is: create a race condition that is _always_ at 100% missed by tsan:
>
> cat lib.c
> /* { dg-do compile } */
> /* { dg-options "-O2 -fno-sanitize=all" } */
> Â
> static volatile int serial = 0;
> Â
> void step (int i)
> {
> Â while (__atomic_load_n (&serial, __ATOMIC_ACQUIRE) != i - 1);
> Â __atomic_store_n (&serial, i, __ATOMIC_RELEASE);
> }
Such busy waiting is not very CPU time friendly in the testsuite, especially
if you have just a single HW thread.
If libtsan is not fixed not to be so racy, perhaps instead of all the sleeps
we could arrange (on x86_64-linux, which is the only target supporting tsan
right now) to make sure the thread runs on the same core/hw thread as the
initial thread using pthread_[gs]etaffinity_np/pthread_attr_setaffinity_np ?
Does tsan then always report the races when the threads can't run
concurrently?
Jakub
next prev parent reply other threads:[~2015-01-06 9:16 UTC|newest]
Thread overview: 49+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-01-03 9:01 Bernd Edlinger
2015-01-03 9:51 ` Mike Stump
2015-01-03 11:20 ` Bernd Edlinger
2015-01-04 17:00 ` Bernd Edlinger
2015-01-04 19:07 ` Mike Stump
2015-01-04 19:17 ` Jakub Jelinek
2015-01-04 19:44 ` Bernd Edlinger
2015-01-04 22:19 ` Mike Stump
2015-01-05 8:49 ` Bernd Edlinger
2015-01-05 20:58 ` Mike Stump
2015-01-05 22:02 ` Mike Stump
2015-01-06 1:07 ` Bernd Edlinger
2015-01-06 9:08 ` Bernd Edlinger
2015-01-06 9:16 ` Jakub Jelinek [this message]
2015-01-06 9:38 ` Bernd Edlinger
2015-01-06 17:45 ` Bernd Edlinger
2015-01-06 19:48 ` Mike Stump
2015-01-06 23:22 ` Bernd Edlinger
2015-01-07 0:33 ` Mike Stump
2015-01-07 7:17 ` Bernd Edlinger
2015-01-07 8:23 ` Jakub Jelinek
2015-01-07 14:55 ` Bernd Edlinger
2015-01-07 15:04 ` Jakub Jelinek
2015-01-07 16:58 ` Mike Stump
2015-01-07 17:00 ` Jakub Jelinek
2015-01-07 18:21 ` Bernd Edlinger
2015-01-07 18:32 ` Jakub Jelinek
2015-01-07 22:44 ` Bernd Edlinger
2015-01-08 19:24 ` Mike Stump
2015-01-08 19:29 ` Jakub Jelinek
2015-01-08 21:07 ` Mike Stump
2015-01-08 21:27 ` Jakub Jelinek
2015-01-08 22:06 ` Mike Stump
2015-01-08 22:23 ` Bernd Edlinger
2015-01-09 15:36 ` Bernd Edlinger
2015-01-09 15:37 ` Jakub Jelinek
2015-01-19 8:53 ` Dmitry Vyukov
2015-01-19 15:16 ` Mike Stump
2015-01-21 8:52 ` Dmitry Vyukov
2015-01-21 9:02 ` Jakub Jelinek
2015-01-21 9:07 ` Dmitry Vyukov
2015-01-08 19:10 ` Mike Stump
2015-01-07 16:36 ` Mike Stump
2015-01-06 21:29 ` Mike Stump
2015-01-04 19:05 ` Mike Stump
2015-01-04 19:15 ` Jakub Jelinek
2015-01-04 21:48 ` Bernd Edlinger
2015-01-04 21:58 ` Mike Stump
2015-01-04 22:20 ` Bernd Edlinger
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=20150106091633.GF1667@tucnak.redhat.com \
--to=jakub@redhat.com \
--cc=bernd.edlinger@hotmail.de \
--cc=dvyukov@google.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=hjl.tools@gmail.com \
--cc=mikestump@comcast.net \
/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).