From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <sourceware-bugzilla@sourceware.org>
Received: by sourceware.org (Postfix, from userid 48)
	id 1DFAB385B535; Fri,  5 May 2023 23:44:34 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 1DFAB385B535
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=sourceware.org;
	s=default; t=1683330275;
	bh=TR0vQEwp2CyP9xxvnmpQtr/ssdREVauKBZheeAXrcac=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=nSeSib8CMrBVYhk7BrcyHmNTcc3yddfOqJqRy6On5xQZcvCvuDYY/RXX9HBewg4hM
	 1U7oeboC1XteKakdbFaxMudiPakUWYH8bLfrw7AMgs/Gvx+6AMqtz93MkQy4pJjgJL
	 voPcsT1SM1nGpzJ3sxe/T7ZA6dv8RUXuT9yGWWHQ=
From: "carlos at redhat dot com" <sourceware-bugzilla@sourceware.org>
To: glibc-bugs@sourceware.org
Subject: [Bug nptl/25847] pthread_cond_signal failed to wake up
 pthread_cond_wait due to a bug in undoing stealing
Date: Fri, 05 May 2023 23:44:32 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: glibc
X-Bugzilla-Component: nptl
X-Bugzilla-Version: 2.27
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: carlos at redhat dot com
X-Bugzilla-Status: ASSIGNED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: carlos at redhat dot com
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-25847-131-Yh5ntPVgzi@http.sourceware.org/bugzilla/>
In-Reply-To: <bug-25847-131@http.sourceware.org/bugzilla/>
References: <bug-25847-131@http.sourceware.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://sourceware.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <glibc-bugs.sourceware.org>

https://sourceware.org/bugzilla/show_bug.cgi?id=3D25847

--- Comment #59 from Carlos O'Donell <carlos at redhat dot com> ---
I can walk the reproducer with 8 threads to the point where the state looks
like this:

$210 =3D "pthread_cond_t" =3D {
__wseq >> 1 =3D 4
__wseq G2 =3D 1
__g1_start >> 1 =3D 2
__g1_start G2 =3D 1
__g1_orig_size >> 2 =3D 2
__g1_orig_size UNLOCKED =3D 0
__wrefs >> 3 =3D 4
__g_refs    =3D { 0 / 0 , 0 / 0 }
__g_signals =3D { 0 / 0, 0 / 0 }
__g_size    =3D { 1 , 0 }
__wrefs CLOCK_REALTIME
__wrefs Private
 =3D }

At this point a pthread_cond_signal will consider a quiesce based on
__g_size[0] (since __wseq G2 is 1, so G1 is 0) and will not because __g_siz=
e[0]
=3D=3D 1.
The thread will not see a __g1_start larger than its own sequence number.

I agree that at this point we seem stuck with a lost signal because __g_siz=
e[0]
=3D=3D 1 out of sync with the delivery of the signal.

--=20
You are receiving this mail because:
You are on the CC list for the bug.=