From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
	id EFB1B3858C5F; Mon, 15 Jan 2024 16:54:36 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org EFB1B3858C5F
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1705337676;
	bh=h20SlR0eAhEUY1CqAWhH8M1aKvDP9GFebTk1vfLl0OY=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=HRytN3/TiEqFCUqQJLaO3EO4fipmWKyyV9R4XkBU9xcYAxL+n5w5biXsG4JOgG9a7
	 nhFBowe6K5iOEXYPZLFMg44eU4OgHH5P3F+oaPS9NqM1OsJeKbA4cJzeY7yHoXO3nX
	 mDFFzSnLZAXmupz7f3V/4AhQaiheBZdXnbhAmw+8=
From: "fw at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libgcc/113401] Memory (resource) leak in -ftrampoline-impl=heap
Date: Mon, 15 Jan 2024 16:54:36 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: libgcc
X-Bugzilla-Version: 14.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: fw at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-113401-4-WF5Vf493E6@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-113401-4@http.gcc.gnu.org/bugzilla/>
References: <bug-113401-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D113401

--- Comment #2 from Florian Weimer <fw at gcc dot gnu.org> ---
(In reply to Iain Sandoe from comment #1)
> (In reply to Florian Weimer from comment #0)
>=20
> > The fix is to register a TLS destructor to
> > deallocate that page, too. On glibc, that also fixes another memory lea=
k for
> > -fno-exceptions compilations (the default for C) if pthread_exit is cal=
led
> > with an active trampoline.
>=20
> Does this mean you have a proposed patch already? (before I start
> investigation)

No, this was a reference to __cxa_thread_atexit, which is unfortunately in
libstdc++ (but is a thin shim around __cxa_thread_atexit_impl for current
glibc). The pthread_key_create fallback probably needs to be duplicated into
libgcc_s.=