Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S932541AbcLSShZ (ORCPT <rfc822;w@1wt.eu>);
        Mon, 19 Dec 2016 13:37:25 -0500
Received: from mail.kernel.org ([198.145.29.136]:48894 "EHLO mail.kernel.org"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S1753183AbcLSShY (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
        Mon, 19 Dec 2016 13:37:24 -0500
Date: Mon, 19 Dec 2016 15:37:16 -0300
From: Arnaldo Carvalho de Melo <acme@kernel.org>
To: Markus Trippelsdorf <markus@trippelsdorf.de>
Cc: linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>
Subject: Re: *** buffer overflow detected ***: /usr/src/linux/tools/perf/perf
 terminated
Message-ID: <20161219183716.GB551@kernel.org>
References: <20161219161821.GA294@x4>
 <20161219165243.GA289@x4>
 <20161219172842.GB289@x4>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20161219172842.GB289@x4>
X-Url: http://acmel.wordpress.com
User-Agent: Mutt/1.7.1 (2016-10-04)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org
Content-Length: 1658
Lines: 46

Em Mon, Dec 19, 2016 at 06:28:42PM +0100, Markus Trippelsdorf escreveu:
> On 2016.12.19 at 17:52 +0100, Markus Trippelsdorf wrote:
> > On 2016.12.19 at 17:18 +0100, Markus Trippelsdorf wrote:
> > > Running the latest kernel git tree, I get buffer overflow warnings when
> > > I try to run "perf top":
> > > 
> > > *** buffer overflow detected ***: /usr/src/linux/tools/perf/perf terminated        
> > 
> > It only happens when I link my kernel with gold.
<SNIP>
> > And perf than chokes when parsing /sys/kernel/notes.
> 
> Started with gold commit:
> 
>  commit bfbf34de2f27672282d4fa4c0916cc41890dcc6c
>  Author: Cary Coutant <ccoutant@gmail.com>
>  Date:   Mon Dec 12 18:51:29 2016 -0800
> 
>     When using linker scripts, place linker-generated sections by the output section name.
> 
> I've opened a gold bug:
> https://sourceware.org/bugzilla/show_bug.cgi?id=20983

I guess we need something along the lines of below, but I think the max
should be set to the filesize, have to spend some more time here, using
stat() in such case, etc, wdyt?

- Arnaldo

diff --git a/tools/perf/util/symbol-elf.c b/tools/perf/util/symbol-elf.c
index 99400b0e8f2a..adbc6c02c3aa 100644
--- a/tools/perf/util/symbol-elf.c
+++ b/tools/perf/util/symbol-elf.c
@@ -537,6 +537,12 @@ int sysfs__read_build_id(const char *filename, void *build_id, size_t size)
 				break;
 		} else {
 			int n = namesz + descsz;
+
+			if (n > (int)sizeof(bf)) {
+				n = sizeof(bf);
+				pr_debug("%s: truncating reading of build id in sysfs file %s: n_namesz=%u, n_descsz=%u.\n",
+					 __func__, filename, nhdr.n_namesz, nhdr.n_descsz);
+			}
 			if (read(fd, bf, n) != n)
 				break;
 		}