From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:50478) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1S2nL3-0005e4-U7 for qemu-devel@nongnu.org; Wed, 29 Feb 2012 12:29:23 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1S2nKx-0007eT-DB for qemu-devel@nongnu.org; Wed, 29 Feb 2012 12:29:17 -0500 Received: from v220110690675601.yourvserver.net ([78.47.199.172]:44337) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1S2nKx-0007dA-6x for qemu-devel@nongnu.org; Wed, 29 Feb 2012 12:29:11 -0500 Message-ID: <4F4E6064.1060608@weilnetz.de> Date: Wed, 29 Feb 2012 18:29:08 +0100 From: Stefan Weil MIME-Version: 1.0 References: <20120229161927.GC8661@redhat.com> In-Reply-To: <20120229161927.GC8661@redhat.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] full valrind report List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Michael S. Tsirkin" Cc: qemu-devel@nongnu.org Am 29.02.2012 17:19, schrieb Michael S. Tsirkin: > Here's a full report of possible leaks: > Any idea? I am invedtigating. Hi Michael, try valgrind with --track-origins=yes. It costs some memory, but improves diagnostics not only for memory leaks. Most important are the leaks marked with "definitely lost". A lot of them are just missing destructors when QEMU terminates. Some QEMU classes provide an init function, but no exit function, for example. If you suspect a leak, you can try to re-run QEMU and look whether it is possible to increase the leak: repeat an action in the QEMU monitor several times, connect to the VNC server more than once, let the emulation run for a long time and so on. This kind of leaks is dangerous for long running QEMU instances or allows denial of service attacks. Good (bug) hunting! Stefan W.