From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jan Kiszka Subject: Re: segmentation fault in qemu-kvm-0.14.0 Date: Wed, 09 Mar 2011 12:44:12 +0100 Message-ID: <4D77680C.2030908@siemens.com> References: <2640D58E-2101-47FA-99B6-28815666651E@dlh.net> <4D772E4C.6020604@web.de> <4D77629A.204@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: Peter Lieven , qemu-devel , kvm@vger.kernel.org To: Paolo Bonzini Return-path: Received: from david.siemens.de ([192.35.17.14]:30787 "EHLO david.siemens.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753951Ab1CILo2 (ORCPT ); Wed, 9 Mar 2011 06:44:28 -0500 In-Reply-To: <4D77629A.204@redhat.com> Sender: kvm-owner@vger.kernel.org List-ID: On 2011-03-09 12:20, Paolo Bonzini wrote: > On 03/09/2011 08:37 AM, Jan Kiszka wrote: >> It's probably worth validating that the iothread lock is >> always held when qemu_set_fd_handler2 is invoked to confirm this race >> theory, adding something like >> >> assert(pthread_mutex_trylock(&qemu_mutex) != 0); >> (that's for qemu-kvm only) > > Alternatively, iohandlers could be a(nother) good place to start > introducing fine-grained locks or rwlocks. Yeah, could be a good idea. It's a fairly confined area here that needs protection. Jan -- Siemens AG, Corporate Technology, CT T DE IT 1 Corporate Competence Center Embedded Linux