From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1M4Z6g-0004QT-9i for qemu-devel@nongnu.org; Thu, 14 May 2009 07:28:10 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1M4Z6b-0004MG-KB for qemu-devel@nongnu.org; Thu, 14 May 2009 07:28:09 -0400 Received: from [199.232.76.173] (port=41975 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1M4Z6b-0004Lu-89 for qemu-devel@nongnu.org; Thu, 14 May 2009 07:28:05 -0400 Received: from mx2.redhat.com ([66.187.237.31]:53718) by monty-python.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1M4Z6a-0004Ab-PU for qemu-devel@nongnu.org; Thu, 14 May 2009 07:28:05 -0400 Message-ID: <4A0C0041.1080809@redhat.com> Date: Thu, 14 May 2009 14:28:01 +0300 From: Avi Kivity MIME-Version: 1.0 References: <1240265600-9469-1-git-send-email-ryanh@us.ibm.com> <1242298581-30587-1-git-send-email-markmc@redhat.com> In-Reply-To: <1242298581-30587-1-git-send-email-markmc@redhat.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: [Qemu-devel] Re: [STABLE][PATCH 0/4] Rebase more aio dma cancellation work List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Mark McLoughlin Cc: Anthony Liguori , qemu-devel@nongnu.org Mark McLoughlin wrote: > Hi, > We've a report of F11 qemu segfaulting when using read-only > images here: > > https://bugzilla.redhat.com/497170 > > Looking at the current code in stable, there is clearly some > borkage: > > - No error handling in bdrv_aio_rw_vector() > > - No error handling in dma_bdrv_cb()/dma_bdrv_io() > > - VectorTranslationState is leaked by bdrv_aio_rw_vector_cb() > > - dma_aio_cancel() is cancelling the wrong aio > > The rest of Avi's aio cancellation patches actually fixes all > these problems. Please apply to stable. > It's a little sad that we have to apply such a non trivial patchset to stable, but I see no way around it. At least it has seen extensive testing in the main branch and in qemu-kvm.git. (Mark, you have to add your signoff if you're submitting a patch, even if you've done no changes to it at all) -- error compiling committee.c: too many arguments to function