From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:34758) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ecGLi-0000jO-3D for qemu-devel@nongnu.org; Thu, 18 Jan 2018 14:59:47 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ecGLf-0006ci-2R for qemu-devel@nongnu.org; Thu, 18 Jan 2018 14:59:46 -0500 Received: from mx1.redhat.com ([209.132.183.28]:45223) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1ecGLe-0006cH-PQ for qemu-devel@nongnu.org; Thu, 18 Jan 2018 14:59:42 -0500 Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.phx2.redhat.com [10.5.11.16]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 258902DA9B8 for ; Thu, 18 Jan 2018 19:59:41 +0000 (UTC) Date: Thu, 18 Jan 2018 19:59:35 +0000 From: "Dr. David Alan Gilbert" Message-ID: <20180118195934.GD2660@work-vm> References: <20180116180408.11279-1-dgilbert@redhat.com> <20180118213321-mutt-send-email-mst@kernel.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180118213321-mutt-send-email-mst@kernel.org> Subject: Re: [Qemu-devel] [PATCH v6 0/7] Rework vhost memory region updates List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: "Michael S. Tsirkin" Cc: qemu-devel@nongnu.org, imammedo@redhat.com, maxime.coquelin@redhat.com * Michael S. Tsirkin (mst@redhat.com) wrote: > On Tue, Jan 16, 2018 at 06:04:01PM +0000, Dr. David Alan Gilbert (git) wrote: > > From: "Dr. David Alan Gilbert" > > > > Hi, > > This patch set reworks the way the vhost code handles changes in > > physical address space layout that came from a discussion with Igor. > > > > Its intention is to simplify a lot of the update code, > > and to make it easier for the postcopy+shared code to > > do the hugepage alignments that are needed. > > > > Instead of inserting/removing each section during the add/del > > callbacks of the listener, we start afresh and build a list > > from the add and nop callbacks, then at the end compare the list > > we've built with the exisiting list. > > > > v6 > > Tidy ups from Igor > > The biggest change is moving the 'Move log_dirty check' to be > > the last patch in the set. > > > > Dr. David Alan Gilbert (7): > > vhost: Build temporary section list and deref after commit > > vhost: Simplify ring verification checks > > vhost: Merge sections added to temporary list > > vhost: Regenerate region list from changed sections list > > vhost: Clean out old vhost_set_memory and friends > > vhost: Merge and delete unused callbacks > > vhost: Move log_dirty check > > > > hw/virtio/trace-events | 6 + > > hw/virtio/vhost.c | 497 ++++++++++++++++------------------------------ > > include/hw/virtio/vhost.h | 5 +- > > 3 files changed, 180 insertions(+), 328 deletions(-) > > > Seems to trigger errors with clang runtime sanitizer: > > /scm/qemu/hw/virtio/vhost.c:425:26: runtime error: null pointer passed as argument 1, which is declared to never be null > /usr/include/string.h:64:33: note: nonnull attribute specified here > /scm/qemu/hw/virtio/vhost.c:425:45: runtime error: null pointer passed as argument 2, which is declared to never be null > /usr/include/string.h:64:33: note: nonnull attribute specified here > /scm/qemu/hw/virtio/vhost.c:425:26: runtime error: null pointer passed as argument 1, which is declared to never be null > /usr/include/string.h:64:33: note: nonnull attribute specified here > /scm/qemu/hw/virtio/vhost.c:425:45: runtime error: null pointer passed as argument 2, which is declared to never be null > /usr/include/string.h:64:33: note: nonnull attribute specified here > /scm/qemu/hw/virtio/vhost.c:425:26: runtime error: null pointer passed as argument 1, which is declared to never be null > /usr/include/string.h:64:33: note: nonnull attribute specified here > /scm/qemu/hw/virtio/vhost.c:425:45: runtime error: null pointer passed as argument 2, which is declared to never be null > /usr/include/string.h:64:33: note: nonnull attribute specified here > /scm/qemu/hw/virtio/vhost.c:425:26: runtime error: null pointer passed as argument 1, which is declared to never be null > /usr/include/string.h:64:33: note: nonnull attribute specified here > /scm/qemu/hw/virtio/vhost.c:425:45: runtime error: null pointer passed as argument 2, which is declared to never be null > /usr/include/string.h:64:33: note: nonnull attribute specified here How are you running that test? Can you add this printf and tell me what it's seeing? /* Same size, lets check the contents */ fprintf(stderr, "%s: %p %p %d\n", __func__, dev->mem_sections, old_sections, n_old_sections); changed = memcmp(dev->mem_sections, old_sections, n_old_sections * sizeof(old_sections[0])) != 0; I'm seeing a bunch of calls where both pointers are NULL, but n_old_sections is 0, which feels legal to me. I guess we could make it: changed = n_old_sections ? memcmp(....) : false; just to shut clang up. Dave > > > > -- > > 2.14.3 -- Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK