From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:34827) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1brr92-00080P-5S for qemu-devel@nongnu.org; Wed, 05 Oct 2016 14:42:21 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1brr8x-000113-9w for qemu-devel@nongnu.org; Wed, 05 Oct 2016 14:42:20 -0400 Received: from mx1.redhat.com ([209.132.183.28]:34720) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1brr8x-00010j-2p for qemu-devel@nongnu.org; Wed, 05 Oct 2016 14:42:15 -0400 Date: Wed, 5 Oct 2016 12:42:13 -0600 From: Alex Williamson Message-ID: <20161005124213.2075ca67@t450s.home> In-Reply-To: <87r37wcktv.fsf@dusky.pond.sub.org> References: <1475498477-2695-1-git-send-email-eric.auger@redhat.com> <87r37wcktv.fsf@dusky.pond.sub.org> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH v4 00/17] Convert VFIO-PCI to realize List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Markus Armbruster Cc: Eric Auger , eric.auger.pro@gmail.com, qemu-devel@nongnu.org On Tue, 04 Oct 2016 15:20:12 +0200 Markus Armbruster wrote: > Eric Auger writes: > > > This series converts VFIO-PCI to realize. It also aims at improving > > the error reporting in case of QMP hot-plug. > > > > Before the series, a device_add failure would have reported: > > {"error": {"class": "GenericError", "desc": "Device initialization > > failed"}}. > > > > Now the actual error cause is reported. > > > > A side effect of this series is the cleanup of the error handling > > in vfio-platform. > > Only a few really minor issues left, so series > Reviewed-by: Markus Armbruster Eric, thanks for doing this conversion and thanks to Markus for the thorough reviews. One tiny nit I see is patch 06/17 does this: > @@ -1189,6 +1189,7 @@ static int vfio_msi_setup(VFIOPCIDevice *vdev, int pos) > > if (pread(vdev->vbasedev.fd, &ctrl, sizeof(ctrl), > vdev->config_offset + pos + PCI_CAP_FLAGS) != sizeof(ctrl)) { > + error_setg_errno(errp, errno, "failed reading PCI_CAP_FLAGS"); > return -errno; > } And that new error string might be more useful if it were "failed reading MSI PCI_CAP_FLAGS". Unless some sort of MSI hint gets added somewhere that I'm missing. Markus noted a few things that could be fixed on commit and several others that I don't want to decide for you, so I'll look forward to a v5 to wrap this up. Thanks! Alex