From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:58743) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1ficFE-0008Kc-AP for qemu-devel@nongnu.org; Thu, 26 Jul 2018 05:07:39 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1ficFB-0004f8-9O for qemu-devel@nongnu.org; Thu, 26 Jul 2018 05:07:36 -0400 Date: Thu, 26 Jul 2018 11:07:27 +0200 From: Cornelia Huck Message-ID: <20180726110727.4d879427.cohuck@redhat.com> In-Reply-To: References: <1532521224-27235-1-git-send-email-thuth@redhat.com> <20180725154837.5f9c6735.cohuck@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Re: [Qemu-devel] [PATCH for-3.1 00/10] Move virtio-ccw devices to separate files List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Thomas Huth Cc: David Hildenbrand , qemu-s390x@nongnu.org, Christian Borntraeger , qemu-devel@nongnu.org, Juan Quintela On Wed, 25 Jul 2018 19:20:59 +0200 Thomas Huth wrote: > On 25.07.2018 15:48, Cornelia Huck wrote: > > On Wed, 25 Jul 2018 14:20:14 +0200 > > Thomas Huth wrote: > > > >> For more fine-grained control over the build process, it would be good > >> to have the possibility to disable single virtio devices, too. We already > >> have CONFIG_VIRTIO_* switches in the Makefiles, but currently all > >> virtio-ccw devices are compiled in anyway. Move them to separate files > >> so we can disable them in the Makefile more easily. > >> > >> NB: I did not move virtio-blk-ccw and virtio-net-ccw to a separate file > >> yet since they are essential for the s390x-virtio machine and thus it > >> does IMHO not make much sense to disable net and blk. But if somebody > >> needs that possibility, too, I can add two more patches on top... > > > > I understand virtio-net-ccw (autogenerated network devices), but why > > virtio-blk-ccw? > > Because we set mc->block_default_type = IF_VIRTIO ==> if you disable > virtio-blk, the CLI options like -hda and -cdrom would not work anymore. I see. Maybe we need some strategic #ifdefs to deal with this.