From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:51954) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WzUTT-00080k-4B for qemu-devel@nongnu.org; Tue, 24 Jun 2014 13:25:46 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1WzUTF-0004qG-1Q for qemu-devel@nongnu.org; Tue, 24 Jun 2014 13:25:39 -0400 Received: from mx1.redhat.com ([209.132.183.28]:25701) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1WzUTE-0004q1-Qg for qemu-devel@nongnu.org; Tue, 24 Jun 2014 13:25:24 -0400 Date: Tue, 24 Jun 2014 20:25:41 +0300 From: "Michael S. Tsirkin" Message-ID: <20140624172541.GD14798@redhat.com> References: <20140624151955.17522.62537.stgit@bahia.local> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: <20140624151955.17522.62537.stgit@bahia.local> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v9 00/22] legacy virtio support for cross-endian targets List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Greg Kurz Cc: Kevin Wolf , Peter Maydell , Stefan Hajnoczi , Juan Quintela , Rusty Russell , qemu-devel@nongnu.org, Alexander Graf , aneesh.kumar@linux.vnet.ibm.com, Anthony Liguori , Amit Shah , Paolo Bonzini , Andreas =?iso-8859-1?Q?F=E4rber?= On Tue, Jun 24, 2014 at 07:06:58PM +0200, Greg Kurz wrote: > The current legacy virtio devices have a fundamental flaw: they all sha= re > data between host and guest with guest endianness ordering. This is ok = for > nearly all architectures that have fixed endianness. Unfortunately, it = breaks > for recent PPC64 and ARM targets that can change endianness at runtime. > The virtio-1.0 specification fixes the issue by enforcing little-endian > ordering. It may take some time though until the code for 1.0 gets avai= lable > and supported, and all the users can migrate. There have been discussio= ns > for some monthes about supporting such oddity: now we have little-endia= n > PPC64 distros available, it is worth to propose something. >=20 > This patch set brings legacy virtio support for cross-endian targets. T= he > rationale is that we add a new device_endianness property to VirtIODevi= ce. > This property is used as a runtime indicator to decide wether we should > do little-endian or big-endian conversion, as opposed to the compile ti= me > choice we have now with TARGTE_WORDS_BIGENDIAN. typo :) no need to repost for this. > The choice was made to > sample the device endianness out of the endianness mode of the guest > CPU that does the reset. It is an evil but logical consequence of the > initial flaw in the virtio specification, and it was agreed that the co= ncept > would be a good common base for ARM and PPC64 enablement at least. Plea= se > note also that this new property is state and must be preserved across > migrations. >=20 > There are several parts in the serie: > - patches 1 and 2 are simple fixes > - patches 3 to 9 introduce VMState based subsections in the virtio > migration code. This is needed because we introduce a new property > in VirtIODevice that we want to migrate without ruining compatibility > efforts > - patches 10 to 13 bring virtio device endianness and memory accessors > to be used by the virtio code > - patches 14 to 20 wire the new memory accessors everywhere accross the > virtio code > - patch 21 is the PPC64 enablement > - patch 22 is a follow-up workaround to disable vhost-net acceleration > in the case the host and guest have different endianness, because > it is not supported for the moment >=20 > Changes since v8 are provided in each patch. >=20 > Cheers. >=20 > --- >=20 > Alexander Graf (1): > virtio-serial: don't migrate the config space >=20 > C=E9dric Le Goater (1): > virtio-net: byteswap virtio-net header >=20 > Greg Kurz (14): > virtio: introduce device specific migration calls > virtio-net: implement per-device migration calls > virtio-blk: implement per-device migration calls > virtio-serial: implement per-device migration calls > virtio-balloon: implement per-device migration calls > virtio-rng: implement per-device migration calls > virtio: add subsections to the migration stream > exec: introduce target_words_bigendian() helper > cpu: introduce CPUClass::virtio_is_big_endian() > virtio: add endian-ambivalent support to VirtIODevice > virtio: memory accessors for endian-ambivalent targets > virtio-9p: use virtio wrappers to access headers > target-ppc: enable virtio endian ambivalent support > vhost-net: disable when cross-endian >=20 > Rusty Russell (6): > virtio: allow byte swapping for vring > virtio-net: use virtio wrappers to access headers > virtio-balloon: use virtio wrappers to access page frame numbers > virtio-blk: use virtio wrappers to access headers > virtio-scsi: use virtio wrappers to access headers > virtio-serial-bus: use virtio wrappers to access headers >=20 >=20 > exec.c | 8 - > hw/9pfs/virtio-9p-device.c | 3 - > hw/block/virtio-blk.c | 62 ++++++----- > hw/char/virtio-serial-bus.c | 94 ++++++++++------ > hw/net/vhost_net.c | 19 +++ > hw/net/virtio-net.c | 56 +++++++--- > hw/scsi/virtio-scsi.c | 40 ++++--- > hw/virtio/virtio-balloon.c | 33 +++--- > hw/virtio/virtio-pci.c | 11 +- > hw/virtio/virtio-rng.c | 12 +- > hw/virtio/virtio.c | 216 ++++++++++++++++++++++++++++-= -------- > include/hw/virtio/virtio-access.h | 170 +++++++++++++++++++++++++++++ > include/hw/virtio/virtio.h | 17 +++ > include/qom/cpu.h | 1=20 > qom/cpu.c | 6 + > target-ppc/cpu.h | 2=20 > target-ppc/translate_init.c | 15 +++ > 17 files changed, 583 insertions(+), 182 deletions(-) > create mode 100644 include/hw/virtio/virtio-access.h >=20 > -- > Greg