[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-devel] [PATCH v3] libxl: Spice vdagent support for upstream qemu
Il 06/05/2013 18:28, Wei Liu ha scritto:
On Mon, May 06, 2013 at 04:17:00PM +0100, Fabio Fantoni wrote:
diff --git a/tools/libxl/libxl_dm.c b/tools/libxl/libxl_dm.c
index d10a58f..bc605e4 100644
--- a/tools/libxl/libxl_dm.c
+++ b/tools/libxl/libxl_dm.c
@@ -465,6 +465,12 @@ static char **
libxl__build_device_model_args_new(libxl__gc *gc,
flexarray_append(dm_args, "-spice");
flexarray_append(dm_args, spiceoptions);
+ if (libxl_defbool_val(b_info->u.hvm.spice.vdagent)) {
+ flexarray_vappend(dm_args, "-device", "virtio-serial",
Dependency on VirtIO? Should this be documented as well? Does this mean
that vdagent can only work with HVM? AFAICT we don't support VirtIO in
PV.
Wei.
Spice vdagent is available only if spice is available, and spice is
available only on hvm domU for now. There is no way for libxl to
check if the qemu build has all the required components before
running it. For example xen build qemu-xen without all the features
supported by libxl (not only the ones insert by me) and debian
experimental qemu build support all libxl features. If there isn't
spice enabled nothing happens, and if spice is enabled but the qemu
build doesn't have it, qemu doesn't start.
We cannot check QEMU features in libxl, but we can control build options
for QEMU.
Do we need to modify tools/Makefile to make QEMU include spice?
Presumably the answer is "no" because you didn't have it in your patch
but it is better to be clear.
Wei.
I did one year ago but all version of my patch was rejected, for example
one of these was:
http://lists.xen.org/archives/html/xen-devel/2012-04/msg00743.html
Attachment:
smime.p7s
Description: Firma crittografica S/MIME
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel
|