This patch introduces 'pre-start' signal and registers
nodedev checking handler to check duplicate USB devices.
If virt-manager can not identify unique usb device any more
before domain startup, it will throw a tip error to tell it
is time to reattach host USB devices to get updated bus/addr info.
When there are multiple usb devices with same vendor/product
in the host device list, the bus/addr is going to be used when
attaching one of usb devices.
Currently is_dup flag is only useful to VirtualHostDeviceUSB.
I put get_nodedevs_number() in connection.py, so the startup
hooks can use it.
When there are multile USB devices, --host-device with bus.addr
can identify unique device, so we need to add bus/addr info to
XML rather than only vendor/product
If there is only one USB device with --host-device bus.addr,
adding bus/addr is no harm under the request of user.
host-model is a bit more scary then previous behavior, since the
model of an existing VM can transparently change across libvirt
upgrade or HW upgrade. This will probably be improved at the libvirt
level, but for now revert.
This reverts commit 17f43e47fb.
Although the code in lookup_nodedev() is general to deal with
hostdev, it has to seperate USB device from PCI device there.
For PCI device, the domain/bus/slot/function is hex.
For USB device, the bus/device is decimal.
This fix makes the label use hostdev pretty_name:
+------------------------------------------------------+
| Physical USB Device |
| Device: 006:032 RSA RSA SecureID (R) Authenticator |
(crobinso: fix some pylint)
There was an example for USB controller, but to show all the
sub-options, there was the master= option used without extra
information about the fact that it is used for companion controllers
only. So this patch adds second example in front of the previous one
and re-words the description a little bit.
When there are mutiple USB devices with identical vendorId and
productId, with --host-device <vendorId:productId>, virt-install
fails to specify one.
Currently the virt-install add the first node device with
these vendorId and productId.
This fix will throw out an error when argument to --host-device
can't specify unique node device.
For example:
virt-install ${other args} \
--host-device <vendorId:productId>
ERROR 15e1:2007 corresponds to multiple node devices
Remote sound works fine these days. If people want to turn the
preference off, they probably want to do it in all cases, and can
'customize before install' to alter the rest.
In case there was '.treeinfo' file available, virt-install was
forcibly getting image information out of the file. For some
distributions (especially older releases, e.g. RHEL 4), there might be
'.treeinfo' file available, but with incomplete information. Allow
fallback to default values even when '.treeinfo' file is found.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=954262
Kind of a maintenance pain, and it's just poorly duplicating info
from the git logs. We could autogenerate this but I don't really
have the interest in figuring it out. Patches welcome :)
-c maps to --cdrom for virt-install, but for other libvirt tools like
virsh it maps to --connect.
Handle -c and --cdrom a little differently: if -c contains '://',
error out with an explanation. User can work around it by using
--cdrom. This could in theory break some users, but the change they
make will be backwards compatible.
If the host-model is selected, disable the cpu model drop down
and features list. They still show what exact configuration the
host-model is using.
For the old libvirt which doesn't support <cpu mode='host-model'/>
virt-manager still copy cpu configs from caps XML to domain XML.
(crobinso: Fix some minor pylint)
The helper function aims to reset cpu related vars to none except
<topology>.
It is named with clear_attrs(). clear() could be saved for later use
of clearing all of them including <topology>.
UPDATE_CPU flag exists after v0.8.0, so libvirt that
supports HOST_MODEL should support UPDATE_CPU,
the earliest version we are going to check is v0.9.10.
After all these changes, it fixes the invisible scrollbar issue. It
was some combination of do_style_set and something else, but I didn't
bother narrowing it down.
