Daniel P. Berrange wrote: > On Mon, Jul 28, 2008 at 11:51:49AM -0400, Cole Robinson wrote: >> The attached patch teaches virt-manager how to parse VM >> character devices, showing them in the details section. >> This patch also implements removing these devices. >> >> There are a couple quirks though: >> >> 1) All the possible character device xml properties >> aren't accounted for. This only parses target port, >> device type, and source path. I figure this covers >> the common cases, and is easily expandable. >> >> 2) Explicit entries are only set for serial and parallel >> devices, not console devices. Since by default, console >> is really just a duplicate of the first serial device, >> we try to detect this and label said serial device as >> appropriately (seen in the second screen shot). >> This simplifies things a good deal, and makes deleting >> easier, since in order to remove the console or >> first serial device, both xml blocks need to be removed >> simultaneously (not sure if this is a libvirt bug or not). > > I think that is probably a bug. Can you re-test with the latest > CVS snapshot of libvirt where I've changed the handling a little > bit. Basically you should never need to add or remove a 'console' > device explicitly. Either it is there permanently (eg, Xen PV > always has a console), or it is automatically added/removed when > the first serial device is added or removed. > > Cool, things seem to work with current upstream. As far as the virt-manager code though, do you think it's a reasonable approach (don't have an explicit console entry)? Thanks, Cole _______________________________________________ et-mgmt-tools mailing list et-mgmt-tools@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/et-mgmt-tools
