This patch addresses the missing device alias assignment for filesystem
and network interface devices in LXC domains.
Before this patch, virsh dumpxml showed filesystem and network
devices without aliases:
<filesystem type='mount' accessmode='passthrough'>
<source dir='/shared6'/>
<target dir='/shared6'/>
</filesystem>
<interface type='bridge'>
<mac address='b8:3f:d2:c9:e8:bd'/>
<source bridge='br0'/>
<target dev='vnet8'/>
<model type='virtio'/>
</interface>
After this patch, filesystem devices have auto-generated aliases:
<filesystem type='mount' accessmode='passthrough'>
<source dir='/shared6'/>
<target dir='/shared6'/>
<alias name='fs1'/> <!-- Now assigned -->
</filesystem>
<interface type='bridge'>
<mac address='b8:3f:d2:c9:e8:bd'/>
<source bridge='br0'/>
<target dev='vnet8'/>
<model type='virtio'/>
<alias name='net4'/> <!-- Now assigned -->
</interface>
Changes in v3:
- Implemented new index allocation algorithm,
scans existing aliases before new index allocation
- Added hotplug support in lxcDomainAttachDeviceNetLive()
- Handles gaps in numbering after device detach/attach
- Follows QEMU's qemuAssignDeviceNetAlias() pattern
- Tested hotplug scenario
The implementation follows the same pattern used by QEMU's
qemuAssignDeviceNetAlias() and qemuAssignDeviceFSAlias().
This is needed as a prerequisite for fixing bug #63
(loop device path length limitation), where device aliases
will be used to compose synthetic lo_file_name references
in the format "libvirt-$UUID-$DEVALIAS".
