On 01/09/2013 01:02 PM, Daniel P. Berrange wrote:
On Wed, Jan 09, 2013 at 12:36:52PM -0500, Laine Stump wrote:
> On 01/09/2013 10:52 AM, Amos Kong wrote:
>> On Wed, Jan 09, 2013 at 08:14:07AM -0700, Eric Blake wrote:
>>> On 01/09/2013 01:39 AM, Amos Kong wrote:
>>>> Current seabios will try to boot from selected devices first,
>>>> if they are all failed, seabios will also try to boot from
>>>> un-selected devices.
>>>>
>>>> We need to make it configurable. I already posted a seabios
>>>> patch to add a new device type to halt booting. Qemu can add
>>>> "HALT" at the end of bootindex string, then seabios will halt
>>>> booting after trying to boot from selected devices.
>>>>
>>>> This option only effects when boot priority is changed by
>>>> bootindex options, the old style(-boot order=..) will still
>>>> try to boot from un-selected devices.
>>>>
>>>> v2: add HALT entry in get_boot_devices_list()
>>>> define boot_strict to bool
>>>>
>>>> Signed-off-by: Amos Kong <akong(a)redhat.com>
>>>> ---
>>> Libvirt will need to expose an attribute that lets the user control
>>> whether to use this new option; how do we probe via QMP whether the new
>>> -boot strict=on command-line option is available?
>> Old style to adjust boot priority by order parameter:
>> -boot order=n,strict=on (BAD, unselected devices will be tried)
>>
>> New style to adjust boot priority by bootindex:
>> -device virtio-net-pci,...,bootindex=1 -boot strict=on (OK)
>>
>> We only want strict option to support new style.
>>
>> (those two styles are implemented in two different way insider
>> seabios, the latest simple patch only changed the bootindex way)
> Just a note about this: as far as I can tell virt-manager currently only
> uses the "old style" of specifying boot order; it needs to be enhanced
> to recognize the presence of bootindex=n ordering, and behave
> accordingly. Looking from the outside, that looks to be not completely
> trivial, as the old style of ordering only allows specifying "hard disk"
> as a single line in the priority order, but the new style has each disk
> specified separately. Not only that, but *which* of the disks is tried
> first under the old order may change depending on whether or not a
> bootmenu is requested (in one case it picks unit=0 on the controller, in
> the other case, it orders the disks alphabetically by target dev name)
Hmm, well libvirt should be using bootindex=n on the QEMU command line
regardless of what applications put in the XML. ie if the application
uses the old style XML, libvirt should translate that into bootindex=n
for them.
Good point. WHerever it's changed though, it will be a bit tricky to
preserve current behavior, due to the idiosyncrasies I noted above.