Re: Using bios-settings-mgr for setting hypervisor network attributes

[Deepak Kodihalli <dkodihal@linux.vnet.ibm.com>]

On 24/09/20 2:56 am, Patrick Williams wrote:
> On Wed, Sep 23, 2020 at 01:51:33PM -0700, Ed Tanous wrote:
>> On Wed, Sep 23, 2020 at 12:24 PM Patrick Williams <patrick@stwcx.xyz> wrote:
>>>
>>> On Tue, Sep 22, 2020 at 02:39:04PM +0530, Ratan Gupta wrote:
>>>
>>> It is unfortunate that org.freedesktop.DBus.Properties doesn't have a
>>> way to set multiple properties as the analogous operation to 'GetAll'.
>>
>> It was proposed we (OpenBMC) add one while back.  I think it muddies
>> the water of what it means to be a method call, and what it means to
>> be a property, especially for the use case that it was being proposed
>> to cover.
> 
> I'm not sure why it would be considered mudding the water.  All property
> Get/Set/GetAll operations really are just a method call under the covers
> anyhow to org.freedesktop.DBus.Properties.  I do think that ideally we'd
> get the method added directly to that interface because then the DBus
> bindings will support it natively.

I had proposed 
https://gerrit.openbmc-project.xyz/#/c/openbmc/phosphor-dbus-interfaces/+/12861/ 
a while back but there were concerns expressed in the review.

> I forgot the mention this again, but another way to solve it is similar
> to xyz.openbmc_project.Inventory.Manager where you take a fully (or
> partially) formed object as a method parameter and the process which
> hosts Inventory.Manager hosts the object.  Settings could be done the
> same way.  The issue is, again, having other processes know when to use
> this new method and when to just update properties.
> 
>>> When all of our DBus objects were serial we likely never had this issue
>>> because the request to read the properties (to send to the hypervisor)
>>> would come behind the signal and subsequent property updates.  Now that
>>> we're moving towards more ASIO we likely will see this kind of issue
>>> more often.  I don't like it but we could certainly proposal a
>>> 'SetMultiple' extension to org.freedesktop or create our own interface.
>>
>> If you have properties that need to be set in lockstep with one
>> another to be valid, I suspect that indicates that properties are not
>> the right tool.  Redfish hits this a lot, where each resource is
>> expected that any property is modifiable independently, and certain
>> implementations need an atomic "unit" of update.  bmcweb doesn't want
>> to have to cache properties that are collectively invalid right now,
>> but might become valid in the future, so there's an impasse.  Who
>> keeps the state while it's invalid?  Thus Far, that falls to the
>> dbus-daemons to store.
> 
> Agreed.  This has also been a general statement  we've made in reviews
> for new interfaces.  "If you need to update multiple properties, use
> a method; if you are just updating a single property, update the property."

Can we add a method to 
https://github.com/openbmc/phosphor-dbus-interfaces/blob/master/xyz/openbmc_project/Network/EthernetInterface.interface.yaml 
to update multiple properties? Also, how does one ensure these updates 
via the method result in a single PropertiesChanged event on D-Bus. Is 
that implicit D-Bus behavior, or does the binding need to enable this?

>>> We could define an interface to implement something like Proposal #1,
>>> but we would need a new interface and not a property we tack onto
>>> existing interfaces.  We'd probably need to revisit a lot of our
>>> interface definitions and see which ones typicallly have multi-property
>>> updates and does an intermediate state leave us in a bad situation.
>>>
>>> Specifically for BIOS/Hypervisor settings, I mentioned before that it
>>> isn't clear to me what the proposal is for applying Pending to Current.
>>> Again, this isn't general, but we could define an interface specific for
>>> BIOS/Hypervisor settings which has a way to indicate 'Pending
>>> transaction is complete' (set by entities like Redfish) and 'Pending
>>> values applied to Current' (set by entities like PLDM).  For the current
>>> settings-style values though, this requires external interfaces to
>>> somehow know that the setting is associated with the Host in order to do
>>> the application, since BMC-owned properties won't have or need this.
>>
>> Dumb question: Does anyone actually need to know the "current" value?
>> Redfish certainly would need to return  the "pending" value in all
>> cases, as it's required so the restful API emulates ACID-like
>> compliance to the user.  Could we just have an optional interface that
>> indicates "values might not be loaded yet" and simplify the dbus API a
>> little?
> 
> I think this is generally for humans in the case of BIOS settings.
>     - "What is the setting my system is currently running with?"
>     - "What will happen next time I reboot?"
> 
> I don't know how this is modeled in Redfish.

I believe there is a Bios.Attributes for currently applied BIOS settings 
and Bios.Settings for what will be applied post a system reset.