Re: [virtio-comment] Re: [PATCH v6 2/5] virtio-net: Add flow filter capabilities read commands

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Nov 27, 2023 at 10:19:49AM +0000, Parav Pandit wrote:
> 
> > From: Michael S. Tsirkin <mst@redhat.com>
> > Sent: Friday, November 24, 2023 5:04 PM
> > 
> > On Fri, Nov 24, 2023 at 06:31:03AM +0000, Parav Pandit wrote:
> > >
> > > > From: Michael S. Tsirkin <mst@redhat.com>
> > > > Sent: Friday, November 24, 2023 11:48 AM
> > > >
> > > > On Fri, Nov 24, 2023 at 05:40:26AM +0000, Parav Pandit wrote:
> > > > > > we
> > > > > > strongly suggest that *drivers* support both old and new
> > > > > > mechanism, and then *devices* will only implement what's required.
> > > > > There are other examples in the same document that makes things
> > > > > worst
> > > > with old and new.
> > > > >
> > > > > Also there is literally no way to enforce that driver supports
> > > > > both and new. It is just sounds like an excuse to force infinite
> > > > > config space.
> > > >
> > > > There is a very simple method though.  We allow devices to expose a
> > > > subset of features when DMA is not used. So drivers that want
> > > > maximum features will always opt for DMA. We can also strongly
> > > > recommend that all drivers support DMA if available.
> > > Yeah, don't see how this is elegant at all with all mixed bits.
> > 
> > It's elegant because simple low end devices can cheaply implement MMIO and
> > not worry about DMA.
> > 
> It is not of much help in this case because any low end cheap device which want to support flow filter commands need to have CVQ anyway.
> And hence reusing the same CVQ is more elegant that already does the DMA.
> 
> So CVQ is fulfilling all the below needs.
> 1. Single interface for the get/set config flow filters
> 2. DMA the data
> 3. Not have any partial issues

I don't know what these are.

> 4. provides consistent structures that provisioning side will be able to use

Problem for provisioning is extra definitions will be needed, in a
device specific way.

> > > Nor do I see any enforcement, single method via cvq still holds strong.
> > 
> > You don't need to enforce things, if people want to put a lot of RAM on device
> > and put it in a register let them.
> > 
> Not enforced. It uses the CVQ for flow group and flow filter life cycles and for the sharing this config as well.
> Also aligns with stats that rest also agreed on.

I am talking about your attempt to generally say "no more config fields
everything must be in CVQ". I think it's wrong definitiely for non
network devices must sometimes for network too and generally we need
a solution for config over DMA. This specific thing - whether it
fits in CVQ is a separate discussion.


> > >
> > > >
> > > >
> > > >
> > > > > The method proposed here is elegant and clearly promote one way to
> > > > > do
> > > > things for driver and device with predictability.
> > > > >
> > > >
> > > > I don't see it as elegant at all. What is elegant is *a single tag*
> > > > that describes each property of the device. And this single tag should be
> > good for everything:
> > > > driver, provisioning, migration. And config space offset serves as such.
> > > The single tag is the set of structures.
> > 
> > I have no idea how this will work. If migration format i started reviewing is
> > anything to go by then there will be a huge elaborate structure nothing single
> > or simple. By comparison there's already a proposal how provisioning can
> > work by supplying config space.
> > it is just a clean model to grasp.
> > 
> The provisioning model is simple is to supply all the configuration.
> To draw parallels to some sw side,
> 
> There is per functionality socket option to set things, instead of one giant structure.
> There is per functionality ethtool option/cmd instead of Set ALL/get ALL enforcement.

I'm not sure how much of a parallel one can draw.
Do not see a lot of similarity.
Devices commonly use register map. Everyone understands
this paradigm.


I am not altogether happy with the way you are making
migration generate duplicate definitions for lots of
things we already have definitions for.
Having a 3rd one for provisioning? Gimme a break.


> > > Provisioning access them via owner device.
> > > Member driver access them via CVQ or 1.2 legacy config space.
> >