Re: [virtio-comment] Re: [PATCH v3 4/4] Add CCW configuration field "indirect_num"

[Christian Schoenebeck <qemu_oss@crudebyte.com>]

On Freitag, 18. März 2022 17:06:25 CET Halil Pasic wrote:
> On Fri, 18 Mar 2022 12:02:31 +0100
> 
> Christian Schoenebeck <qemu_oss@crudebyte.com> wrote:
> > On Donnerstag, 17. März 2022 15:12:42 CET Cornelia Huck wrote:
> > > On Wed, Mar 16 2022, Christian Schoenebeck <qemu_oss@crudebyte.com> 
wrote:
> > > > This new CCW configuration field allows to negotiate a more fine
> > > > graded maximum lenght of indirect descriptor chains.
> > > > 
> > > > Bump CCW virtio revision to 3 and make the existence of this new
> > > > field "indirect_num" dependant on revision 3.
> > > > 
> > > > Fixes: https://github.com/oasis-tcs/virtio-spec/issues/122
> > > > Signed-off-by: Christian Schoenebeck <qemu_oss@crudebyte.com>
> > > > ---
> > > > 
> > > >  content.tex | 14 +++++++++++++-
> > > >  1 file changed, 13 insertions(+), 1 deletion(-)
> > > 
> > > (...)
> > > 
> > > > @@ -2581,12 +2583,17 @@ \subsubsection{Configuring a
> > > > Virtqueue}\label{sec:Virtio Transport Options / Vir>
> > > > 
> > > >  struct vq_config_block {
> > > >  
> > > >          be16 index;
> > > >          be16 max_num;
> > > > 
> > > > +        be32 indirect_num; /* since virtio-ccw rev. 3 */
> > > 
> > > Maybe make it max_indirect_num (to mirror max_num?)
> > 
> > Right, that name appears indeed more appropriate here.
> > 
> > > >  };
> > > >  \end{lstlisting}
> > > >  
> > > >  The requested number of buffers for queue \field{index} is returned
> > > >  in
> > > >  \field{max_num}.
> > > > 
> > > > +Since revision 3, \field{indirect_num} exists, which is supposed to
> > > > reflect the +Queue Indirect Size (i.e. the maximum length of indirect
> > > > descriptor tables) +supported by device for this queue.
> > > 
> > > It still depends on the feature flag, though.
> > > 
> > > Maybe
> > > 
> > > "If revision 2 or lower is set, struct vq_config_block extends up to
> > > \field{max_num}.
> > > 
> > > If revision 3 or higher is set, struct vq_config_block extends to
> > > \field{max_indirect_num}. If VIRTIO_RING_F_INDIRECT_SIZE is negotiated,
> > > \field{indirect_max_num} contains the maximum Queue Indirect Size
> > > (i.e. the maximum length of indirect descriptor tables) supported by the
> > > device for this queue; otherwise, its contents are unpredictable."
> > 
> > Maybe rather something like this instead:
> > 
> > "Actual size of struct vq_config_block depends on the virtio-ccw revision.
> > 
> > If revision 2 or lower is set, struct vq_config_block extends up to
> > *including* \field{max_num}.
> > 
> > If revision 3 or higher is set, struct vq_config_block extends up to
> > *including* \field{max_indirect_num}. If VIRTIO_RING_F_INDIRECT_SIZE is
> > negotiated, \field{indirect_max_num} contains the maximum Queue Indirect
> > Size (i.e. the maximum length of indirect descriptor tables) supported by
> > the device for this queue; otherwise, its contents are *undefined*."
> 
> I agree that the "including" is important, but I'm not sure about the
> "its contents are undefined". I don't really understand why should we use
> a plural here. What speaks against specifying that in SHOULD be stored
> as 0 by the device, and MUST be ignored by the driver?

Both solutions would be viable. Personally I would just use something like 
"Should be zero" if there is a value in recommending that, but I don't see a 
value in recommending to set something to zero and at the same time requiring 
to not access it in the first place.

> Currently we say that \field{max_indirect_num} exists like a be32 field
> even if VIRTIO_RING_F_INDIRECT_SIZE is not negotiated. Which kind of
> implies that at least type invariants should hold. Of course, there is
> none here (i.e. every bits value is also a be32 value), but for something
> like an enum interesting corner cases can pop up.

I can't follow you on that one. What has that do with enums in this case?

Anyway, I won't persist on my suggestion to use the (IMO more compact form) 
"undefined". If you guys prefer the more specific solution "SHOULD be 0 and 
MUST not be accessed" then I will go that way.

Best regards,
Christian Schoenebeck