Re: [PATCH RFC v7 2/8] Introduce group administration commands

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

On Thu, Aug 18, 2022 at 04:46:45PM +0800, Jason Wang wrote:
> On Sat, Aug 13, 2022 at 1:19 AM Michael S. Tsirkin <mst@redhat.com> wrote:
> >
> > From: Max Gurtovoy <mgurtovoy@nvidia.com>
> >
> > These commands are used for essential administrative and management
> > operations.
> >
> > Following patches will introduce an interface for submitting these
> > commands to the device.
> >
> > Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > ---
> >  admin.tex | 99 ++++++++++++++++++++++++++++++++++++++++++++++++++++---
> >  1 file changed, 94 insertions(+), 5 deletions(-)
> >
> > diff --git a/admin.tex b/admin.tex
> > index 6e9434a..4840dd4 100644
> > --- a/admin.tex
> > +++ b/admin.tex
> > @@ -9,8 +9,10 @@ \section{Device groups}\label{sec:Basic Facilities of a Virtio Device / Device g
> >  \item[Parent device]
> >          or parent, the device controlling the group.
> >  \item[Member device]
> > -        a device within a group. Parent device itself may, or may
> > -       not be a member of the group.
> > +        a device within a group. Parent device itself is not
> > +       a member of the group. In the future it is envisoned that
> > +       new group types may be introduced where the parent
> > +       device is a member of the group.
> 
> This part should go with patch 1?
> 
> >  \item[Member identifier]
> >          each member has this indentifier, unique within the group
> >         and used to address it through the parent device.
> > @@ -20,9 +22,10 @@ \section{Device groups}\label{sec:Basic Facilities of a Virtio Device / Device g
> >         and what kind of control does the parent have.
> >  \item[Group identifier]
> >         each group has this identifier, unique within the parent device.
> > -       this specifies the group type and possibly selects the
> > -       group if multiple groups of the same type can be controlled by the same
> > -       parent device.
> > +       This specifies the group type. In the future it is
> > +       envisoned that new group types will be introduced where the group
> > +       identifier also selects the group if multiple groups of the same
> > +       type can be controlled by the same parent device.
> >  \end{description}
> >
> >  A single group type is currently specified:
> > @@ -46,4 +49,90 @@ \section{Device groups}\label{sec:Basic Facilities of a Virtio Device / Device g
> >  PCI transport (see \ref{sec:Virtio Transport Options / Virtio Over PCI Bus}).
> >  \end{description}
> >
> > +\subsection{Group administration commands}\label{sec:Basic Facilities of a Virtio Device / Group administration commands}
> > +
> > +Group administration commands can be issued through a parent
> > +device to control member devices of a group.  This mechanism can
> > +be used, for example, to configure a member device before it is
> > +initialized by its driver.
> > +
> > +All the group administration commands are of the following form:
> > +
> > +\begin{lstlisting}
> > +struct virtio_admin_cmd {
> > +        /* Device-readable part */
> > +        le16 opcode;
> > +        /*
> > +         * 1 - SR-IOV
> > +         * 2 - 65535 reserved
> > +         */
> > +        le16 group_id;
> 
> I wonder about the implication of this field. Does it mean:
> 
> 1) a single admin virtqueue that manage two group of VFs
> 
> or
> 
> 2) a single admin virtqueue that manages both VFs and SFs (then
> 'group_type' should be better?)
> 
> If it's 1, the group id seems not sufficient. If it's 2, I wonder if
> it scales (using a single admin virtqueue to manage both VF and SF).
> 
> Thanks

2 at the moment.  Not sure I understand why wouldn't it scale. I don't
see a fundamental difference between 1 or 2 VQs. We are not going to be
adding 1000s of them,

> > +        /* unused, reserved for future extensions */
> > +        u8 reserved1[12];
> > +        le64 group_member_id;
> > +        u8 command_specific_data[];
> > +
> > +        /* Device-writable part */
> > +        u8 status;
> > +        u8 command_specific_error;
> > +        /* unused, reserved for future extensions */
> > +        u8 reserved2[6];
> > +        u8 command_specific_result[];
> > +};
> > +\end{lstlisting}
> > +
> > +For all commands, \field{opcode}, \field{group_id} and if
> > +necessary \field{group_member_id} and \field{command_specific_data} are
> > +set by the driver, and the parent device sets \field{status} and if
> > +needed \field{command_specific_error} and
> > +\field{command_specific_result}.
> > +
> > +As a rule, any unused device-readable fields are set to zero by the driver
> > +and ignored by the device.  Any unused device-writeable fields are set to zero
> > +by the device and ignored by the driver.
> > +
> > +\field{opcode} specifies the command. The valid
> > +values for \field{opcode} can be found in the following table:
> > +
> > +\begin{tabular}{|l|l|}
> > +\hline
> > +opcode & Command Description \\
> > +\hline \hline
> > +0000h - 7FFFh   & Group administration commands    \\
> > +\hline
> > +8000h - FFFFh   & Reserved    \\
> > +\hline
> > +\end{tabular}
> > +
> > +The \field{group_id} specifies the device group.
> > +The \field{group_member_id} specifies the member device within the group.
> > +See section \ref{sec:Introduction / Terminology / Device group}
> > +for the definition of the group identifier and group member
> > +identifier.
> > +
> > +The following table describes possible \field{status} values:
> > +
> > +\begin{tabular}{|l|l|l|}
> > +\hline
> > +Status & Name & Description \\
> > +\hline \hline
> > +00h   & VIRTIO_ADMIN_STATUS_OK    & successful completion  \\
> > +\hline
> > +01h   & VIRTIO_ADMIN_STATUS_CS_ERR    & command specific error  \\
> > +\hline
> > +02h   & VIRTIO_ADMIN_STATUS_OPCODE_UNSUPPORTED    & unsupported or invalid \field{opcode}  \\
> > +\hline
> > +03h   & VIRTIO_ADMIN_STATUS_INVALID_FIELD    & invalid field within command specific data  \\
> > +\hline
> > +04h   & VIRTIO_ADMIN_STATUS_INVALID_GROUP    & invalid group identifier was set  \\
> > +\hline
> > +05h   & VIRTIO_ADMIN_STATUS_INVALID_MEM    & invalid group member identifier was set  \\
> > +\hline
> > +\end{tabular}
> > +
> > +The \field{command_specific_error} should be inspected by the driver only if \field{status} is set to
> > +VIRTIO_ADMIN_STATUS_CS_ERR by the device. In this case, the
> > +contents of \field{command_specific_error}
> > +holds the command specific error. If \field{status} is not set to VIRTIO_ADMIN_STATUS_CS_ERR, the
> > +\field{command_specific_error} value is undefined and should be ignored by the driver.
> >
> > --
> > MST
> >