From: "Michael S. Tsirkin" <mst@redhat.com>
To: Jason Wang <jasowang@redhat.com>
Cc: virtio-comment@lists.oasis-open.org,
virtio-dev@lists.oasis-open.org, cohuck@redhat.com,
sgarzare@redhat.com, stefanha@redhat.com, nrupal.jani@intel.com,
Piotr.Uminski@intel.com, hang.yuan@intel.com,
virtio@lists.oasis-open.org,
Zhu Lingshan <lingshan.zhu@intel.com>,
pasic@linux.ibm.com, Shahaf Shuler <shahafs@nvidia.com>,
Parav Pandit <parav@nvidia.com>,
Max Gurtovoy <mgurtovoy@nvidia.com>
Subject: Re: [PATCH v9 08/10] admin: command list discovery
Date: Thu, 24 Nov 2022 03:30:41 -0500 [thread overview]
Message-ID: <20221124032505-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <CACGkMEsientWq6sKcuB-0hL0uaPJB4dcp4=8c2BXpNRZrd5RWA@mail.gmail.com>
On Thu, Nov 24, 2022 at 02:40:15PM +0800, Jason Wang wrote:
> On Thu, Nov 24, 2022 at 5:08 AM Michael S. Tsirkin <mst@redhat.com> wrote:
> >
> > Add commands to find out which commands does each group support,
> > as well as enable their use by driver.
> > This will be especially useful once we have multiple group types.
> >
> > An alternative is per-type VQs. This is possible but will
> > require more per-transport work. Discovery through the vq
> > helps keep things contained.
>
> This seems to complicate the migration (compatibility) of the adminq
> itself. We need check both features and command list.
Yes. There's no way around this though - commands need to
be per group and a pci device can host at least both
SIOV and SRIOV children.
> More below.
>
> >
> > Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > ---
> > admin.tex | 82 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-
> > 1 file changed, 81 insertions(+), 1 deletion(-)
> >
> > diff --git a/admin.tex b/admin.tex
> > index eeccd2e..eec12a9 100644
> > --- a/admin.tex
> > +++ b/admin.tex
> > @@ -91,7 +91,9 @@ \subsection{Group administration commands}\label{sec:Basic Facilities of a Virti
> > \hline
> > opcode & Name & Command Description \\
> > \hline \hline
> > -0x0000 - 0x7FFF & - & Group administration commands \\
> > +0x0000 & VIRTIO_ADMIN_CMD_LIST_QUERY & Provides to driver list of commands supported for this group type \\
> > +0x0001 & VIRTIO_ADMIN_CMD_LIST_USE & Provides to device list of commands used for this group type \\
> > +0x0002 - 0x7FFF & - & Group administration commands \\
> > \hline
> > 0x8000 - 0xFFFF & - & Reserved \\
> > \hline
> > @@ -142,6 +144,84 @@ \subsection{Group administration commands}\label{sec:Basic Facilities of a Virti
> > \hline
> > \end{tabular}
> >
> > +Before sending any administration commands to the device, the driver
> > +needs to communicate to the device which commands it is going to
> > +use. Initially (after reset), only two commands are assumed to be used:
> > +VIRTIO_ADMIN_CMD_LIST_QUERY and VIRTIO_ADMIN_CMD_LIST_USE.
> > +
> > +Before sending any other commands for any member of a specific group to
> > +the device, the driver queries the supported commands via
> > +VIRTIO_ADMIN_CMD_LIST_QUERY and sends the commands it will use via
> > +VIRTIO_ADMIN_CMD_LIST_USE.
> > +
> > +Commands VIRTIO_ADMIN_CMD_LIST_QUERY and
> > +VIRTIO_ADMIN_CMD_LIST_USE
> > +both use the following structure describing the
> > +command opcodes:
> > +
> > +\begin{lstlisting}
> > +struct virtio_admin_cmd_list {
> > + /* Indicates which of the below fields were returned
> > + le32 device_admin_cmds[];
> > +};
> > +\end{lstlisting}
> > +
> > +This structure is an array of 32 bit values in little-endian byte
> > +order, in which a bit is set if the specific command opcode
> > +is supported.
>
> Is this simpler if we map group type instead of a single opcode? E.g
> virtqueue transport will introduce 10+ commands, using a bit for a
> specific opcode will complicate things.
I think more flexibility is better, and bits are cheap.
> > +Thus, \field{device_admin_cmds[0]} refers to the first 32-bit value
> > +in this array corresponding to opcodes 0 to 31,
> > +\field{device_admin_cmds[1]} is the second 32-bit value
> > +corresponding to opcodes 32 to 63, etc.
> > +For example, the array of size 2 including
> > +the values 0x3 in \field{device_admin_cmds[0]}
> > +and 0x1 in \field{device_admin_cmds[1]} indicates that only opcodes 0, 1 and 32
> > +are supported.
> > +
> > +Accordingly, bits 0 and 1 corresponding to opcode 0
> > +(VIRTIO_ADMIN_CMD_LIST_QUERY) and 1
> > +(VIRTIO_ADMIN_CMD_LIST_USE) are
> > +always set in \field{device_admin_cmds[0]} returned by VIRTIO_ADMIN_CMD_LIST_QUERY.
> > +
> > +For the command VIRTIO_ADMIN_CMD_LIST_QUERY, \field{opcode} is set to 0x0.
> > +The \field{group_member_id} is unused. It is set to zero by driver.
> > +This command has no command specific data.
> > +The device, upon success, returns a result in
> > +\field{command_specific_result} in the format
> > +\field{struct virtio_admin_cmd_list} describing the
> > +list of administration commands supported for the given group.
> > +
> > +
> > +For the command VIRTIO_ADMIN_CMD_LIST_USE, \field{opcode}
> > +is set to 0x1.
> > +The \field{group_member_id} is unused. It is set to zero by driver.
> > +The \field{command_specific_data} is in the format
> > +\field{struct virtio_admin_cmd_list} describing the
> > +list of administration commands used by the driver.
> > +This command has no command specific result.
> > +
> > +The driver issues the command VIRTIO_ADMIN_CMD_LIST_QUERY to
> > +query the list of commands valid for this group and before sending
> > +any commands for any member of a group.
> > +
> > +The driver then enables use of some of the opcodes by sending to
> > +the device the command VIRTIO_ADMIN_CMD_LIST_USE with a subset
> > +of the list returned by VIRTIO_ADMIN_CMD_LIST_QUERY that is
> > +both understood and used by the driver.
>
> Any reason we need to invent a negotiation process here? (This seems
> only useful when some commands/opcodes are mutually exclusive)
>
> Thanks
yes, features don't help because they are per owner.
this is per group.
imagine vq transport. SET_FEATURES commands sets the features
for the child. we need something to negotiate with owner
but still within the vq.
> > +
> > +If the device supports the command list used by the driver, the
> > +device completes the command with status VIRTIO_ADMIN_STATUS_OK.
> > +If the device does not support the command list, the device
> > +completes the command with status
> > +VIRTIO_ADMIN_STATUS_INVALID_FIELD.
> > +
> > +Note: drivers are assumed not to set bits in device_admin_cmds
> > +if they are not familiar with how the command opcode
> > +is used, since devices could have dependencies between
> > +command opcodes.
> > +
> > +It is assumed that all members in a group support and are used
> > +with the same list of commands.
> > +
> > \section{Administration Virtqueues}\label{sec:Basic Facilities of a Virtio Device / Administration Virtqueues}
> >
> > An administration virtqueue of an owner device is used to submit
> > --
> > MST
> >
next prev parent reply other threads:[~2022-11-24 8:30 UTC|newest]
Thread overview: 62+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-23 21:07 [PATCH v9 00/10] Introduce device group and device management Michael S. Tsirkin
2022-11-23 21:08 ` [PATCH v9 01/10] virtio: document forward compatibility guarantees Michael S. Tsirkin
2022-11-24 4:33 ` Jason Wang
2022-11-24 6:59 ` Michael S. Tsirkin
2022-11-24 7:34 ` Jason Wang
2022-11-24 8:15 ` Michael S. Tsirkin
2022-11-24 12:05 ` [virtio-dev] " Cornelia Huck
2022-11-25 3:17 ` Jason Wang
2022-11-25 10:37 ` [virtio-dev] " Cornelia Huck
2022-11-28 6:14 ` Jason Wang
2022-11-23 21:08 ` [PATCH v9 02/10] admin: introduce device group and related concepts Michael S. Tsirkin
2022-11-24 5:41 ` Jason Wang
2022-11-24 7:08 ` Michael S. Tsirkin
2022-11-24 7:37 ` [virtio-dev] " Jason Wang
2022-11-24 8:18 ` Michael S. Tsirkin
2022-11-24 12:12 ` Cornelia Huck
2022-11-25 3:23 ` Jason Wang
2022-11-25 10:58 ` [virtio] " Cornelia Huck
2022-11-25 12:08 ` Michael S. Tsirkin
2022-11-23 21:08 ` [PATCH v9 03/10] admin: introduce group administration commands Michael S. Tsirkin
2022-11-24 5:52 ` [virtio-dev] " Jason Wang
2022-11-24 7:12 ` Michael S. Tsirkin
2022-11-24 7:42 ` Jason Wang
2022-11-24 8:03 ` Michael S. Tsirkin
2022-11-25 3:24 ` [virtio-comment] " Jason Wang
2022-11-24 12:21 ` [virtio-dev] " Cornelia Huck
2022-11-25 3:54 ` Jason Wang
2022-11-28 13:14 ` [virtio-comment] " Zhu Lingshan
2022-11-23 21:08 ` [PATCH v9 04/10] admin: introduce virtio admin virtqueues Michael S. Tsirkin
2022-11-28 13:12 ` [virtio-comment] " Zhu Lingshan
2022-11-23 21:08 ` [PATCH v9 05/10] pci: add admin vq registers to virtio over pci Michael S. Tsirkin
2022-11-24 6:00 ` Jason Wang
2022-11-24 7:14 ` Michael S. Tsirkin
2022-11-24 7:46 ` Jason Wang
2022-11-24 8:09 ` Michael S. Tsirkin
2022-11-25 3:27 ` Jason Wang
2022-11-23 21:08 ` [PATCH v9 06/10] mmio: document ADMIN_VQ as reserved Michael S. Tsirkin
2022-11-24 6:03 ` Jason Wang
2022-11-24 7:45 ` Michael S. Tsirkin
2022-11-23 21:08 ` [PATCH v9 07/10] ccw: " Michael S. Tsirkin
2022-11-23 21:08 ` [PATCH v9 08/10] admin: command list discovery Michael S. Tsirkin
2022-11-24 6:40 ` Jason Wang
2022-11-24 8:30 ` Michael S. Tsirkin [this message]
2022-11-25 3:38 ` [virtio-comment] " Jason Wang
2022-11-25 11:43 ` Michael S. Tsirkin
2022-11-28 4:34 ` Jason Wang
2022-11-28 7:42 ` Michael S. Tsirkin
2022-11-23 21:08 ` [PATCH v9 09/10] admin: conformance clauses Michael S. Tsirkin
2022-11-24 6:51 ` Jason Wang
2022-11-24 8:36 ` Michael S. Tsirkin
2022-11-25 3:50 ` Jason Wang
2022-11-25 11:42 ` [virtio] " Cornelia Huck
2022-11-25 11:56 ` Michael S. Tsirkin
2022-11-25 12:10 ` [virtio] " Cornelia Huck
2022-11-25 11:47 ` Michael S. Tsirkin
2022-11-28 4:32 ` Jason Wang
2022-11-28 7:39 ` Michael S. Tsirkin
2022-11-24 12:28 ` [virtio] " Cornelia Huck
2022-11-25 3:38 ` Jason Wang
2022-11-23 21:08 ` [PATCH RFC v9 10/10] ccw: document more reserved features Michael S. Tsirkin
2022-11-24 6:53 ` Jason Wang
2022-11-24 8:31 ` Michael S. Tsirkin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20221124032505-mutt-send-email-mst@kernel.org \
--to=mst@redhat.com \
--cc=Piotr.Uminski@intel.com \
--cc=cohuck@redhat.com \
--cc=hang.yuan@intel.com \
--cc=jasowang@redhat.com \
--cc=lingshan.zhu@intel.com \
--cc=mgurtovoy@nvidia.com \
--cc=nrupal.jani@intel.com \
--cc=parav@nvidia.com \
--cc=pasic@linux.ibm.com \
--cc=sgarzare@redhat.com \
--cc=shahafs@nvidia.com \
--cc=stefanha@redhat.com \
--cc=virtio-comment@lists.oasis-open.org \
--cc=virtio-dev@lists.oasis-open.org \
--cc=virtio@lists.oasis-open.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox