Re: Who / what uses QMP command add_client?

[Markus Armbruster]

Marc-André Lureau <marcandre.lureau@redhat.com> writes:

> Hi
>
> On Tue, Nov 29, 2022 at 6:55 PM Markus Armbruster <armbru@redhat.com> wrote:
>
>> QMP command add_client's schema:
>>
>>     ##
>>     # @add_client:
>>     #
>>     # Allow client connections for VNC, Spice and socket based
>>     # character devices to be passed in to QEMU via SCM_RIGHTS.
>>     #
>>     # @protocol: protocol name. Valid names are "vnc", "spice",
>> "@dbus-display" or
>>     #            the name of a character device (eg. from -chardev id=XXXX)
>>     #
>>     # @fdname: file descriptor name previously passed via 'getfd' command
>>     #
>>     # @skipauth: whether to skip authentication. Only applies
>>     #            to "vnc" and "spice" protocols
>>     #
>>     # @tls: whether to perform TLS. Only applies to the "spice"
>>     #       protocol
>>     #
>>     # Returns: nothing on success.
>>     #
>>     # Since: 0.14
>>     #
>>     # Example:
>>     #
>>     # -> { "execute": "add_client", "arguments": { "protocol": "vnc",
>>     #                                              "fdname": "myclient" } }
>>     # <- { "return": {} }
>>     #
>>     ##
>>     { 'command': 'add_client',
>>       'data': { 'protocol': 'str', 'fdname': 'str', '*skipauth': 'bool',
>>                 '*tls': 'bool' } }
>>
>> Spot the design flaw!
>>
>> It's overloaded @protocol.  Two issues.
>>
>> One, character device IDs "vnc", "spice", "@dbus-display" don't work
>> here.  If we ever add another protocol, we make another device ID not
>> work.  Perhaps this is why Marc-André chose "@dbus-display", which
>> otherwise looks like a typo :)
>>
>
> That's right, I tried to avoid conflicting with chardev ID namespace. IDs
> can't start with '@'.
>
> btw, I have a few patches pending to extend add_client for windows sockets.
>
> I also have patches to check if fds are actually sockets, since that
> command doesn't make much sense with other fds.
>
>
>> Two, introspection can't tell us what protocols are supported.
>
> Hmm, not really a big deal I suppose. You would have both compile-time and
> run-time config. There are other means to introspect the display protocol,
> like query-vnc or query-spice. I thought I had something covering dbus as
> well, but I can't find it, I'll look at it.

Defeating introspection like this is careless design.  I agree this one
is no big deal in practice.

> Let me know if you plan to touch that command, it will likely conflict with
> my work. I plan to submit it soon after the release, but I might do it
> earlier.

I looked at the command only because I'm trying to move crap out of
monitor/.

For PCI, that's "[PATCH 00/12] pci: Move and clean up monitor command
code".

add_client is a bit awkward.  It's in qapi/misc.json and
monitor/qmp-cmds.c.  If it was only about VNC and Spice, I'd move it to
qapi/ui.json and ui/.  If it was only about socket character devices, to
qapi/char.json and char/.  But it's both.  Unless we pick one of the two
arbitrarily, it's stuck in monitor/.