Re: [Qemu-devel] [PATCH RFC v5 30/32] qapi: New QMP command query-qmp-schema for QMP introspection

[Michael Roth]

Quoting Markus Armbruster (2015-09-07 05:16:41)
> qapi/introspect.json defines the introspection schema.  It's designed
> for QMP introspection, but should do for similar uses, such as QGA.
> 
> The introspection schema does not reflect all the rules and
> restrictions that apply to QAPI schemata.  A valid QAPI schema has an
> introspection value conforming to the introspection schema, but the
> converse is not true.
> 
> Introspection lowers away a number of schema details, and makes
> implicit things explicit:
> 
> * The built-in types are declared with their JSON type.
> 
>   TODO Should we map all the integer types to just int?
> 
> * Implicit type definitions are made explicit, and given
>   auto-generated names:
> 
>   - Array types, named by appending "List" to the name of their
>     element type, like in generated C.
> 
>   - The enumeration types implicitly defined by simple union types,
>     named by appending "Kind" to the name of their simple union type,
>     like in generated C.
> 
>   - Types that don't occur in generated C.  Their names start with ':'
>     so they don't clash with the user's names.
> 
> * All type references are by name.
> 
> * The struct and union types are generalized into an object type.
> 
> * Base types are flattened.
> 
> * Commands take a single argument and return a single result.
> 
>   Dictionary argument or list result is an implicit type definition.
> 
>   The empty object type is used when a command takes no arguments or
>   produces no results.
> 
>   The argument is always of object type, but the introspection schema
>   doesn't reflect that.
> 
>   The 'gen': false directive is omitted as implementation detail.
> 
>   The 'success-response' directive is omitted as well for now, even
>   though it's not an implementation detail, because it's not used by
>   QMP.
> 
> * Events carry a single data value.
> 
>   Implicit type definition and empty object type use, just like for
>   commands.
> 
>   The value is of object type, but the introspection schema doesn't
>   reflect that.
> 
> * Types not used by commands or events are omitted.
> 
>   Indirect use counts as use.
> 
> * Optional members have a default, which can only be null right now
> 
>   Instead of a mandatory "optional" flag, we have an optional default.
>   No default means mandatory, default null means optional without
>   default value.  Non-null is available for optional with default
>   (possible future extension).
> 
> * Clients should *not* look up types by name, because type names are
>   not ABI.  Look up the command or event you're interested in, then
>   follow the references.
> 
>   TODO Should we hide the type names to eliminate the temptation?
> 
> New generator scripts/qapi-introspect.py computes an introspection
> value for its input, and generates a C variable holding it.
> 
> It can generate awfully long lines.  Marked TODO.

Would be good if we could get this is for initial merge to help
with debugging if anything pops up.

> 
> A new test-qmp-input-visitor test case feeds its result for both
> tests/qapi-schema/qapi-schema-test.json and qapi-schema.json to a
> QmpInputVisitor to verify it actually conforms to the schema.
> 
> New QMP command query-qmp-schema takes its return value from that
> variable.  Its reply is some 85KiBytes for me right now.
> 
> If this turns out to be too much, we have a couple of options:
> 
> * We can use shorter names in the JSON.  Not the QMP style.
> 
> * Optionally return the sub-schema for commands and events given as
>   arguments.
> 
>   Right now qmp_query_schema() sends the string literal computed by
>   qmp-introspect.py.  To compute sub-schema at run time, we'd have to
>   duplicate parts of qapi-introspect.py in C.  Unattractive.
> 
> * Let clients cache the output of query-qmp-schema.
> 
>   It changes only on QEMU upgrades, i.e. rarely.  Provide a command
>   query-qmp-schema-hash.  Clients can have a cache indexed by hash,
>   and re-query the schema only when they don't have it cached.  Even
>   simpler: put the hash in the QMP greeting.

Would probably be easier for management to build up their own structure
for querying caps, so I think a computed hash seems best. But I don't
think either is something that couldn't be added later if need be.

> 
> Signed-off-by: Markus Armbruster <address@hidden>
> ---
>  .gitignore                                      |   1 +
>  Makefile                                        |   9 +-
>  Makefile.objs                                   |   4 +-
>  docs/qapi-code-gen.txt                          | 232 +++++++++++++++++++-
>  monitor.c                                       |  15 ++
>  qapi-schema.json                                |   3 +
>  qapi/introspect.json                            | 276 
> ++++++++++++++++++++++++
>  qmp-commands.hx                                 |  17 ++
>  scripts/qapi-introspect.py                      | 177 +++++++++++++++
>  scripts/qapi.py                                 |  13 +-
>  tests/.gitignore                                |   1 +
>  tests/Makefile                                  |  10 +-
>  tests/qapi-schema/alternate-good.out            |   1 +
>  tests/qapi-schema/args-member-array.out         |   1 +
>  tests/qapi-schema/comments.out                  |   1 +
>  tests/qapi-schema/empty.out                     |   1 +
>  tests/qapi-schema/enum-empty.out                |   1 +
>  tests/qapi-schema/event-case.out                |   1 +
>  tests/qapi-schema/flat-union-reverse-define.out |   1 +
>  tests/qapi-schema/ident-with-escape.out         |   1 +
>  tests/qapi-schema/include-relpath.out           |   1 +
>  tests/qapi-schema/include-repetition.out        |   1 +
>  tests/qapi-schema/include-simple.out            |   1 +
>  tests/qapi-schema/indented-expr.out             |   1 +
>  tests/qapi-schema/qapi-schema-test.out          |   1 +
>  tests/qapi-schema/returns-int.out               |   1 +
>  tests/test-qmp-input-strict.c                   |  55 +++++
>  27 files changed, 816 insertions(+), 11 deletions(-)
>  create mode 100644 qapi/introspect.json
>  create mode 100644 scripts/qapi-introspect.py
> 
<snip>

> diff --git a/monitor.c b/monitor.c
> index e083716..ddf6cb5 100644
> --- a/monitor.c
> +++ b/monitor.c
> @@ -74,6 +74,7 @@
>  #include "block/qapi.h"
>  #include "qapi/qmp-event.h"
>  #include "qapi-event.h"
> +#include "qmp-introspect.h"
>  #include "sysemu/block-backend.h"
> 
>  /* for hmp_info_irq/pic */
> @@ -928,6 +929,20 @@ EventInfoList *qmp_query_events(Error **errp)
>      return ev_list;
>  }
> 
> +/*
> + * Minor hack: generated marshalling suppressed for this command
> + * ('gen': false in the schema) so we can parse the JSON string
> + * directly into QObject instead of first parsing it with
> + * visit_type_SchemaInfoList() into a SchemaInfoList, then marshal it
> + * to QObject with generated output marshallers, every time.  Instead,
> + * we do it in test-qmp-input-visitor.c, just to make sure
> + * qapi-introspect.py's output actually conforms to the schema.
> + */
> +static void qmp_query_schema(QDict *qdict, QObject **ret_data, Error **errp)
> +{
> +    *ret_data = qobject_from_json(qmp_schema_json);
> +}

What if we still did 'gen': false, manually ran it through input/output
visitors, but then cached the result? I think the unit test approach
is reasonable, but currently we only test the input visitor, whereas
QmpOutputVisitor would come into play for the wire response if we
ever moved to using a C-based schema visitor (for querying individual
fields for instance)

> +
>  /* set the current CPU defined by the user */
>  int monitor_set_cpu(int cpu_index)
>  {
> diff --git a/qapi-schema.json b/qapi-schema.json
> index d874c11..8c77220 100644
> --- a/qapi-schema.json
> +++ b/qapi-schema.json
> @@ -14,6 +14,9 @@
>  # Tracing commands
>  { 'include': 'qapi/trace.json' }
> 
> +# QAPI introspection
> +{ 'include': 'qapi/introspect.json' }
> +
>  ##
>  # @LostTickPolicy:
>  #
> diff --git a/qapi/introspect.json b/qapi/introspect.json
> new file mode 100644
> index 0000000..5b4713b
> --- /dev/null
> +++ b/qapi/introspect.json
> @@ -0,0 +1,276 @@
> +# -*- Mode: Python -*-
> +#
> +# QAPI/QMP introspection
> +#
> +# Copyright (C) 2015 Red Hat, Inc.
> +#
> +# Authors:
> +#  Markus Armbruster <address@hidden>
> +#
> +# This work is licensed under the terms of the GNU GPL, version 2 or later.
> +# See the COPYING file in the top-level directory.
> +
> +##
> +# @query-qmp-schema
> +#
> +# Command query-qmp-schema exposes the QMP wire ABI as an array of
> +# SchemaInfo.  This lets QMP clients figure out what commands and
> +# events are available in this QEMU, and their parameters and results.
> +#
> +# However, the SchemaInfo can't reflect all the rules and restrictions
> +# that apply to QMP.  It's interface introspection (figuring out
> +# what's there), not interface specification.  The specification is in
> +# the QAPI schema.

Thanks, the clarification here and in the documentation addresses my
concerns.

> +#
> +# Returns: array of @SchemaInfo, where each element describes an
> +# entity in the ABI: command, event, type, ...
> +#
> +# Note: the QAPI schema is also used to help define *internal*
> +# interfaces, by defining QAPI types.  These are not part of the QMP
> +# wire ABI, and therefore not returned by this command.
> +#
> +# Since: 2.5
> +##
> +{ 'command': 'query-qmp-schema',
> +  'returns': [ 'SchemaInfo' ],
> +  'gen': false }                # just to simplify qmp_query_json()
> +
> +##
> +# @SchemaMetaType
> +#
> +# This is a @SchemaInfo's meta type, i.e. the kind of entity it
> +# describes.
> +#
> +# @builtin: a predefined type such as 'int' or 'bool'.
> +#
> +# @enum: an enumeration type
> +#
> +# @array: an array type
> +#
> +# @object: an object type (struct or union)
> +#
> +# @alternate: an alternate type
> +#
> +# @command: a QMP command
> +#
> +# @event: a QMP event
> +#
> +# Since: 2.5
> +##
> +{ 'enum': 'SchemaMetaType',
> +  'data': [ 'builtin', 'enum', 'array', 'object', 'alternate',
> +            'command', 'event' ] }
> +


Small nit, don't we usually put type definitions before the
commands that reference them?