qemu/qapi/qobject-input-visitor.c

745 lines
20 KiB
C
Raw Normal View History

/*
* Input Visitor
*
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
* Copyright (C) 2012-2017 Red Hat, Inc.
* Copyright IBM, Corp. 2011
*
* Authors:
* Anthony Liguori <aliguori@us.ibm.com>
*
* This work is licensed under the terms of the GNU LGPL, version 2.1 or later.
* See the COPYING.LIB file in the top-level directory.
*
*/
#include "qemu/osdep.h"
#include <math.h>
2016-03-14 16:01:28 +08:00
#include "qapi/error.h"
#include "qapi/qobject-input-visitor.h"
#include "qapi/visitor-impl.h"
#include "qemu/queue.h"
#include "qapi/qmp/qjson.h"
#include "qapi/qmp/qbool.h"
#include "qapi/qmp/qdict.h"
#include "qapi/qmp/qerror.h"
#include "qapi/qmp/qlist.h"
#include "qapi/qmp/qnull.h"
#include "qapi/qmp/qnum.h"
#include "qapi/qmp/qstring.h"
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
#include "qemu/cutils.h"
#include "qemu/option.h"
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
typedef struct StackObject {
const char *name; /* Name of @obj in its parent, if any */
QObject *obj; /* QDict or QList being visited */
2016-06-10 00:48:34 +08:00
void *qapi; /* sanity check that caller uses same pointer */
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
GHashTable *h; /* If @obj is QDict: unvisited keys */
const QListEntry *entry; /* If @obj is QList: unvisited tail */
unsigned index; /* If @obj is QList: list index of @entry */
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
QSLIST_ENTRY(StackObject) node; /* parent */
} StackObject;
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
struct QObjectInputVisitor {
Visitor visitor;
/* Root of visit at visitor creation. */
QObject *root;
bool keyval; /* Assume @root made with keyval_parse() */
/* Stack of objects being visited (all entries will be either
* QDict or QList). */
QSLIST_HEAD(, StackObject) stack;
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
GString *errname; /* Accumulator for full_name() */
};
static QObjectInputVisitor *to_qiv(Visitor *v)
{
return container_of(v, QObjectInputVisitor, visitor);
}
/*
* Find the full name of something @qiv is currently visiting.
* @qiv is visiting something named @name in the stack of containers
* @qiv->stack.
* If @n is zero, return its full name.
* If @n is positive, return the full name of the @n-th container
* counting from the top. The stack of containers must have at least
* @n elements.
* The returned string is valid until the next full_name_nth(@v) or
* destruction of @v.
*/
static const char *full_name_nth(QObjectInputVisitor *qiv, const char *name,
int n)
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
{
StackObject *so;
char buf[32];
if (qiv->errname) {
g_string_truncate(qiv->errname, 0);
} else {
qiv->errname = g_string_new("");
}
QSLIST_FOREACH(so , &qiv->stack, node) {
if (n) {
n--;
} else if (qobject_type(so->obj) == QTYPE_QDICT) {
g_string_prepend(qiv->errname, name ?: "<anonymous>");
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
g_string_prepend_c(qiv->errname, '.');
} else {
snprintf(buf, sizeof(buf),
qiv->keyval ? ".%u" : "[%u]",
so->index);
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
g_string_prepend(qiv->errname, buf);
}
name = so->name;
}
assert(!n);
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
if (name) {
g_string_prepend(qiv->errname, name);
} else if (qiv->errname->str[0] == '.') {
g_string_erase(qiv->errname, 0, 1);
} else if (!qiv->errname->str[0]) {
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
return "<anonymous>";
}
return qiv->errname->str;
}
static const char *full_name(QObjectInputVisitor *qiv, const char *name)
{
return full_name_nth(qiv, name, 0);
}
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
static QObject *qobject_input_try_get_object(QObjectInputVisitor *qiv,
const char *name,
bool consume)
{
StackObject *tos;
QObject *qobj;
QObject *ret;
if (QSLIST_EMPTY(&qiv->stack)) {
/* Starting at root, name is ignored. */
assert(qiv->root);
return qiv->root;
}
/* We are in a container; find the next element. */
tos = QSLIST_FIRST(&qiv->stack);
qobj = tos->obj;
assert(qobj);
if (qobject_type(qobj) == QTYPE_QDICT) {
assert(name);
ret = qdict_get(qobject_to(QDict, qobj), name);
if (tos->h && consume && ret) {
bool removed = g_hash_table_remove(tos->h, name);
assert(removed);
}
} else {
assert(qobject_type(qobj) == QTYPE_QLIST);
assert(!name);
if (tos->entry) {
ret = qlist_entry_obj(tos->entry);
if (consume) {
tos->entry = qlist_next(tos->entry);
}
} else {
ret = NULL;
}
qmp-input: Refactor when list is advanced In the QMP input visitor, visiting a list traverses two objects: the QAPI GenericList of the caller (which gets advanced in visit_next_list() regardless of this patch), and the QList input that we are converting to QAPI. For consistency with QDict visits, we want to consume elements from the input QList during the visit_type_FOO() for the list element; that is, we want ALL the code for consuming an input to live in qmp_input_get_object(), rather than having it split according to whether we are visiting a dict or a list. Making qmp_input_get_object() the common point of consumption will make it easier for a later patch to refactor visit_start_list() to cover the GenericList * head of a QAPI list, and in turn will get rid of the 'first' flag (which lived in qmp_input_next_list() pre-patch, and is hoisted to StackObject by this patch). This patch is therefore altering the post-condition use of 'entry', while keeping what gets visited unchanged, from: start_list next_list type_ELT ... next_list type_ELT next_list end_list visits 1st elt last elt entry NULL 1st elt 1st elt last elt last elt NULL gone where type_ELT() returns (entry ? entry : 1st elt) and next_list() steps entry to this usage: start_list next_list type_ELT ... next_list type_ELT next_list end_list visits 1st elt last elt entry 1st elt 1nd elt 2nd elt last elt NULL NULL gone where type_ELT() steps entry and returns the old entry, and next_list() leaves entry alone. Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-12-git-send-email-eblake@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:19 +08:00
if (consume) {
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
tos->index++;
qmp-input: Refactor when list is advanced In the QMP input visitor, visiting a list traverses two objects: the QAPI GenericList of the caller (which gets advanced in visit_next_list() regardless of this patch), and the QList input that we are converting to QAPI. For consistency with QDict visits, we want to consume elements from the input QList during the visit_type_FOO() for the list element; that is, we want ALL the code for consuming an input to live in qmp_input_get_object(), rather than having it split according to whether we are visiting a dict or a list. Making qmp_input_get_object() the common point of consumption will make it easier for a later patch to refactor visit_start_list() to cover the GenericList * head of a QAPI list, and in turn will get rid of the 'first' flag (which lived in qmp_input_next_list() pre-patch, and is hoisted to StackObject by this patch). This patch is therefore altering the post-condition use of 'entry', while keeping what gets visited unchanged, from: start_list next_list type_ELT ... next_list type_ELT next_list end_list visits 1st elt last elt entry NULL 1st elt 1st elt last elt last elt NULL gone where type_ELT() returns (entry ? entry : 1st elt) and next_list() steps entry to this usage: start_list next_list type_ELT ... next_list type_ELT next_list end_list visits 1st elt last elt entry 1st elt 1nd elt 2nd elt last elt NULL NULL gone where type_ELT() steps entry and returns the old entry, and next_list() leaves entry alone. Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-12-git-send-email-eblake@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:19 +08:00
}
}
return ret;
}
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
static QObject *qobject_input_get_object(QObjectInputVisitor *qiv,
const char *name,
bool consume, Error **errp)
{
QObject *obj = qobject_input_try_get_object(qiv, name, consume);
if (!obj) {
error_setg(errp, QERR_MISSING_PARAMETER, full_name(qiv, name));
}
return obj;
}
static const char *qobject_input_get_keyval(QObjectInputVisitor *qiv,
const char *name,
Error **errp)
{
QObject *qobj;
QString *qstr;
qobj = qobject_input_get_object(qiv, name, true, errp);
if (!qobj) {
return NULL;
}
qstr = qobject_to(QString, qobj);
if (!qstr) {
switch (qobject_type(qobj)) {
case QTYPE_QDICT:
case QTYPE_QLIST:
error_setg(errp, "Parameters '%s.*' are unexpected",
full_name(qiv, name));
return NULL;
default:
/* Non-string scalar (should this be an assertion?) */
error_setg(errp, "Internal error: parameter %s invalid",
full_name(qiv, name));
return NULL;
}
}
return qstring_get_str(qstr);
}
static const QListEntry *qobject_input_push(QObjectInputVisitor *qiv,
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
const char *name,
QObject *obj, void *qapi)
{
GHashTable *h;
StackObject *tos = g_new0(StackObject, 1);
QDict *qdict = qobject_to(QDict, obj);
QList *qlist = qobject_to(QList, obj);
const QDictEntry *entry;
assert(obj);
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
tos->name = name;
tos->obj = obj;
2016-06-10 00:48:34 +08:00
tos->qapi = qapi;
if (qdict) {
h = g_hash_table_new(g_str_hash, g_str_equal);
for (entry = qdict_first(qdict);
entry;
entry = qdict_next(qdict, entry)) {
g_hash_table_insert(h, (void *)qdict_entry_key(entry), NULL);
}
tos->h = h;
} else {
assert(qlist);
tos->entry = qlist_first(qlist);
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
tos->index = -1;
}
QSLIST_INSERT_HEAD(&qiv->stack, tos, node);
qapi: Simplify semantics of visit_next_list() The semantics of the list visit are somewhat baroque, with the following pseudocode when FooList is used: start() for (prev = head; cur = next(prev); prev = &cur) { visit(&cur->value) } Note that these semantics (advance before visit) requires that the first call to next() return the list head, while all other calls return the next element of the list; that is, every visitor implementation is required to track extra state to decide whether to return the input as-is, or to advance. It also requires an argument of 'GenericList **' to next(), solely because the first iteration might need to modify the caller's GenericList head, so that all other calls have to do a layer of dereferencing. Thankfully, we only have two uses of list visits in the entire code base: one in spapr_drc (which completely avoids visit_next_list(), feeding in integers from a different source than uint8List), and one in qapi-visit.py. That is, all other list visitors are generated in qapi-visit.c, and share the same paradigm based on a qapi FooList type, so we can refactor how lists are laid out with minimal churn among clients. We can greatly simplify things by hoisting the special case into the start() routine, and flipping the order in the loop to visit before advance: start(head) for (tail = *head; tail; tail = next(tail)) { visit(&tail->value) } With the simpler semantics, visitors have less state to track, the argument to next() is reduced to 'GenericList *', and it also becomes obvious whether an input visitor is allocating a FooList during visit_start_list() (rather than the old way of not knowing if an allocation happened until the first visit_next_list()). As a minor drawback, we now allocate in two functions instead of one, and have to pass the size to both functions (unless we were to tweak the input visitors to cache the size to start_list for reuse during next_list, but that defeats the goal of less visitor state). The signature of visit_start_list() is chosen to match visit_start_struct(), with the new parameters after 'name'. The spapr_drc case is a virtual visit, done by passing NULL for list, similarly to how NULL is passed to visit_start_struct() when a qapi type is not used in those visits. It was easy to provide these semantics for qmp-output and dealloc visitors, and a bit harder for qmp-input (several prerequisite patches refactored things to make this patch straightforward). But it turned out that the string and opts visitors munge enough other state during visit_next_list() to make it easier to just document and require a GenericList visit for now; an assertion will remind us to adjust things if we need the semantics in the future. Several pre-requisite cleanup patches made the reshuffling of the various visitors easier; particularly the qmp input visitor. Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-24-git-send-email-eblake@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:31 +08:00
return tos->entry;
}
static void qobject_input_check_struct(Visitor *v, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
StackObject *tos = QSLIST_FIRST(&qiv->stack);
GHashTableIter iter;
const char *key;
assert(tos && !tos->entry);
g_hash_table_iter_init(&iter, tos->h);
if (g_hash_table_iter_next(&iter, (void **)&key, NULL)) {
error_setg(errp, "Parameter '%s' is unexpected",
full_name(qiv, key));
qapi: Split visit_end_struct() into pieces As mentioned in previous patches, we want to call visit_end_struct() functions unconditionally, so that visitors can release resources tied up since the matching visit_start_struct() without also having to worry about error priority if more than one error occurs. Even though error_propagate() can be safely used to ignore a second error during cleanup caused by a first error, it is simpler if the cleanup cannot set an error. So, split out the error checking portion (basically, input visitors checking for unvisited keys) into a new function visit_check_struct(), which can be safely skipped if any earlier errors are encountered, and leave the cleanup portion (which never fails, but must be called unconditionally if visit_start_struct() succeeded) in visit_end_struct(). Generated code in qapi-visit.c has diffs resembling: |@@ -59,10 +59,12 @@ void visit_type_ACPIOSTInfo(Visitor *v, | goto out_obj; | } | visit_type_ACPIOSTInfo_members(v, obj, &err); |- error_propagate(errp, err); |- err = NULL; |+ if (err) { |+ goto out_obj; |+ } |+ visit_check_struct(v, &err); | out_obj: |- visit_end_struct(v, &err); |+ visit_end_struct(v); | out: and in qapi-event.c: @@ -47,7 +47,10 @@ void qapi_event_send_acpi_device_ost(ACP | goto out; | } | visit_type_q_obj_ACPI_DEVICE_OST_arg_members(v, &param, &err); |- visit_end_struct(v, err ? NULL : &err); |+ if (!err) { |+ visit_check_struct(v, &err); |+ } |+ visit_end_struct(v); | if (err) { | goto out; Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-20-git-send-email-eblake@redhat.com> [Conflict with a doc fixup resolved] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:27 +08:00
}
}
static void qobject_input_stack_object_free(StackObject *tos)
qapi: Split visit_end_struct() into pieces As mentioned in previous patches, we want to call visit_end_struct() functions unconditionally, so that visitors can release resources tied up since the matching visit_start_struct() without also having to worry about error priority if more than one error occurs. Even though error_propagate() can be safely used to ignore a second error during cleanup caused by a first error, it is simpler if the cleanup cannot set an error. So, split out the error checking portion (basically, input visitors checking for unvisited keys) into a new function visit_check_struct(), which can be safely skipped if any earlier errors are encountered, and leave the cleanup portion (which never fails, but must be called unconditionally if visit_start_struct() succeeded) in visit_end_struct(). Generated code in qapi-visit.c has diffs resembling: |@@ -59,10 +59,12 @@ void visit_type_ACPIOSTInfo(Visitor *v, | goto out_obj; | } | visit_type_ACPIOSTInfo_members(v, obj, &err); |- error_propagate(errp, err); |- err = NULL; |+ if (err) { |+ goto out_obj; |+ } |+ visit_check_struct(v, &err); | out_obj: |- visit_end_struct(v, &err); |+ visit_end_struct(v); | out: and in qapi-event.c: @@ -47,7 +47,10 @@ void qapi_event_send_acpi_device_ost(ACP | goto out; | } | visit_type_q_obj_ACPI_DEVICE_OST_arg_members(v, &param, &err); |- visit_end_struct(v, err ? NULL : &err); |+ if (!err) { |+ visit_check_struct(v, &err); |+ } |+ visit_end_struct(v); | if (err) { | goto out; Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-20-git-send-email-eblake@redhat.com> [Conflict with a doc fixup resolved] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:27 +08:00
{
if (tos->h) {
g_hash_table_unref(tos->h);
}
qapi: Split visit_end_struct() into pieces As mentioned in previous patches, we want to call visit_end_struct() functions unconditionally, so that visitors can release resources tied up since the matching visit_start_struct() without also having to worry about error priority if more than one error occurs. Even though error_propagate() can be safely used to ignore a second error during cleanup caused by a first error, it is simpler if the cleanup cannot set an error. So, split out the error checking portion (basically, input visitors checking for unvisited keys) into a new function visit_check_struct(), which can be safely skipped if any earlier errors are encountered, and leave the cleanup portion (which never fails, but must be called unconditionally if visit_start_struct() succeeded) in visit_end_struct(). Generated code in qapi-visit.c has diffs resembling: |@@ -59,10 +59,12 @@ void visit_type_ACPIOSTInfo(Visitor *v, | goto out_obj; | } | visit_type_ACPIOSTInfo_members(v, obj, &err); |- error_propagate(errp, err); |- err = NULL; |+ if (err) { |+ goto out_obj; |+ } |+ visit_check_struct(v, &err); | out_obj: |- visit_end_struct(v, &err); |+ visit_end_struct(v); | out: and in qapi-event.c: @@ -47,7 +47,10 @@ void qapi_event_send_acpi_device_ost(ACP | goto out; | } | visit_type_q_obj_ACPI_DEVICE_OST_arg_members(v, &param, &err); |- visit_end_struct(v, err ? NULL : &err); |+ if (!err) { |+ visit_check_struct(v, &err); |+ } |+ visit_end_struct(v); | if (err) { | goto out; Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-20-git-send-email-eblake@redhat.com> [Conflict with a doc fixup resolved] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:27 +08:00
g_free(tos);
}
qapi: Split visit_end_struct() into pieces As mentioned in previous patches, we want to call visit_end_struct() functions unconditionally, so that visitors can release resources tied up since the matching visit_start_struct() without also having to worry about error priority if more than one error occurs. Even though error_propagate() can be safely used to ignore a second error during cleanup caused by a first error, it is simpler if the cleanup cannot set an error. So, split out the error checking portion (basically, input visitors checking for unvisited keys) into a new function visit_check_struct(), which can be safely skipped if any earlier errors are encountered, and leave the cleanup portion (which never fails, but must be called unconditionally if visit_start_struct() succeeded) in visit_end_struct(). Generated code in qapi-visit.c has diffs resembling: |@@ -59,10 +59,12 @@ void visit_type_ACPIOSTInfo(Visitor *v, | goto out_obj; | } | visit_type_ACPIOSTInfo_members(v, obj, &err); |- error_propagate(errp, err); |- err = NULL; |+ if (err) { |+ goto out_obj; |+ } |+ visit_check_struct(v, &err); | out_obj: |- visit_end_struct(v, &err); |+ visit_end_struct(v); | out: and in qapi-event.c: @@ -47,7 +47,10 @@ void qapi_event_send_acpi_device_ost(ACP | goto out; | } | visit_type_q_obj_ACPI_DEVICE_OST_arg_members(v, &param, &err); |- visit_end_struct(v, err ? NULL : &err); |+ if (!err) { |+ visit_check_struct(v, &err); |+ } |+ visit_end_struct(v); | if (err) { | goto out; Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-20-git-send-email-eblake@redhat.com> [Conflict with a doc fixup resolved] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:27 +08:00
static void qobject_input_pop(Visitor *v, void **obj)
{
QObjectInputVisitor *qiv = to_qiv(v);
StackObject *tos = QSLIST_FIRST(&qiv->stack);
assert(tos && tos->qapi == obj);
QSLIST_REMOVE_HEAD(&qiv->stack, node);
qobject_input_stack_object_free(tos);
}
static void qobject_input_start_struct(Visitor *v, const char *name, void **obj,
size_t size, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
if (obj) {
*obj = NULL;
}
if (!qobj) {
return;
}
if (qobject_type(qobj) != QTYPE_QDICT) {
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
error_setg(errp, QERR_INVALID_PARAMETER_TYPE,
full_name(qiv, name), "object");
return;
}
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
qobject_input_push(qiv, name, qobj, obj);
if (obj) {
*obj = g_malloc0(size);
}
}
static void qobject_input_end_struct(Visitor *v, void **obj)
{
QObjectInputVisitor *qiv = to_qiv(v);
StackObject *tos = QSLIST_FIRST(&qiv->stack);
assert(qobject_type(tos->obj) == QTYPE_QDICT && tos->h);
qobject_input_pop(v, obj);
}
static void qobject_input_start_list(Visitor *v, const char *name,
GenericList **list, size_t size,
Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
qapi: Simplify semantics of visit_next_list() The semantics of the list visit are somewhat baroque, with the following pseudocode when FooList is used: start() for (prev = head; cur = next(prev); prev = &cur) { visit(&cur->value) } Note that these semantics (advance before visit) requires that the first call to next() return the list head, while all other calls return the next element of the list; that is, every visitor implementation is required to track extra state to decide whether to return the input as-is, or to advance. It also requires an argument of 'GenericList **' to next(), solely because the first iteration might need to modify the caller's GenericList head, so that all other calls have to do a layer of dereferencing. Thankfully, we only have two uses of list visits in the entire code base: one in spapr_drc (which completely avoids visit_next_list(), feeding in integers from a different source than uint8List), and one in qapi-visit.py. That is, all other list visitors are generated in qapi-visit.c, and share the same paradigm based on a qapi FooList type, so we can refactor how lists are laid out with minimal churn among clients. We can greatly simplify things by hoisting the special case into the start() routine, and flipping the order in the loop to visit before advance: start(head) for (tail = *head; tail; tail = next(tail)) { visit(&tail->value) } With the simpler semantics, visitors have less state to track, the argument to next() is reduced to 'GenericList *', and it also becomes obvious whether an input visitor is allocating a FooList during visit_start_list() (rather than the old way of not knowing if an allocation happened until the first visit_next_list()). As a minor drawback, we now allocate in two functions instead of one, and have to pass the size to both functions (unless we were to tweak the input visitors to cache the size to start_list for reuse during next_list, but that defeats the goal of less visitor state). The signature of visit_start_list() is chosen to match visit_start_struct(), with the new parameters after 'name'. The spapr_drc case is a virtual visit, done by passing NULL for list, similarly to how NULL is passed to visit_start_struct() when a qapi type is not used in those visits. It was easy to provide these semantics for qmp-output and dealloc visitors, and a bit harder for qmp-input (several prerequisite patches refactored things to make this patch straightforward). But it turned out that the string and opts visitors munge enough other state during visit_next_list() to make it easier to just document and require a GenericList visit for now; an assertion will remind us to adjust things if we need the semantics in the future. Several pre-requisite cleanup patches made the reshuffling of the various visitors easier; particularly the qmp input visitor. Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-24-git-send-email-eblake@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:31 +08:00
const QListEntry *entry;
if (list) {
*list = NULL;
}
if (!qobj) {
return;
}
if (qobject_type(qobj) != QTYPE_QLIST) {
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
error_setg(errp, QERR_INVALID_PARAMETER_TYPE,
full_name(qiv, name), "array");
return;
}
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
entry = qobject_input_push(qiv, name, qobj, list);
if (entry && list) {
*list = g_malloc0(size);
qapi: Simplify semantics of visit_next_list() The semantics of the list visit are somewhat baroque, with the following pseudocode when FooList is used: start() for (prev = head; cur = next(prev); prev = &cur) { visit(&cur->value) } Note that these semantics (advance before visit) requires that the first call to next() return the list head, while all other calls return the next element of the list; that is, every visitor implementation is required to track extra state to decide whether to return the input as-is, or to advance. It also requires an argument of 'GenericList **' to next(), solely because the first iteration might need to modify the caller's GenericList head, so that all other calls have to do a layer of dereferencing. Thankfully, we only have two uses of list visits in the entire code base: one in spapr_drc (which completely avoids visit_next_list(), feeding in integers from a different source than uint8List), and one in qapi-visit.py. That is, all other list visitors are generated in qapi-visit.c, and share the same paradigm based on a qapi FooList type, so we can refactor how lists are laid out with minimal churn among clients. We can greatly simplify things by hoisting the special case into the start() routine, and flipping the order in the loop to visit before advance: start(head) for (tail = *head; tail; tail = next(tail)) { visit(&tail->value) } With the simpler semantics, visitors have less state to track, the argument to next() is reduced to 'GenericList *', and it also becomes obvious whether an input visitor is allocating a FooList during visit_start_list() (rather than the old way of not knowing if an allocation happened until the first visit_next_list()). As a minor drawback, we now allocate in two functions instead of one, and have to pass the size to both functions (unless we were to tweak the input visitors to cache the size to start_list for reuse during next_list, but that defeats the goal of less visitor state). The signature of visit_start_list() is chosen to match visit_start_struct(), with the new parameters after 'name'. The spapr_drc case is a virtual visit, done by passing NULL for list, similarly to how NULL is passed to visit_start_struct() when a qapi type is not used in those visits. It was easy to provide these semantics for qmp-output and dealloc visitors, and a bit harder for qmp-input (several prerequisite patches refactored things to make this patch straightforward). But it turned out that the string and opts visitors munge enough other state during visit_next_list() to make it easier to just document and require a GenericList visit for now; an assertion will remind us to adjust things if we need the semantics in the future. Several pre-requisite cleanup patches made the reshuffling of the various visitors easier; particularly the qmp input visitor. Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-24-git-send-email-eblake@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:31 +08:00
}
}
static GenericList *qobject_input_next_list(Visitor *v, GenericList *tail,
size_t size)
{
QObjectInputVisitor *qiv = to_qiv(v);
StackObject *tos = QSLIST_FIRST(&qiv->stack);
assert(tos && qobject_to(QList, tos->obj));
if (!tos->entry) {
return NULL;
}
qapi: Simplify semantics of visit_next_list() The semantics of the list visit are somewhat baroque, with the following pseudocode when FooList is used: start() for (prev = head; cur = next(prev); prev = &cur) { visit(&cur->value) } Note that these semantics (advance before visit) requires that the first call to next() return the list head, while all other calls return the next element of the list; that is, every visitor implementation is required to track extra state to decide whether to return the input as-is, or to advance. It also requires an argument of 'GenericList **' to next(), solely because the first iteration might need to modify the caller's GenericList head, so that all other calls have to do a layer of dereferencing. Thankfully, we only have two uses of list visits in the entire code base: one in spapr_drc (which completely avoids visit_next_list(), feeding in integers from a different source than uint8List), and one in qapi-visit.py. That is, all other list visitors are generated in qapi-visit.c, and share the same paradigm based on a qapi FooList type, so we can refactor how lists are laid out with minimal churn among clients. We can greatly simplify things by hoisting the special case into the start() routine, and flipping the order in the loop to visit before advance: start(head) for (tail = *head; tail; tail = next(tail)) { visit(&tail->value) } With the simpler semantics, visitors have less state to track, the argument to next() is reduced to 'GenericList *', and it also becomes obvious whether an input visitor is allocating a FooList during visit_start_list() (rather than the old way of not knowing if an allocation happened until the first visit_next_list()). As a minor drawback, we now allocate in two functions instead of one, and have to pass the size to both functions (unless we were to tweak the input visitors to cache the size to start_list for reuse during next_list, but that defeats the goal of less visitor state). The signature of visit_start_list() is chosen to match visit_start_struct(), with the new parameters after 'name'. The spapr_drc case is a virtual visit, done by passing NULL for list, similarly to how NULL is passed to visit_start_struct() when a qapi type is not used in those visits. It was easy to provide these semantics for qmp-output and dealloc visitors, and a bit harder for qmp-input (several prerequisite patches refactored things to make this patch straightforward). But it turned out that the string and opts visitors munge enough other state during visit_next_list() to make it easier to just document and require a GenericList visit for now; an assertion will remind us to adjust things if we need the semantics in the future. Several pre-requisite cleanup patches made the reshuffling of the various visitors easier; particularly the qmp input visitor. Signed-off-by: Eric Blake <eblake@redhat.com> Message-Id: <1461879932-9020-24-git-send-email-eblake@redhat.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2016-04-29 05:45:31 +08:00
tail->next = g_malloc0(size);
return tail->next;
}
static void qobject_input_check_list(Visitor *v, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
StackObject *tos = QSLIST_FIRST(&qiv->stack);
assert(tos && qobject_to(QList, tos->obj));
if (tos->entry) {
error_setg(errp, "Only %u list elements expected in %s",
tos->index + 1, full_name_nth(qiv, NULL, 1));
}
}
static void qobject_input_end_list(Visitor *v, void **obj)
{
QObjectInputVisitor *qiv = to_qiv(v);
StackObject *tos = QSLIST_FIRST(&qiv->stack);
assert(qobject_type(tos->obj) == QTYPE_QLIST && !tos->h);
qobject_input_pop(v, obj);
}
static void qobject_input_start_alternate(Visitor *v, const char *name,
GenericAlternate **obj, size_t size,
Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, false, errp);
if (!qobj) {
*obj = NULL;
return;
}
*obj = g_malloc0(size);
(*obj)->type = qobject_type(qobj);
}
static void qobject_input_type_int64(Visitor *v, const char *name, int64_t *obj,
Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
QNum *qnum;
if (!qobj) {
return;
}
qnum = qobject_to(QNum, qobj);
if (!qnum || !qnum_get_try_int(qnum, obj)) {
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
error_setg(errp, QERR_INVALID_PARAMETER_TYPE,
full_name(qiv, name), "integer");
}
}
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
static void qobject_input_type_int64_keyval(Visitor *v, const char *name,
int64_t *obj, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
const char *str = qobject_input_get_keyval(qiv, name, errp);
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
if (!str) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
return;
}
if (qemu_strtoi64(str, NULL, 0, obj) < 0) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
/* TODO report -ERANGE more nicely */
error_setg(errp, QERR_INVALID_PARAMETER_VALUE,
full_name(qiv, name), "integer");
}
}
static void qobject_input_type_uint64(Visitor *v, const char *name,
uint64_t *obj, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
QNum *qnum;
int64_t val;
if (!qobj) {
return;
}
qnum = qobject_to(QNum, qobj);
if (!qnum) {
goto err;
}
if (qnum_get_try_uint(qnum, obj)) {
return;
}
/* Need to accept negative values for backward compatibility */
if (qnum_get_try_int(qnum, &val)) {
*obj = val;
return;
}
err:
error_setg(errp, QERR_INVALID_PARAMETER_VALUE,
full_name(qiv, name), "uint64");
}
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
static void qobject_input_type_uint64_keyval(Visitor *v, const char *name,
uint64_t *obj, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
const char *str = qobject_input_get_keyval(qiv, name, errp);
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
if (!str) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
return;
}
if (qemu_strtou64(str, NULL, 0, obj) < 0) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
/* TODO report -ERANGE more nicely */
error_setg(errp, QERR_INVALID_PARAMETER_VALUE,
full_name(qiv, name), "integer");
}
}
static void qobject_input_type_bool(Visitor *v, const char *name, bool *obj,
Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
QBool *qbool;
if (!qobj) {
return;
}
qbool = qobject_to(QBool, qobj);
if (!qbool) {
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
error_setg(errp, QERR_INVALID_PARAMETER_TYPE,
full_name(qiv, name), "boolean");
return;
}
*obj = qbool_get_bool(qbool);
}
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
static void qobject_input_type_bool_keyval(Visitor *v, const char *name,
bool *obj, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
const char *str = qobject_input_get_keyval(qiv, name, errp);
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
if (!str) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
return;
}
if (!strcmp(str, "on")) {
*obj = true;
} else if (!strcmp(str, "off")) {
*obj = false;
} else {
error_setg(errp, QERR_INVALID_PARAMETER_VALUE,
full_name(qiv, name), "'on' or 'off'");
}
}
static void qobject_input_type_str(Visitor *v, const char *name, char **obj,
Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
QString *qstr;
*obj = NULL;
if (!qobj) {
return;
}
qstr = qobject_to(QString, qobj);
if (!qstr) {
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
error_setg(errp, QERR_INVALID_PARAMETER_TYPE,
full_name(qiv, name), "string");
return;
}
*obj = g_strdup(qstring_get_str(qstr));
}
static void qobject_input_type_str_keyval(Visitor *v, const char *name,
char **obj, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
const char *str = qobject_input_get_keyval(qiv, name, errp);
*obj = g_strdup(str);
}
static void qobject_input_type_number(Visitor *v, const char *name, double *obj,
Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
QNum *qnum;
if (!qobj) {
return;
}
qnum = qobject_to(QNum, qobj);
if (!qnum) {
error_setg(errp, QERR_INVALID_PARAMETER_TYPE,
full_name(qiv, name), "number");
return;
}
*obj = qnum_get_double(qnum);
}
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
static void qobject_input_type_number_keyval(Visitor *v, const char *name,
double *obj, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
const char *str = qobject_input_get_keyval(qiv, name, errp);
double val;
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
if (!str) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
return;
}
if (qemu_strtod_finite(str, NULL, &val)) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
/* TODO report -ERANGE more nicely */
error_setg(errp, QERR_INVALID_PARAMETER_TYPE,
full_name(qiv, name), "number");
return;
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
}
*obj = val;
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
}
static void qobject_input_type_any(Visitor *v, const char *name, QObject **obj,
Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
*obj = NULL;
if (!qobj) {
return;
}
*obj = qobject_ref(qobj);
}
static void qobject_input_type_null(Visitor *v, const char *name,
QNull **obj, Error **errp)
2016-04-29 05:45:22 +08:00
{
QObjectInputVisitor *qiv = to_qiv(v);
QObject *qobj = qobject_input_get_object(qiv, name, true, errp);
*obj = NULL;
if (!qobj) {
return;
}
if (qobject_type(qobj) != QTYPE_QNULL) {
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
error_setg(errp, QERR_INVALID_PARAMETER_TYPE,
full_name(qiv, name), "null");
return;
}
*obj = qnull();
2016-04-29 05:45:22 +08:00
}
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
static void qobject_input_type_size_keyval(Visitor *v, const char *name,
uint64_t *obj, Error **errp)
{
QObjectInputVisitor *qiv = to_qiv(v);
const char *str = qobject_input_get_keyval(qiv, name, errp);
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
if (!str) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
return;
}
if (qemu_strtosz(str, NULL, obj) < 0) {
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
/* TODO report -ERANGE more nicely */
error_setg(errp, QERR_INVALID_PARAMETER_VALUE,
full_name(qiv, name), "size");
}
}
static void qobject_input_optional(Visitor *v, const char *name, bool *present)
{
QObjectInputVisitor *qiv = to_qiv(v);
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
QObject *qobj = qobject_input_try_get_object(qiv, name, false);
if (!qobj) {
*present = false;
return;
}
*present = true;
}
static void qobject_input_free(Visitor *v)
{
QObjectInputVisitor *qiv = to_qiv(v);
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
while (!QSLIST_EMPTY(&qiv->stack)) {
StackObject *tos = QSLIST_FIRST(&qiv->stack);
QSLIST_REMOVE_HEAD(&qiv->stack, node);
qobject_input_stack_object_free(tos);
}
qobject_unref(qiv->root);
qapi: Improve qobject input visitor error reporting Error messages refer to nodes of the QObject being visited by name. Trouble is the names are sometimes less than helpful: * The name of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "null". We commonly pass NULL. Not good. Avoiding errors "at the root" mitigates. For instance, visit_start_struct() can only fail when the visited object is not a dictionary, and we commonly ensure it is beforehand. * The name of a QDict's member is the member key. Good enough only when this happens to be unique. * The name of a QList's member is "null". Not good. Improve error messages by referring to nodes by path instead, as follows: * The path of the root QObject is whatever @name argument got passed to the visitor, except NULL gets mapped to "<anonymous>". * The path of a root QDict's member is the member key. * The path of a root QList's member is "[%u]", where %u is the list index, starting at zero. * The path of a non-root QDict's member is the path of the QDict concatenated with "." and the member key. * The path of a non-root QList's member is the path of the QList concatenated with "[%u]", where %u is the list index. For example, the incorrect QMP command { "execute": "blockdev-add", "arguments": { "node-name": "foo", "driver": "raw", "file": {"driver": "file" } } } now fails with {"error": {"class": "GenericError", "desc": "Parameter 'file.filename' is missing"}} instead of {"error": {"class": "GenericError", "desc": "Parameter 'filename' is missing"}} and { "execute": "input-send-event", "arguments": { "device": "bar", "events": [ [] ] } } now fails with {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'events[0]', expected: object"}} instead of {"error": {"class": "GenericError", "desc": "Invalid parameter type for 'null', expected: QDict"}} Aside: calling the thing "parameter" is suboptimal for QMP, because the root object is "arguments" there. The qobject output visitor doesn't have this problem because it should not fail. Same for dealloc and clone visitors. The string visitors don't have this problem because they visit just one value, whose name needs to be passed to the visitor as @name. The string output visitor shouldn't fail anyway. The options visitor uses QemuOpts names. Their name space is flat, so the use of QDict member keys as names is fine. NULL names used with roots and lists could conceivably result in bad error messages. Left for another day. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Message-Id: <1488544368-30622-15-git-send-email-armbru@redhat.com>
2017-03-03 20:32:34 +08:00
if (qiv->errname) {
g_string_free(qiv->errname, TRUE);
}
g_free(qiv);
}
static QObjectInputVisitor *qobject_input_visitor_base_new(QObject *obj)
{
QObjectInputVisitor *v = g_malloc0(sizeof(*v));
assert(obj);
v->visitor.type = VISITOR_INPUT;
v->visitor.start_struct = qobject_input_start_struct;
v->visitor.check_struct = qobject_input_check_struct;
v->visitor.end_struct = qobject_input_end_struct;
v->visitor.start_list = qobject_input_start_list;
v->visitor.next_list = qobject_input_next_list;
v->visitor.check_list = qobject_input_check_list;
v->visitor.end_list = qobject_input_end_list;
v->visitor.start_alternate = qobject_input_start_alternate;
v->visitor.optional = qobject_input_optional;
v->visitor.free = qobject_input_free;
v->root = qobject_ref(obj);
return v;
}
Visitor *qobject_input_visitor_new(QObject *obj)
{
QObjectInputVisitor *v = qobject_input_visitor_base_new(obj);
v->visitor.type_int64 = qobject_input_type_int64;
v->visitor.type_uint64 = qobject_input_type_uint64;
v->visitor.type_bool = qobject_input_type_bool;
v->visitor.type_str = qobject_input_type_str;
v->visitor.type_number = qobject_input_type_number;
v->visitor.type_any = qobject_input_type_any;
v->visitor.type_null = qobject_input_type_null;
return &v->visitor;
}
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
Visitor *qobject_input_visitor_new_keyval(QObject *obj)
{
QObjectInputVisitor *v = qobject_input_visitor_base_new(obj);
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
v->visitor.type_int64 = qobject_input_type_int64_keyval;
v->visitor.type_uint64 = qobject_input_type_uint64_keyval;
v->visitor.type_bool = qobject_input_type_bool_keyval;
v->visitor.type_str = qobject_input_type_str_keyval;
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
v->visitor.type_number = qobject_input_type_number_keyval;
v->visitor.type_any = qobject_input_type_any;
v->visitor.type_null = qobject_input_type_null;
v->visitor.type_size = qobject_input_type_size_keyval;
v->keyval = true;
qapi: qobject input visitor variant for use with keyval_parse() Currently the QObjectInputVisitor assumes that all scalar values are directly represented as the final types declared by the thing being visited. i.e. it assumes an 'int' is using QInt, and a 'bool' is using QBool, etc. This is good when QObjectInputVisitor is fed a QObject that came from a JSON document on the QMP monitor, as it will strictly validate correctness. To allow QObjectInputVisitor to be reused for visiting a QObject originating from keyval_parse(), an alternative mode is needed where all the scalars types are represented as QString and converted on the fly to the final desired type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Message-Id: <1475246744-29302-8-git-send-email-berrange@redhat.com> Rebased, conflicts resolved, commit message updated to refer to keyval_parse(). autocast replaced by keyval in identifiers, noautocast replaced by fail in tests. Fix qobject_input_type_uint64_keyval() not to reject '-', for QemuOpts compatibility: replace parse_uint_full() by open-coded parse_option_number(). The next commit will add suitable tests. Leave out the fancy ERANGE error reporting for now, but add a TODO comment. Add it qobject_input_type_int64_keyval() and qobject_input_type_number_keyval(), too. Open code parse_option_bool() and parse_option_size() so we have to call qobject_input_get_name() only when actually needed. Again, leave out ERANGE error reporting for now. QAPI/QMP downstream extension prefixes __RFQDN_ don't work, because keyval_parse() splits them at '.'. This will be addressed later in the series. qobject_input_type_int64_keyval(), qobject_input_type_uint64_keyval(), qobject_input_type_number_keyval() tweaked for style. Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Kevin Wolf <kwolf@redhat.com> Message-Id: <1488317230-26248-5-git-send-email-armbru@redhat.com>
2017-03-01 05:26:50 +08:00
return &v->visitor;
}
Visitor *qobject_input_visitor_new_str(const char *str,
const char *implied_key,
Error **errp)
{
bool is_json = str[0] == '{';
QObject *obj;
QDict *args;
Visitor *v;
if (is_json) {
obj = qobject_from_json(str, errp);
if (!obj) {
return NULL;
}
args = qobject_to(QDict, obj);
assert(args);
v = qobject_input_visitor_new(QOBJECT(args));
} else {
args = keyval_parse(str, implied_key, errp);
if (!args) {
return NULL;
}
v = qobject_input_visitor_new_keyval(QOBJECT(args));
}
qobject_unref(args);
return v;
}