qemu/tests/check-qom-proplist.c

637 lines
17 KiB
C
Raw Normal View History

/*
* Copyright (C) 2015 Red Hat, Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see
* <http://www.gnu.org/licenses/>.
*
* Author: Daniel P. Berrange <berrange@redhat.com>
*/
#include "qemu/osdep.h"
2016-03-14 11:01:28 +03:00
#include "qapi/error.h"
#include "qom/object.h"
#include "qemu/module.h"
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
#include "qemu/option.h"
#include "qemu/config-file.h"
#include "qom/object_interfaces.h"
#define TYPE_DUMMY "qemu-dummy"
typedef struct DummyObject DummyObject;
typedef struct DummyObjectClass DummyObjectClass;
#define DUMMY_OBJECT(obj) \
OBJECT_CHECK(DummyObject, (obj), TYPE_DUMMY)
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
typedef enum DummyAnimal DummyAnimal;
enum DummyAnimal {
DUMMY_FROG,
DUMMY_ALLIGATOR,
DUMMY_PLATYPUS,
DUMMY_LAST,
};
const QEnumLookup dummy_animal_map = {
.array = (const char *const[]) {
[DUMMY_FROG] = "frog",
[DUMMY_ALLIGATOR] = "alligator",
[DUMMY_PLATYPUS] = "platypus",
},
.size = DUMMY_LAST
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
};
struct DummyObject {
Object parent_obj;
bool bv;
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
DummyAnimal av;
char *sv;
};
struct DummyObjectClass {
ObjectClass parent_class;
};
static void dummy_set_bv(Object *obj,
bool value,
Error **errp)
{
DummyObject *dobj = DUMMY_OBJECT(obj);
dobj->bv = value;
}
static bool dummy_get_bv(Object *obj,
Error **errp)
{
DummyObject *dobj = DUMMY_OBJECT(obj);
return dobj->bv;
}
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
static void dummy_set_av(Object *obj,
int value,
Error **errp)
{
DummyObject *dobj = DUMMY_OBJECT(obj);
dobj->av = value;
}
static int dummy_get_av(Object *obj,
Error **errp)
{
DummyObject *dobj = DUMMY_OBJECT(obj);
return dobj->av;
}
static void dummy_set_sv(Object *obj,
const char *value,
Error **errp)
{
DummyObject *dobj = DUMMY_OBJECT(obj);
g_free(dobj->sv);
dobj->sv = g_strdup(value);
}
static char *dummy_get_sv(Object *obj,
Error **errp)
{
DummyObject *dobj = DUMMY_OBJECT(obj);
return g_strdup(dobj->sv);
}
static void dummy_init(Object *obj)
{
object_property_add_bool(obj, "bv",
dummy_get_bv,
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
dummy_set_bv);
}
static void dummy_class_init(ObjectClass *cls, void *data)
{
object_class_property_add_str(cls, "sv",
dummy_get_sv,
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
dummy_set_sv);
object_class_property_add_enum(cls, "av",
"DummyAnimal",
&dummy_animal_map,
dummy_get_av,
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
dummy_set_av);
}
static void dummy_finalize(Object *obj)
{
DummyObject *dobj = DUMMY_OBJECT(obj);
g_free(dobj->sv);
}
static const TypeInfo dummy_info = {
.name = TYPE_DUMMY,
.parent = TYPE_OBJECT,
.instance_size = sizeof(DummyObject),
.instance_init = dummy_init,
.instance_finalize = dummy_finalize,
.class_size = sizeof(DummyObjectClass),
.class_init = dummy_class_init,
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
.interfaces = (InterfaceInfo[]) {
{ TYPE_USER_CREATABLE },
{ }
}
};
/*
* The following 3 object classes are used to
* simulate the kind of relationships seen in
* qdev, which result in complex object
* property destruction ordering.
*
* DummyDev has a 'bus' child to a DummyBus
* DummyBus has a 'backend' child to a DummyBackend
* DummyDev has a 'backend' link to DummyBackend
*
* When DummyDev is finalized, it unparents the
* DummyBackend, which unparents the DummyDev
* which deletes the 'backend' link from DummyDev
* to DummyBackend. This illustrates that the
* object_property_del_all() method needs to
* cope with the list of properties being changed
* while it iterates over them.
*/
typedef struct DummyDev DummyDev;
typedef struct DummyDevClass DummyDevClass;
typedef struct DummyBus DummyBus;
typedef struct DummyBusClass DummyBusClass;
typedef struct DummyBackend DummyBackend;
typedef struct DummyBackendClass DummyBackendClass;
#define TYPE_DUMMY_DEV "qemu-dummy-dev"
#define TYPE_DUMMY_BUS "qemu-dummy-bus"
#define TYPE_DUMMY_BACKEND "qemu-dummy-backend"
#define DUMMY_DEV(obj) \
OBJECT_CHECK(DummyDev, (obj), TYPE_DUMMY_DEV)
#define DUMMY_BUS(obj) \
OBJECT_CHECK(DummyBus, (obj), TYPE_DUMMY_BUS)
#define DUMMY_BACKEND(obj) \
OBJECT_CHECK(DummyBackend, (obj), TYPE_DUMMY_BACKEND)
struct DummyDev {
Object parent_obj;
DummyBus *bus;
};
struct DummyDevClass {
ObjectClass parent_class;
};
struct DummyBus {
Object parent_obj;
DummyBackend *backend;
};
struct DummyBusClass {
ObjectClass parent_class;
};
struct DummyBackend {
Object parent_obj;
};
struct DummyBackendClass {
ObjectClass parent_class;
};
static void dummy_dev_finalize(Object *obj)
{
DummyDev *dev = DUMMY_DEV(obj);
object_unref(OBJECT(dev->bus));
}
static void dummy_dev_init(Object *obj)
{
DummyDev *dev = DUMMY_DEV(obj);
DummyBus *bus = DUMMY_BUS(object_new(TYPE_DUMMY_BUS));
DummyBackend *backend = DUMMY_BACKEND(object_new(TYPE_DUMMY_BACKEND));
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
object_property_add_child(obj, "bus", OBJECT(bus));
dev->bus = bus;
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
object_property_add_child(OBJECT(bus), "backend", OBJECT(backend));
bus->backend = backend;
object_property_add_link(obj, "backend", TYPE_DUMMY_BACKEND,
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
(Object **)&bus->backend, NULL, 0);
}
static void dummy_dev_unparent(Object *obj)
{
DummyDev *dev = DUMMY_DEV(obj);
object_unparent(OBJECT(dev->bus));
}
static void dummy_dev_class_init(ObjectClass *klass, void *opaque)
{
klass->unparent = dummy_dev_unparent;
}
static void dummy_bus_finalize(Object *obj)
{
DummyBus *bus = DUMMY_BUS(obj);
object_unref(OBJECT(bus->backend));
}
static void dummy_bus_init(Object *obj)
{
}
static void dummy_bus_unparent(Object *obj)
{
DummyBus *bus = DUMMY_BUS(obj);
object_property_del(obj->parent, "backend");
object_unparent(OBJECT(bus->backend));
}
static void dummy_bus_class_init(ObjectClass *klass, void *opaque)
{
klass->unparent = dummy_bus_unparent;
}
static void dummy_backend_init(Object *obj)
{
}
static const TypeInfo dummy_dev_info = {
.name = TYPE_DUMMY_DEV,
.parent = TYPE_OBJECT,
.instance_size = sizeof(DummyDev),
.instance_init = dummy_dev_init,
.instance_finalize = dummy_dev_finalize,
.class_size = sizeof(DummyDevClass),
.class_init = dummy_dev_class_init,
};
static const TypeInfo dummy_bus_info = {
.name = TYPE_DUMMY_BUS,
.parent = TYPE_OBJECT,
.instance_size = sizeof(DummyBus),
.instance_init = dummy_bus_init,
.instance_finalize = dummy_bus_finalize,
.class_size = sizeof(DummyBusClass),
.class_init = dummy_bus_class_init,
};
static const TypeInfo dummy_backend_info = {
.name = TYPE_DUMMY_BACKEND,
.parent = TYPE_OBJECT,
.instance_size = sizeof(DummyBackend),
.instance_init = dummy_backend_init,
.class_size = sizeof(DummyBackendClass),
};
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
static QemuOptsList qemu_object_opts = {
.name = "object",
.implied_opt_name = "qom-type",
.head = QTAILQ_HEAD_INITIALIZER(qemu_object_opts.head),
.desc = {
{ }
},
};
static void test_dummy_createv(void)
{
Error *err = NULL;
Object *parent = object_get_objects_root();
DummyObject *dobj = DUMMY_OBJECT(
object_new_with_props(TYPE_DUMMY,
parent,
"dummy0",
&err,
"bv", "yes",
"sv", "Hiss hiss hiss",
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
"av", "platypus",
NULL));
g_assert(err == NULL);
g_assert_cmpstr(dobj->sv, ==, "Hiss hiss hiss");
g_assert(dobj->bv == true);
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
g_assert(dobj->av == DUMMY_PLATYPUS);
g_assert(object_resolve_path_component(parent, "dummy0")
== OBJECT(dobj));
object_unparent(OBJECT(dobj));
}
static Object *new_helper(Error **errp,
Object *parent,
...)
{
va_list vargs;
Object *obj;
va_start(vargs, parent);
obj = object_new_with_propv(TYPE_DUMMY,
parent,
"dummy0",
errp,
vargs);
va_end(vargs);
return obj;
}
static void test_dummy_createlist(void)
{
Error *err = NULL;
Object *parent = object_get_objects_root();
DummyObject *dobj = DUMMY_OBJECT(
new_helper(&err,
parent,
"bv", "yes",
"sv", "Hiss hiss hiss",
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
"av", "platypus",
NULL));
g_assert(err == NULL);
g_assert_cmpstr(dobj->sv, ==, "Hiss hiss hiss");
g_assert(dobj->bv == true);
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
g_assert(dobj->av == DUMMY_PLATYPUS);
g_assert(object_resolve_path_component(parent, "dummy0")
== OBJECT(dobj));
object_unparent(OBJECT(dobj));
}
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
static void test_dummy_createcmdl(void)
{
QemuOpts *opts;
DummyObject *dobj;
Error *err = NULL;
const char *params = TYPE_DUMMY \
",id=dev0," \
"bv=yes,sv=Hiss hiss hiss,av=platypus";
qemu_add_opts(&qemu_object_opts);
opts = qemu_opts_parse(&qemu_object_opts, params, true, &err);
g_assert(err == NULL);
g_assert(opts);
dobj = DUMMY_OBJECT(user_creatable_add_opts(opts, &err));
g_assert(err == NULL);
g_assert(dobj);
g_assert_cmpstr(dobj->sv, ==, "Hiss hiss hiss");
g_assert(dobj->bv == true);
g_assert(dobj->av == DUMMY_PLATYPUS);
user_creatable_del("dev0", &error_abort);
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
object_unref(OBJECT(dobj));
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
/*
* cmdline-parsing via qemu_opts_parse() results in a QemuOpts entry
* corresponding to the Object's ID to be added to the QemuOptsList
* for objects. To avoid having this entry conflict with future
* Objects using the same ID (which can happen in cases where
* qemu_opts_parse() is used to parse the object params, such as
* with hmp_object_add() at the time of this comment), we need to
* check for this in user_creatable_del() and remove the QemuOpts if
* it is present.
*
monitor: fix object_del for command-line-created objects Currently objects specified on the command-line are only partially cleaned up when 'object_del' is issued in either HMP or QMP: the object itself is fully finalized, but the QemuOpts are not removed. This results in the following behavior: x86_64-softmmu/qemu-system-x86_64 -monitor stdio \ -object memory-backend-ram,id=ram1,size=256M QEMU 2.7.91 monitor - type 'help' for more information (qemu) object_del ram1 (qemu) object_del ram1 object 'ram1' not found (qemu) object_add memory-backend-ram,id=ram1,size=256M Duplicate ID 'ram1' for object Try "help object_add" for more information which can be an issue for use-cases like memory hotplug. This happens on the HMP side because hmp_object_add() attempts to create a temporary QemuOpts entry with ID 'ram1', which ends up conflicting with the command-line-created entry, since it was never cleaned up during the previous hmp_object_del() call. We address this by adding a check in user_creatable_del(), which is called by both qmp_object_del() and hmp_object_del() to handle the actual object cleanup, to determine whether an option group entry matching the object's ID is present and removing it if it is. Note that qmp_object_add() never attempts to create a temporary QemuOpts entry, so it does not encounter the duplicate ID error, which is why this isn't generally visible in libvirt. Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-3-git-send-email-mdroth@linux.vnet.ibm.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:32 +03:00
* The below check ensures this works as expected.
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
*/
monitor: fix object_del for command-line-created objects Currently objects specified on the command-line are only partially cleaned up when 'object_del' is issued in either HMP or QMP: the object itself is fully finalized, but the QemuOpts are not removed. This results in the following behavior: x86_64-softmmu/qemu-system-x86_64 -monitor stdio \ -object memory-backend-ram,id=ram1,size=256M QEMU 2.7.91 monitor - type 'help' for more information (qemu) object_del ram1 (qemu) object_del ram1 object 'ram1' not found (qemu) object_add memory-backend-ram,id=ram1,size=256M Duplicate ID 'ram1' for object Try "help object_add" for more information which can be an issue for use-cases like memory hotplug. This happens on the HMP side because hmp_object_add() attempts to create a temporary QemuOpts entry with ID 'ram1', which ends up conflicting with the command-line-created entry, since it was never cleaned up during the previous hmp_object_del() call. We address this by adding a check in user_creatable_del(), which is called by both qmp_object_del() and hmp_object_del() to handle the actual object cleanup, to determine whether an option group entry matching the object's ID is present and removing it if it is. Note that qmp_object_add() never attempts to create a temporary QemuOpts entry, so it does not encounter the duplicate ID error, which is why this isn't generally visible in libvirt. Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-3-git-send-email-mdroth@linux.vnet.ibm.com> Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:32 +03:00
g_assert_null(qemu_opts_find(&qemu_object_opts, "dev0"));
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
}
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
static void test_dummy_badenum(void)
{
Error *err = NULL;
Object *parent = object_get_objects_root();
Object *dobj =
object_new_with_props(TYPE_DUMMY,
parent,
"dummy0",
&err,
"bv", "yes",
"sv", "Hiss hiss hiss",
"av", "yeti",
NULL);
g_assert(dobj == NULL);
g_assert(err != NULL);
g_assert_cmpstr(error_get_pretty(err), ==,
"Invalid parameter 'yeti'");
g_assert(object_resolve_path_component(parent, "dummy0")
== NULL);
error_free(err);
}
static void test_dummy_getenum(void)
{
Error *err = NULL;
int val;
Object *parent = object_get_objects_root();
DummyObject *dobj = DUMMY_OBJECT(
object_new_with_props(TYPE_DUMMY,
parent,
"dummy0",
&err,
"av", "platypus",
NULL));
g_assert(err == NULL);
g_assert(dobj->av == DUMMY_PLATYPUS);
val = object_property_get_enum(OBJECT(dobj),
"av",
"DummyAnimal",
&error_abort);
g_assert(val == DUMMY_PLATYPUS);
/* A bad enum type name */
val = object_property_get_enum(OBJECT(dobj),
"av",
"BadAnimal",
&err);
error_free_or_abort(&err);
/* A non-enum property name */
val = object_property_get_enum(OBJECT(dobj),
"iv",
"DummyAnimal",
&err);
error_free_or_abort(&err);
object_unparent(OBJECT(dobj));
}
static void test_dummy_prop_iterator(ObjectPropertyIterator *iter,
const char *expected[], int n)
{
ObjectProperty *prop;
int i;
while ((prop = object_property_iter_next(iter))) {
for (i = 0; i < n; i++) {
if (!g_strcmp0(prop->name, expected[i])) {
break;
}
}
g_assert(i < n);
expected[i] = NULL;
}
for (i = 0; i < n; i++) {
g_assert(!expected[i]);
}
}
static void test_dummy_iterator(void)
{
const char *expected[] = {
"type", /* inherited from TYPE_OBJECT */
"sv", "av", /* class properties */
"bv"}; /* instance property */
Object *parent = object_get_objects_root();
DummyObject *dobj = DUMMY_OBJECT(
object_new_with_props(TYPE_DUMMY,
parent,
"dummy0",
&error_abort,
"bv", "yes",
"sv", "Hiss hiss hiss",
"av", "platypus",
NULL));
ObjectPropertyIterator iter;
object_property_iter_init(&iter, OBJECT(dobj));
test_dummy_prop_iterator(&iter, expected, ARRAY_SIZE(expected));
object_unparent(OBJECT(dobj));
}
static void test_dummy_class_iterator(void)
{
const char *expected[] = { "type", "av", "sv" };
ObjectPropertyIterator iter;
ObjectClass *klass = object_class_by_name(TYPE_DUMMY);
object_class_property_iter_init(&iter, klass);
test_dummy_prop_iterator(&iter, expected, ARRAY_SIZE(expected));
}
static void test_dummy_delchild(void)
{
Object *parent = object_get_objects_root();
DummyDev *dev = DUMMY_DEV(
object_new_with_props(TYPE_DUMMY_DEV,
parent,
"dev0",
&error_abort,
NULL));
object_unparent(OBJECT(dev));
}
static void test_qom_partial_path(void)
{
Object *root = object_get_objects_root();
Object *cont1 = container_get(root, "/cont1");
Object *obj1 = object_new(TYPE_DUMMY);
Object *obj2a = object_new(TYPE_DUMMY);
Object *obj2b = object_new(TYPE_DUMMY);
bool ambiguous;
/* Objects created:
* /cont1
* /cont1/obj1
* /cont1/obj2 (obj2a)
* /obj2 (obj2b)
*/
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
object_property_add_child(cont1, "obj1", obj1);
object_unref(obj1);
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
object_property_add_child(cont1, "obj2", obj2a);
object_unref(obj2a);
qom: Drop parameter @errp of object_property_add() & friends The only way object_property_add() can fail is when a property with the same name already exists. Since our property names are all hardcoded, failure is a programming error, and the appropriate way to handle it is passing &error_abort. Same for its variants, except for object_property_add_child(), which additionally fails when the child already has a parent. Parentage is also under program control, so this is a programming error, too. We have a bit over 500 callers. Almost half of them pass &error_abort, slightly fewer ignore errors, one test case handles errors, and the remaining few callers pass them to their own callers. The previous few commits demonstrated once again that ignoring programming errors is a bad idea. Of the few ones that pass on errors, several violate the Error API. The Error ** argument must be NULL, &error_abort, &error_fatal, or a pointer to a variable containing NULL. Passing an argument of the latter kind twice without clearing it in between is wrong: if the first call sets an error, it no longer points to NULL for the second call. ich9_pm_add_properties(), sparc32_ledma_realize(), sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize() are wrong that way. When the one appropriate choice of argument is &error_abort, letting users pick the argument is a bad idea. Drop parameter @errp and assert the preconditions instead. There's one exception to "duplicate property name is a programming error": the way object_property_add() implements the magic (and undocumented) "automatic arrayification". Don't drop @errp there. Instead, rename object_property_add() to object_property_try_add(), and add the obvious wrapper object_property_add(). Signed-off-by: Markus Armbruster <armbru@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-Id: <20200505152926.18877-15-armbru@redhat.com> [Two semantic rebase conflicts resolved]
2020-05-05 18:29:22 +03:00
object_property_add_child(root, "obj2", obj2b);
object_unref(obj2b);
ambiguous = false;
g_assert(!object_resolve_path_type("", TYPE_DUMMY, &ambiguous));
g_assert(ambiguous);
g_assert(!object_resolve_path_type("", TYPE_DUMMY, NULL));
ambiguous = false;
g_assert(!object_resolve_path("obj2", &ambiguous));
g_assert(ambiguous);
g_assert(!object_resolve_path("obj2", NULL));
ambiguous = false;
g_assert(object_resolve_path("obj1", &ambiguous) == obj1);
g_assert(!ambiguous);
g_assert(object_resolve_path("obj1", NULL) == obj1);
object_unparent(obj2b);
object_unparent(cont1);
}
int main(int argc, char **argv)
{
g_test_init(&argc, &argv, NULL);
module_call_init(MODULE_INIT_QOM);
type_register_static(&dummy_info);
type_register_static(&dummy_dev_info);
type_register_static(&dummy_bus_info);
type_register_static(&dummy_backend_info);
g_test_add_func("/qom/proplist/createlist", test_dummy_createlist);
g_test_add_func("/qom/proplist/createv", test_dummy_createv);
tests: check-qom-proplist: add checks for cmdline-created objects check-qom-proplist originally added tests for verifying that object-creation helpers object_new_with_{props,propv} behaved in similar fashion to the "traditional" method involving setting each individual property separately after object creation rather than via a single call. Another similar "helper" for creating Objects exists in the form of objects specified via -object command-line parameters. By that rationale, we extend check-qom-proplist to include similar checks for command-line-created objects by employing the same qemu_opts_parse()-based parsing the vl.c employs. This parser has a side-effect of parsing the object's options into a QemuOpt structure and registering this in the global QemuOptsList using the Object's ID. This can conflict with future Object instances that attempt to use the same ID if we don't ensure this is cleaned up as part of Object finalization, so we include a FIXME stub to test for this case, which will then be resolved in a subsequent patch. Suggested-by: Daniel Berrange <berrange@redhat.com> Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com> Cc: Markus Armbruster <armbru@redhat.com> Cc: Eric Blake <eblake@redhat.com> Cc: Daniel Berrange <berrange@redhat.com> Cc: qemu-stable@nongnu.org Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com> Reviewed-by: Markus Armbruster <armbru@redhat.com> Message-Id: <1496531612-22166-2-git-send-email-mdroth@linux.vnet.ibm.com> [Comment formatting tidied up] Signed-off-by: Markus Armbruster <armbru@redhat.com>
2017-06-04 02:13:31 +03:00
g_test_add_func("/qom/proplist/createcmdline", test_dummy_createcmdl);
qom: Add an object_property_add_enum() helper function A QOM property can be parsed as enum using the visit_type_enum() helper function, but this forces callers to use the more complex generic object_property_add() method when registering it. It also requires that users of that object have access to the string map when they want to read the property value. This patch introduces a specialized object_property_add_enum() method which simplifies the use of enum properties, so the setters/getters directly get passed the int value. typedef enum { MYDEV_TYPE_FROG, MYDEV_TYPE_ALLIGATOR, MYDEV_TYPE_PLATYPUS, MYDEV_TYPE_LAST } MyDevType; Then provide a table of enum <-> string mappings static const char *const mydevtypemap[MYDEV_TYPE_LAST + 1] = { [MYDEV_TYPE_FROG] = "frog", [MYDEV_TYPE_ALLIGATOR] = "alligator", [MYDEV_TYPE_PLATYPUS] = "platypus", [MYDEV_TYPE_LAST] = NULL, }; Assuming an object struct of typedef struct { Object parent_obj; MyDevType devtype; ...other fields... } MyDev; The property can then be registered as follows: static int mydev_prop_get_devtype(Object *obj, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); return dev->devtype; } static void mydev_prop_set_devtype(Object *obj, int value, Error **errp G_GNUC_UNUSED) { MyDev *dev = MYDEV(obj); dev->devtype = value; } object_property_add_enum(obj, "devtype", mydevtypemap, "MyDevType", mydev_prop_get_devtype, mydev_prop_set_devtype, NULL); Note there is no need to check the range of 'value' in the setter, because the string->enum conversion code will have already done that and reported an error as required. Signed-off-by: Daniel P. Berrange <berrange@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Signed-off-by: Andreas Färber <afaerber@suse.de>
2015-05-13 19:14:08 +03:00
g_test_add_func("/qom/proplist/badenum", test_dummy_badenum);
g_test_add_func("/qom/proplist/getenum", test_dummy_getenum);
g_test_add_func("/qom/proplist/iterator", test_dummy_iterator);
g_test_add_func("/qom/proplist/class_iterator", test_dummy_class_iterator);
g_test_add_func("/qom/proplist/delchild", test_dummy_delchild);
g_test_add_func("/qom/resolve/partial", test_qom_partial_path);
return g_test_run();
}