Skip to contentSkip to conversation

Remove some qdev_get_machine() calls from CONFIG_USER_ONLY

PATCH In review Eduardo Habkost 6 participants 14 messages Apr 25, 2019 – Nov 1, 2021

lore
Share & tools
lore
#0 lore raw

cover Remove some qdev_get_machine() calls from CONFIG_USER_ONLY

This series moves some qdev code outside qdev.o, so it can be compiled only in CONFIG_SOFTMMU.

The code being moved includes two qdev_get_machine() calls, so this will make it easier to move qdev_get_machine() to CONFIG_SOFTMMU later.

After this series, there's one remaining qdev_get_machine() call that seems more difficult to remove:

    static void device_set_realized(Object *obj, bool value, Error **errp)
    {
        /* [...] */
        if (!obj->parent) {
            gchar *name = g_strdup_printf("device[%d]", unattached_count++);

            object_property_add_child(container_get(qdev_get_machine(),
                                                    "/unattached"),
                                      name, obj, &error_abort);
            unattached_parent = true;
            g_free(name);
        }
        /* [...] */
    }

This one is tricky because on system emulation mode it needs "/machine" to already exist, but in user-only mode it needs to implicitly create a "/machine" container.

Eduardo Habkost (4):
machine: Move gpio code to hw/core/gpio.c
move qdev hotplug code to qdev-hotplug.c
qdev: Don't compile hotplug code in user-mode emulation
qdev-hotplug: Don't check type of qdev_get_machine()

 hw/core/bus.c                |  11 --
 hw/core/gpio.c               | 206 ++++++++++++++++++++++++++++++++
 hw/core/qdev-hotplug-stubs.c |  44 +++++++
 hw/core/qdev-hotplug.c       |  64 ++++++++++
 hw/core/qdev.c               | 219 -----------------------------------
 hw/core/Makefile.objs        |   5 +-
 tests/Makefile.include       |   3 +-
 7 files changed, 320 insertions(+), 232 deletions(-)
 create mode 100644 hw/core/gpio.c
 create mode 100644 hw/core/qdev-hotplug-stubs.c
 create mode 100644 hw/core/qdev-hotplug.c
Signature · 1 line
2.18.0.rc1.1.g3f1ff2140
#1 lore raw

PATCH 1/4 machine: Move gpio code to hw/core/gpio.c

Only softmmu code uses gpio, so move gpio code from qdev.c to gpio.c and compile it only on softmmu mode.

---
 hw/core/gpio.c        | 206 ++++++++++++++++++++++++++++++++++++++++++
 hw/core/qdev.c        | 184 -------------------------------------
 hw/core/Makefile.objs |   1 +
 3 files changed, 207 insertions(+), 184 deletions(-)
 create mode 100644 hw/core/gpio.c
hw/core/gpio.c new file +206 -0
new file mode 100644
index 0000000000..2724b4d7e7
@@ -0,0 +1,206 @@
+/*
+ * qdev GPIO helpers
+ *
+ *  Copyright (c) 2009 CodeSourcery
+ *
+ * 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 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/>.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/qdev.h"
+#include "qapi/error.h"
+
+static NamedGPIOList *qdev_get_named_gpio_list(DeviceState *dev,
+                                               const char *name)
+{
+    NamedGPIOList *ngl;
+
+    QLIST_FOREACH(ngl, &dev->gpios, node) {
+        /* NULL is a valid and matchable name, otherwise do a normal
+         * strcmp match.
+         */
+        if ((!ngl->name && !name) ||
+                (name && ngl->name && strcmp(name, ngl->name) == 0)) {
+            return ngl;
+        }
+    }
+
+    ngl = g_malloc0(sizeof(*ngl));
+    ngl->name = g_strdup(name);
+    QLIST_INSERT_HEAD(&dev->gpios, ngl, node);
+    return ngl;
+}
+
+void qdev_init_gpio_in_named_with_opaque(DeviceState *dev,
+                                         qemu_irq_handler handler,
+                                         void *opaque,
+                                         const char *name, int n)
+{
+    int i;
+    NamedGPIOList *gpio_list = qdev_get_named_gpio_list(dev, name);
+
+    assert(gpio_list->num_out == 0 || !name);
+    gpio_list->in = qemu_extend_irqs(gpio_list->in, gpio_list->num_in, handler,
+                                     opaque, n);
+
+    if (!name) {
+        name = "unnamed-gpio-in";
+    }
+    for (i = gpio_list->num_in; i < gpio_list->num_in + n; i++) {
+        gchar *propname = g_strdup_printf("%s[%u]", name, i);
+
+        object_property_add_child(OBJECT(dev), propname,
+                                  OBJECT(gpio_list->in[i]), &error_abort);
+        g_free(propname);
+    }
+
+    gpio_list->num_in += n;
+}
+
+void qdev_init_gpio_in(DeviceState *dev, qemu_irq_handler handler, int n)
+{
+    qdev_init_gpio_in_named(dev, handler, NULL, n);
+}
+
+void qdev_init_gpio_out_named(DeviceState *dev, qemu_irq *pins,
+                              const char *name, int n)
+{
+    int i;
+    NamedGPIOList *gpio_list = qdev_get_named_gpio_list(dev, name);
+
+    assert(gpio_list->num_in == 0 || !name);
+
+    if (!name) {
+        name = "unnamed-gpio-out";
+    }
+    memset(pins, 0, sizeof(*pins) * n);
+    for (i = 0; i < n; ++i) {
+        gchar *propname = g_strdup_printf("%s[%u]", name,
+                                          gpio_list->num_out + i);
+
+        object_property_add_link(OBJECT(dev), propname, TYPE_IRQ,
+                                 (Object **)&pins[i],
+                                 object_property_allow_set_link,
+                                 OBJ_PROP_LINK_STRONG,
+                                 &error_abort);
+        g_free(propname);
+    }
+    gpio_list->num_out += n;
+}
+
+void qdev_init_gpio_out(DeviceState *dev, qemu_irq *pins, int n)
+{
+    qdev_init_gpio_out_named(dev, pins, NULL, n);
+}
+
+qemu_irq qdev_get_gpio_in_named(DeviceState *dev, const char *name, int n)
+{
+    NamedGPIOList *gpio_list = qdev_get_named_gpio_list(dev, name);
+
+    assert(n >= 0 && n < gpio_list->num_in);
+    return gpio_list->in[n];
+}
+
+qemu_irq qdev_get_gpio_in(DeviceState *dev, int n)
+{
+    return qdev_get_gpio_in_named(dev, NULL, n);
+}
+
+void qdev_connect_gpio_out_named(DeviceState *dev, const char *name, int n,
+                                 qemu_irq pin)
+{
+    char *propname = g_strdup_printf("%s[%d]",
+                                     name ? name : "unnamed-gpio-out", n);
+    if (pin) {
+        /* We need a name for object_property_set_link to work.  If the
+         * object has a parent, object_property_add_child will come back
+         * with an error without doing anything.  If it has none, it will
+         * never fail.  So we can just call it with a NULL Error pointer.
+         */
+        object_property_add_child(container_get(qdev_get_machine(),
+                                                "/unattached"),
+                                  "non-qdev-gpio[*]", OBJECT(pin), NULL);
+    }
+    object_property_set_link(OBJECT(dev), OBJECT(pin), propname, &error_abort);
+    g_free(propname);
+}
+
+qemu_irq qdev_get_gpio_out_connector(DeviceState *dev, const char *name, int n)
+{
+    char *propname = g_strdup_printf("%s[%d]",
+                                     name ? name : "unnamed-gpio-out", n);
+
+    qemu_irq ret = (qemu_irq)object_property_get_link(OBJECT(dev), propname,
+                                                      NULL);
+
+    return ret;
+}
+
+/* disconnect a GPIO output, returning the disconnected input (if any) */
+
+static qemu_irq qdev_disconnect_gpio_out_named(DeviceState *dev,
+                                               const char *name, int n)
+{
+    char *propname = g_strdup_printf("%s[%d]",
+                                     name ? name : "unnamed-gpio-out", n);
+
+    qemu_irq ret = (qemu_irq)object_property_get_link(OBJECT(dev), propname,
+                                                      NULL);
+    if (ret) {
+        object_property_set_link(OBJECT(dev), NULL, propname, NULL);
+    }
+    g_free(propname);
+    return ret;
+}
+
+qemu_irq qdev_intercept_gpio_out(DeviceState *dev, qemu_irq icpt,
+                                 const char *name, int n)
+{
+    qemu_irq disconnected = qdev_disconnect_gpio_out_named(dev, name, n);
+    qdev_connect_gpio_out_named(dev, name, n, icpt);
+    return disconnected;
+}
+
+void qdev_connect_gpio_out(DeviceState * dev, int n, qemu_irq pin)
+{
+    qdev_connect_gpio_out_named(dev, NULL, n, pin);
+}
+
+void qdev_pass_gpios(DeviceState *dev, DeviceState *container,
+                     const char *name)
+{
+    int i;
+    NamedGPIOList *ngl = qdev_get_named_gpio_list(dev, name);
+
+    for (i = 0; i < ngl->num_in; i++) {
+        const char *nm = ngl->name ? ngl->name : "unnamed-gpio-in";
+        char *propname = g_strdup_printf("%s[%d]", nm, i);
+
+        object_property_add_alias(OBJECT(container), propname,
+                                  OBJECT(dev), propname,
+                                  &error_abort);
+        g_free(propname);
+    }
+    for (i = 0; i < ngl->num_out; i++) {
+        const char *nm = ngl->name ? ngl->name : "unnamed-gpio-out";
+        char *propname = g_strdup_printf("%s[%d]", nm, i);
+
+        object_property_add_alias(OBJECT(container), propname,
+                                  OBJECT(dev), propname,
+                                  &error_abort);
+        g_free(propname);
+    }
+    QLIST_REMOVE(ngl, node);
+    QLIST_INSERT_HEAD(&container->gpios, ngl, node);
+}
hw/core/qdev.c +0 -184

190 lines not on this page · raw message

hw/core/Makefile.objs +1 -0
index a799c83815..6789154807 100644
@@ -13,6 +13,7 @@ common-obj-$(CONFIG_XILINX_AXI) += stream.o
 common-obj-$(CONFIG_PTIMER) += ptimer.o
 common-obj-$(CONFIG_SOFTMMU) += sysbus.o
 common-obj-$(CONFIG_SOFTMMU) += machine.o
+common-obj-$(CONFIG_SOFTMMU) += gpio.o
 common-obj-$(CONFIG_SOFTMMU) += loader.o
 common-obj-$(CONFIG_FITLOADER) += loader-fit.o
 common-obj-$(CONFIG_SOFTMMU) += qdev-properties-system.o
Signature · 1 line
2.18.0.rc1.1.g3f1ff2140
Reviewed-by Tested-by on v1 1/4
#2 lore raw

Re: [Qemu-devel] [PATCH 1/4] machine: Move gpio code to hw/core/gpio.c

Quoted 10 lines · Eduardo Habkost

Only softmmu code uses gpio, so move gpio code from qdev.c to gpio.c and compile it only on softmmu mode.

---
 hw/core/gpio.c        | 206 ++++++++++++++++++++++++++++++++++++++++++
 hw/core/qdev.c        | 184 -------------------------------------
 hw/core/Makefile.objs |   1 +
 3 files changed, 207 insertions(+), 184 deletions(-)
 create mode 100644 hw/core/gpio.c
#3 lore raw

PATCH 2/4 move qdev hotplug code to qdev-hotplug.c

In reply to Eduardo Habkost

The qdev hotplug code is used only in softmmu mode, so move it to a separate file so we can eventually avoid compiling it in user-only mode.

---
 hw/core/bus.c          | 11 -------
 hw/core/qdev-hotplug.c | 69 ++++++++++++++++++++++++++++++++++++++++++
 hw/core/qdev.c         | 35 ---------------------
 hw/core/Makefile.objs  |  2 +-
 4 files changed, 70 insertions(+), 47 deletions(-)
 create mode 100644 hw/core/qdev-hotplug.c
hw/core/bus.c +0 -11
index e09843f6ab..35e042416c 100644
@@ -22,17 +22,6 @@
 #include "hw/qdev.h"
 #include "qapi/error.h"
 
-void qbus_set_hotplug_handler(BusState *bus, Object *handler, Error **errp)
-{
-    object_property_set_link(OBJECT(bus), OBJECT(handler),
-                             QDEV_HOTPLUG_HANDLER_PROPERTY, errp);
-}
-
-void qbus_set_bus_hotplug_handler(BusState *bus, Error **errp)
-{
-    qbus_set_hotplug_handler(bus, OBJECT(bus), errp);
-}
-
 int qbus_walk_children(BusState *bus,
                        qdev_walkerfn *pre_devfn, qbus_walkerfn *pre_busfn,
                        qdev_walkerfn *post_devfn, qbus_walkerfn *post_busfn,
hw/core/qdev-hotplug.c new file +69 -0

69 lines not on this page · raw message

hw/core/qdev.c +0 -35
index f73e7ded1a..3015da0ac9 100644
@@ -219,41 +219,6 @@ void qdev_set_legacy_instance_id(DeviceState *dev, int alias_id,
     dev->alias_required_for_version = required_for_version;
 }
 
-HotplugHandler *qdev_get_machine_hotplug_handler(DeviceState *dev)
-{
-    MachineState *machine;
-    MachineClass *mc;
-    Object *m_obj = qdev_get_machine();
-
-    if (object_dynamic_cast(m_obj, TYPE_MACHINE)) {
-        machine = MACHINE(m_obj);
-        mc = MACHINE_GET_CLASS(machine);
-        if (mc->get_hotplug_handler) {
-            return mc->get_hotplug_handler(machine, dev);
-        }
-    }
-
-    return NULL;
-}
-
-HotplugHandler *qdev_get_bus_hotplug_handler(DeviceState *dev)
-{
-    if (dev->parent_bus) {
-        return dev->parent_bus->hotplug_handler;
-    }
-    return NULL;
-}
-
-HotplugHandler *qdev_get_hotplug_handler(DeviceState *dev)
-{
-    HotplugHandler *hotplug_ctrl = qdev_get_machine_hotplug_handler(dev);
-
-    if (hotplug_ctrl == NULL && dev->parent_bus) {
-        hotplug_ctrl = qdev_get_bus_hotplug_handler(dev);
-    }
-    return hotplug_ctrl;
-}
-
 static int qdev_reset_one(DeviceState *dev, void *opaque)
 {
     device_reset(dev);
hw/core/Makefile.objs +1 -1
index 6789154807..9c4f953716 100644
@@ -1,6 +1,6 @@
 # core qdev-related obj files, also used by *-user:
 common-obj-y += qdev.o qdev-properties.o
-common-obj-y += bus.o reset.o
+common-obj-y += bus.o reset.o qdev-hotplug.o
 common-obj-$(CONFIG_SOFTMMU) += qdev-fw.o
 common-obj-$(CONFIG_SOFTMMU) += fw-path-provider.o
 # irq.o needed for qdev GPIO handling:
Signature · 1 line
2.18.0.rc1.1.g3f1ff2140
#4 lore raw

PATCH 3/4 qdev: Don't compile hotplug code in user-mode emulation

In reply to Eduardo Habkost

Move qdev-hotplug.o and hotplug.o to CONFIG_SOFTMMU because the code is not necessary for *-user.

Provide a few stubs in qdev-hotplug-stubs.c to replace initialization functions that are called from qdev.c.

---
 hw/core/qdev-hotplug-stubs.c | 44 ++++++++++++++++++++++++++++++++++++
 hw/core/Makefile.objs        |  6 +++--
 tests/Makefile.include       |  3 ++-
 3 files changed, 50 insertions(+), 3 deletions(-)
 create mode 100644 hw/core/qdev-hotplug-stubs.c
hw/core/qdev-hotplug-stubs.c new file +44 -0
new file mode 100644
index 0000000000..c710f23388
@@ -0,0 +1,44 @@
+/*
+ * qdev hotplug handler stubs (for user-mode emulation and unit tests)
+ *
+ *  Copyright (c) 2019 Red Hat Inc
+ *
+ * Authors:
+ *  Eduardo Habkost <[email protected]>
+ *
+ * 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 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/>.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/qdev-core.h"
+#include "hw/hotplug.h"
+
+HotplugHandler *qdev_get_hotplug_handler(DeviceState *dev)
+{
+    return NULL;
+}
+
+void hotplug_handler_pre_plug(HotplugHandler *plug_handler,
+                              DeviceState *plugged_dev,
+                              Error **errp)
+{
+    assert(plug_handler == NULL);
+}
+
+void hotplug_handler_plug(HotplugHandler *plug_handler,
+                          DeviceState *plugged_dev,
+                          Error **errp)
+{
+    assert(plug_handler == NULL);
+}
hw/core/Makefile.objs +4 -2
index 9c4f953716..284fc306c5 100644
@@ -1,11 +1,10 @@
 # core qdev-related obj files, also used by *-user:
 common-obj-y += qdev.o qdev-properties.o
-common-obj-y += bus.o reset.o qdev-hotplug.o
+common-obj-y += bus.o reset.o
 common-obj-$(CONFIG_SOFTMMU) += qdev-fw.o
 common-obj-$(CONFIG_SOFTMMU) += fw-path-provider.o
 # irq.o needed for qdev GPIO handling:
 common-obj-y += irq.o
-common-obj-y += hotplug.o
 common-obj-$(CONFIG_SOFTMMU) += nmi.o
 
 common-obj-$(CONFIG_EMPTY_SLOT) += empty_slot.o
@@ -13,6 +12,7 @@ common-obj-$(CONFIG_XILINX_AXI) += stream.o
 common-obj-$(CONFIG_PTIMER) += ptimer.o
 common-obj-$(CONFIG_SOFTMMU) += sysbus.o
 common-obj-$(CONFIG_SOFTMMU) += machine.o
+common-obj-$(CONFIG_SOFTMMU) += hotplug.o qdev-hotplug.o
 common-obj-$(CONFIG_SOFTMMU) += gpio.o
 common-obj-$(CONFIG_SOFTMMU) += loader.o
 common-obj-$(CONFIG_FITLOADER) += loader-fit.o
@@ -23,3 +23,5 @@ common-obj-$(CONFIG_SOFTMMU) += split-irq.o
 common-obj-$(CONFIG_PLATFORM_BUS) += platform-bus.o
 common-obj-$(CONFIG_SOFTMMU) += generic-loader.o
 common-obj-$(CONFIG_SOFTMMU) += null-machine.o
+
+common-obj-$(CONFIG_USER_ONLY) += qdev-hotplug-stubs.o
tests/Makefile.include +2 -1
index e2432d5e77..fd5e7e0c77 100644
@@ -568,7 +568,8 @@ tests/fp/%:
 	$(MAKE) -C $(dir $@) $(notdir $@)
 
 tests/test-qdev-global-props$(EXESUF): tests/test-qdev-global-props.o \
-	hw/core/qdev.o hw/core/qdev-properties.o hw/core/hotplug.o\
+	hw/core/qdev.o hw/core/qdev-properties.o \
+	hw/core/qdev-hotplug-stubs.o \
 	hw/core/bus.o \
 	hw/core/irq.o \
 	hw/core/fw-path-provider.o \
Signature · 1 line
2.18.0.rc1.1.g3f1ff2140
#5 lore raw

Re: [Qemu-devel] [PATCH 3/4] qdev: Don't compile hotplug code in user-mode emulation

hw/core/qdev-hotplug-stubs.c line 2

Quoted 8 lines · Eduardo Habkost
hw/core/qdev-hotplug-stubs.c new file +2 -0
new file mode 100644
index 0000000000..c710f23388
@@ -0,0 +1,44 @@
+/*
+ * qdev hotplug handler stubs (for user-mode emulation and unit tests)

Can you explain the issue with unit tests in the commit message?

Quoted 31 lines
hw/core/qdev-hotplug-stubs.c new file +31 -0
+ *  Copyright (c) 2019 Red Hat Inc
+ *
+ * Authors:
+ *  Eduardo Habkost <[email protected]>
+ *
+ * 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 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/>.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/qdev-core.h"
+#include "hw/hotplug.h"
+
+HotplugHandler *qdev_get_hotplug_handler(DeviceState *dev)
+{
+    return NULL;
+}
+
+void hotplug_handler_pre_plug(HotplugHandler *plug_handler,
+                              DeviceState *plugged_dev,
+                              Error **errp)

hw/core/qdev-hotplug-stubs.c line 44

Quoted 10 lines
hw/core/qdev-hotplug-stubs.c new file +10 -0
+{
+    assert(plug_handler == NULL);
+}
+
+void hotplug_handler_plug(HotplugHandler *plug_handler,
+                          DeviceState *plugged_dev,
+                          Error **errp)
+{
+    assert(plug_handler == NULL);
+}

Would it work if you instead make these functions (and the others in hw/core/hotplug.c) inlines in include/hw/hotplug.h?

Then all that remains is qdev_get_hotplug_handler.

Paolo

#6 lore raw

Re: [Qemu-devel] [PATCH 3/4] qdev: Don't compile hotplug code in user-mode emulation

Quoted 11 lines · Paolo Bonzini

hw/core/qdev-hotplug-stubs.c line 2

Quoted 8 lines · Eduardo Habkost
hw/core/qdev-hotplug-stubs.c new file +2 -0
new file mode 100644
index 0000000000..c710f23388
@@ -0,0 +1,44 @@
+/*
+ * qdev hotplug handler stubs (for user-mode emulation and unit tests)

Can you explain the issue with unit tests in the commit message?

I don't think there are issues with unit tests. I just thought it was pointless to link qdev-hotplug.o into unit tests binaries if no hotplug handler object is created by them.

What do you think? Should I keep qdev-hotplug.o and hotplug.o in the unit tests binaries?

Quoted 46 lines
Quoted 32 lines
hw/core/qdev-hotplug-stubs.c new file +31 -0
+ *  Copyright (c) 2019 Red Hat Inc
+ *
+ * Authors:
+ *  Eduardo Habkost <[email protected]>
+ *
+ * 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 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/>.
+ */
+
+#include "qemu/osdep.h"
+#include "hw/qdev-core.h"
+#include "hw/hotplug.h"
+
+HotplugHandler *qdev_get_hotplug_handler(DeviceState *dev)
+{
+    return NULL;
+}
+
+void hotplug_handler_pre_plug(HotplugHandler *plug_handler,
+                              DeviceState *plugged_dev,
+                              Error **errp)

hw/core/qdev-hotplug-stubs.c line 44

Quoted 10 lines
hw/core/qdev-hotplug-stubs.c new file +10 -0
+{
+    assert(plug_handler == NULL);
+}
+
+void hotplug_handler_plug(HotplugHandler *plug_handler,
+                          DeviceState *plugged_dev,
+                          Error **errp)
+{
+    assert(plug_handler == NULL);
+}

Would it work if you instead make these functions (and the others in hw/core/hotplug.c) inlines in include/hw/hotplug.h?

Quoted 1 line

Then all that remains is qdev_get_hotplug_handler.

I think that would work. I'll do that in v2, thanks!

Signature · 1 line
Eduardo
#7 lore raw

PATCH 4/4 qdev-hotplug: Don't check type of qdev_get_machine()

In reply to Eduardo Habkost

Now that qdev_get_machine_hotplug_handler() is only compiled in softmmu mode, we don't need to check if qdev_gt_machine() is really of type TYPE_MACHINE.

---
 hw/core/qdev-hotplug.c | 13 ++++---------
 1 file changed, 4 insertions(+), 9 deletions(-)
hw/core/qdev-hotplug.c +4 -9
index 8ab31043a7..017500c1e1 100644
@@ -35,16 +35,11 @@ void qbus_set_bus_hotplug_handler(BusState *bus, Error **errp)
 
 HotplugHandler *qdev_get_machine_hotplug_handler(DeviceState *dev)
 {
-    MachineState *machine;
-    MachineClass *mc;
-    Object *m_obj = qdev_get_machine();
+    MachineState *machine = MACHINE(qdev_get_machine());
+    MachineClass *mc = MACHINE_GET_CLASS(machine);
 
-    if (object_dynamic_cast(m_obj, TYPE_MACHINE)) {
-        machine = MACHINE(m_obj);
-        mc = MACHINE_GET_CLASS(machine);
-        if (mc->get_hotplug_handler) {
-            return mc->get_hotplug_handler(machine, dev);
-        }
+    if (mc->get_hotplug_handler) {
+        return mc->get_hotplug_handler(machine, dev);
     }
 
     return NULL;
Signature · 1 line
2.18.0.rc1.1.g3f1ff2140
#8 lore raw

Re: [Qemu-devel] [PATCH 4/4] qdev-hotplug: Don't check type of qdev_get_machine()

Quoted 25 lines · Eduardo Habkost

Now that qdev_get_machine_hotplug_handler() is only compiled in softmmu mode, we don't need to check if qdev_gt_machine() is really of type TYPE_MACHINE.

---
 hw/core/qdev-hotplug.c | 13 ++++---------
 1 file changed, 4 insertions(+), 9 deletions(-)
hw/core/qdev-hotplug.c +2 -4
index 8ab31043a7..017500c1e1 100644
@@ -35,16 +35,11 @@ void qbus_set_bus_hotplug_handler(BusState *bus, Error **errp)
 
 HotplugHandler *qdev_get_machine_hotplug_handler(DeviceState *dev)
 {
-    MachineState *machine;
-    MachineClass *mc;
-    Object *m_obj = qdev_get_machine();
+    MachineState *machine = MACHINE(qdev_get_machine());
+    MachineClass *mc = MACHINE_GET_CLASS(machine);
 
-    if (object_dynamic_cast(m_obj, TYPE_MACHINE)) {

hw/core/qdev-hotplug.c void qbus_set_bus_hotplug_handler(BusState *bus, Error **errp) · line 45

Quoted 10 lines
hw/core/qdev-hotplug.c +2 -5
-        machine = MACHINE(m_obj);
-        mc = MACHINE_GET_CLASS(machine);
-        if (mc->get_hotplug_handler) {
-            return mc->get_hotplug_handler(machine, dev);
-        }
+    if (mc->get_hotplug_handler) {
+        return mc->get_hotplug_handler(machine, dev);
     }
 
     return NULL;

I rebased this patch and tried it, but it fails 'check-unit':

Running test test-qdev-global-props
** ERROR:../tests/unit/test-qdev-global-props.c:89:test_static_prop: child process (/qdev/properties/static/default/subprocess [12678]) failed unexpectedly
ERROR test-qdev-global-props - Bail out! ERROR:../tests/unit/test-qdev-global-props.c:89:test_static_prop: child process (/qdev/properties/static/default/subprocess [12678]) failed unexpectedly
make: *** [Makefile.mtest:640: run-test-78] Error 1

Here the object is of type "container":

  Object *object_get_root(void)
  {
      static Object *root;
      if (!root) {
          root = object_new("container");
      }

      return root;

}

#9 lore raw
In reply to Eduardo Habkost
Patchew URL: https://patchew.org/QEMU/[email protected]/

Patchew URL: https://patchew.org/QEMU/[email protected]/

Hi,

This series seems to have some coding style problems. See output below for more information:

Type: series
Message-id: [email protected]
Subject: [Qemu-devel] [PATCH 0/4] Remove some qdev_get_machine() calls from CONFIG_USER_ONLY

=== TEST SCRIPT BEGIN ===
#!/bin/bash
git rev-parse base > /dev/null || exit 0
git config --local diff.renamelimit 0
git config --local diff.renames True
git config --local diff.algorithm histogram
./scripts/checkpatch.pl --mailback base..
=== TEST SCRIPT END ===

Updating 3c8cf5a9c21ff8782164d1def7f44bd888713384
From https://github.com/patchew-project/qemu
t [tag update] patchew/[email protected] -> patchew/[email protected]
* [new tag] patchew/[email protected] -> patchew/[email protected]
Switched to a new branch 'test'
c5c77ea2cc qdev-hotplug: Don't check type of qdev_get_machine() 854ded711d qdev: Don't compile hotplug code in user-mode emulation 1bd33307be move qdev hotplug code to qdev-hotplug.c 02b1f5db6f machine: Move gpio code to hw/core/gpio.c

=== OUTPUT BEGIN ===
1/4 Checking commit 02b1f5db6fcd (machine: Move gpio code to hw/core/gpio.c)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #26:
new file mode 100644

WARNING: Block comments use a leading /* on a separate line #60: FILE: hw/core/gpio.c:30:
+ /* NULL is a valid and matchable name, otherwise do a normal

WARNING: Block comments use a leading /* on a separate line #156: FILE: hw/core/gpio.c:126:
+ /* We need a name for object_property_set_link to work. If the

ERROR: "foo * bar" should be "foo *bar"
#205: FILE: hw/core/gpio.c:175:
+void qdev_connect_gpio_out(DeviceState * dev, int n, qemu_irq pin)

total: 1 errors, 3 warnings, 403 lines checked

Patch 1/4 has style problems, please review. If any of these errors are false positives report them to the maintainer, see CHECKPATCH in MAINTAINERS.

2/4 Checking commit 1bd33307be11 (move qdev hotplug code to qdev-hotplug.c)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #49:
new file mode 100644

total: 0 errors, 1 warnings, 134 lines checked

Patch 2/4 has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
3/4 Checking commit 854ded711daa (qdev: Don't compile hotplug code in user-mode emulation)
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating? #48:
new file mode 100644

total: 0 errors, 1 warnings, 77 lines checked

Patch 3/4 has style problems, please review. If any of these errors
are false positives report them to the maintainer, see
CHECKPATCH in MAINTAINERS.
4/4 Checking commit c5c77ea2ccde (qdev-hotplug: Don't check type of qdev_get_machine()) === OUTPUT END ===

Test command exited with code: 1

The full log is available at
http://patchew.org/logs/[email protected]/testing.checkpatch/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]

#10 lore raw
In reply to Eduardo Habkost
Quoted 10 lines · Eduardo Habkost

This series moves some qdev code outside qdev.o, so it can be compiled only in CONFIG_SOFTMMU.

The code being moved includes two qdev_get_machine() calls, so this will make it easier to move qdev_get_machine() to CONFIG_SOFTMMU later.

After this series, there's one remaining qdev_get_machine() call that seems more difficult to remove:

Quoted 15 lines
     static void device_set_realized(Object *obj, bool value, Error **errp)
     {
         /* [...] */
         if (!obj->parent) {
             gchar *name = g_strdup_printf("device[%d]", unattached_count++);

             object_property_add_child(container_get(qdev_get_machine(),
                                                     "/unattached"),
                                       name, obj, &error_abort);
             unattached_parent = true;
             g_free(name);
         }
         /* [...] */
     }

I may have an experimental patch to fix device_set_realized issue:

1. in qdev_get_machine():
replace

	dev = container_get(object_get_root(), "/machine");
with
	dev = object_resolve_path("/machine", NULL);

2. in device_set_realized():

Using

	Object *container = qdev_get_machine() ?
		qdev_get_machine() : object_get_root();
and pass it to
	object_property_add_child(
		container_get(container, "/unattached"),
		name, obj, &error_abort);

With this fix, we could say the qdev_get_machine() does return the "/machine" object (or null) not a confused "/container".

We could continue to use qdev_get_machine() in system emulation mode, getting rid of its surprising side effect as Markus said.

The return value of qdev_get_machine() in user-only mode is the same object returned by object_get_root(), so no semantic changes.

Quoted 22 lines

This one is tricky because on system emulation mode it needs "/machine" to already exist, but in user-only mode it needs to implicitly create a "/machine" container.

Eduardo Habkost (4):
machine: Move gpio code to hw/core/gpio.c
move qdev hotplug code to qdev-hotplug.c
qdev: Don't compile hotplug code in user-mode emulation
qdev-hotplug: Don't check type of qdev_get_machine()

  hw/core/bus.c                |  11 --
  hw/core/gpio.c               | 206 ++++++++++++++++++++++++++++++++
  hw/core/qdev-hotplug-stubs.c |  44 +++++++
  hw/core/qdev-hotplug.c       |  64 ++++++++++
  hw/core/qdev.c               | 219 -----------------------------------
  hw/core/Makefile.objs        |   5 +-
  tests/Makefile.include       |   3 +-
  7 files changed, 320 insertions(+), 232 deletions(-)

create mode 100644 hw/core/gpio.c
create mode 100644 hw/core/qdev-hotplug-stubs.c
create mode 100644 hw/core/qdev-hotplug.c

#11 lore raw
Quoted 37 lines · Like Xu
Quoted 10 lines · Eduardo Habkost

This series moves some qdev code outside qdev.o, so it can be compiled only in CONFIG_SOFTMMU.

The code being moved includes two qdev_get_machine() calls, so this will make it easier to move qdev_get_machine() to CONFIG_SOFTMMU later.

After this series, there's one remaining qdev_get_machine() call that seems more difficult to remove:

Quoted 15 lines
     static void device_set_realized(Object *obj, bool value, Error **errp)
     {
         /* [...] */
         if (!obj->parent) {
             gchar *name = g_strdup_printf("device[%d]", unattached_count++);

             object_property_add_child(container_get(qdev_get_machine(),
                                                     "/unattached"),
                                       name, obj, &error_abort);
             unattached_parent = true;
             g_free(name);
         }
         /* [...] */
     }

I may have an experimental patch to fix device_set_realized issue:

1. in qdev_get_machine():
replace

	dev = container_get(object_get_root(), "/machine");
with
	dev = object_resolve_path("/machine", NULL);

2. in device_set_realized():

Quoted 7 lines

Using

	Object *container = qdev_get_machine() ?
		qdev_get_machine() : object_get_root();
and pass it to
	object_property_add_child(
		container_get(container, "/unattached"),
		name, obj, &error_abort);

I wouldn't like to call qdev_get_machine() here (see below[1]), but trying "/machine" first and falling back to object_get_root() sounds good, for two reasons:

* It won't require "/machine" to exist at all, on *-user;
* It will allow machines to create unattached devices on instance_init without crashing (which is not possible today).

Quoted 10 lines

With this fix, we could say the qdev_get_machine() does return the "/machine" object (or null) not a confused "/container".

We could continue to use qdev_get_machine() in system emulation mode, getting rid of its surprising side effect as Markus said.

The return value of qdev_get_machine() in user-only mode is the same object returned by object_get_root(), so no semantic changes.

[1] I wouldn't like to have a different qdev_get_machine() function in user-only mode. I would like to simply not call qdev_get_machine() at all unless we're in system emulation mode.

Signature · 1 line
Eduardo
#12 lore raw
In reply to Eduardo Habkost
Quoted 25 lines · Eduardo Habkost

This series moves some qdev code outside qdev.o, so it can be compiled only in CONFIG_SOFTMMU.

The code being moved includes two qdev_get_machine() calls, so this will make it easier to move qdev_get_machine() to CONFIG_SOFTMMU later.

After this series, there's one remaining qdev_get_machine() call that seems more difficult to remove:

    static void device_set_realized(Object *obj, bool value, Error **errp)
    {
        /* [...] */
        if (!obj->parent) {
            gchar *name = g_strdup_printf("device[%d]", unattached_count++);

            object_property_add_child(container_get(qdev_get_machine(),
                                                    "/unattached"),
                                      name, obj, &error_abort);
            unattached_parent = true;
            g_free(name);
        }
        /* [...] */
    }
Quoted 3 lines

This one is tricky because on system emulation mode it needs "/machine" to already exist, but in user-only mode it needs to implicitly create a "/machine" container.

The patches look good to me on a glance. Looking forward to v2.

#13 lore raw
In reply to Eduardo Habkost
Quoted 5 lines · Eduardo Habkost

Eduardo Habkost (4):
machine: Move gpio code to hw/core/gpio.c
move qdev hotplug code to qdev-hotplug.c
qdev: Don't compile hotplug code in user-mode emulation
qdev-hotplug: Don't check type of qdev_get_machine()

Patch 1 queued to machine-next.