[Qemu-devel] [PATCH for-3.2 v10 0/3] wakeup-from-suspend and system_wakeup changes

[Qemu-devel] [PATCH for-3.2 v10 0/3] wakeup-from-suspend and system_wakeup changes

[Daniel Henrique Barboza]

Changes in v10:
- added qemu_register_wakeup_support call in hw/i386/xen/xen-hvm.c
- explained the effects of not checking the query-current-machine
before issuing guest-suspend-ram/hybrid in QGA documentation
- changed qemu_system_wakeup_request to allow the runstate check
error to be propagated to callers
- previous series link:
https://lists.gnu.org/archive/html/qemu-devel/2018-09/msg02171.html


Daniel Henrique Barboza (3):
  qmp: query-current-machine with wakeup-suspend-support
  qga: update guest-suspend-ram and guest-suspend-hybrid descriptions
  qmp hmp: Make system_wakeup check wake-up support and run state

 hmp.c                   |  5 ++++-
 hw/acpi/core.c          | 10 +++++++++-
 hw/char/serial.c        |  3 ++-
 hw/i386/xen/xen-hvm.c   |  5 +++++
 hw/input/ps2.c          |  9 ++++++---
 hw/timer/mc146818rtc.c  |  3 ++-
 include/sysemu/sysemu.h |  4 +++-
 migration/migration.c   |  7 +++++--
 qapi/misc.json          | 32 +++++++++++++++++++++++++++++++-
 qga/qapi-schema.json    | 16 ++++++++++------
 qmp.c                   | 13 ++++++++++++-
 vl.c                    | 23 ++++++++++++++++++++++-
 12 files changed, 111 insertions(+), 19 deletions(-)

-- 
2.19.1

[Qemu-devel] [PATCH for-3.2 v10 1/3] qmp: query-current-machine with wakeup-suspend-support

[Daniel Henrique Barboza]

When issuing the qmp/hmp 'system_wakeup' command, what happens in a
nutshell is:

- qmp_system_wakeup_request set runstate to RUNNING, sets a wakeup_reason
and notify the event
- in the main_loop, all vcpus are paused, a system reset is issued, all
subscribers of wakeup_notifiers receives a notification, vcpus are then
resumed and the wake up QAPI event is fired

Note that this procedure alone doesn't ensure that the guest will awake
from SUSPENDED state - the subscribers of the wake up event must take
action to resume the guest, otherwise the guest will simply reboot. At
this moment, only the ACPI machines via acpi_pm1_cnt_init has wake-up
from suspend support.

However, only the presence of 'system_wakeup' is required for QGA to
support 'guest-suspend-ram' and 'guest-suspend-hybrid' at this moment.
This means that the user/management will expect to suspend the guest using
one of those suspend commands and then resume execution using system_wakeup,
regardless of the support offered in system_wakeup in the first place.

This patch creates a new API called query-current-machine [1], that holds
a new flag called 'wakeup-suspend-support' that indicates if the guest
supports wake up from suspend via system_wakeup. The machine is considered
to implement wake-up support if a call to a new 'qemu_register_wakeup_support'
is made during its init, as it is now being done inside acpi_pm1_cnt_init.
This allows for any other machine type to declare wake-up support regardless
of ACPI state or wakeup_notifiers subscription, making easier for
newer implementations that might have its own mechanisms in the future.

This is the expected output of query-current-machine when running a x86
guest:

{"execute" : "query-current-machine"}
{"return": {"wakeup-suspend-support": true}}

Running the same x86 guest, but with the --no-acpi option:

{"execute" : "query-current-machine"}
{"return": {"wakeup-suspend-support": false}}

This is the output when running a pseries guest:

{"execute" : "query-current-machine"}
{"return": {"wakeup-suspend-support": false}}

With this extra tool, management can avoid situations where a guest
that does not have proper suspend/wake capabilities ends up in
inconsistent state (e.g.
https://github.com/open-power-host-os/qemu/issues/31).

[1] the decision of creating the query-current-machine API is based
on discussions in the QEMU mailing list where it was decided that
query-target wasn't a proper place to store the wake-up flag, neither
was query-machines because this isn't a static property of the
machine object. This new API can then be used to store other
dynamic machine properties that are scattered around the code
ATM. More info at:
https://lists.gnu.org/archive/html/qemu-devel/2018-05/msg04235.html

Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
---
 hw/acpi/core.c          |  6 ++++++
 hw/i386/xen/xen-hvm.c   |  5 +++++
 include/sysemu/sysemu.h |  1 +
 qapi/misc.json          | 24 ++++++++++++++++++++++++
 vl.c                    | 19 +++++++++++++++++++
 5 files changed, 55 insertions(+)

diff --git a/hw/acpi/core.c b/hw/acpi/core.c
index aafdc61648..52e18d7810 100644
--- a/hw/acpi/core.c
+++ b/hw/acpi/core.c
@@ -617,6 +617,12 @@ void acpi_pm1_cnt_init(ACPIREGS *ar, MemoryRegion *parent,
     ar->pm1.cnt.s4_val = s4_val;
     ar->wakeup.notify = acpi_notify_wakeup;
     qemu_register_wakeup_notifier(&ar->wakeup);
+
+    /*
+     * Register wake-up support in QMP query-current-machine API
+     */
+    qemu_register_wakeup_support();
+
     memory_region_init_io(&ar->pm1.cnt.io, memory_region_owner(parent),
                           &acpi_pm_cnt_ops, ar, "acpi-cnt", 2);
     memory_region_add_subregion(parent, 4, &ar->pm1.cnt.io);
diff --git a/hw/i386/xen/xen-hvm.c b/hw/i386/xen/xen-hvm.c
index 935a3676c8..2143d33b18 100644
--- a/hw/i386/xen/xen-hvm.c
+++ b/hw/i386/xen/xen-hvm.c
@@ -1405,6 +1405,11 @@ void xen_hvm_init(PCMachineState *pcms, MemoryRegion **ram_memory)
     state->wakeup.notify = xen_wakeup_notifier;
     qemu_register_wakeup_notifier(&state->wakeup);
 
+    /*
+     * Register wake-up support in QMP query-current-machine API
+     */
+    qemu_register_wakeup_support();
+
     rc = xen_map_ioreq_server(state);
     if (rc < 0) {
         goto err;
diff --git a/include/sysemu/sysemu.h b/include/sysemu/sysemu.h
index 8d6095d98b..0446adacc6 100644
--- a/include/sysemu/sysemu.h
+++ b/include/sysemu/sysemu.h
@@ -77,6 +77,7 @@ void qemu_register_suspend_notifier(Notifier *notifier);
 void qemu_system_wakeup_request(WakeupReason reason);
 void qemu_system_wakeup_enable(WakeupReason reason, bool enabled);
 void qemu_register_wakeup_notifier(Notifier *notifier);
+void qemu_register_wakeup_support(void);
 void qemu_system_shutdown_request(ShutdownCause reason);
 void qemu_system_powerdown_request(void);
 void qemu_register_powerdown_notifier(Notifier *notifier);
diff --git a/qapi/misc.json b/qapi/misc.json
index 6c1c5c0a37..9b86576e61 100644
--- a/qapi/misc.json
+++ b/qapi/misc.json
@@ -2008,6 +2008,30 @@
 ##
 { 'command': 'query-machines', 'returns': ['MachineInfo'] }
 
+##
+# @CurrentMachineParams:
+#
+# Information describing the running machine parameters.
+#
+# @wakeup-suspend-support: true if the machine supports wake up from
+#                          suspend
+#
+# Since: 3.2.0
+##
+{ 'struct': 'CurrentMachineParams',
+  'data': { 'wakeup-suspend-support': 'bool'} }
+
+##
+# @query-current-machine:
+#
+# Return information on the current virtual machine.
+#
+# Returns: CurrentMachineParams
+#
+# Since: 3.2.0
+##
+{ 'command': 'query-current-machine', 'returns': 'CurrentMachineParams' }
+
 ##
 # @CpuDefinitionInfo:
 #
diff --git a/vl.c b/vl.c
index 55bab005b6..5813a8785f 100644
--- a/vl.c
+++ b/vl.c
@@ -191,6 +191,7 @@ bool boot_strict;
 uint8_t *boot_splash_filedata;
 size_t boot_splash_filedata_size;
 uint8_t qemu_extra_params_fw[2];
+bool wakeup_suspend_enabled;
 
 int icount_align_option;
 
@@ -1782,6 +1783,24 @@ void qemu_register_wakeup_notifier(Notifier *notifier)
     notifier_list_add(&wakeup_notifiers, notifier);
 }
 
+void qemu_register_wakeup_support(void)
+{
+    wakeup_suspend_enabled = true;
+}
+
+static bool qemu_wakeup_suspend_enabled(void)
+{
+    return wakeup_suspend_enabled;
+}
+
+CurrentMachineParams *qmp_query_current_machine(Error **errp)
+{
+    CurrentMachineParams *params = g_malloc0(sizeof(*params));
+    params->wakeup_suspend_support = qemu_wakeup_suspend_enabled();
+
+    return params;
+}
+
 void qemu_system_killed(int signal, pid_t pid)
 {
     shutdown_signal = signal;
-- 
2.19.1

[Qemu-devel] [PATCH for-3.2 v10 2/3] qga: update guest-suspend-ram and guest-suspend-hybrid descriptions

[Daniel Henrique Barboza]

This patch updates the descriptions of 'guest-suspend-ram' and
'guest-suspend-hybrid' to mention that both commands relies now
on the proper support for wake up from suspend, retrieved by the
'wakeup-suspend-support' attribute of the 'query-current-machine'
QMP command.

Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
---
 qga/qapi-schema.json | 16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/qga/qapi-schema.json b/qga/qapi-schema.json
index c6725b3ec8..78a439263e 100644
--- a/qga/qapi-schema.json
+++ b/qga/qapi-schema.json
@@ -567,9 +567,11 @@
 # For the best results it's strongly recommended to have the pm-utils
 # package installed in the guest.
 #
-# IMPORTANT: guest-suspend-ram requires QEMU to support the 'system_wakeup'
-# command.  Thus, it's *required* to query QEMU for the presence of the
-# 'system_wakeup' command before issuing guest-suspend-ram.
+# IMPORTANT: guest-suspend-ram requires working wakeup support in
+# QEMU. You *must* check QMP command query-current-machine returns
+# wakeup-suspend-support: true before issuing this command. Failure in
+# doing so can result in a suspended guest that QEMU will not be able to
+# awaken, forcing the user to power cycle the guest to bring it back.
 #
 # This command does NOT return a response on success. There are two options
 # to check for success:
@@ -594,9 +596,11 @@
 #
 # This command requires the pm-utils package to be installed in the guest.
 #
-# IMPORTANT: guest-suspend-hybrid requires QEMU to support the 'system_wakeup'
-# command.  Thus, it's *required* to query QEMU for the presence of the
-# 'system_wakeup' command before issuing guest-suspend-hybrid.
+# IMPORTANT: guest-suspend-hybrid requires working wakeup support in
+# QEMU. You *must* check QMP command query-current-machine returns
+# wakeup-suspend-support: true before issuing this command. Failure in
+# doing so can result in a suspended guest that QEMU will not be able to
+# awaken, forcing the user to power cycle the guest to bring it back.
 #
 # This command does NOT return a response on success. There are two options
 # to check for success:
-- 
2.19.1

[Qemu-devel] [PATCH for-3.2 v10 3/3] qmp hmp: Make system_wakeup check wake-up support and run state

[Daniel Henrique Barboza]

The qmp/hmp command 'system_wakeup' is simply a direct call to
'qemu_system_wakeup_request' from vl.c. This function verifies if
runstate is SUSPENDED and if the wake up reason is valid before
proceeding. However, no error or warning is thrown if any of those
pre-requirements isn't met. There is no way for the caller to
differentiate between a successful wakeup or an error state caused
when trying to wake up a guest that wasn't suspended.

This means that system_wakeup is silently failing, which can be
considered a bug. Adding error handling isn't an API break in this
case - applications that didn't check the result will remain broken,
the ones that check it will have a chance to deal with it.

Adding to that, the commit before previous created a new QMP API called
query-current-machine, with a new flag called wakeup-suspend-support,
that indicates if the guest has the capability of waking up from suspended
state. Although such guest will never reach SUSPENDED state and erroring
it out in this scenario would suffice, it is more informative for the user
to differentiate between a failure because the guest isn't suspended versus
a failure because the guest does not have support for wake up at all.

All this considered, this patch changes qmp_system_wakeup to check if
the guest is capable of waking up from suspend, and if it is suspended.
After this patch, this is the output of system_wakeup in a guest that
does not have wake-up from suspend support (ppc64):

(qemu) system_wakeup
wake-up from suspend is not supported by this guest
(qemu)

And this is the output of system_wakeup in a x86 guest that has the
support but isn't suspended:

(qemu) system_wakeup
Unable to wake up: guest is not in suspended state
(qemu)

Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
---
 hmp.c                   |  5 ++++-
 hw/acpi/core.c          |  4 +++-
 hw/char/serial.c        |  3 ++-
 hw/input/ps2.c          |  9 ++++++---
 hw/timer/mc146818rtc.c  |  3 ++-
 include/sysemu/sysemu.h |  3 ++-
 migration/migration.c   |  7 +++++--
 qapi/misc.json          |  8 +++++++-
 qmp.c                   | 13 ++++++++++++-
 vl.c                    |  6 ++++--
 10 files changed, 47 insertions(+), 14 deletions(-)

diff --git a/hmp.c b/hmp.c
index 7828f93a39..0f5d943413 100644
--- a/hmp.c
+++ b/hmp.c
@@ -1220,7 +1220,10 @@ void hmp_cont(Monitor *mon, const QDict *qdict)
 
 void hmp_system_wakeup(Monitor *mon, const QDict *qdict)
 {
-    qmp_system_wakeup(NULL);
+    Error *err = NULL;
+
+    qmp_system_wakeup(&err);
+    hmp_handle_error(mon, &err);
 }
 
 void hmp_nmi(Monitor *mon, const QDict *qdict)
diff --git a/hw/acpi/core.c b/hw/acpi/core.c
index 52e18d7810..a7073dd435 100644
--- a/hw/acpi/core.c
+++ b/hw/acpi/core.c
@@ -514,7 +514,9 @@ static uint32_t acpi_pm_tmr_get(ACPIREGS *ar)
 static void acpi_pm_tmr_timer(void *opaque)
 {
     ACPIREGS *ar = opaque;
-    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER);
+    Error *err = NULL;
+
+    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER, &err);
     ar->tmr.update_sci(ar);
 }
 
diff --git a/hw/char/serial.c b/hw/char/serial.c
index 02463e3388..23c5d2870c 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -609,9 +609,10 @@ static int serial_can_receive1(void *opaque)
 static void serial_receive1(void *opaque, const uint8_t *buf, int size)
 {
     SerialState *s = opaque;
+    Error *err = NULL;
 
     if (s->wakeup) {
-        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
+        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
     }
     if(s->fcr & UART_FCR_FE) {
         int i;
diff --git a/hw/input/ps2.c b/hw/input/ps2.c
index 6c43fc2912..fc742b0061 100644
--- a/hw/input/ps2.c
+++ b/hw/input/ps2.c
@@ -253,9 +253,10 @@ void ps2_queue_4(PS2State *s, int b1, int b2, int b3, int b4)
 static void ps2_put_keycode(void *opaque, int keycode)
 {
     PS2KbdState *s = opaque;
+    Error *err = NULL;
 
     trace_ps2_put_keycode(opaque, keycode);
-    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
+    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
 
     if (s->translate) {
         if (keycode == 0xf0) {
@@ -276,6 +277,7 @@ static void ps2_keyboard_event(DeviceState *dev, QemuConsole *src,
 {
     PS2KbdState *s = (PS2KbdState *)dev;
     InputKeyEvent *key = evt->u.key.data;
+    Error *err = NULL;
     int qcode;
     uint16_t keycode = 0;
     int mod;
@@ -285,7 +287,7 @@ static void ps2_keyboard_event(DeviceState *dev, QemuConsole *src,
         return;
     }
 
-    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
+    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
     assert(evt->type == INPUT_EVENT_KIND_KEY);
     qcode = qemu_input_key_value_to_qcode(key->key);
 
@@ -741,6 +743,7 @@ static void ps2_mouse_event(DeviceState *dev, QemuConsole *src,
 static void ps2_mouse_sync(DeviceState *dev)
 {
     PS2MouseState *s = (PS2MouseState *)dev;
+    Error *err = NULL;
 
     /* do not sync while disabled to prevent stream corruption */
     if (!(s->mouse_status & MOUSE_STATUS_ENABLED)) {
@@ -748,7 +751,7 @@ static void ps2_mouse_sync(DeviceState *dev)
     }
 
     if (s->mouse_buttons) {
-        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
+        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
     }
     if (!(s->mouse_status & MOUSE_STATUS_REMOTE)) {
         /* if not remote, send event. Multiple events are sent if
diff --git a/hw/timer/mc146818rtc.c b/hw/timer/mc146818rtc.c
index e4e4de8b8a..e65962cdb4 100644
--- a/hw/timer/mc146818rtc.c
+++ b/hw/timer/mc146818rtc.c
@@ -443,6 +443,7 @@ static uint64_t get_next_alarm(RTCState *s)
 static void rtc_update_timer(void *opaque)
 {
     RTCState *s = opaque;
+    Error *err = NULL;
     int32_t irqs = REG_C_UF;
     int32_t new_irqs;
 
@@ -455,7 +456,7 @@ static void rtc_update_timer(void *opaque)
     if (qemu_clock_get_ns(rtc_clock) >= s->next_alarm_time) {
         irqs |= REG_C_AF;
         if (s->cmos_data[RTC_REG_B] & REG_B_AIE) {
-            qemu_system_wakeup_request(QEMU_WAKEUP_REASON_RTC);
+            qemu_system_wakeup_request(QEMU_WAKEUP_REASON_RTC, &err);
         }
     }
 
diff --git a/include/sysemu/sysemu.h b/include/sysemu/sysemu.h
index 0446adacc6..162c4b16d9 100644
--- a/include/sysemu/sysemu.h
+++ b/include/sysemu/sysemu.h
@@ -74,7 +74,8 @@ void qemu_exit_preconfig_request(void);
 void qemu_system_reset_request(ShutdownCause reason);
 void qemu_system_suspend_request(void);
 void qemu_register_suspend_notifier(Notifier *notifier);
-void qemu_system_wakeup_request(WakeupReason reason);
+bool qemu_wakeup_suspend_enabled(void);
+void qemu_system_wakeup_request(WakeupReason reason, Error **errp);
 void qemu_system_wakeup_enable(WakeupReason reason, bool enabled);
 void qemu_register_wakeup_notifier(Notifier *notifier);
 void qemu_register_wakeup_support(void);
diff --git a/migration/migration.c b/migration/migration.c
index b261c1e4ce..8e9b4af71c 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -2390,6 +2390,8 @@ static int postcopy_start(MigrationState *ms)
     int64_t bandwidth = migrate_max_postcopy_bandwidth();
     bool restart_block = false;
     int cur_state = MIGRATION_STATUS_ACTIVE;
+    Error *err = NULL;
+
     if (!migrate_pause_before_switchover()) {
         migrate_set_state(&ms->state, MIGRATION_STATUS_ACTIVE,
                           MIGRATION_STATUS_POSTCOPY_ACTIVE);
@@ -2399,7 +2401,7 @@ static int postcopy_start(MigrationState *ms)
     qemu_mutex_lock_iothread();
     trace_postcopy_start_set_run();
 
-    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
+    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
     global_state_store();
     ret = vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);
     if (ret < 0) {
@@ -2597,13 +2599,14 @@ static int migration_maybe_pause(MigrationState *s,
  */
 static void migration_completion(MigrationState *s)
 {
+    Error *err = NULL;
     int ret;
     int current_active_state = s->state;
 
     if (s->state == MIGRATION_STATUS_ACTIVE) {
         qemu_mutex_lock_iothread();
         s->downtime_start = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
-        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
+        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
         s->vm_was_running = runstate_is_running();
         ret = global_state_store();
 
diff --git a/qapi/misc.json b/qapi/misc.json
index 9b86576e61..471f6e680c 100644
--- a/qapi/misc.json
+++ b/qapi/misc.json
@@ -1235,12 +1235,18 @@
 ##
 # @system_wakeup:
 #
-# Wakeup guest from suspend.  Does nothing in case the guest isn't suspended.
+# Wake-up guest from suspend. If the guest has wake-up from suspend
+# support enabled (wakeup-suspend-support flag from
+# query-current-machine), wake-up guest from suspend if the guest is
+# in SUSPENDED state. Returns an error otherwise.
 #
 # Since:  1.1
 #
 # Returns:  nothing.
 #
+# Note: prior to 3.2, this command does nothing in case the guest
+# isn't suspended.
+#
 # Example:
 #
 # -> { "execute": "system_wakeup" }
diff --git a/qmp.c b/qmp.c
index e7c0a2fd60..2265aa48b9 100644
--- a/qmp.c
+++ b/qmp.c
@@ -183,7 +183,18 @@ void qmp_cont(Error **errp)
 
 void qmp_system_wakeup(Error **errp)
 {
-    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
+    Error *local_err = NULL;
+
+    if (!qemu_wakeup_suspend_enabled()) {
+        error_setg(errp,
+                   "wake-up from suspend is not supported by this guest");
+        return;
+    }
+
+    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &local_err);
+    if (local_err) {
+        error_propagate(errp, local_err);
+    }
 }
 
 ObjectPropertyInfoList *qmp_qom_list(const char *path, Error **errp)
diff --git a/vl.c b/vl.c
index 5813a8785f..b057796e48 100644
--- a/vl.c
+++ b/vl.c
@@ -1754,11 +1754,13 @@ void qemu_register_suspend_notifier(Notifier *notifier)
     notifier_list_add(&suspend_notifiers, notifier);
 }
 
-void qemu_system_wakeup_request(WakeupReason reason)
+void qemu_system_wakeup_request(WakeupReason reason, Error **errp)
 {
     trace_system_wakeup_request(reason);
 
     if (!runstate_check(RUN_STATE_SUSPENDED)) {
+        error_setg(errp,
+                   "Unable to wake up: guest is not in suspended state");
         return;
     }
     if (!(wakeup_reason_mask & (1 << reason))) {
@@ -1788,7 +1790,7 @@ void qemu_register_wakeup_support(void)
     wakeup_suspend_enabled = true;
 }
 
-static bool qemu_wakeup_suspend_enabled(void)
+bool qemu_wakeup_suspend_enabled(void)
 {
     return wakeup_suspend_enabled;
 }
-- 
2.19.1

Re: [Qemu-devel] [PATCH for-3.2 v10 1/3] qmp: query-current-machine with wakeup-suspend-support

[Markus Armbruster]

Daniel Henrique Barboza <danielhb413@gmail.com> writes:

> When issuing the qmp/hmp 'system_wakeup' command, what happens in a
> nutshell is:
>
> - qmp_system_wakeup_request set runstate to RUNNING, sets a wakeup_reason
> and notify the event
> - in the main_loop, all vcpus are paused, a system reset is issued, all
> subscribers of wakeup_notifiers receives a notification, vcpus are then
> resumed and the wake up QAPI event is fired
>
> Note that this procedure alone doesn't ensure that the guest will awake
> from SUSPENDED state - the subscribers of the wake up event must take
> action to resume the guest, otherwise the guest will simply reboot. At
> this moment, only the ACPI machines via acpi_pm1_cnt_init has wake-up
> from suspend support.

the ACPI machines via acpi_pm1_cnt_init have

>
> However, only the presence of 'system_wakeup' is required for QGA to
> support 'guest-suspend-ram' and 'guest-suspend-hybrid' at this moment.
> This means that the user/management will expect to suspend the guest using
> one of those suspend commands and then resume execution using system_wakeup,
> regardless of the support offered in system_wakeup in the first place.
>
> This patch creates a new API called query-current-machine [1], that holds
> a new flag called 'wakeup-suspend-support' that indicates if the guest
> supports wake up from suspend via system_wakeup. The machine is considered
> to implement wake-up support if a call to a new 'qemu_register_wakeup_support'
> is made during its init, as it is now being done inside acpi_pm1_cnt_init.

In xen_hvm_init(), too.

Doesn't that contradict "only the ACPI machines via acpi_pm1_cnt_init
has wake-up from suspend support"?

> This allows for any other machine type to declare wake-up support regardless
> of ACPI state or wakeup_notifiers subscription, making easier for
> newer implementations that might have its own mechanisms in the future.

their own mechanisms

>
> This is the expected output of query-current-machine when running a x86
> guest:
>
> {"execute" : "query-current-machine"}
> {"return": {"wakeup-suspend-support": true}}
>
> Running the same x86 guest, but with the --no-acpi option:
>
> {"execute" : "query-current-machine"}
> {"return": {"wakeup-suspend-support": false}}
>
> This is the output when running a pseries guest:
>
> {"execute" : "query-current-machine"}
> {"return": {"wakeup-suspend-support": false}}
>
> With this extra tool, management can avoid situations where a guest
> that does not have proper suspend/wake capabilities ends up in
> inconsistent state (e.g.
> https://github.com/open-power-host-os/qemu/issues/31).
>
> [1] the decision of creating the query-current-machine API is based
> on discussions in the QEMU mailing list where it was decided that
> query-target wasn't a proper place to store the wake-up flag, neither
> was query-machines because this isn't a static property of the
> machine object. This new API can then be used to store other
> dynamic machine properties that are scattered around the code
> ATM. More info at:
> https://lists.gnu.org/archive/html/qemu-devel/2018-05/msg04235.html
>
> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
> ---
>  hw/acpi/core.c          |  6 ++++++
>  hw/i386/xen/xen-hvm.c   |  5 +++++
>  include/sysemu/sysemu.h |  1 +
>  qapi/misc.json          | 24 ++++++++++++++++++++++++
>  vl.c                    | 19 +++++++++++++++++++
>  5 files changed, 55 insertions(+)
>
> diff --git a/hw/acpi/core.c b/hw/acpi/core.c
> index aafdc61648..52e18d7810 100644
> --- a/hw/acpi/core.c
> +++ b/hw/acpi/core.c
> @@ -617,6 +617,12 @@ void acpi_pm1_cnt_init(ACPIREGS *ar, MemoryRegion *parent,
>      ar->pm1.cnt.s4_val = s4_val;
>      ar->wakeup.notify = acpi_notify_wakeup;
>      qemu_register_wakeup_notifier(&ar->wakeup);
> +
> +    /*
> +     * Register wake-up support in QMP query-current-machine API
> +     */
> +    qemu_register_wakeup_support();
> +

Not sure we really need the comment.  Up to the maintainer.

>      memory_region_init_io(&ar->pm1.cnt.io, memory_region_owner(parent),
>                            &acpi_pm_cnt_ops, ar, "acpi-cnt", 2);
>      memory_region_add_subregion(parent, 4, &ar->pm1.cnt.io);
> diff --git a/hw/i386/xen/xen-hvm.c b/hw/i386/xen/xen-hvm.c
> index 935a3676c8..2143d33b18 100644
> --- a/hw/i386/xen/xen-hvm.c
> +++ b/hw/i386/xen/xen-hvm.c
> @@ -1405,6 +1405,11 @@ void xen_hvm_init(PCMachineState *pcms, MemoryRegion **ram_memory)
>      state->wakeup.notify = xen_wakeup_notifier;
>      qemu_register_wakeup_notifier(&state->wakeup);
>  
> +    /*
> +     * Register wake-up support in QMP query-current-machine API
> +     */
> +    qemu_register_wakeup_support();
> +

Likewise.

>      rc = xen_map_ioreq_server(state);
>      if (rc < 0) {
>          goto err;
> diff --git a/include/sysemu/sysemu.h b/include/sysemu/sysemu.h
> index 8d6095d98b..0446adacc6 100644
> --- a/include/sysemu/sysemu.h
> +++ b/include/sysemu/sysemu.h
> @@ -77,6 +77,7 @@ void qemu_register_suspend_notifier(Notifier *notifier);
>  void qemu_system_wakeup_request(WakeupReason reason);
>  void qemu_system_wakeup_enable(WakeupReason reason, bool enabled);
>  void qemu_register_wakeup_notifier(Notifier *notifier);
> +void qemu_register_wakeup_support(void);
>  void qemu_system_shutdown_request(ShutdownCause reason);
>  void qemu_system_powerdown_request(void);
>  void qemu_register_powerdown_notifier(Notifier *notifier);
> diff --git a/qapi/misc.json b/qapi/misc.json
> index 6c1c5c0a37..9b86576e61 100644
> --- a/qapi/misc.json
> +++ b/qapi/misc.json
> @@ -2008,6 +2008,30 @@
>  ##
>  { 'command': 'query-machines', 'returns': ['MachineInfo'] }
>  
> +##
> +# @CurrentMachineParams:
> +#
> +# Information describing the running machine parameters.
> +#
> +# @wakeup-suspend-support: true if the machine supports wake up from
> +#                          suspend
> +#
> +# Since: 3.2.0

The next release is going to be named 4.0.  Perhaps the maintainer can
fix that up for you.

> +##
> +{ 'struct': 'CurrentMachineParams',
> +  'data': { 'wakeup-suspend-support': 'bool'} }
> +
> +##
> +# @query-current-machine:
> +#
> +# Return information on the current virtual machine.
> +#
> +# Returns: CurrentMachineParams
> +#
> +# Since: 3.2.0
> +##
> +{ 'command': 'query-current-machine', 'returns': 'CurrentMachineParams' }
> +
>  ##
>  # @CpuDefinitionInfo:
>  #
> diff --git a/vl.c b/vl.c
> index 55bab005b6..5813a8785f 100644
> --- a/vl.c
> +++ b/vl.c
> @@ -191,6 +191,7 @@ bool boot_strict;
>  uint8_t *boot_splash_filedata;
>  size_t boot_splash_filedata_size;
>  uint8_t qemu_extra_params_fw[2];
> +bool wakeup_suspend_enabled;
>  
>  int icount_align_option;
>  
> @@ -1782,6 +1783,24 @@ void qemu_register_wakeup_notifier(Notifier *notifier)
>      notifier_list_add(&wakeup_notifiers, notifier);
>  }
>  
> +void qemu_register_wakeup_support(void)
> +{
> +    wakeup_suspend_enabled = true;
> +}
> +
> +static bool qemu_wakeup_suspend_enabled(void)
> +{
> +    return wakeup_suspend_enabled;
> +}
> +
> +CurrentMachineParams *qmp_query_current_machine(Error **errp)
> +{
> +    CurrentMachineParams *params = g_malloc0(sizeof(*params));
> +    params->wakeup_suspend_support = qemu_wakeup_suspend_enabled();
> +
> +    return params;
> +}
> +
>  void qemu_system_killed(int signal, pid_t pid)
>  {
>      shutdown_signal = signal;

Patch looks good to me.

Re: [Qemu-devel] [PATCH for-3.2 v10 2/3] qga: update guest-suspend-ram and guest-suspend-hybrid descriptions

[Markus Armbruster]

Daniel Henrique Barboza <danielhb413@gmail.com> writes:

> This patch updates the descriptions of 'guest-suspend-ram' and
> 'guest-suspend-hybrid' to mention that both commands relies now
> on the proper support for wake up from suspend, retrieved by the
> 'wakeup-suspend-support' attribute of the 'query-current-machine'
> QMP command.
>
> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
> Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
> ---
>  qga/qapi-schema.json | 16 ++++++++++------
>  1 file changed, 10 insertions(+), 6 deletions(-)
>
> diff --git a/qga/qapi-schema.json b/qga/qapi-schema.json
> index c6725b3ec8..78a439263e 100644
> --- a/qga/qapi-schema.json
> +++ b/qga/qapi-schema.json
> @@ -567,9 +567,11 @@
>  # For the best results it's strongly recommended to have the pm-utils
>  # package installed in the guest.
>  #
> -# IMPORTANT: guest-suspend-ram requires QEMU to support the 'system_wakeup'
> -# command.  Thus, it's *required* to query QEMU for the presence of the
> -# 'system_wakeup' command before issuing guest-suspend-ram.
> +# IMPORTANT: guest-suspend-ram requires working wakeup support in
> +# QEMU. You *must* check QMP command query-current-machine returns
> +# wakeup-suspend-support: true before issuing this command. Failure in
> +# doing so can result in a suspended guest that QEMU will not be able to
> +# awaken, forcing the user to power cycle the guest to bring it back.
>  #
>  # This command does NOT return a response on success. There are two options
>  # to check for success:
> @@ -594,9 +596,11 @@
>  #
>  # This command requires the pm-utils package to be installed in the guest.
>  #
> -# IMPORTANT: guest-suspend-hybrid requires QEMU to support the 'system_wakeup'
> -# command.  Thus, it's *required* to query QEMU for the presence of the
> -# 'system_wakeup' command before issuing guest-suspend-hybrid.
> +# IMPORTANT: guest-suspend-hybrid requires working wakeup support in
> +# QEMU. You *must* check QMP command query-current-machine returns
> +# wakeup-suspend-support: true before issuing this command. Failure in
> +# doing so can result in a suspended guest that QEMU will not be able to
> +# awaken, forcing the user to power cycle the guest to bring it back.
>  #
>  # This command does NOT return a response on success. There are two options
>  # to check for success:

Nice improvement.  However, I'd tone down "You *must*" to "You should".
This is just one of many ways to screw up the guest with the guest
agent.  While serious, it's not as bad as compromising the sandbox, or
killing QEMU.

Re: [Qemu-devel] [PATCH for-3.2 v10 3/3] qmp hmp: Make system_wakeup check wake-up support and run state

[Markus Armbruster]

Daniel Henrique Barboza <danielhb413@gmail.com> writes:

> The qmp/hmp command 'system_wakeup' is simply a direct call to
> 'qemu_system_wakeup_request' from vl.c. This function verifies if
> runstate is SUSPENDED and if the wake up reason is valid before
> proceeding. However, no error or warning is thrown if any of those
> pre-requirements isn't met. There is no way for the caller to
> differentiate between a successful wakeup or an error state caused
> when trying to wake up a guest that wasn't suspended.
>
> This means that system_wakeup is silently failing, which can be
> considered a bug. Adding error handling isn't an API break in this
> case - applications that didn't check the result will remain broken,
> the ones that check it will have a chance to deal with it.
>
> Adding to that, the commit before previous created a new QMP API called
> query-current-machine, with a new flag called wakeup-suspend-support,
> that indicates if the guest has the capability of waking up from suspended
> state. Although such guest will never reach SUSPENDED state and erroring
> it out in this scenario would suffice, it is more informative for the user
> to differentiate between a failure because the guest isn't suspended versus
> a failure because the guest does not have support for wake up at all.
>
> All this considered, this patch changes qmp_system_wakeup to check if
> the guest is capable of waking up from suspend, and if it is suspended.
> After this patch, this is the output of system_wakeup in a guest that
> does not have wake-up from suspend support (ppc64):
>
> (qemu) system_wakeup
> wake-up from suspend is not supported by this guest
> (qemu)
>
> And this is the output of system_wakeup in a x86 guest that has the
> support but isn't suspended:
>
> (qemu) system_wakeup
> Unable to wake up: guest is not in suspended state
> (qemu)
>
> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
> ---
>  hmp.c                   |  5 ++++-
>  hw/acpi/core.c          |  4 +++-
>  hw/char/serial.c        |  3 ++-
>  hw/input/ps2.c          |  9 ++++++---
>  hw/timer/mc146818rtc.c  |  3 ++-
>  include/sysemu/sysemu.h |  3 ++-
>  migration/migration.c   |  7 +++++--
>  qapi/misc.json          |  8 +++++++-
>  qmp.c                   | 13 ++++++++++++-
>  vl.c                    |  6 ++++--
>  10 files changed, 47 insertions(+), 14 deletions(-)
>
> diff --git a/hmp.c b/hmp.c
> index 7828f93a39..0f5d943413 100644
> --- a/hmp.c
> +++ b/hmp.c
> @@ -1220,7 +1220,10 @@ void hmp_cont(Monitor *mon, const QDict *qdict)
>  
>  void hmp_system_wakeup(Monitor *mon, const QDict *qdict)
>  {
> -    qmp_system_wakeup(NULL);
> +    Error *err = NULL;
> +
> +    qmp_system_wakeup(&err);
> +    hmp_handle_error(mon, &err);
>  }
>  
>  void hmp_nmi(Monitor *mon, const QDict *qdict)
> diff --git a/hw/acpi/core.c b/hw/acpi/core.c
> index 52e18d7810..a7073dd435 100644
> --- a/hw/acpi/core.c
> +++ b/hw/acpi/core.c
> @@ -514,7 +514,9 @@ static uint32_t acpi_pm_tmr_get(ACPIREGS *ar)
>  static void acpi_pm_tmr_timer(void *opaque)
>  {
>      ACPIREGS *ar = opaque;
> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER);
> +    Error *err = NULL;
> +
> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER, &err);
>      ar->tmr.update_sci(ar);
>  }
>  
> diff --git a/hw/char/serial.c b/hw/char/serial.c
> index 02463e3388..23c5d2870c 100644
> --- a/hw/char/serial.c
> +++ b/hw/char/serial.c
> @@ -609,9 +609,10 @@ static int serial_can_receive1(void *opaque)
>  static void serial_receive1(void *opaque, const uint8_t *buf, int size)
>  {
>      SerialState *s = opaque;
> +    Error *err = NULL;
>  
>      if (s->wakeup) {
> -        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
> +        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
>      }
>      if(s->fcr & UART_FCR_FE) {
>          int i;
> diff --git a/hw/input/ps2.c b/hw/input/ps2.c
> index 6c43fc2912..fc742b0061 100644
> --- a/hw/input/ps2.c
> +++ b/hw/input/ps2.c
> @@ -253,9 +253,10 @@ void ps2_queue_4(PS2State *s, int b1, int b2, int b3, int b4)
>  static void ps2_put_keycode(void *opaque, int keycode)
>  {
>      PS2KbdState *s = opaque;
> +    Error *err = NULL;
>  
>      trace_ps2_put_keycode(opaque, keycode);
> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
>  
>      if (s->translate) {
>          if (keycode == 0xf0) {
> @@ -276,6 +277,7 @@ static void ps2_keyboard_event(DeviceState *dev, QemuConsole *src,
>  {
>      PS2KbdState *s = (PS2KbdState *)dev;
>      InputKeyEvent *key = evt->u.key.data;
> +    Error *err = NULL;
>      int qcode;
>      uint16_t keycode = 0;
>      int mod;
> @@ -285,7 +287,7 @@ static void ps2_keyboard_event(DeviceState *dev, QemuConsole *src,
>          return;
>      }
>  
> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
>      assert(evt->type == INPUT_EVENT_KIND_KEY);
>      qcode = qemu_input_key_value_to_qcode(key->key);
>  
> @@ -741,6 +743,7 @@ static void ps2_mouse_event(DeviceState *dev, QemuConsole *src,
>  static void ps2_mouse_sync(DeviceState *dev)
>  {
>      PS2MouseState *s = (PS2MouseState *)dev;
> +    Error *err = NULL;
>  
>      /* do not sync while disabled to prevent stream corruption */
>      if (!(s->mouse_status & MOUSE_STATUS_ENABLED)) {
> @@ -748,7 +751,7 @@ static void ps2_mouse_sync(DeviceState *dev)
>      }
>  
>      if (s->mouse_buttons) {
> -        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
> +        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
>      }
>      if (!(s->mouse_status & MOUSE_STATUS_REMOTE)) {
>          /* if not remote, send event. Multiple events are sent if
> diff --git a/hw/timer/mc146818rtc.c b/hw/timer/mc146818rtc.c
> index e4e4de8b8a..e65962cdb4 100644
> --- a/hw/timer/mc146818rtc.c
> +++ b/hw/timer/mc146818rtc.c
> @@ -443,6 +443,7 @@ static uint64_t get_next_alarm(RTCState *s)
>  static void rtc_update_timer(void *opaque)
>  {
>      RTCState *s = opaque;
> +    Error *err = NULL;
>      int32_t irqs = REG_C_UF;
>      int32_t new_irqs;
>  
> @@ -455,7 +456,7 @@ static void rtc_update_timer(void *opaque)
>      if (qemu_clock_get_ns(rtc_clock) >= s->next_alarm_time) {
>          irqs |= REG_C_AF;
>          if (s->cmos_data[RTC_REG_B] & REG_B_AIE) {
> -            qemu_system_wakeup_request(QEMU_WAKEUP_REASON_RTC);
> +            qemu_system_wakeup_request(QEMU_WAKEUP_REASON_RTC, &err);
>          }
>      }
>  
> diff --git a/include/sysemu/sysemu.h b/include/sysemu/sysemu.h
> index 0446adacc6..162c4b16d9 100644
> --- a/include/sysemu/sysemu.h
> +++ b/include/sysemu/sysemu.h
> @@ -74,7 +74,8 @@ void qemu_exit_preconfig_request(void);
>  void qemu_system_reset_request(ShutdownCause reason);
>  void qemu_system_suspend_request(void);
>  void qemu_register_suspend_notifier(Notifier *notifier);
> -void qemu_system_wakeup_request(WakeupReason reason);
> +bool qemu_wakeup_suspend_enabled(void);
> +void qemu_system_wakeup_request(WakeupReason reason, Error **errp);
>  void qemu_system_wakeup_enable(WakeupReason reason, bool enabled);
>  void qemu_register_wakeup_notifier(Notifier *notifier);
>  void qemu_register_wakeup_support(void);
> diff --git a/migration/migration.c b/migration/migration.c
> index b261c1e4ce..8e9b4af71c 100644
> --- a/migration/migration.c
> +++ b/migration/migration.c
> @@ -2390,6 +2390,8 @@ static int postcopy_start(MigrationState *ms)
>      int64_t bandwidth = migrate_max_postcopy_bandwidth();
>      bool restart_block = false;
>      int cur_state = MIGRATION_STATUS_ACTIVE;
> +    Error *err = NULL;
> +
>      if (!migrate_pause_before_switchover()) {
>          migrate_set_state(&ms->state, MIGRATION_STATUS_ACTIVE,
>                            MIGRATION_STATUS_POSTCOPY_ACTIVE);
> @@ -2399,7 +2401,7 @@ static int postcopy_start(MigrationState *ms)
>      qemu_mutex_lock_iothread();
>      trace_postcopy_start_set_run();
>  
> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
>      global_state_store();
>      ret = vm_stop_force_state(RUN_STATE_FINISH_MIGRATE);
>      if (ret < 0) {
> @@ -2597,13 +2599,14 @@ static int migration_maybe_pause(MigrationState *s,
>   */
>  static void migration_completion(MigrationState *s)
>  {
> +    Error *err = NULL;
>      int ret;
>      int current_active_state = s->state;
>  
>      if (s->state == MIGRATION_STATUS_ACTIVE) {
>          qemu_mutex_lock_iothread();
>          s->downtime_start = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
> -        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
> +        qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &err);
>          s->vm_was_running = runstate_is_running();
>          ret = global_state_store();
>  
> diff --git a/qapi/misc.json b/qapi/misc.json
> index 9b86576e61..471f6e680c 100644
> --- a/qapi/misc.json
> +++ b/qapi/misc.json
> @@ -1235,12 +1235,18 @@
>  ##
>  # @system_wakeup:
>  #
> -# Wakeup guest from suspend.  Does nothing in case the guest isn't suspended.
> +# Wake-up guest from suspend. If the guest has wake-up from suspend
> +# support enabled (wakeup-suspend-support flag from
> +# query-current-machine), wake-up guest from suspend if the guest is

"wake up guest", because wake-up is the noun, the verb is spelled wake
up.

> +# in SUSPENDED state. Returns an error otherwise.

Imperative mood, please: "Return an error".

>  #
>  # Since:  1.1
>  #
>  # Returns:  nothing.
>  #
> +# Note: prior to 3.2, this command does nothing in case the guest
> +# isn't suspended.

The next relase will be 4.0.

> +#
>  # Example:
>  #
>  # -> { "execute": "system_wakeup" }
> diff --git a/qmp.c b/qmp.c
> index e7c0a2fd60..2265aa48b9 100644
> --- a/qmp.c
> +++ b/qmp.c
> @@ -183,7 +183,18 @@ void qmp_cont(Error **errp)
>  
>  void qmp_system_wakeup(Error **errp)
>  {
> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER);
> +    Error *local_err = NULL;
> +
> +    if (!qemu_wakeup_suspend_enabled()) {
> +        error_setg(errp,
> +                   "wake-up from suspend is not supported by this guest");
> +        return;
> +    }
> +
> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, &local_err);
> +    if (local_err) {
> +        error_propagate(errp, local_err);
> +    }

Make that

       qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, errp);

See error.h's big comment:

 * Receive an error and pass it on to the caller:
 *     Error *err = NULL;
 *     foo(arg, &err);
 *     if (err) {
 *         handle the error...
 *         error_propagate(errp, err);
 *     }
 * where Error **errp is a parameter, by convention the last one.
[...]
 * But when all you do with the error is pass it on, please use
 *     foo(arg, errp);
 * for readability.

>  }
>  
>  ObjectPropertyInfoList *qmp_qom_list(const char *path, Error **errp)
> diff --git a/vl.c b/vl.c
> index 5813a8785f..b057796e48 100644
> --- a/vl.c
> +++ b/vl.c
> @@ -1754,11 +1754,13 @@ void qemu_register_suspend_notifier(Notifier *notifier)
>      notifier_list_add(&suspend_notifiers, notifier);
>  }
>  
> -void qemu_system_wakeup_request(WakeupReason reason)
> +void qemu_system_wakeup_request(WakeupReason reason, Error **errp)
>  {
>      trace_system_wakeup_request(reason);
>  
>      if (!runstate_check(RUN_STATE_SUSPENDED)) {
> +        error_setg(errp,
> +                   "Unable to wake up: guest is not in suspended state");
>          return;
>      }
>      if (!(wakeup_reason_mask & (1 << reason))) {
> @@ -1788,7 +1790,7 @@ void qemu_register_wakeup_support(void)
>      wakeup_suspend_enabled = true;
>  }
>  
> -static bool qemu_wakeup_suspend_enabled(void)
> +bool qemu_wakeup_suspend_enabled(void)
>  {
>      return wakeup_suspend_enabled;
>  }

Re: [Qemu-devel] [PATCH for-3.2 v10 3/3] qmp hmp: Make system_wakeup check wake-up support and run state

[Markus Armbruster]

One more thing...

Markus Armbruster <armbru@redhat.com> writes:

> Daniel Henrique Barboza <danielhb413@gmail.com> writes:
>
>> The qmp/hmp command 'system_wakeup' is simply a direct call to
>> 'qemu_system_wakeup_request' from vl.c. This function verifies if
>> runstate is SUSPENDED and if the wake up reason is valid before
>> proceeding. However, no error or warning is thrown if any of those
>> pre-requirements isn't met. There is no way for the caller to
>> differentiate between a successful wakeup or an error state caused
>> when trying to wake up a guest that wasn't suspended.
>>
>> This means that system_wakeup is silently failing, which can be
>> considered a bug. Adding error handling isn't an API break in this
>> case - applications that didn't check the result will remain broken,
>> the ones that check it will have a chance to deal with it.
>>
>> Adding to that, the commit before previous created a new QMP API called
>> query-current-machine, with a new flag called wakeup-suspend-support,
>> that indicates if the guest has the capability of waking up from suspended
>> state. Although such guest will never reach SUSPENDED state and erroring
>> it out in this scenario would suffice, it is more informative for the user
>> to differentiate between a failure because the guest isn't suspended versus
>> a failure because the guest does not have support for wake up at all.
>>
>> All this considered, this patch changes qmp_system_wakeup to check if
>> the guest is capable of waking up from suspend, and if it is suspended.
>> After this patch, this is the output of system_wakeup in a guest that
>> does not have wake-up from suspend support (ppc64):
>>
>> (qemu) system_wakeup
>> wake-up from suspend is not supported by this guest
>> (qemu)
>>
>> And this is the output of system_wakeup in a x86 guest that has the
>> support but isn't suspended:
>>
>> (qemu) system_wakeup
>> Unable to wake up: guest is not in suspended state
>> (qemu)
>>
>> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
>> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
>> ---
>>  hmp.c                   |  5 ++++-
>>  hw/acpi/core.c          |  4 +++-
>>  hw/char/serial.c        |  3 ++-
>>  hw/input/ps2.c          |  9 ++++++---
>>  hw/timer/mc146818rtc.c  |  3 ++-
>>  include/sysemu/sysemu.h |  3 ++-
>>  migration/migration.c   |  7 +++++--
>>  qapi/misc.json          |  8 +++++++-
>>  qmp.c                   | 13 ++++++++++++-
>>  vl.c                    |  6 ++++--
>>  10 files changed, 47 insertions(+), 14 deletions(-)
>>
>> diff --git a/hmp.c b/hmp.c
>> index 7828f93a39..0f5d943413 100644
>> --- a/hmp.c
>> +++ b/hmp.c
>> @@ -1220,7 +1220,10 @@ void hmp_cont(Monitor *mon, const QDict *qdict)
>>  
>>  void hmp_system_wakeup(Monitor *mon, const QDict *qdict)
>>  {
>> -    qmp_system_wakeup(NULL);
>> +    Error *err = NULL;
>> +
>> +    qmp_system_wakeup(&err);
>> +    hmp_handle_error(mon, &err);
>>  }
>>  
>>  void hmp_nmi(Monitor *mon, const QDict *qdict)
>> diff --git a/hw/acpi/core.c b/hw/acpi/core.c
>> index 52e18d7810..a7073dd435 100644
>> --- a/hw/acpi/core.c
>> +++ b/hw/acpi/core.c
>> @@ -514,7 +514,9 @@ static uint32_t acpi_pm_tmr_get(ACPIREGS *ar)
>>  static void acpi_pm_tmr_timer(void *opaque)
>>  {
>>      ACPIREGS *ar = opaque;
>> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER);
>> +    Error *err = NULL;
>> +
>> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER, &err);
>>      ar->tmr.update_sci(ar);
>>  }

Leaks the error object when qemu_system_wakeup_request() fails.

If it cannot fail here, pass &error_abort.

If it can fail, but you want to ignore failure, pass NULL.

More of the same elsewhere.

[...]

Re: [Qemu-devel] [PATCH for-3.2 v10 2/3] qga: update guest-suspend-ram and guest-suspend-hybrid descriptions

[Daniel Henrique Barboza]

On 11/29/18 10:18 AM, Markus Armbruster wrote:
> Daniel Henrique Barboza <danielhb413@gmail.com> writes:
>
>> This patch updates the descriptions of 'guest-suspend-ram' and
>> 'guest-suspend-hybrid' to mention that both commands relies now
>> on the proper support for wake up from suspend, retrieved by the
>> 'wakeup-suspend-support' attribute of the 'query-current-machine'
>> QMP command.
>>
>> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
>> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
>> Reviewed-by: Michael Roth <mdroth@linux.vnet.ibm.com>
>> ---
>>   qga/qapi-schema.json | 16 ++++++++++------
>>   1 file changed, 10 insertions(+), 6 deletions(-)
>>
>> diff --git a/qga/qapi-schema.json b/qga/qapi-schema.json
>> index c6725b3ec8..78a439263e 100644
>> --- a/qga/qapi-schema.json
>> +++ b/qga/qapi-schema.json
>> @@ -567,9 +567,11 @@
>>   # For the best results it's strongly recommended to have the pm-utils
>>   # package installed in the guest.
>>   #
>> -# IMPORTANT: guest-suspend-ram requires QEMU to support the 'system_wakeup'
>> -# command.  Thus, it's *required* to query QEMU for the presence of the
>> -# 'system_wakeup' command before issuing guest-suspend-ram.
>> +# IMPORTANT: guest-suspend-ram requires working wakeup support in
>> +# QEMU. You *must* check QMP command query-current-machine returns
>> +# wakeup-suspend-support: true before issuing this command. Failure in
>> +# doing so can result in a suspended guest that QEMU will not be able to
>> +# awaken, forcing the user to power cycle the guest to bring it back.
>>   #
>>   # This command does NOT return a response on success. There are two options
>>   # to check for success:
>> @@ -594,9 +596,11 @@
>>   #
>>   # This command requires the pm-utils package to be installed in the guest.
>>   #
>> -# IMPORTANT: guest-suspend-hybrid requires QEMU to support the 'system_wakeup'
>> -# command.  Thus, it's *required* to query QEMU for the presence of the
>> -# 'system_wakeup' command before issuing guest-suspend-hybrid.
>> +# IMPORTANT: guest-suspend-hybrid requires working wakeup support in
>> +# QEMU. You *must* check QMP command query-current-machine returns
>> +# wakeup-suspend-support: true before issuing this command. Failure in
>> +# doing so can result in a suspended guest that QEMU will not be able to
>> +# awaken, forcing the user to power cycle the guest to bring it back.
>>   #
>>   # This command does NOT return a response on success. There are two options
>>   # to check for success:
> Nice improvement.  However, I'd tone down "You *must*" to "You should".
> This is just one of many ways to screw up the guest with the guest
> agent.  While serious, it's not as bad as compromising the sandbox, or
> killing QEMU.

I'll tone down in the next spin. Thanks!

Re: [Qemu-devel] [PATCH for-3.2 v10 3/3] qmp hmp: Make system_wakeup check wake-up support and run state

[Daniel Henrique Barboza]

On 11/29/18 4:55 PM, Markus Armbruster wrote:
> One more thing...
>
> Markus Armbruster <armbru@redhat.com> writes:
>
>> Daniel Henrique Barboza <danielhb413@gmail.com> writes:
>>
>>> The qmp/hmp command 'system_wakeup' is simply a direct call to
>>> 'qemu_system_wakeup_request' from vl.c. This function verifies if
>>> runstate is SUSPENDED and if the wake up reason is valid before
>>> proceeding. However, no error or warning is thrown if any of those
>>> pre-requirements isn't met. There is no way for the caller to
>>> differentiate between a successful wakeup or an error state caused
>>> when trying to wake up a guest that wasn't suspended.
>>>
>>> This means that system_wakeup is silently failing, which can be
>>> considered a bug. Adding error handling isn't an API break in this
>>> case - applications that didn't check the result will remain broken,
>>> the ones that check it will have a chance to deal with it.
>>>
>>> Adding to that, the commit before previous created a new QMP API called
>>> query-current-machine, with a new flag called wakeup-suspend-support,
>>> that indicates if the guest has the capability of waking up from suspended
>>> state. Although such guest will never reach SUSPENDED state and erroring
>>> it out in this scenario would suffice, it is more informative for the user
>>> to differentiate between a failure because the guest isn't suspended versus
>>> a failure because the guest does not have support for wake up at all.
>>>
>>> All this considered, this patch changes qmp_system_wakeup to check if
>>> the guest is capable of waking up from suspend, and if it is suspended.
>>> After this patch, this is the output of system_wakeup in a guest that
>>> does not have wake-up from suspend support (ppc64):
>>>
>>> (qemu) system_wakeup
>>> wake-up from suspend is not supported by this guest
>>> (qemu)
>>>
>>> And this is the output of system_wakeup in a x86 guest that has the
>>> support but isn't suspended:
>>>
>>> (qemu) system_wakeup
>>> Unable to wake up: guest is not in suspended state
>>> (qemu)
>>>
>>> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
>>> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
>>> ---
>>>   hmp.c                   |  5 ++++-
>>>   hw/acpi/core.c          |  4 +++-
>>>   hw/char/serial.c        |  3 ++-
>>>   hw/input/ps2.c          |  9 ++++++---
>>>   hw/timer/mc146818rtc.c  |  3 ++-
>>>   include/sysemu/sysemu.h |  3 ++-
>>>   migration/migration.c   |  7 +++++--
>>>   qapi/misc.json          |  8 +++++++-
>>>   qmp.c                   | 13 ++++++++++++-
>>>   vl.c                    |  6 ++++--
>>>   10 files changed, 47 insertions(+), 14 deletions(-)
>>>
>>> diff --git a/hmp.c b/hmp.c
>>> index 7828f93a39..0f5d943413 100644
>>> --- a/hmp.c
>>> +++ b/hmp.c
>>> @@ -1220,7 +1220,10 @@ void hmp_cont(Monitor *mon, const QDict *qdict)
>>>   
>>>   void hmp_system_wakeup(Monitor *mon, const QDict *qdict)
>>>   {
>>> -    qmp_system_wakeup(NULL);
>>> +    Error *err = NULL;
>>> +
>>> +    qmp_system_wakeup(&err);
>>> +    hmp_handle_error(mon, &err);
>>>   }
>>>   
>>>   void hmp_nmi(Monitor *mon, const QDict *qdict)
>>> diff --git a/hw/acpi/core.c b/hw/acpi/core.c
>>> index 52e18d7810..a7073dd435 100644
>>> --- a/hw/acpi/core.c
>>> +++ b/hw/acpi/core.c
>>> @@ -514,7 +514,9 @@ static uint32_t acpi_pm_tmr_get(ACPIREGS *ar)
>>>   static void acpi_pm_tmr_timer(void *opaque)
>>>   {
>>>       ACPIREGS *ar = opaque;
>>> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER);
>>> +    Error *err = NULL;
>>> +
>>> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER, &err);
>>>       ar->tmr.update_sci(ar);
>>>   }
> Leaks the error object when qemu_system_wakeup_request() fails.
>
> If it cannot fail here, pass &error_abort.
>
> If it can fail, but you want to ignore failure, pass NULL.

Good point. I'll simply pass NULL to all callers that didn't care
for the error prior to this change.



Thanks,


Daniel


>
> More of the same elsewhere.
>
> [...]

Re: [Qemu-devel] [PATCH for-3.2 v10 3/3] qmp hmp: Make system_wakeup check wake-up support and run state

[Eduardo Habkost]

On Tue, Dec 04, 2018 at 04:36:56PM -0200, Daniel Henrique Barboza wrote:
> 
> 
> On 11/29/18 4:55 PM, Markus Armbruster wrote:
> > One more thing...
> > 
> > Markus Armbruster <armbru@redhat.com> writes:
> > 
> > > Daniel Henrique Barboza <danielhb413@gmail.com> writes:
> > > 
> > > > The qmp/hmp command 'system_wakeup' is simply a direct call to
> > > > 'qemu_system_wakeup_request' from vl.c. This function verifies if
> > > > runstate is SUSPENDED and if the wake up reason is valid before
> > > > proceeding. However, no error or warning is thrown if any of those
> > > > pre-requirements isn't met. There is no way for the caller to
> > > > differentiate between a successful wakeup or an error state caused
> > > > when trying to wake up a guest that wasn't suspended.
> > > > 
> > > > This means that system_wakeup is silently failing, which can be
> > > > considered a bug. Adding error handling isn't an API break in this
> > > > case - applications that didn't check the result will remain broken,
> > > > the ones that check it will have a chance to deal with it.
> > > > 
> > > > Adding to that, the commit before previous created a new QMP API called
> > > > query-current-machine, with a new flag called wakeup-suspend-support,
> > > > that indicates if the guest has the capability of waking up from suspended
> > > > state. Although such guest will never reach SUSPENDED state and erroring
> > > > it out in this scenario would suffice, it is more informative for the user
> > > > to differentiate between a failure because the guest isn't suspended versus
> > > > a failure because the guest does not have support for wake up at all.
> > > > 
> > > > All this considered, this patch changes qmp_system_wakeup to check if
> > > > the guest is capable of waking up from suspend, and if it is suspended.
> > > > After this patch, this is the output of system_wakeup in a guest that
> > > > does not have wake-up from suspend support (ppc64):
> > > > 
> > > > (qemu) system_wakeup
> > > > wake-up from suspend is not supported by this guest
> > > > (qemu)
> > > > 
> > > > And this is the output of system_wakeup in a x86 guest that has the
> > > > support but isn't suspended:
> > > > 
> > > > (qemu) system_wakeup
> > > > Unable to wake up: guest is not in suspended state
> > > > (qemu)
> > > > 
> > > > Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
> > > > Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
> > > > ---
> > > >   hmp.c                   |  5 ++++-
> > > >   hw/acpi/core.c          |  4 +++-
> > > >   hw/char/serial.c        |  3 ++-
> > > >   hw/input/ps2.c          |  9 ++++++---
> > > >   hw/timer/mc146818rtc.c  |  3 ++-
> > > >   include/sysemu/sysemu.h |  3 ++-
> > > >   migration/migration.c   |  7 +++++--
> > > >   qapi/misc.json          |  8 +++++++-
> > > >   qmp.c                   | 13 ++++++++++++-
> > > >   vl.c                    |  6 ++++--
> > > >   10 files changed, 47 insertions(+), 14 deletions(-)
> > > > 
> > > > diff --git a/hmp.c b/hmp.c
> > > > index 7828f93a39..0f5d943413 100644
> > > > --- a/hmp.c
> > > > +++ b/hmp.c
> > > > @@ -1220,7 +1220,10 @@ void hmp_cont(Monitor *mon, const QDict *qdict)
> > > >   void hmp_system_wakeup(Monitor *mon, const QDict *qdict)
> > > >   {
> > > > -    qmp_system_wakeup(NULL);
> > > > +    Error *err = NULL;
> > > > +
> > > > +    qmp_system_wakeup(&err);
> > > > +    hmp_handle_error(mon, &err);
> > > >   }
> > > >   void hmp_nmi(Monitor *mon, const QDict *qdict)
> > > > diff --git a/hw/acpi/core.c b/hw/acpi/core.c
> > > > index 52e18d7810..a7073dd435 100644
> > > > --- a/hw/acpi/core.c
> > > > +++ b/hw/acpi/core.c
> > > > @@ -514,7 +514,9 @@ static uint32_t acpi_pm_tmr_get(ACPIREGS *ar)
> > > >   static void acpi_pm_tmr_timer(void *opaque)
> > > >   {
> > > >       ACPIREGS *ar = opaque;
> > > > -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER);
> > > > +    Error *err = NULL;
> > > > +
> > > > +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER, &err);
> > > >       ar->tmr.update_sci(ar);
> > > >   }
> > Leaks the error object when qemu_system_wakeup_request() fails.
> > 
> > If it cannot fail here, pass &error_abort.
> > 
> > If it can fail, but you want to ignore failure, pass NULL.
> 
> Good point. I'll simply pass NULL to all callers that didn't care
> for the error prior to this change.

Most times I saw QEMU code ignoring errors, it actually didn't
expect any errors to happen and was supposed to be using
&error_abort instead.

This makes NULL errp always look like a bug to be fixed.  If you
are sure you really want to ignore an error, I'd recommend adding
a comment making your intention explicit.

-- 
Eduardo

Re: [Qemu-devel] [PATCH for-3.2 v10 3/3] qmp hmp: Make system_wakeup check wake-up support and run state

[Daniel Henrique Barboza]

On 12/4/18 5:15 PM, Eduardo Habkost wrote:
> On Tue, Dec 04, 2018 at 04:36:56PM -0200, Daniel Henrique Barboza wrote:
>>
>> On 11/29/18 4:55 PM, Markus Armbruster wrote:
>>> One more thing...
>>>
>>> Markus Armbruster <armbru@redhat.com> writes:
>>>
>>>> Daniel Henrique Barboza <danielhb413@gmail.com> writes:
>>>>
>>>>> The qmp/hmp command 'system_wakeup' is simply a direct call to
>>>>> 'qemu_system_wakeup_request' from vl.c. This function verifies if
>>>>> runstate is SUSPENDED and if the wake up reason is valid before
>>>>> proceeding. However, no error or warning is thrown if any of those
>>>>> pre-requirements isn't met. There is no way for the caller to
>>>>> differentiate between a successful wakeup or an error state caused
>>>>> when trying to wake up a guest that wasn't suspended.
>>>>>
>>>>> This means that system_wakeup is silently failing, which can be
>>>>> considered a bug. Adding error handling isn't an API break in this
>>>>> case - applications that didn't check the result will remain broken,
>>>>> the ones that check it will have a chance to deal with it.
>>>>>
>>>>> Adding to that, the commit before previous created a new QMP API called
>>>>> query-current-machine, with a new flag called wakeup-suspend-support,
>>>>> that indicates if the guest has the capability of waking up from suspended
>>>>> state. Although such guest will never reach SUSPENDED state and erroring
>>>>> it out in this scenario would suffice, it is more informative for the user
>>>>> to differentiate between a failure because the guest isn't suspended versus
>>>>> a failure because the guest does not have support for wake up at all.
>>>>>
>>>>> All this considered, this patch changes qmp_system_wakeup to check if
>>>>> the guest is capable of waking up from suspend, and if it is suspended.
>>>>> After this patch, this is the output of system_wakeup in a guest that
>>>>> does not have wake-up from suspend support (ppc64):
>>>>>
>>>>> (qemu) system_wakeup
>>>>> wake-up from suspend is not supported by this guest
>>>>> (qemu)
>>>>>
>>>>> And this is the output of system_wakeup in a x86 guest that has the
>>>>> support but isn't suspended:
>>>>>
>>>>> (qemu) system_wakeup
>>>>> Unable to wake up: guest is not in suspended state
>>>>> (qemu)
>>>>>
>>>>> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
>>>>> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
>>>>> ---
>>>>>    hmp.c                   |  5 ++++-
>>>>>    hw/acpi/core.c          |  4 +++-
>>>>>    hw/char/serial.c        |  3 ++-
>>>>>    hw/input/ps2.c          |  9 ++++++---
>>>>>    hw/timer/mc146818rtc.c  |  3 ++-
>>>>>    include/sysemu/sysemu.h |  3 ++-
>>>>>    migration/migration.c   |  7 +++++--
>>>>>    qapi/misc.json          |  8 +++++++-
>>>>>    qmp.c                   | 13 ++++++++++++-
>>>>>    vl.c                    |  6 ++++--
>>>>>    10 files changed, 47 insertions(+), 14 deletions(-)
>>>>>
>>>>> diff --git a/hmp.c b/hmp.c
>>>>> index 7828f93a39..0f5d943413 100644
>>>>> --- a/hmp.c
>>>>> +++ b/hmp.c
>>>>> @@ -1220,7 +1220,10 @@ void hmp_cont(Monitor *mon, const QDict *qdict)
>>>>>    void hmp_system_wakeup(Monitor *mon, const QDict *qdict)
>>>>>    {
>>>>> -    qmp_system_wakeup(NULL);
>>>>> +    Error *err = NULL;
>>>>> +
>>>>> +    qmp_system_wakeup(&err);
>>>>> +    hmp_handle_error(mon, &err);
>>>>>    }
>>>>>    void hmp_nmi(Monitor *mon, const QDict *qdict)
>>>>> diff --git a/hw/acpi/core.c b/hw/acpi/core.c
>>>>> index 52e18d7810..a7073dd435 100644
>>>>> --- a/hw/acpi/core.c
>>>>> +++ b/hw/acpi/core.c
>>>>> @@ -514,7 +514,9 @@ static uint32_t acpi_pm_tmr_get(ACPIREGS *ar)
>>>>>    static void acpi_pm_tmr_timer(void *opaque)
>>>>>    {
>>>>>        ACPIREGS *ar = opaque;
>>>>> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER);
>>>>> +    Error *err = NULL;
>>>>> +
>>>>> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER, &err);
>>>>>        ar->tmr.update_sci(ar);
>>>>>    }
>>> Leaks the error object when qemu_system_wakeup_request() fails.
>>>
>>> If it cannot fail here, pass &error_abort.
>>>
>>> If it can fail, but you want to ignore failure, pass NULL.
>> Good point. I'll simply pass NULL to all callers that didn't care
>> for the error prior to this change.
> Most times I saw QEMU code ignoring errors, it actually didn't
> expect any errors to happen and was supposed to be using
> &error_abort instead.

In this particular case, the existing code wouldn't be expecting errors 
because
qemu_system_wakeup_request wasn't reporting any. Prior to this patch, 
the function
would either change the runstate and notify the wake up event or fail 
silently.

The idea of the patch is to add the extra Error pointer into 
qemu_system_wakeup_request,
so qmp_system_wakeup can propagate the runstate_check error back to 
hmp_system_wakeup,
instead of duplicating this runstate verification inside 
qmp_system_wakeup (like I was
doing in some earlier version). With this idea in mind, passing NULL in 
the errp
of the remaining qemu_system_wakeup_request calls will not improve the 
existing usage
or fix potential bugs, sure, but doesn't make it worse either.

I don't see any problems with re-evaluating every existing 
qemu_system_wakeup_request
call and judge f qemu should error_abort out of it in case of error. 
It's just out of scope of
this patch series, IMO.


Thanks,


Daniel


>
> This makes NULL errp always look like a bug to be fixed.  If you
> are sure you really want to ignore an error, I'd recommend adding
> a comment making your intention explicit.
>

Re: [Qemu-devel] [PATCH for-3.2 v10 3/3] qmp hmp: Make system_wakeup check wake-up support and run state

[Markus Armbruster]

Daniel Henrique Barboza <danielhb413@gmail.com> writes:

> On 12/4/18 5:15 PM, Eduardo Habkost wrote:
>> On Tue, Dec 04, 2018 at 04:36:56PM -0200, Daniel Henrique Barboza wrote:
>>>
>>> On 11/29/18 4:55 PM, Markus Armbruster wrote:
>>>> One more thing...
>>>>
>>>> Markus Armbruster <armbru@redhat.com> writes:
>>>>
>>>>> Daniel Henrique Barboza <danielhb413@gmail.com> writes:
>>>>>
>>>>>> The qmp/hmp command 'system_wakeup' is simply a direct call to
>>>>>> 'qemu_system_wakeup_request' from vl.c. This function verifies if
>>>>>> runstate is SUSPENDED and if the wake up reason is valid before
>>>>>> proceeding. However, no error or warning is thrown if any of those
>>>>>> pre-requirements isn't met. There is no way for the caller to
>>>>>> differentiate between a successful wakeup or an error state caused
>>>>>> when trying to wake up a guest that wasn't suspended.
>>>>>>
>>>>>> This means that system_wakeup is silently failing, which can be
>>>>>> considered a bug. Adding error handling isn't an API break in this
>>>>>> case - applications that didn't check the result will remain broken,
>>>>>> the ones that check it will have a chance to deal with it.
>>>>>>
>>>>>> Adding to that, the commit before previous created a new QMP API called
>>>>>> query-current-machine, with a new flag called wakeup-suspend-support,
>>>>>> that indicates if the guest has the capability of waking up from suspended
>>>>>> state. Although such guest will never reach SUSPENDED state and erroring
>>>>>> it out in this scenario would suffice, it is more informative for the user
>>>>>> to differentiate between a failure because the guest isn't suspended versus
>>>>>> a failure because the guest does not have support for wake up at all.
>>>>>>
>>>>>> All this considered, this patch changes qmp_system_wakeup to check if
>>>>>> the guest is capable of waking up from suspend, and if it is suspended.
>>>>>> After this patch, this is the output of system_wakeup in a guest that
>>>>>> does not have wake-up from suspend support (ppc64):
>>>>>>
>>>>>> (qemu) system_wakeup
>>>>>> wake-up from suspend is not supported by this guest
>>>>>> (qemu)
>>>>>>
>>>>>> And this is the output of system_wakeup in a x86 guest that has the
>>>>>> support but isn't suspended:
>>>>>>
>>>>>> (qemu) system_wakeup
>>>>>> Unable to wake up: guest is not in suspended state
>>>>>> (qemu)
>>>>>>
>>>>>> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
>>>>>> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
>>>>>> ---
>>>>>>    hmp.c                   |  5 ++++-
>>>>>>    hw/acpi/core.c          |  4 +++-
>>>>>>    hw/char/serial.c        |  3 ++-
>>>>>>    hw/input/ps2.c          |  9 ++++++---
>>>>>>    hw/timer/mc146818rtc.c  |  3 ++-
>>>>>>    include/sysemu/sysemu.h |  3 ++-
>>>>>>    migration/migration.c   |  7 +++++--
>>>>>>    qapi/misc.json          |  8 +++++++-
>>>>>>    qmp.c                   | 13 ++++++++++++-
>>>>>>    vl.c                    |  6 ++++--
>>>>>>    10 files changed, 47 insertions(+), 14 deletions(-)
>>>>>>
>>>>>> diff --git a/hmp.c b/hmp.c
>>>>>> index 7828f93a39..0f5d943413 100644
>>>>>> --- a/hmp.c
>>>>>> +++ b/hmp.c
>>>>>> @@ -1220,7 +1220,10 @@ void hmp_cont(Monitor *mon, const QDict *qdict)
>>>>>>    void hmp_system_wakeup(Monitor *mon, const QDict *qdict)
>>>>>>    {
>>>>>> -    qmp_system_wakeup(NULL);
>>>>>> +    Error *err = NULL;
>>>>>> +
>>>>>> +    qmp_system_wakeup(&err);
>>>>>> +    hmp_handle_error(mon, &err);
>>>>>>    }
>>>>>>    void hmp_nmi(Monitor *mon, const QDict *qdict)
>>>>>> diff --git a/hw/acpi/core.c b/hw/acpi/core.c
>>>>>> index 52e18d7810..a7073dd435 100644
>>>>>> --- a/hw/acpi/core.c
>>>>>> +++ b/hw/acpi/core.c
>>>>>> @@ -514,7 +514,9 @@ static uint32_t acpi_pm_tmr_get(ACPIREGS *ar)
>>>>>>    static void acpi_pm_tmr_timer(void *opaque)
>>>>>>    {
>>>>>>        ACPIREGS *ar = opaque;
>>>>>> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER);
>>>>>> +    Error *err = NULL;
>>>>>> +
>>>>>> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER, &err);
>>>>>>        ar->tmr.update_sci(ar);
>>>>>>    }
>>>> Leaks the error object when qemu_system_wakeup_request() fails.
>>>>
>>>> If it cannot fail here, pass &error_abort.
>>>>
>>>> If it can fail, but you want to ignore failure, pass NULL.
>>> Good point. I'll simply pass NULL to all callers that didn't care
>>> for the error prior to this change.
>> Most times I saw QEMU code ignoring errors, it actually didn't
>> expect any errors to happen and was supposed to be using
>> &error_abort instead.
>>
>> This makes NULL errp always look like a bug to be fixed.  If you
>> are sure you really want to ignore an error, I'd recommend adding
>> a comment making your intention explicit.

I agree the "ignore errors" feature is widely abused.  I don't agree
with adding "I mean it" comments.  I think the proper remedy is to flag
abuses in patch review, and clean them up wherever we find them.

> In this particular case, the existing code wouldn't be expecting
> errors because
> qemu_system_wakeup_request wasn't reporting any. Prior to this patch,
> the function
> would either change the runstate and notify the wake up event or fail
> silently.
>
> The idea of the patch is to add the extra Error pointer into
> qemu_system_wakeup_request,
> so qmp_system_wakeup can propagate the runstate_check error back to
> hmp_system_wakeup,
> instead of duplicating this runstate verification inside
> qmp_system_wakeup (like I was
> doing in some earlier version). With this idea in mind, passing NULL
> in the errp
> of the remaining qemu_system_wakeup_request calls will not improve the
> existing usage
> or fix potential bugs, sure, but doesn't make it worse either.
>
> I don't see any problems with re-evaluating every existing
> qemu_system_wakeup_request
> call and judge f qemu should error_abort out of it in case of
> error. It's just out of scope of
> this patch series, IMO.

Okay.  I asked you to eliminate the code duplication, and I don't want
to punish your good deed there by demanding further cleanup.  That said,
further cleanup is always welcome :)

Re: [Qemu-devel] [PATCH for-3.2 v10 3/3] qmp hmp: Make system_wakeup check wake-up support and run state

[Daniel Henrique Barboza]

On 12/5/18 5:35 AM, Markus Armbruster wrote:
> Daniel Henrique Barboza <danielhb413@gmail.com> writes:
>
>> On 12/4/18 5:15 PM, Eduardo Habkost wrote:
>>> On Tue, Dec 04, 2018 at 04:36:56PM -0200, Daniel Henrique Barboza wrote:
>>>> On 11/29/18 4:55 PM, Markus Armbruster wrote:
>>>>> One more thing...
>>>>>
>>>>> Markus Armbruster <armbru@redhat.com> writes:
>>>>>
>>>>>> Daniel Henrique Barboza <danielhb413@gmail.com> writes:
>>>>>>
>>>>>>> The qmp/hmp command 'system_wakeup' is simply a direct call to
>>>>>>> 'qemu_system_wakeup_request' from vl.c. This function verifies if
>>>>>>> runstate is SUSPENDED and if the wake up reason is valid before
>>>>>>> proceeding. However, no error or warning is thrown if any of those
>>>>>>> pre-requirements isn't met. There is no way for the caller to
>>>>>>> differentiate between a successful wakeup or an error state caused
>>>>>>> when trying to wake up a guest that wasn't suspended.
>>>>>>>
>>>>>>> This means that system_wakeup is silently failing, which can be
>>>>>>> considered a bug. Adding error handling isn't an API break in this
>>>>>>> case - applications that didn't check the result will remain broken,
>>>>>>> the ones that check it will have a chance to deal with it.
>>>>>>>
>>>>>>> Adding to that, the commit before previous created a new QMP API called
>>>>>>> query-current-machine, with a new flag called wakeup-suspend-support,
>>>>>>> that indicates if the guest has the capability of waking up from suspended
>>>>>>> state. Although such guest will never reach SUSPENDED state and erroring
>>>>>>> it out in this scenario would suffice, it is more informative for the user
>>>>>>> to differentiate between a failure because the guest isn't suspended versus
>>>>>>> a failure because the guest does not have support for wake up at all.
>>>>>>>
>>>>>>> All this considered, this patch changes qmp_system_wakeup to check if
>>>>>>> the guest is capable of waking up from suspend, and if it is suspended.
>>>>>>> After this patch, this is the output of system_wakeup in a guest that
>>>>>>> does not have wake-up from suspend support (ppc64):
>>>>>>>
>>>>>>> (qemu) system_wakeup
>>>>>>> wake-up from suspend is not supported by this guest
>>>>>>> (qemu)
>>>>>>>
>>>>>>> And this is the output of system_wakeup in a x86 guest that has the
>>>>>>> support but isn't suspended:
>>>>>>>
>>>>>>> (qemu) system_wakeup
>>>>>>> Unable to wake up: guest is not in suspended state
>>>>>>> (qemu)
>>>>>>>
>>>>>>> Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
>>>>>>> Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
>>>>>>> ---
>>>>>>>     hmp.c                   |  5 ++++-
>>>>>>>     hw/acpi/core.c          |  4 +++-
>>>>>>>     hw/char/serial.c        |  3 ++-
>>>>>>>     hw/input/ps2.c          |  9 ++++++---
>>>>>>>     hw/timer/mc146818rtc.c  |  3 ++-
>>>>>>>     include/sysemu/sysemu.h |  3 ++-
>>>>>>>     migration/migration.c   |  7 +++++--
>>>>>>>     qapi/misc.json          |  8 +++++++-
>>>>>>>     qmp.c                   | 13 ++++++++++++-
>>>>>>>     vl.c                    |  6 ++++--
>>>>>>>     10 files changed, 47 insertions(+), 14 deletions(-)
>>>>>>>
>>>>>>> diff --git a/hmp.c b/hmp.c
>>>>>>> index 7828f93a39..0f5d943413 100644
>>>>>>> --- a/hmp.c
>>>>>>> +++ b/hmp.c
>>>>>>> @@ -1220,7 +1220,10 @@ void hmp_cont(Monitor *mon, const QDict *qdict)
>>>>>>>     void hmp_system_wakeup(Monitor *mon, const QDict *qdict)
>>>>>>>     {
>>>>>>> -    qmp_system_wakeup(NULL);
>>>>>>> +    Error *err = NULL;
>>>>>>> +
>>>>>>> +    qmp_system_wakeup(&err);
>>>>>>> +    hmp_handle_error(mon, &err);
>>>>>>>     }
>>>>>>>     void hmp_nmi(Monitor *mon, const QDict *qdict)
>>>>>>> diff --git a/hw/acpi/core.c b/hw/acpi/core.c
>>>>>>> index 52e18d7810..a7073dd435 100644
>>>>>>> --- a/hw/acpi/core.c
>>>>>>> +++ b/hw/acpi/core.c
>>>>>>> @@ -514,7 +514,9 @@ static uint32_t acpi_pm_tmr_get(ACPIREGS *ar)
>>>>>>>     static void acpi_pm_tmr_timer(void *opaque)
>>>>>>>     {
>>>>>>>         ACPIREGS *ar = opaque;
>>>>>>> -    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER);
>>>>>>> +    Error *err = NULL;
>>>>>>> +
>>>>>>> +    qemu_system_wakeup_request(QEMU_WAKEUP_REASON_PMTIMER, &err);
>>>>>>>         ar->tmr.update_sci(ar);
>>>>>>>     }
>>>>> Leaks the error object when qemu_system_wakeup_request() fails.
>>>>>
>>>>> If it cannot fail here, pass &error_abort.
>>>>>
>>>>> If it can fail, but you want to ignore failure, pass NULL.
>>>> Good point. I'll simply pass NULL to all callers that didn't care
>>>> for the error prior to this change.
>>> Most times I saw QEMU code ignoring errors, it actually didn't
>>> expect any errors to happen and was supposed to be using
>>> &error_abort instead.
>>>
>>> This makes NULL errp always look like a bug to be fixed.  If you
>>> are sure you really want to ignore an error, I'd recommend adding
>>> a comment making your intention explicit.
> I agree the "ignore errors" feature is widely abused.  I don't agree
> with adding "I mean it" comments.  I think the proper remedy is to flag
> abuses in patch review, and clean them up wherever we find them.
>
>> In this particular case, the existing code wouldn't be expecting
>> errors because
>> qemu_system_wakeup_request wasn't reporting any. Prior to this patch,
>> the function
>> would either change the runstate and notify the wake up event or fail
>> silently.
>>
>> The idea of the patch is to add the extra Error pointer into
>> qemu_system_wakeup_request,
>> so qmp_system_wakeup can propagate the runstate_check error back to
>> hmp_system_wakeup,
>> instead of duplicating this runstate verification inside
>> qmp_system_wakeup (like I was
>> doing in some earlier version). With this idea in mind, passing NULL
>> in the errp
>> of the remaining qemu_system_wakeup_request calls will not improve the
>> existing usage
>> or fix potential bugs, sure, but doesn't make it worse either.
>>
>> I don't see any problems with re-evaluating every existing
>> qemu_system_wakeup_request
>> call and judge f qemu should error_abort out of it in case of
>> error. It's just out of scope of
>> this patch series, IMO.
> Okay.  I asked you to eliminate the code duplication, and I don't want
> to punish your good deed there by demanding further cleanup.  That said,
> further cleanup is always welcome :)

I agree! Let's see if we can get this series pushed first, then we can
get back to all these new "ignore errors" we just added here.


Daniel