[Qemu-devel] [PATCH v8 0/5] cpus: Add generic "nmi" monitor command support

[Qemu-devel] [PATCH v8 0/5] cpus: Add generic "nmi" monitor command support

[Alexey Kardashevskiy]

This adds an "nmi" monitor command handler per CPUs.
x86, s390 and ppc CPUS are supported.

Please comment. Thanks.

Changes:
v8:
* adjusted commit log for s390 patch
* fixed missing "static" in 1/4

v7:
* fixed typenames, function names, copyrights
* s390x QOM'ed for later addition of the NMI interface
* Since none of x86 machines is QOM'ed, postpone migration to new interface for x86

v6:
* back 5 steps and make it an interface again

v5:
* added Error** to the callback
* fixed some comments

v4:
* now it is not nmi() but nmi_monitor_handler() to avoid confusion

v3:
* patches reorganized
* comments from v2 addressed, more details are in individual commit logs

v2:
* moved from machine interface to CPUClass callback
* s390 and x86 moved to target-s390/target-i386
* x86 handler delivers to the current CPU only now




Alexey Kardashevskiy (5):
  cpus: Define callback for QEMU "nmi" command
  s390x: Convert QEMUMachine to MachineClass
  s390x: Migrate to new NMI interface
  spapr: Add support for new NMI interface
  pc_piix: Migrate to new NMI interface

 cpus.c                     | 31 ++---------------
 hmp-commands.hx            |  6 ++--
 hw/core/Makefile.objs      |  1 +
 hw/core/nmi.c              | 84 ++++++++++++++++++++++++++++++++++++++++++++++
 hw/i386/pc_piix.c          | 42 +++++++++++++++++++++++
 hw/ppc/spapr.c             | 21 ++++++++++++
 hw/s390x/s390-virtio-ccw.c | 49 ++++++++++++++++++---------
 hw/s390x/s390-virtio.c     | 59 ++++++++++++++++++++++----------
 hw/s390x/s390-virtio.h     |  3 ++
 include/hw/nmi.h           | 49 +++++++++++++++++++++++++++
 qapi-schema.json           |  4 +--
 qmp-commands.hx            |  3 +-
 target-ppc/cpu-qom.h       |  1 +
 target-ppc/excp_helper.c   |  8 +++++
 14 files changed, 291 insertions(+), 70 deletions(-)
 create mode 100644 hw/core/nmi.c
 create mode 100644 include/hw/nmi.h

-- 
2.0.0

[Qemu-devel] [PATCH v8 1/5] cpus: Define callback for QEMU "nmi" command

[Alexey Kardashevskiy]

This introduces an NMI (Non Maskable Interrupt) interface with
a single nmi_monitor_handler() method. A machine or a device can
implement it. This searches for an QOM object with this interface
and if it is implemented, calls it. The callback implements an action
required to cause debug crash dump on in-kernel debugger invocation.
The callback returns Error**.

This adds a nmi_monitor_handle() helper which walks through
all objects to find the interface. The interface method is called
for all found instances.

This adds support for it in qmp_inject_nmi(). Since no architecture
supports it at the moment, there is no change in behaviour.

This changes inject-nmi command description for HMP and QMP.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Alexander Graf <agraf@suse.de>
---
Changes:
v8:
* added missing "static" to nmi_children() definition

v7:
* s/NMI/NMIState/
* added copyrighs to new files
* s/an Non/a Non/
* s/nmi()/nmi_monitor_handle()/

v6:
* NMI is an interface again

v5:
* s/given guest's (CPU|VCPU)/default CPU/
* nmi_monitor_handler() now returns Error**

v4:
* s/\<nmi\>/nmi_monitor_handler/

v3:
* actual nmi() enablement moved from last patch to first patch
* changed description for QMP command too
---
 cpus.c                |  3 +-
 hmp-commands.hx       |  6 ++--
 hw/core/Makefile.objs |  1 +
 hw/core/nmi.c         | 84 +++++++++++++++++++++++++++++++++++++++++++++++++++
 include/hw/nmi.h      | 49 ++++++++++++++++++++++++++++++
 qapi-schema.json      |  4 +--
 qmp-commands.hx       |  3 +-
 7 files changed, 141 insertions(+), 9 deletions(-)
 create mode 100644 hw/core/nmi.c
 create mode 100644 include/hw/nmi.h

diff --git a/cpus.c b/cpus.c
index 2b5c0bd..37d90f4 100644
--- a/cpus.c
+++ b/cpus.c
@@ -40,6 +40,7 @@
 #include "qemu/bitmap.h"
 #include "qemu/seqlock.h"
 #include "qapi-event.h"
+#include "hw/nmi.h"
 
 #ifndef _WIN32
 #include "qemu/compatfd.h"
@@ -1551,7 +1552,7 @@ void qmp_inject_nmi(Error **errp)
         }
     }
 #else
-    error_set(errp, QERR_UNSUPPORTED);
+    nmi_monitor_handle(monitor_get_cpu_index(), errp);
 #endif
 }
 
diff --git a/hmp-commands.hx b/hmp-commands.hx
index d0943b1..f859f8d 100644
--- a/hmp-commands.hx
+++ b/hmp-commands.hx
@@ -832,19 +832,17 @@ The values that can be specified here depend on the machine type, but are
 the same that can be specified in the @code{-boot} command line option.
 ETEXI
 
-#if defined(TARGET_I386) || defined(TARGET_S390X)
     {
         .name       = "nmi",
         .args_type  = "",
         .params     = "",
-        .help       = "inject an NMI on all guest's CPUs",
+        .help       = "inject an NMI",
         .mhandler.cmd = hmp_inject_nmi,
     },
-#endif
 STEXI
 @item nmi @var{cpu}
 @findex nmi
-Inject an NMI (x86) or RESTART (s390x) on the given CPU.
+Inject an NMI on the default CPU (x86/s390) or all CPUs (ppc64).
 
 ETEXI
 
diff --git a/hw/core/Makefile.objs b/hw/core/Makefile.objs
index 5377d05..17845df 100644
--- a/hw/core/Makefile.objs
+++ b/hw/core/Makefile.objs
@@ -4,6 +4,7 @@ common-obj-y += fw-path-provider.o
 # irq.o needed for qdev GPIO handling:
 common-obj-y += irq.o
 common-obj-y += hotplug.o
+common-obj-y += nmi.o
 
 common-obj-$(CONFIG_EMPTY_SLOT) += empty_slot.o
 common-obj-$(CONFIG_XILINX_AXI) += stream.o
diff --git a/hw/core/nmi.c b/hw/core/nmi.c
new file mode 100644
index 0000000..3dff020
--- /dev/null
+++ b/hw/core/nmi.c
@@ -0,0 +1,84 @@
+/*
+ *  NMI monitor handler class and helpers.
+ *
+ *  Copyright IBM Corp., 2014
+ *
+ *  Author: Alexey Kardashevskiy <aik@ozlabs.ru>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License,
+ *  or (at your option) any later version.
+ *
+ *  This program 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 General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include "hw/nmi.h"
+#include "qapi/qmp/qerror.h"
+
+struct do_nmi_s {
+    int cpu_index;
+    Error *errp;
+    bool handled;
+};
+
+static void nmi_children(Object *o, struct do_nmi_s *ns);
+
+static int do_nmi(Object *o, void *opaque)
+{
+    struct do_nmi_s *ns = opaque;
+    NMIState *n = (NMIState *) object_dynamic_cast(o, TYPE_NMI);
+
+    if (n) {
+        NMIClass *nc = NMI_GET_CLASS(n);
+
+        ns->handled = true;
+        nc->nmi_monitor_handler(n, ns->cpu_index, &ns->errp);
+        if (ns->errp) {
+            return -1;
+        }
+    }
+    nmi_children(o, ns);
+
+    return 0;
+}
+
+static void nmi_children(Object *o, struct do_nmi_s *ns)
+{
+    object_child_foreach(o, do_nmi, ns);
+}
+
+void nmi_monitor_handle(int cpu_index, Error **errp)
+{
+    struct do_nmi_s ns = {
+        .cpu_index = cpu_index,
+        .errp = NULL,
+        .handled = false
+    };
+
+    nmi_children(object_get_root(), &ns);
+    if (ns.handled) {
+        error_propagate(errp, ns.errp);
+    } else {
+        error_set(errp, QERR_UNSUPPORTED);
+    }
+}
+
+static const TypeInfo nmi_info = {
+    .name          = TYPE_NMI,
+    .parent        = TYPE_INTERFACE,
+    .class_size    = sizeof(NMIClass),
+};
+
+static void nmi_register_types(void)
+{
+    type_register_static(&nmi_info);
+}
+
+type_init(nmi_register_types)
diff --git a/include/hw/nmi.h b/include/hw/nmi.h
new file mode 100644
index 0000000..b541772
--- /dev/null
+++ b/include/hw/nmi.h
@@ -0,0 +1,49 @@
+/*
+ *  NMI monitor handler class and helpers definitions.
+ *
+ *  Copyright IBM Corp., 2014
+ *
+ *  Author: Alexey Kardashevskiy <aik@ozlabs.ru>
+ *
+ *  This program is free software; you can redistribute it and/or modify
+ *  it under the terms of the GNU General Public License as published by
+ *  the Free Software Foundation; either version 2 of the License,
+ *  or (at your option) any later version.
+ *
+ *  This program 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 General Public License for more details.
+ *
+ *  You should have received a copy of the GNU General Public License
+ *  along with this program; if not, see <http://www.gnu.org/licenses/>.
+ */
+
+#ifndef NMI_H
+#define NMI_H 1
+
+#include "qemu-common.h"
+#include "qom/object.h"
+
+#define TYPE_NMI "nmi"
+
+#define NMI_CLASS(klass) \
+     OBJECT_CLASS_CHECK(NMIClass, (klass), TYPE_NMI)
+#define NMI_GET_CLASS(obj) \
+    OBJECT_GET_CLASS(NMIClass, (obj), TYPE_NMI)
+#define NMI(obj) \
+     INTERFACE_CHECK(NMI, (obj), TYPE_NMI)
+
+typedef struct NMIState {
+    Object parent_obj;
+} NMIState;
+
+typedef struct NMIClass {
+    InterfaceClass parent_class;
+
+    void (*nmi_monitor_handler)(NMIState *n, int cpu_index, Error **errp);
+} NMIClass;
+
+void nmi_monitor_handle(int cpu_index, Error **errp);
+
+#endif /* NMI_H */
diff --git a/qapi-schema.json b/qapi-schema.json
index 341f417..689b548 100644
--- a/qapi-schema.json
+++ b/qapi-schema.json
@@ -1186,13 +1186,13 @@
 ##
 # @inject-nmi:
 #
-# Injects an Non-Maskable Interrupt into all guest's VCPUs.
+# Injects a Non-Maskable Interrupt into the default CPU (x86/s390) or all CPUs (ppc64).
 #
 # Returns:  If successful, nothing
 #
 # Since:  0.14.0
 #
-# Notes: Only x86 Virtual Machines support this command.
+# Note: prior to 2.1, this command was only supported for x86 and s390 VMs
 ##
 { 'command': 'inject-nmi' }
 
diff --git a/qmp-commands.hx b/qmp-commands.hx
index 4be4765..7658d4b 100644
--- a/qmp-commands.hx
+++ b/qmp-commands.hx
@@ -477,7 +477,7 @@ SQMP
 inject-nmi
 ----------
 
-Inject an NMI on guest's CPUs.
+Inject an NMI on the default CPU (x86/s390) or all CPUs (ppc64).
 
 Arguments: None.
 
@@ -487,7 +487,6 @@ Example:
 <- { "return": {} }
 
 Note: inject-nmi fails when the guest doesn't support injecting.
-      Currently, only x86 (NMI) and s390x (RESTART) guests do.
 
 EQMP
 
-- 
2.0.0

[Qemu-devel] [PATCH v8 2/5] s390x: Convert QEMUMachine to MachineClass

[Alexey Kardashevskiy]

This converts s390-virtio and s390-ccw-virtio machines to QOM MachineClass.
This brings ability to add interfaces to the machine classes. The first
interface for addition will be NMI.

The patch is mechanical so no change in behavior is expected.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Alexander Graf <agraf@suse.de>
---
 hw/s390x/s390-virtio-ccw.c | 43 +++++++++++++++++++++++++++----------------
 hw/s390x/s390-virtio.c     | 44 +++++++++++++++++++++++++++-----------------
 2 files changed, 54 insertions(+), 33 deletions(-)

diff --git a/hw/s390x/s390-virtio-ccw.c b/hw/s390x/s390-virtio-ccw.c
index 42f5cec..05311b8 100644
--- a/hw/s390x/s390-virtio-ccw.c
+++ b/hw/s390x/s390-virtio-ccw.c
@@ -18,6 +18,8 @@
 #include "css.h"
 #include "virtio-ccw.h"
 
+#define TYPE_S390_CCW_MACHINE               "s390-ccw-machine"
+
 void io_subsystem_reset(void)
 {
     DeviceState *css, *sclp, *flic;
@@ -134,24 +136,33 @@ static void ccw_init(MachineState *machine)
     s390_create_virtio_net(BUS(css_bus), "virtio-net-ccw");
 }
 
-static QEMUMachine ccw_machine = {
-    .name = "s390-ccw-virtio",
-    .alias = "s390-ccw",
-    .desc = "VirtIO-ccw based S390 machine",
-    .init = ccw_init,
-    .block_default_type = IF_VIRTIO,
-    .no_cdrom = 1,
-    .no_floppy = 1,
-    .no_serial = 1,
-    .no_parallel = 1,
-    .no_sdcard = 1,
-    .use_sclp = 1,
-    .max_cpus = 255,
+static void ccw_machine_class_init(ObjectClass *oc, void *data)
+{
+    MachineClass *mc = MACHINE_CLASS(oc);
+
+    mc->name = "s390-ccw-virtio";
+    mc->alias = "s390-ccw";
+    mc->desc = "VirtIO-ccw based S390 machine";
+    mc->init = ccw_init;
+    mc->block_default_type = IF_VIRTIO;
+    mc->no_cdrom = 1;
+    mc->no_floppy = 1;
+    mc->no_serial = 1;
+    mc->no_parallel = 1;
+    mc->no_sdcard = 1;
+    mc->use_sclp = 1,
+    mc->max_cpus = 255;
+}
+
+static const TypeInfo ccw_machine_info = {
+    .name          = TYPE_S390_CCW_MACHINE,
+    .parent        = TYPE_MACHINE,
+    .class_init    = ccw_machine_class_init,
 };
 
-static void ccw_machine_init(void)
+static void ccw_machine_register_types(void)
 {
-    qemu_register_machine(&ccw_machine);
+    type_register_static(&ccw_machine_info);
 }
 
-machine_init(ccw_machine_init)
+type_init(ccw_machine_register_types)
diff --git a/hw/s390x/s390-virtio.c b/hw/s390x/s390-virtio.c
index 93c7ace..f1e0dbc 100644
--- a/hw/s390x/s390-virtio.c
+++ b/hw/s390x/s390-virtio.c
@@ -51,6 +51,7 @@
 
 #define MAX_BLK_DEVS                    10
 #define ZIPL_FILENAME                   "s390-zipl.rom"
+#define TYPE_S390_MACHINE               "s390-machine"
 
 static VirtIOS390Bus *s390_bus;
 static S390CPU **ipi_states;
@@ -279,25 +280,34 @@ static void s390_init(MachineState *machine)
     s390_create_virtio_net((BusState *)s390_bus, "virtio-net-s390");
 }
 
-static QEMUMachine s390_machine = {
-    .name = "s390-virtio",
-    .alias = "s390",
-    .desc = "VirtIO based S390 machine",
-    .init = s390_init,
-    .block_default_type = IF_VIRTIO,
-    .no_cdrom = 1,
-    .no_floppy = 1,
-    .no_serial = 1,
-    .no_parallel = 1,
-    .no_sdcard = 1,
-    .use_virtcon = 1,
-    .max_cpus = 255,
-    .is_default = 1,
+static void s390_machine_class_init(ObjectClass *oc, void *data)
+{
+    MachineClass *mc = MACHINE_CLASS(oc);
+
+    mc->name = "s390-virtio";
+    mc->alias = "s390";
+    mc->desc = "VirtIO based S390 machine";
+    mc->init = s390_init;
+    mc->block_default_type = IF_VIRTIO;
+    mc->max_cpus = 255;
+    mc->no_serial = 1;
+    mc->no_parallel = 1;
+    mc->use_virtcon = 1;
+    mc->no_floppy = 1;
+    mc->no_cdrom = 1;
+    mc->no_sdcard = 1;
+    mc->is_default = 1;
+}
+
+static const TypeInfo s390_machine_info = {
+    .name          = TYPE_S390_MACHINE,
+    .parent        = TYPE_MACHINE,
+    .class_init    = s390_machine_class_init,
 };
 
-static void s390_machine_init(void)
+static void s390_machine_register_types(void)
 {
-    qemu_register_machine(&s390_machine);
+    type_register_static(&s390_machine_info);
 }
 
-machine_init(s390_machine_init);
+type_init(s390_machine_register_types)
-- 
2.0.0

[Qemu-devel] [PATCH v8 3/5] s390x: Migrate to new NMI interface

[Alexey Kardashevskiy]

This implements an NMI interface for s390 and s390-ccw machines.

This removes #ifdef s390 branch in qmp_inject_nmi so new s390's
nmi_monitor_handler() callback is going to be used for NMI.

Since nmi_monitor_handler()-calling code is platform independent,
CPUState::cpu_index is used instead of S390CPU::env.cpu_num.
There should not be any change in behaviour as both @cpu_index and
@cpu_num are global CPU numbers.

Note that s390_cpu_restart() already takes care of the specified cpu,
so we don't need to schedule via async_run_on_cpu().

Since the only error s390_cpu_restart() can return is ENOSYS, convert
it to QERR_UNSUPPORTED.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
---
Changes:
v8:
* adjusted commit log

v7:
* since now s390' are QOM'ed, add interface to them and do not create
new object

v6:
* supported NMI interface

v5:
* added ENOSYS -> QERR_UNSUPPORTED, qapi/qmp/qerror.h was added for this

v4:
* s/\<nmi\>/nmi_monitor_handler/

v3:
* now contains both old code removal and new code insertion, easier to
track changes
---
 cpus.c                     | 14 --------------
 hw/s390x/s390-virtio-ccw.c |  6 ++++++
 hw/s390x/s390-virtio.c     | 15 +++++++++++++++
 hw/s390x/s390-virtio.h     |  3 +++
 4 files changed, 24 insertions(+), 14 deletions(-)

diff --git a/cpus.c b/cpus.c
index 37d90f4..eb1ac85 100644
--- a/cpus.c
+++ b/cpus.c
@@ -1537,20 +1537,6 @@ void qmp_inject_nmi(Error **errp)
             apic_deliver_nmi(cpu->apic_state);
         }
     }
-#elif defined(TARGET_S390X)
-    CPUState *cs;
-    S390CPU *cpu;
-
-    CPU_FOREACH(cs) {
-        cpu = S390_CPU(cs);
-        if (cpu->env.cpu_num == monitor_get_cpu_index()) {
-            if (s390_cpu_restart(S390_CPU(cs)) == -1) {
-                error_set(errp, QERR_UNSUPPORTED);
-                return;
-            }
-            break;
-        }
-    }
 #else
     nmi_monitor_handle(monitor_get_cpu_index(), errp);
 #endif
diff --git a/hw/s390x/s390-virtio-ccw.c b/hw/s390x/s390-virtio-ccw.c
index 05311b8..004b2c2 100644
--- a/hw/s390x/s390-virtio-ccw.c
+++ b/hw/s390x/s390-virtio-ccw.c
@@ -139,6 +139,7 @@ static void ccw_init(MachineState *machine)
 static void ccw_machine_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    NMIClass *nc = NMI_CLASS(oc);
 
     mc->name = "s390-ccw-virtio";
     mc->alias = "s390-ccw";
@@ -152,12 +153,17 @@ static void ccw_machine_class_init(ObjectClass *oc, void *data)
     mc->no_sdcard = 1;
     mc->use_sclp = 1,
     mc->max_cpus = 255;
+    nc->nmi_monitor_handler = s390_nmi;
 }
 
 static const TypeInfo ccw_machine_info = {
     .name          = TYPE_S390_CCW_MACHINE,
     .parent        = TYPE_MACHINE,
     .class_init    = ccw_machine_class_init,
+    .interfaces = (InterfaceInfo[]) {
+        { TYPE_NMI },
+        { }
+    },
 };
 
 static void ccw_machine_register_types(void)
diff --git a/hw/s390x/s390-virtio.c b/hw/s390x/s390-virtio.c
index f1e0dbc..1a75a1c 100644
--- a/hw/s390x/s390-virtio.c
+++ b/hw/s390x/s390-virtio.c
@@ -280,9 +280,19 @@ static void s390_init(MachineState *machine)
     s390_create_virtio_net((BusState *)s390_bus, "virtio-net-s390");
 }
 
+void s390_nmi(NMIState *n, int cpu_index, Error **errp)
+{
+    CPUState *cs = qemu_get_cpu(cpu_index);
+
+    if (s390_cpu_restart(S390_CPU(cs))) {
+        error_set(errp, QERR_UNSUPPORTED);
+    }
+}
+
 static void s390_machine_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
+    NMIClass *nc = NMI_CLASS(oc);
 
     mc->name = "s390-virtio";
     mc->alias = "s390";
@@ -297,12 +307,17 @@ static void s390_machine_class_init(ObjectClass *oc, void *data)
     mc->no_cdrom = 1;
     mc->no_sdcard = 1;
     mc->is_default = 1;
+    nc->nmi_monitor_handler = s390_nmi;
 }
 
 static const TypeInfo s390_machine_info = {
     .name          = TYPE_S390_MACHINE,
     .parent        = TYPE_MACHINE,
     .class_init    = s390_machine_class_init,
+    .interfaces = (InterfaceInfo[]) {
+        { TYPE_NMI },
+        { }
+    },
 };
 
 static void s390_machine_register_types(void)
diff --git a/hw/s390x/s390-virtio.h b/hw/s390x/s390-virtio.h
index 5c405e7..33847ae 100644
--- a/hw/s390x/s390-virtio.h
+++ b/hw/s390x/s390-virtio.h
@@ -12,6 +12,8 @@
 #ifndef HW_S390_VIRTIO_H
 #define HW_S390_VIRTIO_H 1
 
+#include "hw/nmi.h"
+
 #define KVM_S390_VIRTIO_NOTIFY          0
 #define KVM_S390_VIRTIO_RESET           1
 #define KVM_S390_VIRTIO_SET_STATUS      2
@@ -26,4 +28,5 @@ void s390_init_ipl_dev(const char *kernel_filename,
                        const char *initrd_filename,
                        const char *firmware);
 void s390_create_virtio_net(BusState *bus, const char *name);
+void s390_nmi(NMIState *n, int cpu_index, Error **errp);
 #endif
-- 
2.0.0

[Qemu-devel] [PATCH v8 4/5] spapr: Add support for new NMI interface

[Alexey Kardashevskiy]

This implements an NMI interface POWERPC SPAPR machine.
This enables an "nmi" HMP/QMP command supported on SPAPR.

This calls POWERPC_EXCP_RESET (vector 0x100) in the guest to deliver NMI
to every CPU. The expected result is XMON (in-kernel debugger) invocation.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Alexander Graf <agraf@suse.de>
---
Changes:
v7:
* s/ppc_cpu_do_nmi/ppc_cpu_do_system_reset/

v6:
* support NMI interface

v4:
* s/\<nmi\>/nmi_monitor_handler/
* added note about XMON into commit log

v3:
* ppc_cpu_do_nmi() is exported from excp_helper.c instead of powerpc_excp()
---
 hw/ppc/spapr.c           | 21 +++++++++++++++++++++
 target-ppc/cpu-qom.h     |  1 +
 target-ppc/excp_helper.c |  8 ++++++++
 3 files changed, 30 insertions(+)

diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index 5c92707..c619c3b 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -55,6 +55,7 @@
 #include "qemu/config-file.h"
 #include "qemu/error-report.h"
 #include "trace.h"
+#include "hw/nmi.h"
 
 #include <libfdt.h>
 
@@ -1639,10 +1640,28 @@ static void spapr_machine_initfn(Object *obj)
                             spapr_get_kvm_type, spapr_set_kvm_type, NULL);
 }
 
+static void ppc_cpu_do_nmi_on_cpu(void *arg)
+{
+    CPUState *cs = arg;
+
+    cpu_synchronize_state(cs);
+    ppc_cpu_do_system_reset(cs);
+}
+
+static void spapr_nmi(NMIState *n, int cpu_index, Error **errp)
+{
+    CPUState *cs;
+
+    CPU_FOREACH(cs) {
+        async_run_on_cpu(cs, ppc_cpu_do_nmi_on_cpu, cs);
+    }
+}
+
 static void spapr_machine_class_init(ObjectClass *oc, void *data)
 {
     MachineClass *mc = MACHINE_CLASS(oc);
     FWPathProviderClass *fwc = FW_PATH_PROVIDER_CLASS(oc);
+    NMIClass *nc = NMI_CLASS(oc);
 
     mc->name = "pseries";
     mc->desc = "pSeries Logical Partition (PAPR compliant)";
@@ -1656,6 +1675,7 @@ static void spapr_machine_class_init(ObjectClass *oc, void *data)
     mc->kvm_type = spapr_kvm_type;
 
     fwc->get_dev_path = spapr_get_fw_dev_path;
+    nc->nmi_monitor_handler = spapr_nmi;
 }
 
 static const TypeInfo spapr_machine_info = {
@@ -1666,6 +1686,7 @@ static const TypeInfo spapr_machine_info = {
     .class_init    = spapr_machine_class_init,
     .interfaces = (InterfaceInfo[]) {
         { TYPE_FW_PATH_PROVIDER },
+        { TYPE_NMI },
         { }
     },
 };
diff --git a/target-ppc/cpu-qom.h b/target-ppc/cpu-qom.h
index 0fee36f..a379f79 100644
--- a/target-ppc/cpu-qom.h
+++ b/target-ppc/cpu-qom.h
@@ -127,6 +127,7 @@ int ppc64_cpu_write_elf64_qemunote(WriteCoreDumpFunction f,
 int ppc64_cpu_write_elf64_note(WriteCoreDumpFunction f, CPUState *cs,
                                int cpuid, void *opaque);
 #ifndef CONFIG_USER_ONLY
+void ppc_cpu_do_system_reset(CPUState *cs);
 extern const struct VMStateDescription vmstate_ppc_cpu;
 
 typedef struct PPCTimebase {
diff --git a/target-ppc/excp_helper.c b/target-ppc/excp_helper.c
index be71590..922e86d 100644
--- a/target-ppc/excp_helper.c
+++ b/target-ppc/excp_helper.c
@@ -810,6 +810,14 @@ void ppc_hw_interrupt(CPUPPCState *env)
         }
     }
 }
+
+void ppc_cpu_do_system_reset(CPUState *cs)
+{
+    PowerPCCPU *cpu = POWERPC_CPU(cs);
+    CPUPPCState *env = &cpu->env;
+
+    powerpc_excp(cpu, env->excp_model, POWERPC_EXCP_RESET);
+}
 #endif /* !CONFIG_USER_ONLY */
 
 #if defined(DEBUG_OP)
-- 
2.0.0

[Qemu-devel] [PATCH v8 5/5] pc_piix: Migrate to new NMI interface

[Alexey Kardashevskiy]

This implements an NMI interface for i386 PC machines.

This removes #ifdef I386 branch in qmp_inject_nmi so new i386's nmi()
callback is going to be used for NMI.

This changes code to inject NMI on the current CPU instead of injecting
it on every CPU. However that does not seem to be an issue.

Since kvm_apic_external_nmi() takes care of preforming operations in
the specific CPU thread so no extra measure is required here.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Alexander Graf <agraf@suse.de>
---
Changes:
v6:
* make use of NMI interface

v5:
* make use of NMI interface

v4:
* s/\<nmi\>/nmi_monitor_handler/

v3:
* now contains both old code removal and new code insertion, easier to
track changes
* fixed compile for linux-user
---
 cpus.c            | 14 --------------
 hw/i386/pc_piix.c | 42 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 42 insertions(+), 14 deletions(-)

diff --git a/cpus.c b/cpus.c
index eb1ac85..f57fe5f 100644
--- a/cpus.c
+++ b/cpus.c
@@ -1525,21 +1525,7 @@ exit:
 
 void qmp_inject_nmi(Error **errp)
 {
-#if defined(TARGET_I386)
-    CPUState *cs;
-
-    CPU_FOREACH(cs) {
-        X86CPU *cpu = X86_CPU(cs);
-
-        if (!cpu->apic_state) {
-            cpu_interrupt(cs, CPU_INTERRUPT_NMI);
-        } else {
-            apic_deliver_nmi(cpu->apic_state);
-        }
-    }
-#else
     nmi_monitor_handle(monitor_get_cpu_index(), errp);
-#endif
 }
 
 void dump_drift_info(FILE *f, fprintf_function cpu_fprintf)
diff --git a/hw/i386/pc_piix.c b/hw/i386/pc_piix.c
index 4f22be8..125e609 100644
--- a/hw/i386/pc_piix.c
+++ b/hw/i386/pc_piix.c
@@ -49,11 +49,13 @@
 #include "hw/acpi/acpi.h"
 #include "cpu.h"
 #include "qemu/error-report.h"
+#include "hw/nmi.h"
 #ifdef CONFIG_XEN
 #  include <xen/hvm/hvm_info_table.h>
 #endif
 
 #define MAX_IDE_BUS 2
+#define TYPE_NMI_X86        "x86-nmi"
 
 static const int ide_iobase[MAX_IDE_BUS] = { 0x1f0, 0x170 };
 static const int ide_iobase2[MAX_IDE_BUS] = { 0x3f6, 0x376 };
@@ -297,6 +299,9 @@ static void pc_init1(MachineState *machine,
     if (pci_enabled) {
         pc_pci_device_init(pci_bus);
     }
+
+    object_property_add_child(OBJECT(machine), "nmi",
+                              object_new(TYPE_NMI_X86), NULL);
 }
 
 static void pc_init_pci(MachineState *machine)
@@ -926,3 +931,40 @@ static void pc_machine_init(void)
 }
 
 machine_init(pc_machine_init);
+
+static void x86_nmi(NMIState *n, int cpu_index, Error **errp)
+{
+    CPUState *cs = qemu_get_cpu(cpu_index);
+    X86CPU *cpu = X86_CPU(cs);
+
+    if (!cpu->apic_state) {
+        cpu_interrupt(cs, CPU_INTERRUPT_NMI);
+#ifndef CONFIG_USER_ONLY
+    } else {
+        apic_deliver_nmi(cpu->apic_state);
+#endif
+    }
+}
+
+static void x86_nmi_class_init(ObjectClass *oc, void *data)
+{
+    NMIClass *nc = NMI_CLASS(oc);
+    nc->nmi_monitor_handler = x86_nmi;
+}
+
+static const TypeInfo x86_nmi_info = {
+    .name          = TYPE_NMI_X86,
+    .parent        = TYPE_OBJECT,
+    .class_init    = x86_nmi_class_init,
+    .interfaces = (InterfaceInfo[]) {
+        { TYPE_NMI },
+        { }
+    },
+};
+
+static void x86_nmi_register_types(void)
+{
+    type_register_static(&x86_nmi_info);
+}
+
+type_init(x86_nmi_register_types)
-- 
2.0.0

Re: [Qemu-devel] [PATCH v8 0/5] cpus: Add generic "nmi" monitor command support

[Paolo Bonzini]

Il 20/08/2014 14:16, Alexey Kardashevskiy ha scritto:
> This adds an "nmi" monitor command handler per CPUs.
> x86, s390 and ppc CPUS are supported.
> 
> Please comment. Thanks.
> 
> Changes:
> v8:
> * adjusted commit log for s390 patch
> * fixed missing "static" in 1/4
> 
> v7:
> * fixed typenames, function names, copyrights
> * s390x QOM'ed for later addition of the NMI interface
> * Since none of x86 machines is QOM'ed, postpone migration to new interface for x86
> 
> v6:
> * back 5 steps and make it an interface again
> 
> v5:
> * added Error** to the callback
> * fixed some comments
> 
> v4:
> * now it is not nmi() but nmi_monitor_handler() to avoid confusion
> 
> v3:
> * patches reorganized
> * comments from v2 addressed, more details are in individual commit logs
> 
> v2:
> * moved from machine interface to CPUClass callback
> * s390 and x86 moved to target-s390/target-i386
> * x86 handler delivers to the current CPU only now
> 
> 
> 
> 
> Alexey Kardashevskiy (5):
>   cpus: Define callback for QEMU "nmi" command
>   s390x: Convert QEMUMachine to MachineClass
>   s390x: Migrate to new NMI interface
>   spapr: Add support for new NMI interface
>   pc_piix: Migrate to new NMI interface
> 
>  cpus.c                     | 31 ++---------------
>  hmp-commands.hx            |  6 ++--
>  hw/core/Makefile.objs      |  1 +
>  hw/core/nmi.c              | 84 ++++++++++++++++++++++++++++++++++++++++++++++
>  hw/i386/pc_piix.c          | 42 +++++++++++++++++++++++
>  hw/ppc/spapr.c             | 21 ++++++++++++
>  hw/s390x/s390-virtio-ccw.c | 49 ++++++++++++++++++---------
>  hw/s390x/s390-virtio.c     | 59 ++++++++++++++++++++++----------
>  hw/s390x/s390-virtio.h     |  3 ++
>  include/hw/nmi.h           | 49 +++++++++++++++++++++++++++
>  qapi-schema.json           |  4 +--
>  qmp-commands.hx            |  3 +-
>  target-ppc/cpu-qom.h       |  1 +
>  target-ppc/excp_helper.c   |  8 +++++
>  14 files changed, 291 insertions(+), 70 deletions(-)
>  create mode 100644 hw/core/nmi.c
>  create mode 100644 include/hw/nmi.h
> 

Applying patches 1-4 to uq/master.  PC will wait for the QOM machine
conversion.

Paolo