[PATCH 0/4] hw/nmi: Safer vCPU delivery

[PATCH 0/4] hw/nmi: Safer vCPU delivery

[Philippe Mathieu-Daudé]

Hi Marc-André,

I don't remember what I suggested particularly for
20260626-qemu-no-hmp-v2-7-8af31bc54c61@redhat.com,
but this series is what I had in mind back then.

If you find it worthy, please consider including it
in your "Make HMP optional" work.

Regard,

Phil.

Philippe Mathieu-Daudé (4):
  hw/nmi: Allow delivering to first available vCPU (not by index)
  hw/nmi: Introduce nmi_cpu_handler() for vcpu-specific delivery
  hw/s390x: Deliver per-vcpu NMI using nmi_cpu_handler
  hw/nmi: Remove unused @cpu_index and @errp arguments

 include/hw/core/nmi.h      | 10 +++++++++-
 hw/core/nmi.c              | 25 ++++++++++++++++++++++---
 hw/hppa/machine.c          |  4 ++--
 hw/i386/x86.c              |  5 ++---
 hw/intc/m68k_irqc.c        |  4 ++--
 hw/ipmi/ipmi.c             |  2 +-
 hw/m68k/q800-glue.c        |  4 ++--
 hw/misc/macio/gpio.c       |  4 ++--
 hw/ppc/pnv.c               |  7 +++----
 hw/ppc/spapr.c             |  4 ++--
 hw/s390x/s390-virtio-ccw.c |  6 ++----
 hw/watchdog/watchdog.c     |  2 +-
 12 files changed, 50 insertions(+), 27 deletions(-)

-- 
2.53.0

[PATCH 1/4] hw/nmi: Allow delivering to first available vCPU (not by index)

[Philippe Mathieu-Daudé]

Allow callers to deliver on any vCPU (for example when
the vCPU index is not known).

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
---
 include/hw/core/nmi.h  | 7 +++++++
 hw/core/nmi.c          | 5 +++++
 hw/ipmi/ipmi.c         | 2 +-
 hw/watchdog/watchdog.c | 2 +-
 4 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/include/hw/core/nmi.h b/include/hw/core/nmi.h
index fff41bebc69..6136290f4fe 100644
--- a/include/hw/core/nmi.h
+++ b/include/hw/core/nmi.h
@@ -40,6 +40,13 @@ struct NMIClass {
     void (*nmi_monitor_handler)(NMIState *n, int cpu_index, Error **errp);
 };
 
+/**
+ * nmi_monitor_handle:
+ * @cpu_index: Index of the CPU to deliver the NMI
+ * @errp: pointer to a NULL-initialized error object
+ *
+ * Note: cpu_index = -1 can be used to deliver to the first vCPU available.
+ */
 void nmi_monitor_handle(int cpu_index, Error **errp);
 
 #endif /* NMI_H */
diff --git a/hw/core/nmi.c b/hw/core/nmi.c
index 4b447e126ba..67213631c04 100644
--- a/hw/core/nmi.c
+++ b/hw/core/nmi.c
@@ -20,6 +20,7 @@
  */
 
 #include "qemu/osdep.h"
+#include "hw/core/cpu.h"
 #include "hw/core/nmi.h"
 #include "qapi/error.h"
 #include "qemu/module.h"
@@ -42,6 +43,10 @@ static int do_nmi(Object *o, void *opaque)
         NMIClass *nc = NMI_GET_CLASS(n);
 
         ns->handled = true;
+        if (ns->cpu_index == -1) {
+            /* Any vCPU is OK, take the first one */
+            ns->cpu_index = first_cpu->cpu_index;
+        }
         nc->nmi_monitor_handler(n, ns->cpu_index, &ns->err);
         if (ns->err) {
             return -1;
diff --git a/hw/ipmi/ipmi.c b/hw/ipmi/ipmi.c
index 74818ff3cea..8c492271f90 100644
--- a/hw/ipmi/ipmi.c
+++ b/hw/ipmi/ipmi.c
@@ -60,7 +60,7 @@ static int ipmi_do_hw_op(IPMIInterface *s, enum ipmi_op op, int checkonly)
             return 0;
         }
         /* We don't care what CPU we use. */
-        nmi_monitor_handle(0, NULL);
+        nmi_monitor_handle(-1, NULL);
         return 0;
 
     case IPMI_SHUTDOWN_VIA_ACPI_OVERTEMP:
diff --git a/hw/watchdog/watchdog.c b/hw/watchdog/watchdog.c
index 0842fe373ae..0c1b94a62f3 100644
--- a/hw/watchdog/watchdog.c
+++ b/hw/watchdog/watchdog.c
@@ -81,7 +81,7 @@ void watchdog_perform_action(void)
 
     case WATCHDOG_ACTION_INJECT_NMI:
         qapi_event_send_watchdog(WATCHDOG_ACTION_INJECT_NMI);
-        nmi_monitor_handle(0, NULL);
+        nmi_monitor_handle(-1, NULL);
         break;
 
     default:
-- 
2.53.0

[PATCH 2/4] hw/nmi: Introduce nmi_cpu_handler() for vcpu-specific delivery

[Philippe Mathieu-Daudé]

No need to have handlers check for existing CPU index,
lookup the CPUState once in the core do_nmi() code and
pass along.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
---
 include/hw/core/nmi.h |  1 +
 hw/core/nmi.c         | 29 +++++++++++++++++++++++------
 2 files changed, 24 insertions(+), 6 deletions(-)

diff --git a/include/hw/core/nmi.h b/include/hw/core/nmi.h
index 6136290f4fe..2f554d58b7f 100644
--- a/include/hw/core/nmi.h
+++ b/include/hw/core/nmi.h
@@ -38,6 +38,7 @@ struct NMIClass {
     InterfaceClass parent_class;
 
     void (*nmi_monitor_handler)(NMIState *n, int cpu_index, Error **errp);
+    void (*nmi_cpu_handler)(NMIState *ns, CPUState *cs);
 };
 
 /**
diff --git a/hw/core/nmi.c b/hw/core/nmi.c
index 67213631c04..fb623fd00be 100644
--- a/hw/core/nmi.c
+++ b/hw/core/nmi.c
@@ -43,13 +43,30 @@ static int do_nmi(Object *o, void *opaque)
         NMIClass *nc = NMI_GET_CLASS(n);
 
         ns->handled = true;
-        if (ns->cpu_index == -1) {
-            /* Any vCPU is OK, take the first one */
-            ns->cpu_index = first_cpu->cpu_index;
+
+        /* Generic handler */
+        if (nc->nmi_monitor_handler) {
+            nc->nmi_monitor_handler(n, ns->cpu_index, &ns->err);
+            if (ns->err) {
+                return -1;
+            }
         }
-        nc->nmi_monitor_handler(n, ns->cpu_index, &ns->err);
-        if (ns->err) {
-            return -1;
+        /* Per vCPU handler */
+        if (nc->nmi_cpu_handler) {
+            CPUState *cs;
+
+            if (ns->cpu_index >= 0) {
+                /* Specific vCPU requested */
+                cs = qemu_get_cpu(ns->cpu_index);
+            } else {
+                /* Any vCPU is OK, take the first one */
+                cs = first_cpu;
+            }
+            if (!cs) {
+                error_setg(&ns->err, "CPU %d does not exist", ns->cpu_index);
+                return -1;
+            }
+            nc->nmi_cpu_handler(n, cs);
         }
     }
     nmi_children(o, ns);
-- 
2.53.0

[PATCH 3/4] hw/s390x: Deliver per-vcpu NMI using nmi_cpu_handler

[Philippe Mathieu-Daudé]

NMIClass::nmi_cpu_handler() handler is safer because
the CPUState argument has been checked to be valid.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
---
 hw/s390x/s390-virtio-ccw.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/hw/s390x/s390-virtio-ccw.c b/hw/s390x/s390-virtio-ccw.c
index 25a9fa49554..b71086c254a 100644
--- a/hw/s390x/s390-virtio-ccw.c
+++ b/hw/s390x/s390-virtio-ccw.c
@@ -725,10 +725,8 @@ static HotplugHandler *s390_get_hotplug_handler(MachineState *machine,
     return NULL;
 }
 
-static void s390_nmi(NMIState *n, int cpu_index, Error **errp)
+static void s390_nmi(NMIState *ns, CPUState *cs)
 {
-    CPUState *cs = qemu_get_cpu(cpu_index);
-
     s390_cpu_restart(S390_CPU(cs));
 }
 
@@ -832,7 +830,7 @@ static void ccw_machine_class_init(ObjectClass *oc, const void *data)
     hc->plug = s390_machine_device_plug;
     hc->unplug_request = s390_machine_device_unplug_request;
     hc->unplug = s390_machine_device_unplug;
-    nc->nmi_monitor_handler = s390_nmi;
+    nc->nmi_cpu_handler = s390_nmi;
     mc->default_ram_id = "s390.ram";
     mc->default_nic = "virtio-net-ccw";
     dsi->qmp_dump_skeys = s390_qmp_dump_skeys;
-- 
2.53.0

[PATCH 4/4] hw/nmi: Remove unused @cpu_index and @errp arguments

[Philippe Mathieu-Daudé]

NMI delivery never fails: remove the unused @errp argument.
nmi_handler() neither use the @cpu_index argument, remove it.

Signed-off-by: Philippe Mathieu-Daudé <philmd@oss.qualcomm.com>
---
 include/hw/core/nmi.h | 2 +-
 hw/core/nmi.c         | 7 ++-----
 hw/hppa/machine.c     | 4 ++--
 hw/i386/x86.c         | 5 ++---
 hw/intc/m68k_irqc.c   | 4 ++--
 hw/m68k/q800-glue.c   | 4 ++--
 hw/misc/macio/gpio.c  | 4 ++--
 hw/ppc/pnv.c          | 7 +++----
 hw/ppc/spapr.c        | 4 ++--
 9 files changed, 18 insertions(+), 23 deletions(-)

diff --git a/include/hw/core/nmi.h b/include/hw/core/nmi.h
index 2f554d58b7f..476b5d02fa1 100644
--- a/include/hw/core/nmi.h
+++ b/include/hw/core/nmi.h
@@ -37,7 +37,7 @@ typedef struct NMIState NMIState;
 struct NMIClass {
     InterfaceClass parent_class;
 
-    void (*nmi_monitor_handler)(NMIState *n, int cpu_index, Error **errp);
+    void (*nmi_handler)(NMIState *ns);
     void (*nmi_cpu_handler)(NMIState *ns, CPUState *cs);
 };
 
diff --git a/hw/core/nmi.c b/hw/core/nmi.c
index fb623fd00be..3f4ef5b2094 100644
--- a/hw/core/nmi.c
+++ b/hw/core/nmi.c
@@ -45,11 +45,8 @@ static int do_nmi(Object *o, void *opaque)
         ns->handled = true;
 
         /* Generic handler */
-        if (nc->nmi_monitor_handler) {
-            nc->nmi_monitor_handler(n, ns->cpu_index, &ns->err);
-            if (ns->err) {
-                return -1;
-            }
+        if (nc->nmi_handler) {
+            nc->nmi_handler(n);
         }
         /* Per vCPU handler */
         if (nc->nmi_cpu_handler) {
diff --git a/hw/hppa/machine.c b/hw/hppa/machine.c
index d762163ddf5..93874b5d5c9 100644
--- a/hw/hppa/machine.c
+++ b/hw/hppa/machine.c
@@ -829,7 +829,7 @@ static void hppa_machine_reset(MachineState *ms, ResetType type)
     cpu[0]->env.cmdline_or_bootorder = 'c';
 }
 
-static void hppa_nmi(NMIState *n, int cpu_index, Error **errp)
+static void hppa_nmi(NMIState *ns)
 {
     CPUState *cs;
 
@@ -851,7 +851,7 @@ static void hppa_machine_common_class_init(ObjectClass *oc, const void *data)
     mc->default_ram_id = "hppa.ram";
     mc->default_nic = "tulip";
 
-    nc->nmi_monitor_handler = hppa_nmi;
+    nc->nmi_handler = hppa_nmi;
 }
 
 static void HP_B160L_machine_init_class_init(ObjectClass *oc, const void *data)
diff --git a/hw/i386/x86.c b/hw/i386/x86.c
index dc7f0d56b01..201746945d5 100644
--- a/hw/i386/x86.c
+++ b/hw/i386/x86.c
@@ -147,9 +147,8 @@ static const CPUArchIdList *x86_possible_cpu_arch_ids(MachineState *ms)
     return ms->possible_cpus;
 }
 
-static void x86_nmi(NMIState *n, int cpu_index, Error **errp)
+static void x86_nmi(NMIState *ns)
 {
-    /* cpu index isn't used */
     CPUState *cs;
 
     CPU_FOREACH(cs) {
@@ -389,7 +388,7 @@ static void x86_machine_class_init(ObjectClass *oc, const void *data)
     mc->get_default_cpu_node_id = x86_get_default_cpu_node_id;
     mc->possible_cpu_arch_ids = x86_possible_cpu_arch_ids;
     mc->kvm_type = x86_kvm_type;
-    nc->nmi_monitor_handler = x86_nmi;
+    nc->nmi_handler = x86_nmi;
 
     object_class_property_add(oc, X86_MACHINE_SMM, "OnOffAuto",
         x86_machine_get_smm, x86_machine_set_smm,
diff --git a/hw/intc/m68k_irqc.c b/hw/intc/m68k_irqc.c
index 68ddb5351be..5d4e5a10ab8 100644
--- a/hw/intc/m68k_irqc.c
+++ b/hw/intc/m68k_irqc.c
@@ -70,7 +70,7 @@ static void m68k_irqc_instance_init(Object *obj)
     qdev_init_gpio_in(DEVICE(obj), m68k_set_irq, M68K_IRQC_LEVEL_NUM);
 }
 
-static void m68k_nmi(NMIState *n, int cpu_index, Error **errp)
+static void m68k_nmi(NMIState *n)
 {
     m68k_set_irq(n, M68K_IRQC_LEVEL_7, 1);
 }
@@ -97,7 +97,7 @@ static void m68k_irqc_class_init(ObjectClass *oc, const void *data)
     InterruptStatsProviderClass *ic = INTERRUPT_STATS_PROVIDER_CLASS(oc);
 
     device_class_set_props(dc, m68k_irqc_properties);
-    nc->nmi_monitor_handler = m68k_nmi;
+    nc->nmi_handler = m68k_nmi;
     device_class_set_legacy_reset(dc, m68k_irqc_reset);
     dc->vmsd = &vmstate_m68k_irqc;
     ic->get_statistics = m68k_irqc_get_statistics;
diff --git a/hw/m68k/q800-glue.c b/hw/m68k/q800-glue.c
index ac9062c6488..6a9592182a4 100644
--- a/hw/m68k/q800-glue.c
+++ b/hw/m68k/q800-glue.c
@@ -159,7 +159,7 @@ static void glue_auxmode_set_irq(void *opaque, int irq, int level)
     s->auxmode = level;
 }
 
-static void glue_nmi(NMIState *n, int cpu_index, Error **errp)
+static void glue_nmi(NMIState *n)
 {
     GLUEState *s = GLUE(n);
 
@@ -237,7 +237,7 @@ static void glue_class_init(ObjectClass *klass, const void *data)
     dc->vmsd = &vmstate_glue;
     device_class_set_props(dc, glue_properties);
     rc->phases.hold = glue_reset_hold;
-    nc->nmi_monitor_handler = glue_nmi;
+    nc->nmi_handler = glue_nmi;
 }
 
 static const TypeInfo glue_info_types[] = {
diff --git a/hw/misc/macio/gpio.c b/hw/misc/macio/gpio.c
index 1a7c534d652..f6fce5b3532 100644
--- a/hw/misc/macio/gpio.c
+++ b/hw/misc/macio/gpio.c
@@ -188,7 +188,7 @@ static void macio_gpio_reset(DeviceState *dev)
     macio_set_gpio(s, 1, true);
 }
 
-static void macio_gpio_nmi(NMIState *n, int cpu_index, Error **errp)
+static void macio_gpio_nmi(NMIState *n)
 {
     macio_set_gpio(MACIO_GPIO(n), 9, true);
     macio_set_gpio(MACIO_GPIO(n), 9, false);
@@ -201,7 +201,7 @@ static void macio_gpio_class_init(ObjectClass *oc, const void *data)
 
     device_class_set_legacy_reset(dc, macio_gpio_reset);
     dc->vmsd = &vmstate_macio_gpio;
-    nc->nmi_monitor_handler = macio_gpio_nmi;
+    nc->nmi_handler = macio_gpio_nmi;
 }
 
 static const TypeInfo macio_gpio_init_info = {
diff --git a/hw/ppc/pnv.c b/hw/ppc/pnv.c
index c0cb45dbfbc..817f2ed7ae4 100644
--- a/hw/ppc/pnv.c
+++ b/hw/ppc/pnv.c
@@ -3552,12 +3552,11 @@ static void pnv_cpu_do_nmi(PnvChip *chip, PowerPCCPU *cpu, void *opaque)
     async_run_on_cpu(CPU(cpu), pnv_cpu_do_nmi_on_cpu, RUN_ON_CPU_HOST_INT(0));
 }
 
-static void pnv_nmi(NMIState *n, int cpu_index, Error **errp)
+static void pnv_nmi(NMIState *ns)
 {
     PnvMachineState *pnv = PNV_MACHINE(qdev_get_machine());
-    int i;
 
-    for (i = 0; i < pnv->num_chips; i++) {
+    for (int i = 0; i < pnv->num_chips; i++) {
         pnv_chip_foreach_cpu(pnv->chips[i], pnv_cpu_do_nmi, NULL);
     }
 }
@@ -3583,7 +3582,7 @@ static void pnv_machine_class_init(ObjectClass *oc, const void *data)
     mc->default_ram_size = 1 * GiB;
     mc->default_ram_id = "pnv.ram";
     ispc->print_info = pnv_pic_print_info;
-    nc->nmi_monitor_handler = pnv_nmi;
+    nc->nmi_handler = pnv_nmi;
 
     object_class_property_add_bool(oc, "hb-mode",
                                    pnv_machine_get_hb, pnv_machine_set_hb);
diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
index b79828b4e90..a2195796efe 100644
--- a/hw/ppc/spapr.c
+++ b/hw/ppc/spapr.c
@@ -3550,7 +3550,7 @@ void spapr_do_system_reset_on_cpu(CPUState *cs, run_on_cpu_data arg)
     }
 }
 
-static void spapr_nmi(NMIState *n, int cpu_index, Error **errp)
+static void spapr_nmi(NMIState *ns)
 {
     CPUState *cs;
 
@@ -4652,7 +4652,7 @@ static void spapr_machine_class_init(ObjectClass *oc, const void *data)
     mc->nvdimm_supported = true;
     smc->resize_hpt_default = SPAPR_RESIZE_HPT_ENABLED;
     fwc->get_dev_path = spapr_get_fw_dev_path;
-    nc->nmi_monitor_handler = spapr_nmi;
+    nc->nmi_handler = spapr_nmi;
     vhc->cpu_in_nested = spapr_cpu_in_nested;
     vhc->deliver_hv_excp = spapr_exit_nested;
     vhc->hypercall = emulate_spapr_hypercall;
-- 
2.53.0

Re: [PATCH 0/4] hw/nmi: Safer vCPU delivery

[Philippe Mathieu-Daudé]

On 10/8/26 23:37, Philippe Mathieu-Daudé wrote:
> Hi Marc-André,
> 
> I don't remember what I suggested particularly for
> 20260626-qemu-no-hmp-v2-7-8af31bc54c61@redhat.com,
> but this series is what I had in mind back then.
> 
> If you find it worthy, please consider including it
> in your "Make HMP optional" work.

Not well explained, this series disentangle HMP from
the generic NMI class.

> 
> Regard,
> 
> Phil.
> 
> Philippe Mathieu-Daudé (4):
>    hw/nmi: Allow delivering to first available vCPU (not by index)
>    hw/nmi: Introduce nmi_cpu_handler() for vcpu-specific delivery
>    hw/s390x: Deliver per-vcpu NMI using nmi_cpu_handler
>    hw/nmi: Remove unused @cpu_index and @errp arguments
> 
>   include/hw/core/nmi.h      | 10 +++++++++-
>   hw/core/nmi.c              | 25 ++++++++++++++++++++++---
>   hw/hppa/machine.c          |  4 ++--
>   hw/i386/x86.c              |  5 ++---
>   hw/intc/m68k_irqc.c        |  4 ++--
>   hw/ipmi/ipmi.c             |  2 +-
>   hw/m68k/q800-glue.c        |  4 ++--
>   hw/misc/macio/gpio.c       |  4 ++--
>   hw/ppc/pnv.c               |  7 +++----
>   hw/ppc/spapr.c             |  4 ++--
>   hw/s390x/s390-virtio-ccw.c |  6 ++----
>   hw/watchdog/watchdog.c     |  2 +-
>   12 files changed, 50 insertions(+), 27 deletions(-)
> 

Re: [PATCH 0/4] hw/nmi: Safer vCPU delivery

[Philippe Mathieu-Daudé]

On 11/8/26 10:44, Philippe Mathieu-Daudé wrote:
> On 10/8/26 23:37, Philippe Mathieu-Daudé wrote:
>> Hi Marc-André,
>>
>> I don't remember what I suggested particularly for
>> 20260626-qemu-no-hmp-v2-7-8af31bc54c61@redhat.com,
>> but this series is what I had in mind back then.
>>
>> If you find it worthy, please consider including it
>> in your "Make HMP optional" work.
> 
> Not well explained, this series disentangle HMP from
> the generic NMI class.

Oh no, one patch is missing!

> 
>>
>> Regard,
>>
>> Phil.
>>
>> Philippe Mathieu-Daudé (4):
>>    hw/nmi: Allow delivering to first available vCPU (not by index)
>>    hw/nmi: Introduce nmi_cpu_handler() for vcpu-specific delivery
>>    hw/s390x: Deliver per-vcpu NMI using nmi_cpu_handler
>>    hw/nmi: Remove unused @cpu_index and @errp arguments
>>
>>   include/hw/core/nmi.h      | 10 +++++++++-
>>   hw/core/nmi.c              | 25 ++++++++++++++++++++++---
>>   hw/hppa/machine.c          |  4 ++--
>>   hw/i386/x86.c              |  5 ++---
>>   hw/intc/m68k_irqc.c        |  4 ++--
>>   hw/ipmi/ipmi.c             |  2 +-
>>   hw/m68k/q800-glue.c        |  4 ++--
>>   hw/misc/macio/gpio.c       |  4 ++--
>>   hw/ppc/pnv.c               |  7 +++----
>>   hw/ppc/spapr.c             |  4 ++--
>>   hw/s390x/s390-virtio-ccw.c |  6 ++----
>>   hw/watchdog/watchdog.c     |  2 +-
>>   12 files changed, 50 insertions(+), 27 deletions(-)
>>
> 
> 

Re: [PATCH 0/4] hw/nmi: Safer vCPU delivery

[Philippe Mathieu-Daudé]

On 11/8/26 10:50, Philippe Mathieu-Daudé wrote:
> On 11/8/26 10:44, Philippe Mathieu-Daudé wrote:
>> On 10/8/26 23:37, Philippe Mathieu-Daudé wrote:
>>> Hi Marc-André,
>>>
>>> I don't remember what I suggested particularly for
>>> 20260626-qemu-no-hmp-v2-7-8af31bc54c61@redhat.com,
>>> but this series is what I had in mind back then.
>>>
>>> If you find it worthy, please consider including it
>>> in your "Make HMP optional" work.
>>
>> Not well explained, this series disentangle HMP from
>> the generic NMI class.

Actually previous version is better:
https://lore.kernel.org/qemu-devel/20240220150833.13674-1-philmd@linaro.org/

> 
> Oh no, one patch is missing!
> 
>>
>>>
>>> Regard,
>>>
>>> Phil.
>>>
>>> Philippe Mathieu-Daudé (4):
>>>    hw/nmi: Allow delivering to first available vCPU (not by index)
>>>    hw/nmi: Introduce nmi_cpu_handler() for vcpu-specific delivery
>>>    hw/s390x: Deliver per-vcpu NMI using nmi_cpu_handler
>>>    hw/nmi: Remove unused @cpu_index and @errp arguments
>>>
>>>   include/hw/core/nmi.h      | 10 +++++++++-
>>>   hw/core/nmi.c              | 25 ++++++++++++++++++++++---
>>>   hw/hppa/machine.c          |  4 ++--
>>>   hw/i386/x86.c              |  5 ++---
>>>   hw/intc/m68k_irqc.c        |  4 ++--
>>>   hw/ipmi/ipmi.c             |  2 +-
>>>   hw/m68k/q800-glue.c        |  4 ++--
>>>   hw/misc/macio/gpio.c       |  4 ++--
>>>   hw/ppc/pnv.c               |  7 +++----
>>>   hw/ppc/spapr.c             |  4 ++--
>>>   hw/s390x/s390-virtio-ccw.c |  6 ++----
>>>   hw/watchdog/watchdog.c     |  2 +-
>>>   12 files changed, 50 insertions(+), 27 deletions(-)
>>>
>>
>>
> 
> 

Re: [PATCH 0/4] hw/nmi: Safer vCPU delivery

[Daniel P. Berrangé]

On Tue, Aug 11, 2026 at 11:36:59AM +0200, Philippe Mathieu-Daudé wrote:
> On 11/8/26 10:50, Philippe Mathieu-Daudé wrote:
> > On 11/8/26 10:44, Philippe Mathieu-Daudé wrote:
> > > On 10/8/26 23:37, Philippe Mathieu-Daudé wrote:
> > > > Hi Marc-André,
> > > > 
> > > > I don't remember what I suggested particularly for
> > > > 20260626-qemu-no-hmp-v2-7-8af31bc54c61@redhat.com,
> > > > but this series is what I had in mind back then.
> > > > 
> > > > If you find it worthy, please consider including it
> > > > in your "Make HMP optional" work.
> > > 
> > > Not well explained, this series disentangle HMP from
> > > the generic NMI class.
> 
> Actually previous version is better:
> https://lore.kernel.org/qemu-devel/20240220150833.13674-1-philmd@linaro.org/

Are you going to rebase & resend that ?   The old thread has a lot
of commentary that I don't think it is worth reviewing the old code
again until updated.


With regards,
Daniel
-- 
|: https://berrange.com       ~~        https://hachyderm.io/@berrange :|
|: https://libvirt.org          ~~          https://entangle-photo.org :|
|: https://pixelfed.art/berrange   ~~    https://fstop138.berrange.com :|

Re: [PATCH 0/4] hw/nmi: Safer vCPU delivery

[Philippe Mathieu-Daudé]

On 11/8/26 11:49, Daniel P. Berrangé wrote:
> On Tue, Aug 11, 2026 at 11:36:59AM +0200, Philippe Mathieu-Daudé wrote:
>> On 11/8/26 10:50, Philippe Mathieu-Daudé wrote:
>>> On 11/8/26 10:44, Philippe Mathieu-Daudé wrote:
>>>> On 10/8/26 23:37, Philippe Mathieu-Daudé wrote:
>>>>> Hi Marc-André,
>>>>>
>>>>> I don't remember what I suggested particularly for
>>>>> 20260626-qemu-no-hmp-v2-7-8af31bc54c61@redhat.com,
>>>>> but this series is what I had in mind back then.
>>>>>
>>>>> If you find it worthy, please consider including it
>>>>> in your "Make HMP optional" work.
>>>>
>>>> Not well explained, this series disentangle HMP from
>>>> the generic NMI class.
>>
>> Actually previous version is better:
>> https://lore.kernel.org/qemu-devel/20240220150833.13674-1-philmd@linaro.org/
> 
> Are you going to rebase & resend that ?   The old thread has a lot
> of commentary that I don't think it is worth reviewing the old code
> again until updated.

Please discard this series.

Old series rebased and reposted (addressing comments) as v3 here:
https://lore.kernel.org/qemu-devel/20260811105425.7429-1-philmd@oss.qualcomm.com/