[PATCH 2/3] nmi: make cpu-index argument optional

[PATCH 2/3] nmi: make cpu-index argument optional

[Lai Jiangshan]

When the argument "cpu-index" is not given,
then "nmi" command will inject NMI on all CPUs.

This simulate the nmi button on physical machine.

Signed-off-by:  Lai Jiangshan <laijs@cn.fujitsu.com>
---
diff --git a/hmp-commands.hx b/hmp-commands.hx
index 8de7aa3..d8fe4c0 100644
--- a/hmp-commands.hx
+++ b/hmp-commands.hx
@@ -721,9 +721,9 @@ ETEXI
 #if defined(TARGET_I386)
     {
         .name       = "nmi",
-        .args_type  = "cpu-index:i",
-        .params     = "cpu",
-        .help       = "inject an NMI on the given CPU",
+        .args_type  = "cpu-index:i?",
+        .params     = "[cpu]",
+        .help       = "inject an NMI on all CPUs or the given CPU",
         .mhandler.cmd = do_inject_nmi,
     },
 #endif
diff --git a/monitor.c b/monitor.c
index c16b39d..45a8dc2 100644
--- a/monitor.c
+++ b/monitor.c
@@ -2410,7 +2410,13 @@ static void do_wav_capture(Monitor *mon, const QDict *qdict)
 static void do_inject_nmi(Monitor *mon, const QDict *qdict)
 {
     CPUState *env;
-    int cpu_index = qdict_get_int(qdict, "cpu-index");
+    int cpu_index = qdict_get_try_int(qdict, "cpu-index", -1);
+
+    if (cpu_index == -1) {
+        for (env = first_cpu; env != NULL; env = env->next_cpu)
+            cpu_interrupt(env, CPU_INTERRUPT_NMI);
+        return;
+    }
 
     for (env = first_cpu; env != NULL; env = env->next_cpu)
         if (env->cpu_index == cpu_index) {

Re: [Qemu-devel] [PATCH 2/3] nmi: make cpu-index argument optional

[Markus Armbruster]

Lai Jiangshan <laijs@cn.fujitsu.com> writes:

> When the argument "cpu-index" is not given,
> then "nmi" command will inject NMI on all CPUs.
>
> This simulate the nmi button on physical machine.
>
> Signed-off-by:  Lai Jiangshan <laijs@cn.fujitsu.com>
> ---
> diff --git a/hmp-commands.hx b/hmp-commands.hx
> index 8de7aa3..d8fe4c0 100644
> --- a/hmp-commands.hx
> +++ b/hmp-commands.hx
> @@ -721,9 +721,9 @@ ETEXI
>  #if defined(TARGET_I386)
>      {
>          .name       = "nmi",
> -        .args_type  = "cpu-index:i",
> -        .params     = "cpu",
> -        .help       = "inject an NMI on the given CPU",
> +        .args_type  = "cpu-index:i?",
> +        .params     = "[cpu]",
> +        .help       = "inject an NMI on all CPUs or the given CPU",
>          .mhandler.cmd = do_inject_nmi,
>      },
>  #endif
> diff --git a/monitor.c b/monitor.c
> index c16b39d..45a8dc2 100644
> --- a/monitor.c
> +++ b/monitor.c
> @@ -2410,7 +2410,13 @@ static void do_wav_capture(Monitor *mon, const QDict *qdict)
>  static void do_inject_nmi(Monitor *mon, const QDict *qdict)
>  {
>      CPUState *env;
> -    int cpu_index = qdict_get_int(qdict, "cpu-index");
> +    int cpu_index = qdict_get_try_int(qdict, "cpu-index", -1);

Note that we interpret "nmi -1" as "nmi" (no argument).  Doesn't matter
now.  Does matter after the next patch, because that patch makes the
command when the argument doesn't make sense.  Except it doesn't for the
non-sensical argument -1.  Same problem with every existing use of
qdict_get_try_int() in the monitor.  Do we care?

> +
> +    if (cpu_index == -1) {
> +        for (env = first_cpu; env != NULL; env = env->next_cpu)
> +            cpu_interrupt(env, CPU_INTERRUPT_NMI);
> +        return;
> +    }
>  
>      for (env = first_cpu; env != NULL; env = env->next_cpu)
>          if (env->cpu_index == cpu_index) {

Re: [Qemu-devel] [PATCH 2/3] nmi: make cpu-index argument optional

[Lai Jiangshan]

On 12/21/2010 12:58 AM, Markus Armbruster wrote:
> Lai Jiangshan <laijs@cn.fujitsu.com> writes:
> 
>> When the argument "cpu-index" is not given,
>> then "nmi" command will inject NMI on all CPUs.
>>
>> This simulate the nmi button on physical machine.
>>
>> Signed-off-by:  Lai Jiangshan <laijs@cn.fujitsu.com>
>> ---
>> diff --git a/hmp-commands.hx b/hmp-commands.hx
>> index 8de7aa3..d8fe4c0 100644
>> --- a/hmp-commands.hx
>> +++ b/hmp-commands.hx
>> @@ -721,9 +721,9 @@ ETEXI
>>  #if defined(TARGET_I386)
>>      {
>>          .name       = "nmi",
>> -        .args_type  = "cpu-index:i",
>> -        .params     = "cpu",
>> -        .help       = "inject an NMI on the given CPU",
>> +        .args_type  = "cpu-index:i?",
>> +        .params     = "[cpu]",
>> +        .help       = "inject an NMI on all CPUs or the given CPU",
>>          .mhandler.cmd = do_inject_nmi,
>>      },
>>  #endif
>> diff --git a/monitor.c b/monitor.c
>> index c16b39d..45a8dc2 100644
>> --- a/monitor.c
>> +++ b/monitor.c
>> @@ -2410,7 +2410,13 @@ static void do_wav_capture(Monitor *mon, const QDict *qdict)
>>  static void do_inject_nmi(Monitor *mon, const QDict *qdict)
>>  {
>>      CPUState *env;
>> -    int cpu_index = qdict_get_int(qdict, "cpu-index");
>> +    int cpu_index = qdict_get_try_int(qdict, "cpu-index", -1);
> 
> Note that we interpret "nmi -1" as "nmi" (no argument).  Doesn't matter
> now.  Does matter after the next patch, because that patch makes the
> command when the argument doesn't make sense.  Except it doesn't for the
> non-sensical argument -1.  Same problem with every existing use of
> qdict_get_try_int() in the monitor.  Do we care?
> 

Good catch!! I will fix it.

Tanks,
Lai.