Re: [Qemu-devel] [RFC][PATCH v2 09/17] qmp proxy: core code for proxying qmp requests to guest

[Michael Roth <mdroth@linux.vnet.ibm.com>]

On 04/21/2011 03:30 AM, Jes Sorensen wrote:
> On 04/18/11 17:02, Michael Roth wrote:
>> diff --git a/qmp-core.c b/qmp-core.c
>> index 9f3d182..dab50a1 100644
>> --- a/qmp-core.c
>> +++ b/qmp-core.c
>> @@ -937,7 +937,15 @@ void qmp_async_complete_command(QmpCommandState *cmd, QObject *retval, Error *er
>>       qemu_free(cmd);
>>   }
>>
>> +extern QmpProxy *qmp_proxy_default;
>
> Please put this in a header file.
>
>> +static void qmp_proxy_process_control_event(QmpProxy *p, const QDict *evt)
>> +{
>> +    const char *cmd;
>> +    int host_sid, guest_sid;
>> +
>> +    cmd = qdict_get_try_str(evt, "_control_event");
>> +    if (!cmd) {
>> +        fprintf(stderr, "received NULL control event\n");
>> +    } else if (strcmp(cmd, "guest_ack") == 0) {
>> +        host_sid = qdict_get_try_int(evt, "_control_arg_host_sid", 0);
>> +        if (!host_sid) {
>> +            fprintf(stderr, "invalid guest_ack: wrong host sid\n");
>> +            return;
>> +        }
>> +        /* guest responded to host_init, return a host_ack */
>> +        /* reset outstanding requests, then send an ack with the
>> +         * session id they passed us
>> +         */
>> +        guest_sid = qdict_get_try_int(evt, "_control_arg_guest_sid", 0);
>
> I am wondering if it would make sense to put these arguments in a header
> file as #define's to make sure you don't have to chase down a typo on
> one side at some point? Just an idea, dunno if it is worth it.

That's probably a good idea.

>
> Cheers,
> Jes