Re: [PATCH 4/5] qapi: introduce x-query-registers QMP command

[Markus Armbruster]

Daniel P. Berrangé <berrange@redhat.com> writes:

> This is a counterpart to the HMP "info registers" command. It is being
> added with an "x-" prefix because this QMP command is intended as an
> adhoc debugging tool and will thus not be modelled in QAPI as fully
> structured data, nor will it have long term guaranteed stability.
>
> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
> ---
>  hw/core/machine-qmp-cmds.c | 28 ++++++++++++++++++++++++++++
>  qapi/machine.json          | 37 +++++++++++++++++++++++++++++++++++++
>  2 files changed, 65 insertions(+)
>
> diff --git a/hw/core/machine-qmp-cmds.c b/hw/core/machine-qmp-cmds.c
> index 216fdfaf3a..0d9943ff60 100644
> --- a/hw/core/machine-qmp-cmds.c
> +++ b/hw/core/machine-qmp-cmds.c
> @@ -204,3 +204,31 @@ MemdevList *qmp_query_memdev(Error **errp)
>      object_child_foreach(obj, query_memdev, &list);
>      return list;
>  }
> +
> +RegisterInfo *qmp_x_query_registers(bool has_cpu, int64_t cpu, Error **errp)
> +{
> +    RegisterInfo *info = g_new0(RegisterInfo, 1);
> +    g_autoptr(GString) buf = g_string_new("");
> +    CPUState *cs = NULL, *tmp;
> +
> +    if (has_cpu) {
> +        CPU_FOREACH(tmp) {
> +            if (cpu == tmp->cpu_index) {
> +                cs = tmp;
> +            }
> +        }
> +        if (!cs) {
> +            error_setg(errp, "CPU %"PRId64" not available", cpu);
> +            return NULL;
> +        }
> +        cpu_format_state(cs, buf, CPU_DUMP_FPU);
> +    } else {
> +        CPU_FOREACH(cs) {
> +            g_string_append_printf(buf, "\nCPU#%d\n", cs->cpu_index);
> +            cpu_format_state(cs, buf, CPU_DUMP_FPU);
> +        }
> +    }
> +
> +    info->state = g_steal_pointer(&buf->str);
> +    return info;
> +}
> diff --git a/qapi/machine.json b/qapi/machine.json
> index 157712f006..27b922f2ce 100644
> --- a/qapi/machine.json
> +++ b/qapi/machine.json
> @@ -1312,3 +1312,40 @@
>       '*cores': 'int',
>       '*threads': 'int',
>       '*maxcpus': 'int' } }
> +
> +##
> +# @RegisterParams:
> +#
> +# Information about the CPU to query state of
> +#
> +# @cpu: the CPU number to query. If omitted, queries all CPUs
> +#
> +# Since: 6.2.0
> +#
> +##
> +{ 'struct': 'RegisterParams', 'data': {'*cpu': 'int' } }
> +
> +##
> +# @RegisterInfo:
> +#
> +# Information about the CPU state
> +#
> +# @state: the CPU state in an architecture specific format
> +#
> +# Since: 6.2.0
> +#
> +##
> +{ 'struct': 'RegisterInfo', 'data': {'state': 'str' } }
> +
> +##
> +# @x-query-registers:
> +#
> +# Return information on the CPU registers
> +#
> +# Returns: the CPU state
> +#
> +# Since: 6.2.0
> +##
> +{ 'command': 'x-query-registers',
> +  'data': 'RegisterParams',

Unless you have further uses of RegisterParams in mind, use an implicit
type:

     'data': { '*cpu': 'int' } }

> +  'returns': 'RegisterInfo' }

I'd like us to adopt a convention for commands returning formatted text
for human consumption.  Like so:

     'returns': 'HumanReadableText' }

with the obvious

   ##
   # @HumanReadableText:
   #
   # @human-readable-text: Formatted output intended for humans.
   #
   # Since: 6.2.0
   ##
   { 'struct': 'HumanReadableText',
     'data': { 'human-readable-text': 'str' } }

When the output needs explaining, do that in the command's doc string.
I figure

   ##
   # @x-query-registers:
   #
   # Returns information about the CPU state
   #
   # Returns: CPU state in an architecture-specific format
   #
   # Since: 6.2.0
   ##

would do in this case.