On Fri, Apr 18, 2014 at 2:53 PM, Paolo Bonzini <pbonz...@redhat.com> wrote:
> Il 18/04/2014 00:25, Peter Crosthwaite ha scritto:
>
>> This "nofail" (i.e. does not return NULL) mechanism driving
>> qemu_get_machine_opts() does not need to be specific to machine opts
>> - its applicable to other types of opts. Generalise and re-implement
>> qemu_get_machine_opts() as a caller of the generalisation.
>>
>> Signed-off-by: Peter Crosthwaite <peter.crosthwa...@xilinx.com>
>> ---
>>
>> vl.c | 20 +++++++++++++-------
>> 1 file changed, 13 insertions(+), 7 deletions(-)
>>
>> diff --git a/vl.c b/vl.c
>> index 9975e5a..bc12d0f 100644
>> --- a/vl.c
>> +++ b/vl.c
>> @@ -510,17 +510,12 @@ static QemuOptsList qemu_name_opts = {
>> },
>> };
>>
>> -/**
>> - * Get machine options
>> - *
>> - * Returns: machine options (never null).
>> - */
>> -QemuOpts *qemu_get_machine_opts(void)
>> +static QemuOpts *qemu_get_opts_nofail(const char *type)
>> {
>> QemuOptsList *list;
>> QemuOpts *opts;
>>
>> - list = qemu_find_opts("machine");
>> + list = qemu_find_opts(type);
>> assert(list);
>> opts = qemu_opts_find(list, NULL);
>> if (!opts) {
>> @@ -529,6 +524,17 @@ QemuOpts *qemu_get_machine_opts(void)
>> return opts;
>> }
>>
>> +/**
>> + * Get machine options
>> + *
>> + * Returns: machine options (never null).
>> + */
>> +
>> +QemuOpts *qemu_get_machine_opts(void)
>> +{
>> + return qemu_get_opts_nofail("machine");
>> +}
>> +
>> const char *qemu_get_vm_name(void)
>> {
>> return qemu_name;
>>
>
> This is already planned for 2.1 as qemu_find_opts_singleton, which either
> Igor or Hu will send.
>
I see. Anything else in that work that you see conflicting with other
components of this series? AFAICT its just a trivial change to P2 to
back onto the qemu_find_opts_singleton work and the rest stands as-is.
Regards,
Peter
> Paolo
>
