Pradeep Jagadeesh <pradeep.jagade...@huawei.com> writes:
> On 9/8/2017 12:02 PM, Markus Armbruster wrote:
>> Pradeep Jagadeesh <pradeepkiruv...@gmail.com> writes:
[...]
>>> diff --git a/qmp.c b/qmp.c
>>> index b86201e..eed91e5 100644
>>> --- a/qmp.c
>>> +++ b/qmp.c
>>> @@ -130,6 +130,20 @@ void qmp_cpu_add(int64_t id, Error **errp)
>>> }
>>> }
>>>
>>> +#if defined(_WIN64) || defined(_WIN32) || defined(__FreeBSD__)
>>> +
>>> +void qmp_fsdev_set_io_throttle(IOThrottle *arg, Error **errp)
>>> +{
>>> + return;
>>> +}
>>> +
>>> +IOThrottleList *qmp_query_fsdev_io_throttle(Error **errp)
>>> +{
>>> + return NULL;
>>> +}
>>> +
>>> +#endif
>>> +
>>> #ifndef CONFIG_VNC
>>> /* If VNC support is enabled, the "true" query-vnc command is
>>> defined in the VNC subsystem */
>>
>> Why do we need *two* stubs for these functions, one here and one in
>> fsdev/qemu-fsdev-dummy.c?
> At two different platforms the build fails. So, there are needed.
We don't add stubs to random places until the linker succeeds for all
the configurations we happen to test. We figure out *how* the linker
fails to correct our idea of where a certain stub belongs until we find
the one appropriate home for it.
If you can't figure that out, that's okay, just tell us how exactly the
linker fails for you, and we'll be happy to help.
