"Dr. David Alan Gilbert" <dgilb...@redhat.com> wrote:
> * Juan Quintela (quint...@redhat.com) wrote:
>> We set the x-multifd-page-count and x-multifd-channels.
>>
>> Signed-off-by: Juan Quintela <quint...@redhat.com>
>
> <snip>
> We *must* find a way to share all the boiler plate I've snipped
> out; even this test_migrate function is almost identical to
> your other test series with just those few extra parameter sets.
See my migration-test series, I changed this a lot. I think that
without using function pointers, it is not easy to share more code that
what I did there (and I am not sure that using function pointers makes
things easier to understand).
>> + /* 200ms downtime */
>> + rsp = qmp("{ 'execute': 'migrate-set-parameters',"
>> + "'arguments': { 'downtime-limit': 300 } }");
>
> Note 200 vs 300 !
Fixed when I merged with the previous cod.
>> + /* Destination still running, wait for a byte to change */
>> + do {
>> + qtest_memread(to, start_address, &dest_byte_b, 1);
>> + usleep(10 * 1000);
>> + } while (dest_byte_a == dest_byte_b);
>
> Are there any multifd stats we can check? I guess the migration
> wont start until all channels are connected, so we know they're all
> in use in theory, so there's nothing specific.
We can check that all threads have been created, for instance.
Later, Juan.
