Re: [PATCH v3 09/14] migration/multifd: Isolate ram pages packet data

[Fabiano Rosas]

Peter Xu <peterx@redhat.com> writes:

> On Thu, Aug 22, 2024 at 11:13:36AM -0300, Fabiano Rosas wrote:
>> Peter Xu <peterx@redhat.com> writes:
>> 
>> > On Thu, Aug 01, 2024 at 09:35:11AM -0300, Fabiano Rosas wrote:
>> >> @@ -1554,7 +1577,6 @@ static void *multifd_recv_thread(void *opaque)
>> >>                  qemu_sem_wait(&p->sem_sync);
>> >>              }
>> >>          } else {
>> >> -            p->total_normal_pages += p->data->size / 
>> >> qemu_target_page_size();
>> >
>> > Is this line dropped by accident?
>> >
>> 
>> No, this was just used in the tracepoint below. I stopped including this
>> information there.
>
> But this will cause socket / file paths not doing the same thing, since
> this counter should still be increamented in socket path (and this is the
> file path).
>
> Either we keep it the same as before, or.. if we want to drop it, shouldn't
> we remove all instead (along with the two variables "total_normal_pages /
> total_zero_pages")?

I'll just remove them. There's not much point in these anyway since
they are per-channel.