Re: [PATCH] replication: compile out some staff when replication is not configured

[Vladimir Sementsov-Ogievskiy]

On 17.04.23 20:19, Vladimir Sementsov-Ogievskiy wrote:
> On 16.04.23 21:44, Zhang, Chen wrote:
> > > -----Original Message-----
> > > From: Vladimir Sementsov-Ogievskiy <vsementsov@yandex-team.ru>
> > > Sent: Friday, April 14, 2023 5:51 PM
> > > To: Zhang, Chen <chen.zhang@intel.com>; qemu-devel@nongnu.org
> > > Cc: qemu-block@nongnu.org; pbonzini@redhat.com; armbru@redhat.com;
> > > eblake@redhat.com; jasowang@redhat.com; dgilbert@redhat.com;
> > > quintela@redhat.com; hreitz@redhat.com; kwolf@redhat.com; Zhang,
> > > Hailiang <zhanghailiang@xfusion.com>; lizhijian@fujitsu.com;
> > > wencongyang2@huawei.com; xiechanglong.d@gmail.com; den-
> > > plotnikov@yandex-team.ru
> > > Subject: Re: [PATCH] replication: compile out some staff when replication is
> > > not configured
> > >
> > > On 14.04.23 04:24, Zhang, Chen wrote:
> > > > > So, if I want to have an option to disable all COLO modules, do you
> > > > > mean it should be additional --disable-colo option? Or better keep
> > > > > one option -- disable-replication (and, maybe just rename to to --disable-
> > > colo)?
> > > > I think keep the option --disable-replication is enough.
> > > > Generally speaking, these three modules do not belong to COLO, It has
> > > been decoupled at the time of design.
> > > > Instead, COLO is formed when these three modules are used in
> > > combination.
> > >
> > > But it's not enough to me, I want to have a possibility to not build the
> > > subsystem I don't need.
> >
> > As I said, COLO not a specific subsystem, It is a usage of three general 
> > subsystems.
> > Let's back to this patch, it try to not build block replication when not 
> > configured. It's OK.
> > Although COLO may be the only user of replication,  but can't assume all the 
> > COLO used subsystem
> > not needed, even have a --disable-colo. For example in this patch disabled the 
> > net/filter-mirror/redirector....
> > Qemu network filter is a general framework with many submodules: 
> > filter-buffer/replay/mirror/rewriter.....
> > Logically speaking, It is completely irrelevant with COLO.
>
> OK, understand now, thanks for explanation!

Hmm, OK, filters are separate.

But what's the sense in COLO subsystem when replication is disabled?

We have an explicit check in migration.c in migrate_caps_check():

#ifndef CONFIG_REPLICATION
    if (cap_list[MIGRATION_CAPABILITY_X_COLO]) {
        error_setg(errp, "QEMU compiled without replication module"
                   " can't enable COLO");
        error_append_hint(errp, "Please enable replication before COLO.\n");
        return false;
    }
#endif


So, x-colo capability - an entry point to COLO can't be used anyway. So, no 
reason to build it in?

--
Best regards,
Vladimir