Re: [PATCH] net/colo: check vnet_hdr_support flag when using virtio-net

[Tao Xu]

On 8/17/2021 2:01 PM, Tao Xu wrote:
> On 8/16/2021 10:58 AM, Jason Wang wrote:
> > 在 2021/8/6 下午2:08, Tao Xu 写道:
> > > When COLO use only one vnet_hdr_support parameter between
> > > COLO network filter(filter-mirror, filter-redirector or
> > > filter-rewriter and colo-compare, packet will not be parsed
> > > correctly. Acquire network driver related to COLO, if it is
> > > nirtio-net, check vnet_hdr_support flag of COLO network filter
> > > and colo-compare.
> > >
> > > Signed-off-by: Tao Xu <tao3.xu@intel.com>
> > > Signed-off-by: Zhang Chen <chen.zhang@intel.com>
> > > ---
> > >    net/colo-compare.c    | 25 +++++++++++++++++++++++++
> > >    net/colo.c            | 20 ++++++++++++++++++++
> > >    net/colo.h            |  4 ++++
> > >    net/filter-mirror.c   | 17 +++++++++++++++++
> > >    net/filter-rewriter.c |  9 +++++++++
> > >    5 files changed, 75 insertions(+)
> > >
> > > diff --git a/net/colo-compare.c b/net/colo-compare.c
> > > index b100e7b51f..bc1cc951c0 100644
> > > --- a/net/colo-compare.c
> > > +++ b/net/colo-compare.c
> > > @@ -838,6 +838,23 @@ static int compare_chr_can_read(void *opaque)
> > >        return COMPARE_READ_LEN_MAX;
> > >    }
> > > +/* check vnet_hdr_support flag through COLO filter modules */
> > > +static int colo_vnet_driver_check(void *opaque, QemuOpts *opts, 
> > > Error **errp)
> > > +{
> > > +    const char *colo_obj_type;
> > > +
> > > +    colo_obj_type = qemu_opt_get(opts, "qom-type");
> > > +
> > > +    if (strcmp(colo_obj_type, "filter-mirror") == 0 ||
> > > +        strcmp(colo_obj_type, "filter-redirector") == 0 ||
> > > +        strcmp(colo_obj_type, "filter-rewriter") == 0) {
> > > +        if (qemu_opt_get(opts, "vnet_hdr_support")) {
> > > +            return 1;
> > > +        }
> > > +    }
> > > +    return 0;
> > > +}
> > > +
> > >    /*
> > >     * Called from the main thread on the primary for packets
> > >     * arriving over the socket from the primary.
> > > @@ -1289,6 +1306,14 @@ static void 
> > > colo_compare_complete(UserCreatable *uc, Error **errp)
> > >            return;
> > >        }
> > > +    if (!s->vnet_hdr &&
> > > +        qemu_opts_foreach(qemu_find_opts("object"),
> > > +                          colo_vnet_driver_check, NULL, NULL)) {
> > > +        error_setg(errp, "colo compare needs 'vnet_hdr_support' "
> > > +                   "when colo filter modules work on virtio-net");
> > > +        return;
> > > +    }
> >
> >
> > I wonder if we can detect virtio-net and apply vnet_hdr automatically.
> >
> > Thanks
> For filter-mirror, filter-redirector and filter-rewriter, we can detect 
> and add it automatically, because these netfilter is attached to netdev, 
> for example,
>
>      if (!s->vnet_hdr &&
>          qemu_opts_foreach(qemu_find_opts("device"),
>                           vnet_driver_check, nf->netdev_id, NULL)) {
>          s->vnet_hdr = true.
>      }
>
>
> But for colo-compare, it isn't attached to netdev, only can check colo 
> netfilter to check vnet_hdr_support. In this situation, if all netfilter 
> vnet_hdr_support is missing, colo_vnet_driver_check() will return 0, it 
> can't find vnet_hdr_support is missing.
>
> So can we apply vnet_hdr automatically for filter-mirror, 
> filter-redirector and filter-rewriter? And keep report error for 
> colo-compare?

Sorry, I find the solution for colo-compare apply vnet_hdr 
automatically, I will submit V2 later.