Linux Ethernet Bridging - Aug 2019 - [Bridge] [PATCH v2 0/3] Add NETIF_F_HW_BR_CAP feature

The 08/26/2019 14:38, Andrew Lunn wrote:
> External E-Mail
> 
> 
> On Mon, Aug 26, 2019 at 10:11:12AM +0200, Horatiu Vultur wrote:
> > When a network port is added to a bridge then the port is added in
> > promisc mode. Some HW that has bridge capabilities(can learn, forward,
> > flood etc the frames) they are disabling promisc mode in the network
> > driver when the port is added to the SW bridge.
> > 
> > This patch adds the feature NETIF_F_HW_BR_CAP so that the network
ports
> > that have this feature will not be set in promisc mode when they are
> > added to a SW bridge.
> > 
> > In this way the HW that has bridge capabilities don't need to send
all the
> > traffic to the CPU and can also implement the promisc mode and toggle
it
> > using the command 'ip link set dev swp promisc on'
> 
> Hi Horatiu
Hi Andrew,
> 
> I'm still not convinced this is needed. The model is, the hardware is
> there to accelerate what Linux can do in software. Any peculiarities
> of the accelerator should be hidden in the driver.  If the accelerator
> can do its job without needing promisc mode, do that in the driver.
Thanks for the model description. I will keep in my mind for the next
patches that I will do.
> 
> So you are trying to differentiate between promisc mode because the
> interface is a member of a bridge, and promisc mode because some
> application, like pcap, has asked for promisc mode.
> 
> dev->promiscuity is a counter. So what you can do it look at its
> value, and how the interface is being used. If the interface is not a
> member of a bridge, and the count > 0, enable promisc mode in the
> accelerator. If the interface is a member of a bridge, and the count >
> 1, enable promisc mode in the accelerator.
That sounds like a great idea. I was expecting to add this logic in the
set_rx_mode function of the driver. But unfortunetly, I got the calls to
this function before the dev->promiscuity is updated or not to get the
call at all. For example in case the port is member of a bridge and I try
to enable the promisc mode.
> 
>    Andrew
> 
> 
-- 
/Horatiu

> That sounds like a great idea. I was expecting to add this logic in the
> set_rx_mode function of the driver. But unfortunetly, I got the calls to
> this function before the dev->promiscuity is updated or not to get the
> call at all. For example in case the port is member of a bridge and I try
> to enable the promisc mode.
Hi Horatiu

What about the notifier? Is it called in all the conditions you need
to know about?

Or, you could consider adding a new switchdev call to pass this
information to any switchdev driver which is interested in the
information.

At the moment, the DSA driver core does not pass onto the driver it
should put a port into promisc mode. So pcap etc, will only see
traffic directed to the CPU, not all the traffic ingressing the
interface. If you put the needed core infrastructure into place, we
could plumb it down from the DSA core to DSA drivers.

Having said that, i don't actually know if the Marvell switches
support this. Forward using the ATU and send a copy to the CPU?  What
switches tend to support is port mirroring, sending all the traffic
out another port. A couple of DSA drivers support that, via TC.

	Andrew