Rasmus Villemoes
2021-Jan-18 18:13 UTC
[Bridge] [PATCH net] net: mrp: use stp state as substitute for unimplemented mrp state
When using MRP with hardware that does understand the concept of
blocked or forwarding ports, but not the full MRP offload, we
currently fail to tell the hardware what state it should put the port
in when the ring is closed - resulting in a ring of forwarding ports
and all the trouble that comes with that.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes at prevas.dk>
---
I don't really understand why SWITCHDEV_ATTR_ID_MRP_PORT_STATE even
has to exist seperately from SWITCHDEV_ATTR_ID_PORT_STP_STATE, and
it's hard to tell what the difference might be since no kernel code
implements the former.
net/bridge/br_mrp_switchdev.c | 18 ++++++++++++++++++
1 file changed, 18 insertions(+)
diff --git a/net/bridge/br_mrp_switchdev.c b/net/bridge/br_mrp_switchdev.c
index ed547e03ace1..8a1c7953e57a 100644
--- a/net/bridge/br_mrp_switchdev.c
+++ b/net/bridge/br_mrp_switchdev.c
@@ -180,6 +180,24 @@ int br_mrp_port_switchdev_set_state(struct net_bridge_port
*p,
int err;
err = switchdev_port_attr_set(p->dev, &attr);
+ if (err == -EOPNOTSUPP) {
+ attr.id = SWITCHDEV_ATTR_ID_PORT_STP_STATE;
+ switch (state) {
+ case BR_MRP_PORT_STATE_DISABLED:
+ case BR_MRP_PORT_STATE_NOT_CONNECTED:
+ attr.u.stp_state = BR_STATE_DISABLED;
+ break;
+ case BR_MRP_PORT_STATE_BLOCKED:
+ attr.u.stp_state = BR_STATE_BLOCKING;
+ break;
+ case BR_MRP_PORT_STATE_FORWARDING:
+ attr.u.stp_state = BR_STATE_FORWARDING;
+ break;
+ default:
+ return err;
+ };
+ err = switchdev_port_attr_set(p->dev, &attr);
+ }
if (err && err != -EOPNOTSUPP)
br_warn(p->br, "error setting offload MRP state on port
%u(%s)\n",
(unsigned int)p->port_no, p->dev->name);
--
2.23.0
Horatiu Vultur
2021-Jan-18 18:56 UTC
[Bridge] [PATCH net] net: mrp: use stp state as substitute for unimplemented mrp state
The 01/18/2021 19:13, Rasmus Villemoes wrote:>Hi Rasmus,> When using MRP with hardware that does understand the concept of > blocked or forwarding ports, but not the full MRP offload, we > currently fail to tell the hardware what state it should put the port > in when the ring is closed - resulting in a ring of forwarding ports > and all the trouble that comes with that.But why don't you implement the SWITCHDEV_ATTR_ID_MRP_PORT_STATE in your driver? if already the HW understands the concept of block or forwarding?> > Signed-off-by: Rasmus Villemoes <rasmus.villemoes at prevas.dk> > --- > > I don't really understand why SWITCHDEV_ATTR_ID_MRP_PORT_STATE even > has to exist seperately from SWITCHDEV_ATTR_ID_PORT_STP_STATE, and > it's hard to tell what the difference might be since no kernel code > implements the former.The reason was to stay away from STP, because you can't run these two protocols at the same time. Even though in SW, we reuse port's state. In our driver(which is not upstreamed), we currently implement SWITCHDEV_ATTR_ID_MRP_PORT_STATE and just call the SWITCHDEV_ATTR_ID_PORT_STP_STATE.> > net/bridge/br_mrp_switchdev.c | 18 ++++++++++++++++++ > 1 file changed, 18 insertions(+) > > diff --git a/net/bridge/br_mrp_switchdev.c b/net/bridge/br_mrp_switchdev.c > index ed547e03ace1..8a1c7953e57a 100644 > --- a/net/bridge/br_mrp_switchdev.c > +++ b/net/bridge/br_mrp_switchdev.c > @@ -180,6 +180,24 @@ int br_mrp_port_switchdev_set_state(struct net_bridge_port *p, > int err; > > err = switchdev_port_attr_set(p->dev, &attr); > + if (err == -EOPNOTSUPP) { > + attr.id = SWITCHDEV_ATTR_ID_PORT_STP_STATE; > + switch (state) { > + case BR_MRP_PORT_STATE_DISABLED: > + case BR_MRP_PORT_STATE_NOT_CONNECTED: > + attr.u.stp_state = BR_STATE_DISABLED; > + break; > + case BR_MRP_PORT_STATE_BLOCKED: > + attr.u.stp_state = BR_STATE_BLOCKING; > + break; > + case BR_MRP_PORT_STATE_FORWARDING: > + attr.u.stp_state = BR_STATE_FORWARDING; > + break; > + default: > + return err; > + }; > + err = switchdev_port_attr_set(p->dev, &attr); > + } > if (err && err != -EOPNOTSUPP) > br_warn(p->br, "error setting offload MRP state on port %u(%s)\n", > (unsigned int)p->port_no, p->dev->name); > -- > 2.23.0 >-- /Horatiu