Commit b3612ccdf284 ("net: dsa: microchip: implement multi-bridge support") plugged a packet leak between ports that were members of different bridges. Unfortunately, this broke another use case, namely that of more than two ports that are members of the same bridge. After that commit, when a port is added to a bridge, hardware bridging between other member ports of that bridge will be cleared, preventing packet exchange between them. Fix by ensuring that the Port VLAN Membership bitmap includes any existing ports in the bridge, not just the port being added. Upstream commit 3d00827a90db6f79abc7cdc553887f89a2e0a184, backported to 5.16. Fixes: b3612ccdf284 ("net: dsa: microchip: implement multi-bridge support") Signed-off-by: Svenning Sørensen <sss@xxxxxxxxxxx> Tested-by: Oleksij Rempel <o.rempel@xxxxxxxxxxxxxx> Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx> --- drivers/net/dsa/microchip/ksz_common.c | 26 +++++++++++++++++++++++--- 1 file changed, 23 insertions(+), 3 deletions(-) diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c index 8a04302018dc..7ab9ab58de65 100644 --- a/drivers/net/dsa/microchip/ksz_common.c +++ b/drivers/net/dsa/microchip/ksz_common.c @@ -26,7 +26,7 @@ void ksz_update_port_member(struct ksz_device *dev, int port) struct dsa_switch *ds = dev->ds; u8 port_member = 0, cpu_port; const struct dsa_port *dp; - int i; + int i, j; if (!dsa_is_user_port(ds, port)) return; @@ -45,13 +45,33 @@ void ksz_update_port_member(struct ksz_device *dev, int port) continue; if (!dp->bridge_dev || dp->bridge_dev != other_dp->bridge_dev) continue; + if (other_p->stp_state != BR_STATE_FORWARDING) + continue; - if (other_p->stp_state == BR_STATE_FORWARDING && - p->stp_state == BR_STATE_FORWARDING) { + if (p->stp_state == BR_STATE_FORWARDING) { val |= BIT(port); port_member |= BIT(i); } + /* Retain port [i]'s relationship to other ports than [port] */ + for (j = 0; j < ds->num_ports; j++) { + const struct dsa_port *third_dp; + struct ksz_port *third_p; + + if (j == i) + continue; + if (j == port) + continue; + if (!dsa_is_user_port(ds, j)) + continue; + third_p = &dev->ports[j]; + if (third_p->stp_state != BR_STATE_FORWARDING) + continue; + third_dp = dsa_to_port(ds, j); + if (third_dp->bridge_dev == dp->bridge_dev) + val |= BIT(j); + } + dev->dev_ops->cfg_port_member(dev, i, val | cpu_port); } -- 2.20.1 Hi Greg, I'm sorry about the crippled patch in my last message. Unfortunately, I have no other means than the MS webmail client to send mail, and it seems it corrupts plain text when replying to a thread. Therefore, I send this as a new message - I really hope it will work this time. Best regards, Svenning