在 2018/4/16 22:03, Parav Pandit 写道: > Hi, > >> -----Original Message----- >> From: linux-rdma-owner@xxxxxxxxxxxxxxx [mailto:linux-rdma- >> owner@xxxxxxxxxxxxxxx] On Behalf Of oulijun >> Sent: Monday, April 16, 2018 2:19 AM >> To: linux-rdma <linux-rdma@xxxxxxxxxxxxxxx> >> Cc: Linuxarm <linuxarm@xxxxxxxxxx> >> Subject: 【Ask for help】A question for connection management with vlan >> device >> >> Hi，All >> I am developing the cm for vlan device. I have a question need to help from >> you. > Cm for vlan device is support for IPv4 addresses. > Did it fail for you? If so, which client-server test commands did you try? > Yes, The test flow as follows: 1. I configure the vlan device ip; ifconfig eth0 192.168.1.100; vconfig add eth0 100; ifconfig eth0.100 192.168.1.1 root@(none)# ifconfig eth0 Link encap:Ethernet HWaddr A2:AF:57:53:A5:39 inet addr:192.168.1.100 Bcast:192.168.1.255 Mask:255.255.255.0 inet6 addr: fe80::a0af:57ff:fe53:a539/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:22 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 B) TX bytes:1732 (1.6 KiB) eth0.100 Link encap:Ethernet HWaddr A2:AF:57:53:A5:39 inet addr:192.168.1.1 Bcast:192.168.1.255 Mask:255.255.255.0 inet6 addr: fe80::a0af:57ff:fe53:a539/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:11 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:0 (0.0 B) TX bytes:866 (866.0 B) 2. I use perftest tool, the cmd as follows: server: ./ib_send_bw -n 5 -x 4 -R client: ./ib_send_bw -n 5 -x 4 -R 192.168.1.1 I also check the gid_attr, the info as follows: root@(none)# cat /sys/class/infiniband/hns_0/ports/1/gid_attrs/types/4 IB/RoCE v1 root@(none)# cat /sys/class/infiniband/hns_0/ports/1/gid_attrs/ndevs/4 eth0.100 thanks Lijun Ou > Cm for vlan device for non-link local IPv6 is also supported. > >> I grap the packet content for REQ packet. the value of primary local port GID >> field is the same when use vlan device with non-vlan device. > GID value will be same, but GID attribute will be different once ib_cache_gid_add() will be fixed to remove the default GID check for IPv6. > Till that time, please use non-link local IPv6 address. > >> I think that it is incorrect. I have a question after analysised the flow of cm. >> why the rec->sgid is fixed and the value is translated by the non-vlan device >> when >> use vlan device. >> >> The interface and code as follows: >> int ib_init_ah_attr_from_path(struct ib_device *device, u8 port_num, >> struct sa_path_rec *rec, >> struct rdma_ah_attr *ah_attr) >> { >> int ret; >> u16 gid_index; >> int use_roce; >> struct net_device *ndev = NULL; >> >> memset(ah_attr, 0, sizeof *ah_attr); >> ah_attr->type = rdma_ah_find_type(device, port_num); >> >> rdma_ah_set_dlid(ah_attr, be32_to_cpu(sa_path_get_dlid(rec))); >> >> if ((ah_attr->type == RDMA_AH_ATTR_TYPE_OPA) && >> (rdma_ah_get_dlid(ah_attr) == be16_to_cpu(IB_LID_PERMISSIVE))) >> rdma_ah_set_make_grd(ah_attr, true); >> >> rdma_ah_set_sl(ah_attr, rec->sl); >> rdma_ah_set_path_bits(ah_attr, be32_to_cpu(sa_path_get_slid(rec)) & >> get_src_path_mask(device, port_num)); >> rdma_ah_set_port_num(ah_attr, port_num); >> rdma_ah_set_static_rate(ah_attr, rec->rate); >> use_roce = rdma_cap_eth_ah(device, port_num); >> >> if (use_roce) { >> struct net_device *idev; >> struct net_device *resolved_dev; >> struct rdma_dev_addr dev_addr = { >> .bound_dev_if = ((sa_path_get_ifindex(rec) >= 0) ? >> sa_path_get_ifindex(rec) : 0), >> .net = sa_path_get_ndev(rec) ? >> sa_path_get_ndev(rec) : >> &init_net >> }; >> union { >> struct sockaddr _sockaddr; >> struct sockaddr_in _sockaddr_in; >> struct sockaddr_in6 _sockaddr_in6; >> } sgid_addr, dgid_addr; >> >> if (!device->get_netdev) >> return -EOPNOTSUPP; >> >> rdma_gid2ip(&sgid_addr._sockaddr, &rec->sgid); >> rdma_gid2ip(&dgid_addr._sockaddr, &rec->dgid); >> >> ... >> ... >> ... >> } >> >> From my point, I think that it is different with vlan device and non vlan device. >> >> Thanks >> Lijun Ou >> >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body >> of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at >> http://vger.kernel.org/majordomo-info.html > > -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
