Re: [vpp-dev] Query regarding MPLS Tunnel
Hi Neale, Sure, thanks a lot. Thanks & Regards, Sreejith On Fri, May 27, 2022 at 4:56 AM Neale Ranns wrote: > > > Hi Sreejith, > > > > No good reason. Please provide a patch to add that capability. > > > > /neale > > > > *From: *vpp-dev@lists.fd.io on behalf of sreejith n > via lists.fd.io > *Date: *Friday, 27 May 2022 at 00:34 > *To: *vpp-dev@lists.fd.io > *Subject: *[vpp-dev] Query regarding MPLS Tunnel > > Hi All, > > I have a query regarding Mpls Tunnel delete option. > > I have observed we can delete the mpls tunnel created using the CLI by > only passing the tunnel index. > > *CLI (To delete mpls tunnel):* > > > > *mpls tunnel del mpls-tunnel0* > > > > But in API option I have observed we cannot delete the mpls tunnel by > passing tunnel index alone we need to pass rpath details (mandatory) . > > > > *API:* > > *vl_api_mpls_tunnel_add_del_t_handler* > > > > *if (!vnet_mpls_tunnel_path_remove (tunnel_sw_if_index, rpaths))* >* vnet_mpls_tunnel_del (tunnel_sw_if_index);* > > > > I wanted to ask is there any specific reason to pass the rpath details in > the API option, can we delete the tunnel directly using the tunnel index > similar to CLI. > > > > In the CLI API, I have observed it is handled as below, the tunnel is > deleted directly using tunnel index if rpath is NULL: > > > > *CLI API:* > > *vnet_create_mpls_tunnel_command_fn * > > > > > > > > > > > *if (NULL == rpaths) { vnet_mpls_tunnel_del(sw_if_index); } > else if (!vnet_mpls_tunnel_path_remove(sw_if_index, rpaths)) { > vnet_mpls_tunnel_del(sw_if_index); }* > > > > Can we consider a similar handling for API option in > "vl_api_mpls_tunnel_add_del_t_handler". > > > > Thanks & Regards, > > Sreejith > > > > > > -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21476): https://lists.fd.io/g/vpp-dev/message/21476 Mute This Topic: https://lists.fd.io/mt/91355659/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
Re: [vpp-dev] Query regarding MPLS Tunnel
Hi Sreejith, No good reason. Please provide a patch to add that capability. /neale From: vpp-dev@lists.fd.io on behalf of sreejith n via lists.fd.io Date: Friday, 27 May 2022 at 00:34 To: vpp-dev@lists.fd.io Subject: [vpp-dev] Query regarding MPLS Tunnel Hi All, I have a query regarding Mpls Tunnel delete option. I have observed we can delete the mpls tunnel created using the CLI by only passing the tunnel index. CLI (To delete mpls tunnel): mpls tunnel del mpls-tunnel0 But in API option I have observed we cannot delete the mpls tunnel by passing tunnel index alone we need to pass rpath details (mandatory) . API: vl_api_mpls_tunnel_add_del_t_handler if (!vnet_mpls_tunnel_path_remove (tunnel_sw_if_index, rpaths)) vnet_mpls_tunnel_del (tunnel_sw_if_index); I wanted to ask is there any specific reason to pass the rpath details in the API option, can we delete the tunnel directly using the tunnel index similar to CLI. In the CLI API, I have observed it is handled as below, the tunnel is deleted directly using tunnel index if rpath is NULL: CLI API: vnet_create_mpls_tunnel_command_fn if (NULL == rpaths) { vnet_mpls_tunnel_del(sw_if_index); } else if (!vnet_mpls_tunnel_path_remove(sw_if_index, rpaths)) { vnet_mpls_tunnel_del(sw_if_index); } Can we consider a similar handling for API option in "vl_api_mpls_tunnel_add_del_t_handler". Thanks & Regards, Sreejith -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21474): https://lists.fd.io/g/vpp-dev/message/21474 Mute This Topic: https://lists.fd.io/mt/91355659/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
Re: [vpp-dev] Regarding DES support in VPP
Hi Vijay, It sounds like the SA you programme did not install. As you say, DES is insecure, so we don’t even test it anymore. I would suggest you start with a UT in VPP and go from there. Maybe extend the algos in MyParameters in test/test_ipsec_esp.py /neale From: vpp-dev@lists.fd.io on behalf of Vijay Kumar via lists.fd.io Date: Thursday, 26 May 2022 at 21:52 To: vpp-dev Subject: [vpp-dev] Regarding DES support in VPP Hi Neale/Benoit, I know we must not talk about DES and MD5 these days as they are insecure and must not be configured. My QA has raised an issue that DES is not working. I have myself not tested it as the customers would never configure it. The QA says the "show ipsec sa" command does not show anything if DES is configured. Also the show node counters had this counter incremented for DES. "4 ipsec4-tun-input no matching tunnel" Not sure if I am missing something. NOTE: == 1) We don't use the vpp ikev2 plugin. We have our own IKE stack that programs the VPP with IPSEC SA. Basically our application receives the SA and calls the ipsec_sa_add_and_lock() API to install the SA. 2) We have tested AES128, ASE256, 3DES and they were working fine. The code to receive keys from IKE stack and program the vnet/ipsec is the same. Regards, Vijay Kumar N. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21473): https://lists.fd.io/g/vpp-dev/message/21473 Mute This Topic: https://lists.fd.io/mt/91352430/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
Re: [vpp-dev] VPP v22.10-rc0~1 can't inject routes from Linux Control Plane.
On Thu, May 26, 2022 at 08:03 AM, Fabio Bizzi wrote: > > GigabitEthernet6/0/0 Hi, How did you manage to get GigabitEthernet6/0/0 in VPP? As there is no dpdk{} section for whitelisting kernel interfaces. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21472): https://lists.fd.io/g/vpp-dev/message/21472 Mute This Topic: https://lists.fd.io/mt/91356280/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
[vpp-dev] VPP v22.10-rc0~1 can't inject routes from Linux Control Plane.
Hello, I'm in trouble with VPP and frr/static routing. My setup is the following: - O.S. Ubuntu 20.04LTS server - VPP from https://gerrit.fd.io/r/vpp (git clone) v22.10-rc0~1 - FRR from frr deb repository To build/configure all I followed the great Pim van Pelt guide at ipng.ch, adapting it to the 22.10 version. ("ip netns exec dataplane bash --login" typed at the beginning) This is my /etc/vpp/startup.conf root@vpp-router:~# cat /etc/vpp/startup.conf unix { nodaemon log /var/log/vpp/vpp.log full-coredump cli-listen /run/vpp/cli.sock gid vpp ## This makes VPP sleep 1ms between each DPDK poll, greatly ## reducing CPU usage, at the expense of latency/throughput. poll-sleep-usec 1000 ## Execute all CLI commands from this file upon startup exec /etc/vpp/bootstrap.vpp } api-trace { on } api-segment { gid vpp } socksvr { default } memory { main-heap-size 512M main-heap-page-size default-hugepage } buffers { buffers-per-numa 128000 default data-size 2048 page-size default-hugepage } statseg { size 512M page-size default-hugepage per-node-counters off } plugins { plugin linux_nl_plugin.so { enable } plugin linux_cp_plugin.so { enable } } logging { default-log-level info default-syslog-log-level notice class linux-cp/if { rate-limit 1 level debug syslog-level debug } class linux-cp/nl { rate-limit 1 level debug syslog-level debug } } and this is my /etc/vpp/bootstrap.vpp root@vpp-router:~# cat /etc/vpp/bootstrap.vpp comment { Set lcp defaults } lcp default netns dataplane lcp lcp-sync on lcp lcp-auto-subint on comment { Create a loopback interface } create loopback interface instance 0 lcp create loop0 host-if loop0 set interface state loop0 up set interface ip address loop0 2a02:4d80:602:11::1/128 set interface ip address loop0 10.0.0.1/32 comment { Create Linux Control Plane interfaces } lcp create GigabitEthernet6/0/0 host-if gbe0 lcp create GigabitEthernet6/0/1 host-if gbe1 lcp create GigabitEthernet7/0/0 host-if gbe2 lcp create GigabitEthernet7/0/1 host-if gbe3 set interface state GigabitEthernet6/0/0 up set interface state GigabitEthernet7/0/1 up comment { Create Subinterfaces and set IP Addresses } create sub GigabitEthernet6/0/0 100 set interface state GigabitEthernet6/0/0.100 up set interface ip address GigabitEthernet6/0/0.100 172.16.1.2/24 create sub GigabitEthernet7/0/1 100 set interface state GigabitEthernet7/0/1.100 up set interface ip address GigabitEthernet7/0/1.100 192.168.1.1/24 All works fine and under the kernel dataplane namespace and under vpp I see all the interfaces, the one with ip addresses ping their gateways. Frr (started in the dataplane netns) has a bgp peer on gbe0.100 interface and it learns the 192.168.4.0/24 network from it's peer and inject the route in the dataplane RIB: root@vpp-router:~# vtysh Hello, this is FRRouting (version 8.2.2). Copyright 1996-2005 Kunihiro Ishiguro, et al. vpp-router# show ip bgp summary IPv4 Unicast Summary (VRF default): BGP router identifier 10.0.0.1, local AS number 65000 vrf-id 0 BGP table version 2 RIB entries 3, using 552 bytes of memory Peers 1, using 723 KiB of memory NeighborV AS MsgRcvd MsgSent TblVer InQ OutQ Up/Down State/PfxRcd PfxSnt Desc 172.16.1.1 4 89686971000 01:06:1012 N/A Total number of neighbors 1 vpp-router# show ip bgp BGP table version is 2, local router ID is 10.0.0.1, vrf id 0 Default local pref 100, local AS 65000 Status codes: s suppressed, d damped, h history, * valid, > best, = multipath, i internal, r RIB-failure, S Stale, R Removed Nexthop codes: @NNN nexthop's vrf id, < announce-nh-self Origin codes: i - IGP, e - EGP, ? - incomplete RPKI validation codes: V valid, I invalid, N Not found Network Next HopMetric LocPrf Weight Path *> 192.168.1.0/24 0.0.0.0 0 32768 i *> 192.168.4.0/24 172.16.1.1 0 0 8968 i Displayed 2 routes and 2 total paths vpp-router# q root@vpp-router:~# ip route show 172.16.1.0/24 dev gbe0.100 proto kernel scope link src 172.16.1.2 192.168.1.0/24 dev gbe3.100 proto kernel scope link src 192.168.1.1 192.168.4.0/24 nhid 19 via 172.16.1.1 dev gbe0.100 proto bgp metric 20 root@vpp-router:~# But when I try to look at the vpp fib, the entry for 192.168.4.0/24 is missing. :( vpp# show ip fib 192.168.4.0/24 ipv4-VRF:0, fib_index:0, flow hash:[src dst sport dport proto flowlabel ] epoch:0 flags:none locks:[adjacency:1, default-route:1, lcp-rt:1, ] 0.0.0.0/0 fib:0 index:0 locks:2 default-route refs:1 entry-flags:drop, src-flags:added,contributing,active, path-list:[0] locks:2 flags:drop, uPRF-list:0 len:0 itfs:[] path:[0] pl-index:0 ip4 weight=1 pref=0 special: cfg-flags:drop, [@0]: dpo-drop ip4 forwarding: unicast-ip4-chain [@0]: dpo-load-balance: [proto:ip4 index:1 buckets:1 uRPF:0 to:[0:0]]
[vpp-dev] New CSIT dashboard with VPP and DPDK performance data
Hi, Following from Tibor's demo of the new CSIT dashboard on the VPP call last Tuesday, here the links to what was presented. Note that the backend and frontend apps are still being tuned, so please treat this as an engineering preview. Glitches and disruption are likely to occur :) - Current landing page: http://csit.fd.io/ + Performance trending - dynamic dashboard with user defined views. + Job statistics - csit perf jobs execution data, richer then jenkins. - Functionality highlights + User selectable inputs * DUT, Infra, Area, Test * Frame size, Number of cores, Test type + Graphs with detected anomaly highlights * Throughput graphs for MRR, PDR and NDR test types * Latency graphs for PDR test types * Data point click for off-canvas display of details + Saving options * Custom view encoded in the URL * Custom view test data in csv file - Some example views with recent anomalies + avf ip4 2n-clx - regression on 04/18 [1] + avf ip4 2n-clx 2n-icx - regression 04/18 [2] + added dpdk 2n-clx ip4 - progression 03/15 [3] + dpdk ip4 base 1c 2c 4c 3n-alt - multi-core speed-up [4] + dpdk ip6 3n-alt - regression 05/17 [5] Please use csit-...@lists.fd.io for all feedback and comments. Cheers, Maciek [1] avf ip4 2n-clx, http://csit.fd.io/trending/#eNrVkk0KwjAQhU9TNzKQjq1dufDnHhLTUYtNDUks1dObSmEs6E4kXWRC-B4zk8dz_mpp76heJfkmKTYJFlUZSrJYz8PVGgPYgKo7wPxEaNKua4tUgGyPUJkMltkBUgXkz_0rHKdkTSjEBbS1fSPc9o3Kmx91ZWLOdybfZrFcWpKsDwMZeXJvQz6sxNKjlZpc9SDWh58wV8EVRqkaD_F380aHfxa7l-KnFtqmjNDGfq1pWRlfGC9TMzDGKE4viTo2D_XE7IswhvofKcx3M-el9SsUiCAQEOdCBNlQZtSUA8sBsxF7AkMkH_s [2] avf ip4 2n-clx 2n-icx, http://csit.fd.io/trending/#eNrVkEEOgjAQRU-DGzNJGUFWLFTuYWoZhQjYtJWgp7cYkpFEd0ZhQUnzfue3z7qLob2lKg3ibZBsA0zK3C_BarP0v1ZrwAZU1QHGJ0Iddl2bhAJke4RSR7CODhAqIFf0O_9ZJStCIc5gmhxqY_phuOuH5Vc3msxEFzcmn_o4Lg1JzvtSRo7sS8mHa3H8aGRNtrwTn_EvYq68HUahGhe5m36hw1uT7Jn4nsppmpyfyHqKHut_ayzVLzW-bZuLxjhbWCeNS1EggkBAXArhY8OyoCYfWAwYjdgD2tl11g [3] added dpdk 2n-clx ip4, http://csit.fd.io/trending/#eNrVkk0OgjAQhU-DGzJJGUFWLkTvYWo7KuHHpq0EPb3FkIws3BkDi07TfK_zOi91_mbp6KjeRlkR5UWEealDida7OGydMYAtqLoHzC6EJun7Lk8EyO4MpUlhk54gUUD-OpzCckrWhEJUYFsNjbVDM9wPzfTdTzozMdcHk29-LJeWJOuDKSNP7sPky7NYfrayIVc-ie-EiZirkA6jRE2N_MN80HHW_PBW_C7KeSa5vCCbOebYzC9GbXT1zw85-C3mR2aHlfPS-i0KRBAIiLEQQTaWFbV6ZBlgOmEv6bp3BA [4] dpdk ip4 base 1c 2c 4c 3n-alt, http://csit.fd.io/trending/#eNrdkMEOgjAQRL8GL2STshQ5eRD5D1PpKkTApq1G_HqLIVk5mBgTLxy6TfMmM9tx_mJp76jdRFkR5UWEeaPDiNJtHK6bMZD2oFoPUpwITXJv80SANvoMjZGwlgdIKiBfj69wDsoRdNaOHrgbPfTVzwyZmHpg8imG5cqSYn3IYuTJvYXMt2HV0aqOXPMglob9mVehC0ZJNff3g3mj0xfz8qX4rThcZnH49-LkMouTXxaXlSvnlfUbFIggEBBjIYJsGivq9cQyQDljTy_DDDc [5] dpdk ip6 3n-alt, http://csit.fd.io/trending/#eNrVkM0OgjAQhJ8GL2aTsoL1wsGf9zCVrkJEbdpqxKe3GOLSgzdN8NBtmm8ysx3nL5a2jpoiyVeJXCUoax1GMltOw3UzBmZnUI2HTBwITXpvZCpAG32E2sxBLnaQlkC-6l7h7JQjOFnbeeC689BXHxkyMVXL5FMMy5UlxfqQxciTG4TE27Bqb9WJXP0glob9mZehC0ZpGfv71gxo_0W5eSm-UJwrVUMoxHE87b1X-q8Kx9fgzwvMNxPnlfUFCkQQCIhTIYKsHxM6657lgFnEnq0vEbk -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21469): https://lists.fd.io/g/vpp-dev/message/21469 Mute This Topic: https://lists.fd.io/mt/91352683/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
Re: [vpp-dev] Help needed in bringing up memif with shm_open.
Hi , Thanks Ben and Daniel for looking into this. Both the issues which I was facing is resolved now. steps that i took was : I added "shm_unlink" after "shm_open call". For the rx issue which I was getting , as suggested by Ben i used gdb to debug into the libmemif. and I realised a "memif_refill_queue" call is required on slave before master can transmit data, calling memif_refill_queue() on slave provides buffers for master. Are you able to run with shm_open without using shm_unlink ? If you could please post what was your observation. Thank You Pragya Nand On Wed, May 25, 2022 at 1:16 PM Daniel Béreš via lists.fd.io wrote: > Hi, > > To sync up, can you run loopback example? > > I tried to reproduce your problem on 3.10.0-229.1.2.rt56.141.2.el7_1.x86_64 > but I got different result. > > diff --git a/extras/libmemif/src/CMakeLists.txt > b/extras/libmemif/src/CMakeLists.txt > index ddb8a52..a11a65f 100644 > --- a/extras/libmemif/src/CMakeLists.txt > +++ b/extras/libmemif/src/CMakeLists.txt > @@ -41,6 +41,12 @@ if(LIB_BSD) >target_link_libraries(memif ${LIB_BSD}) > endif() > > +find_library(LIB_LRT rt) > +if(LIB_BSD) > + add_compile_definitions(HAS_LIB_LRT) > + target_link_libraries(memif ${LIB_LRT}) > +endif() > + > foreach(file ${MEMIF_HEADERS}) >get_filename_component(dir ${file} DIRECTORY) > install( > diff --git a/extras/libmemif/src/main.c b/extras/libmemif/src/main.c > index 21b3994..132ce8d 100644 > --- a/extras/libmemif/src/main.c > +++ b/extras/libmemif/src/main.c > @@ -184,6 +184,7 @@ memif_get_version_str () > > #define DBG_TX_BUF (0) > #define DBG_RX_BUF (1) > +#define MEMIF_DBG_SHM 1 > > #ifdef MEMIF_DBG_SHM > static void > @@ -1234,13 +1235,17 @@ memif_add_region (memif_connection_t *conn, uint8_t > has_buffers) >r->region_size = (has_buffers == 0) ? r->buffer_offset : r->buffer_offset + > conn->run_args.buffer_size * (1 << conn->run_args.log2_ring_size) * > (conn->run_args.num_s2m_rings + conn->run_args.num_m2s_rings); > - > +/* >if ((r->fd = memfd_create ("memif region 0", MFD_ALLOW_SEALING)) == -1) > return memif_syscall_error_handler (errno); > +*/ > + if ((r->fd = shm_open ("memif region 0", O_RDWR | O_CREAT, S_IRWXU)) == -1) > +return memif_syscall_error_handler (errno); > > +/* >if ((fcntl (r->fd, F_ADD_SEALS, F_SEAL_SHRINK)) == -1) > return memif_syscall_error_handler (errno); > - > +*/ >if ((ftruncate (r->fd, r->region_size)) == -1) > return memif_syscall_error_handler (errno); > > > > > -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21468): https://lists.fd.io/g/vpp-dev/message/21468 Mute This Topic: https://lists.fd.io/mt/91224655/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
[vpp-dev] Regarding DES support in VPP
Hi Neale/Benoit, I know we must not talk about DES and MD5 these days as they are insecure and must not be configured. My QA has raised an issue that DES is not working. I have myself not tested it as the customers would never configure it. The QA says the "show ipsec sa" command does not show anything if DES is configured. Also the show node counters had this counter incremented for DES. "4 ipsec4-tun-input no matching tunnel" Not sure if I am missing something. NOTE: == 1) We don't use the vpp ikev2 plugin. We have our own IKE stack that programs the VPP with IPSEC SA. Basically our application receives the SA and calls the ipsec_sa_add_and_lock() API to install the SA. 2) We have tested AES128, ASE256, 3DES and they were working fine. The code to receive keys from IKE stack and program the vnet/ipsec is the same. Regards, Vijay Kumar N. -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21467): https://lists.fd.io/g/vpp-dev/message/21467 Mute This Topic: https://lists.fd.io/mt/91352430/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
Re: [vpp-dev] Committer update - Out hiking
Ole Troan writes: > Guys, > > Apologies for the noise on the list, but just wanted to let you know that I'm > out hiking from now until end of October. > Doing the PCT. At least that's the plan. > > I expect the VPP project will happily continue without me, but please expect > a little delay in reviews and merges (i.e. about 5 months. :-)) > > Best regards and see you on the other side, > Ole Have a great trip Ole! > -- Regards, Ray K -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#21466): https://lists.fd.io/g/vpp-dev/message/21466 Mute This Topic: https://lists.fd.io/mt/91336628/21656 Group Owner: vpp-dev+ow...@lists.fd.io Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-