[vpp-dev] using vpp as pppoe client
Hello, I saw Hongjun.ni upload some plugin code and python case about using vpp for pppoe. But from my understanding, it looks like vpp is used as a pppoe server. Could you please give some detail info about how to use vpp as a pppoe client? If vpp is not able to do that, is there a roadmap to achieve it in your plan? And how to do nat through pppoe_session if vpp is used as a pppoe client? Your answer will be greatly appriciated! 谢谢祁晋 ___ vpp-dev mailing list vpp-dev@lists.fd.io https://lists.fd.io/mailman/listinfo/vpp-dev
Re: [vpp-dev] memmove fault
Decided it was the correct fix and just merged it, thanks. Chris. From:on behalf of "Luke, Chris" Date: Wednesday, September 20, 2017 at 07:04 To: "Yoann Desmouceaux (ydesmouc)" , Florin Coras , 薛欣颖 Cc: vpp-dev Subject: Re: [vpp-dev] memmove fault Thanks Yoann. I’ll review it properly in a bit, have a bit of a drive this morning before I can get to it properly however. On first reading though it seems that it always places the cursor at the end of the line when you cursor up or down. This is consistent with “bash” behavior. Will ponder if that is the behavior we want; I think I tried to deliberately avoid that, but this would be a simple range check if so. Thoughts anyone? Chris. From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On Behalf Of Yoann Desmouceaux (ydesmouc) Sent: Wednesday, September 20, 2017 4:22 To: Florin Coras ; 薛欣颖 Cc: vpp-dev Subject: Re: [vpp-dev] memmove fault Hello, I also stumbled upon this bug a few times. I just submitted https://gerrit.fd.io/r/#/c/8468, hopefully this fixes this issue. Cheers, Yoann From: on behalf of Florin Coras Date: Wednesday, 20 September 2017 at 08:10 To: 薛欣颖 Cc: vpp-dev Subject: Re: [vpp-dev] memmove fault Hi, Could this be a result of: type a command in the cli and don’t execute it, press up-arrow, try to input a new command in the empty prompt line, sigsegv? Florin P.S. Chris, apologies, I plead guilty for not trying to debug it … On Sep 19, 2017, at 10:52 PM, 薛欣颖 wrote: Hi Chris, Until today, the problem only appeared again: Program received signal SIGSEGV, Segmentation fault. __memmove_ssse3_back () at ../sysdeps/x86_64/multiarch/memcpy-ssse3-back.S:1663 1663../sysdeps/x86_64/multiarch/memcpy-ssse3-back.S: No such file or directory. (gdb) bt #0 __memmove_ssse3_back () at ../sysdeps/x86_64/multiarch/memcpy-ssse3-back.S:1663 #1 0x7796531f in unix_cli_line_process_one (cm=0x77ba6d80 , um=0x77ba77e0 , cf=0x7fffb5e5bf80, uf=0x7fffb5e5bf24, input=115 's', action=UNIX_CLI_PARSE_ACTION_NOMATCH) at /home/vpp/17.07_x64/vpp/build-data/../src/vlib/unix/cli.c:2065 #2 0x77965911 in unix_cli_line_edit (cm=0x77ba6d80 , um=0x77ba77e0 , cf=0x7fffb5e5bf80) at /home/vpp/17.07_x64/vpp/build-data/../src/vlib/unix/cli.c:2154 #3 0x77965c49 in unix_cli_process_input (cm=0x77ba6d80 , cli_file_index=0) at /home/vpp/17.07_x64/vpp/build-data/../src/vlib/unix/cli.c:2194 #4 0x77966814 in unix_cli_process (vm=0x77ba7080 , rt=0x7fffb5e5c000, f=0x0) at /home/vpp/17.07_x64/vpp/build-data/../src/vlib/unix/cli.c:2320 #5 0x77928699 in vlib_process_bootstrap (_a=140736239855696) at /home/vpp/17.07_x64/vpp/build-data/../src/vlib/main.c:1275 #6 0x7695fddc in clib_calljmp () at /home/vpp/17.07_x64/vpp/build-data/../src/vppinfra/longjmp.S:128 #7 0x7fffb5956820 in ?? () #8 0x779287ce in vlib_process_startup (vm=0x7795c5b2 , p=0x7fffb5956880, f=0x8496528d2a6d3500) at /home/vpp/17.07_x64/vpp/build-data/../src/vlib/main.c:1300 Backtrace stopped: previous frame inner to this frame (corrupt stack?) (gdb) f 1 #1 0x7796531f in unix_cli_line_process_one (cm=0x77ba6d80 , um=0x77ba77e0 , cf=0x7fffb5e5bf80, uf=0x7fffb5e5bf24, input=115 's', action=UNIX_CLI_PARSE_ACTION_NOMATCH) at /home/vpp/17.07_x64/vpp/build-data/../src/vlib/unix/cli.c:2065 2065 memmove (cf->current_command + cf->cursor + 1, (gdb) p cf->current_command $36 = (u8 *) 0x7fffb5e5b740 "Ahow interface ipsec-gre " (gdb) p cf->cursor $37 = 15 (gdb) p j $38 = -15 (gdb) Until now I have not figured out the rule that it appears. Thanks, xyxue From: Luke, Chris Date: 2017-08-17 04:54 To: 薛欣颖; vpp-dev@lists.fd.io Subject: RE: [vpp-dev] memmove fault Can you be more specific about the steps that led up to this issue? I won’t be able to work on it until September (about to go on vacation) but I do want to try to capture the details so I can try to recreate it. Chris. From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On Behalf Of ??? Sent: Tuesday, August 15, 2017 5:14 To: vpp-dev@lists.fd.io Subject: [vpp-dev] memmove fault Hi, When I failed to configure a command, then click the up key , there is a Segmentation fault(not every time): Program received signal SIGSEGV, Segmentation fault. __memmove_ssse3_back () at ../sysdeps/x86_64/multiarch/memcpy-ssse3-back.S:1664 1664../sysdeps/x86_64/multiarch/memcpy-ssse3-back.S: No such file or directory. (gdb) bt
Re: [vpp-dev] CentOS 7.4
After update you need reboot to install new kernel and make sure kernel headers are consistent with new kernel. I think the update "should" update all installed RPMs including those that were previously installed with make install-dep. nasm for Centos is updated outside of yum because nasm 2.12+ is still not available in downstream Centos distro. On 09/19/2017 04:03 PM, Burt Silverman wrote: Problem solved after 1)rebooting, 2) cd ~/vpp; rm -rf *; git checkout . 3) make install-dep; make bootstrap; make build-release. Possibly it was just the reboot and make install-dep, I am not sure, but I saw nasm getting updated at that point. Thanks, Tom, for hints. Burt On Mon, Sep 18, 2017 at 10:59 PM, Burt Silverman> wrote: I updated my CentOS using "yum update" and that leaves me with CentOS 7.4. So I get those fPIC errors that others have reported on non CentOS systems but that have glibc 2.17. Are others seeing the same thing? It looks to me that the official builds under CentOS are being done WITHOUT "yum update". Thanks. Burt -- *Thomas F Herbert* NFV and Fast Data Planes Office of Technology *Red Hat* ___ vpp-dev mailing list vpp-dev@lists.fd.io https://lists.fd.io/mailman/listinfo/vpp-dev
Re: [vpp-dev] Failed to build vpp with vppsb
Adding vppsb mailing list. Burt On Tue, Sep 19, 2017 at 6:05 PM, Burt Silvermanwrote: > Same problem here. I guess the vlib code was updated but the vppsb has not > caught up. > > Burt > > On Mon, Sep 18, 2017 at 12:56 PM, Paweł Staszewski > wrote: > >> Trying to add netlink from vppsb to the latest vpp git but : >> >> >> Building netlink in /git/vpp/build-root/build-vpp_debug-native/netlink >> >> make[1]: Entering directory '/git/vpp/build-root/build-vpp >> _debug-native/netlink' >> CC librtnl/netns.lo >> CC librtnl/rtnl.lo >> CC librtnl/mapper.lo >> CC test/test.lo >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:105:42: error: ‘struct >> unix_file’ declared inside parameter list will not be visible outside of >> this definition or declaration [-Werror] >> static clib_error_t *rtnl_read_cb(struct unix_file * f) >> ^ >> /git/vpp/build-data/../netlink/librtnl/rtnl.c: In function >> ‘rtnl_read_cb’: >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:109:33: error: >> dereferencing pointer to incomplete type ‘struct unix_file’ >>rtnl_ns_t *ns = >streams[f->private_data]; >> ^~ >> /git/vpp/build-data/../netlink/librtnl/rtnl.c: In function >> ‘rtnl_socket_close’: >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:141:3: error: implicit >> declaration of function ‘unix_file_del’ [-Werror=implicit-function-dec >> laration] >>unix_file_del(_main, _main.file_pool[ns->unix_index]); >>^ >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:141:39: error: >> ‘unix_main_t {aka struct }’ has no member named ‘file_pool’ >>unix_file_del(_main, _main.file_pool[ns->unix_index]); >>^ >> /git/vpp/build-data/../netlink/librtnl/rtnl.c: In function >> ‘rtnl_socket_open’: >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:262:3: error: unknown type >> name ‘unix_file_t’ >>unix_file_t template = {0}; >>^~~ >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:263:11: error: request for >> member ‘read_function’ in something not a structure or union >>template.read_function = rtnl_read_cb; >>^ >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:264:11: error: request for >> member ‘file_descriptor’ in something not a structure or union >>template.file_descriptor = ns->rtnl_socket; >>^ >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:265:11: error: request for >> member ‘private_data’ in something not a structure or union >>template.private_data = (uword) (ns - rm->streams); >>^ >> /git/vpp/build-data/../netlink/librtnl/rtnl.c:266:20: error: implicit >> declaration of function ‘unix_file_add’ [-Werror=implicit-function-dec >> laration] >>ns->unix_index = unix_file_add (_main, ); >> ^ >> cc1: all warnings being treated as errors >> Makefile:520: recipe for target 'librtnl/rtnl.lo' failed >> make[1]: *** [librtnl/rtnl.lo] Error 1 >> make[1]: *** Waiting for unfinished jobs >> /git/vpp/build-data/../netlink/test/test.c: In function >> ‘mapper_ns_add_command_fn’: >> /git/vpp/build-data/../netlink/test/test.c:125:14: error: implicit >> declaration of function ‘ip4_fib_index_from_table_id’ >> [-Werror=implicit-function-declaration] >>u32 fib4 = ip4_fib_index_from_table_id(table_id); >> ^~~ >> /git/vpp/build-data/../netlink/test/test.c:126:14: error: implicit >> declaration of function ‘ip6_fib_index_from_table_id’ >> [-Werror=implicit-function-declaration] >>u32 fib6 = ip6_fib_index_from_table_id(table_id); >> ^~~ >> cc1: all warnings being treated as errors >> Makefile:520: recipe for target 'test/test.lo' failed >> make[1]: *** [test/test.lo] Error 1 >> make[1]: Leaving directory '/git/vpp/build-root/build-vpp >> _debug-native/netlink' >> Makefile:698: recipe for target 'netlink-build' failed >> make: *** [netlink-build] Error 2 >> >> >> >> >> >> W dniu 2017-09-17 o 16:27, Paweł Staszewski pisze: >> >> Also latest vpp from git without including vppsb >>> >>> Sep 17 16:28:02 ubuntu systemd[1]: vpp.service: Service hold-off time >>> over, scheduling restart. >>> Sep 17 16:28:02 ubuntu systemd[1]: Stopped vector packet processing >>> engine. >>> Sep 17 16:28:02 ubuntu systemd[1]: Starting vector packet processing >>> engine... >>> Sep 17 16:28:02 ubuntu systemd[1]: Started vector packet processing >>> engine. >>> Sep 17 16:28:02 ubuntu vpp[1557]: vlib_plugin_early_init:356: plugin >>> path /usr/lib/vpp_plugins >>> Sep 17 16:28:02 ubuntu vpp[1557]: load_one_plugin:184: Loaded plugin: >>> acl_plugin.so (Access Control Lists) >>> Sep 17 16:28:02 ubuntu vpp[1557]: load_one_plugin:142: >>> /usr/lib/vpp_plugins/dpdk_plugin.so: undefined symbol: >>> aes_gcm_enc_128_sse >>> Sep 17 16:28:02 ubuntu vpp[1557]: load_one_plugin:143: Failed to load
Re: [vpp-dev] some issue about using unformat %u
Varargs functions effectively bypass strong type-checking. It can’t be helped. From: vpp-dev-boun...@lists.fd.io [mailto:vpp-dev-boun...@lists.fd.io] On Behalf Of wang.hu...@zte.com.cn Sent: Tuesday, September 19, 2017 11:34 PM To: vpp-dev@lists.fd.io Cc: gu.ji...@zte.com.cn; wang.ju...@zte.com.cn Subject: [vpp-dev] some issue about using unformat %u hi all: we found some common using issues about the use of CLI unformat, as follow: u16 out_port = 0; u32 vrf_id = 0, protocol; else if (unformat (line_input, "%U %u", unformat_ip4_address, _addr, _port)) when inputing u16 or u8 type param(not u32), the local variable which behind of "out_port" in stack will be overwrited, is that right? is there some Notes about this? I think the bellow code maybe cause that issue. unformat->va_unformat->do_percent->unformat_integer-> *(u32 *) v = value; 王辉 wanghui IT开发工程师 IT Development Engineer 虚拟化南京四部/无线研究院/无线产品经营部 NIV Nanjing Dept. IV/Wireless Product R&D Institute/Wireless Product Operation Division ___ vpp-dev mailing list vpp-dev@lists.fd.io https://lists.fd.io/mailman/listinfo/vpp-dev