Hi Bob, I still can not reproduce your issue with similar setup. I build the kernel module with make rpm-fedora-kmod with vanilla 2.11.1. I think O2 is enabled by default. Here are some logs on my testbed.
[root@localhost openvswitch]# uname -a Linux localhost.localdomain 3.10.0-514.el7.x86_64 #1 SMP Tue Nov 22 16:42:41 UTC 2016 x86_64 x86_64 x86_64 GNU/Linux [root@localhost openvswitch]# gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/libexec/gcc/x86_64-redhat-linux/4.8.5/lto-wrapper Target: x86_64-redhat-linux Configured with: ../configure --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-boot strap --enable-shared --enable-threads=posix --enable-checking=release --with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions --enable-gnu-u nique-object --enable-linker-build-id --with-linker-hash-style=gnu --enable-languages=c,c++,objc,obj-c++,java,fortran,ada,go,lto --enable-plugin --enable-ini tfini-array --disable-libgcj --with-isl=/builddir/build/BUILD/gcc-4.8.5-20150702/obj-x86_64-redhat-linux/isl-install --with-cloog=/builddir/build/BUILD/gcc-4 .8.5-20150702/obj-x86_64-redhat-linux/cloog-install --enable-gnu-indirect-function --with-tune=generic --with-arch_32=x86-64 --build=x86_64-redhat-linux Thread model: posix gcc version 4.8.5 20150623 (Red Hat 4.8.5-16) (GCC) [root@localhost openvswitch]# modinfo openvswitch filename: /lib/modules/3.10.0-514.el7.x86_64/extra/openvswitch/openvswitch.ko alias: net-pf-16-proto-16-family-ovs_ct_limit alias: net-pf-16-proto-16-family-ovs_meter alias: net-pf-16-proto-16-family-ovs_packet alias: net-pf-16-proto-16-family-ovs_flow alias: net-pf-16-proto-16-family-ovs_vport alias: net-pf-16-proto-16-family-ovs_datapath version: 2.11.1 license: GPL description: Open vSwitch switching datapath rhelversion: 7.3 srcversion: 04A91192E0D402BB3726A12 depends: nf_conntrack,tunnel6,nf_nat,nf_defrag_ipv6,libcrc32c,nf_nat_ipv6,nf_nat_ipv4 vermagic: 3.10.0-514.el7.x86_64 SMP mod_unload modversions parm: udp_port:Destination UDP port (ushort) [root@localhost openvswitch]# dmesg | tail -30 [ 23.781617] XFS (sda1): Mounting V5 Filesystem [ 23.908973] XFS (sda1): Ending clean mount [ 25.005344] floppy0: no floppy controllers found [ 25.912844] type=1305 audit(1578099898.496:4): audit_pid=818 old=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:auditd_t:s0 res=1 [ 38.493488] nf_conntrack version 0.5.0 (16384 buckets, 65536 max) [ 40.156435] NET: Registered protocol family 40 [ 45.531089] IPv6: ADDRCONF(NETDEV_UP): ens33: link is not ready [ 45.534915] e1000: ens33 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None [ 45.538020] IPv6: ADDRCONF(NETDEV_UP): ens33: link is not ready [ 45.538037] IPv6: ADDRCONF(NETDEV_CHANGE): ens33: link becomes ready [ 70.625134] ip6_tables: (C) 2000-2006 Netfilter Core Team [ 71.584401] Ebtables v2.0 registered [ 76.300066] bridge: automatic filtering via arp/ip/ip6tables has been deprecated. Update your scripts to load br_netfilter if you need this. [ 76.763055] tun: Universal TUN/TAP device driver, 1.6 [ 76.763058] tun: (C) 1999-2004 Max Krasnyansky <[email protected]> [ 76.768141] device virbr0-nic entered promiscuous mode [ 80.673242] virbr0: port 1(virbr0-nic) entered listening state [ 80.673248] virbr0: port 1(virbr0-nic) entered listening state [ 80.673374] IPv6: ADDRCONF(NETDEV_UP): virbr0: link is not ready [ 81.049899] virbr0: port 1(virbr0-nic) entered disabled state [ 713.096440] openvswitch: loading out-of-tree module taints kernel. [ 713.096606] openvswitch: module verification failed: signature and/or required key missing - tainting kernel [ 713.099132] openvswitch: Open vSwitch switching datapath 2.11.1 [ 713.099482] openvswitch: LISP tunneling driver [ 713.099483] GRE over IPv4 demultiplexor driver [ 713.100254] openvswitch: GRE over IPv4 tunneling driver [ 713.100598] openvswitch: GRE over IPv6 tunneling driver [ 713.100929] openvswitch: Geneve tunneling driver [ 713.100938] openvswitch: VxLAN tunneling driver [ 713.100940] openvswitch: STT tunneling driver -Yi-Hung On Mon, Dec 30, 2019 at 6:25 PM 杜勃勃 Bob Du <[email protected]> wrote: > > Hi Yi-Hung, > My build method is as follows: > 1. ./configure --with-linux=/lib/modules/$(uname -r)/build --prefix=/usr > --localstatedir=/var --sysconfdir=/etc > 2. make rpm-fedora > 3. make rpm-fedora-kmod > the kernel ovs ko load operate is: modprobe openvswitch > > > 在 2019/12/31 上午10:10，“杜勃勃 Bob Du”<[email protected]> 写入: > > Hi Yi-Hung, > My os is CentOS Linux release 7.2.1511 (Core),the kernel version is > 3.10.0-514 with the arch of x86_64,the gcc is gcc-4.8.5-16.el7.x86_64. > I think the problem is triggered by the gcc internal inline issue which > the default gcc option with O2 has the problem. > However, the O1 option has no problem. > > 在 2019/12/31 上午3:46，“Yi-Hung Wei”<[email protected]> 写入: > > Hi Bob, > > Thanks for reporting this issue. I can not reproduce your issue on my > Ubuntu 16.04 VM with 4.4.0-170 kernel. > > Can you report the distribution and kernel version of your host > system, and how you build and load the kernel module? > > -Yi-Hung > > > On Wed, Dec 25, 2019 at 8:53 AM 杜勃勃 Bob Du <[email protected]> > wrote: > > > > Hi ovs-dev, > > > > After we upgrade ovs to the 2.11.1, because our program enable > the config of CONFIG_NF_CONNTRACK=m, our program can not be loaded after do > it, the error info as follows: > > > > unable to find symbol 'ovs_ct_get_labels' > > > > > > > > we find that openvswitch.ko has no ovs_ct_get_labels, Is this a bug > in ovs 2.11.1? > > > > > > > > _______________________________________________ > > discuss mailing list > > [email protected] > > https://mail.openvswitch.org/mailman/listinfo/ovs-discuss > > > > _______________________________________________ discuss mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-discuss
