[Bug 1206674] New: No network (wicked) after 23-12-2022 updates
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 Bug ID: 1206674 Summary: No network (wicked) after 23-12-2022 updates Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: x86-64 OS: openSUSE Tumbleweed Status: NEW Severity: Major Priority: P5 - None Component: Network Assignee: screening-team-bugs@suse.de Reporter: alessandro.sturniolo@gmail.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- Created attachment 863670 --> http://bugzilla.opensuse.org/attachment.cgi?id=863670&action=edit Network related log files and configurations. Today (23-12-2022) I have updated my system, and after reboot (there was a new kernel version), network no longer worked. After a while I realized that the interfaces remained down, so if I manually bring them up (ip link set devName up), network came back to works. Currently, after every system restart, I have to bring up manually my network interfaces, to get the network to works. Obviously before today updates, my network it always worked without problems for years. My network setup is based on wicked (no NetworkManager), has two network interfaces (enp3s0 and enp4s0) configured as slaves of a bonding device (bond0). I followed these instructions https://en.opensuse.org/index.php?title=openSUSE:Bugreport_wicked and I've collected some files that I attach to this report. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 Alessandro Sturniolo <alessandro.sturniolo@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P5 - None |P2 - High -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 Andreas Stieger <Andreas.Stieger@gmx.de> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P2 - High |P5 - None -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c1 Alexey Bobrov <alex.bobrov@mail.ru> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |alex.bobrov@mail.ru --- Comment #1 from Alexey Bobrov <alex.bobrov@mail.ru> --- Similarly # journalctl | grep bond0 ������ 28 12:04:41 srv-mail wicked[1801]: received new ready device bond0 (/org/opensuse/Network/Interface/5) ������ 28 12:04:41 srv-mail wicked[1801]: bond0: refresh <client-state> control: persistent=false, usercontrol=false, require-link=auto ������ 28 12:04:41 srv-mail wicked[1801]: bond0: refresh <client-state> config: origin=compat:suse:/etc/sysconfig/network/ifcfg-bond0, uuid=e483643d-7a31-5cd2-8932-39fd7ab4793b, owner-uid=4294967295 ������ 28 12:04:41 srv-mail wicked[1801]: bond0: changed state none -> device-exists ������ 28 12:04:41 srv-mail wicked[1801]: ni_sysconfig_read(/etc/sysconfig/network/ifcfg-bond0) ������ 28 12:04:41 srv-mail wicked[1801]: ni_compat_generate_interfaces: compat:suse:/etc/sysconfig/network/ifcfg-bond0:0 ������ 28 12:04:41 srv-mail wicked[1801]: bond0: <device> identified device as "enp15s0f1" ������ 28 12:04:41 srv-mail wicked[1801]: bond0: resolved reference bond/slaves/slave/device to subordinate device enp15s0f1 ������ 28 12:04:41 srv-mail wicked[1801]: enp15s0f1 (<orphan xml node>): setting master device to bond0 ������ 28 12:04:41 srv-mail wicked[1801]: bond0 [5] ������ 28 12:04:41 srv-mail wicked[1801]: bond0: hiring nanny ������ 28 12:04:41 srv-mail wicked[1801]: bond0: converting config into policy 'policy__bond0' ������ 28 12:04:41 srv-mail wicked[1801]: bond0: adding policy policy__bond0 to nanny ������ 28 12:04:41 srv-mail wicked[1801]: Calling /org/opensuse/Network/Nanny.createPolicy(policy__bond0) ������ 28 12:04:41 srv-mail wicked[1801]: Call to /org/opensuse/Network/Nanny.createPolicy(policy__bond0) failed: Policy already exists ������ 28 12:04:41 srv-mail wicked[1801]: Calling /org/opensuse/Network/Nanny/Policy/policy__bond0.update() ������ 28 12:04:41 srv-mail wicked[1801]: Successfully updated nanny policy /org/opensuse/Network/Nanny/Policy/policy__bond0 ������ 28 12:04:41 srv-mail wicked[1801]: bond0: nanny hired! ������ 28 12:04:41 srv-mail wicked[1801]: bond0: configuration applied to nanny ������ 28 12:04:48 srv-mail wicked[1801]: received refresh for ready device bond0 (/org/opensuse/Network/Interface/5) ������ 28 12:04:48 srv-mail wicked[1801]: bond0: refresh <client-state> control: persistent=false, usercontrol=false, require-link=auto ������ 28 12:04:48 srv-mail wicked[1801]: bond0: refresh <client-state> config: origin=compat:suse:/etc/sysconfig/network/ifcfg-bond0, uuid=e483643d-7a31-5cd2-8932-39fd7ab4793b, owner-uid=4294967295 ������ 28 12:04:48 srv-mail wicked[1801]: received refresh for ready device bond0 (/org/opensuse/Network/Interface/5) ������ 28 12:04:48 srv-mail wicked[1801]: bond0: refresh <client-state> control: persistent=false, usercontrol=false, require-link=auto ������ 28 12:04:48 srv-mail wicked[1801]: bond0: refresh <client-state> config: origin=compat:suse:/etc/sysconfig/network/ifcfg-bond0, uuid=e483643d-7a31-5cd2-8932-39fd7ab4793b, owner-uid=4294967295 # ip a 3: enp15s0f1: <BROADCAST,MULTICAST,SLAVE> mtu 1500 qdisc pfifo_fast master bond0 state DOWN group default qlen 1000 link/ether 00:e0:81:b1:c0:1d brd ff:ff:ff:ff:ff:ff ............ 5: bond0: <BROADCAST,MULTICAST,MASTER> mtu 1500 qdisc noop state DOWN group default qlen 1000 link/ether 7a:fa:1d:9a:b8:0f brd ff:ff:ff:ff:ff:ff -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c6 Graeme Spence <email@graemespence.co.uk> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |email@graemespence.co.uk --- Comment #6 from Graeme Spence <email@graemespence.co.uk> --- https://bugzilla.kernel.org/show_bug.cgi?id=216844 Is it a Mediatek MT7921? -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c7 --- Comment #7 from Andrei Borzenkov <arvidjaar@gmail.com> --- (In reply to Graeme Spence from comment #6)
Is it a Mediatek MT7921?
No. It is trivially reproducible in QEMU with virtio network. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c8 M Westermann <wmicha@gmail.com> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |wmicha@gmail.com --- Comment #8 from M Westermann <wmicha@gmail.com> --- Hello. Same here. Kernel 6.1.2 bond with enp4s0f0 and enp4s0f1 fails Kernel 6.0.12 works fine. Kernel 6.1.2: echidna:/var/log # journalctl -b -1 | egrep '(wicked|igb)' Jan 05 07:06:49 echidna kernel: igb: Intel(R) Gigabit Ethernet Network Driver Jan 05 07:06:49 echidna kernel: igb: Copyright (c) 2007-2014 Intel Corporation. Jan 05 07:06:49 echidna kernel: igb 0000:04:00.0: added PHC on eth0 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.0: Intel(R) Gigabit Ethernet Network Connection Jan 05 07:06:49 echidna kernel: igb 0000:04:00.0: eth0: (PCIe:5.0Gb/s:Width x4) a0:36:9f:6a:39:b0 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.0: eth0: PBA No: G15139-002 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.0: Using MSI-X interrupts. 8 rx queue(s), 8 tx queue(s) Jan 05 07:06:49 echidna kernel: igb 0000:04:00.1: added PHC on eth1 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.1: Intel(R) Gigabit Ethernet Network Connection Jan 05 07:06:49 echidna kernel: igb 0000:04:00.1: eth1: (PCIe:5.0Gb/s:Width x4) a0:36:9f:6a:39:b1 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.1: eth1: PBA No: G15139-002 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.1: Using MSI-X interrupts. 8 rx queue(s), 8 tx queue(s) Jan 05 07:06:49 echidna kernel: igb 0000:04:00.2: added PHC on eth2 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.2: Intel(R) Gigabit Ethernet Network Connection Jan 05 07:06:49 echidna kernel: igb 0000:04:00.2: eth2: (PCIe:5.0Gb/s:Width x4) a0:36:9f:6a:39:b2 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.2: eth2: PBA No: G15139-002 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.2: Using MSI-X interrupts. 8 rx queue(s), 8 tx queue(s) Jan 05 07:06:49 echidna kernel: igb 0000:04:00.3: added PHC on eth4 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.3: Intel(R) Gigabit Ethernet Network Connection Jan 05 07:06:49 echidna kernel: igb 0000:04:00.3: eth4: (PCIe:5.0Gb/s:Width x4) a0:36:9f:6a:39:b3 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.3: eth4: PBA No: G15139-002 Jan 05 07:06:49 echidna kernel: igb 0000:04:00.3: Using MSI-X interrupts. 8 rx queue(s), 8 tx queue(s) Jan 05 07:06:50 echidna kernel: igb 0000:04:00.0 enp4s0f0: renamed from eth0 Jan 05 07:06:50 echidna kernel: igb 0000:04:00.1 enp4s0f1: renamed from eth1 Jan 05 07:06:50 echidna kernel: igb 0000:04:00.2 enp4s0f2: renamed from eth2 Jan 05 07:06:50 echidna kernel: igb 0000:04:00.3 enp4s0f3: renamed from eth4 Jan 05 07:08:08 echidna systemd[1]: Starting wicked AutoIPv4 supplicant service... Jan 05 07:08:08 echidna systemd[1]: Starting wicked DHCPv4 supplicant service... Jan 05 07:08:08 echidna systemd[1]: Starting wicked DHCPv6 supplicant service... Jan 05 07:08:08 echidna systemd[1]: Started wicked AutoIPv4 supplicant service. Jan 05 07:08:08 echidna systemd[1]: Started wicked DHCPv4 supplicant service. Jan 05 07:08:08 echidna systemd[1]: Started wicked DHCPv6 supplicant service. Jan 05 07:08:08 echidna systemd[1]: Starting wicked network management service daemon... Jan 05 07:08:08 echidna systemd[1]: Started wicked network management service daemon. Jan 05 07:08:08 echidna systemd[1]: Starting wicked network nanny service... Jan 05 07:08:08 echidna systemd[1]: Started wicked network nanny service. Jan 05 07:08:08 echidna systemd[1]: Starting wicked managed network interfaces... Jan 05 07:08:38 echidna wicked[1306]: lo up Jan 05 07:08:38 echidna wicked[1306]: enp4s0f0 device-not-running Jan 05 07:08:38 echidna wicked[1306]: enp4s0f1 device-not-running Jan 05 07:08:38 echidna wicked[1306]: enp4s0f2 setup-in-progress Jan 05 07:08:38 echidna wicked[1306]: enp0s31f6 setup-in-progress Jan 05 07:08:38 echidna wicked[1306]: br0 device-not-running Jan 05 07:08:38 echidna wicked[1306]: bond0 device-not-running Jan 05 07:08:38 echidna systemd[1]: Finished wicked managed network interfaces. Kernel 6.0.12: echidna:/var/log # journalctl -b 0 | egrep '(wicked|igb)' Jan 05 07:15:12 echidna kernel: igb: Intel(R) Gigabit Ethernet Network Driver Jan 05 07:15:12 echidna kernel: igb: Copyright (c) 2007-2014 Intel Corporation. Jan 05 07:15:12 echidna kernel: igb 0000:04:00.0: added PHC on eth0 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.0: Intel(R) Gigabit Ethernet Network Connection Jan 05 07:15:12 echidna kernel: igb 0000:04:00.0: eth0: (PCIe:5.0Gb/s:Width x4) a0:36:9f:6a:39:b0 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.0: eth0: PBA No: G15139-002 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.0: Using MSI-X interrupts. 8 rx queue(s), 8 tx queue(s) Jan 05 07:15:12 echidna kernel: igb 0000:04:00.1: added PHC on eth1 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.1: Intel(R) Gigabit Ethernet Network Connection Jan 05 07:15:12 echidna kernel: igb 0000:04:00.1: eth1: (PCIe:5.0Gb/s:Width x4) a0:36:9f:6a:39:b1 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.1: eth1: PBA No: G15139-002 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.1: Using MSI-X interrupts. 8 rx queue(s), 8 tx queue(s) Jan 05 07:15:12 echidna kernel: igb 0000:04:00.2: added PHC on eth2 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.2: Intel(R) Gigabit Ethernet Network Connection Jan 05 07:15:12 echidna kernel: igb 0000:04:00.2: eth2: (PCIe:5.0Gb/s:Width x4) a0:36:9f:6a:39:b2 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.2: eth2: PBA No: G15139-002 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.2: Using MSI-X interrupts. 8 rx queue(s), 8 tx queue(s) Jan 05 07:15:12 echidna kernel: igb 0000:04:00.3: added PHC on eth4 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.3: Intel(R) Gigabit Ethernet Network Connection Jan 05 07:15:12 echidna kernel: igb 0000:04:00.3: eth4: (PCIe:5.0Gb/s:Width x4) a0:36:9f:6a:39:b3 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.3: eth4: PBA No: G15139-002 Jan 05 07:15:12 echidna kernel: igb 0000:04:00.3: Using MSI-X interrupts. 8 rx queue(s), 8 tx queue(s) Jan 05 07:15:13 echidna kernel: igb 0000:04:00.0 enp4s0f0: renamed from eth0 Jan 05 07:15:13 echidna kernel: igb 0000:04:00.1 enp4s0f1: renamed from eth1 Jan 05 07:15:13 echidna kernel: igb 0000:04:00.3 enp4s0f3: renamed from eth4 Jan 05 07:15:13 echidna kernel: igb 0000:04:00.2 enp4s0f2: renamed from eth2 Jan 05 07:16:31 echidna systemd[1]: Starting wicked AutoIPv4 supplicant service... Jan 05 07:16:31 echidna systemd[1]: Starting wicked DHCPv4 supplicant service... Jan 05 07:16:31 echidna systemd[1]: Starting wicked DHCPv6 supplicant service... Jan 05 07:16:31 echidna systemd[1]: Started wicked DHCPv4 supplicant service. Jan 05 07:16:31 echidna systemd[1]: Started wicked AutoIPv4 supplicant service. Jan 05 07:16:31 echidna systemd[1]: Started wicked DHCPv6 supplicant service. Jan 05 07:16:31 echidna systemd[1]: Starting wicked network management service daemon... Jan 05 07:16:31 echidna systemd[1]: Started wicked network management service daemon. Jan 05 07:16:31 echidna systemd[1]: Starting wicked network nanny service... Jan 05 07:16:31 echidna systemd[1]: Started wicked network nanny service. Jan 05 07:16:31 echidna systemd[1]: Starting wicked managed network interfaces... Jan 05 07:16:37 echidna kernel: igb 0000:04:00.1 enp4s0f1: igb: enp4s0f1 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX Jan 05 07:16:37 echidna kernel: igb 0000:04:00.0 enp4s0f0: igb: enp4s0f0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX Jan 05 07:16:37 echidna wickedd-dhcp4[1259]: br0: Request to acquire DHCPv4 lease with UUID 3f6bb663-0d42-0b00-f304-00000e000000 Jan 05 07:16:37 echidna wickedd-dhcp6[1261]: br0: Request to acquire DHCPv6 lease with UUID 3f6bb663-0d42-0b00-f304-00000f000000 in mode auto Jan 05 07:16:49 echidna wickedd-dhcp4[1259]: br0: Committed DHCPv4 lease with address 192.168.200.10 (lease time 863988, renew in 431988 sec, rebind in 755988 sec) Jan 05 07:16:54 echidna wickedd-dhcp6[1261]: br0: Committing DHCPv6 lease with: Jan 05 07:16:54 echidna wickedd-dhcp6[1261]: br0 +ia-na.address 2003:f1:f71f:f100:60d7:f1ff:fed9:ab55/0, pref-lft 3598, valid-lft 7198 Jan 05 07:16:54 echidna wickedd[1267]: br0: address 2003:f1:f71f:f100:60d7:f1ff:fed9:ab55 covered by a auto lease Jan 05 07:17:01 echidna wicked[1279]: lo up Jan 05 07:17:01 echidna wicked[1279]: enp4s0f0 enslaved Jan 05 07:17:01 echidna wicked[1279]: enp4s0f1 enslaved Jan 05 07:17:01 echidna wicked[1279]: enp4s0f2 setup-in-progress Jan 05 07:17:01 echidna wicked[1279]: enp0s31f6 setup-in-progress Jan 05 07:17:01 echidna wicked[1279]: br0 up Jan 05 07:17:01 echidna wicked[1279]: bond0 enslaved -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c9 --- Comment #9 from Marius Tomaschewski <mt@suse.com> --- Created attachment 864293 --> http://bugzilla.opensuse.org/attachment.cgi?id=864293&action=edit Test script: bond-ip.sh [-1 ethX -2 ethY] [-m mode] [one|two|] This script sets up a bonding to illustrate the behavior/rtnetlink enslave processing change in 6.1 kernel -- regression to 4.x (sle12) - 6.0.x kernels. It is using the `ip link` commands, stop/shutdown (bond0 in) wicked before. Use the optional -1 and -2 options to specify the slave interface names; it is setting up a bond0 in active-backup mode to work with any switch. * bond-ip.sh -1 ethX -2 ethY one Is the variant wicked is using and setting the link down and enslaving to bond in a single `ip link set down master bond dev ethX` call aka rtnetlink message that were working with all 4.x .. 6.0.x kernels and stopped to work with 6.1 kernel. [�] A slave has to be set down before enslave or the kernel will reject it. * bond-ip.sh -1 ethX -2 ethY two Possible workaround -- as two instead of one `ip link set` rtnetlink message any more, thus it _may_ cause some race conditions or side effect... The 1st ip link is setting the slave down, 2nd enslaves to the master. * bond-ip.sh [-1 ethX -2 ethY] This is a [�] setup failure case without to set the links down before enslave to show that setting/ensuring that the slave links are down is required. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c10 --- Comment #10 from Marius Tomaschewski <mt@suse.com> --- Created attachment 864294 --> http://bugzilla.opensuse.org/attachment.cgi?id=864294&action=edit Test script outputs with 4.x, 6.0 and 6.1 cases $ grep ^=== bond-ip.txt === Working case "one" with SLE-12-SP5 4.x kernel ====================================== === Working case "two" with SLE-12-SP5 4.x kernel ====================================== === Failure case "fail" with SLE-12-SP5 4.x kernel ====================================== === Regression case "one" with Tumbleweed 6.1 kernel =================================== === Working case "two" with Tumbleweed 6.1 kernel ====================================== === Failure case "fail" with Tumbleweed 6.1 kernel ====================================== === Working case "one" with Tumbleweed 6.0 kernel ====================================== === Working case "two" with Tumbleweed 6.0 kernel ====================================== === Failure case "fail" with Tumbleweed 6.0 kernel ===================================== -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c11 Marius Tomaschewski <mt@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|No network (wicked) after |No network (wicked) after |23-12-2022 updates |23-12-2022 updates - bond | |change in kernel Flags| |needinfo?(kernel-bugs@suse. | |de) --- Comment #11 from Marius Tomaschewski <mt@suse.com> --- (In reply to Andrei Borzenkov from comment #5)
(In reply to Alessandro Sturniolo from comment #4)
kernel-default-6.0.10-1.1.x86_64 kernel-default-6.1.0-1.1.x86_64
Yes, I can reproduce it with kernel 6.1.1 and the problem goes away when I boot kernel 6.0.3 without any change in configuration. Apparently something changed between kernel 6.0 and 6.1. Whether it is a bug that must be fixed in kernel or behavior change that must be fixed in wicked I do not know.
Exactly. The enslave `ip link set down master bond0 dev ethX` in a single call works with all kernels since 4.x .. testsuite worked on SLE-12-SP2, I just don't have it installed on my test machine any more, thus 12-SP5 in the test outputs above. So this is a question to out kernel-maintainers: is this a bug or feature? We could change to send two separate messages, but it would be better when the kernel would get a fix to work with one down+enslave message again as sending two isn't "atomic" any more. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c12 --- Comment #12 from Andrei Borzenkov <arvidjaar@gmail.com> --- (In reply to Marius Tomaschewski from comment #11)
So this is a question to out kernel-maintainers: is this a bug or feature?
You can't please 'em all ... commit a4abfa627c3865c37e036bccb681619a50d3d93c Author: Phil Sutter <phil@nwl.cc> Date: Wed Sep 14 17:06:23 2022 +0200 net: rtnetlink: Enslave device before bringing it up Unlike with bridges, one can't add an interface to a bond and set it up at the same time: | # ip link set dummy0 down | # ip link set dummy0 master bond0 up | Error: Device can not be enslaved while up. Of all drivers with ndo_add_slave callback, bond and team decline if IFF_UP flag is set, vrf cycles the interface (i.e., sets it down and immediately up again) and the others just don't care. Support the common notion of setting the interface up after enslaving it by sorting the operations accordingly.
We could change to send two separate messages, but it would be better when the kernel would get a fix to work with one down+enslave message again as sending two isn't "atomic" any more.
Well ... I have a feeling that as long as interface must be down to enslave it this commit indeed opens up race condition without really fixing anything. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c13 --- Comment #13 from Marius Tomaschewski <mt@suse.com> ---
ip link set dummy0 master bond0 up
This is IMO a faulty call for a bond. bond != bridge Note also: it is IMO not a good idea to set the slaves UP yourself, e.g.: ip link set master bond0 dev ethX ## enslave to bond0 (must be down) ip link set up dev ethX ## "hack" to set the slave UP after enslave The interface needs to be down while enslave (set master). The bonding driver is not only setting it up itself, but also performs several actions on the slaves (inherits MTU, MAC, vlan filter, ... offloads). When we set the slave UP after enslave, we'll most probable disturb or break these actions -- some nic drivers may even reset the HW (and remove the administrative UP flag .. at least for a while) when they're applying offloads... ==> they may need the slave in down state (especially the expensive, full featured offload NICs IMO tend to behave like this). -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c14 --- Comment #14 from Marius Tomaschewski <mt@suse.com> --- Andrei, I've prepared a workaround for this kernel behavior regression to not set the DOWN flag any more in wicked bond enslave: https://github.com/openSUSE/wicked/pull/952 Once our CI tests passed, we'll apply it to: https://download.opensuse.org/repositories/network:/wicked:/testing/ -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c15 --- Comment #15 from Alexey Bobrov <alex.bobrov@mail.ru> --- repo https://download.opensuse.org/repositories/network:/wicked:/testing/ kernel 6.0.12 - bonding working kernel 6.1.8 - not working -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c16 --- Comment #16 from Marius Tomaschewski <mt@suse.com> --- It never arrived to the testing branch project -- seems, it was outdated there (still 0.6.71) ... whatever the reason were [CI hackweek?]. Sorry! I've manually prepared a test package with pull/952 applied as patch: https://download.opensuse.org/repositories/home:/mtomaschewski:/branches:/ne... * Tue Feb 07 2023 mt@suse.com - Applied pull request https://github.com/openSUSE/wicked/pull/952 to workaround kernel 6.1 enslave processing regression breaking `ip link set down master bond0 dev eth0` like rtnetlink enslave, that worked with all kernels from 4.12 up to 6.0 (boo#1206674). [+ 0001-bond-workaround-6.1-enslave-regression-boo-1206674.patch] -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.opensuse.org/show_bug.cgi?id=1206674 http://bugzilla.opensuse.org/show_bug.cgi?id=1206674#c17 --- Comment #17 from Alexey Bobrov <alex.bobrov@mail.ru> --- repo https://download.opensuse.org/repositories/home:/mtomaschewski:/branches:/ne... wicked-0.6.72-20.1 kernel 6.0.12 - bonding working kernel 6.1.8 - bonding working -- You are receiving this mail because: You are on the CC list for the bug.
participants (1)
-
bugzilla_noreply＠suse.com