[Bug 1091762] New: Regression 4.14=>4.16: data corruption with FCoE/FIP packets over bridged VLAN
http://bugzilla.suse.com/show_bug.cgi?id=1091762 Bug ID: 1091762 Summary: Regression 4.14=>4.16: data corruption with FCoE/FIP packets over bridged VLAN Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: x86-64 OS: Other Status: NEW Severity: Major Priority: P5 - None Component: Kernel Assignee: kernel-maintainers@forge.provo.novell.com Reporter: martin.wilck@suse.com QA Contact: qa-bugs@suse.de Found By: --- Blocker: --- I have a test setup for FCoE / multipath storage using two virtual machines. They are connected via a dedicated bridge. All virtual interfaces and the bridge use pfifo_fast qdisc and MTU 9000. FCoE is running in VN2VN mode. One VM is configured as a LIO server exporting some storage via FCoE. The other one is configured as an FCoE client. On both VMs, two interfaces are connected to the bridge, and I create two 802.1q VLANs over both, so that both VMs have 4 participating interfaces. The server has matching ACLs for the client for each LUN it exports. When FCoE is enabled on the client VM, I expect the server LUNs to be detected and mounted automatically. This has worked alright for almost 2 years now. I recently tested this setup again and started pulling my hair because no FC remote ports were being detected on either VM. I spent several hours trying to debug the VM configuration, until I finally realized that the FIP packets the server received from the bridge in the discovery process were corrupt. Wireshark running on the host the bridge device itself would see correct packets, but packets dumped using tcpdump on the server VM would see the last 4 bytes of many packets nullified. This was with a 4.16.4 kernel on the host (the VMs are both running SLE12-SP2). I downgraded the host kernel to v4.14 and the problem is gone, thus I'm pretty certain that this is a kernel problem and not qemu or whatever. -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c1
--- Comment #1 from Martin Wilck
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c2
Michal Kubeček
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c3
--- Comment #3 from Martin Wilck
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c4
--- Comment #4 from Martin Wilck
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c5
--- Comment #5 from Martin Wilck
even though that were sent by server VM on the other interface. s/though/those/
-- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c6
--- Comment #6 from Martin Wilck
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c7
--- Comment #7 from Michal Kubeček
The last 4 bytes, which are part of the FIP VN_Port info "port name" field, are missing. (my description in comment 0 was wrong in this respect too, sorry).
Makes commit 81c895072d29 look even more promising. :-) -- You are receiving this mail because: You are on the CC list for the bug.
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c8
--- Comment #8 from Martin Wilck
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c9
--- Comment #9 from Martin Wilck
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c10
Michal Kubeček
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c11
Michal Kubeček
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c12
--- Comment #12 from Martin Wilck
http://bugzilla.suse.com/show_bug.cgi?id=1091762
http://bugzilla.suse.com/show_bug.cgi?id=1091762#c13
--- Comment #13 from Martin Wilck
participants (1)
-
bugzilla_noreply@novell.com