According to the trace reported in syslog the problem happened in btrfs_get_or_create_delayed_node(), specifically the faulty function was prefetch_freepointer() defined in mm/slub.c.
Considering that btrfs_get_or_create_delayed_node() is allocating a "node" object using kmem_cache_zalloc() from a local kmem_cache pool called "delayed_node_cache", I would say that most likely that pool was corrupted due to a bug in btrfs (because only btrfs is using that pool). If I understand correctly this problem can be reproduced fairly easily (it's happening every few days running a specific test suite). It would be interesting to understand if it is a regression or not. Do you know if the problem started to happen recently from a specific kernel version? Have you tried to run the same test suite on an older kernel? Ideally if we can find a "good" kernel and a "bad" kernel we could try to bisect the problem. -- You received this bug notification because you are a member of Kernel Packages, which is subscribed to linux in Ubuntu. https://bugs.launchpad.net/bugs/1820275 Title: btrfs-kernel oops from btrfs subvolume delete command Status in Ubuntu on IBM z Systems: Triaged Status in linux package in Ubuntu: New Bug description: This report show a kernel oops in the btrfs lugin , happened every few days when running a test suite on a test system. It looks like their systemd/Journal failed to kick off a dump immediately following the panic. The exploiter is using the kernel 4.15.0-42 We're not sure it's the btrfs subvolume delete that's causing it yet Following infos are attached: - syslog - sosreport - kdump output here : https://ibm.ent.box.com/folder/70219521934 To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu-z-systems/+bug/1820275/+subscriptions -- Mailing list: https://launchpad.net/~kernel-packages Post to : kernel-packages@lists.launchpad.net Unsubscribe : https://launchpad.net/~kernel-packages More help : https://help.launchpad.net/ListHelp
