Verification done for linux on xenial-proposed.
The inodes for file, xattr dir, and xattr child are all evicted at file
removal time, not making it to any disposal list after file removal.
So the window/scenario for the problem to occur is not present anymore.
Log
---
$ uname -rv
4.4.0-160-generic #188-Ubuntu SMP Wed Aug 14 04:21:43 UTC 2019
$ modinfo zfs | head
filename: /lib/modules/4.4.0-160-generic/kernel/zfs/zfs/zfs.ko
version: 0.6.5.6-0ubuntu28
...
srcversion: 99F1D0FED2F291CA7AED0C6
$ sudo apt-get install zfsutils-linux attr
$ sudo ./zfs-mount.sh
$ echo 2 | sudo tee /proc/sys/vm/drop_caches
2
$ sudo ./zfs-kprobes.sh
$ sudo cat /sys/kernel/debug/tracing/trace_pipe &
$ touch /zfs/file
touch-10656 [001] d... 359.615887: p_zfs_mknode_0:
(zfs_mknode+0x0/0xe00 [zfs]) flag=0x0 dzp=0xffff8800b9875940
touch-10656 [001] d... 359.616184: p_zfs_znode_alloc_0:
(zfs_znode_alloc+0x0/0x520 [zfs]) obj=0xa
touch-10656 [001] d... 359.616339: r_zfs_znode_alloc_0:
(zfs_mknode+0x8a3/0xe00 [zfs] <- zfs_znode_alloc) zpp=0xffff880036f48440
$ setfattr -n user.debug -v 1 /zfs/file
setfattr-10657 [000] d... 361.507063: p_zfs_mknode_0:
(zfs_mknode+0x0/0xe00 [zfs]) flag=0x2 dzp=0xffff880036f48440
setfattr-10657 [000] d... 361.507265: p_zfs_znode_alloc_0:
(zfs_znode_alloc+0x0/0x520 [zfs]) obj=0xb
setfattr-10657 [000] d... 361.507402: r_zfs_znode_alloc_0:
(zfs_mknode+0x8a3/0xe00 [zfs] <- zfs_znode_alloc) zpp=0xffff880139d09980
setfattr-10657 [000] d... 361.507665: p_zfs_zget_0:
(zfs_zget+0x0/0x230 [zfs]) zsb=0xffff8800bb12c000 obj=0xa
setfattr-10657 [000] d... 361.507792: r_zfs_zget_0:
(zfs_zaccess+0x12b/0x220 [zfs] <- zfs_zget)
setfattr-10657 [000] d... 361.507981: p_zfs_zget_0:
(zfs_zget+0x0/0x230 [zfs]) zsb=0xffff8800bb12c000 obj=0xa
setfattr-10657 [000] d... 361.508104: r_zfs_zget_0:
(zfs_zaccess+0x12b/0x220 [zfs] <- zfs_zget)
setfattr-10657 [000] d... 361.508692: p_zfs_zget_0:
(zfs_zget+0x0/0x230 [zfs]) zsb=0xffff8800bb12c000 obj=0xa
setfattr-10657 [000] d... 361.508821: r_zfs_zget_0:
(zfs_zaccess+0x12b/0x220 [zfs] <- zfs_zget)
setfattr-10657 [000] d... 361.509022: p_zfs_mknode_0:
(zfs_mknode+0x0/0xe00 [zfs]) flag=0x0 dzp=0xffff880139d09980
setfattr-10657 [000] d... 361.509170: p_zfs_znode_alloc_0:
(zfs_znode_alloc+0x0/0x520 [zfs]) obj=0xc
setfattr-10657 [000] d... 361.509302: r_zfs_znode_alloc_0:
(zfs_mknode+0x8a3/0xe00 [zfs] <- zfs_znode_alloc) zpp=0xffff880139d09100
$ rm /zfs/file
rm-10658 [001] d... 363.216716: p_zfs_zget_0:
(zfs_zget+0x0/0x230 [zfs]) zsb=0xffff8800bb12c000 obj=0xa
rm-10658 [001] d... 363.216882: r_zfs_zget_0:
(zfs_dirent_lock+0x56c/0x6c0 [zfs] <- zfs_zget)
rm-10658 [001] d... 363.217130: p_zfs_zget_0:
(zfs_zget+0x0/0x230 [zfs]) zsb=0xffff8800bb12c000 obj=0xb
rm-10658 [001] d... 363.217271: r_zfs_zget_0:
(zfs_remove+0x22b/0x4c0 [zfs] <- zfs_zget)
rm-10658 [001] d... 363.217567: p_zpl_evict_inode_0:
(zpl_evict_inode+0x0/0x60 [zfs]) inode=0xffff880036f48650
rm-10658 [001] d... 363.217715: p_zfs_inactive_0:
(zfs_inactive+0x0/0x270 [zfs]) inode=0xffff880036f48650
rm-10658 [001] d... 363.217835: p_zfs_zinactive_0:
(zfs_zinactive+0x0/0xe0 [zfs]) znode=0xffff880036f48440 obj=0xa
rm-10658 [001] d... 363.217963: p_zfs_rmnode_0:
(zfs_rmnode+0x0/0x360 [zfs]) znode=0xffff880036f48440
rm-10658 [001] d... 363.218102: p_zfs_zget_0:
(zfs_zget+0x0/0x230 [zfs]) zsb=0xffff8800bb12c000 obj=0xb
rm-10658 [001] d... 363.218232: r_zfs_zget_0:
(zfs_rmnode+0x25b/0x360 [zfs] <- zfs_zget)
rm-10658 [001] d... 363.218464: p_zfs_iput_async_0:
(zfs_iput_async+0x0/0x60 [zfs]) inode=0xffff880139d09b90 obj=0x0
<...>-10308 [003] d... 363.218496: p_zpl_evict_inode_0:
(zpl_evict_inode+0x0/0x60 [zfs]) inode=0xffff880139d09b90
z_iput-10308 [003] d... 363.218503: p_zfs_inactive_0:
(zfs_inactive+0x0/0x270 [zfs]) inode=0xffff880139d09b90
z_iput-10308 [003] d... 363.218505: p_zfs_zinactive_0:
(zfs_zinactive+0x0/0xe0 [zfs]) znode=0xffff880139d09980 obj=0xb
z_iput-10308 [003] d... 363.218509: p_zfs_rmnode_0:
(zfs_rmnode+0x0/0x360 [zfs]) znode=0xffff880139d09980
z_iput-10308 [003] d... 363.218512: p_zfs_purgedir_0:
(zfs_purgedir+0x0/0x230 [zfs]) znode=0xffff880139d09980
z_iput-10308 [003] d... 363.218560: p_zfs_zget_0:
(zfs_zget+0x0/0x230 [zfs]) zsb=0xffff8800bb12c000 obj=0xc
z_iput-10308 [003] d... 363.218566: r_zfs_zget_0:
(zfs_purgedir+0xb4/0x230 [zfs] <- zfs_zget)
z_iput-10308 [003] d... 363.218606: p_zfs_iput_async_0:
(zfs_iput_async+0x0/0x60 [zfs]) inode=0xffff880139d09310 obj=0x0
z_iput-10308 [003] d... 363.218626: r_zfs_purgedir_0:
(zfs_rmnode+0x272/0x360 [zfs] <- zfs_purgedir)
z_iput-10308 [003] d... 363.218663: p_zpl_evict_inode_0:
(zpl_evict_inode+0x0/0x60 [zfs]) inode=0xffff880139d09310
z_iput-10308 [003] d... 363.218665: p_zfs_inactive_0:
(zfs_inactive+0x0/0x270 [zfs]) inode=0xffff880139d09310
z_iput-10308 [003] d... 363.218665: p_zfs_zinactive_0:
(zfs_zinactive+0x0/0xe0 [zfs]) znode=0xffff880139d09100 obj=0xc
z_iput-10308 [003] d... 363.218667: p_zfs_rmnode_0:
(zfs_rmnode+0x0/0x360 [zfs]) znode=0xffff880139d09100
$ echo 2 | sudo tee /proc/sys/vm/drop_caches
2
<...>-10661 [003] d... 372.513411: p_dispose_list_0:
(dispose_list+0x0/0x50)
tee-10661 [003] d... 372.513525: p_dispose_list_0:
(dispose_list+0x0/0x50)
tee-10661 [003] d... 372.513836: p_dispose_list_0:
(dispose_list+0x0/0x50)
tee-10661 [003] d... 372.514065: p_dispose_list_0:
(dispose_list+0x0/0x50)
tee-10661 [003] d... 372.516630: p_dispose_list_0:
(dispose_list+0x0/0x50)
tee-10661 [003] d... 372.516691: p_dispose_list_0:
(dispose_list+0x0/0x50)
tee-10661 [003] d... 372.516830: p_dispose_list_0:
(dispose_list+0x0/0x50)
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to zfs-linux in Ubuntu.
https://bugs.launchpad.net/bugs/1839521
Title:
Xenial: ZFS deadlock in shrinker path with xattrs
Status in zfs-linux package in Ubuntu:
Fix Released
Status in zfs-linux source package in Xenial:
Fix Committed
Bug description:
[Impact]
* Xenial's ZFS can deadlock in the memory shrinker path
after removing files with extended attributes (xattr).
* Extended attributes are enabled by default, but are
_not_ used by default, which reduces the likelyhood.
* It's very difficult/rare to reproduce this problem,
due to file/xattr/remove/shrinker/lru order/timing
circumstances required. (weeks for a reporter user)
but a synthetic test-case has been found for tests.
[Test Case]
* A synthetic reproducer is available for this LP,
with a few steps to touch/setfattr/rm/drop_caches
plus a kernel module to massage the disposal list.
(comment #8)
* In the original ZFS module:
the xattr dir inode is not purged immediately on
file removal, but possibly purged _two_ shrinker
invocations later. This allows for other thread
started before file remove to call zfs_zget() on
the xattr child inode and iput() it, so it makes
to the same disposal list as the xattr dir inode.
(comment #3)
* In the modified ZFS module:
the xattr dir inode is purged immediately on file
removal not possibly later on shrinker invocation,
so the problem window above doesn't exist anymore.
(comment #12)
[Regression Potential]
* Low. The patches are confined to extended attributes
in ZFS, specifically node removal/purge, and another
change how an xattr child inode tracks its xattr dir
(parent) inode, so that it can be purged immediately
on removal.
* The ZFS test-suite has been run on original/modified
zfs-dkms package/kernel modules, with no regressions.
(comment #11)
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/zfs-linux/+bug/1839521/+subscriptions
--
Mailing list: https://launchpad.net/~kernel-packages
Post to : [email protected]
Unsubscribe : https://launchpad.net/~kernel-packages
More help : https://help.launchpad.net/ListHelp
