]> git.proxmox.com Git - mirror_ubuntu-bionic-kernel.git/commit
cgroup: Iterate tasks that did not finish do_exit()
authorMichal Koutný <mkoutny@suse.com>
Fri, 24 Jan 2020 11:40:15 +0000 (12:40 +0100)
committerKleber Sacilotto de Souza <kleber.souza@canonical.com>
Tue, 7 Apr 2020 08:51:52 +0000 (10:51 +0200)
commit3c5bef308798e77c36ba6e1430765381db7c1de8
tree4978f2d5ab0dc0d842864c58bf5ad5fb07ba9dcb
parent4a9f0f54dc62b0d1036fa542f4dcd4af4563f05e
cgroup: Iterate tasks that did not finish do_exit()

BugLink: https://bugs.launchpad.net/bugs/1869732
commit 9c974c77246460fa6a92c18554c3311c8c83c160 upstream.

PF_EXITING is set earlier than actual removal from css_set when a task
is exitting. This can confuse cgroup.procs readers who see no PF_EXITING
tasks, however, rmdir is checking against css_set membership so it can
transitionally fail with EBUSY.

Fix this by listing tasks that weren't unlinked from css_set active
lists.
It may happen that other users of the task iterator (without
CSS_TASK_ITER_PROCS) spot a PF_EXITING task before cgroup_exit(). This
is equal to the state before commit c03cd7738a83 ("cgroup: Include dying
leaders with live threads in PROCS iterations") but it may be reviewed
later.

Reported-by: Suren Baghdasaryan <surenb@google.com>
Fixes: c03cd7738a83 ("cgroup: Include dying leaders with live threads in PROCS iterations")
Signed-off-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Kamal Mostafa <kamal@canonical.com>
Signed-off-by: Khalid Elmously <khalid.elmously@canonical.com>
include/linux/cgroup.h
kernel/cgroup/cgroup.c