From 77f88796cee819b9c4562b0b6b44691b3b7755b1 Mon Sep 17 00:00:00 2001
From: Tejun Heo <tj@kernel.org>
Date: Thu, 16 Mar 2017 16:54:24 -0400
Subject: cgroup, kthread: close race window where new kthreads can be migrated
to non-root cgroups
Creation of a kthread goes through a couple interlocked stages between
the kthread itself and its creator. Once the new kthread starts
running, it initializes itself and wakes up the creator. The creator
then can further configure the kthread and then let it start doing its
job by waking it up.
In this configuration-by-creator stage, the creator is the only one
that can wake it up but the kthread is visible to userland. When
altering the kthread's attributes from userland is allowed, this is
fine; however, for cases where CPU affinity is critical,
kthread_bind() is used to first disable affinity changes from userland
and then set the affinity. This also prevents the kthread from being
migrated into non-root cgroups as that can affect the CPU affinity and
many other things.
Unfortunately, the cgroup side of protection is racy. While the
PF_NO_SETAFFINITY flag prevents further migrations, userland can win
the race before the creator sets the flag with kthread_bind() and put
the kthread in a non-root cgroup, which can lead to all sorts of
problems including incorrect CPU affinity and starvation.
This bug got triggered by userland which periodically tries to migrate
all processes in the root cpuset cgroup to a non-root one. Per-cpu
workqueue workers got caught while being created and ended up with
incorrected CPU affinity breaking concurrency management and sometimes
stalling workqueue execution.
This patch adds task->no_cgroup_migration which disallows the task to
be migrated by userland. kthreadd starts with the flag set making
every child kthread start in the root cgroup with migration
disallowed. The flag is cleared after the kthread finishes
initialization by which time PF_NO_SETAFFINITY is set if the kthread
should stay in the root cgroup.
It'd be better to wait for the initialization instead of failing but I
couldn't think of a way of implementing that without adding either a
new PF flag, or sleeping and retrying from waiting side. Even if
userland depends on changing cgroup membership of a kthread, it either
has to be synchronized with kthread_create() or periodically repeat,
so it's unlikely that this would break anything.
v2: Switch to a simpler implementation using a new task_struct bit
field suggested by Oleg.
Signed-off-by: Tejun Heo <tj@kernel.org>
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Reported-and-debugged-by: Chris Mason <clm@fb.com>
Cc: stable@vger.kernel.org # v4.3+ (we can't close the race on < v4.3)
Signed-off-by: Tejun Heo <tj@kernel.org>
---
kernel/cgroup/cgroup.c | 9 +++++----
1 file changed, 5 insertions(+), 4 deletions(-)
(limited to 'kernel/cgroup')
diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c
index 0125589c7428..638ef7568495 100644
--- a/kernel/cgroup/cgroup.c
+++ b/kernel/cgroup/cgroup.c
@@ -2425,11 +2425,12 @@ ssize_t __cgroup_procs_write(struct kernfs_open_file *of, char *buf,
tsk = tsk->group_leader;
/*
- * Workqueue threads may acquire PF_NO_SETAFFINITY and become
- * trapped in a cpuset, or RT worker may be born in a cgroup
- * with no rt_runtime allocated. Just say no.
+ * kthreads may acquire PF_NO_SETAFFINITY during initialization.
+ * If userland migrates such a kthread to a non-root cgroup, it can
+ * become trapped in a cpuset, or RT kthread may be born in a
+ * cgroup with no rt_runtime allocated. Just say no.
*/
- if (tsk == kthreadd_task || (tsk->flags & PF_NO_SETAFFINITY)) {
+ if (tsk->no_cgroup_migration || (tsk->flags & PF_NO_SETAFFINITY)) {
ret = -EINVAL;
goto out_unlock_rcu;
}
--
cgit v1.2.3
From bfb0b80db5f9dca5ac0a5fd0edb765ee555e5a8e Mon Sep 17 00:00:00 2001
From: Zefan Li <lizefan@huawei.com>
Date: Fri, 7 Apr 2017 16:51:55 +0800
Subject: cgroup: avoid attaching a cgroup root to two different superblocks
Run this:
touch file0
for ((; ;))
{
mount -t cpuset xxx file0
}
And this concurrently:
touch file1
for ((; ;))
{
mount -t cpuset xxx file1
}
We'll trigger a warning like this:
------------[ cut here ]------------
WARNING: CPU: 1 PID: 4675 at lib/percpu-refcount.c:317 percpu_ref_kill_and_confirm+0x92/0xb0
percpu_ref_kill_and_confirm called more than once on css_release!
CPU: 1 PID: 4675 Comm: mount Not tainted 4.11.0-rc5+ #5
Hardware name: Bochs Bochs, BIOS Bochs 01/01/2007
Call Trace:
dump_stack+0x63/0x84
__warn+0xd1/0xf0
warn_slowpath_fmt+0x5f/0x80
percpu_ref_kill_and_confirm+0x92/0xb0
cgroup_kill_sb+0x95/0xb0
deactivate_locked_super+0x43/0x70
deactivate_super+0x46/0x60
...
---[ end trace a79f61c2a2633700 ]---
Here's a race:
Thread A Thread B
cgroup1_mount()
# alloc a new cgroup root
cgroup_setup_root()
cgroup1_mount()
# no sb yet, returns NULL
kernfs_pin_sb()
# but succeeds in getting the refcnt,
# so re-use cgroup root
percpu_ref_tryget_live()
# alloc sb with cgroup root
cgroup_do_mount()
cgroup_kill_sb()
# alloc another sb with same root
cgroup_do_mount()
cgroup_kill_sb()
We end up using the same cgroup root for two different superblocks,
so percpu_ref_kill() will be called twice on the same root when the
two superblocks are destroyed.
We should fix to make sure the superblock pinning is really successful.
Cc: stable@vger.kernel.org # 3.16+
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Zefan Li <lizefan@huawei.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
---
kernel/cgroup/cgroup-v1.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
(limited to 'kernel/cgroup')
diff --git a/kernel/cgroup/cgroup-v1.c b/kernel/cgroup/cgroup-v1.c
index 1dc22f6b49f5..12e19f0636ea 100644
--- a/kernel/cgroup/cgroup-v1.c
+++ b/kernel/cgroup/cgroup-v1.c
@@ -1146,7 +1146,7 @@ struct dentry *cgroup1_mount(struct file_system_type *fs_type, int flags,
* path is super cold. Let's just sleep a bit and retry.
*/
pinned_sb = kernfs_pin_sb(root->kf_root, NULL);
- if (IS_ERR(pinned_sb) ||
+ if (IS_ERR_OR_NULL(pinned_sb) ||
!percpu_ref_tryget_live(&root->cgrp.self.refcnt)) {
mutex_unlock(&cgroup_mutex);
if (!IS_ERR_OR_NULL(pinned_sb))
--
cgit v1.2.3