From aa32362f011c6e863132b16c1761487166a4bad2 Mon Sep 17 00:00:00 2001 From: Li Zefan Date: Thu, 4 Sep 2014 14:43:38 +0800 Subject: [PATCH] cgroup: check cgroup liveliness before unbreaking kernfs MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When cgroup_kn_lock_live() is called through some kernfs operation and another thread is calling cgroup_rmdir(), we'll trigger the warning in cgroup_get(). ------------[ cut here ]------------ WARNING: CPU: 1 PID: 1228 at kernel/cgroup.c:1034 cgroup_get+0x89/0xa0() ... Call Trace: [] dump_stack+0x41/0x52 [] warn_slowpath_common+0x7f/0xa0 [] warn_slowpath_null+0x1d/0x20 [] cgroup_get+0x89/0xa0 [] cgroup_kn_lock_live+0x28/0x70 [] __cgroup_procs_write.isra.26+0x51/0x230 [] cgroup_tasks_write+0x12/0x20 [] cgroup_file_write+0x40/0x130 [] kernfs_fop_write+0xd1/0x160 [] vfs_write+0x98/0x1e0 [] SyS_write+0x4d/0xa0 [] sysenter_do_call+0x12/0x12 ---[ end trace 6f2e0c38c2108a74 ]--- Fix this by calling css_tryget() instead of cgroup_get(). v2: - move cgroup_tryget() right below cgroup_get() definition. (Tejun) Cc: # 3.15+ Reported-by: Toralf Förster Signed-off-by: Zefan Li Signed-off-by: Tejun Heo --- kernel/cgroup.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/kernel/cgroup.c b/kernel/cgroup.c index bf30076664ca..940aced4ed00 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c @@ -1035,6 +1035,11 @@ static void cgroup_get(struct cgroup *cgrp) css_get(&cgrp->self); } +static bool cgroup_tryget(struct cgroup *cgrp) +{ + return css_tryget(&cgrp->self); +} + static void cgroup_put(struct cgroup *cgrp) { css_put(&cgrp->self); @@ -1147,7 +1152,8 @@ static struct cgroup *cgroup_kn_lock_live(struct kernfs_node *kn) * protection against removal. Ensure @cgrp stays accessible and * break the active_ref protection. */ - cgroup_get(cgrp); + if (!cgroup_tryget(cgrp)) + return NULL; kernfs_break_active_protection(kn); mutex_lock(&cgroup_mutex);