Merge branch 'for-4.7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup

Pull cgroup fixes from Tejun Heo:
 "Three fix patches.  Two are for cgroup / css init failure path.  The
  last one makes css_set_lock irq-safe as the deadline scheduler ends up
  calling put_css_set() from irq context"

* 'for-4.7-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
  cgroup: Disable IRQs while holding css_set_lock
  cgroup: set css->id to -1 during init
  cgroup: remove redundant cleanup in css_create
This commit is contained in:
Linus Torvalds 2016-06-29 10:04:42 -07:00
commit 52827f389b

View File

@ -837,6 +837,8 @@ static void put_css_set_locked(struct css_set *cset)
static void put_css_set(struct css_set *cset)
{
unsigned long flags;
/*
* Ensure that the refcount doesn't hit zero while any readers
* can see it. Similar to atomic_dec_and_lock(), but for an
@ -845,9 +847,9 @@ static void put_css_set(struct css_set *cset)
if (atomic_add_unless(&cset->refcount, -1, 1))
return;
spin_lock_bh(&css_set_lock);
spin_lock_irqsave(&css_set_lock, flags);
put_css_set_locked(cset);
spin_unlock_bh(&css_set_lock);
spin_unlock_irqrestore(&css_set_lock, flags);
}
/*
@ -1070,11 +1072,11 @@ static struct css_set *find_css_set(struct css_set *old_cset,
/* First see if we already have a cgroup group that matches
* the desired set */
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
cset = find_existing_css_set(old_cset, cgrp, template);
if (cset)
get_css_set(cset);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
if (cset)
return cset;
@ -1102,7 +1104,7 @@ static struct css_set *find_css_set(struct css_set *old_cset,
* find_existing_css_set() */
memcpy(cset->subsys, template, sizeof(cset->subsys));
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
/* Add reference counts and links from the new css_set. */
list_for_each_entry(link, &old_cset->cgrp_links, cgrp_link) {
struct cgroup *c = link->cgrp;
@ -1128,7 +1130,7 @@ static struct css_set *find_css_set(struct css_set *old_cset,
css_get(css);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
return cset;
}
@ -1192,7 +1194,7 @@ static void cgroup_destroy_root(struct cgroup_root *root)
* Release all the links from cset_links to this hierarchy's
* root cgroup
*/
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_for_each_entry_safe(link, tmp_link, &cgrp->cset_links, cset_link) {
list_del(&link->cset_link);
@ -1200,7 +1202,7 @@ static void cgroup_destroy_root(struct cgroup_root *root)
kfree(link);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
if (!list_empty(&root->root_list)) {
list_del(&root->root_list);
@ -1600,11 +1602,11 @@ static int rebind_subsystems(struct cgroup_root *dst_root, u16 ss_mask)
ss->root = dst_root;
css->cgroup = dcgrp;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
hash_for_each(css_set_table, i, cset, hlist)
list_move_tail(&cset->e_cset_node[ss->id],
&dcgrp->e_csets[ss->id]);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
/* default hierarchy doesn't enable controllers by default */
dst_root->subsys_mask |= 1 << ssid;
@ -1640,10 +1642,10 @@ static int cgroup_show_path(struct seq_file *sf, struct kernfs_node *kf_node,
if (!buf)
return -ENOMEM;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
ns_cgroup = current_cgns_cgroup_from_root(kf_cgroot);
len = kernfs_path_from_node(kf_node, ns_cgroup->kn, buf, PATH_MAX);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
if (len >= PATH_MAX)
len = -ERANGE;
@ -1897,7 +1899,7 @@ static void cgroup_enable_task_cg_lists(void)
{
struct task_struct *p, *g;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
if (use_task_css_set_links)
goto out_unlock;
@ -1922,8 +1924,12 @@ static void cgroup_enable_task_cg_lists(void)
* entry won't be deleted though the process has exited.
* Do it while holding siglock so that we don't end up
* racing against cgroup_exit().
*
* Interrupts were already disabled while acquiring
* the css_set_lock, so we do not need to disable it
* again when acquiring the sighand->siglock here.
*/
spin_lock_irq(&p->sighand->siglock);
spin_lock(&p->sighand->siglock);
if (!(p->flags & PF_EXITING)) {
struct css_set *cset = task_css_set(p);
@ -1932,11 +1938,11 @@ static void cgroup_enable_task_cg_lists(void)
list_add_tail(&p->cg_list, &cset->tasks);
get_css_set(cset);
}
spin_unlock_irq(&p->sighand->siglock);
spin_unlock(&p->sighand->siglock);
} while_each_thread(g, p);
read_unlock(&tasklist_lock);
out_unlock:
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
}
static void init_cgroup_housekeeping(struct cgroup *cgrp)
@ -2043,13 +2049,13 @@ static int cgroup_setup_root(struct cgroup_root *root, u16 ss_mask)
* Link the root cgroup in this hierarchy into all the css_set
* objects.
*/
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
hash_for_each(css_set_table, i, cset, hlist) {
link_css_set(&tmp_links, cset, root_cgrp);
if (css_set_populated(cset))
cgroup_update_populated(root_cgrp, true);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
BUG_ON(!list_empty(&root_cgrp->self.children));
BUG_ON(atomic_read(&root->nr_cgrps) != 1);
@ -2256,11 +2262,11 @@ out_mount:
struct cgroup *cgrp;
mutex_lock(&cgroup_mutex);
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
cgrp = cset_cgroup_from_root(ns->root_cset, root);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
mutex_unlock(&cgroup_mutex);
nsdentry = kernfs_node_dentry(cgrp->kn, dentry->d_sb);
@ -2337,11 +2343,11 @@ char *cgroup_path_ns(struct cgroup *cgrp, char *buf, size_t buflen,
char *ret;
mutex_lock(&cgroup_mutex);
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
ret = cgroup_path_ns_locked(cgrp, buf, buflen, ns);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
mutex_unlock(&cgroup_mutex);
return ret;
@ -2369,7 +2375,7 @@ char *task_cgroup_path(struct task_struct *task, char *buf, size_t buflen)
char *path = NULL;
mutex_lock(&cgroup_mutex);
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
root = idr_get_next(&cgroup_hierarchy_idr, &hierarchy_id);
@ -2382,7 +2388,7 @@ char *task_cgroup_path(struct task_struct *task, char *buf, size_t buflen)
path = buf;
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
mutex_unlock(&cgroup_mutex);
return path;
}
@ -2557,7 +2563,7 @@ static int cgroup_taskset_migrate(struct cgroup_taskset *tset,
* the new cgroup. There are no failure cases after here, so this
* is the commit point.
*/
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_for_each_entry(cset, &tset->src_csets, mg_node) {
list_for_each_entry_safe(task, tmp_task, &cset->mg_tasks, cg_list) {
struct css_set *from_cset = task_css_set(task);
@ -2568,7 +2574,7 @@ static int cgroup_taskset_migrate(struct cgroup_taskset *tset,
put_css_set_locked(from_cset);
}
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
/*
* Migration is committed, all target tasks are now on dst_csets.
@ -2597,13 +2603,13 @@ out_cancel_attach:
}
} while_each_subsys_mask();
out_release_tset:
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_splice_init(&tset->dst_csets, &tset->src_csets);
list_for_each_entry_safe(cset, tmp_cset, &tset->src_csets, mg_node) {
list_splice_tail_init(&cset->mg_tasks, &cset->tasks);
list_del_init(&cset->mg_node);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
return ret;
}
@ -2634,7 +2640,7 @@ static void cgroup_migrate_finish(struct list_head *preloaded_csets)
lockdep_assert_held(&cgroup_mutex);
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_for_each_entry_safe(cset, tmp_cset, preloaded_csets, mg_preload_node) {
cset->mg_src_cgrp = NULL;
cset->mg_dst_cgrp = NULL;
@ -2642,7 +2648,7 @@ static void cgroup_migrate_finish(struct list_head *preloaded_csets)
list_del_init(&cset->mg_preload_node);
put_css_set_locked(cset);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
}
/**
@ -2783,7 +2789,7 @@ static int cgroup_migrate(struct task_struct *leader, bool threadgroup,
* already PF_EXITING could be freed from underneath us unless we
* take an rcu_read_lock.
*/
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
rcu_read_lock();
task = leader;
do {
@ -2792,7 +2798,7 @@ static int cgroup_migrate(struct task_struct *leader, bool threadgroup,
break;
} while_each_thread(leader, task);
rcu_read_unlock();
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
return cgroup_taskset_migrate(&tset, root);
}
@ -2816,7 +2822,7 @@ static int cgroup_attach_task(struct cgroup *dst_cgrp,
return -EBUSY;
/* look up all src csets */
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
rcu_read_lock();
task = leader;
do {
@ -2826,7 +2832,7 @@ static int cgroup_attach_task(struct cgroup *dst_cgrp,
break;
} while_each_thread(leader, task);
rcu_read_unlock();
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
/* prepare dst csets and commit */
ret = cgroup_migrate_prepare_dst(&preloaded_csets);
@ -2859,9 +2865,9 @@ static int cgroup_procs_write_permission(struct task_struct *task,
struct cgroup *cgrp;
struct inode *inode;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
cgrp = task_cgroup_from_root(task, &cgrp_dfl_root);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
while (!cgroup_is_descendant(dst_cgrp, cgrp))
cgrp = cgroup_parent(cgrp);
@ -2962,9 +2968,9 @@ int cgroup_attach_task_all(struct task_struct *from, struct task_struct *tsk)
if (root == &cgrp_dfl_root)
continue;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
from_cgrp = task_cgroup_from_root(from, root);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
retval = cgroup_attach_task(from_cgrp, tsk, false);
if (retval)
@ -3080,7 +3086,7 @@ static int cgroup_update_dfl_csses(struct cgroup *cgrp)
percpu_down_write(&cgroup_threadgroup_rwsem);
/* look up all csses currently attached to @cgrp's subtree */
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
cgroup_for_each_live_descendant_pre(dsct, d_css, cgrp) {
struct cgrp_cset_link *link;
@ -3088,14 +3094,14 @@ static int cgroup_update_dfl_csses(struct cgroup *cgrp)
cgroup_migrate_add_src(link->cset, dsct,
&preloaded_csets);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
/* NULL dst indicates self on default hierarchy */
ret = cgroup_migrate_prepare_dst(&preloaded_csets);
if (ret)
goto out_finish;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_for_each_entry(src_cset, &preloaded_csets, mg_preload_node) {
struct task_struct *task, *ntask;
@ -3107,7 +3113,7 @@ static int cgroup_update_dfl_csses(struct cgroup *cgrp)
list_for_each_entry_safe(task, ntask, &src_cset->tasks, cg_list)
cgroup_taskset_add(task, &tset);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
ret = cgroup_taskset_migrate(&tset, cgrp->root);
out_finish:
@ -3908,10 +3914,10 @@ static int cgroup_task_count(const struct cgroup *cgrp)
int count = 0;
struct cgrp_cset_link *link;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_for_each_entry(link, &cgrp->cset_links, cset_link)
count += atomic_read(&link->cset->refcount);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
return count;
}
@ -4249,7 +4255,7 @@ void css_task_iter_start(struct cgroup_subsys_state *css,
memset(it, 0, sizeof(*it));
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
it->ss = css->ss;
@ -4262,7 +4268,7 @@ void css_task_iter_start(struct cgroup_subsys_state *css,
css_task_iter_advance_css_set(it);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
}
/**
@ -4280,7 +4286,7 @@ struct task_struct *css_task_iter_next(struct css_task_iter *it)
it->cur_task = NULL;
}
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
if (it->task_pos) {
it->cur_task = list_entry(it->task_pos, struct task_struct,
@ -4289,7 +4295,7 @@ struct task_struct *css_task_iter_next(struct css_task_iter *it)
css_task_iter_advance(it);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
return it->cur_task;
}
@ -4303,10 +4309,10 @@ struct task_struct *css_task_iter_next(struct css_task_iter *it)
void css_task_iter_end(struct css_task_iter *it)
{
if (it->cur_cset) {
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_del(&it->iters_node);
put_css_set_locked(it->cur_cset);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
}
if (it->cur_task)
@ -4338,10 +4344,10 @@ int cgroup_transfer_tasks(struct cgroup *to, struct cgroup *from)
mutex_lock(&cgroup_mutex);
/* all tasks in @from are being moved, all csets are source */
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_for_each_entry(link, &from->cset_links, cset_link)
cgroup_migrate_add_src(link->cset, to, &preloaded_csets);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
ret = cgroup_migrate_prepare_dst(&preloaded_csets);
if (ret)
@ -5063,6 +5069,7 @@ static void init_and_link_css(struct cgroup_subsys_state *css,
memset(css, 0, sizeof(*css));
css->cgroup = cgrp;
css->ss = ss;
css->id = -1;
INIT_LIST_HEAD(&css->sibling);
INIT_LIST_HEAD(&css->children);
css->serial_nr = css_serial_nr_next++;
@ -5150,7 +5157,7 @@ static struct cgroup_subsys_state *css_create(struct cgroup *cgrp,
err = cgroup_idr_alloc(&ss->css_idr, NULL, 2, 0, GFP_KERNEL);
if (err < 0)
goto err_free_percpu_ref;
goto err_free_css;
css->id = err;
/* @css is ready to be brought online now, make it visible */
@ -5174,9 +5181,6 @@ static struct cgroup_subsys_state *css_create(struct cgroup *cgrp,
err_list_del:
list_del_rcu(&css->sibling);
cgroup_idr_remove(&ss->css_idr, css->id);
err_free_percpu_ref:
percpu_ref_exit(&css->refcnt);
err_free_css:
call_rcu(&css->rcu_head, css_free_rcu_fn);
return ERR_PTR(err);
@ -5451,10 +5455,10 @@ static int cgroup_destroy_locked(struct cgroup *cgrp)
*/
cgrp->self.flags &= ~CSS_ONLINE;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_for_each_entry(link, &cgrp->cset_links, cset_link)
link->cset->dead = true;
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
/* initiate massacre of all css's */
for_each_css(css, ssid, cgrp)
@ -5725,7 +5729,7 @@ int proc_cgroup_show(struct seq_file *m, struct pid_namespace *ns,
goto out;
mutex_lock(&cgroup_mutex);
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
for_each_root(root) {
struct cgroup_subsys *ss;
@ -5778,7 +5782,7 @@ int proc_cgroup_show(struct seq_file *m, struct pid_namespace *ns,
retval = 0;
out_unlock:
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
mutex_unlock(&cgroup_mutex);
kfree(buf);
out:
@ -5923,13 +5927,13 @@ void cgroup_post_fork(struct task_struct *child)
if (use_task_css_set_links) {
struct css_set *cset;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
cset = task_css_set(current);
if (list_empty(&child->cg_list)) {
get_css_set(cset);
css_set_move_task(child, NULL, cset, false);
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
}
/*
@ -5974,9 +5978,9 @@ void cgroup_exit(struct task_struct *tsk)
cset = task_css_set(tsk);
if (!list_empty(&tsk->cg_list)) {
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
css_set_move_task(tsk, cset, NULL, false);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
} else {
get_css_set(cset);
}
@ -6044,9 +6048,9 @@ static void cgroup_release_agent(struct work_struct *work)
if (!pathbuf || !agentbuf)
goto out;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
path = cgroup_path_ns_locked(cgrp, pathbuf, PATH_MAX, &init_cgroup_ns);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
if (!path)
goto out;
@ -6306,12 +6310,12 @@ struct cgroup_namespace *copy_cgroup_ns(unsigned long flags,
return ERR_PTR(-EPERM);
mutex_lock(&cgroup_mutex);
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
cset = task_css_set(current);
get_css_set(cset);
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
mutex_unlock(&cgroup_mutex);
new_ns = alloc_cgroup_ns();
@ -6435,7 +6439,7 @@ static int current_css_set_cg_links_read(struct seq_file *seq, void *v)
if (!name_buf)
return -ENOMEM;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
rcu_read_lock();
cset = rcu_dereference(current->cgroups);
list_for_each_entry(link, &cset->cgrp_links, cgrp_link) {
@ -6446,7 +6450,7 @@ static int current_css_set_cg_links_read(struct seq_file *seq, void *v)
c->root->hierarchy_id, name_buf);
}
rcu_read_unlock();
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
kfree(name_buf);
return 0;
}
@ -6457,7 +6461,7 @@ static int cgroup_css_links_read(struct seq_file *seq, void *v)
struct cgroup_subsys_state *css = seq_css(seq);
struct cgrp_cset_link *link;
spin_lock_bh(&css_set_lock);
spin_lock_irq(&css_set_lock);
list_for_each_entry(link, &css->cgroup->cset_links, cset_link) {
struct css_set *cset = link->cset;
struct task_struct *task;
@ -6480,7 +6484,7 @@ static int cgroup_css_links_read(struct seq_file *seq, void *v)
overflow:
seq_puts(seq, " ...\n");
}
spin_unlock_bh(&css_set_lock);
spin_unlock_irq(&css_set_lock);
return 0;
}