[PATCH stable 5.10.y] proc: fix a dentry lock race between release_task and lookup

[PATCH stable 5.10.y] proc: fix a dentry lock race between release_task and lookup

[Suraj Jitindar Singh]

From: Zhihao Cheng <chengzhihao1@huawei.com>

commit d919a1e79bac890421537cf02ae773007bf55e6b upstream.

Commit 7bc3e6e55acf06 ("proc: Use a list of inodes to flush from proc")
moved proc_flush_task() behind __exit_signal().  Then, process systemd can
take long period high cpu usage during releasing task in following
concurrent processes:

  systemd                                 ps
kernel_waitid                 stat(/proc/tgid)
  do_wait                       filename_lookup
    wait_consider_task            lookup_fast
      release_task
        __exit_signal
          __unhash_process
            detach_pid
              __change_pid // remove task->pid_links
                                     d_revalidate -> pid_revalidate  // 0
                                     d_invalidate(/proc/tgid)
                                       shrink_dcache_parent(/proc/tgid)
                                         d_walk(/proc/tgid)
                                           spin_lock_nested(/proc/tgid/fd)
                                           // iterating opened fd
        proc_flush_pid                                    |
           d_invalidate (/proc/tgid/fd)                   |
              shrink_dcache_parent(/proc/tgid/fd)         |
                shrink_dentry_list(subdirs)               ↓
                  shrink_lock_dentry(/proc/tgid/fd) --> race on dentry lock

Function d_invalidate() will remove dentry from hash firstly, but why does
proc_flush_pid() process dentry '/proc/tgid/fd' before dentry
'/proc/tgid'?  That's because proc_pid_make_inode() adds proc inode in
reverse order by invoking hlist_add_head_rcu().  But proc should not add
any inodes under '/proc/tgid' except '/proc/tgid/task/pid', fix it by
adding inode into 'pid->inodes' only if the inode is /proc/tgid or
/proc/tgid/task/pid.

Performance regression:
Create 200 tasks, each task open one file for 50,000 times. Kill all
tasks when opened files exceed 10,000,000 (cat /proc/sys/fs/file-nr).

Before fix:
$ time killall -wq aa
  real    4m40.946s   # During this period, we can see 'ps' and 'systemd'
			taking high cpu usage.

After fix:
$ time killall -wq aa
  real    1m20.732s   # During this period, we can see 'systemd' taking
			high cpu usage.

Link: https://lkml.kernel.org/r/20220713130029.4133533-1-chengzhihao1@huawei.com
Fixes: 7bc3e6e55acf06 ("proc: Use a list of inodes to flush from proc")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=216054
Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Suggested-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Baoquan He <bhe@redhat.com>
Cc: Kalesh Singh <kaleshsingh@google.com>
Cc: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[ bp: Context adjustments ]
Signed-off-by: Suraj Jitindar Singh <surajjs@amazon.com>
---
 fs/proc/base.c | 46 ++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 38 insertions(+), 8 deletions(-)

diff --git a/fs/proc/base.c b/fs/proc/base.c
index a484c30bd5cf..712948e97991 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -1881,7 +1881,7 @@ void proc_pid_evict_inode(struct proc_inode *ei)
 	put_pid(pid);
 }
 
-struct inode *proc_pid_make_inode(struct super_block * sb,
+struct inode *proc_pid_make_inode(struct super_block *sb,
 				  struct task_struct *task, umode_t mode)
 {
 	struct inode * inode;
@@ -1910,11 +1910,6 @@ struct inode *proc_pid_make_inode(struct super_block * sb,
 
 	/* Let the pid remember us for quick removal */
 	ei->pid = pid;
-	if (S_ISDIR(mode)) {
-		spin_lock(&pid->lock);
-		hlist_add_head_rcu(&ei->sibling_inodes, &pid->inodes);
-		spin_unlock(&pid->lock);
-	}
 
 	task_dump_owner(task, 0, &inode->i_uid, &inode->i_gid);
 	security_task_to_inode(task, inode);
@@ -1927,6 +1922,39 @@ struct inode *proc_pid_make_inode(struct super_block * sb,
 	return NULL;
 }
 
+/*
+ * Generating an inode and adding it into @pid->inodes, so that task will
+ * invalidate inode's dentry before being released.
+ *
+ * This helper is used for creating dir-type entries under '/proc' and
+ * '/proc/<tgid>/task'. Other entries(eg. fd, stat) under '/proc/<tgid>'
+ * can be released by invalidating '/proc/<tgid>' dentry.
+ * In theory, dentries under '/proc/<tgid>/task' can also be released by
+ * invalidating '/proc/<tgid>' dentry, we reserve it to handle single
+ * thread exiting situation: Any one of threads should invalidate its
+ * '/proc/<tgid>/task/<pid>' dentry before released.
+ */
+static struct inode *proc_pid_make_base_inode(struct super_block *sb,
+				struct task_struct *task, umode_t mode)
+{
+	struct inode *inode;
+	struct proc_inode *ei;
+	struct pid *pid;
+
+	inode = proc_pid_make_inode(sb, task, mode);
+	if (!inode)
+		return NULL;
+
+	/* Let proc_flush_pid find this directory inode */
+	ei = PROC_I(inode);
+	pid = ei->pid;
+	spin_lock(&pid->lock);
+	hlist_add_head_rcu(&ei->sibling_inodes, &pid->inodes);
+	spin_unlock(&pid->lock);
+
+	return inode;
+}
+
 int pid_getattr(const struct path *path, struct kstat *stat,
 		u32 request_mask, unsigned int query_flags)
 {
@@ -3341,7 +3369,8 @@ static struct dentry *proc_pid_instantiate(struct dentry * dentry,
 {
 	struct inode *inode;
 
-	inode = proc_pid_make_inode(dentry->d_sb, task, S_IFDIR | S_IRUGO | S_IXUGO);
+	inode = proc_pid_make_base_inode(dentry->d_sb, task,
+					 S_IFDIR | S_IRUGO | S_IXUGO);
 	if (!inode)
 		return ERR_PTR(-ENOENT);
 
@@ -3637,7 +3666,8 @@ static struct dentry *proc_task_instantiate(struct dentry *dentry,
 	struct task_struct *task, const void *ptr)
 {
 	struct inode *inode;
-	inode = proc_pid_make_inode(dentry->d_sb, task, S_IFDIR | S_IRUGO | S_IXUGO);
+	inode = proc_pid_make_base_inode(dentry->d_sb, task,
+					 S_IFDIR | S_IRUGO | S_IXUGO);
 	if (!inode)
 		return ERR_PTR(-ENOENT);
 
-- 
2.34.1

Re: [PATCH stable 5.10.y] proc: fix a dentry lock race between release_task and lookup

[Greg KH]

On Tue, Sep 19, 2023 at 04:33:35PM -0700, Suraj Jitindar Singh wrote:
> From: Zhihao Cheng <chengzhihao1@huawei.com>
> 
> commit d919a1e79bac890421537cf02ae773007bf55e6b upstream.
> 
> Commit 7bc3e6e55acf06 ("proc: Use a list of inodes to flush from proc")
> moved proc_flush_task() behind __exit_signal().  Then, process systemd can
> take long period high cpu usage during releasing task in following
> concurrent processes:
> 
>   systemd                                 ps
> kernel_waitid                 stat(/proc/tgid)
>   do_wait                       filename_lookup
>     wait_consider_task            lookup_fast
>       release_task
>         __exit_signal
>           __unhash_process
>             detach_pid
>               __change_pid // remove task->pid_links
>                                      d_revalidate -> pid_revalidate  // 0
>                                      d_invalidate(/proc/tgid)
>                                        shrink_dcache_parent(/proc/tgid)
>                                          d_walk(/proc/tgid)
>                                            spin_lock_nested(/proc/tgid/fd)
>                                            // iterating opened fd
>         proc_flush_pid                                    |
>            d_invalidate (/proc/tgid/fd)                   |
>               shrink_dcache_parent(/proc/tgid/fd)         |
>                 shrink_dentry_list(subdirs)               ↓
>                   shrink_lock_dentry(/proc/tgid/fd) --> race on dentry lock
> 
> Function d_invalidate() will remove dentry from hash firstly, but why does
> proc_flush_pid() process dentry '/proc/tgid/fd' before dentry
> '/proc/tgid'?  That's because proc_pid_make_inode() adds proc inode in
> reverse order by invoking hlist_add_head_rcu().  But proc should not add
> any inodes under '/proc/tgid' except '/proc/tgid/task/pid', fix it by
> adding inode into 'pid->inodes' only if the inode is /proc/tgid or
> /proc/tgid/task/pid.
> 
> Performance regression:
> Create 200 tasks, each task open one file for 50,000 times. Kill all
> tasks when opened files exceed 10,000,000 (cat /proc/sys/fs/file-nr).
> 
> Before fix:
> $ time killall -wq aa
>   real    4m40.946s   # During this period, we can see 'ps' and 'systemd'
> 			taking high cpu usage.
> 
> After fix:
> $ time killall -wq aa
>   real    1m20.732s   # During this period, we can see 'systemd' taking
> 			high cpu usage.
> 
> Link: https://lkml.kernel.org/r/20220713130029.4133533-1-chengzhihao1@huawei.com
> Fixes: 7bc3e6e55acf06 ("proc: Use a list of inodes to flush from proc")
> Link: https://bugzilla.kernel.org/show_bug.cgi?id=216054
> Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
> Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
> Suggested-by: Brian Foster <bfoster@redhat.com>
> Reviewed-by: Brian Foster <bfoster@redhat.com>
> Cc: Al Viro <viro@zeniv.linux.org.uk>
> Cc: Alexey Dobriyan <adobriyan@gmail.com>
> Cc: Eric Biederman <ebiederm@xmission.com>
> Cc: Matthew Wilcox <willy@infradead.org>
> Cc: Baoquan He <bhe@redhat.com>
> Cc: Kalesh Singh <kaleshsingh@google.com>
> Cc: Yu Kuai <yukuai3@huawei.com>
> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
> [ bp: Context adjustments ]
> Signed-off-by: Suraj Jitindar Singh <surajjs@amazon.com>

both now queued up, thanks.

greg k-h