procfs: make errno values consistent when open pident vs exit(2) race…

… occurs

proc_pident_instantiate() has following call flow.

proc_pident_lookup()
  proc_pident_instantiate()
    proc_pid_make_inode()

And, proc_pident_lookup() has following error handling.

	const struct pid_entry *p, *last;
	error = ERR_PTR(-ENOENT);
	if (!task)
		goto out_no_task;

Then, proc_pident_instantiate should return ENOENT too when racing against
exit(2) occur.

EINAL has two bad reason.
  - it implies caller is wrong. bad the race isn't caller's mistake.
  - man 2 open don't explain EINVAL. user often don't handle it.

Note: Other proc_pid_make_inode() caller already use ENOENT properly.

Acked-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs/proc/base.c

@@ -1956,7 +1956,7 @@ static struct dentry *proc_pident_instantiate(struct inode *dir,
 	const struct pid_entry *p = ptr;
 	struct inode *inode;
 	struct proc_inode *ei;
-	struct dentry *error = ERR_PTR(-EINVAL);
+	struct dentry *error = ERR_PTR(-ENOENT);
 
 	inode = proc_pid_make_inode(dir->i_sb, task);
 	if (!inode)