Thread (4 messages) 4 messages, 4 authors, 2019-02-01

Re: general protection fault in debugfs_create_files

From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Date: 2019-01-31 07:01:02
Also in: lkml

On Thu, Jan 31, 2019 at 02:34:56PM +0900, Tetsuo Handa wrote:
quoted hunk ↗ jump to hunk
Hello, again.

syzbot is hitting a similar crash due to debugfs_create_dir() returning -EEXIST.
Should debugfs_create_dir() return NULL as well? Or should the caller use IS_ERR_OR_NULL() ?
--- a/block/blk-mq-debugfs.c
+++ b/block/blk-mq-debugfs.c
@@ -861,6 +861,8 @@ int blk_mq_debugfs_register(struct request_queue *q)
                                            blk_debugfs_root);
        if (!q->debugfs_dir)
                return -ENOMEM;
+       if (IS_ERR(q->debugfs_dir))
+               printk("debugfs_create_dir=%ld\n", PTR_ERR(q->debugfs_dir));

        if (!debugfs_create_files(q->debugfs_dir, q,
                                  blk_mq_debugfs_queue_attrs))
I already posted this patch last Wednesday:
	https://lore.kernel.org/lkml/20190123134854.GA25906@kroah.com/ (local)
to solve this problem.

I guess I should queue it up in my tree as well, to handle this issue.
I'll go do that now.

thanks,

greg k-h
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help