Re: [PATCH] configfs: fix a race in configfs_lookup()

[Greg KH <gregkh@linuxfoundation.org>]

On Sat, Sep 02, 2023 at 01:20:36PM -0700, Kyle Zeng wrote:
> commit c42dd069be8dfc9b2239a5c89e73bbd08ab35de0 upstream.
> Backporting the patch to stable-v5.10.y to avoid race condition between configfs_dir_lseek and
> configfs_lookup since they both operate ->s_childre and configfs_lookup
> forgets to obtain the lock.
> The patch deviates from the original patch because of code change.
> The idea is to hold the configfs_dirent_lock when traversing
> ->s_children, which follows the core idea of the original patch.
> 
> 
> Signed-off-by: Kyle Zeng <zengyhkyle@gmail.com>
> ---
>  fs/configfs/dir.c | 2 ++
>  1 file changed, 2 insertions(+)

You lost all the original signed-off-by lines of the original, AND you
lost the authorship of the original commit.  And you didn't cc: anyone
involved in the original patch, to get their review, or objection to it
being backported.

Take a look at many of the backports that happen on the stable list for
examples of how to do this properly.

Here are 2 examples from this weekend alone that are good examples of
how to do this properly:
	https://lore.kernel.org/r/20230902151000.3817-1-konishi.ryusuke@gmail.com
	https://lore.kernel.org/r/cover.1693593288.git.luizcap@amazon.com

Also, how did you test this change?  is this something that you have
actually hit in real life?

thanks,

greg k-h