linux-kernel - Re: [PATCH] exportfs: do not read dentry after free

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAOQ4uxj74Xh6ZbJp7zPZOxOY-gkrm5K1yU31+DW1YMYzOWtKgA@mail.gmail.com>
Date:   Fri, 23 Nov 2018 07:58:15 +0200
From:   Amir Goldstein <amir73il@...il.com>
To:     bianpan2016@....com
Cc:     Miklos Szeredi <mszeredi@...hat.com>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        Al Viro <viro@...iv.linux.org.uk>,
        "J. Bruce Fields" <bfields@...ldses.org>,
        Christoph Hellwig <hch@....de>
Subject: Re: [PATCH] exportfs: do not read dentry after free

On Fri, Nov 23, 2018 at 5:16 AM Pan Bian <bianpan2016@....com> wrote:
>
> The function dentry_connected calls dput(dentry) to drop the previously
> acquired reference to dentry. In this case, dentry can be released.
> After that, IS_ROOT(dentry) checks the condition
> (dentry == dentry->d_parent), which may result in a use-after-free bug.
> This patch directly compares dentry with its parent obtained before
> dropping the reference.
>
> Fixes: a056cc8934c("exportfs: stop retrying once we race with
> rename/remove")
>

CC Fixes patch author/reviewers

How did you find this? by code review or did this actually happen?

Normally a IS_ROOT dentry would be either DCACHE_DISCONNECTED or
pinned to some super block, but I guess there may be corner cases?

> Signed-off-by: Pan Bian <bianpan2016@....com>
> ---
>  fs/exportfs/expfs.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fs/exportfs/expfs.c b/fs/exportfs/expfs.c
> index 645158d..a69aaf5 100644
> --- a/fs/exportfs/expfs.c
> +++ b/fs/exportfs/expfs.c
> @@ -77,7 +77,7 @@ static bool dentry_connected(struct dentry *dentry)
>                 struct dentry *parent = dget_parent(dentry);
>
>                 dput(dentry);
> -               if (IS_ROOT(dentry)) {
> +               if (dentry == parent) { /* is root entry */
>                         dput(parent);
>                         return false;
>                 }

The change itself looks right, but the name IS_ROOT is confusing
enough as it is. The explicit comment is just plain wrong.
If it was really a root dentry, it wouldn't have been DCACHE_DISCONNECTED
(unless it is a filesystem bug).

Thanks,
Amir.