linux-hardening - Re: [PATCH 2/2] ocfs2: Replace strlcpy with strscpy

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]

Date:   Wed, 30 Aug 2023 16:06:00 -0700
From:   Kees Cook <keescook@...omium.org>
To:     Azeem Shaikh <azeemshaikh38@...il.com>
Cc:     Mark Fasheh <mark@...heh.com>, Joel Becker <jlbec@...lplan.org>,
        Joseph Qi <joseph.qi@...ux.alibaba.com>,
        linux-hardening@...r.kernel.org, ocfs2-devel@...ts.linux.dev,
        linux-kernel@...r.kernel.org,
        Christian Brauner <brauner@...nel.org>,
        Dave Chinner <dchinner@...hat.com>,
        Jeff Layton <jlayton@...nel.org>
Subject: Re: [PATCH 2/2] ocfs2: Replace strlcpy with strscpy

On Wed, Aug 30, 2023 at 09:54:26PM +0000, Azeem Shaikh wrote:
> strlcpy() reads the entire source buffer first.
> This read may exceed the destination size limit.
> This is both inefficient and can lead to linear read
> overflows if a source string is not NUL-terminated [1].
> In an effort to remove strlcpy() completely [2], replace
> strlcpy() here with strscpy().
> 
> Direct replacement is assumed to be safe here since
> it's ok for `kernel_param_ops.get()` to return -errno [3].
> This changes the behavior such that instead of silently ignoring the
> case when sizeof(@buffer) < DLMFS_CAPABILITIES, we now return error.
> 
> [1] https://www.kernel.org/doc/html/latest/process/deprecated.html#strlcpy
> [2] https://github.com/KSPP/linux/issues/89
> [3] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L52
> 
> Signed-off-by: Azeem Shaikh <azeemshaikh38@...il.com>
> ---
>  fs/ocfs2/dlmfs/dlmfs.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/ocfs2/dlmfs/dlmfs.c b/fs/ocfs2/dlmfs/dlmfs.c
> index 33e529de93b2..b001eccdd2f3 100644
> --- a/fs/ocfs2/dlmfs/dlmfs.c
> +++ b/fs/ocfs2/dlmfs/dlmfs.c
> @@ -80,7 +80,7 @@ static int param_set_dlmfs_capabilities(const char *val,
>  static int param_get_dlmfs_capabilities(char *buffer,
>  					const struct kernel_param *kp)
>  {
> -	return strlcpy(buffer, DLMFS_CAPABILITIES,
> +	return strscpy(buffer, DLMFS_CAPABILITIES,
>  		       strlen(DLMFS_CAPABILITIES) + 1);
>  }

This is another case of "accidentally correct".


param->get() is hooked here, in the sysfs "show" callback:

static ssize_t param_attr_show(struct module_attribute *mattr,
                               struct module_kobject *mk, char *buf)
{
        int count;
        struct param_attribute *attribute = to_param_attr(mattr);

        if (!attribute->param->ops->get)
                return -EPERM;

        kernel_param_lock(mk->mod);
        count = attribute->param->ops->get(buf, attribute->param);
        kernel_param_unlock(mk->mod);
        return count;
}

Meaning ultimately this will show up here, if I'm reading names right:
/sys/module/ocfs/parameters/dlmfs_capabilities

Anyway, the "count" being returned would be quite bad if
DLMFS_CAPABILITIES were dynamic and larger than PAGE_SIZE (the size of
the sysfs buffer).

For this case, I would say replace strlcpy with sysfs_emit:

	return sysfs_emit(buffer, DLMFS_CAPABILITIES);

(Also, ew, existing code doesn't include a trailing "\n". Oh well.)

-Kees

-- 
Kees Cook