Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220822115257.7457-1-goriainov@ispras.ru>
In-Reply-To: <20220822115257.7457-1-goriainov@ispras.ru>
From:   Miklos Szeredi <miklos@szeredi.hu>
Date:   Mon, 22 Aug 2022 17:24:24 +0200
Message-ID: <CAJfpeguyD-znkZVwmiYZCK6tMsoJc+UzMKnkWxb7TToT1DFb4Q@mail.gmail.com>
Subject: Re: [PATCH] ovl: Fix potential memory leak
To:     Stanislav Goriainov <goriainov@ispras.ru>
Cc:     linux-unionfs@vger.kernel.org, linux-kernel@vger.kernel.org,
        ldv-project@linuxtesting.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, 22 Aug 2022 at 13:53, Stanislav Goriainov <goriainov@ispras.ru> wrote:
>
> ovl: Fix potential memory leak in ovl_lookup()
>
> If memory for uperredirect was allocated with kstrdup()
> in upperdir != NULL and d.redirect != NULL path,
> it may be lost when upperredirect is reassigned later.

Can't happen because the first assignment of upperredirect will only
happen if upperdentry is non-NULL, while second one will only happen
if upperdentry is NULL.   I understand why static checker fails to see
this: it doesn't know that dentry->d_name will never contain '/'.  In
this case the looped call to ovl_lookup_single() can be ignored and it
is trivial to prove that d.redirect can only be set if *ret is
non-NULL.

Thanks,
Miklos