Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.31 as permitted sender) client-ip=23.128.96.31;
Date:   Wed, 18 Oct 2023 18:25:55 -0700
From:   Jakub Kicinski <kuba@kernel.org>
To:     Juntong Deng <juntong.deng@outlook.com>
Cc:     borisp@nvidia.com, john.fastabend@gmail.com, davem@davemloft.net,
        edumazet@google.com, pabeni@redhat.com, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        linux-kernel-mentees@lists.linuxfoundation.org,
        syzbot+29c22ea2d6b2c5fd2eae@syzkaller.appspotmail.com,
        Sabrina Dubroca <sd@queasysnail.net>
Subject: Re: [PATCH v2] net/tls: Fix slab-use-after-free in tls_encrypt_done
Message-ID: <20231018182555.28f1a774@kernel.org>
In-Reply-To: <VI1P193MB0752321F24623E024C87886A99D6A@VI1P193MB0752.EURP193.PROD.OUTLOOK.COM>
References: <VI1P193MB0752321F24623E024C87886A99D6A@VI1P193MB0752.EURP193.PROD.OUTLOOK.COM>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Wed, 18 Oct 2023 00:22:15 +0800 Juntong Deng wrote:
> In the current implementation, ctx->async_wait.completion is completed
> after spin_lock_bh, which causes tls_sw_release_resources_tx to
> continue executing and return to tls_sk_proto_cleanup, then return
> to tls_sk_proto_close, and after that enter tls_sw_free_ctx_tx to kfree
> the entire struct tls_sw_context_tx (including ctx->encrypt_compl_lock).
> 
> Since ctx->encrypt_compl_lock has been freed, subsequent spin_unlock_bh
> will result in slab-use-after-free error. Due to SMP, even using
> spin_lock_bh does not prevent tls_sw_release_resources_tx from continuing
> on other CPUs. After tls_sw_release_resources_tx is woken up, there is no
> attempt to hold ctx->encrypt_compl_lock again, therefore everything
> described above is possible.

Whoever triggered the Tx should wait for all outstanding encryption 
to finish before exiting sendmsg() (or alike).  This looks like 
a band-aid. Sabrina is working on fixes for the async code, lets
get those in first before attempting spot fixes.
-- 
pw-bot: cr