linux-kernel - Re: [PATCH net] net/tls: fix slab-out-of-bounds bug in decrypt

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] [day] [month] [year] [list]

Message-ID: <20220330195228.21616546@kernel.org>
Date:   Wed, 30 Mar 2022 19:52:28 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     "Ziyang Xuan (William)" <william.xuanziyang@...wei.com>
Cc:     Ard Biesheuvel <ardb@...nel.org>,
        Eric Biggers <ebiggers@...gle.com>,
        Herbert Xu <herbert@...dor.apana.org.au>, <borisp@...dia.com>,
        <john.fastabend@...il.com>, <daniel@...earbox.net>,
        <davem@...emloft.net>, <pabeni@...hat.com>,
        <netdev@...r.kernel.org>, <vakul.garg@....com>,
        <davejwatson@...com>, <linux-kernel@...r.kernel.org>,
        Vadim Fedorenko <vfedorenko@...ek.ru>,
        <linux-crypto@...r.kernel.org>
Subject: Re: [PATCH net] net/tls: fix slab-out-of-bounds bug in
 decrypt_internal

On Thu, 31 Mar 2022 10:35:41 +0800 Ziyang Xuan (William) wrote:
> I am thinking about is skb_copy_bits() necessary in non-TLS_1_3_VERSION
> and non-TLS_CIPHER_CHACHA20_POLY1305 scenarios?

It's not necessary there, but we should not make that change be part of
the fix, the fix should be minimal. I'll send a separate patch to move
the skb_copy_bits() call later on.

I think for the fix all you should do is replace the
	crypto_aead_ivsize(ctx->aead_recv));
line with
	prot->iv_size + prot->salt_size);

> If the inital iv+salt negotiated configuration for tx/rx offload is right
> and reliable, what is the reason why we have to extract the iv value from
> received skb instead if using the negotiated iv value? Does it can be
> modified or just follow spec that versions below TLS_1_3_VERSION?

TLS 1.3 does not send the nonce as part of the record. Instead 
the record number is always used as nonce in crypto.