linux-kernel - Re: [PATCH v4 3/8] crypto: x86/camellia: Use new glue function macros

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: <201911121452.AE2672AECB@keescook>
Date:   Tue, 12 Nov 2019 14:56:44 -0800
From:   Kees Cook <keescook@...omium.org>
To:     Herbert Xu <herbert@...dor.apana.org.au>
Cc:     Stephan Müller <smueller@...onox.de>,
        João Moreira <joao.moreira@....ic.unicamp.br>,
        Sami Tolvanen <samitolvanen@...gle.com>,
        "David S. Miller" <davem@...emloft.net>,
        Ard Biesheuvel <ard.biesheuvel@...aro.org>, x86@...nel.org,
        linux-crypto@...r.kernel.org, linux-kernel@...r.kernel.org,
        kernel-hardening@...ts.openwall.com
Subject: Re: [PATCH v4 3/8] crypto: x86/camellia: Use new glue function macros

On Tue, Nov 12, 2019 at 11:16:35AM +0800, Herbert Xu wrote:
> On Mon, Nov 11, 2019 at 07:14:17PM -0800, Eric Biggers wrote:
> >
> > Also, I don't see the point of the macros, other than to obfuscate things.  To
> > keep things straightforward, I think we should keep the explicit function
> > prototypes for each algorithm.
> 
> I agree.  Kees, please get rid of the macros.

Okay, if we do that, then we'll likely be dropping a lot of union logic
(since ecb and cbc end up with identical params and ctr and xts do too):

typedef void (*common_glue_func_t)(void *ctx, u8 *dst, const u8 *src);
typedef void (*common_glue_cbc_func_t)(void *ctx, u128 *dst, const u128 *src);
typedef void (*common_glue_ctr_func_t)(void *ctx, u128 *dst, const u128 *src,
                                       le128 *iv);
typedef void (*common_glue_xts_func_t)(void *ctx, u128 *dst, const u128 *src,
                                       le128 *iv);
...
struct common_glue_func_entry {
        unsigned int num_blocks; /* number of blocks that @fn will process */
        union { 
                common_glue_func_t ecb;
                common_glue_cbc_func_t cbc;
                common_glue_ctr_func_t ctr;
                common_glue_xts_func_t xts;
        } fn_u;
};

These would end up being just:

typedef void (*common_glue_func_t)(void *ctx, u8 *dst, const u8 *src);
typedef void (*common_glue_iv_func_t)(void *ctx, u8 *dst, const u8 *src,
                                       le128 *iv);
...
struct common_glue_func_entry {
        unsigned int num_blocks; /* number of blocks that @fn will process */
        union { 
                common_glue_func_t func;
                common_glue_iv_func_t iv_func;
        } fn_u;

Is that reasonable?

-- 
Kees Cook