linux-kernel - Re: [PATCH 3/6] crypto: sha256 - Move lib/sha256.c to lib/crypto

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:   Sat, 17 Aug 2019 10:28:04 +0200
From:   Hans de Goede <hdegoede@...hat.com>
To:     Herbert Xu <herbert@...dor.apana.org.au>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        "H . Peter Anvin" <hpa@...or.com>,
        Heiko Carstens <heiko.carstens@...ibm.com>,
        Vasily Gorbik <gor@...ux.ibm.com>,
        Christian Borntraeger <borntraeger@...ibm.com>,
        Ard Biesheuvel <ard.biesheuvel@...aro.org>,
        linux-crypto@...r.kernel.org, x86@...nel.org,
        linux-s390@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 3/6] crypto: sha256 - Move lib/sha256.c to lib/crypto

Hi,

On 17-08-19 07:19, Eric Biggers wrote:
> On Fri, Aug 16, 2019 at 11:16:08PM +0200, Hans de Goede wrote:
>> diff --git a/include/linux/sha256.h b/include/crypto/sha256.h
>> similarity index 100%
>> rename from include/linux/sha256.h
>> rename to include/crypto/sha256.h
> 
> <crypto/sha.h> already has the declarations for both SHA-1 and SHA-2, including
> SHA-256.  So I'm not sure a separate sha256.h is appropriate.  How about putting
> these declarations in <crypto/sha.h>?

The problems with that is that the sha256_init, etc. names are quite generic
and they have not been reserved before, so a lot of the crypto hw-accel
drivers use them, for private file-local (static) code, e.g.:

[hans@...lem linux]$ ack -l sha256_init
include/crypto/sha256.h
drivers/crypto/marvell/hash.c
drivers/crypto/ccp/ccp-ops.c
drivers/crypto/nx/nx-sha256.c
drivers/crypto/ux500/hash/hash_core.c
drivers/crypto/inside-secure/safexcel_hash.c
drivers/crypto/chelsio/chcr_algo.h
drivers/crypto/stm32/stm32-hash.c
drivers/crypto/omap-sham.c
drivers/crypto/padlock-sha.c
drivers/crypto/n2_core.c
drivers/crypto/atmel-aes.c
drivers/crypto/axis/artpec6_crypto.c
drivers/crypto/mediatek/mtk-sha.c
drivers/crypto/qat/qat_common/qat_algs.c
drivers/crypto/img-hash.c
drivers/crypto/ccree/cc_hash.c
lib/crypto/sha256.c
arch/powerpc/crypto/sha256-spe-glue.c
arch/mips/cavium-octeon/crypto/octeon-sha256.c
arch/x86/purgatory/purgatory.c
arch/s390/crypto/sha256_s390.c
arch/s390/purgatory/purgatory.c

(in case you do not know ack is a smarter grep, which skips .o files, etc.)

All these do include crypto/sha.h and putting the stuff which is in what
was linux/sha256.h into crypto/sha.h leads to name collisions which causes
more churn then I would like this series to cause.

I guess we could do a cleanup afterwards, with one patch per file above
to fix the name collision issue, and then merge the 2 headers. I do not
want to do that for this series, as I want to keep this series as KISS
as possible since it is messing with somewhat sensitive stuff.

And TBH I even wonder if a follow-up series is worth the churn...

Regards,

Hans