[FFmpeg-devel] [PATCH] SHA-256 support

[Michael Niedermayer]

On Tue, Jun 09, 2009 at 09:32:24PM +0300, Kostya wrote:
> On Tue, Jun 09, 2009 at 04:47:09PM +0200, Michael Niedermayer wrote:
> > On Tue, Jun 09, 2009 at 08:13:40AM +0300, Kostya wrote:
> [...]
> > > /* Hash a single 512-bit block. This is the core of the algorithm. */
> > > 
> > > static void transform(uint32_t state[5], const uint8_t buffer[64])
> > 
> > doxy?
>  
> no

make it a yes please


>  
> > > {
> > >     unsigned int i, a, b, c, d, e, f, g, h;
> > >     uint32_t block[80];
> > >     uint32_t T1, T2;
> > > 
> > >     a = state[0];
> > >     b = state[1];
> > >     c = state[2];
> > >     d = state[3];
> > >     e = state[4];
> > >     f = state[5];
> > >     g = state[6];
> > >     h = state[7];
> > > 
> > >     for (i = 0; i < 64; i++) {
> > >         if (i < 16) {
> > >             T1 = block[i] = be2me_32(((const uint32_t*)buffer)[i]);
> > >         } else {
> > >             block[i & 0xF] +=            block[(i +  9) & 0xF]
> > >                             + sigma0_256(block[(i +  1) & 0xF])
> > >                             + sigma1_256(block[(i + 14) & 0xF]);
> > >             T1 = block[i & 0xF];
> > >         }
> > 
> > you have lost some optimizations from our sha1 here, 
> > 
> > also sha224 should be supported because its almost identical
> 
> Here you are. Seems to work fine on x86, should be fine on BE systems as
> well (will test tomorrow). 
> 
> > [...]
> > 
> > -- 
> > Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

>  Makefile |    6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 9f3b5755dfe516571ef2196e1460c648e15423b1  sha2.patch
> Index: libavutil/Makefile
> ===================================================================
> --- libavutil/Makefile	(revision 18770)
> +++ libavutil/Makefile	(working copy)
> @@ -17,7 +17,8 @@
>            mem.h                                                         \
>            pixfmt.h                                                      \
>            rational.h                                                    \
> -          sha1.h
> +          sha1.h                                                        \
> +          sha2.h

the sha2.h should also have a \ at the end so future patches are dont need to
move the lack of a \ down


>  
>  OBJS = adler32.o                                                        \
>         aes.o                                                            \
> @@ -38,10 +39,11 @@
>         rational.o                                                       \
>         rc4.o                                                            \
>         sha1.o                                                           \
> +       sha2.o                                                           \
>         tree.o                                                           \
>         utils.o                                                          \
>  
> -TESTPROGS = adler32 aes base64 crc des lls md5 pca sha1 softfloat tree
> +TESTPROGS = adler32 aes base64 crc des lls md5 pca sha1 sha2 softfloat tree
>  TESTPROGS-$(HAVE_LZO1X_999_COMPRESS) += lzo
>  
>  DIRS = arm bfin sh4 x86

> /*
>  * Copyright (C) 2009 Konstantin Shishkov
>  *
>  * This file is part of FFmpeg.
>  *
>  * FFmpeg is free software; you can redistribute it and/or
>  * modify it under the terms of the GNU Lesser General Public
>  * License as published by the Free Software Foundation; either
>  * version 2.1 of the License, or (at your option) any later version.
>  *
>  * FFmpeg is distributed in the hope that it will be useful,
>  * but WITHOUT ANY WARRANTY; without even the implied warranty of
>  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
>  * Lesser General Public License for more details.
>  *
>  * You should have received a copy of the GNU Lesser General Public
>  * License along with FFmpeg; if not, write to the Free Software
>  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
>  */
> 
> #ifndef AVUTIL_SHA2_H
> #define AVUTIL_SHA2_H
> 
> #include <stdint.h>
> 
> extern const int av_sha256_size;
> 
> struct AVSHA256;
> 
> void av_sha224_init(struct AVSHA256* context);
> #define av_sha224_update av_sha_256_update
> void av_sha224_final(struct AVSHA256* context, uint8_t digest[28]);
> 
> void av_sha256_init(struct AVSHA256* context);
> void av_sha256_update(struct AVSHA256* context, const uint8_t* data, unsigned int len);
> void av_sha256_final(struct AVSHA256* context, uint8_t digest[32]);

i think a int bits is nicer than 2 sets of functions
it also might safe you from some comments about code duplication


[...]

> static void transform(uint32_t state[5], const uint8_t buffer[64])
> {
>     unsigned int i, a, b, c, d, e, f, g, h;
>     uint32_t block[16];
>     uint32_t T1, T2;
> 
>     a = state[0];
>     b = state[1];
>     c = state[2];
>     d = state[3];
>     e = state[4];
>     f = state[5];
>     g = state[6];
>     h = state[7];
> #if CONFIG_SMALL
>     for (i = 0; i < 64; i++) {
>         if (i < 16) {
>             T1 = block[i] = be2me_32(((const uint32_t*)buffer)[i]);
>         } else {
>             block[i & 0xF] +=            block[(i +  9) & 0xF]
>                             + sigma0_256(block[(i +  1) & 0xF])
>                             + sigma1_256(block[(i + 14) & 0xF]);
>             T1 = block[i & 0xF];

now you removed the other part of the optimization
please restore the size of the block array and get rid of the & 0xF

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Its not that you shouldnt use gotos but rather that you should write
readable code and code with gotos often but not always is less readable
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20090609/3635752e/attachment.pgp>