From: Denys Vlasenko <vda.linux@googlemail.com>
To: Herbert Xu <herbert@gondor.apana.org.au>
Cc: linux-crypto@vger.kernel.org
Subject: [PATCH 4/5] camellia: de-unrolling
Date: Thu, 25 Oct 2007 12:47:16 +0100 [thread overview]
Message-ID: <200710251247.16843.vda.linux@googlemail.com> (raw)
In-Reply-To: <200710251243.58701.vda.linux@googlemail.com>
[-- Attachment #1: Type: text/plain, Size: 494 bytes --]
On Thursday 25 October 2007 12:43, Denys Vlasenko wrote:
> Hi Hervert,
>
> Please review and maybe propagate upstream following patches.
>
> camellia4.diff
> Move huge unrolled pieces of code (3 screenfuls) at the end of
> 128/256 key setup routines into common camellia_setup_tail(),
> convert it to loop there.
> Loop is still unrolled six times, so performance hit is very small,
> code size win is big.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
--
vda
[-- Attachment #2: camellia4.diff --]
[-- Type: text/x-diff, Size: 6849 bytes --]
--- linux-2.6.23.src/crypto/camellia3.c 2007-10-24 19:03:27.000000000 +0100
+++ linux-2.6.23.src/crypto/camellia.c 2007-10-24 19:03:57.000000000 +0100
@@ -424,6 +424,27 @@ static const u32 camellia_sp4404[256] =
#define SUBKEY_L(INDEX) (subkey[(INDEX)*2])
#define SUBKEY_R(INDEX) (subkey[(INDEX)*2 + 1])
+static void camellia_setup_tail(u32 *subkey, int max)
+{
+ u32 dw;
+ int i = 2;
+ do {
+ dw = SUBKEY_L(i + 0) ^ SUBKEY_R(i + 0); dw = ROL8(dw);/* round 1 */
+ SUBKEY_R(i + 0) = SUBKEY_L(i + 0) ^ dw; SUBKEY_L(i + 0) = dw;
+ dw = SUBKEY_L(i + 1) ^ SUBKEY_R(i + 1); dw = ROL8(dw);/* round 2 */
+ SUBKEY_R(i + 1) = SUBKEY_L(i + 1) ^ dw; SUBKEY_L(i + 1) = dw;
+ dw = SUBKEY_L(i + 2) ^ SUBKEY_R(i + 2); dw = ROL8(dw);/* round 3 */
+ SUBKEY_R(i + 2) = SUBKEY_L(i + 2) ^ dw; SUBKEY_L(i + 2) = dw;
+ dw = SUBKEY_L(i + 3) ^ SUBKEY_R(i + 3); dw = ROL8(dw);/* round 4 */
+ SUBKEY_R(i + 3) = SUBKEY_L(i + 3) ^ dw; SUBKEY_L(i + 3) = dw;
+ dw = SUBKEY_L(i + 4) ^ SUBKEY_R(i + 4); dw = ROL8(dw);/* round 5 */
+ SUBKEY_R(i + 4) = SUBKEY_L(i + 4) ^ dw; SUBKEY_L(i + 4) = dw;
+ dw = SUBKEY_L(i + 5) ^ SUBKEY_R(i + 5); dw = ROL8(dw);/* round 6 */
+ SUBKEY_R(i + 5) = SUBKEY_L(i + 5) ^ dw; SUBKEY_L(i + 5) = dw;
+ i += 8;
+ } while (i < max);
+}
+
static void camellia_setup128(const unsigned char *key, u32 *subkey)
{
u32 kll, klr, krl, krr;
@@ -650,42 +671,7 @@ static void camellia_setup128(const unsi
SUBKEY_R(24) = subR[24] ^ subR[23];
/* apply the inverse of the last half of P-function */
- dw = SUBKEY_L(2) ^ SUBKEY_R(2); dw = ROL8(dw);/* round 1 */
- SUBKEY_R(2) = SUBKEY_L(2) ^ dw; SUBKEY_L(2) = dw;
- dw = SUBKEY_L(3) ^ SUBKEY_R(3); dw = ROL8(dw);/* round 2 */
- SUBKEY_R(3) = SUBKEY_L(3) ^ dw; SUBKEY_L(3) = dw;
- dw = SUBKEY_L(4) ^ SUBKEY_R(4); dw = ROL8(dw);/* round 3 */
- SUBKEY_R(4) = SUBKEY_L(4) ^ dw; SUBKEY_L(4) = dw;
- dw = SUBKEY_L(5) ^ SUBKEY_R(5); dw = ROL8(dw);/* round 4 */
- SUBKEY_R(5) = SUBKEY_L(5) ^ dw; SUBKEY_L(5) = dw;
- dw = SUBKEY_L(6) ^ SUBKEY_R(6); dw = ROL8(dw);/* round 5 */
- SUBKEY_R(6) = SUBKEY_L(6) ^ dw; SUBKEY_L(6) = dw;
- dw = SUBKEY_L(7) ^ SUBKEY_R(7); dw = ROL8(dw);/* round 6 */
- SUBKEY_R(7) = SUBKEY_L(7) ^ dw; SUBKEY_L(7) = dw;
- dw = SUBKEY_L(10) ^ SUBKEY_R(10); dw = ROL8(dw);/* round 7 */
- SUBKEY_R(10) = SUBKEY_L(10) ^ dw; SUBKEY_L(10) = dw;
- dw = SUBKEY_L(11) ^ SUBKEY_R(11); dw = ROL8(dw);/* round 8 */
- SUBKEY_R(11) = SUBKEY_L(11) ^ dw; SUBKEY_L(11) = dw;
- dw = SUBKEY_L(12) ^ SUBKEY_R(12); dw = ROL8(dw);/* round 9 */
- SUBKEY_R(12) = SUBKEY_L(12) ^ dw; SUBKEY_L(12) = dw;
- dw = SUBKEY_L(13) ^ SUBKEY_R(13); dw = ROL8(dw);/* round 10 */
- SUBKEY_R(13) = SUBKEY_L(13) ^ dw; SUBKEY_L(13) = dw;
- dw = SUBKEY_L(14) ^ SUBKEY_R(14); dw = ROL8(dw);/* round 11 */
- SUBKEY_R(14) = SUBKEY_L(14) ^ dw; SUBKEY_L(14) = dw;
- dw = SUBKEY_L(15) ^ SUBKEY_R(15); dw = ROL8(dw);/* round 12 */
- SUBKEY_R(15) = SUBKEY_L(15) ^ dw; SUBKEY_L(15) = dw;
- dw = SUBKEY_L(18) ^ SUBKEY_R(18); dw = ROL8(dw);/* round 13 */
- SUBKEY_R(18) = SUBKEY_L(18) ^ dw; SUBKEY_L(18) = dw;
- dw = SUBKEY_L(19) ^ SUBKEY_R(19); dw = ROL8(dw);/* round 14 */
- SUBKEY_R(19) = SUBKEY_L(19) ^ dw; SUBKEY_L(19) = dw;
- dw = SUBKEY_L(20) ^ SUBKEY_R(20); dw = ROL8(dw);/* round 15 */
- SUBKEY_R(20) = SUBKEY_L(20) ^ dw; SUBKEY_L(20) = dw;
- dw = SUBKEY_L(21) ^ SUBKEY_R(21); dw = ROL8(dw);/* round 16 */
- SUBKEY_R(21) = SUBKEY_L(21) ^ dw; SUBKEY_L(21) = dw;
- dw = SUBKEY_L(22) ^ SUBKEY_R(22); dw = ROL8(dw);/* round 17 */
- SUBKEY_R(22) = SUBKEY_L(22) ^ dw; SUBKEY_L(22) = dw;
- dw = SUBKEY_L(23) ^ SUBKEY_R(23); dw = ROL8(dw);/* round 18 */
- SUBKEY_R(23) = SUBKEY_L(23) ^ dw; SUBKEY_L(23) = dw;
+ camellia_setup_tail(subkey, 24);
}
static void camellia_setup256(const unsigned char *key, u32 *subkey)
@@ -995,54 +981,7 @@ static void camellia_setup256(const unsi
SUBKEY_R(32) = subR[32] ^ subR[31];
/* apply the inverse of the last half of P-function */
- dw = SUBKEY_L(2) ^ SUBKEY_R(2); dw = ROL8(dw);/* round 1 */
- SUBKEY_R(2) = SUBKEY_L(2) ^ dw; SUBKEY_L(2) = dw;
- dw = SUBKEY_L(3) ^ SUBKEY_R(3); dw = ROL8(dw);/* round 2 */
- SUBKEY_R(3) = SUBKEY_L(3) ^ dw; SUBKEY_L(3) = dw;
- dw = SUBKEY_L(4) ^ SUBKEY_R(4); dw = ROL8(dw);/* round 3 */
- SUBKEY_R(4) = SUBKEY_L(4) ^ dw; SUBKEY_L(4) = dw;
- dw = SUBKEY_L(5) ^ SUBKEY_R(5); dw = ROL8(dw);/* round 4 */
- SUBKEY_R(5) = SUBKEY_L(5) ^ dw; SUBKEY_L(5) = dw;
- dw = SUBKEY_L(6) ^ SUBKEY_R(6); dw = ROL8(dw);/* round 5 */
- SUBKEY_R(6) = SUBKEY_L(6) ^ dw; SUBKEY_L(6) = dw;
- dw = SUBKEY_L(7) ^ SUBKEY_R(7); dw = ROL8(dw);/* round 6 */
- SUBKEY_R(7) = SUBKEY_L(7) ^ dw; SUBKEY_L(7) = dw;
- dw = SUBKEY_L(10) ^ SUBKEY_R(10); dw = ROL8(dw);/* round 7 */
- SUBKEY_R(10) = SUBKEY_L(10) ^ dw; SUBKEY_L(10) = dw;
- dw = SUBKEY_L(11) ^ SUBKEY_R(11); dw = ROL8(dw);/* round 8 */
- SUBKEY_R(11) = SUBKEY_L(11) ^ dw; SUBKEY_L(11) = dw;
- dw = SUBKEY_L(12) ^ SUBKEY_R(12); dw = ROL8(dw);/* round 9 */
- SUBKEY_R(12) = SUBKEY_L(12) ^ dw; SUBKEY_L(12) = dw;
- dw = SUBKEY_L(13) ^ SUBKEY_R(13); dw = ROL8(dw);/* round 10 */
- SUBKEY_R(13) = SUBKEY_L(13) ^ dw; SUBKEY_L(13) = dw;
- dw = SUBKEY_L(14) ^ SUBKEY_R(14); dw = ROL8(dw);/* round 11 */
- SUBKEY_R(14) = SUBKEY_L(14) ^ dw; SUBKEY_L(14) = dw;
- dw = SUBKEY_L(15) ^ SUBKEY_R(15); dw = ROL8(dw);/* round 12 */
- SUBKEY_R(15) = SUBKEY_L(15) ^ dw; SUBKEY_L(15) = dw;
- dw = SUBKEY_L(18) ^ SUBKEY_R(18); dw = ROL8(dw);/* round 13 */
- SUBKEY_R(18) = SUBKEY_L(18) ^ dw; SUBKEY_L(18) = dw;
- dw = SUBKEY_L(19) ^ SUBKEY_R(19); dw = ROL8(dw);/* round 14 */
- SUBKEY_R(19) = SUBKEY_L(19) ^ dw; SUBKEY_L(19) = dw;
- dw = SUBKEY_L(20) ^ SUBKEY_R(20); dw = ROL8(dw);/* round 15 */
- SUBKEY_R(20) = SUBKEY_L(20) ^ dw; SUBKEY_L(20) = dw;
- dw = SUBKEY_L(21) ^ SUBKEY_R(21); dw = ROL8(dw);/* round 16 */
- SUBKEY_R(21) = SUBKEY_L(21) ^ dw; SUBKEY_L(21) = dw;
- dw = SUBKEY_L(22) ^ SUBKEY_R(22); dw = ROL8(dw);/* round 17 */
- SUBKEY_R(22) = SUBKEY_L(22) ^ dw; SUBKEY_L(22) = dw;
- dw = SUBKEY_L(23) ^ SUBKEY_R(23); dw = ROL8(dw);/* round 18 */
- SUBKEY_R(23) = SUBKEY_L(23) ^ dw; SUBKEY_L(23) = dw;
- dw = SUBKEY_L(26) ^ SUBKEY_R(26); dw = ROL8(dw);/* round 19 */
- SUBKEY_R(26) = SUBKEY_L(26) ^ dw; SUBKEY_L(26) = dw;
- dw = SUBKEY_L(27) ^ SUBKEY_R(27); dw = ROL8(dw);/* round 20 */
- SUBKEY_R(27) = SUBKEY_L(27) ^ dw; SUBKEY_L(27) = dw;
- dw = SUBKEY_L(28) ^ SUBKEY_R(28); dw = ROL8(dw);/* round 21 */
- SUBKEY_R(28) = SUBKEY_L(28) ^ dw; SUBKEY_L(28) = dw;
- dw = SUBKEY_L(29) ^ SUBKEY_R(29); dw = ROL8(dw);/* round 22 */
- SUBKEY_R(29) = SUBKEY_L(29) ^ dw; SUBKEY_L(29) = dw;
- dw = SUBKEY_L(30) ^ SUBKEY_R(30); dw = ROL8(dw);/* round 23 */
- SUBKEY_R(30) = SUBKEY_L(30) ^ dw; SUBKEY_L(30) = dw;
- dw = SUBKEY_L(31) ^ SUBKEY_R(31); dw = ROL8(dw);/* round 24 */
- SUBKEY_R(31) = SUBKEY_L(31) ^ dw; SUBKEY_L(31) = dw;
+ camellia_setup_tail(subkey, 32);
}
static void camellia_setup192(const unsigned char *key, u32 *subkey)
next prev parent reply other threads:[~2007-10-25 11:47 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-10-25 11:43 [PATCH0/5] camellia: cleanup, de-unrolling, and 64bit-ization Denys Vlasenko
2007-10-25 11:45 ` [PATCH 1/5] camellia: cleanup Denys Vlasenko
2007-10-26 8:43 ` Noriaki TAKAMIYA
2007-11-06 14:17 ` Herbert Xu
2007-10-25 11:45 ` [PATCH 2/5] " Denys Vlasenko
2007-10-26 8:44 ` Noriaki TAKAMIYA
2007-11-06 14:19 ` Herbert Xu
2007-10-25 11:46 ` [PATCH 3/5] " Denys Vlasenko
2007-10-26 8:44 ` Noriaki TAKAMIYA
2007-11-06 14:21 ` Herbert Xu
2007-10-25 11:47 ` Denys Vlasenko [this message]
2007-10-26 8:45 ` [PATCH 4/5] camellia: de-unrolling Noriaki TAKAMIYA
2007-11-06 14:21 ` Herbert Xu
2007-10-25 11:48 ` [PATCH 5/5] camellia: de-unrolling, 64bit-ization Denys Vlasenko
2007-10-26 8:45 ` Noriaki TAKAMIYA
2007-11-06 14:23 ` Herbert Xu
2007-11-07 13:22 ` Denys Vlasenko
2007-11-08 13:30 ` Herbert Xu
2007-11-13 6:07 ` Noriaki TAKAMIYA
2007-11-13 6:25 ` [camellia-oss:00952] " Noriaki TAKAMIYA
2007-11-13 22:34 ` Denys Vlasenko
2007-11-14 1:41 ` David Miller
2007-11-14 2:47 ` Denys Vlasenko
2007-11-14 3:49 ` David Miller
2007-11-14 5:30 ` Denys Vlasenko
2007-11-14 6:10 ` David Miller
2007-11-14 7:38 ` Denys Vlasenko
2007-11-14 7:15 ` Denys Vlasenko
2007-11-14 14:14 ` Herbert Xu
2007-11-14 21:28 ` Denys Vlasenko
2007-11-18 13:21 ` Herbert Xu
2007-11-19 4:30 ` Denys Vlasenko
2007-11-19 18:49 ` Noriaki TAKAMIYA
2007-11-21 2:44 ` Denys Vlasenko
2007-11-21 3:53 ` Herbert Xu
2007-11-21 8:08 ` Denys Vlasenko
2007-11-21 8:12 ` Herbert Xu
2007-11-21 8:38 ` Denys Vlasenko
2007-11-14 4:18 ` Noriaki TAKAMIYA
2007-10-25 11:57 ` [PATCH0/5] camellia: cleanup, de-unrolling, and 64bit-ization Denys Vlasenko
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=200710251247.16843.vda.linux@googlemail.com \
--to=vda.linux@googlemail.com \
--cc=herbert@gondor.apana.org.au \
--cc=linux-crypto@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).