monero/src/crypto/aesb.c

187 lines
7.1 KiB
C
Raw Normal View History

2014-05-17 23:25:58 +00:00
/*
---------------------------------------------------------------------------
Copyright (c) 1998-2013, Brian Gladman, Worcester, UK. All rights reserved.
The redistribution and use of this software (with or without changes)
is allowed without the payment of fees or royalties provided that:
source code distributions include the above copyright notice, this
list of conditions and the following disclaimer;
binary distributions include the above copyright notice, this list
of conditions and the following disclaimer in their documentation.
This software is provided 'as is' with no explicit or implied warranties
in respect of its operation, including, but not limited to, correctness
and fitness for purpose.
---------------------------------------------------------------------------
Issue Date: 20/12/2007
*/
#include <stdint.h>
#include "common/int-util.h"
2014-05-17 23:25:58 +00:00
#if defined(__cplusplus)
extern "C"
{
#endif
#define TABLE_ALIGN 32
#define WPOLY 0x011b
#define N_COLS 4
#define AES_BLOCK_SIZE 16
#define RC_LENGTH (5 * (AES_BLOCK_SIZE / 4 - 2))
2014-05-17 23:25:58 +00:00
#if defined(_MSC_VER)
#define ALIGN __declspec(align(TABLE_ALIGN))
#elif defined(__GNUC__)
#define ALIGN __attribute__ ((aligned(16)))
#else
#define ALIGN
#endif
#define rf1(r,c) (r)
#define word_in(x,c) (*((uint32_t*)(x)+(c)))
#define word_out(x,c,v) (*((uint32_t*)(x)+(c)) = (v))
#define s(x,c) x[c]
#define si(y,x,c) (s(y,c) = word_in(x, c))
#define so(y,x,c) word_out(y, c, s(x,c))
#define state_in(y,x) si(y,x,0); si(y,x,1); si(y,x,2); si(y,x,3)
#define state_out(y,x) so(y,x,0); so(y,x,1); so(y,x,2); so(y,x,3)
#define round(rm,y,x,k) rm(y,x,k,0); rm(y,x,k,1); rm(y,x,k,2); rm(y,x,k,3)
#define to_byte(x) ((x) & 0xff)
#define bval(x,n) to_byte(SWAP32LE(x) >> (8 * (n)))
2014-05-17 23:25:58 +00:00
#define fwd_var(x,r,c)\
( r == 0 ? ( c == 0 ? s(x,0) : c == 1 ? s(x,1) : c == 2 ? s(x,2) : s(x,3))\
: r == 1 ? ( c == 0 ? s(x,1) : c == 1 ? s(x,2) : c == 2 ? s(x,3) : s(x,0))\
: r == 2 ? ( c == 0 ? s(x,2) : c == 1 ? s(x,3) : c == 2 ? s(x,0) : s(x,1))\
: ( c == 0 ? s(x,3) : c == 1 ? s(x,0) : c == 2 ? s(x,1) : s(x,2)))
#define fwd_rnd(y,x,k,c) (s(y,c) = (k)[c] ^ SWAP32LE(four_tables(x,t_use(f,n),fwd_var,rf1,c)))
2014-05-17 23:25:58 +00:00
#define sb_data(w) {\
w(0x63), w(0x7c), w(0x77), w(0x7b), w(0xf2), w(0x6b), w(0x6f), w(0xc5),\
w(0x30), w(0x01), w(0x67), w(0x2b), w(0xfe), w(0xd7), w(0xab), w(0x76),\
w(0xca), w(0x82), w(0xc9), w(0x7d), w(0xfa), w(0x59), w(0x47), w(0xf0),\
w(0xad), w(0xd4), w(0xa2), w(0xaf), w(0x9c), w(0xa4), w(0x72), w(0xc0),\
w(0xb7), w(0xfd), w(0x93), w(0x26), w(0x36), w(0x3f), w(0xf7), w(0xcc),\
w(0x34), w(0xa5), w(0xe5), w(0xf1), w(0x71), w(0xd8), w(0x31), w(0x15),\
w(0x04), w(0xc7), w(0x23), w(0xc3), w(0x18), w(0x96), w(0x05), w(0x9a),\
w(0x07), w(0x12), w(0x80), w(0xe2), w(0xeb), w(0x27), w(0xb2), w(0x75),\
w(0x09), w(0x83), w(0x2c), w(0x1a), w(0x1b), w(0x6e), w(0x5a), w(0xa0),\
w(0x52), w(0x3b), w(0xd6), w(0xb3), w(0x29), w(0xe3), w(0x2f), w(0x84),\
w(0x53), w(0xd1), w(0x00), w(0xed), w(0x20), w(0xfc), w(0xb1), w(0x5b),\
w(0x6a), w(0xcb), w(0xbe), w(0x39), w(0x4a), w(0x4c), w(0x58), w(0xcf),\
w(0xd0), w(0xef), w(0xaa), w(0xfb), w(0x43), w(0x4d), w(0x33), w(0x85),\
w(0x45), w(0xf9), w(0x02), w(0x7f), w(0x50), w(0x3c), w(0x9f), w(0xa8),\
w(0x51), w(0xa3), w(0x40), w(0x8f), w(0x92), w(0x9d), w(0x38), w(0xf5),\
w(0xbc), w(0xb6), w(0xda), w(0x21), w(0x10), w(0xff), w(0xf3), w(0xd2),\
w(0xcd), w(0x0c), w(0x13), w(0xec), w(0x5f), w(0x97), w(0x44), w(0x17),\
w(0xc4), w(0xa7), w(0x7e), w(0x3d), w(0x64), w(0x5d), w(0x19), w(0x73),\
w(0x60), w(0x81), w(0x4f), w(0xdc), w(0x22), w(0x2a), w(0x90), w(0x88),\
w(0x46), w(0xee), w(0xb8), w(0x14), w(0xde), w(0x5e), w(0x0b), w(0xdb),\
w(0xe0), w(0x32), w(0x3a), w(0x0a), w(0x49), w(0x06), w(0x24), w(0x5c),\
w(0xc2), w(0xd3), w(0xac), w(0x62), w(0x91), w(0x95), w(0xe4), w(0x79),\
w(0xe7), w(0xc8), w(0x37), w(0x6d), w(0x8d), w(0xd5), w(0x4e), w(0xa9),\
w(0x6c), w(0x56), w(0xf4), w(0xea), w(0x65), w(0x7a), w(0xae), w(0x08),\
w(0xba), w(0x78), w(0x25), w(0x2e), w(0x1c), w(0xa6), w(0xb4), w(0xc6),\
w(0xe8), w(0xdd), w(0x74), w(0x1f), w(0x4b), w(0xbd), w(0x8b), w(0x8a),\
w(0x70), w(0x3e), w(0xb5), w(0x66), w(0x48), w(0x03), w(0xf6), w(0x0e),\
w(0x61), w(0x35), w(0x57), w(0xb9), w(0x86), w(0xc1), w(0x1d), w(0x9e),\
w(0xe1), w(0xf8), w(0x98), w(0x11), w(0x69), w(0xd9), w(0x8e), w(0x94),\
w(0x9b), w(0x1e), w(0x87), w(0xe9), w(0xce), w(0x55), w(0x28), w(0xdf),\
w(0x8c), w(0xa1), w(0x89), w(0x0d), w(0xbf), w(0xe6), w(0x42), w(0x68),\
w(0x41), w(0x99), w(0x2d), w(0x0f), w(0xb0), w(0x54), w(0xbb), w(0x16) }
2014-05-17 23:25:58 +00:00
#define rc_data(w) {\
w(0x01), w(0x02), w(0x04), w(0x08), w(0x10),w(0x20), w(0x40), w(0x80),\
w(0x1b), w(0x36) }
2014-05-17 23:25:58 +00:00
#define bytes2word(b0, b1, b2, b3) (((uint32_t)(b3) << 24) | \
((uint32_t)(b2) << 16) | ((uint32_t)(b1) << 8) | (b0))
2014-05-17 23:25:58 +00:00
#define h0(x) (x)
#define w0(p) bytes2word(p, 0, 0, 0)
#define w1(p) bytes2word(0, p, 0, 0)
#define w2(p) bytes2word(0, 0, p, 0)
#define w3(p) bytes2word(0, 0, 0, p)
#define u0(p) bytes2word(f2(p), p, p, f3(p))
#define u1(p) bytes2word(f3(p), f2(p), p, p)
#define u2(p) bytes2word(p, f3(p), f2(p), p)
#define u3(p) bytes2word(p, p, f3(p), f2(p))
#define v0(p) bytes2word(fe(p), f9(p), fd(p), fb(p))
#define v1(p) bytes2word(fb(p), fe(p), f9(p), fd(p))
#define v2(p) bytes2word(fd(p), fb(p), fe(p), f9(p))
#define v3(p) bytes2word(f9(p), fd(p), fb(p), fe(p))
#define f2(x) ((x<<1) ^ (((x>>7) & 1) * WPOLY))
#define f4(x) ((x<<2) ^ (((x>>6) & 1) * WPOLY) ^ (((x>>6) & 2) * WPOLY))
#define f8(x) ((x<<3) ^ (((x>>5) & 1) * WPOLY) ^ (((x>>5) & 2) * WPOLY) ^ (((x>>5) & 4) * WPOLY))
#define f3(x) (f2(x) ^ x)
#define f9(x) (f8(x) ^ x)
#define fb(x) (f8(x) ^ f2(x) ^ x)
#define fd(x) (f8(x) ^ f4(x) ^ x)
#define fe(x) (f8(x) ^ f4(x) ^ f2(x))
#define t_dec(m,n) t_##m##n
#define t_set(m,n) t_##m##n
#define t_use(m,n) t_##m##n
#define d_4(t,n,b,e,f,g,h) ALIGN const t n[4][256] = { b(e), b(f), b(g), b(h) }
#define four_tables(x,tab,vf,rf,c) \
(tab[0][bval(vf(x,0,c),rf(0,c))] \
^ tab[1][bval(vf(x,1,c),rf(1,c))] \
^ tab[2][bval(vf(x,2,c),rf(2,c))] \
^ tab[3][bval(vf(x,3,c),rf(3,c))])
2014-05-17 23:25:58 +00:00
d_4(uint32_t, t_dec(f,n), sb_data, u0, u1, u2, u3);
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY) Bockchain: 1. Optim: Multi-thread long-hash computation when encountering groups of blocks. 2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible. 3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible. 4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks. 5. Optim: Multi-thread signature computation whenever possible. 6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD) 7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???). 8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads). Berkeley-DB: 1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc). 2. Fix: Unable to pop blocks on reorganize due to transaction errors. 3. Patch: Large number of transaction aborts when running multi-threaded bulk queries. 4. Patch: Insufficient locks error when running full sync. 5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation. 6. Optim: Add bulk queries to get output global indices. 7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3) 8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key 9. Optim: Added thread-safe buffers used when multi-threading bulk queries. 10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details) 11. Mod: Added checkpoint thread and auto-remove-logs option. 12. *Now usable on 32-bit systems like RPI2. LMDB: 1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect) 2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3) 3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key 4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details) 5. Mod: Auto resize to +1GB instead of multiplier x1.5 ETC: 1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete. 2. Fix: 32-bit saturation bug when computing next difficulty on large blocks. [PENDING ISSUES] 1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization. This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD. 2. Berkeley db, possible bug "unable to allocate memory". TBD. [NEW OPTIONS] (*Currently all enabled for testing purposes) 1. --fast-block-sync arg=[0:1] (default: 1) a. 0 = Compute long hash per block (may take a while depending on CPU) b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence) 2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000) a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions. b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache. Fast - Write meta-data but defer data flush. Fastest - Defer meta-data and data flush. Sync - Flush data after nblocks_per_sync and wait. Async - Flush data after nblocks_per_sync but do not wait for the operation to finish. 3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower) Max number of threads to use when computing long-hash in groups. 4. --show-time-stats arg=[0:1] (default: 1) Show benchmark related time stats. 5. --db-auto-remove-logs arg=[0:1] (default: 1) For berkeley-db only. Auto remove logs if enabled. **Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version. At the moment, you need a full resync to use this optimized version. [PERFORMANCE COMPARISON] **Some figures are approximations only. Using a baseline machine of an i7-2600K+SSD+(with full pow computation): 1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain. 2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain. 3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain. Averate procesing times (with full pow computation): lmdb-optimized: 1. tx_ave = 2.5 ms / tx 2. block_ave = 5.87 ms / block memory-official-repo: 1. tx_ave = 8.85 ms / tx 2. block_ave = 19.68 ms / block lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e) 1. tx_ave = 47.8 ms / tx 2. block_ave = 64.2 ms / block **Note: The following data denotes processing times only (does not include p2p download time) lmdb-optimized processing times (with full pow computation): 1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000). 2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000). 3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000). lmdb-optimized processing times (with per-block-checkpoint) 1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000). berkeley-db optimized processing times (with full pow computation) 1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000). 2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000). berkeley-db optimized processing times (with per-block-checkpoint) 1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 20:09:32 +00:00
#if !defined(STATIC)
#define STATIC
#endif
#if !defined(INLINE)
#define INLINE
#endif
STATIC INLINE void aesb_single_round(const uint8_t *in, uint8_t *out, uint8_t *expandedKey)
{
uint32_t b0[4], b1[4];
const uint32_t *kp = (uint32_t *) expandedKey;
state_in(b0, in);
2014-05-17 23:25:58 +00:00
round(fwd_rnd, b1, b0, kp);
2014-05-17 23:25:58 +00:00
state_out(out, b1);
2014-05-17 23:25:58 +00:00
}
** CHANGES ARE EXPERIMENTAL (FOR TESTING ONLY) Bockchain: 1. Optim: Multi-thread long-hash computation when encountering groups of blocks. 2. Optim: Cache verified txs and return result from cache instead of re-checking whenever possible. 3. Optim: Preload output-keys when encoutering groups of blocks. Sort by amount and global-index before bulk querying database and multi-thread when possible. 4. Optim: Disable double spend check on block verification, double spend is already detected when trying to add blocks. 5. Optim: Multi-thread signature computation whenever possible. 6. Patch: Disable locking (recursive mutex) on called functions from check_tx_inputs which causes slowdowns (only seems to happen on ubuntu/VMs??? Reason: TBD) 7. Optim: Removed looped full-tx hash computation when retrieving transactions from pool (???). 8. Optim: Cache difficulty/timestamps (735 blocks) for next-difficulty calculations so that only 2 db reads per new block is needed when a new block arrives (instead of 1470 reads). Berkeley-DB: 1. Fix: 32-bit data errors causing wrong output global indices and failure to send blocks to peers (etc). 2. Fix: Unable to pop blocks on reorganize due to transaction errors. 3. Patch: Large number of transaction aborts when running multi-threaded bulk queries. 4. Patch: Insufficient locks error when running full sync. 5. Patch: Incorrect db stats when returning from an immediate exit from "pop block" operation. 6. Optim: Add bulk queries to get output global indices. 7. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3) 8. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key 9. Optim: Added thread-safe buffers used when multi-threading bulk queries. 10. Optim: Added support for nosync/write_nosync options for improved performance (*see --db-sync-mode option for details) 11. Mod: Added checkpoint thread and auto-remove-logs option. 12. *Now usable on 32-bit systems like RPI2. LMDB: 1. Optim: Added custom comparison for 256-bit key tables (minor speed-up, TBD: get actual effect) 2. Optim: Modified output_keys table to store public_key+unlock_time+height for single transaction lookup (vs 3) 3. Optim: Used output_keys table retrieve public_keys instead of going through output_amounts->output_txs+output_indices->txs->output:public_key 4. Optim: Added support for sync/writemap options for improved performance (*see --db-sync-mode option for details) 5. Mod: Auto resize to +1GB instead of multiplier x1.5 ETC: 1. Minor optimizations for slow-hash for ARM (RPI2). Incomplete. 2. Fix: 32-bit saturation bug when computing next difficulty on large blocks. [PENDING ISSUES] 1. Berkely db has a very slow "pop-block" operation. This is very noticeable on the RPI2 as it sometimes takes > 10 MINUTES to pop a block during reorganization. This does not happen very often however, most reorgs seem to take a few seconds but it possibly depends on the number of outputs present. TBD. 2. Berkeley db, possible bug "unable to allocate memory". TBD. [NEW OPTIONS] (*Currently all enabled for testing purposes) 1. --fast-block-sync arg=[0:1] (default: 1) a. 0 = Compute long hash per block (may take a while depending on CPU) b. 1 = Skip long-hash and verify blocks based on embedded known good block hashes (faster, minimal CPU dependence) 2. --db-sync-mode arg=[[safe|fast|fastest]:[sync|async]:[nblocks_per_sync]] (default: fastest:async:1000) a. safe = fdatasync/fsync (or equivalent) per stored block. Very slow, but safest option to protect against power-out/crash conditions. b. fast/fastest = Enables asynchronous fdatasync/fsync (or equivalent). Useful for battery operated devices or STABLE systems with UPS and/or systems with battery backed write cache/solid state cache. Fast - Write meta-data but defer data flush. Fastest - Defer meta-data and data flush. Sync - Flush data after nblocks_per_sync and wait. Async - Flush data after nblocks_per_sync but do not wait for the operation to finish. 3. --prep-blocks-threads arg=[n] (default: 4 or system max threads, whichever is lower) Max number of threads to use when computing long-hash in groups. 4. --show-time-stats arg=[0:1] (default: 1) Show benchmark related time stats. 5. --db-auto-remove-logs arg=[0:1] (default: 1) For berkeley-db only. Auto remove logs if enabled. **Note: lmdb and berkeley-db have changes to the tables and are not compatible with official git head version. At the moment, you need a full resync to use this optimized version. [PERFORMANCE COMPARISON] **Some figures are approximations only. Using a baseline machine of an i7-2600K+SSD+(with full pow computation): 1. The optimized lmdb/blockhain core can process blocks up to 585K for ~1.25 hours + download time, so it usually takes 2.5 hours to sync the full chain. 2. The current head with memory can process blocks up to 585K for ~4.2 hours + download time, so it usually takes 5.5 hours to sync the full chain. 3. The current head with lmdb can process blocks up to 585K for ~32 hours + download time and usually takes 36 hours to sync the full chain. Averate procesing times (with full pow computation): lmdb-optimized: 1. tx_ave = 2.5 ms / tx 2. block_ave = 5.87 ms / block memory-official-repo: 1. tx_ave = 8.85 ms / tx 2. block_ave = 19.68 ms / block lmdb-official-repo (0f4a036437fd41a5498ee5e74e2422ea6177aa3e) 1. tx_ave = 47.8 ms / tx 2. block_ave = 64.2 ms / block **Note: The following data denotes processing times only (does not include p2p download time) lmdb-optimized processing times (with full pow computation): 1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.25 hours processing time (--db-sync-mode=fastest:async:1000). 2. Laptop, Dual-core / 4-threads U4200 (3Mb) - 4.90 hours processing time (--db-sync-mode=fastest:async:1000). 3. Embedded, Quad-core / 4-threads Z3735F (2x1Mb) - 12.0 hours processing time (--db-sync-mode=fastest:async:1000). lmdb-optimized processing times (with per-block-checkpoint) 1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 10 minutes processing time (--db-sync-mode=fastest:async:1000). berkeley-db optimized processing times (with full pow computation) 1. Desktop, Quad-core / 8-threads 2600k (8Mb) - 1.8 hours processing time (--db-sync-mode=fastest:async:1000). 2. RPI2. Improved from estimated 3 months(???) into 2.5 days (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000). berkeley-db optimized processing times (with per-block-checkpoint) 1. RPI2. 12-15 hours (*Need 2AMP supply + Clock:1Ghz + [usb+ssd] to achieve this speed) (--db-sync-mode=fastest:async:1000).
2015-07-10 20:09:32 +00:00
STATIC INLINE void aesb_pseudo_round(const uint8_t *in, uint8_t *out, uint8_t *expandedKey)
{
uint32_t b0[4], b1[4];
const uint32_t *kp = (uint32_t *) expandedKey;
state_in(b0, in);
round(fwd_rnd, b1, b0, kp);
round(fwd_rnd, b0, b1, kp + 1 * N_COLS);
round(fwd_rnd, b1, b0, kp + 2 * N_COLS);
round(fwd_rnd, b0, b1, kp + 3 * N_COLS);
round(fwd_rnd, b1, b0, kp + 4 * N_COLS);
round(fwd_rnd, b0, b1, kp + 5 * N_COLS);
round(fwd_rnd, b1, b0, kp + 6 * N_COLS);
round(fwd_rnd, b0, b1, kp + 7 * N_COLS);
round(fwd_rnd, b1, b0, kp + 8 * N_COLS);
round(fwd_rnd, b0, b1, kp + 9 * N_COLS);
state_out(out, b0);
2014-05-17 23:25:58 +00:00
}
#if defined(__cplusplus)
}
#endif