574 lines
17 KiB
C
574 lines
17 KiB
C
/*
|
|
* Synopsis : Error Correction Codes (ECC) Algorithms.
|
|
*
|
|
* Copyright (c) 2008 STMicroelectronics Limited. All right reserved.
|
|
*
|
|
* See stm_nand_ecc.h for a description of this module.
|
|
*
|
|
* May be copied or modified under the terms of the GNU General Public
|
|
* License Version 2.0 only. See linux/COPYING for more information.
|
|
*
|
|
* Changelog:
|
|
*
|
|
* 2009-12-14 Angus Clark <angus.clark@st.com>
|
|
* - Added S/W ECC for LP1617 bits, as required by AFM controller
|
|
* 2009-02-25 Angus Clark <angus.clark@st.com>
|
|
* - Renamed, formatted, and edited for linux compatibility
|
|
* - Added clause in stm_ecc_correct() to hanlde inconsistency between
|
|
* data and expected ecc for freshly erased page.
|
|
*
|
|
*/
|
|
|
|
#include <linux/types.h>
|
|
#include <linux/kernel.h>
|
|
#include <linux/module.h>
|
|
#include "stm_nand_ecc.h"
|
|
|
|
static const uint8_t byte_parity_table[] = /* Parity look up table */
|
|
{
|
|
0x00, 0x2B, 0x2D, 0x06, 0x33, 0x18, 0x1E, 0x35,
|
|
0x35, 0x1E, 0x18, 0x33, 0x06, 0x2D, 0x2B, 0x00,
|
|
0x4B, 0x60, 0x66, 0x4D, 0x78, 0x53, 0x55, 0x7E,
|
|
0x7E, 0x55, 0x53, 0x78, 0x4D, 0x66, 0x60, 0x4B,
|
|
0x4D, 0x66, 0x60, 0x4B, 0x7E, 0x55, 0x53, 0x78,
|
|
0x78, 0x53, 0x55, 0x7E, 0x4B, 0x60, 0x66, 0x4D,
|
|
0x06, 0x2D, 0x2B, 0x00, 0x35, 0x1E, 0x18, 0x33,
|
|
0x33, 0x18, 0x1E, 0x35, 0x00, 0x2B, 0x2D, 0x06,
|
|
0x53, 0x78, 0x7E, 0x55, 0x60, 0x4B, 0x4D, 0x66,
|
|
0x66, 0x4D, 0x4B, 0x60, 0x55, 0x7E, 0x78, 0x53,
|
|
0x18, 0x33, 0x35, 0x1E, 0x2B, 0x00, 0x06, 0x2D,
|
|
0x2D, 0x06, 0x00, 0x2B, 0x1E, 0x35, 0x33, 0x18,
|
|
0x1E, 0x35, 0x33, 0x18, 0x2D, 0x06, 0x00, 0x2B,
|
|
0x2B, 0x00, 0x06, 0x2D, 0x18, 0x33, 0x35, 0x1E,
|
|
0x55, 0x7E, 0x78, 0x53, 0x66, 0x4D, 0x4B, 0x60,
|
|
0x60, 0x4B, 0x4D, 0x66, 0x53, 0x78, 0x7E, 0x55,
|
|
0x55, 0x7E, 0x78, 0x53, 0x66, 0x4D, 0x4B, 0x60,
|
|
0x60, 0x4B, 0x4D, 0x66, 0x53, 0x78, 0x7E, 0x55,
|
|
0x1E, 0x35, 0x33, 0x18, 0x2D, 0x06, 0x00, 0x2B,
|
|
0x2B, 0x00, 0x06, 0x2D, 0x18, 0x33, 0x35, 0x1E,
|
|
0x18, 0x33, 0x35, 0x1E, 0x2B, 0x00, 0x06, 0x2D,
|
|
0x2D, 0x06, 0x00, 0x2B, 0x1E, 0x35, 0x33, 0x18,
|
|
0x53, 0x78, 0x7E, 0x55, 0x60, 0x4B, 0x4D, 0x66,
|
|
0x66, 0x4D, 0x4B, 0x60, 0x55, 0x7E, 0x78, 0x53,
|
|
0x06, 0x2D, 0x2B, 0x00, 0x35, 0x1E, 0x18, 0x33,
|
|
0x33, 0x18, 0x1E, 0x35, 0x00, 0x2B, 0x2D, 0x06,
|
|
0x4D, 0x66, 0x60, 0x4B, 0x7E, 0x55, 0x53, 0x78,
|
|
0x78, 0x53, 0x55, 0x7E, 0x4B, 0x60, 0x66, 0x4D,
|
|
0x4B, 0x60, 0x66, 0x4D, 0x78, 0x53, 0x55, 0x7E,
|
|
0x7E, 0x55, 0x53, 0x78, 0x4D, 0x66, 0x60, 0x4B,
|
|
0x00, 0x2B, 0x2D, 0x06, 0x33, 0x18, 0x1E, 0x35,
|
|
0x35, 0x1E, 0x18, 0x33, 0x06, 0x2D, 0x2B, 0x00
|
|
};
|
|
|
|
static const uint8_t bit_count_table[] = /* Parity look up table */
|
|
{
|
|
0, 1, 1, 2, 1, 2, 2, 3,
|
|
1, 2, 2, 3, 2, 3, 3, 4,
|
|
1, 2, 2, 3, 2, 3, 3, 4,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
1, 2, 2, 3, 2, 3, 3, 4,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
1, 2, 2, 3, 2, 3, 3, 4,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
4, 5, 5, 6, 5, 6, 6, 7,
|
|
1, 2, 2, 3, 2, 3, 3, 4,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
4, 5, 5, 6, 5, 6, 6, 7,
|
|
2, 3, 3, 4, 3, 4, 4, 5,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
4, 5, 5, 6, 5, 6, 6, 7,
|
|
3, 4, 4, 5, 4, 5, 5, 6,
|
|
4, 5, 5, 6, 5, 6, 6, 7,
|
|
4, 5, 5, 6, 5, 6, 6, 7,
|
|
5, 6, 6, 7, 6, 7, 7, 8
|
|
};
|
|
|
|
/******************************************************************************/
|
|
#define COL_LOOP_STEP(c__f, c__e, c__o, c__t) \
|
|
do { \
|
|
c__o ^= (c__f ? c__t : 1); \
|
|
c__e ^= (c__f ? 1 : c__t); \
|
|
} while (0)
|
|
|
|
/* Generate 3 byte ECC code for ecc_size block p_data.
|
|
"p_data" is a pointer to the date and must be 4-byte aligned.
|
|
"size" gives length of "p_data" - one of enum ecc_size.
|
|
*/
|
|
void stm_ecc_gen(const uint8_t *p_data, uint8_t *ecc, enum ecc_size size)
|
|
{
|
|
uint32_t *p_data_long = (uint32_t *)p_data;
|
|
uint32_t parity_bits[18]; /* maximum number */
|
|
uint32_t reg32;
|
|
uint32_t temp;
|
|
uint32_t int_cnt;
|
|
uint32_t bit_cnt;
|
|
|
|
unsigned int num_parity_bits;
|
|
|
|
uint8_t *p_byt;
|
|
uint8_t byte_reg;
|
|
uint8_t byte_a;
|
|
uint8_t byte_b;
|
|
uint8_t byte_c;
|
|
uint8_t byte_d;
|
|
|
|
ecc[0] = 0;
|
|
ecc[1] = 0;
|
|
ecc[2] = 0;
|
|
|
|
switch (size) {
|
|
case ECC_128:
|
|
num_parity_bits = 14;
|
|
break;
|
|
case ECC_256:
|
|
num_parity_bits = 16;
|
|
break;
|
|
case ECC_512:
|
|
num_parity_bits = 18;
|
|
break;
|
|
default:
|
|
printk(KERN_ERR "Internal error in ecc_gen: unknown format\n");
|
|
BUG();
|
|
return;
|
|
}
|
|
|
|
/* Initialize variables */
|
|
byte_reg = 0;
|
|
reg32 = 0;
|
|
|
|
for (bit_cnt = 0; bit_cnt < num_parity_bits; bit_cnt ++)
|
|
parity_bits[bit_cnt] = 0;
|
|
|
|
/* Build up column parity */
|
|
for (int_cnt = 0; int_cnt < size/sizeof(uint32_t); int_cnt++) {
|
|
temp = p_data_long[int_cnt];
|
|
|
|
switch (size) {
|
|
case ECC_512:
|
|
COL_LOOP_STEP((int_cnt & 0x40), parity_bits[16],
|
|
parity_bits[17], temp);
|
|
/* fall through */
|
|
case ECC_256:
|
|
COL_LOOP_STEP((int_cnt & 0x20), parity_bits[14],
|
|
parity_bits[15], temp);
|
|
/* fall through */
|
|
case ECC_128:
|
|
COL_LOOP_STEP((int_cnt & 0x01), parity_bits[4],
|
|
parity_bits[5], temp);
|
|
COL_LOOP_STEP((int_cnt & 0x02), parity_bits[6],
|
|
parity_bits[7], temp);
|
|
COL_LOOP_STEP((int_cnt & 0x04), parity_bits[8],
|
|
parity_bits[9], temp);
|
|
COL_LOOP_STEP((int_cnt & 0x08), parity_bits[10],
|
|
parity_bits[11], temp);
|
|
COL_LOOP_STEP((int_cnt & 0x10), parity_bits[12],
|
|
parity_bits[13], temp);
|
|
}
|
|
}
|
|
|
|
reg32 = parity_bits[12] ^ parity_bits[13];
|
|
|
|
p_byt = (uint8_t *)®32;
|
|
#if __LITTLE_ENDIAN__
|
|
byte_a = p_byt[3];
|
|
byte_b = p_byt[2];
|
|
byte_c = p_byt[1];
|
|
byte_d = p_byt[0];
|
|
#else
|
|
byte_a = p_byt[0];
|
|
byte_b = p_byt[1];
|
|
byte_c = p_byt[2];
|
|
byte_d = p_byt[3];
|
|
#endif
|
|
|
|
byte_reg = byte_a ^ byte_b ^ byte_c ^ byte_d;
|
|
byte_reg = byte_parity_table[byte_reg] >> 1;
|
|
|
|
/* Create line parity */
|
|
parity_bits[0] = byte_d ^ byte_b;
|
|
parity_bits[1] = byte_c ^ byte_a;
|
|
parity_bits[2] = byte_d ^ byte_c;
|
|
parity_bits[3] = byte_b ^ byte_a;
|
|
|
|
for (bit_cnt = 4; bit_cnt < num_parity_bits; bit_cnt++) {
|
|
p_byt = (uint8_t *)(parity_bits + bit_cnt);
|
|
/* NB Only LS Byte of parity_bits used from now on */
|
|
p_byt[0] ^= (p_byt[1] ^ p_byt[2] ^ p_byt[3]);
|
|
}
|
|
|
|
/* Calculate final ECC code */
|
|
for (bit_cnt = 0; bit_cnt < 8; bit_cnt++)
|
|
ecc[0] |=
|
|
(byte_parity_table[(uint8_t)parity_bits[bit_cnt]] &
|
|
0x01) << bit_cnt;
|
|
for (; bit_cnt < 16 && bit_cnt < num_parity_bits; bit_cnt++)
|
|
ecc[1] |=
|
|
(byte_parity_table[(uint8_t)parity_bits[bit_cnt]] &
|
|
0x01) << (bit_cnt - 8);
|
|
for (; bit_cnt < num_parity_bits; bit_cnt++)
|
|
ecc[2] |=
|
|
(byte_parity_table[(uint8_t)parity_bits[bit_cnt]] &
|
|
0x01) << (bit_cnt - 16);
|
|
|
|
ecc[2] = (uint8_t)(byte_reg << 2) | (ecc[2] & 0x03);
|
|
}
|
|
EXPORT_SYMBOL_GPL(stm_ecc_gen);
|
|
|
|
/******************************************************************************/
|
|
/* Detect and correct a 1 bit error in a 128, 256 or 512 byte block.
|
|
"p_data" is a pointer to the data.
|
|
"old_ecc" is the proper ECC for the data.
|
|
"new_ecc" is the ECC generated from the (possibly) corrupted data.
|
|
The size of the block is given in "size".
|
|
|
|
Returns whether the data needed correcting, or was not correctable.
|
|
If the result code is E_D1_CHK, then the data will have been modified.
|
|
*/
|
|
enum ecc_check stm_ecc_correct(uint8_t *p_data,
|
|
uint8_t *old_ecc,
|
|
uint8_t *new_ecc,
|
|
enum ecc_size size)
|
|
{
|
|
uint8_t bit_cnt02;
|
|
uint8_t bit_addr02;
|
|
unsigned int byte_addr02;
|
|
|
|
uint8_t ecc_xor[3];
|
|
|
|
uint8_t error_bit_count;
|
|
|
|
switch (size) {
|
|
case ECC_128:
|
|
error_bit_count = 10;
|
|
break;
|
|
case ECC_256:
|
|
error_bit_count = 11;
|
|
break;
|
|
case ECC_512:
|
|
error_bit_count = 12;
|
|
break;
|
|
default:
|
|
printk(KERN_ERR "Internal error in ecc_gen: unknown format\n");
|
|
BUG();
|
|
return E_UN_CHK;
|
|
}
|
|
|
|
/* Basic Error Detection phase */
|
|
ecc_xor[0] = new_ecc[0] ^ old_ecc[0];
|
|
ecc_xor[1] = new_ecc[1] ^ old_ecc[1];
|
|
ecc_xor[2] = new_ecc[2] ^ old_ecc[2];
|
|
|
|
if ((ecc_xor[0] | ecc_xor[1] | ecc_xor[2]) == 0) {
|
|
return E_NO_CHK; /* No errors */
|
|
}
|
|
/* If we get here then there were errors */
|
|
if (ecc_xor[0] == 0xff &&
|
|
ecc_xor[1] == 0xff &&
|
|
ecc_xor[2] == 0xff) {
|
|
/* Probably a freshly erased page: All 0xff's, including OOB,
|
|
but expecting 0x00, 0x00, 0x00 for ECC data
|
|
*/
|
|
return E_NO_CHK;
|
|
}
|
|
if (size == ECC_512) {
|
|
/* 512-byte error correction requires a little more than 128 or
|
|
256. If there is a correctable error then the xor will have
|
|
12 bits set, but there can also be 12 bits set in some
|
|
uncorrectable errors. This can be solved by xoring the odd
|
|
and even numbered bits.
|
|
|
|
0xAA = 10101010
|
|
0x55 = 01010101
|
|
*/
|
|
bit_cnt02 = bit_count_table[((ecc_xor[0] & 0xAA) >> 1) ^
|
|
(ecc_xor[0] & 0x55)];
|
|
bit_cnt02 += bit_count_table[((ecc_xor[1] & 0xAA) >> 1) ^
|
|
(ecc_xor[1] & 0x55)];
|
|
bit_cnt02 += bit_count_table[((ecc_xor[2] & 0xAA) >> 1) ^
|
|
(ecc_xor[2] & 0x55)];
|
|
} else {
|
|
/* Counts the number of bits set in ecc code */
|
|
bit_cnt02 = bit_count_table[ecc_xor[0]];
|
|
bit_cnt02 += bit_count_table[ecc_xor[1]];
|
|
bit_cnt02 += bit_count_table[ecc_xor[2]];
|
|
}
|
|
|
|
if (bit_cnt02 == error_bit_count) {
|
|
/* Set the bit address */
|
|
bit_addr02 = ((ecc_xor[2] >> 3) & 0x01) |
|
|
((ecc_xor[2] >> 4) & 0x02) |
|
|
((ecc_xor[2] >> 5) & 0x04);
|
|
|
|
/* Evaluate 2 LS bits of address */
|
|
byte_addr02 = ((ecc_xor[0] >> 1) & 0x01) |
|
|
((ecc_xor[0] >> 2) & 0x02);
|
|
|
|
/* Add in remaining bits of address */
|
|
switch (size) {
|
|
case ECC_512:
|
|
byte_addr02 |= (((unsigned int)ecc_xor[2]) << 7)
|
|
& 0x100;
|
|
/* Fall through */
|
|
case ECC_256:
|
|
byte_addr02 |= (ecc_xor[1] & 0x80);
|
|
/* Fall through */
|
|
case ECC_128:
|
|
byte_addr02 |= ((ecc_xor[0] >> 3) & 0x04) |
|
|
((ecc_xor[0] >> 4) & 0x08) |
|
|
((ecc_xor[1] << 3) & 0x10) |
|
|
((ecc_xor[1] << 2) & 0x20) |
|
|
((ecc_xor[1] << 1) & 0x40);
|
|
}
|
|
printk(KERN_DEBUG "%s: correcting bit [bit %d, byte %d, ",
|
|
__func__, bit_addr02, byte_addr02);
|
|
|
|
/* Correct bit error in the data */
|
|
p_data[byte_addr02] ^= (0x01 << bit_addr02);
|
|
|
|
/* NB p_old_code is okay, p_new_code is corrupt */
|
|
return E_D1_CHK; /* Data had 1-bit error (now corrected) */
|
|
} else {
|
|
if (bit_cnt02 == 1) {
|
|
printk(KERN_DEBUG "%s: ignoring error in ECC, "
|
|
"data ok: [", __func__);
|
|
return E_C1_CHK; /* ECC has 1-bit error, data is ok */
|
|
} else {
|
|
printk(KERN_ERR "%s: uncorrectable error: [", __func__);
|
|
|
|
return E_UN_CHK; /* Uncorrectable Error */
|
|
}
|
|
}
|
|
}
|
|
EXPORT_SYMBOL_GPL(stm_ecc_correct);
|
|
|
|
/*****************************************************************************
|
|
* The STMicroelectronics ECC engine requires S/W generated LP16 and LP17 bits.
|
|
* The following code segment is based on nand_ecc.c:nand_calculate_ecc(),
|
|
* stripped-down to calculate only LP16 and LP17 parity bits. Please see
|
|
* nand_ecc.c for further details.
|
|
*****************************************************************************/
|
|
|
|
/*
|
|
* invparity is a 256 byte table that contains the odd parity
|
|
* for each byte. So if the number of bits in a byte is even,
|
|
* the array element is 1, and when the number of bits is odd
|
|
* the array eleemnt is 0.
|
|
*/
|
|
static const char invparity[256] = {
|
|
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1,
|
|
0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0,
|
|
0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0,
|
|
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1,
|
|
0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0,
|
|
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1,
|
|
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1,
|
|
0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0,
|
|
0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0,
|
|
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1,
|
|
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1,
|
|
0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0,
|
|
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1,
|
|
0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0,
|
|
0, 1, 1, 0, 1, 0, 0, 1, 1, 0, 0, 1, 0, 1, 1, 0,
|
|
1, 0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 1
|
|
};
|
|
|
|
/*
|
|
* bitsperbyte contains the number of bits per byte
|
|
* this is only used for testing and repairing parity
|
|
* (a precalculated value slightly improves performance)
|
|
*/
|
|
static const char bitsperbyte[256] = {
|
|
0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4,
|
|
1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
|
|
1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
|
|
2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
|
|
1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
|
|
2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
|
|
2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
|
|
3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
|
|
1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5,
|
|
2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
|
|
2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
|
|
3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
|
|
2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6,
|
|
3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
|
|
3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7,
|
|
4, 5, 5, 6, 5, 6, 6, 7, 5, 6, 6, 7, 6, 7, 7, 8,
|
|
};
|
|
|
|
/*
|
|
* addressbits is a lookup table to filter out the bits from the xor-ed
|
|
* ecc data that identify the faulty location.
|
|
* this is only used for repairing parity
|
|
* see the comments in nand_correct_data for more details
|
|
*/
|
|
static const char addressbits[256] = {
|
|
0x00, 0x00, 0x01, 0x01, 0x00, 0x00, 0x01, 0x01,
|
|
0x02, 0x02, 0x03, 0x03, 0x02, 0x02, 0x03, 0x03,
|
|
0x00, 0x00, 0x01, 0x01, 0x00, 0x00, 0x01, 0x01,
|
|
0x02, 0x02, 0x03, 0x03, 0x02, 0x02, 0x03, 0x03,
|
|
0x04, 0x04, 0x05, 0x05, 0x04, 0x04, 0x05, 0x05,
|
|
0x06, 0x06, 0x07, 0x07, 0x06, 0x06, 0x07, 0x07,
|
|
0x04, 0x04, 0x05, 0x05, 0x04, 0x04, 0x05, 0x05,
|
|
0x06, 0x06, 0x07, 0x07, 0x06, 0x06, 0x07, 0x07,
|
|
0x00, 0x00, 0x01, 0x01, 0x00, 0x00, 0x01, 0x01,
|
|
0x02, 0x02, 0x03, 0x03, 0x02, 0x02, 0x03, 0x03,
|
|
0x00, 0x00, 0x01, 0x01, 0x00, 0x00, 0x01, 0x01,
|
|
0x02, 0x02, 0x03, 0x03, 0x02, 0x02, 0x03, 0x03,
|
|
0x04, 0x04, 0x05, 0x05, 0x04, 0x04, 0x05, 0x05,
|
|
0x06, 0x06, 0x07, 0x07, 0x06, 0x06, 0x07, 0x07,
|
|
0x04, 0x04, 0x05, 0x05, 0x04, 0x04, 0x05, 0x05,
|
|
0x06, 0x06, 0x07, 0x07, 0x06, 0x06, 0x07, 0x07,
|
|
0x08, 0x08, 0x09, 0x09, 0x08, 0x08, 0x09, 0x09,
|
|
0x0a, 0x0a, 0x0b, 0x0b, 0x0a, 0x0a, 0x0b, 0x0b,
|
|
0x08, 0x08, 0x09, 0x09, 0x08, 0x08, 0x09, 0x09,
|
|
0x0a, 0x0a, 0x0b, 0x0b, 0x0a, 0x0a, 0x0b, 0x0b,
|
|
0x0c, 0x0c, 0x0d, 0x0d, 0x0c, 0x0c, 0x0d, 0x0d,
|
|
0x0e, 0x0e, 0x0f, 0x0f, 0x0e, 0x0e, 0x0f, 0x0f,
|
|
0x0c, 0x0c, 0x0d, 0x0d, 0x0c, 0x0c, 0x0d, 0x0d,
|
|
0x0e, 0x0e, 0x0f, 0x0f, 0x0e, 0x0e, 0x0f, 0x0f,
|
|
0x08, 0x08, 0x09, 0x09, 0x08, 0x08, 0x09, 0x09,
|
|
0x0a, 0x0a, 0x0b, 0x0b, 0x0a, 0x0a, 0x0b, 0x0b,
|
|
0x08, 0x08, 0x09, 0x09, 0x08, 0x08, 0x09, 0x09,
|
|
0x0a, 0x0a, 0x0b, 0x0b, 0x0a, 0x0a, 0x0b, 0x0b,
|
|
0x0c, 0x0c, 0x0d, 0x0d, 0x0c, 0x0c, 0x0d, 0x0d,
|
|
0x0e, 0x0e, 0x0f, 0x0f, 0x0e, 0x0e, 0x0f, 0x0f,
|
|
0x0c, 0x0c, 0x0d, 0x0d, 0x0c, 0x0c, 0x0d, 0x0d,
|
|
0x0e, 0x0e, 0x0f, 0x0f, 0x0e, 0x0e, 0x0f, 0x0f
|
|
};
|
|
|
|
/* The STMicroelectronics ECC engine requires S/W generated LP16 and LP17 bits.
|
|
* The following code segment is based on 'nand_calculate_ecc', stripped_down to
|
|
* generate only LP16 and LP17 parity bits.
|
|
*
|
|
* Angus Clark <angus.clark@st.com>
|
|
*/
|
|
unsigned char stm_afm_lp1617(const unsigned char *buf)
|
|
{
|
|
int i;
|
|
const uint32_t *bp = (uint32_t *)buf;
|
|
const uint32_t eccsize_mult = 2;
|
|
uint32_t cur; /* current value in buffer */
|
|
uint32_t rp16;
|
|
uint32_t uninitialized_var(rp17); /* to make compiler happy */
|
|
uint32_t par; /* the cumulative parity for all data */
|
|
uint32_t tmppar; /* the cumulative parity for this iteration;
|
|
for rp12, rp14 and rp16 at the end of the
|
|
loop */
|
|
par = 0;
|
|
rp16 = 0;
|
|
|
|
/*
|
|
* The loop is unrolled a number of times;
|
|
* This avoids if statements to decide on which rp value to update
|
|
* Also we process the data by longwords.
|
|
* Note: passing unaligned data might give a performance penalty.
|
|
* It is assumed that the buffers are aligned.
|
|
* tmppar is the cumulative sum of this iteration.
|
|
* needed for calculating rp12, rp14, rp16 and par
|
|
* also used as a performance improvement for rp6, rp8 and rp10
|
|
*/
|
|
for (i = 0; i < eccsize_mult << 2; i++) {
|
|
cur = *bp++;
|
|
tmppar = cur;
|
|
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
|
|
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
|
|
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
cur = *bp++;
|
|
tmppar ^= cur;
|
|
|
|
par ^= tmppar;
|
|
|
|
if ((i & 0x4) == 0)
|
|
rp16 ^= tmppar;
|
|
}
|
|
|
|
/*
|
|
* handle the fact that we use longword operations
|
|
* we'll bring rp4..rp14..rp16 back to single byte entities by
|
|
* shifting and xoring first fold the upper and lower 16 bits,
|
|
* then the upper and lower 8 bits.
|
|
*/
|
|
rp16 ^= (rp16 >> 16);
|
|
rp16 ^= (rp16 >> 8);
|
|
rp16 &= 0xff;
|
|
|
|
/* reduce par to 16 bits then calculate rp1 and rp0 */
|
|
par ^= (par >> 16);
|
|
|
|
/* finally reduce par to 8 bits */
|
|
par ^= (par >> 8);
|
|
par &= 0xff;
|
|
|
|
/*
|
|
* and calculate rp5..rp15..rp17
|
|
* note that par = rp4 ^ rp5 and due to the commutative property
|
|
* of the ^ operator we can say:
|
|
* rp5 = (par ^ rp4);
|
|
* The & 0xff seems superfluous, but benchmarking learned that
|
|
* leaving it out gives slightly worse results. No idea why, probably
|
|
* it has to do with the way the pipeline in pentium is organized.
|
|
*/
|
|
rp17 = (par ^ rp16) & 0xff;
|
|
|
|
return ~((invparity[rp17] << 1) | (invparity[rp16] << 0)) & 0x3;
|
|
|
|
}
|
|
EXPORT_SYMBOL_GPL(stm_afm_lp1617);
|
|
|
|
/*****************************************************************************/
|
|
|
|
|
|
MODULE_LICENSE("GPL");
|
|
MODULE_AUTHOR("Angus Clark");
|
|
MODULE_DESCRIPTION("STMicroelectronics NAND ECC support");
|
|
|