OpenCloudOS-Kernel/drivers/mtd/nand/nand_bbt.c

1453 lines
40 KiB
C
Raw Normal View History

/*
* Overview:
* Bad block table support for the NAND driver
*
* Copyright © 2004 Thomas Gleixner (tglx@linutronix.de)
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
* Description:
*
* When nand_scan_bbt is called, then it tries to find the bad block table
* depending on the options in the BBT descriptor(s). If no flash based BBT
* (NAND_BBT_USE_FLASH) is specified then the device is scanned for factory
* marked good / bad blocks. This information is used to create a memory BBT.
* Once a new bad block is discovered then the "factory" information is updated
* on the device.
* If a flash based BBT is specified then the function first tries to find the
* BBT on flash. If a BBT is found then the contents are read and the memory
* based BBT is created. If a mirrored BBT is selected then the mirror is
* searched too and the versions are compared. If the mirror has a greater
* version number, then the mirror BBT is used to build the memory based BBT.
* If the tables are not versioned, then we "or" the bad block information.
* If one of the BBTs is out of date or does not exist it is (re)created.
* If no BBT exists at all then the device is scanned for factory marked
* good / bad blocks and the bad block tables are created.
*
* For manufacturer created BBTs like the one found on M-SYS DOC devices
* the BBT is searched and read but never created
*
* The auto generated bad block table is located in the last good blocks
* of the device. The table is mirrored, so it can be updated eventually.
* The table is marked in the OOB area with an ident pattern and a version
* number which indicates which of both tables is more up to date. If the NAND
* controller needs the complete OOB area for the ECC information then the
* option NAND_BBT_NO_OOB should be used (along with NAND_BBT_USE_FLASH, of
* course): it moves the ident pattern and the version byte into the data area
* and the OOB area will remain untouched.
*
* The table uses 2 bits per block
* 11b: block is good
* 00b: block is factory marked bad
* 01b, 10b: block is marked bad due to wear
*
* The memory bad block table uses the following scheme:
* 00b: block is good
* 01b: block is marked bad due to wear
* 10b: block is reserved (to protect the bbt area)
* 11b: block is factory marked bad
*
* Multichip devices like DOC store the bad block info per floor.
*
* Following assumptions are made:
* - bbts start at a page boundary, if autolocated on a block boundary
* - the space necessary for a bbt in FLASH does not exceed a block boundary
*
*/
#include <linux/slab.h>
#include <linux/types.h>
#include <linux/mtd/mtd.h>
#include <linux/mtd/bbm.h>
#include <linux/mtd/rawnand.h>
#include <linux/bitops.h>
#include <linux/delay.h>
#include <linux/vmalloc.h>
#include <linux/export.h>
#include <linux/string.h>
#define BBT_BLOCK_GOOD 0x00
#define BBT_BLOCK_WORN 0x01
#define BBT_BLOCK_RESERVED 0x02
#define BBT_BLOCK_FACTORY_BAD 0x03
#define BBT_ENTRY_MASK 0x03
#define BBT_ENTRY_SHIFT 2
static int nand_update_bbt(struct mtd_info *mtd, loff_t offs);
static inline uint8_t bbt_get_entry(struct nand_chip *chip, int block)
{
uint8_t entry = chip->bbt[block >> BBT_ENTRY_SHIFT];
entry >>= (block & BBT_ENTRY_MASK) * 2;
return entry & BBT_ENTRY_MASK;
}
static inline void bbt_mark_entry(struct nand_chip *chip, int block,
uint8_t mark)
{
uint8_t msk = (mark & BBT_ENTRY_MASK) << ((block & BBT_ENTRY_MASK) * 2);
chip->bbt[block >> BBT_ENTRY_SHIFT] |= msk;
}
static int check_pattern_no_oob(uint8_t *buf, struct nand_bbt_descr *td)
{
if (memcmp(buf, td->pattern, td->len))
return -1;
return 0;
}
/**
* check_pattern - [GENERIC] check if a pattern is in the buffer
* @buf: the buffer to search
* @len: the length of buffer to search
* @paglen: the pagelength
* @td: search pattern descriptor
*
* Check for a pattern at the given place. Used to search bad block tables and
* good / bad block identifiers.
*/
static int check_pattern(uint8_t *buf, int len, int paglen, struct nand_bbt_descr *td)
{
if (td->options & NAND_BBT_NO_OOB)
return check_pattern_no_oob(buf, td);
/* Compare the pattern */
if (memcmp(buf + paglen + td->offs, td->pattern, td->len))
return -1;
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
return 0;
}
/**
* check_short_pattern - [GENERIC] check if a pattern is in the buffer
* @buf: the buffer to search
* @td: search pattern descriptor
*
* Check for a pattern at the given place. Used to search bad block tables and
* good / bad block identifiers. Same as check_pattern, but no optional empty
* check.
*/
static int check_short_pattern(uint8_t *buf, struct nand_bbt_descr *td)
{
/* Compare the pattern */
if (memcmp(buf + td->offs, td->pattern, td->len))
return -1;
return 0;
}
/**
* add_marker_len - compute the length of the marker in data area
* @td: BBT descriptor used for computation
*
* The length will be 0 if the marker is located in OOB area.
*/
static u32 add_marker_len(struct nand_bbt_descr *td)
{
u32 len;
if (!(td->options & NAND_BBT_NO_OOB))
return 0;
len = td->len;
if (td->options & NAND_BBT_VERSION)
len++;
return len;
}
/**
* read_bbt - [GENERIC] Read the bad block table starting from page
* @mtd: MTD device structure
* @buf: temporary buffer
* @page: the starting page
* @num: the number of bbt descriptors to read
* @td: the bbt describtion table
* @offs: block number offset in the table
*
* Read the bad block table starting from page.
*/
static int read_bbt(struct mtd_info *mtd, uint8_t *buf, int page, int num,
struct nand_bbt_descr *td, int offs)
{
int res, ret = 0, i, j, act = 0;
struct nand_chip *this = mtd_to_nand(mtd);
size_t retlen, len, totlen;
loff_t from;
int bits = td->options & NAND_BBT_NRBITS_MSK;
uint8_t msk = (uint8_t)((1 << bits) - 1);
u32 marker_len;
int reserved_block_code = td->reserved_block_code;
totlen = (num * bits) >> 3;
marker_len = add_marker_len(td);
from = ((loff_t)page) << this->page_shift;
while (totlen) {
len = min(totlen, (size_t)(1 << this->bbt_erase_shift));
if (marker_len) {
/*
* In case the BBT marker is not in the OOB area it
* will be just in the first page.
*/
len -= marker_len;
from += marker_len;
marker_len = 0;
}
res = mtd_read(mtd, from, len, &retlen, buf);
if (res < 0) {
if (mtd_is_eccerr(res)) {
pr_info("nand_bbt: ECC error in BBT at 0x%012llx\n",
from & ~mtd->writesize);
return res;
} else if (mtd_is_bitflip(res)) {
pr_info("nand_bbt: corrected error in BBT at 0x%012llx\n",
from & ~mtd->writesize);
ret = res;
} else {
pr_info("nand_bbt: error reading BBT\n");
return res;
}
}
/* Analyse data */
for (i = 0; i < len; i++) {
uint8_t dat = buf[i];
for (j = 0; j < 8; j += bits, act++) {
uint8_t tmp = (dat >> j) & msk;
if (tmp == msk)
continue;
if (reserved_block_code && (tmp == reserved_block_code)) {
pr_info("nand_read_bbt: reserved block at 0x%012llx\n",
(loff_t)(offs + act) <<
this->bbt_erase_shift);
bbt_mark_entry(this, offs + act,
BBT_BLOCK_RESERVED);
mtd->ecc_stats.bbtblocks++;
continue;
}
/*
* Leave it for now, if it's matured we can
* move this message to pr_debug.
*/
pr_info("nand_read_bbt: bad block at 0x%012llx\n",
(loff_t)(offs + act) <<
this->bbt_erase_shift);
/* Factory marked bad or worn out? */
if (tmp == 0)
bbt_mark_entry(this, offs + act,
BBT_BLOCK_FACTORY_BAD);
else
bbt_mark_entry(this, offs + act,
BBT_BLOCK_WORN);
mtd->ecc_stats.badblocks++;
}
}
totlen -= len;
from += len;
}
return ret;
}
/**
* read_abs_bbt - [GENERIC] Read the bad block table starting at a given page
* @mtd: MTD device structure
* @buf: temporary buffer
* @td: descriptor for the bad block table
* @chip: read the table for a specific chip, -1 read all chips; applies only if
* NAND_BBT_PERCHIP option is set
*
* Read the bad block table for all chips starting at a given page. We assume
* that the bbt bits are in consecutive order.
*/
static int read_abs_bbt(struct mtd_info *mtd, uint8_t *buf, struct nand_bbt_descr *td, int chip)
{
struct nand_chip *this = mtd_to_nand(mtd);
int res = 0, i;
if (td->options & NAND_BBT_PERCHIP) {
int offs = 0;
for (i = 0; i < this->numchips; i++) {
if (chip == -1 || chip == i)
res = read_bbt(mtd, buf, td->pages[i],
this->chipsize >> this->bbt_erase_shift,
td, offs);
if (res)
return res;
offs += this->chipsize >> this->bbt_erase_shift;
}
} else {
res = read_bbt(mtd, buf, td->pages[0],
mtd->size >> this->bbt_erase_shift, td, 0);
if (res)
return res;
}
return 0;
}
/* BBT marker is in the first page, no OOB */
static int scan_read_data(struct mtd_info *mtd, uint8_t *buf, loff_t offs,
struct nand_bbt_descr *td)
{
size_t retlen;
size_t len;
len = td->len;
if (td->options & NAND_BBT_VERSION)
len++;
return mtd_read(mtd, offs, len, &retlen, buf);
}
/**
* scan_read_oob - [GENERIC] Scan data+OOB region to buffer
* @mtd: MTD device structure
* @buf: temporary buffer
* @offs: offset at which to scan
* @len: length of data region to read
*
* Scan read data from data+OOB. May traverse multiple pages, interleaving
* page,OOB,page,OOB,... in buf. Completes transfer and returns the "strongest"
* ECC condition (error or bitflip). May quit on the first (non-ECC) error.
*/
static int scan_read_oob(struct mtd_info *mtd, uint8_t *buf, loff_t offs,
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
size_t len)
{
struct mtd_oob_ops ops;
int res, ret = 0;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
ops.mode = MTD_OPS_PLACE_OOB;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
ops.ooboffs = 0;
ops.ooblen = mtd->oobsize;
while (len > 0) {
ops.datbuf = buf;
ops.len = min(len, (size_t)mtd->writesize);
ops.oobbuf = buf + ops.len;
res = mtd_read_oob(mtd, offs, &ops);
if (res) {
if (!mtd_is_bitflip_or_eccerr(res))
return res;
else if (mtd_is_eccerr(res) || !ret)
ret = res;
}
buf += mtd->oobsize + mtd->writesize;
len -= mtd->writesize;
offs += mtd->writesize;
}
return ret;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
}
static int scan_read(struct mtd_info *mtd, uint8_t *buf, loff_t offs,
size_t len, struct nand_bbt_descr *td)
{
if (td->options & NAND_BBT_NO_OOB)
return scan_read_data(mtd, buf, offs, td);
else
return scan_read_oob(mtd, buf, offs, len);
}
/* Scan write data with oob to flash */
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
static int scan_write_bbt(struct mtd_info *mtd, loff_t offs, size_t len,
uint8_t *buf, uint8_t *oob)
{
struct mtd_oob_ops ops;
ops.mode = MTD_OPS_PLACE_OOB;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
ops.ooboffs = 0;
ops.ooblen = mtd->oobsize;
ops.datbuf = buf;
ops.oobbuf = oob;
ops.len = len;
return mtd_write_oob(mtd, offs, &ops);
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
}
static u32 bbt_get_ver_offs(struct mtd_info *mtd, struct nand_bbt_descr *td)
{
u32 ver_offs = td->veroffs;
if (!(td->options & NAND_BBT_NO_OOB))
ver_offs += mtd->writesize;
return ver_offs;
}
/**
* read_abs_bbts - [GENERIC] Read the bad block table(s) for all chips starting at a given page
* @mtd: MTD device structure
* @buf: temporary buffer
* @td: descriptor for the bad block table
* @md: descriptor for the bad block table mirror
*
* Read the bad block table(s) for all chips starting at a given page. We
* assume that the bbt bits are in consecutive order.
*/
static void read_abs_bbts(struct mtd_info *mtd, uint8_t *buf,
struct nand_bbt_descr *td, struct nand_bbt_descr *md)
{
struct nand_chip *this = mtd_to_nand(mtd);
/* Read the primary version, if available */
if (td->options & NAND_BBT_VERSION) {
scan_read(mtd, buf, (loff_t)td->pages[0] << this->page_shift,
mtd->writesize, td);
td->version[0] = buf[bbt_get_ver_offs(mtd, td)];
pr_info("Bad block table at page %d, version 0x%02X\n",
td->pages[0], td->version[0]);
}
/* Read the mirror version, if available */
if (md && (md->options & NAND_BBT_VERSION)) {
scan_read(mtd, buf, (loff_t)md->pages[0] << this->page_shift,
mtd->writesize, md);
md->version[0] = buf[bbt_get_ver_offs(mtd, md)];
pr_info("Bad block table at page %d, version 0x%02X\n",
md->pages[0], md->version[0]);
}
}
/* Scan a given block partially */
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
static int scan_block_fast(struct mtd_info *mtd, struct nand_bbt_descr *bd,
loff_t offs, uint8_t *buf, int numpages)
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
{
struct mtd_oob_ops ops;
int j, ret;
ops.ooblen = mtd->oobsize;
ops.oobbuf = buf;
ops.ooboffs = 0;
ops.datbuf = NULL;
ops.mode = MTD_OPS_PLACE_OOB;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
for (j = 0; j < numpages; j++) {
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
/*
* Read the full oob until read_oob is fixed to handle single
* byte reads for 16 bit buswidth.
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
*/
ret = mtd_read_oob(mtd, offs, &ops);
/* Ignore ECC errors when checking for BBM */
if (ret && !mtd_is_bitflip_or_eccerr(ret))
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
return ret;
if (check_short_pattern(buf, bd))
return 1;
offs += mtd->writesize;
}
return 0;
}
/**
* create_bbt - [GENERIC] Create a bad block table by scanning the device
* @mtd: MTD device structure
* @buf: temporary buffer
* @bd: descriptor for the good/bad block search pattern
* @chip: create the table for a specific chip, -1 read all chips; applies only
* if NAND_BBT_PERCHIP option is set
*
* Create a bad block table by scanning the device for the given good/bad block
* identify pattern.
*/
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
static int create_bbt(struct mtd_info *mtd, uint8_t *buf,
struct nand_bbt_descr *bd, int chip)
{
struct nand_chip *this = mtd_to_nand(mtd);
int i, numblocks, numpages;
int startblock;
loff_t from;
pr_info("Scanning device for bad blocks\n");
if (bd->options & NAND_BBT_SCAN2NDPAGE)
numpages = 2;
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
else
numpages = 1;
if (chip == -1) {
numblocks = mtd->size >> this->bbt_erase_shift;
startblock = 0;
from = 0;
} else {
if (chip >= this->numchips) {
pr_warn("create_bbt(): chipnr (%d) > available chips (%d)\n",
chip + 1, this->numchips);
return -EINVAL;
}
numblocks = this->chipsize >> this->bbt_erase_shift;
startblock = chip * numblocks;
numblocks += startblock;
from = (loff_t)startblock << this->bbt_erase_shift;
}
if (this->bbt_options & NAND_BBT_SCANLASTPAGE)
from += mtd->erasesize - (mtd->writesize * numpages);
mtd: nand: support alternate BB marker locations on MLC This is a slightly modified version of a patch submitted last year by Reuben Dowle <reuben.dowle@navico.com>. His original comments follow: This patch adds support for some MLC NAND flashes that place the BB marker in the LAST page of the bad block rather than the FIRST page used for SLC NAND and other types of MLC nand. Lifted from Samsung datasheet for K9LG8G08U0A (1Gbyte MLC NAND): " Identifying Initial Invalid Block(s) All device locations are erased(FFh) except locations where the initial invalid block(s) information is written prior to shipping. The initial invalid block(s) status is defined by the 1st byte in the spare area. Samsung makes sure that the last page of every initial invalid block has non-FFh data at the column address of 2,048. ... " As far as I can tell, this is the same for all Samsung MLC nand, and in fact the samsung bsp for the processor used in our project (s3c6410) actually contained a hack similar to this patch but less portable to enable use of their NAND parts. I discovered this problem when trying to use a Micron NAND which does not used this layout - I wish samsung would put their stuff in main-line to avoid this type of problem. Currently this patch causes all MLC nand with manufacturer codes from Samsung and ST(Numonyx) to use this alternative location, since these are the manufactures that I know of that use this layout. Signed-off-by: Kevin Cernekee <cernekee@gmail.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-05-05 11:58:10 +08:00
for (i = startblock; i < numblocks; i++) {
int ret;
BUG_ON(bd->options & NAND_BBT_NO_OOB);
ret = scan_block_fast(mtd, bd, from, buf, numpages);
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
if (ret < 0)
return ret;
if (ret) {
bbt_mark_entry(this, i, BBT_BLOCK_FACTORY_BAD);
pr_warn("Bad eraseblock %d at 0x%012llx\n",
i, (unsigned long long)from);
mtd->ecc_stats.badblocks++;
}
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
from += (1 << this->bbt_erase_shift);
}
return 0;
}
/**
* search_bbt - [GENERIC] scan the device for a specific bad block table
* @mtd: MTD device structure
* @buf: temporary buffer
* @td: descriptor for the bad block table
*
* Read the bad block table by searching for a given ident pattern. Search is
* preformed either from the beginning up or from the end of the device
* downwards. The search starts always at the start of a block. If the option
* NAND_BBT_PERCHIP is given, each chip is searched for a bbt, which contains
* the bad block information of this chip. This is necessary to provide support
* for certain DOC devices.
*
* The bbt ident pattern resides in the oob area of the first page in a block.
*/
static int search_bbt(struct mtd_info *mtd, uint8_t *buf, struct nand_bbt_descr *td)
{
struct nand_chip *this = mtd_to_nand(mtd);
int i, chips;
int startblock, block, dir;
int scanlen = mtd->writesize + mtd->oobsize;
int bbtblocks;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
int blocktopage = this->bbt_erase_shift - this->page_shift;
/* Search direction top -> down? */
if (td->options & NAND_BBT_LASTBLOCK) {
startblock = (mtd->size >> this->bbt_erase_shift) - 1;
dir = -1;
} else {
startblock = 0;
dir = 1;
}
/* Do we have a bbt per chip? */
if (td->options & NAND_BBT_PERCHIP) {
chips = this->numchips;
bbtblocks = this->chipsize >> this->bbt_erase_shift;
startblock &= bbtblocks - 1;
} else {
chips = 1;
bbtblocks = mtd->size >> this->bbt_erase_shift;
}
for (i = 0; i < chips; i++) {
/* Reset version information */
td->version[i] = 0;
td->pages[i] = -1;
/* Scan the maximum number of blocks */
for (block = 0; block < td->maxblocks; block++) {
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
int actblock = startblock + dir * block;
loff_t offs = (loff_t)actblock << this->bbt_erase_shift;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
/* Read first page */
scan_read(mtd, buf, offs, mtd->writesize, td);
if (!check_pattern(buf, scanlen, mtd->writesize, td)) {
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
td->pages[i] = actblock << blocktopage;
if (td->options & NAND_BBT_VERSION) {
offs = bbt_get_ver_offs(mtd, td);
td->version[i] = buf[offs];
}
break;
}
}
startblock += this->chipsize >> this->bbt_erase_shift;
}
/* Check, if we found a bbt for each requested chip */
for (i = 0; i < chips; i++) {
if (td->pages[i] == -1)
pr_warn("Bad block table not found for chip %d\n", i);
else
pr_info("Bad block table found at page %d, version 0x%02X\n",
td->pages[i], td->version[i]);
}
return 0;
}
/**
* search_read_bbts - [GENERIC] scan the device for bad block table(s)
* @mtd: MTD device structure
* @buf: temporary buffer
* @td: descriptor for the bad block table
* @md: descriptor for the bad block table mirror
*
* Search and read the bad block table(s).
*/
static void search_read_bbts(struct mtd_info *mtd, uint8_t *buf,
struct nand_bbt_descr *td,
struct nand_bbt_descr *md)
{
/* Search the primary table */
search_bbt(mtd, buf, td);
/* Search the mirror table */
if (md)
search_bbt(mtd, buf, md);
}
/**
* get_bbt_block - Get the first valid eraseblock suitable to store a BBT
* @this: the NAND device
* @td: the BBT description
* @md: the mirror BBT descriptor
* @chip: the CHIP selector
*
* This functions returns a positive block number pointing a valid eraseblock
* suitable to store a BBT (i.e. in the range reserved for BBT), or -ENOSPC if
* all blocks are already used of marked bad. If td->pages[chip] was already
* pointing to a valid block we re-use it, otherwise we search for the next
* valid one.
*/
static int get_bbt_block(struct nand_chip *this, struct nand_bbt_descr *td,
struct nand_bbt_descr *md, int chip)
{
int startblock, dir, page, numblocks, i;
/*
* There was already a version of the table, reuse the page. This
* applies for absolute placement too, as we have the page number in
* td->pages.
*/
if (td->pages[chip] != -1)
return td->pages[chip] >>
(this->bbt_erase_shift - this->page_shift);
numblocks = (int)(this->chipsize >> this->bbt_erase_shift);
if (!(td->options & NAND_BBT_PERCHIP))
numblocks *= this->numchips;
/*
* Automatic placement of the bad block table. Search direction
* top -> down?
*/
if (td->options & NAND_BBT_LASTBLOCK) {
startblock = numblocks * (chip + 1) - 1;
dir = -1;
} else {
startblock = chip * numblocks;
dir = 1;
}
for (i = 0; i < td->maxblocks; i++) {
int block = startblock + dir * i;
/* Check, if the block is bad */
switch (bbt_get_entry(this, block)) {
case BBT_BLOCK_WORN:
case BBT_BLOCK_FACTORY_BAD:
continue;
}
page = block << (this->bbt_erase_shift - this->page_shift);
/* Check, if the block is used by the mirror table */
if (!md || md->pages[chip] != page)
return block;
}
return -ENOSPC;
}
/**
* mark_bbt_block_bad - Mark one of the block reserved for BBT bad
* @this: the NAND device
* @td: the BBT description
* @chip: the CHIP selector
* @block: the BBT block to mark
*
* Blocks reserved for BBT can become bad. This functions is an helper to mark
* such blocks as bad. It takes care of updating the in-memory BBT, marking the
* block as bad using a bad block marker and invalidating the associated
* td->pages[] entry.
*/
static void mark_bbt_block_bad(struct nand_chip *this,
struct nand_bbt_descr *td,
int chip, int block)
{
struct mtd_info *mtd = nand_to_mtd(this);
loff_t to;
int res;
bbt_mark_entry(this, block, BBT_BLOCK_WORN);
to = (loff_t)block << this->bbt_erase_shift;
res = this->block_markbad(mtd, to);
if (res)
pr_warn("nand_bbt: error %d while marking block %d bad\n",
res, block);
td->pages[chip] = -1;
}
/**
* write_bbt - [GENERIC] (Re)write the bad block table
* @mtd: MTD device structure
* @buf: temporary buffer
* @td: descriptor for the bad block table
* @md: descriptor for the bad block table mirror
* @chipsel: selector for a specific chip, -1 for all
*
* (Re)write the bad block table.
*/
static int write_bbt(struct mtd_info *mtd, uint8_t *buf,
struct nand_bbt_descr *td, struct nand_bbt_descr *md,
int chipsel)
{
struct nand_chip *this = mtd_to_nand(mtd);
struct erase_info einfo;
int i, res, chip = 0;
int bits, page, offs, numblocks, sft, sftmsk;
int nrchips, pageoffs, ooboffs;
uint8_t msk[4];
uint8_t rcode = td->reserved_block_code;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
size_t retlen, len = 0;
loff_t to;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
struct mtd_oob_ops ops;
ops.ooblen = mtd->oobsize;
ops.ooboffs = 0;
ops.datbuf = NULL;
ops.mode = MTD_OPS_PLACE_OOB;
if (!rcode)
rcode = 0xff;
/* Write bad block table per chip rather than per device? */
if (td->options & NAND_BBT_PERCHIP) {
numblocks = (int)(this->chipsize >> this->bbt_erase_shift);
/* Full device write or specific chip? */
if (chipsel == -1) {
nrchips = this->numchips;
} else {
nrchips = chipsel + 1;
chip = chipsel;
}
} else {
numblocks = (int)(mtd->size >> this->bbt_erase_shift);
nrchips = 1;
}
/* Loop through the chips */
while (chip < nrchips) {
int block;
block = get_bbt_block(this, td, md, chip);
if (block < 0) {
pr_err("No space left to write bad block table\n");
res = block;
goto outerr;
}
/*
* get_bbt_block() returns a block number, shift the value to
* get a page number.
*/
page = block << (this->bbt_erase_shift - this->page_shift);
/* Set up shift count and masks for the flash table */
bits = td->options & NAND_BBT_NRBITS_MSK;
msk[2] = ~rcode;
switch (bits) {
case 1: sft = 3; sftmsk = 0x07; msk[0] = 0x00; msk[1] = 0x01;
msk[3] = 0x01;
break;
case 2: sft = 2; sftmsk = 0x06; msk[0] = 0x00; msk[1] = 0x01;
msk[3] = 0x03;
break;
case 4: sft = 1; sftmsk = 0x04; msk[0] = 0x00; msk[1] = 0x0C;
msk[3] = 0x0f;
break;
case 8: sft = 0; sftmsk = 0x00; msk[0] = 0x00; msk[1] = 0x0F;
msk[3] = 0xff;
break;
default: return -EINVAL;
}
to = ((loff_t)page) << this->page_shift;
/* Must we save the block contents? */
if (td->options & NAND_BBT_SAVECONTENT) {
/* Make it block aligned */
to &= ~(((loff_t)1 << this->bbt_erase_shift) - 1);
len = 1 << this->bbt_erase_shift;
res = mtd_read(mtd, to, len, &retlen, buf);
if (res < 0) {
if (retlen != len) {
pr_info("nand_bbt: error reading block for writing the bad block table\n");
return res;
}
pr_warn("nand_bbt: ECC error while reading block for writing bad block table\n");
}
/* Read oob data */
ops.ooblen = (len >> this->page_shift) * mtd->oobsize;
[MTD] Rework the out of band handling completely Hopefully the last iteration on this! The handling of out of band data on NAND was accompanied by tons of fruitless discussions and halfarsed patches to make it work for a particular problem. Sufficiently annoyed by I all those "I know it better" mails and the resonable amount of discarded "it solves my problem" patches, I finally decided to go for the big rework. After removing the _ecc variants of mtd read/write functions the solution to satisfy the various requirements was to refactor the read/write _oob functions in mtd. The major change is that read/write_oob now takes a pointer to an operation descriptor structure "struct mtd_oob_ops".instead of having a function with at least seven arguments. read/write_oob which should probably renamed to a more descriptive name, can do the following tasks: - read/write out of band data - read/write data content and out of band data - read/write raw data content and out of band data (ecc disabled) struct mtd_oob_ops has a mode field, which determines the oob handling mode. Aside of the MTD_OOB_RAW mode, which is intended to be especially for diagnostic purposes and some internal functions e.g. bad block table creation, the other two modes are for mtd clients: MTD_OOB_PLACE puts/gets the given oob data exactly to/from the place which is described by the ooboffs and ooblen fields of the mtd_oob_ops strcuture. It's up to the caller to make sure that the byte positions are not used by the ECC placement algorithms. MTD_OOB_AUTO puts/gets the given oob data automaticaly to/from the places in the out of band area which are described by the oobfree tuples in the ecclayout data structre which is associated to the devicee. The decision whether data plus oob or oob only handling is done depends on the setting of the datbuf member of the data structure. When datbuf == NULL then the internal read/write_oob functions are selected, otherwise the read/write data routines are invoked. Tested on a few platforms with all variants. Please be aware of possible regressions for your particular device / application scenario Disclaimer: Any whining will be ignored from those who just contributed "hot air blurb" and never sat down to tackle the underlying problem of the mess in the NAND driver grown over time and the big chunk of work to fix up the existing users. The problem was not the holiness of the existing MTD interfaces. The problems was the lack of time to go for the big overhaul. It's easy to add more mess to the existing one, but it takes alot of effort to go for a real solution. Improvements and bugfixes are welcome! Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
2006-05-29 09:26:58 +08:00
ops.oobbuf = &buf[len];
res = mtd_read_oob(mtd, to + mtd->writesize, &ops);
if (res < 0 || ops.oobretlen != ops.ooblen)
goto outerr;
/* Calc the byte offset in the buffer */
pageoffs = page - (int)(to >> this->page_shift);
offs = pageoffs << this->page_shift;
/* Preset the bbt area with 0xff */
memset(&buf[offs], 0xff, (size_t)(numblocks >> sft));
ooboffs = len + (pageoffs * mtd->oobsize);
} else if (td->options & NAND_BBT_NO_OOB) {
ooboffs = 0;
offs = td->len;
/* The version byte */
if (td->options & NAND_BBT_VERSION)
offs++;
/* Calc length */
len = (size_t)(numblocks >> sft);
len += offs;
/* Make it page aligned! */
len = ALIGN(len, mtd->writesize);
/* Preset the buffer with 0xff */
memset(buf, 0xff, len);
/* Pattern is located at the begin of first page */
memcpy(buf, td->pattern, td->len);
} else {
/* Calc length */
len = (size_t)(numblocks >> sft);
/* Make it page aligned! */
len = ALIGN(len, mtd->writesize);
/* Preset the buffer with 0xff */
memset(buf, 0xff, len +
(len >> this->page_shift)* mtd->oobsize);
offs = 0;
ooboffs = len;
/* Pattern is located in oob area of first page */
memcpy(&buf[ooboffs + td->offs], td->pattern, td->len);
}
if (td->options & NAND_BBT_VERSION)
buf[ooboffs + td->veroffs] = td->version[chip];
/* Walk through the memory table */
for (i = 0; i < numblocks; i++) {
uint8_t dat;
int sftcnt = (i << (3 - sft)) & sftmsk;
dat = bbt_get_entry(this, chip * numblocks + i);
/* Do not store the reserved bbt blocks! */
buf[offs + (i >> sft)] &= ~(msk[dat] << sftcnt);
}
memset(&einfo, 0, sizeof(einfo));
einfo.mtd = mtd;
einfo.addr = to;
einfo.len = 1 << this->bbt_erase_shift;
res = nand_erase_nand(mtd, &einfo, 1);
if (res < 0) {
pr_warn("nand_bbt: error while erasing BBT block %d\n",
res);
mark_bbt_block_bad(this, td, chip, block);
continue;
}
res = scan_write_bbt(mtd, to, len, buf,
td->options & NAND_BBT_NO_OOB ? NULL :
&buf[len]);
if (res < 0) {
pr_warn("nand_bbt: error while writing BBT block %d\n",
res);
mark_bbt_block_bad(this, td, chip, block);
continue;
}
pr_info("Bad block table written to 0x%012llx, version 0x%02X\n",
(unsigned long long)to, td->version[chip]);
/* Mark it as used */
td->pages[chip++] = page;
}
return 0;
outerr:
pr_warn("nand_bbt: error while writing bad block table %d\n", res);
return res;
}
/**
* nand_memory_bbt - [GENERIC] create a memory based bad block table
* @mtd: MTD device structure
* @bd: descriptor for the good/bad block search pattern
*
* The function creates a memory based bbt by scanning the device for
* manufacturer / software marked good / bad blocks.
*/
static inline int nand_memory_bbt(struct mtd_info *mtd, struct nand_bbt_descr *bd)
{
struct nand_chip *this = mtd_to_nand(mtd);
return create_bbt(mtd, this->data_buf, bd, -1);
}
/**
* check_create - [GENERIC] create and write bbt(s) if necessary
* @mtd: MTD device structure
* @buf: temporary buffer
* @bd: descriptor for the good/bad block search pattern
*
* The function checks the results of the previous call to read_bbt and creates
* / updates the bbt(s) if necessary. Creation is necessary if no bbt was found
* for the chip/device. Update is necessary if one of the tables is missing or
* the version nr. of one table is less than the other.
*/
static int check_create(struct mtd_info *mtd, uint8_t *buf, struct nand_bbt_descr *bd)
{
int i, chips, writeops, create, chipsel, res, res2;
struct nand_chip *this = mtd_to_nand(mtd);
struct nand_bbt_descr *td = this->bbt_td;
struct nand_bbt_descr *md = this->bbt_md;
struct nand_bbt_descr *rd, *rd2;
/* Do we have a bbt per chip? */
if (td->options & NAND_BBT_PERCHIP)
chips = this->numchips;
else
chips = 1;
for (i = 0; i < chips; i++) {
writeops = 0;
create = 0;
rd = NULL;
rd2 = NULL;
res = res2 = 0;
/* Per chip or per device? */
chipsel = (td->options & NAND_BBT_PERCHIP) ? i : -1;
/* Mirrored table available? */
if (md) {
if (td->pages[i] == -1 && md->pages[i] == -1) {
create = 1;
writeops = 0x03;
} else if (td->pages[i] == -1) {
rd = md;
writeops = 0x01;
} else if (md->pages[i] == -1) {
rd = td;
writeops = 0x02;
} else if (td->version[i] == md->version[i]) {
rd = td;
if (!(td->options & NAND_BBT_VERSION))
rd2 = md;
} else if (((int8_t)(td->version[i] - md->version[i])) > 0) {
rd = td;
writeops = 0x02;
} else {
rd = md;
writeops = 0x01;
}
} else {
if (td->pages[i] == -1) {
create = 1;
writeops = 0x01;
} else {
rd = td;
}
}
if (create) {
/* Create the bad block table by scanning the device? */
if (!(td->options & NAND_BBT_CREATE))
continue;
/* Create the table in memory by scanning the chip(s) */
if (!(this->bbt_options & NAND_BBT_CREATE_EMPTY))
create_bbt(mtd, buf, bd, chipsel);
td->version[i] = 1;
if (md)
md->version[i] = 1;
}
/* Read back first? */
if (rd) {
res = read_abs_bbt(mtd, buf, rd, chipsel);
if (mtd_is_eccerr(res)) {
/* Mark table as invalid */
rd->pages[i] = -1;
rd->version[i] = 0;
i--;
continue;
}
}
/* If they weren't versioned, read both */
if (rd2) {
res2 = read_abs_bbt(mtd, buf, rd2, chipsel);
if (mtd_is_eccerr(res2)) {
/* Mark table as invalid */
rd2->pages[i] = -1;
rd2->version[i] = 0;
i--;
continue;
}
}
/* Scrub the flash table(s)? */
if (mtd_is_bitflip(res) || mtd_is_bitflip(res2))
writeops = 0x03;
/* Update version numbers before writing */
if (md) {
td->version[i] = max(td->version[i], md->version[i]);
md->version[i] = td->version[i];
}
/* Write the bad block table to the device? */
if ((writeops & 0x01) && (td->options & NAND_BBT_WRITE)) {
res = write_bbt(mtd, buf, td, md, chipsel);
if (res < 0)
return res;
}
/* Write the mirror bad block table to the device? */
if ((writeops & 0x02) && md && (md->options & NAND_BBT_WRITE)) {
res = write_bbt(mtd, buf, md, td, chipsel);
if (res < 0)
return res;
}
}
return 0;
}
/**
* mark_bbt_regions - [GENERIC] mark the bad block table regions
* @mtd: MTD device structure
* @td: bad block table descriptor
*
* The bad block table regions are marked as "bad" to prevent accidental
* erasures / writes. The regions are identified by the mark 0x02.
*/
static void mark_bbt_region(struct mtd_info *mtd, struct nand_bbt_descr *td)
{
struct nand_chip *this = mtd_to_nand(mtd);
int i, j, chips, block, nrblocks, update;
uint8_t oldval;
/* Do we have a bbt per chip? */
if (td->options & NAND_BBT_PERCHIP) {
chips = this->numchips;
nrblocks = (int)(this->chipsize >> this->bbt_erase_shift);
} else {
chips = 1;
nrblocks = (int)(mtd->size >> this->bbt_erase_shift);
}
for (i = 0; i < chips; i++) {
if ((td->options & NAND_BBT_ABSPAGE) ||
!(td->options & NAND_BBT_WRITE)) {
if (td->pages[i] == -1)
continue;
block = td->pages[i] >> (this->bbt_erase_shift - this->page_shift);
oldval = bbt_get_entry(this, block);
bbt_mark_entry(this, block, BBT_BLOCK_RESERVED);
if ((oldval != BBT_BLOCK_RESERVED) &&
td->reserved_block_code)
nand_update_bbt(mtd, (loff_t)block <<
this->bbt_erase_shift);
continue;
}
update = 0;
if (td->options & NAND_BBT_LASTBLOCK)
block = ((i + 1) * nrblocks) - td->maxblocks;
else
block = i * nrblocks;
for (j = 0; j < td->maxblocks; j++) {
oldval = bbt_get_entry(this, block);
bbt_mark_entry(this, block, BBT_BLOCK_RESERVED);
if (oldval != BBT_BLOCK_RESERVED)
update = 1;
block++;
}
/*
* If we want reserved blocks to be recorded to flash, and some
* new ones have been marked, then we need to update the stored
* bbts. This should only happen once.
*/
if (update && td->reserved_block_code)
nand_update_bbt(mtd, (loff_t)(block - 1) <<
this->bbt_erase_shift);
}
}
/**
* verify_bbt_descr - verify the bad block description
* @mtd: MTD device structure
* @bd: the table to verify
*
* This functions performs a few sanity checks on the bad block description
* table.
*/
static void verify_bbt_descr(struct mtd_info *mtd, struct nand_bbt_descr *bd)
{
struct nand_chip *this = mtd_to_nand(mtd);
u32 pattern_len;
u32 bits;
u32 table_size;
if (!bd)
return;
pattern_len = bd->len;
bits = bd->options & NAND_BBT_NRBITS_MSK;
BUG_ON((this->bbt_options & NAND_BBT_NO_OOB) &&
!(this->bbt_options & NAND_BBT_USE_FLASH));
BUG_ON(!bits);
if (bd->options & NAND_BBT_VERSION)
pattern_len++;
if (bd->options & NAND_BBT_NO_OOB) {
BUG_ON(!(this->bbt_options & NAND_BBT_USE_FLASH));
BUG_ON(!(this->bbt_options & NAND_BBT_NO_OOB));
BUG_ON(bd->offs);
if (bd->options & NAND_BBT_VERSION)
BUG_ON(bd->veroffs != bd->len);
BUG_ON(bd->options & NAND_BBT_SAVECONTENT);
}
if (bd->options & NAND_BBT_PERCHIP)
table_size = this->chipsize >> this->bbt_erase_shift;
else
table_size = mtd->size >> this->bbt_erase_shift;
table_size >>= 3;
table_size *= bits;
if (bd->options & NAND_BBT_NO_OOB)
table_size += pattern_len;
BUG_ON(table_size > (1 << this->bbt_erase_shift));
}
/**
* nand_scan_bbt - [NAND Interface] scan, find, read and maybe create bad block table(s)
* @mtd: MTD device structure
* @bd: descriptor for the good/bad block search pattern
*
* The function checks, if a bad block table(s) is/are already available. If
* not it scans the device for manufacturer marked good / bad blocks and writes
* the bad block table(s) to the selected place.
*
* The bad block table memory is allocated here. It must be freed by calling
* the nand_free_bbt function.
*/
static int nand_scan_bbt(struct mtd_info *mtd, struct nand_bbt_descr *bd)
{
struct nand_chip *this = mtd_to_nand(mtd);
int len, res;
uint8_t *buf;
struct nand_bbt_descr *td = this->bbt_td;
struct nand_bbt_descr *md = this->bbt_md;
mtd: nand_bbt: set the smallest size of bbt table When using nandsim to simulate a 128K block nand with `overridesize = 1', the size of mtd device is too small (mtd_size = 4 * block_size) to get the right length of bbt. Then when creating bbt, kzmalloc() will return ZERO_SIZE_PTR. This causes a NULL pointer oops when scanning bbt. [ 952.156166] BUG: unable to handle kernel NULL pointer dereference at 0000000000000010 [ 952.157064] IP: [<ffffffff8148ad4a>] nand_isreserved_bbt+0x2a/0x40 [ 952.157064] PGD 0 [ 952.157064] Oops: 0000 [#1] SMP [ 952.157064] Modules linked in: nandsim(+) [last unloaded: nandsim] [ 952.157064] CPU: 1 PID: 7103 Comm: modprobe Not tainted 4.2.0-rc3-next-20150724 #4 [ 952.157064] Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006 [ 952.157064] task: ffff88003e24b980 ti: ffff88003d274000 task.ti: ffff88003d274000 [ 952.157064] RIP: 0010:[<ffffffff8148ad4a>] [<ffffffff8148ad4a>] nand_isreserved_bbt+0x2a/0x40 [ 952.157064] RSP: 0018:ffff88003d277b90 EFLAGS: 00010246 [ 952.157064] RAX: 0000000000000010 RBX: ffff88003d5a1000 RCX: 0000000000000000 [ 952.157064] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffff88003d919000 [ 952.157064] RBP: ffff88003d277b98 R08: 0000000000020000 R09: 0000000000000000 [ 952.157064] R10: 0000000000000000 R11: 0000000000000195 R12: ffff88003d919000 [ 952.157064] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000 [ 952.157064] FS: 00007fada4d07700(0000) GS:ffff88003fd00000(0000) knlGS:0000000000000000 [ 952.157064] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [ 952.157064] CR2: 0000000000000010 CR3: 0000000037924000 CR4: 00000000000006a0 [ 952.157064] Stack: [ 952.157064] ffffffff814851ec ffff88003d277ba8 ffffffff8147e35f ffff88003d277bf8 [ 952.157064] ffffffff814816f3 ffff88003d277c08 ffff88003d277bc8 0000000000000282 [ 952.157064] 0000000000000001 0000000000000000 ffff88003d209540 0000000000000001 [ 952.157064] Call Trace: [ 952.157064] [<ffffffff814851ec>] ? nand_block_isreserved+0x1c/0x20 [ 952.157064] [<ffffffff8147e35f>] mtd_block_isreserved+0x1f/0x30 [ 952.157064] [<ffffffff814816f3>] allocate_partition+0x463/0x6a0 [ 952.157064] [<ffffffff81481b3b>] add_mtd_partitions+0x4b/0xe0 [ 952.157064] [<ffffffff8147f14c>] mtd_device_parse_register+0x4c/0xe0 [ 952.157064] [<ffffffffa0013daf>] ns_init_module+0xdaf/0xde4 [nandsim] [ 952.157064] [<ffffffff8128d7c8>] ? kasprintf+0x38/0x40 [ 952.157064] [<ffffffffa0013000>] ? 0xffffffffa0013000 [ 952.157064] [<ffffffff810002c3>] do_one_initcall+0x83/0x1b0 [ 952.157064] [<ffffffff8113afab>] ? kmem_cache_alloc_trace+0x6b/0x120 [ 952.157064] [<ffffffff8160b503>] do_init_module+0x5c/0x1dd [ 952.157064] [<ffffffff810aa4db>] load_module+0x1bbb/0x20b0 [ 952.157064] [<ffffffff810a6fc0>] ? __symbol_put+0x30/0x30 [ 952.157064] [<ffffffff810aaac9>] SyS_init_module+0xf9/0x110 [ 952.157064] [<ffffffff810aa9d1>] ? SyS_init_module+0x1/0x110 [ 952.157064] [<ffffffff81615f57>] entry_SYSCALL_64_fastpath+0x12/0x6a [ 952.157064] Code: 00 55 48 8b 87 80 01 00 00 48 89 e5 8b 88 cc 00 00 00 48 8b 80 f0 03 00 00 5d 48 d3 fe 89 f2 83 e6 03 c1 fa 02 8d 0c 36 48 63 d2 <0f> b6 04 10 d3 f8 83 e0 03 3c 02 0f 94 c0 0f b6 c0 c3 0f 1f 40 [ 952.157064] RIP [<ffffffff8148ad4a>] nand_isreserved_bbt+0x2a/0x40 [ 952.157064] RSP <ffff88003d277b90> [ 952.157064] CR2: 0000000000000010 [ 952.204010] ---[ end trace 6ca2e1c041fdba36 ]--- This patch gives a smallest length to bbt, 1 byte, which is enough to represent up to 4 blocks. Signed-off-by: Sheng Yong <shengyong1@huawei.com> Signed-off-by: Brian Norris <computersforpeace@gmail.com>
2015-07-31 09:12:44 +08:00
len = (mtd->size >> (this->bbt_erase_shift + 2)) ? : 1;
/*
* Allocate memory (2bit per block) and clear the memory bad block
* table.
*/
this->bbt = kzalloc(len, GFP_KERNEL);
if (!this->bbt)
return -ENOMEM;
/*
* If no primary table decriptor is given, scan the device to build a
* memory based bad block table.
*/
if (!td) {
if ((res = nand_memory_bbt(mtd, bd))) {
pr_err("nand_bbt: can't scan flash and build the RAM-based BBT\n");
goto err;
}
return 0;
}
verify_bbt_descr(mtd, td);
verify_bbt_descr(mtd, md);
/* Allocate a temporary buffer for one eraseblock incl. oob */
len = (1 << this->bbt_erase_shift);
len += (len >> this->page_shift) * mtd->oobsize;
buf = vmalloc(len);
if (!buf) {
res = -ENOMEM;
goto err;
}
/* Is the bbt at a given page? */
if (td->options & NAND_BBT_ABSPAGE) {
read_abs_bbts(mtd, buf, td, md);
} else {
/* Search the bad block table using a pattern in oob */
search_read_bbts(mtd, buf, td, md);
}
res = check_create(mtd, buf, bd);
if (res)
goto err;
/* Prevent the bbt regions from erasing / writing */
mark_bbt_region(mtd, td);
if (md)
mark_bbt_region(mtd, md);
vfree(buf);
return 0;
err:
kfree(this->bbt);
this->bbt = NULL;
return res;
}
/**
* nand_update_bbt - update bad block table(s)
* @mtd: MTD device structure
* @offs: the offset of the newly marked block
*
* The function updates the bad block table(s).
*/
static int nand_update_bbt(struct mtd_info *mtd, loff_t offs)
{
struct nand_chip *this = mtd_to_nand(mtd);
int len, res = 0;
int chip, chipsel;
uint8_t *buf;
struct nand_bbt_descr *td = this->bbt_td;
struct nand_bbt_descr *md = this->bbt_md;
if (!this->bbt || !td)
return -EINVAL;
/* Allocate a temporary buffer for one eraseblock incl. oob */
len = (1 << this->bbt_erase_shift);
len += (len >> this->page_shift) * mtd->oobsize;
buf = kmalloc(len, GFP_KERNEL);
if (!buf)
return -ENOMEM;
/* Do we have a bbt per chip? */
if (td->options & NAND_BBT_PERCHIP) {
chip = (int)(offs >> this->chip_shift);
chipsel = chip;
} else {
chip = 0;
chipsel = -1;
}
td->version[chip]++;
if (md)
md->version[chip]++;
/* Write the bad block table to the device? */
if (td->options & NAND_BBT_WRITE) {
res = write_bbt(mtd, buf, td, md, chipsel);
if (res < 0)
goto out;
}
/* Write the mirror bad block table to the device? */
if (md && (md->options & NAND_BBT_WRITE)) {
res = write_bbt(mtd, buf, md, td, chipsel);
}
out:
kfree(buf);
return res;
}
/*
* Define some generic bad / good block scan pattern which are used
* while scanning a device for factory marked good / bad blocks.
*/
static uint8_t scan_ff_pattern[] = { 0xff, 0xff };
/* Generic flash bbt descriptors */
static uint8_t bbt_pattern[] = {'B', 'b', 't', '0' };
static uint8_t mirror_pattern[] = {'1', 't', 'b', 'B' };
static struct nand_bbt_descr bbt_main_descr = {
.options = NAND_BBT_LASTBLOCK | NAND_BBT_CREATE | NAND_BBT_WRITE
| NAND_BBT_2BIT | NAND_BBT_VERSION | NAND_BBT_PERCHIP,
.offs = 8,
.len = 4,
.veroffs = 12,
.maxblocks = NAND_BBT_SCAN_MAXBLOCKS,
.pattern = bbt_pattern
};
static struct nand_bbt_descr bbt_mirror_descr = {
.options = NAND_BBT_LASTBLOCK | NAND_BBT_CREATE | NAND_BBT_WRITE
| NAND_BBT_2BIT | NAND_BBT_VERSION | NAND_BBT_PERCHIP,
.offs = 8,
.len = 4,
.veroffs = 12,
.maxblocks = NAND_BBT_SCAN_MAXBLOCKS,
.pattern = mirror_pattern
};
static struct nand_bbt_descr bbt_main_no_oob_descr = {
.options = NAND_BBT_LASTBLOCK | NAND_BBT_CREATE | NAND_BBT_WRITE
| NAND_BBT_2BIT | NAND_BBT_VERSION | NAND_BBT_PERCHIP
| NAND_BBT_NO_OOB,
.len = 4,
.veroffs = 4,
.maxblocks = NAND_BBT_SCAN_MAXBLOCKS,
.pattern = bbt_pattern
};
static struct nand_bbt_descr bbt_mirror_no_oob_descr = {
.options = NAND_BBT_LASTBLOCK | NAND_BBT_CREATE | NAND_BBT_WRITE
| NAND_BBT_2BIT | NAND_BBT_VERSION | NAND_BBT_PERCHIP
| NAND_BBT_NO_OOB,
.len = 4,
.veroffs = 4,
.maxblocks = NAND_BBT_SCAN_MAXBLOCKS,
.pattern = mirror_pattern
};
#define BADBLOCK_SCAN_MASK (~NAND_BBT_NO_OOB)
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
/**
* nand_create_badblock_pattern - [INTERN] Creates a BBT descriptor structure
* @this: NAND chip to create descriptor for
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
*
* This function allocates and initializes a nand_bbt_descr for BBM detection
* based on the properties of @this. The new descriptor is stored in
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
* this->badblock_pattern. Thus, this->badblock_pattern should be NULL when
* passed to this function.
*/
static int nand_create_badblock_pattern(struct nand_chip *this)
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
{
struct nand_bbt_descr *bd;
if (this->badblock_pattern) {
pr_warn("Bad block pattern already allocated; not replacing\n");
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
return -EINVAL;
}
bd = kzalloc(sizeof(*bd), GFP_KERNEL);
if (!bd)
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
return -ENOMEM;
bd->options = this->bbt_options & BADBLOCK_SCAN_MASK;
mtd: nand: more BB Detection refactoring and dynamic scan options This is a revision to PATCH 2/2 that I sent. Link: http://lists.infradead.org/pipermail/linux-mtd/2010-July/030911.html Added new flag for scanning of both bytes 1 and 6 of the OOB for a BB marker (instead of simply one or the other). The "check_pattern" and "check_short_pattern" functions were updated to include support for scanning the two different locations in the OOB. In order to handle increases in variety of necessary scanning patterns, I implemented dynamic memory allocation of nand_bbt_descr structs in new function 'nand_create_default_bbt_descr()'. This replaces some increasingly-unwieldy, statically-declared descriptors. It can replace several more (e.g. "flashbased" structs). However, I do not test the flashbased options personally. How this was tested: I referenced 30+ data sheets (covering 100+ parts), and I tested a selection of 10 different chips to varying degrees. Particularly, I tested the creation of bad-block descriptors and basic BB scanning on three parts: ST NAND04GW3B2D, 2K page ST NAND128W3A, 512B page Samsung K9F1G08U0A, 2K page To test these, I wrote some fake bad block markers to the flash (in OOB bytes 1, 6, and elsewhere) to see if the scanning routine would detect them properly. However, this method was somewhat limited because the driver I am using has some bugs in its OOB write functionality. Signed-off-by: Brian Norris <norris@broadcom.com> Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com> Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
2010-07-16 03:15:44 +08:00
bd->offs = this->badblockpos;
bd->len = (this->options & NAND_BUSWIDTH_16) ? 2 : 1;
bd->pattern = scan_ff_pattern;
bd->options |= NAND_BBT_DYNAMICSTRUCT;
this->badblock_pattern = bd;
return 0;
}
/**
* nand_default_bbt - [NAND Interface] Select a default bad block table for the device
* @mtd: MTD device structure
*
* This function selects the default bad block table support for the device and
* calls the nand_scan_bbt function.
*/
int nand_default_bbt(struct mtd_info *mtd)
{
struct nand_chip *this = mtd_to_nand(mtd);
int ret;
/* Is a flash based bad block table requested? */
if (this->bbt_options & NAND_BBT_USE_FLASH) {
/* Use the default pattern descriptors */
if (!this->bbt_td) {
if (this->bbt_options & NAND_BBT_NO_OOB) {
this->bbt_td = &bbt_main_no_oob_descr;
this->bbt_md = &bbt_mirror_no_oob_descr;
} else {
this->bbt_td = &bbt_main_descr;
this->bbt_md = &bbt_mirror_descr;
}
}
} else {
this->bbt_td = NULL;
this->bbt_md = NULL;
}
if (!this->badblock_pattern) {
ret = nand_create_badblock_pattern(this);
if (ret)
return ret;
}
return nand_scan_bbt(mtd, this->badblock_pattern);
}
/**
* nand_isreserved_bbt - [NAND Interface] Check if a block is reserved
* @mtd: MTD device structure
* @offs: offset in the device
*/
int nand_isreserved_bbt(struct mtd_info *mtd, loff_t offs)
{
struct nand_chip *this = mtd_to_nand(mtd);
int block;
block = (int)(offs >> this->bbt_erase_shift);
return bbt_get_entry(this, block) == BBT_BLOCK_RESERVED;
}
/**
* nand_isbad_bbt - [NAND Interface] Check if a block is bad
* @mtd: MTD device structure
* @offs: offset in the device
* @allowbbt: allow access to bad block table region
*/
int nand_isbad_bbt(struct mtd_info *mtd, loff_t offs, int allowbbt)
{
struct nand_chip *this = mtd_to_nand(mtd);
int block, res;
block = (int)(offs >> this->bbt_erase_shift);
res = bbt_get_entry(this, block);
pr_debug("nand_isbad_bbt(): bbt info for offs 0x%08x: (block %d) 0x%02x\n",
(unsigned int)offs, block, res);
switch (res) {
case BBT_BLOCK_GOOD:
return 0;
case BBT_BLOCK_WORN:
return 1;
case BBT_BLOCK_RESERVED:
return allowbbt ? 0 : 1;
}
return 1;
}
/**
* nand_markbad_bbt - [NAND Interface] Mark a block bad in the BBT
* @mtd: MTD device structure
* @offs: offset of the bad block
*/
int nand_markbad_bbt(struct mtd_info *mtd, loff_t offs)
{
struct nand_chip *this = mtd_to_nand(mtd);
int block, ret = 0;
block = (int)(offs >> this->bbt_erase_shift);
/* Mark bad block in memory */
bbt_mark_entry(this, block, BBT_BLOCK_WORN);
/* Update flash-based bad block table */
if (this->bbt_options & NAND_BBT_USE_FLASH)
ret = nand_update_bbt(mtd, offs);
return ret;
}