linux/drivers/mtd/devices/doc2000.c

1202 lines
32 KiB
C
Raw Normal View History

/*
* Linux driver for Disk-On-Chip 2000 and Millennium
* (c) 1999 Machine Vision Holdings, Inc.
* (c) 1999, 2000 David Woodhouse <dwmw2@infradead.org>
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <asm/errno.h>
#include <asm/io.h>
#include <asm/uaccess.h>
#include <linux/delay.h>
#include <linux/slab.h>
#include <linux/sched.h>
#include <linux/init.h>
#include <linux/types.h>
#include <linux/bitops.h>
#include <linux/mutex.h>
#include <linux/mtd/mtd.h>
#include <linux/mtd/nand.h>
#include <linux/mtd/doc2000.h>
#define DOC_SUPPORT_2000
#define DOC_SUPPORT_2000TSOP
#define DOC_SUPPORT_MILLENNIUM
#ifdef DOC_SUPPORT_2000
#define DoC_is_2000(doc) (doc->ChipID == DOC_ChipID_Doc2k)
#else
#define DoC_is_2000(doc) (0)
#endif
#if defined(DOC_SUPPORT_2000TSOP) || defined(DOC_SUPPORT_MILLENNIUM)
#define DoC_is_Millennium(doc) (doc->ChipID == DOC_ChipID_DocMil)
#else
#define DoC_is_Millennium(doc) (0)
#endif
/* #define ECC_DEBUG */
/* I have no idea why some DoC chips can not use memcpy_from|to_io().
* This may be due to the different revisions of the ASIC controller built-in or
* simplily a QA/Bug issue. Who knows ?? If you have trouble, please uncomment
* this:
#undef USE_MEMCPY
*/
static int doc_read(struct mtd_info *mtd, loff_t from, size_t len,
size_t *retlen, u_char *buf);
static int doc_write(struct mtd_info *mtd, loff_t to, size_t len,
size_t *retlen, const u_char *buf);
[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 doc_read_oob(struct mtd_info *mtd, loff_t ofs,
struct mtd_oob_ops *ops);
static int doc_write_oob(struct mtd_info *mtd, loff_t ofs,
struct mtd_oob_ops *ops);
static int doc_write_oob_nolock(struct mtd_info *mtd, loff_t ofs, size_t len,
size_t *retlen, const u_char *buf);
static int doc_erase (struct mtd_info *mtd, struct erase_info *instr);
static struct mtd_info *doc2klist = NULL;
/* Perform the required delay cycles by reading from the appropriate register */
static void DoC_Delay(struct DiskOnChip *doc, unsigned short cycles)
{
volatile char dummy;
int i;
for (i = 0; i < cycles; i++) {
if (DoC_is_Millennium(doc))
dummy = ReadDOC(doc->virtadr, NOP);
else
dummy = ReadDOC(doc->virtadr, DOCStatus);
}
}
/* DOC_WaitReady: Wait for RDY line to be asserted by the flash chip */
static int _DoC_WaitReady(struct DiskOnChip *doc)
{
void __iomem *docptr = doc->virtadr;
unsigned long timeo = jiffies + (HZ * 10);
DEBUG(MTD_DEBUG_LEVEL3,
"_DoC_WaitReady called for out-of-line wait\n");
/* Out-of-line routine to wait for chip response */
while (!(ReadDOC(docptr, CDSNControl) & CDSN_CTRL_FR_B)) {
/* issue 2 read from NOP register after reading from CDSNControl register
see Software Requirement 11.4 item 2. */
DoC_Delay(doc, 2);
if (time_after(jiffies, timeo)) {
DEBUG(MTD_DEBUG_LEVEL2, "_DoC_WaitReady timed out.\n");
return -EIO;
}
udelay(1);
cond_resched();
}
return 0;
}
static inline int DoC_WaitReady(struct DiskOnChip *doc)
{
void __iomem *docptr = doc->virtadr;
/* This is inline, to optimise the common case, where it's ready instantly */
int ret = 0;
/* 4 read form NOP register should be issued in prior to the read from CDSNControl
see Software Requirement 11.4 item 2. */
DoC_Delay(doc, 4);
if (!(ReadDOC(docptr, CDSNControl) & CDSN_CTRL_FR_B))
/* Call the out-of-line routine to wait */
ret = _DoC_WaitReady(doc);
/* issue 2 read from NOP register after reading from CDSNControl register
see Software Requirement 11.4 item 2. */
DoC_Delay(doc, 2);
return ret;
}
/* DoC_Command: Send a flash command to the flash chip through the CDSN Slow IO register to
bypass the internal pipeline. Each of 4 delay cycles (read from the NOP register) is
required after writing to CDSN Control register, see Software Requirement 11.4 item 3. */
static int DoC_Command(struct DiskOnChip *doc, unsigned char command,
unsigned char xtraflags)
{
void __iomem *docptr = doc->virtadr;
if (DoC_is_2000(doc))
xtraflags |= CDSN_CTRL_FLASH_IO;
/* Assert the CLE (Command Latch Enable) line to the flash chip */
WriteDOC(xtraflags | CDSN_CTRL_CLE | CDSN_CTRL_CE, docptr, CDSNControl);
DoC_Delay(doc, 4); /* Software requirement 11.4.3 for Millennium */
if (DoC_is_Millennium(doc))
WriteDOC(command, docptr, CDSNSlowIO);
/* Send the command */
WriteDOC_(command, docptr, doc->ioreg);
if (DoC_is_Millennium(doc))
WriteDOC(command, docptr, WritePipeTerm);
/* Lower the CLE line */
WriteDOC(xtraflags | CDSN_CTRL_CE, docptr, CDSNControl);
DoC_Delay(doc, 4); /* Software requirement 11.4.3 for Millennium */
/* Wait for the chip to respond - Software requirement 11.4.1 (extended for any command) */
return DoC_WaitReady(doc);
}
/* DoC_Address: Set the current address for the flash chip through the CDSN Slow IO register to
bypass the internal pipeline. Each of 4 delay cycles (read from the NOP register) is
required after writing to CDSN Control register, see Software Requirement 11.4 item 3. */
static int DoC_Address(struct DiskOnChip *doc, int numbytes, unsigned long ofs,
unsigned char xtraflags1, unsigned char xtraflags2)
{
int i;
void __iomem *docptr = doc->virtadr;
if (DoC_is_2000(doc))
xtraflags1 |= CDSN_CTRL_FLASH_IO;
/* Assert the ALE (Address Latch Enable) line to the flash chip */
WriteDOC(xtraflags1 | CDSN_CTRL_ALE | CDSN_CTRL_CE, docptr, CDSNControl);
DoC_Delay(doc, 4); /* Software requirement 11.4.3 for Millennium */
/* Send the address */
/* Devices with 256-byte page are addressed as:
Column (bits 0-7), Page (bits 8-15, 16-23, 24-31)
* there is no device on the market with page256
and more than 24 bits.
Devices with 512-byte page are addressed as:
Column (bits 0-7), Page (bits 9-16, 17-24, 25-31)
* 25-31 is sent only if the chip support it.
* bit 8 changes the read command to be sent
(NAND_CMD_READ0 or NAND_CMD_READ1).
*/
if (numbytes == ADDR_COLUMN || numbytes == ADDR_COLUMN_PAGE) {
if (DoC_is_Millennium(doc))
WriteDOC(ofs & 0xff, docptr, CDSNSlowIO);
WriteDOC_(ofs & 0xff, docptr, doc->ioreg);
}
if (doc->page256) {
ofs = ofs >> 8;
} else {
ofs = ofs >> 9;
}
if (numbytes == ADDR_PAGE || numbytes == ADDR_COLUMN_PAGE) {
for (i = 0; i < doc->pageadrlen; i++, ofs = ofs >> 8) {
if (DoC_is_Millennium(doc))
WriteDOC(ofs & 0xff, docptr, CDSNSlowIO);
WriteDOC_(ofs & 0xff, docptr, doc->ioreg);
}
}
if (DoC_is_Millennium(doc))
WriteDOC(ofs & 0xff, docptr, WritePipeTerm);
DoC_Delay(doc, 2); /* Needed for some slow flash chips. mf. */
/* FIXME: The SlowIO's for millennium could be replaced by
a single WritePipeTerm here. mf. */
/* Lower the ALE line */
WriteDOC(xtraflags1 | xtraflags2 | CDSN_CTRL_CE, docptr,
CDSNControl);
DoC_Delay(doc, 4); /* Software requirement 11.4.3 for Millennium */
/* Wait for the chip to respond - Software requirement 11.4.1 */
return DoC_WaitReady(doc);
}
/* Read a buffer from DoC, taking care of Millennium odditys */
static void DoC_ReadBuf(struct DiskOnChip *doc, u_char * buf, int len)
{
volatile int dummy;
int modulus = 0xffff;
void __iomem *docptr = doc->virtadr;
int i;
if (len <= 0)
return;
if (DoC_is_Millennium(doc)) {
/* Read the data via the internal pipeline through CDSN IO register,
see Pipelined Read Operations 11.3 */
dummy = ReadDOC(docptr, ReadPipeInit);
/* Millennium should use the LastDataRead register - Pipeline Reads */
len--;
/* This is needed for correctly ECC calculation */
modulus = 0xff;
}
for (i = 0; i < len; i++)
buf[i] = ReadDOC_(docptr, doc->ioreg + (i & modulus));
if (DoC_is_Millennium(doc)) {
buf[i] = ReadDOC(docptr, LastDataRead);
}
}
/* Write a buffer to DoC, taking care of Millennium odditys */
static void DoC_WriteBuf(struct DiskOnChip *doc, const u_char * buf, int len)
{
void __iomem *docptr = doc->virtadr;
int i;
if (len <= 0)
return;
for (i = 0; i < len; i++)
WriteDOC_(buf[i], docptr, doc->ioreg + i);
if (DoC_is_Millennium(doc)) {
WriteDOC(0x00, docptr, WritePipeTerm);
}
}
/* DoC_SelectChip: Select a given flash chip within the current floor */
static inline int DoC_SelectChip(struct DiskOnChip *doc, int chip)
{
void __iomem *docptr = doc->virtadr;
/* Software requirement 11.4.4 before writing DeviceSelect */
/* Deassert the CE line to eliminate glitches on the FCE# outputs */
WriteDOC(CDSN_CTRL_WP, docptr, CDSNControl);
DoC_Delay(doc, 4); /* Software requirement 11.4.3 for Millennium */
/* Select the individual flash chip requested */
WriteDOC(chip, docptr, CDSNDeviceSelect);
DoC_Delay(doc, 4);
/* Reassert the CE line */
WriteDOC(CDSN_CTRL_CE | CDSN_CTRL_FLASH_IO | CDSN_CTRL_WP, docptr,
CDSNControl);
DoC_Delay(doc, 4); /* Software requirement 11.4.3 for Millennium */
/* Wait for it to be ready */
return DoC_WaitReady(doc);
}
/* DoC_SelectFloor: Select a given floor (bank of flash chips) */
static inline int DoC_SelectFloor(struct DiskOnChip *doc, int floor)
{
void __iomem *docptr = doc->virtadr;
/* Select the floor (bank) of chips required */
WriteDOC(floor, docptr, FloorSelect);
/* Wait for the chip to be ready */
return DoC_WaitReady(doc);
}
/* DoC_IdentChip: Identify a given NAND chip given {floor,chip} */
static int DoC_IdentChip(struct DiskOnChip *doc, int floor, int chip)
{
int mfr, id, i, j;
volatile char dummy;
/* Page in the required floor/chip */
DoC_SelectFloor(doc, floor);
DoC_SelectChip(doc, chip);
/* Reset the chip */
if (DoC_Command(doc, NAND_CMD_RESET, CDSN_CTRL_WP)) {
DEBUG(MTD_DEBUG_LEVEL2,
"DoC_Command (reset) for %d,%d returned true\n",
floor, chip);
return 0;
}
/* Read the NAND chip ID: 1. Send ReadID command */
if (DoC_Command(doc, NAND_CMD_READID, CDSN_CTRL_WP)) {
DEBUG(MTD_DEBUG_LEVEL2,
"DoC_Command (ReadID) for %d,%d returned true\n",
floor, chip);
return 0;
}
/* Read the NAND chip ID: 2. Send address byte zero */
DoC_Address(doc, ADDR_COLUMN, 0, CDSN_CTRL_WP, 0);
/* Read the manufacturer and device id codes from the device */
if (DoC_is_Millennium(doc)) {
DoC_Delay(doc, 2);
dummy = ReadDOC(doc->virtadr, ReadPipeInit);
mfr = ReadDOC(doc->virtadr, LastDataRead);
DoC_Delay(doc, 2);
dummy = ReadDOC(doc->virtadr, ReadPipeInit);
id = ReadDOC(doc->virtadr, LastDataRead);
} else {
/* CDSN Slow IO register see Software Req 11.4 item 5. */
dummy = ReadDOC(doc->virtadr, CDSNSlowIO);
DoC_Delay(doc, 2);
mfr = ReadDOC_(doc->virtadr, doc->ioreg);
/* CDSN Slow IO register see Software Req 11.4 item 5. */
dummy = ReadDOC(doc->virtadr, CDSNSlowIO);
DoC_Delay(doc, 2);
id = ReadDOC_(doc->virtadr, doc->ioreg);
}
/* No response - return failure */
if (mfr == 0xff || mfr == 0)
return 0;
/* Check it's the same as the first chip we identified.
* M-Systems say that any given DiskOnChip device should only
* contain _one_ type of flash part, although that's not a
* hardware restriction. */
if (doc->mfr) {
if (doc->mfr == mfr && doc->id == id)
return 1; /* This is the same as the first */
else
printk(KERN_WARNING
"Flash chip at floor %d, chip %d is different:\n",
floor, chip);
}
/* Print and store the manufacturer and ID codes. */
for (i = 0; nand_flash_ids[i].name != NULL; i++) {
if (id == nand_flash_ids[i].id) {
/* Try to identify manufacturer */
for (j = 0; nand_manuf_ids[j].id != 0x0; j++) {
if (nand_manuf_ids[j].id == mfr)
break;
}
printk(KERN_INFO
"Flash chip found: Manufacturer ID: %2.2X, "
"Chip ID: %2.2X (%s:%s)\n", mfr, id,
nand_manuf_ids[j].name, nand_flash_ids[i].name);
if (!doc->mfr) {
doc->mfr = mfr;
doc->id = id;
doc->chipshift =
ffs((nand_flash_ids[i].chipsize << 20)) - 1;
doc->page256 = (nand_flash_ids[i].pagesize == 256) ? 1 : 0;
doc->pageadrlen = doc->chipshift > 25 ? 3 : 2;
doc->erasesize =
nand_flash_ids[i].erasesize;
return 1;
}
return 0;
}
}
/* We haven't fully identified the chip. Print as much as we know. */
printk(KERN_WARNING "Unknown flash chip found: %2.2X %2.2X\n",
id, mfr);
printk(KERN_WARNING "Please report to dwmw2@infradead.org\n");
return 0;
}
/* DoC_ScanChips: Find all NAND chips present in a DiskOnChip, and identify them */
static void DoC_ScanChips(struct DiskOnChip *this, int maxchips)
{
int floor, chip;
int numchips[MAX_FLOORS];
int ret = 1;
this->numchips = 0;
this->mfr = 0;
this->id = 0;
/* For each floor, find the number of valid chips it contains */
for (floor = 0; floor < MAX_FLOORS; floor++) {
ret = 1;
numchips[floor] = 0;
for (chip = 0; chip < maxchips && ret != 0; chip++) {
ret = DoC_IdentChip(this, floor, chip);
if (ret) {
numchips[floor]++;
this->numchips++;
}
}
}
/* If there are none at all that we recognise, bail */
if (!this->numchips) {
printk(KERN_NOTICE "No flash chips recognised.\n");
return;
}
/* Allocate an array to hold the information for each chip */
this->chips = kmalloc(sizeof(struct Nand) * this->numchips, GFP_KERNEL);
if (!this->chips) {
printk(KERN_NOTICE "No memory for allocating chip info structures\n");
return;
}
ret = 0;
/* Fill out the chip array with {floor, chipno} for each
* detected chip in the device. */
for (floor = 0; floor < MAX_FLOORS; floor++) {
for (chip = 0; chip < numchips[floor]; chip++) {
this->chips[ret].floor = floor;
this->chips[ret].chip = chip;
this->chips[ret].curadr = 0;
this->chips[ret].curmode = 0x50;
ret++;
}
}
/* Calculate and print the total size of the device */
this->totlen = this->numchips * (1 << this->chipshift);
printk(KERN_INFO "%d flash chips found. Total DiskOnChip size: %ld MiB\n",
this->numchips, this->totlen >> 20);
}
static int DoC2k_is_alias(struct DiskOnChip *doc1, struct DiskOnChip *doc2)
{
int tmp1, tmp2, retval;
if (doc1->physadr == doc2->physadr)
return 1;
/* Use the alias resolution register which was set aside for this
* purpose. If it's value is the same on both chips, they might
* be the same chip, and we write to one and check for a change in
* the other. It's unclear if this register is usuable in the
* DoC 2000 (it's in the Millennium docs), but it seems to work. */
tmp1 = ReadDOC(doc1->virtadr, AliasResolution);
tmp2 = ReadDOC(doc2->virtadr, AliasResolution);
if (tmp1 != tmp2)
return 0;
WriteDOC((tmp1 + 1) % 0xff, doc1->virtadr, AliasResolution);
tmp2 = ReadDOC(doc2->virtadr, AliasResolution);
if (tmp2 == (tmp1 + 1) % 0xff)
retval = 1;
else
retval = 0;
/* Restore register contents. May not be necessary, but do it just to
* be safe. */
WriteDOC(tmp1, doc1->virtadr, AliasResolution);
return retval;
}
/* This routine is found from the docprobe code by symbol_get(),
* which will bump the use count of this module. */
void DoC2k_init(struct mtd_info *mtd)
{
struct DiskOnChip *this = mtd->priv;
struct DiskOnChip *old = NULL;
int maxchips;
/* We must avoid being called twice for the same device. */
if (doc2klist)
old = doc2klist->priv;
while (old) {
if (DoC2k_is_alias(old, this)) {
printk(KERN_NOTICE
"Ignoring DiskOnChip 2000 at 0x%lX - already configured\n",
this->physadr);
iounmap(this->virtadr);
kfree(mtd);
return;
}
if (old->nextdoc)
old = old->nextdoc->priv;
else
old = NULL;
}
switch (this->ChipID) {
case DOC_ChipID_Doc2kTSOP:
mtd->name = "DiskOnChip 2000 TSOP";
this->ioreg = DoC_Mil_CDSN_IO;
/* Pretend it's a Millennium */
this->ChipID = DOC_ChipID_DocMil;
maxchips = MAX_CHIPS;
break;
case DOC_ChipID_Doc2k:
mtd->name = "DiskOnChip 2000";
this->ioreg = DoC_2k_CDSN_IO;
maxchips = MAX_CHIPS;
break;
case DOC_ChipID_DocMil:
mtd->name = "DiskOnChip Millennium";
this->ioreg = DoC_Mil_CDSN_IO;
maxchips = MAX_CHIPS_MIL;
break;
default:
printk("Unknown ChipID 0x%02x\n", this->ChipID);
kfree(mtd);
iounmap(this->virtadr);
return;
}
printk(KERN_NOTICE "%s found at address 0x%lX\n", mtd->name,
this->physadr);
mtd->type = MTD_NANDFLASH;
mtd->flags = MTD_CAP_NANDFLASH;
mtd->size = 0;
mtd->erasesize = 0;
mtd->writesize = 512;
mtd->oobsize = 16;
mtd->owner = THIS_MODULE;
mtd->erase = doc_erase;
mtd->point = NULL;
mtd->unpoint = NULL;
mtd->read = doc_read;
mtd->write = doc_write;
mtd->read_oob = doc_read_oob;
mtd->write_oob = doc_write_oob;
mtd->sync = NULL;
this->totlen = 0;
this->numchips = 0;
this->curfloor = -1;
this->curchip = -1;
mutex_init(&this->lock);
/* Ident all the chips present. */
DoC_ScanChips(this, maxchips);
if (!this->totlen) {
kfree(mtd);
iounmap(this->virtadr);
} else {
this->nextdoc = doc2klist;
doc2klist = mtd;
mtd->size = this->totlen;
mtd->erasesize = this->erasesize;
add_mtd_device(mtd);
return;
}
}
EXPORT_SYMBOL_GPL(DoC2k_init);
static int doc_read(struct mtd_info *mtd, loff_t from, size_t len,
size_t * retlen, u_char * buf)
{
struct DiskOnChip *this = mtd->priv;
void __iomem *docptr = this->virtadr;
struct Nand *mychip;
unsigned char syndrome[6], eccbuf[6];
volatile char dummy;
int i, len256 = 0, ret=0;
size_t left = len;
/* Don't allow read past end of device */
if (from >= this->totlen)
return -EINVAL;
mutex_lock(&this->lock);
*retlen = 0;
while (left) {
len = left;
/* Don't allow a single read to cross a 512-byte block boundary */
if (from + len > ((from | 0x1ff) + 1))
len = ((from | 0x1ff) + 1) - from;
/* The ECC will not be calculated correctly if less than 512 is read */
if (len != 0x200)
printk(KERN_WARNING
"ECC needs a full sector read (adr: %lx size %lx)\n",
(long) from, (long) len);
/* printk("DoC_Read (adr: %lx size %lx)\n", (long) from, (long) len); */
/* Find the chip which is to be used and select it */
mychip = &this->chips[from >> (this->chipshift)];
if (this->curfloor != mychip->floor) {
DoC_SelectFloor(this, mychip->floor);
DoC_SelectChip(this, mychip->chip);
} else if (this->curchip != mychip->chip) {
DoC_SelectChip(this, mychip->chip);
}
this->curfloor = mychip->floor;
this->curchip = mychip->chip;
DoC_Command(this,
(!this->page256
&& (from & 0x100)) ? NAND_CMD_READ1 : NAND_CMD_READ0,
CDSN_CTRL_WP);
DoC_Address(this, ADDR_COLUMN_PAGE, from, CDSN_CTRL_WP,
CDSN_CTRL_ECC_IO);
/* Prime the ECC engine */
WriteDOC(DOC_ECC_RESET, docptr, ECCConf);
WriteDOC(DOC_ECC_EN, docptr, ECCConf);
/* treat crossing 256-byte sector for 2M x 8bits devices */
if (this->page256 && from + len > (from | 0xff) + 1) {
len256 = (from | 0xff) + 1 - from;
DoC_ReadBuf(this, buf, len256);
DoC_Command(this, NAND_CMD_READ0, CDSN_CTRL_WP);
DoC_Address(this, ADDR_COLUMN_PAGE, from + len256,
CDSN_CTRL_WP, CDSN_CTRL_ECC_IO);
}
DoC_ReadBuf(this, &buf[len256], len - len256);
/* Let the caller know we completed it */
*retlen += len;
/* Read the ECC data through the DiskOnChip ECC logic */
/* Note: this will work even with 2M x 8bit devices as */
/* they have 8 bytes of OOB per 256 page. mf. */
DoC_ReadBuf(this, eccbuf, 6);
/* Flush the pipeline */
if (DoC_is_Millennium(this)) {
dummy = ReadDOC(docptr, ECCConf);
dummy = ReadDOC(docptr, ECCConf);
i = ReadDOC(docptr, ECCConf);
} else {
dummy = ReadDOC(docptr, 2k_ECCStatus);
dummy = ReadDOC(docptr, 2k_ECCStatus);
i = ReadDOC(docptr, 2k_ECCStatus);
}
/* Check the ECC Status */
if (i & 0x80) {
int nb_errors;
/* There was an ECC error */
#ifdef ECC_DEBUG
printk(KERN_ERR "DiskOnChip ECC Error: Read at %lx\n", (long)from);
#endif
/* Read the ECC syndrom through the DiskOnChip ECC
logic. These syndrome will be all ZERO when there
is no error */
for (i = 0; i < 6; i++) {
syndrome[i] =
ReadDOC(docptr, ECCSyndrome0 + i);
}
nb_errors = doc_decode_ecc(buf, syndrome);
#ifdef ECC_DEBUG
printk(KERN_ERR "Errors corrected: %x\n", nb_errors);
#endif
if (nb_errors < 0) {
/* We return error, but have actually done the
read. Not that this can be told to
user-space, via sys_read(), but at least
MTD-aware stuff can know about it by
checking *retlen */
ret = -EIO;
}
}
#ifdef PSYCHO_DEBUG
printk(KERN_DEBUG "ECC DATA at %lxB: %2.2X %2.2X %2.2X %2.2X %2.2X %2.2X\n",
(long)from, eccbuf[0], eccbuf[1], eccbuf[2],
eccbuf[3], eccbuf[4], eccbuf[5]);
#endif
/* disable the ECC engine */
WriteDOC(DOC_ECC_DIS, docptr , ECCConf);
/* according to 11.4.1, we need to wait for the busy line
* drop if we read to the end of the page. */
if(0 == ((from + len) & 0x1ff))
{
DoC_WaitReady(this);
}
from += len;
left -= len;
buf += len;
}
mutex_unlock(&this->lock);
return ret;
}
static int doc_write(struct mtd_info *mtd, loff_t to, size_t len,
size_t * retlen, const u_char * buf)
{
struct DiskOnChip *this = mtd->priv;
int di; /* Yes, DI is a hangover from when I was disassembling the binary driver */
void __iomem *docptr = this->virtadr;
unsigned char eccbuf[6];
volatile char dummy;
int len256 = 0;
struct Nand *mychip;
size_t left = len;
int status;
/* Don't allow write past end of device */
if (to >= this->totlen)
return -EINVAL;
mutex_lock(&this->lock);
*retlen = 0;
while (left) {
len = left;
/* Don't allow a single write to cross a 512-byte block boundary */
if (to + len > ((to | 0x1ff) + 1))
len = ((to | 0x1ff) + 1) - to;
/* The ECC will not be calculated correctly if less than 512 is written */
/* DBB-
if (len != 0x200 && eccbuf)
printk(KERN_WARNING
"ECC needs a full sector write (adr: %lx size %lx)\n",
(long) to, (long) len);
-DBB */
/* printk("DoC_Write (adr: %lx size %lx)\n", (long) to, (long) len); */
/* Find the chip which is to be used and select it */
mychip = &this->chips[to >> (this->chipshift)];
if (this->curfloor != mychip->floor) {
DoC_SelectFloor(this, mychip->floor);
DoC_SelectChip(this, mychip->chip);
} else if (this->curchip != mychip->chip) {
DoC_SelectChip(this, mychip->chip);
}
this->curfloor = mychip->floor;
this->curchip = mychip->chip;
/* Set device to main plane of flash */
DoC_Command(this, NAND_CMD_RESET, CDSN_CTRL_WP);
DoC_Command(this,
(!this->page256
&& (to & 0x100)) ? NAND_CMD_READ1 : NAND_CMD_READ0,
CDSN_CTRL_WP);
DoC_Command(this, NAND_CMD_SEQIN, 0);
DoC_Address(this, ADDR_COLUMN_PAGE, to, 0, CDSN_CTRL_ECC_IO);
/* Prime the ECC engine */
WriteDOC(DOC_ECC_RESET, docptr, ECCConf);
WriteDOC(DOC_ECC_EN | DOC_ECC_RW, docptr, ECCConf);
/* treat crossing 256-byte sector for 2M x 8bits devices */
if (this->page256 && to + len > (to | 0xff) + 1) {
len256 = (to | 0xff) + 1 - to;
DoC_WriteBuf(this, buf, len256);
DoC_Command(this, NAND_CMD_PAGEPROG, 0);
DoC_Command(this, NAND_CMD_STATUS, CDSN_CTRL_WP);
/* There's an implicit DoC_WaitReady() in DoC_Command */
dummy = ReadDOC(docptr, CDSNSlowIO);
DoC_Delay(this, 2);
if (ReadDOC_(docptr, this->ioreg) & 1) {
printk(KERN_ERR "Error programming flash\n");
/* Error in programming */
*retlen = 0;
mutex_unlock(&this->lock);
return -EIO;
}
DoC_Command(this, NAND_CMD_SEQIN, 0);
DoC_Address(this, ADDR_COLUMN_PAGE, to + len256, 0,
CDSN_CTRL_ECC_IO);
}
DoC_WriteBuf(this, &buf[len256], len - len256);
WriteDOC(CDSN_CTRL_ECC_IO | CDSN_CTRL_CE, docptr, CDSNControl);
if (DoC_is_Millennium(this)) {
WriteDOC(0, docptr, NOP);
WriteDOC(0, docptr, NOP);
WriteDOC(0, docptr, NOP);
} else {
WriteDOC_(0, docptr, this->ioreg);
WriteDOC_(0, docptr, this->ioreg);
WriteDOC_(0, docptr, this->ioreg);
}
WriteDOC(CDSN_CTRL_ECC_IO | CDSN_CTRL_FLASH_IO | CDSN_CTRL_CE, docptr,
CDSNControl);
/* Read the ECC data through the DiskOnChip ECC logic */
for (di = 0; di < 6; di++) {
eccbuf[di] = ReadDOC(docptr, ECCSyndrome0 + di);
}
/* Reset the ECC engine */
WriteDOC(DOC_ECC_DIS, docptr, ECCConf);
#ifdef PSYCHO_DEBUG
printk
("OOB data at %lx is %2.2X %2.2X %2.2X %2.2X %2.2X %2.2X\n",
(long) to, eccbuf[0], eccbuf[1], eccbuf[2], eccbuf[3],
eccbuf[4], eccbuf[5]);
#endif
DoC_Command(this, NAND_CMD_PAGEPROG, 0);
DoC_Command(this, NAND_CMD_STATUS, CDSN_CTRL_WP);
/* There's an implicit DoC_WaitReady() in DoC_Command */
if (DoC_is_Millennium(this)) {
ReadDOC(docptr, ReadPipeInit);
status = ReadDOC(docptr, LastDataRead);
} else {
dummy = ReadDOC(docptr, CDSNSlowIO);
DoC_Delay(this, 2);
status = ReadDOC_(docptr, this->ioreg);
}
if (status & 1) {
printk(KERN_ERR "Error programming flash\n");
/* Error in programming */
*retlen = 0;
mutex_unlock(&this->lock);
return -EIO;
}
/* Let the caller know we completed it */
*retlen += len;
{
unsigned char x[8];
size_t dummy;
int ret;
/* Write the ECC data to flash */
for (di=0; di<6; di++)
x[di] = eccbuf[di];
x[6]=0x55;
x[7]=0x55;
ret = doc_write_oob_nolock(mtd, to, 8, &dummy, x);
if (ret) {
mutex_unlock(&this->lock);
return ret;
}
}
to += len;
left -= len;
buf += len;
}
mutex_unlock(&this->lock);
return 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
static int doc_read_oob(struct mtd_info *mtd, loff_t ofs,
struct mtd_oob_ops *ops)
{
struct DiskOnChip *this = mtd->priv;
int len256 = 0, ret;
struct Nand *mychip;
[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
uint8_t *buf = ops->oobbuf;
size_t len = ops->len;
BUG_ON(ops->mode != MTD_OOB_PLACE);
ofs += ops->ooboffs;
mutex_lock(&this->lock);
mychip = &this->chips[ofs >> this->chipshift];
if (this->curfloor != mychip->floor) {
DoC_SelectFloor(this, mychip->floor);
DoC_SelectChip(this, mychip->chip);
} else if (this->curchip != mychip->chip) {
DoC_SelectChip(this, mychip->chip);
}
this->curfloor = mychip->floor;
this->curchip = mychip->chip;
/* update address for 2M x 8bit devices. OOB starts on the second */
/* page to maintain compatibility with doc_read_ecc. */
if (this->page256) {
if (!(ofs & 0x8))
ofs += 0x100;
else
ofs -= 0x8;
}
DoC_Command(this, NAND_CMD_READOOB, CDSN_CTRL_WP);
DoC_Address(this, ADDR_COLUMN_PAGE, ofs, CDSN_CTRL_WP, 0);
/* treat crossing 8-byte OOB data for 2M x 8bit devices */
/* Note: datasheet says it should automaticaly wrap to the */
/* next OOB block, but it didn't work here. mf. */
if (this->page256 && ofs + len > (ofs | 0x7) + 1) {
len256 = (ofs | 0x7) + 1 - ofs;
DoC_ReadBuf(this, buf, len256);
DoC_Command(this, NAND_CMD_READOOB, CDSN_CTRL_WP);
DoC_Address(this, ADDR_COLUMN_PAGE, ofs & (~0x1ff),
CDSN_CTRL_WP, 0);
}
DoC_ReadBuf(this, &buf[len256], len - len256);
[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->retlen = len;
/* Reading the full OOB data drops us off of the end of the page,
* causing the flash device to go into busy mode, so we need
* to wait until ready 11.4.1 and Toshiba TC58256FT docs */
ret = DoC_WaitReady(this);
mutex_unlock(&this->lock);
return ret;
}
static int doc_write_oob_nolock(struct mtd_info *mtd, loff_t ofs, size_t len,
size_t * retlen, const u_char * buf)
{
struct DiskOnChip *this = mtd->priv;
int len256 = 0;
void __iomem *docptr = this->virtadr;
struct Nand *mychip = &this->chips[ofs >> this->chipshift];
volatile int dummy;
int status;
// printk("doc_write_oob(%lx, %d): %2.2X %2.2X %2.2X %2.2X ... %2.2X %2.2X .. %2.2X %2.2X\n",(long)ofs, len,
// buf[0], buf[1], buf[2], buf[3], buf[8], buf[9], buf[14],buf[15]);
/* Find the chip which is to be used and select it */
if (this->curfloor != mychip->floor) {
DoC_SelectFloor(this, mychip->floor);
DoC_SelectChip(this, mychip->chip);
} else if (this->curchip != mychip->chip) {
DoC_SelectChip(this, mychip->chip);
}
this->curfloor = mychip->floor;
this->curchip = mychip->chip;
/* disable the ECC engine */
WriteDOC (DOC_ECC_RESET, docptr, ECCConf);
WriteDOC (DOC_ECC_DIS, docptr, ECCConf);
/* Reset the chip, see Software Requirement 11.4 item 1. */
DoC_Command(this, NAND_CMD_RESET, CDSN_CTRL_WP);
/* issue the Read2 command to set the pointer to the Spare Data Area. */
DoC_Command(this, NAND_CMD_READOOB, CDSN_CTRL_WP);
/* update address for 2M x 8bit devices. OOB starts on the second */
/* page to maintain compatibility with doc_read_ecc. */
if (this->page256) {
if (!(ofs & 0x8))
ofs += 0x100;
else
ofs -= 0x8;
}
/* issue the Serial Data In command to initial the Page Program process */
DoC_Command(this, NAND_CMD_SEQIN, 0);
DoC_Address(this, ADDR_COLUMN_PAGE, ofs, 0, 0);
/* treat crossing 8-byte OOB data for 2M x 8bit devices */
/* Note: datasheet says it should automaticaly wrap to the */
/* next OOB block, but it didn't work here. mf. */
if (this->page256 && ofs + len > (ofs | 0x7) + 1) {
len256 = (ofs | 0x7) + 1 - ofs;
DoC_WriteBuf(this, buf, len256);
DoC_Command(this, NAND_CMD_PAGEPROG, 0);
DoC_Command(this, NAND_CMD_STATUS, 0);
/* DoC_WaitReady() is implicit in DoC_Command */
if (DoC_is_Millennium(this)) {
ReadDOC(docptr, ReadPipeInit);
status = ReadDOC(docptr, LastDataRead);
} else {
dummy = ReadDOC(docptr, CDSNSlowIO);
DoC_Delay(this, 2);
status = ReadDOC_(docptr, this->ioreg);
}
if (status & 1) {
printk(KERN_ERR "Error programming oob data\n");
/* There was an error */
*retlen = 0;
return -EIO;
}
DoC_Command(this, NAND_CMD_SEQIN, 0);
DoC_Address(this, ADDR_COLUMN_PAGE, ofs & (~0x1ff), 0, 0);
}
DoC_WriteBuf(this, &buf[len256], len - len256);
DoC_Command(this, NAND_CMD_PAGEPROG, 0);
DoC_Command(this, NAND_CMD_STATUS, 0);
/* DoC_WaitReady() is implicit in DoC_Command */
if (DoC_is_Millennium(this)) {
ReadDOC(docptr, ReadPipeInit);
status = ReadDOC(docptr, LastDataRead);
} else {
dummy = ReadDOC(docptr, CDSNSlowIO);
DoC_Delay(this, 2);
status = ReadDOC_(docptr, this->ioreg);
}
if (status & 1) {
printk(KERN_ERR "Error programming oob data\n");
/* There was an error */
*retlen = 0;
return -EIO;
}
*retlen = len;
return 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
static int doc_write_oob(struct mtd_info *mtd, loff_t ofs,
struct mtd_oob_ops *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
struct DiskOnChip *this = mtd->priv;
int 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
BUG_ON(ops->mode != MTD_OOB_PLACE);
[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
mutex_lock(&this->lock);
ret = doc_write_oob_nolock(mtd, ofs + ops->ooboffs, ops->len,
&ops->retlen, ops->oobbuf);
mutex_unlock(&this->lock);
return ret;
}
static int doc_erase(struct mtd_info *mtd, struct erase_info *instr)
{
struct DiskOnChip *this = mtd->priv;
__u32 ofs = instr->addr;
__u32 len = instr->len;
volatile int dummy;
void __iomem *docptr = this->virtadr;
struct Nand *mychip;
int status;
mutex_lock(&this->lock);
if (ofs & (mtd->erasesize-1) || len & (mtd->erasesize-1)) {
mutex_unlock(&this->lock);
return -EINVAL;
}
instr->state = MTD_ERASING;
/* FIXME: Do this in the background. Use timers or schedule_task() */
while(len) {
mychip = &this->chips[ofs >> this->chipshift];
if (this->curfloor != mychip->floor) {
DoC_SelectFloor(this, mychip->floor);
DoC_SelectChip(this, mychip->chip);
} else if (this->curchip != mychip->chip) {
DoC_SelectChip(this, mychip->chip);
}
this->curfloor = mychip->floor;
this->curchip = mychip->chip;
DoC_Command(this, NAND_CMD_ERASE1, 0);
DoC_Address(this, ADDR_PAGE, ofs, 0, 0);
DoC_Command(this, NAND_CMD_ERASE2, 0);
DoC_Command(this, NAND_CMD_STATUS, CDSN_CTRL_WP);
if (DoC_is_Millennium(this)) {
ReadDOC(docptr, ReadPipeInit);
status = ReadDOC(docptr, LastDataRead);
} else {
dummy = ReadDOC(docptr, CDSNSlowIO);
DoC_Delay(this, 2);
status = ReadDOC_(docptr, this->ioreg);
}
if (status & 1) {
printk(KERN_ERR "Error erasing at 0x%x\n", ofs);
/* There was an error */
instr->state = MTD_ERASE_FAILED;
goto callback;
}
ofs += mtd->erasesize;
len -= mtd->erasesize;
}
instr->state = MTD_ERASE_DONE;
callback:
mtd_erase_callback(instr);
mutex_unlock(&this->lock);
return 0;
}
/****************************************************************************
*
* Module stuff
*
****************************************************************************/
static void __exit cleanup_doc2000(void)
{
struct mtd_info *mtd;
struct DiskOnChip *this;
while ((mtd = doc2klist)) {
this = mtd->priv;
doc2klist = this->nextdoc;
del_mtd_device(mtd);
iounmap(this->virtadr);
kfree(this->chips);
kfree(mtd);
}
}
module_exit(cleanup_doc2000);
MODULE_LICENSE("GPL");
MODULE_AUTHOR("David Woodhouse <dwmw2@infradead.org> et al.");
MODULE_DESCRIPTION("MTD driver for DiskOnChip 2000 and Millennium");