diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-07 18:49:45 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-07 18:49:45 +0000 |
commit | 2c3c1048746a4622d8c89a29670120dc8fab93c4 (patch) | |
tree | 848558de17fb3008cdf4d861b01ac7781903ce39 /drivers/mtd/nand/raw/diskonchip.c | |
parent | Initial commit. (diff) | |
download | linux-2c3c1048746a4622d8c89a29670120dc8fab93c4.tar.xz linux-2c3c1048746a4622d8c89a29670120dc8fab93c4.zip |
Adding upstream version 6.1.76.upstream/6.1.76
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'drivers/mtd/nand/raw/diskonchip.c')
-rw-r--r-- | drivers/mtd/nand/raw/diskonchip.c | 1579 |
1 files changed, 1579 insertions, 0 deletions
diff --git a/drivers/mtd/nand/raw/diskonchip.c b/drivers/mtd/nand/raw/diskonchip.c new file mode 100644 index 000000000..5d2ddb037 --- /dev/null +++ b/drivers/mtd/nand/raw/diskonchip.c @@ -0,0 +1,1579 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * (C) 2003 Red Hat, Inc. + * (C) 2004 Dan Brown <dan_brown@ieee.org> + * (C) 2004 Kalev Lember <kalev@smartlink.ee> + * + * Author: David Woodhouse <dwmw2@infradead.org> + * Additional Diskonchip 2000 and Millennium support by Dan Brown <dan_brown@ieee.org> + * Diskonchip Millennium Plus support by Kalev Lember <kalev@smartlink.ee> + * + * Error correction code lifted from the old docecc code + * Author: Fabrice Bellard (fabrice.bellard@netgem.com) + * Copyright (C) 2000 Netgem S.A. + * converted to the generic Reed-Solomon library by Thomas Gleixner <tglx@linutronix.de> + * + * Interface to generic NAND code for M-Systems DiskOnChip devices + */ + +#include <linux/kernel.h> +#include <linux/init.h> +#include <linux/sched.h> +#include <linux/delay.h> +#include <linux/rslib.h> +#include <linux/moduleparam.h> +#include <linux/slab.h> +#include <linux/io.h> + +#include <linux/mtd/mtd.h> +#include <linux/mtd/rawnand.h> +#include <linux/mtd/doc2000.h> +#include <linux/mtd/partitions.h> +#include <linux/mtd/inftl.h> +#include <linux/module.h> + +/* Where to look for the devices? */ +#ifndef CONFIG_MTD_NAND_DISKONCHIP_PROBE_ADDRESS +#define CONFIG_MTD_NAND_DISKONCHIP_PROBE_ADDRESS 0 +#endif + +static unsigned long doc_locations[] __initdata = { +#if defined (__alpha__) || defined(__i386__) || defined(__x86_64__) +#ifdef CONFIG_MTD_NAND_DISKONCHIP_PROBE_HIGH + 0xfffc8000, 0xfffca000, 0xfffcc000, 0xfffce000, + 0xfffd0000, 0xfffd2000, 0xfffd4000, 0xfffd6000, + 0xfffd8000, 0xfffda000, 0xfffdc000, 0xfffde000, + 0xfffe0000, 0xfffe2000, 0xfffe4000, 0xfffe6000, + 0xfffe8000, 0xfffea000, 0xfffec000, 0xfffee000, +#else + 0xc8000, 0xca000, 0xcc000, 0xce000, + 0xd0000, 0xd2000, 0xd4000, 0xd6000, + 0xd8000, 0xda000, 0xdc000, 0xde000, + 0xe0000, 0xe2000, 0xe4000, 0xe6000, + 0xe8000, 0xea000, 0xec000, 0xee000, +#endif +#endif + 0xffffffff }; + +static struct mtd_info *doclist = NULL; + +struct doc_priv { + struct nand_controller base; + void __iomem *virtadr; + unsigned long physadr; + u_char ChipID; + u_char CDSNControl; + int chips_per_floor; /* The number of chips detected on each floor */ + int curfloor; + int curchip; + int mh0_page; + int mh1_page; + struct rs_control *rs_decoder; + struct mtd_info *nextdoc; + bool supports_32b_reads; + + /* Handle the last stage of initialization (BBT scan, partitioning) */ + int (*late_init)(struct mtd_info *mtd); +}; + +/* This is the ecc value computed by the HW ecc generator upon writing an empty + page, one with all 0xff for data. */ +static u_char empty_write_ecc[6] = { 0x4b, 0x00, 0xe2, 0x0e, 0x93, 0xf7 }; + +#define INFTL_BBT_RESERVED_BLOCKS 4 + +#define DoC_is_MillenniumPlus(doc) ((doc)->ChipID == DOC_ChipID_DocMilPlus16 || (doc)->ChipID == DOC_ChipID_DocMilPlus32) +#define DoC_is_Millennium(doc) ((doc)->ChipID == DOC_ChipID_DocMil) +#define DoC_is_2000(doc) ((doc)->ChipID == DOC_ChipID_Doc2k) + +static int debug = 0; +module_param(debug, int, 0); + +static int try_dword = 1; +module_param(try_dword, int, 0); + +static int no_ecc_failures = 0; +module_param(no_ecc_failures, int, 0); + +static int no_autopart = 0; +module_param(no_autopart, int, 0); + +static int show_firmware_partition = 0; +module_param(show_firmware_partition, int, 0); + +#ifdef CONFIG_MTD_NAND_DISKONCHIP_BBTWRITE +static int inftl_bbt_write = 1; +#else +static int inftl_bbt_write = 0; +#endif +module_param(inftl_bbt_write, int, 0); + +static unsigned long doc_config_location = CONFIG_MTD_NAND_DISKONCHIP_PROBE_ADDRESS; +module_param(doc_config_location, ulong, 0); +MODULE_PARM_DESC(doc_config_location, "Physical memory address at which to probe for DiskOnChip"); + +/* Sector size for HW ECC */ +#define SECTOR_SIZE 512 +/* The sector bytes are packed into NB_DATA 10 bit words */ +#define NB_DATA (((SECTOR_SIZE + 1) * 8 + 6) / 10) +/* Number of roots */ +#define NROOTS 4 +/* First consective root */ +#define FCR 510 +/* Number of symbols */ +#define NN 1023 + +/* + * The HW decoder in the DoC ASIC's provides us a error syndrome, + * which we must convert to a standard syndrome usable by the generic + * Reed-Solomon library code. + * + * Fabrice Bellard figured this out in the old docecc code. I added + * some comments, improved a minor bit and converted it to make use + * of the generic Reed-Solomon library. tglx + */ +static int doc_ecc_decode(struct rs_control *rs, uint8_t *data, uint8_t *ecc) +{ + int i, j, nerr, errpos[8]; + uint8_t parity; + uint16_t ds[4], s[5], tmp, errval[8], syn[4]; + struct rs_codec *cd = rs->codec; + + memset(syn, 0, sizeof(syn)); + /* Convert the ecc bytes into words */ + ds[0] = ((ecc[4] & 0xff) >> 0) | ((ecc[5] & 0x03) << 8); + ds[1] = ((ecc[5] & 0xfc) >> 2) | ((ecc[2] & 0x0f) << 6); + ds[2] = ((ecc[2] & 0xf0) >> 4) | ((ecc[3] & 0x3f) << 4); + ds[3] = ((ecc[3] & 0xc0) >> 6) | ((ecc[0] & 0xff) << 2); + parity = ecc[1]; + + /* Initialize the syndrome buffer */ + for (i = 0; i < NROOTS; i++) + s[i] = ds[0]; + /* + * Evaluate + * s[i] = ds[3]x^3 + ds[2]x^2 + ds[1]x^1 + ds[0] + * where x = alpha^(FCR + i) + */ + for (j = 1; j < NROOTS; j++) { + if (ds[j] == 0) + continue; + tmp = cd->index_of[ds[j]]; + for (i = 0; i < NROOTS; i++) + s[i] ^= cd->alpha_to[rs_modnn(cd, tmp + (FCR + i) * j)]; + } + + /* Calc syn[i] = s[i] / alpha^(v + i) */ + for (i = 0; i < NROOTS; i++) { + if (s[i]) + syn[i] = rs_modnn(cd, cd->index_of[s[i]] + (NN - FCR - i)); + } + /* Call the decoder library */ + nerr = decode_rs16(rs, NULL, NULL, 1019, syn, 0, errpos, 0, errval); + + /* Incorrectable errors ? */ + if (nerr < 0) + return nerr; + + /* + * Correct the errors. The bitpositions are a bit of magic, + * but they are given by the design of the de/encoder circuit + * in the DoC ASIC's. + */ + for (i = 0; i < nerr; i++) { + int index, bitpos, pos = 1015 - errpos[i]; + uint8_t val; + if (pos >= NB_DATA && pos < 1019) + continue; + if (pos < NB_DATA) { + /* extract bit position (MSB first) */ + pos = 10 * (NB_DATA - 1 - pos) - 6; + /* now correct the following 10 bits. At most two bytes + can be modified since pos is even */ + index = (pos >> 3) ^ 1; + bitpos = pos & 7; + if ((index >= 0 && index < SECTOR_SIZE) || index == (SECTOR_SIZE + 1)) { + val = (uint8_t) (errval[i] >> (2 + bitpos)); + parity ^= val; + if (index < SECTOR_SIZE) + data[index] ^= val; + } + index = ((pos >> 3) + 1) ^ 1; + bitpos = (bitpos + 10) & 7; + if (bitpos == 0) + bitpos = 8; + if ((index >= 0 && index < SECTOR_SIZE) || index == (SECTOR_SIZE + 1)) { + val = (uint8_t) (errval[i] << (8 - bitpos)); + parity ^= val; + if (index < SECTOR_SIZE) + data[index] ^= val; + } + } + } + /* If the parity is wrong, no rescue possible */ + return parity ? -EBADMSG : nerr; +} + +static void DoC_Delay(struct doc_priv *doc, unsigned short cycles) +{ + volatile char __always_unused dummy; + int i; + + for (i = 0; i < cycles; i++) { + if (DoC_is_Millennium(doc)) + dummy = ReadDOC(doc->virtadr, NOP); + else if (DoC_is_MillenniumPlus(doc)) + dummy = ReadDOC(doc->virtadr, Mplus_NOP); + else + dummy = ReadDOC(doc->virtadr, DOCStatus); + } + +} + +#define CDSN_CTRL_FR_B_MASK (CDSN_CTRL_FR_B0 | CDSN_CTRL_FR_B1) + +/* DOC_WaitReady: Wait for RDY line to be asserted by the flash chip */ +static int _DoC_WaitReady(struct doc_priv *doc) +{ + void __iomem *docptr = doc->virtadr; + unsigned long timeo = jiffies + (HZ * 10); + + if (debug) + printk("_DoC_WaitReady...\n"); + /* Out-of-line routine to wait for chip response */ + if (DoC_is_MillenniumPlus(doc)) { + while ((ReadDOC(docptr, Mplus_FlashControl) & CDSN_CTRL_FR_B_MASK) != CDSN_CTRL_FR_B_MASK) { + if (time_after(jiffies, timeo)) { + printk("_DoC_WaitReady timed out.\n"); + return -EIO; + } + udelay(1); + cond_resched(); + } + } else { + while (!(ReadDOC(docptr, CDSNControl) & CDSN_CTRL_FR_B)) { + if (time_after(jiffies, timeo)) { + printk("_DoC_WaitReady timed out.\n"); + return -EIO; + } + udelay(1); + cond_resched(); + } + } + + return 0; +} + +static inline int DoC_WaitReady(struct doc_priv *doc) +{ + void __iomem *docptr = doc->virtadr; + int ret = 0; + + if (DoC_is_MillenniumPlus(doc)) { + DoC_Delay(doc, 4); + + if ((ReadDOC(docptr, Mplus_FlashControl) & CDSN_CTRL_FR_B_MASK) != CDSN_CTRL_FR_B_MASK) + /* Call the out-of-line routine to wait */ + ret = _DoC_WaitReady(doc); + } else { + DoC_Delay(doc, 4); + + if (!(ReadDOC(docptr, CDSNControl) & CDSN_CTRL_FR_B)) + /* Call the out-of-line routine to wait */ + ret = _DoC_WaitReady(doc); + DoC_Delay(doc, 2); + } + + if (debug) + printk("DoC_WaitReady OK\n"); + return ret; +} + +static void doc2000_write_byte(struct nand_chip *this, u_char datum) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + + if (debug) + printk("write_byte %02x\n", datum); + WriteDOC(datum, docptr, CDSNSlowIO); + WriteDOC(datum, docptr, 2k_CDSN_IO); +} + +static void doc2000_writebuf(struct nand_chip *this, const u_char *buf, + int len) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + int i; + if (debug) + printk("writebuf of %d bytes: ", len); + for (i = 0; i < len; i++) { + WriteDOC_(buf[i], docptr, DoC_2k_CDSN_IO + i); + if (debug && i < 16) + printk("%02x ", buf[i]); + } + if (debug) + printk("\n"); +} + +static void doc2000_readbuf(struct nand_chip *this, u_char *buf, int len) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + u32 *buf32 = (u32 *)buf; + int i; + + if (debug) + printk("readbuf of %d bytes: ", len); + + if (!doc->supports_32b_reads || + ((((unsigned long)buf) | len) & 3)) { + for (i = 0; i < len; i++) + buf[i] = ReadDOC(docptr, 2k_CDSN_IO + i); + } else { + for (i = 0; i < len / 4; i++) + buf32[i] = readl(docptr + DoC_2k_CDSN_IO + i); + } +} + +/* + * We need our own readid() here because it's called before the NAND chip + * has been initialized, and calling nand_op_readid() would lead to a NULL + * pointer exception when dereferencing the NAND timings. + */ +static void doc200x_readid(struct nand_chip *this, unsigned int cs, u8 *id) +{ + u8 addr = 0; + struct nand_op_instr instrs[] = { + NAND_OP_CMD(NAND_CMD_READID, 0), + NAND_OP_ADDR(1, &addr, 50), + NAND_OP_8BIT_DATA_IN(2, id, 0), + }; + + struct nand_operation op = NAND_OPERATION(cs, instrs); + + if (!id) + op.ninstrs--; + + this->controller->ops->exec_op(this, &op, false); +} + +static uint16_t __init doc200x_ident_chip(struct mtd_info *mtd, int nr) +{ + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + uint16_t ret; + u8 id[2]; + + doc200x_readid(this, nr, id); + + ret = ((u16)id[0] << 8) | id[1]; + + if (doc->ChipID == DOC_ChipID_Doc2k && try_dword && !nr) { + /* First chip probe. See if we get same results by 32-bit access */ + union { + uint32_t dword; + uint8_t byte[4]; + } ident; + void __iomem *docptr = doc->virtadr; + + doc200x_readid(this, nr, NULL); + + ident.dword = readl(docptr + DoC_2k_CDSN_IO); + if (((ident.byte[0] << 8) | ident.byte[1]) == ret) { + pr_info("DiskOnChip 2000 responds to DWORD access\n"); + doc->supports_32b_reads = true; + } + } + + return ret; +} + +static void __init doc2000_count_chips(struct mtd_info *mtd) +{ + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + uint16_t mfrid; + int i; + + /* Max 4 chips per floor on DiskOnChip 2000 */ + doc->chips_per_floor = 4; + + /* Find out what the first chip is */ + mfrid = doc200x_ident_chip(mtd, 0); + + /* Find how many chips in each floor. */ + for (i = 1; i < 4; i++) { + if (doc200x_ident_chip(mtd, i) != mfrid) + break; + } + doc->chips_per_floor = i; + pr_debug("Detected %d chips per floor.\n", i); +} + +static void doc2001_write_byte(struct nand_chip *this, u_char datum) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + + WriteDOC(datum, docptr, CDSNSlowIO); + WriteDOC(datum, docptr, Mil_CDSN_IO); + WriteDOC(datum, docptr, WritePipeTerm); +} + +static void doc2001_writebuf(struct nand_chip *this, const u_char *buf, int len) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + int i; + + for (i = 0; i < len; i++) + WriteDOC_(buf[i], docptr, DoC_Mil_CDSN_IO + i); + /* Terminate write pipeline */ + WriteDOC(0x00, docptr, WritePipeTerm); +} + +static void doc2001_readbuf(struct nand_chip *this, u_char *buf, int len) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + int i; + + /* Start read pipeline */ + ReadDOC(docptr, ReadPipeInit); + + for (i = 0; i < len - 1; i++) + buf[i] = ReadDOC(docptr, Mil_CDSN_IO + (i & 0xff)); + + /* Terminate read pipeline */ + buf[i] = ReadDOC(docptr, LastDataRead); +} + +static void doc2001plus_writebuf(struct nand_chip *this, const u_char *buf, int len) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + int i; + + if (debug) + printk("writebuf of %d bytes: ", len); + for (i = 0; i < len; i++) { + WriteDOC_(buf[i], docptr, DoC_Mil_CDSN_IO + i); + if (debug && i < 16) + printk("%02x ", buf[i]); + } + if (debug) + printk("\n"); +} + +static void doc2001plus_readbuf(struct nand_chip *this, u_char *buf, int len) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + int i; + + if (debug) + printk("readbuf of %d bytes: ", len); + + /* Start read pipeline */ + ReadDOC(docptr, Mplus_ReadPipeInit); + ReadDOC(docptr, Mplus_ReadPipeInit); + + for (i = 0; i < len - 2; i++) { + buf[i] = ReadDOC(docptr, Mil_CDSN_IO); + if (debug && i < 16) + printk("%02x ", buf[i]); + } + + /* Terminate read pipeline */ + if (len >= 2) { + buf[len - 2] = ReadDOC(docptr, Mplus_LastDataRead); + if (debug && i < 16) + printk("%02x ", buf[len - 2]); + } + + buf[len - 1] = ReadDOC(docptr, Mplus_LastDataRead); + if (debug && i < 16) + printk("%02x ", buf[len - 1]); + if (debug) + printk("\n"); +} + +static void doc200x_write_control(struct doc_priv *doc, u8 value) +{ + WriteDOC(value, doc->virtadr, CDSNControl); + /* 11.4.3 -- 4 NOPs after CSDNControl write */ + DoC_Delay(doc, 4); +} + +static void doc200x_exec_instr(struct nand_chip *this, + const struct nand_op_instr *instr) +{ + struct doc_priv *doc = nand_get_controller_data(this); + unsigned int i; + + switch (instr->type) { + case NAND_OP_CMD_INSTR: + doc200x_write_control(doc, CDSN_CTRL_CE | CDSN_CTRL_CLE); + doc2000_write_byte(this, instr->ctx.cmd.opcode); + break; + + case NAND_OP_ADDR_INSTR: + doc200x_write_control(doc, CDSN_CTRL_CE | CDSN_CTRL_ALE); + for (i = 0; i < instr->ctx.addr.naddrs; i++) { + u8 addr = instr->ctx.addr.addrs[i]; + + if (DoC_is_2000(doc)) + doc2000_write_byte(this, addr); + else + doc2001_write_byte(this, addr); + } + break; + + case NAND_OP_DATA_IN_INSTR: + doc200x_write_control(doc, CDSN_CTRL_CE); + if (DoC_is_2000(doc)) + doc2000_readbuf(this, instr->ctx.data.buf.in, + instr->ctx.data.len); + else + doc2001_readbuf(this, instr->ctx.data.buf.in, + instr->ctx.data.len); + break; + + case NAND_OP_DATA_OUT_INSTR: + doc200x_write_control(doc, CDSN_CTRL_CE); + if (DoC_is_2000(doc)) + doc2000_writebuf(this, instr->ctx.data.buf.out, + instr->ctx.data.len); + else + doc2001_writebuf(this, instr->ctx.data.buf.out, + instr->ctx.data.len); + break; + + case NAND_OP_WAITRDY_INSTR: + DoC_WaitReady(doc); + break; + } + + if (instr->delay_ns) + ndelay(instr->delay_ns); +} + +static int doc200x_exec_op(struct nand_chip *this, + const struct nand_operation *op, + bool check_only) +{ + struct doc_priv *doc = nand_get_controller_data(this); + unsigned int i; + + if (check_only) + return true; + + doc->curchip = op->cs % doc->chips_per_floor; + doc->curfloor = op->cs / doc->chips_per_floor; + + WriteDOC(doc->curfloor, doc->virtadr, FloorSelect); + WriteDOC(doc->curchip, doc->virtadr, CDSNDeviceSelect); + + /* Assert CE pin */ + doc200x_write_control(doc, CDSN_CTRL_CE); + + for (i = 0; i < op->ninstrs; i++) + doc200x_exec_instr(this, &op->instrs[i]); + + /* De-assert CE pin */ + doc200x_write_control(doc, 0); + + return 0; +} + +static void doc2001plus_write_pipe_term(struct doc_priv *doc) +{ + WriteDOC(0x00, doc->virtadr, Mplus_WritePipeTerm); + WriteDOC(0x00, doc->virtadr, Mplus_WritePipeTerm); +} + +static void doc2001plus_exec_instr(struct nand_chip *this, + const struct nand_op_instr *instr) +{ + struct doc_priv *doc = nand_get_controller_data(this); + unsigned int i; + + switch (instr->type) { + case NAND_OP_CMD_INSTR: + WriteDOC(instr->ctx.cmd.opcode, doc->virtadr, Mplus_FlashCmd); + doc2001plus_write_pipe_term(doc); + break; + + case NAND_OP_ADDR_INSTR: + for (i = 0; i < instr->ctx.addr.naddrs; i++) { + u8 addr = instr->ctx.addr.addrs[i]; + + WriteDOC(addr, doc->virtadr, Mplus_FlashAddress); + } + doc2001plus_write_pipe_term(doc); + /* deassert ALE */ + WriteDOC(0, doc->virtadr, Mplus_FlashControl); + break; + + case NAND_OP_DATA_IN_INSTR: + doc2001plus_readbuf(this, instr->ctx.data.buf.in, + instr->ctx.data.len); + break; + case NAND_OP_DATA_OUT_INSTR: + doc2001plus_writebuf(this, instr->ctx.data.buf.out, + instr->ctx.data.len); + doc2001plus_write_pipe_term(doc); + break; + case NAND_OP_WAITRDY_INSTR: + DoC_WaitReady(doc); + break; + } + + if (instr->delay_ns) + ndelay(instr->delay_ns); +} + +static int doc2001plus_exec_op(struct nand_chip *this, + const struct nand_operation *op, + bool check_only) +{ + struct doc_priv *doc = nand_get_controller_data(this); + unsigned int i; + + if (check_only) + return true; + + doc->curchip = op->cs % doc->chips_per_floor; + doc->curfloor = op->cs / doc->chips_per_floor; + + /* Assert ChipEnable and deassert WriteProtect */ + WriteDOC(DOC_FLASH_CE, doc->virtadr, Mplus_FlashSelect); + + for (i = 0; i < op->ninstrs; i++) + doc2001plus_exec_instr(this, &op->instrs[i]); + + /* De-assert ChipEnable */ + WriteDOC(0, doc->virtadr, Mplus_FlashSelect); + + return 0; +} + +static void doc200x_enable_hwecc(struct nand_chip *this, int mode) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + + /* Prime the ECC engine */ + switch (mode) { + case NAND_ECC_READ: + WriteDOC(DOC_ECC_RESET, docptr, ECCConf); + WriteDOC(DOC_ECC_EN, docptr, ECCConf); + break; + case NAND_ECC_WRITE: + WriteDOC(DOC_ECC_RESET, docptr, ECCConf); + WriteDOC(DOC_ECC_EN | DOC_ECC_RW, docptr, ECCConf); + break; + } +} + +static void doc2001plus_enable_hwecc(struct nand_chip *this, int mode) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + + /* Prime the ECC engine */ + switch (mode) { + case NAND_ECC_READ: + WriteDOC(DOC_ECC_RESET, docptr, Mplus_ECCConf); + WriteDOC(DOC_ECC_EN, docptr, Mplus_ECCConf); + break; + case NAND_ECC_WRITE: + WriteDOC(DOC_ECC_RESET, docptr, Mplus_ECCConf); + WriteDOC(DOC_ECC_EN | DOC_ECC_RW, docptr, Mplus_ECCConf); + break; + } +} + +/* This code is only called on write */ +static int doc200x_calculate_ecc(struct nand_chip *this, const u_char *dat, + unsigned char *ecc_code) +{ + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + int i; + int __always_unused emptymatch = 1; + + /* flush the pipeline */ + if (DoC_is_2000(doc)) { + WriteDOC(doc->CDSNControl & ~CDSN_CTRL_FLASH_IO, docptr, CDSNControl); + WriteDOC(0, docptr, 2k_CDSN_IO); + WriteDOC(0, docptr, 2k_CDSN_IO); + WriteDOC(0, docptr, 2k_CDSN_IO); + WriteDOC(doc->CDSNControl, docptr, CDSNControl); + } else if (DoC_is_MillenniumPlus(doc)) { + WriteDOC(0, docptr, Mplus_NOP); + WriteDOC(0, docptr, Mplus_NOP); + WriteDOC(0, docptr, Mplus_NOP); + } else { + WriteDOC(0, docptr, NOP); + WriteDOC(0, docptr, NOP); + WriteDOC(0, docptr, NOP); + } + + for (i = 0; i < 6; i++) { + if (DoC_is_MillenniumPlus(doc)) + ecc_code[i] = ReadDOC_(docptr, DoC_Mplus_ECCSyndrome0 + i); + else + ecc_code[i] = ReadDOC_(docptr, DoC_ECCSyndrome0 + i); + if (ecc_code[i] != empty_write_ecc[i]) + emptymatch = 0; + } + if (DoC_is_MillenniumPlus(doc)) + WriteDOC(DOC_ECC_DIS, docptr, Mplus_ECCConf); + else + WriteDOC(DOC_ECC_DIS, docptr, ECCConf); +#if 0 + /* If emptymatch=1, we might have an all-0xff data buffer. Check. */ + if (emptymatch) { + /* Note: this somewhat expensive test should not be triggered + often. It could be optimized away by examining the data in + the writebuf routine, and remembering the result. */ + for (i = 0; i < 512; i++) { + if (dat[i] == 0xff) + continue; + emptymatch = 0; + break; + } + } + /* If emptymatch still =1, we do have an all-0xff data buffer. + Return all-0xff ecc value instead of the computed one, so + it'll look just like a freshly-erased page. */ + if (emptymatch) + memset(ecc_code, 0xff, 6); +#endif + return 0; +} + +static int doc200x_correct_data(struct nand_chip *this, u_char *dat, + u_char *read_ecc, u_char *isnull) +{ + int i, ret = 0; + struct doc_priv *doc = nand_get_controller_data(this); + void __iomem *docptr = doc->virtadr; + uint8_t calc_ecc[6]; + volatile u_char dummy; + + /* flush the pipeline */ + if (DoC_is_2000(doc)) { + dummy = ReadDOC(docptr, 2k_ECCStatus); + dummy = ReadDOC(docptr, 2k_ECCStatus); + dummy = ReadDOC(docptr, 2k_ECCStatus); + } else if (DoC_is_MillenniumPlus(doc)) { + dummy = ReadDOC(docptr, Mplus_ECCConf); + dummy = ReadDOC(docptr, Mplus_ECCConf); + dummy = ReadDOC(docptr, Mplus_ECCConf); + } else { + dummy = ReadDOC(docptr, ECCConf); + dummy = ReadDOC(docptr, ECCConf); + dummy = ReadDOC(docptr, ECCConf); + } + + /* Error occurred ? */ + if (dummy & 0x80) { + for (i = 0; i < 6; i++) { + if (DoC_is_MillenniumPlus(doc)) + calc_ecc[i] = ReadDOC_(docptr, DoC_Mplus_ECCSyndrome0 + i); + else + calc_ecc[i] = ReadDOC_(docptr, DoC_ECCSyndrome0 + i); + } + + ret = doc_ecc_decode(doc->rs_decoder, dat, calc_ecc); + if (ret > 0) + pr_err("doc200x_correct_data corrected %d errors\n", + ret); + } + if (DoC_is_MillenniumPlus(doc)) + WriteDOC(DOC_ECC_DIS, docptr, Mplus_ECCConf); + else + WriteDOC(DOC_ECC_DIS, docptr, ECCConf); + if (no_ecc_failures && mtd_is_eccerr(ret)) { + pr_err("suppressing ECC failure\n"); + ret = 0; + } + return ret; +} + +//u_char mydatabuf[528]; + +static int doc200x_ooblayout_ecc(struct mtd_info *mtd, int section, + struct mtd_oob_region *oobregion) +{ + if (section) + return -ERANGE; + + oobregion->offset = 0; + oobregion->length = 6; + + return 0; +} + +static int doc200x_ooblayout_free(struct mtd_info *mtd, int section, + struct mtd_oob_region *oobregion) +{ + if (section > 1) + return -ERANGE; + + /* + * The strange out-of-order free bytes definition is a (possibly + * unneeded) attempt to retain compatibility. It used to read: + * .oobfree = { {8, 8} } + * Since that leaves two bytes unusable, it was changed. But the + * following scheme might affect existing jffs2 installs by moving the + * cleanmarker: + * .oobfree = { {6, 10} } + * jffs2 seems to handle the above gracefully, but the current scheme + * seems safer. The only problem with it is that any code retrieving + * free bytes position must be able to handle out-of-order segments. + */ + if (!section) { + oobregion->offset = 8; + oobregion->length = 8; + } else { + oobregion->offset = 6; + oobregion->length = 2; + } + + return 0; +} + +static const struct mtd_ooblayout_ops doc200x_ooblayout_ops = { + .ecc = doc200x_ooblayout_ecc, + .free = doc200x_ooblayout_free, +}; + +/* Find the (I)NFTL Media Header, and optionally also the mirror media header. + On successful return, buf will contain a copy of the media header for + further processing. id is the string to scan for, and will presumably be + either "ANAND" or "BNAND". If findmirror=1, also look for the mirror media + header. The page #s of the found media headers are placed in mh0_page and + mh1_page in the DOC private structure. */ +static int __init find_media_headers(struct mtd_info *mtd, u_char *buf, const char *id, int findmirror) +{ + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + unsigned offs; + int ret; + size_t retlen; + + for (offs = 0; offs < mtd->size; offs += mtd->erasesize) { + ret = mtd_read(mtd, offs, mtd->writesize, &retlen, buf); + if (retlen != mtd->writesize) + continue; + if (ret) { + pr_warn("ECC error scanning DOC at 0x%x\n", offs); + } + if (memcmp(buf, id, 6)) + continue; + pr_info("Found DiskOnChip %s Media Header at 0x%x\n", id, offs); + if (doc->mh0_page == -1) { + doc->mh0_page = offs >> this->page_shift; + if (!findmirror) + return 1; + continue; + } + doc->mh1_page = offs >> this->page_shift; + return 2; + } + if (doc->mh0_page == -1) { + pr_warn("DiskOnChip %s Media Header not found.\n", id); + return 0; + } + /* Only one mediaheader was found. We want buf to contain a + mediaheader on return, so we'll have to re-read the one we found. */ + offs = doc->mh0_page << this->page_shift; + ret = mtd_read(mtd, offs, mtd->writesize, &retlen, buf); + if (retlen != mtd->writesize) { + /* Insanity. Give up. */ + pr_err("Read DiskOnChip Media Header once, but can't reread it???\n"); + return 0; + } + return 1; +} + +static inline int __init nftl_partscan(struct mtd_info *mtd, struct mtd_partition *parts) +{ + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + struct nand_memory_organization *memorg; + int ret = 0; + u_char *buf; + struct NFTLMediaHeader *mh; + const unsigned psize = 1 << this->page_shift; + int numparts = 0; + unsigned blocks, maxblocks; + int offs, numheaders; + + memorg = nanddev_get_memorg(&this->base); + + buf = kmalloc(mtd->writesize, GFP_KERNEL); + if (!buf) { + return 0; + } + if (!(numheaders = find_media_headers(mtd, buf, "ANAND", 1))) + goto out; + mh = (struct NFTLMediaHeader *)buf; + + le16_to_cpus(&mh->NumEraseUnits); + le16_to_cpus(&mh->FirstPhysicalEUN); + le32_to_cpus(&mh->FormattedSize); + + pr_info(" DataOrgID = %s\n" + " NumEraseUnits = %d\n" + " FirstPhysicalEUN = %d\n" + " FormattedSize = %d\n" + " UnitSizeFactor = %d\n", + mh->DataOrgID, mh->NumEraseUnits, + mh->FirstPhysicalEUN, mh->FormattedSize, + mh->UnitSizeFactor); + + blocks = mtd->size >> this->phys_erase_shift; + maxblocks = min(32768U, mtd->erasesize - psize); + + if (mh->UnitSizeFactor == 0x00) { + /* Auto-determine UnitSizeFactor. The constraints are: + - There can be at most 32768 virtual blocks. + - There can be at most (virtual block size - page size) + virtual blocks (because MediaHeader+BBT must fit in 1). + */ + mh->UnitSizeFactor = 0xff; + while (blocks > maxblocks) { + blocks >>= 1; + maxblocks = min(32768U, (maxblocks << 1) + psize); + mh->UnitSizeFactor--; + } + pr_warn("UnitSizeFactor=0x00 detected. Correct value is assumed to be 0x%02x.\n", mh->UnitSizeFactor); + } + + /* NOTE: The lines below modify internal variables of the NAND and MTD + layers; variables with have already been configured by nand_scan. + Unfortunately, we didn't know before this point what these values + should be. Thus, this code is somewhat dependent on the exact + implementation of the NAND layer. */ + if (mh->UnitSizeFactor != 0xff) { + this->bbt_erase_shift += (0xff - mh->UnitSizeFactor); + memorg->pages_per_eraseblock <<= (0xff - mh->UnitSizeFactor); + mtd->erasesize <<= (0xff - mh->UnitSizeFactor); + pr_info("Setting virtual erase size to %d\n", mtd->erasesize); + blocks = mtd->size >> this->bbt_erase_shift; + maxblocks = min(32768U, mtd->erasesize - psize); + } + + if (blocks > maxblocks) { + pr_err("UnitSizeFactor of 0x%02x is inconsistent with device size. Aborting.\n", mh->UnitSizeFactor); + goto out; + } + + /* Skip past the media headers. */ + offs = max(doc->mh0_page, doc->mh1_page); + offs <<= this->page_shift; + offs += mtd->erasesize; + + if (show_firmware_partition == 1) { + parts[0].name = " DiskOnChip Firmware / Media Header partition"; + parts[0].offset = 0; + parts[0].size = offs; + numparts = 1; + } + + parts[numparts].name = " DiskOnChip BDTL partition"; + parts[numparts].offset = offs; + parts[numparts].size = (mh->NumEraseUnits - numheaders) << this->bbt_erase_shift; + + offs += parts[numparts].size; + numparts++; + + if (offs < mtd->size) { + parts[numparts].name = " DiskOnChip Remainder partition"; + parts[numparts].offset = offs; + parts[numparts].size = mtd->size - offs; + numparts++; + } + + ret = numparts; + out: + kfree(buf); + return ret; +} + +/* This is a stripped-down copy of the code in inftlmount.c */ +static inline int __init inftl_partscan(struct mtd_info *mtd, struct mtd_partition *parts) +{ + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + int ret = 0; + u_char *buf; + struct INFTLMediaHeader *mh; + struct INFTLPartition *ip; + int numparts = 0; + int blocks; + int vshift, lastvunit = 0; + int i; + int end = mtd->size; + + if (inftl_bbt_write) + end -= (INFTL_BBT_RESERVED_BLOCKS << this->phys_erase_shift); + + buf = kmalloc(mtd->writesize, GFP_KERNEL); + if (!buf) { + return 0; + } + + if (!find_media_headers(mtd, buf, "BNAND", 0)) + goto out; + doc->mh1_page = doc->mh0_page + (4096 >> this->page_shift); + mh = (struct INFTLMediaHeader *)buf; + + le32_to_cpus(&mh->NoOfBootImageBlocks); + le32_to_cpus(&mh->NoOfBinaryPartitions); + le32_to_cpus(&mh->NoOfBDTLPartitions); + le32_to_cpus(&mh->BlockMultiplierBits); + le32_to_cpus(&mh->FormatFlags); + le32_to_cpus(&mh->PercentUsed); + + pr_info(" bootRecordID = %s\n" + " NoOfBootImageBlocks = %d\n" + " NoOfBinaryPartitions = %d\n" + " NoOfBDTLPartitions = %d\n" + " BlockMultiplierBits = %d\n" + " FormatFlgs = %d\n" + " OsakVersion = %d.%d.%d.%d\n" + " PercentUsed = %d\n", + mh->bootRecordID, mh->NoOfBootImageBlocks, + mh->NoOfBinaryPartitions, + mh->NoOfBDTLPartitions, + mh->BlockMultiplierBits, mh->FormatFlags, + ((unsigned char *) &mh->OsakVersion)[0] & 0xf, + ((unsigned char *) &mh->OsakVersion)[1] & 0xf, + ((unsigned char *) &mh->OsakVersion)[2] & 0xf, + ((unsigned char *) &mh->OsakVersion)[3] & 0xf, + mh->PercentUsed); + + vshift = this->phys_erase_shift + mh->BlockMultiplierBits; + + blocks = mtd->size >> vshift; + if (blocks > 32768) { + pr_err("BlockMultiplierBits=%d is inconsistent with device size. Aborting.\n", mh->BlockMultiplierBits); + goto out; + } + + blocks = doc->chips_per_floor << (this->chip_shift - this->phys_erase_shift); + if (inftl_bbt_write && (blocks > mtd->erasesize)) { + pr_err("Writeable BBTs spanning more than one erase block are not yet supported. FIX ME!\n"); + goto out; + } + + /* Scan the partitions */ + for (i = 0; (i < 4); i++) { + ip = &(mh->Partitions[i]); + le32_to_cpus(&ip->virtualUnits); + le32_to_cpus(&ip->firstUnit); + le32_to_cpus(&ip->lastUnit); + le32_to_cpus(&ip->flags); + le32_to_cpus(&ip->spareUnits); + le32_to_cpus(&ip->Reserved0); + + pr_info(" PARTITION[%d] ->\n" + " virtualUnits = %d\n" + " firstUnit = %d\n" + " lastUnit = %d\n" + " flags = 0x%x\n" + " spareUnits = %d\n", + i, ip->virtualUnits, ip->firstUnit, + ip->lastUnit, ip->flags, + ip->spareUnits); + + if ((show_firmware_partition == 1) && + (i == 0) && (ip->firstUnit > 0)) { + parts[0].name = " DiskOnChip IPL / Media Header partition"; + parts[0].offset = 0; + parts[0].size = mtd->erasesize * ip->firstUnit; + numparts = 1; + } + + if (ip->flags & INFTL_BINARY) + parts[numparts].name = " DiskOnChip BDK partition"; + else + parts[numparts].name = " DiskOnChip BDTL partition"; + parts[numparts].offset = ip->firstUnit << vshift; + parts[numparts].size = (1 + ip->lastUnit - ip->firstUnit) << vshift; + numparts++; + if (ip->lastUnit > lastvunit) + lastvunit = ip->lastUnit; + if (ip->flags & INFTL_LAST) + break; + } + lastvunit++; + if ((lastvunit << vshift) < end) { + parts[numparts].name = " DiskOnChip Remainder partition"; + parts[numparts].offset = lastvunit << vshift; + parts[numparts].size = end - parts[numparts].offset; + numparts++; + } + ret = numparts; + out: + kfree(buf); + return ret; +} + +static int __init nftl_scan_bbt(struct mtd_info *mtd) +{ + int ret, numparts; + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + struct mtd_partition parts[2]; + + memset((char *)parts, 0, sizeof(parts)); + /* On NFTL, we have to find the media headers before we can read the + BBTs, since they're stored in the media header eraseblocks. */ + numparts = nftl_partscan(mtd, parts); + if (!numparts) + return -EIO; + this->bbt_td->options = NAND_BBT_ABSPAGE | NAND_BBT_8BIT | + NAND_BBT_SAVECONTENT | NAND_BBT_WRITE | + NAND_BBT_VERSION; + this->bbt_td->veroffs = 7; + this->bbt_td->pages[0] = doc->mh0_page + 1; + if (doc->mh1_page != -1) { + this->bbt_md->options = NAND_BBT_ABSPAGE | NAND_BBT_8BIT | + NAND_BBT_SAVECONTENT | NAND_BBT_WRITE | + NAND_BBT_VERSION; + this->bbt_md->veroffs = 7; + this->bbt_md->pages[0] = doc->mh1_page + 1; + } else { + this->bbt_md = NULL; + } + + ret = nand_create_bbt(this); + if (ret) + return ret; + + return mtd_device_register(mtd, parts, no_autopart ? 0 : numparts); +} + +static int __init inftl_scan_bbt(struct mtd_info *mtd) +{ + int ret, numparts; + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + struct mtd_partition parts[5]; + + if (nanddev_ntargets(&this->base) > doc->chips_per_floor) { + pr_err("Multi-floor INFTL devices not yet supported.\n"); + return -EIO; + } + + if (DoC_is_MillenniumPlus(doc)) { + this->bbt_td->options = NAND_BBT_2BIT | NAND_BBT_ABSPAGE; + if (inftl_bbt_write) + this->bbt_td->options |= NAND_BBT_WRITE; + this->bbt_td->pages[0] = 2; + this->bbt_md = NULL; + } else { + this->bbt_td->options = NAND_BBT_LASTBLOCK | NAND_BBT_8BIT | NAND_BBT_VERSION; + if (inftl_bbt_write) + this->bbt_td->options |= NAND_BBT_WRITE; + this->bbt_td->offs = 8; + this->bbt_td->len = 8; + this->bbt_td->veroffs = 7; + this->bbt_td->maxblocks = INFTL_BBT_RESERVED_BLOCKS; + this->bbt_td->reserved_block_code = 0x01; + this->bbt_td->pattern = "MSYS_BBT"; + + this->bbt_md->options = NAND_BBT_LASTBLOCK | NAND_BBT_8BIT | NAND_BBT_VERSION; + if (inftl_bbt_write) + this->bbt_md->options |= NAND_BBT_WRITE; + this->bbt_md->offs = 8; + this->bbt_md->len = 8; + this->bbt_md->veroffs = 7; + this->bbt_md->maxblocks = INFTL_BBT_RESERVED_BLOCKS; + this->bbt_md->reserved_block_code = 0x01; + this->bbt_md->pattern = "TBB_SYSM"; + } + + ret = nand_create_bbt(this); + if (ret) + return ret; + + memset((char *)parts, 0, sizeof(parts)); + numparts = inftl_partscan(mtd, parts); + /* At least for now, require the INFTL Media Header. We could probably + do without it for non-INFTL use, since all it gives us is + autopartitioning, but I want to give it more thought. */ + if (!numparts) + return -EIO; + return mtd_device_register(mtd, parts, no_autopart ? 0 : numparts); +} + +static inline int __init doc2000_init(struct mtd_info *mtd) +{ + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + + doc->late_init = nftl_scan_bbt; + + doc->CDSNControl = CDSN_CTRL_FLASH_IO | CDSN_CTRL_ECC_IO; + doc2000_count_chips(mtd); + mtd->name = "DiskOnChip 2000 (NFTL Model)"; + return (4 * doc->chips_per_floor); +} + +static inline int __init doc2001_init(struct mtd_info *mtd) +{ + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + + ReadDOC(doc->virtadr, ChipID); + ReadDOC(doc->virtadr, ChipID); + ReadDOC(doc->virtadr, ChipID); + if (ReadDOC(doc->virtadr, ChipID) != DOC_ChipID_DocMil) { + /* It's not a Millennium; it's one of the newer + DiskOnChip 2000 units with a similar ASIC. + Treat it like a Millennium, except that it + can have multiple chips. */ + doc2000_count_chips(mtd); + mtd->name = "DiskOnChip 2000 (INFTL Model)"; + doc->late_init = inftl_scan_bbt; + return (4 * doc->chips_per_floor); + } else { + /* Bog-standard Millennium */ + doc->chips_per_floor = 1; + mtd->name = "DiskOnChip Millennium"; + doc->late_init = nftl_scan_bbt; + return 1; + } +} + +static inline int __init doc2001plus_init(struct mtd_info *mtd) +{ + struct nand_chip *this = mtd_to_nand(mtd); + struct doc_priv *doc = nand_get_controller_data(this); + + doc->late_init = inftl_scan_bbt; + this->ecc.hwctl = doc2001plus_enable_hwecc; + + doc->chips_per_floor = 1; + mtd->name = "DiskOnChip Millennium Plus"; + + return 1; +} + +static int doc200x_attach_chip(struct nand_chip *chip) +{ + if (chip->ecc.engine_type != NAND_ECC_ENGINE_TYPE_ON_HOST) + return 0; + + chip->ecc.placement = NAND_ECC_PLACEMENT_INTERLEAVED; + chip->ecc.size = 512; + chip->ecc.bytes = 6; + chip->ecc.strength = 2; + chip->ecc.options = NAND_ECC_GENERIC_ERASED_CHECK; + chip->ecc.hwctl = doc200x_enable_hwecc; + chip->ecc.calculate = doc200x_calculate_ecc; + chip->ecc.correct = doc200x_correct_data; + + return 0; +} + +static const struct nand_controller_ops doc200x_ops = { + .exec_op = doc200x_exec_op, + .attach_chip = doc200x_attach_chip, +}; + +static const struct nand_controller_ops doc2001plus_ops = { + .exec_op = doc2001plus_exec_op, + .attach_chip = doc200x_attach_chip, +}; + +static int __init doc_probe(unsigned long physadr) +{ + struct nand_chip *nand = NULL; + struct doc_priv *doc = NULL; + unsigned char ChipID; + struct mtd_info *mtd; + void __iomem *virtadr; + unsigned char save_control; + unsigned char tmp, tmpb, tmpc; + int reg, len, numchips; + int ret = 0; + + if (!request_mem_region(physadr, DOC_IOREMAP_LEN, "DiskOnChip")) + return -EBUSY; + virtadr = ioremap(physadr, DOC_IOREMAP_LEN); + if (!virtadr) { + pr_err("Diskonchip ioremap failed: 0x%x bytes at 0x%lx\n", + DOC_IOREMAP_LEN, physadr); + ret = -EIO; + goto error_ioremap; + } + + /* It's not possible to cleanly detect the DiskOnChip - the + * bootup procedure will put the device into reset mode, and + * it's not possible to talk to it without actually writing + * to the DOCControl register. So we store the current contents + * of the DOCControl register's location, in case we later decide + * that it's not a DiskOnChip, and want to put it back how we + * found it. + */ + save_control = ReadDOC(virtadr, DOCControl); + + /* Reset the DiskOnChip ASIC */ + WriteDOC(DOC_MODE_CLR_ERR | DOC_MODE_MDWREN | DOC_MODE_RESET, virtadr, DOCControl); + WriteDOC(DOC_MODE_CLR_ERR | DOC_MODE_MDWREN | DOC_MODE_RESET, virtadr, DOCControl); + + /* Enable the DiskOnChip ASIC */ + WriteDOC(DOC_MODE_CLR_ERR | DOC_MODE_MDWREN | DOC_MODE_NORMAL, virtadr, DOCControl); + WriteDOC(DOC_MODE_CLR_ERR | DOC_MODE_MDWREN | DOC_MODE_NORMAL, virtadr, DOCControl); + + ChipID = ReadDOC(virtadr, ChipID); + + switch (ChipID) { + case DOC_ChipID_Doc2k: + reg = DoC_2k_ECCStatus; + break; + case DOC_ChipID_DocMil: + reg = DoC_ECCConf; + break; + case DOC_ChipID_DocMilPlus16: + case DOC_ChipID_DocMilPlus32: + case 0: + /* Possible Millennium Plus, need to do more checks */ + /* Possibly release from power down mode */ + for (tmp = 0; (tmp < 4); tmp++) + ReadDOC(virtadr, Mplus_Power); + + /* Reset the Millennium Plus ASIC */ + tmp = DOC_MODE_RESET | DOC_MODE_MDWREN | DOC_MODE_RST_LAT | DOC_MODE_BDECT; + WriteDOC(tmp, virtadr, Mplus_DOCControl); + WriteDOC(~tmp, virtadr, Mplus_CtrlConfirm); + + usleep_range(1000, 2000); + /* Enable the Millennium Plus ASIC */ + tmp = DOC_MODE_NORMAL | DOC_MODE_MDWREN | DOC_MODE_RST_LAT | DOC_MODE_BDECT; + WriteDOC(tmp, virtadr, Mplus_DOCControl); + WriteDOC(~tmp, virtadr, Mplus_CtrlConfirm); + usleep_range(1000, 2000); + + ChipID = ReadDOC(virtadr, ChipID); + + switch (ChipID) { + case DOC_ChipID_DocMilPlus16: + reg = DoC_Mplus_Toggle; + break; + case DOC_ChipID_DocMilPlus32: + pr_err("DiskOnChip Millennium Plus 32MB is not supported, ignoring.\n"); + fallthrough; + default: + ret = -ENODEV; + goto notfound; + } + break; + + default: + ret = -ENODEV; + goto notfound; + } + /* Check the TOGGLE bit in the ECC register */ + tmp = ReadDOC_(virtadr, reg) & DOC_TOGGLE_BIT; + tmpb = ReadDOC_(virtadr, reg) & DOC_TOGGLE_BIT; + tmpc = ReadDOC_(virtadr, reg) & DOC_TOGGLE_BIT; + if ((tmp == tmpb) || (tmp != tmpc)) { + pr_warn("Possible DiskOnChip at 0x%lx failed TOGGLE test, dropping.\n", physadr); + ret = -ENODEV; + goto notfound; + } + + for (mtd = doclist; mtd; mtd = doc->nextdoc) { + unsigned char oldval; + unsigned char newval; + nand = mtd_to_nand(mtd); + doc = nand_get_controller_data(nand); + /* Use the alias resolution register to determine if this is + in fact the same DOC aliased to a new address. If writes + to one chip's alias resolution register change the value on + the other chip, they're the same chip. */ + if (ChipID == DOC_ChipID_DocMilPlus16) { + oldval = ReadDOC(doc->virtadr, Mplus_AliasResolution); + newval = ReadDOC(virtadr, Mplus_AliasResolution); + } else { + oldval = ReadDOC(doc->virtadr, AliasResolution); + newval = ReadDOC(virtadr, AliasResolution); + } + if (oldval != newval) + continue; + if (ChipID == DOC_ChipID_DocMilPlus16) { + WriteDOC(~newval, virtadr, Mplus_AliasResolution); + oldval = ReadDOC(doc->virtadr, Mplus_AliasResolution); + WriteDOC(newval, virtadr, Mplus_AliasResolution); // restore it + } else { + WriteDOC(~newval, virtadr, AliasResolution); + oldval = ReadDOC(doc->virtadr, AliasResolution); + WriteDOC(newval, virtadr, AliasResolution); // restore it + } + newval = ~newval; + if (oldval == newval) { + pr_debug("Found alias of DOC at 0x%lx to 0x%lx\n", + doc->physadr, physadr); + goto notfound; + } + } + + pr_notice("DiskOnChip found at 0x%lx\n", physadr); + + len = sizeof(struct nand_chip) + sizeof(struct doc_priv) + + (2 * sizeof(struct nand_bbt_descr)); + nand = kzalloc(len, GFP_KERNEL); + if (!nand) { + ret = -ENOMEM; + goto fail; + } + + /* + * Allocate a RS codec instance + * + * Symbolsize is 10 (bits) + * Primitve polynomial is x^10+x^3+1 + * First consecutive root is 510 + * Primitve element to generate roots = 1 + * Generator polinomial degree = 4 + */ + doc = (struct doc_priv *) (nand + 1); + doc->rs_decoder = init_rs(10, 0x409, FCR, 1, NROOTS); + if (!doc->rs_decoder) { + pr_err("DiskOnChip: Could not create a RS codec\n"); + ret = -ENOMEM; + goto fail; + } + + nand_controller_init(&doc->base); + if (ChipID == DOC_ChipID_DocMilPlus16) + doc->base.ops = &doc2001plus_ops; + else + doc->base.ops = &doc200x_ops; + + mtd = nand_to_mtd(nand); + nand->bbt_td = (struct nand_bbt_descr *) (doc + 1); + nand->bbt_md = nand->bbt_td + 1; + + mtd->owner = THIS_MODULE; + mtd_set_ooblayout(mtd, &doc200x_ooblayout_ops); + + nand->controller = &doc->base; + nand_set_controller_data(nand, doc); + nand->bbt_options = NAND_BBT_USE_FLASH; + /* Skip the automatic BBT scan so we can run it manually */ + nand->options |= NAND_SKIP_BBTSCAN | NAND_NO_BBM_QUIRK; + + doc->physadr = physadr; + doc->virtadr = virtadr; + doc->ChipID = ChipID; + doc->curfloor = -1; + doc->curchip = -1; + doc->mh0_page = -1; + doc->mh1_page = -1; + doc->nextdoc = doclist; + + if (ChipID == DOC_ChipID_Doc2k) + numchips = doc2000_init(mtd); + else if (ChipID == DOC_ChipID_DocMilPlus16) + numchips = doc2001plus_init(mtd); + else + numchips = doc2001_init(mtd); + + if ((ret = nand_scan(nand, numchips)) || (ret = doc->late_init(mtd))) { + /* DBB note: i believe nand_cleanup is necessary here, as + buffers may have been allocated in nand_base. Check with + Thomas. FIX ME! */ + nand_cleanup(nand); + goto fail; + } + + /* Success! */ + doclist = mtd; + return 0; + + notfound: + /* Put back the contents of the DOCControl register, in case it's not + actually a DiskOnChip. */ + WriteDOC(save_control, virtadr, DOCControl); + fail: + if (doc) + free_rs(doc->rs_decoder); + kfree(nand); + iounmap(virtadr); + +error_ioremap: + release_mem_region(physadr, DOC_IOREMAP_LEN); + + return ret; +} + +static void release_nanddoc(void) +{ + struct mtd_info *mtd, *nextmtd; + struct nand_chip *nand; + struct doc_priv *doc; + int ret; + + for (mtd = doclist; mtd; mtd = nextmtd) { + nand = mtd_to_nand(mtd); + doc = nand_get_controller_data(nand); + + nextmtd = doc->nextdoc; + ret = mtd_device_unregister(mtd); + WARN_ON(ret); + nand_cleanup(nand); + iounmap(doc->virtadr); + release_mem_region(doc->physadr, DOC_IOREMAP_LEN); + free_rs(doc->rs_decoder); + kfree(nand); + } +} + +static int __init init_nanddoc(void) +{ + int i, ret = 0; + + if (doc_config_location) { + pr_info("Using configured DiskOnChip probe address 0x%lx\n", + doc_config_location); + ret = doc_probe(doc_config_location); + if (ret < 0) + return ret; + } else { + for (i = 0; (doc_locations[i] != 0xffffffff); i++) { + doc_probe(doc_locations[i]); + } + } + /* No banner message any more. Print a message if no DiskOnChip + found, so the user knows we at least tried. */ + if (!doclist) { + pr_info("No valid DiskOnChip devices found\n"); + ret = -ENODEV; + } + return ret; +} + +static void __exit cleanup_nanddoc(void) +{ + /* Cleanup the nand/DoC resources */ + release_nanddoc(); +} + +module_init(init_nanddoc); +module_exit(cleanup_nanddoc); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("David Woodhouse <dwmw2@infradead.org>"); +MODULE_DESCRIPTION("M-Systems DiskOnChip 2000, Millennium and Millennium Plus device driver"); |