wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 1 | NAND FLASH commands and notes |
| 2 | |
Wolfgang Denk | ac53b70 | 2006-03-06 11:25:22 +0100 | [diff] [blame] | 3 | See NOTE below!!! |
| 4 | |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 5 | # (C) Copyright 2003 |
| 6 | # Dave Ellis, SIXNET, dge@sixnetio.com |
| 7 | # |
| 8 | # See file CREDITS for list of people who contributed to this |
| 9 | # project. |
| 10 | # |
| 11 | # This program is free software; you can redistribute it and/or |
| 12 | # modify it under the terms of the GNU General Public License as |
| 13 | # published by the Free Software Foundation; either version 2 of |
| 14 | # the License, or (at your option) any later version. |
| 15 | # |
| 16 | # This program is distributed in the hope that it will be useful, |
| 17 | # but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 18 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 19 | # GNU General Public License for more details. |
| 20 | # |
| 21 | # You should have received a copy of the GNU General Public License |
| 22 | # along with this program; if not, write to the Free Software |
| 23 | # Foundation, Inc., 59 Temple Place, Suite 330, Boston, |
| 24 | # MA 02111-1307 USA |
| 25 | |
| 26 | Commands: |
| 27 | |
| 28 | nand bad |
| 29 | Print a list of all of the bad blocks in the current device. |
| 30 | |
| 31 | nand device |
| 32 | Print information about the current NAND device. |
| 33 | |
| 34 | nand device num |
| 35 | Make device `num' the current device and print information about it. |
| 36 | |
Stefan Roese | 198b23e | 2006-10-28 15:55:52 +0200 | [diff] [blame] | 37 | nand erase off|partition size |
| 38 | nand erase clean [off|partition size] |
| 39 | Erase `size' bytes starting at offset `off'. Alternatively partition |
| 40 | name can be specified, in this case size will be eventually limited |
| 41 | to not exceed partition size (this behaviour applies also to read |
| 42 | and write commands). Only complete erase blocks can be erased. |
| 43 | |
| 44 | If `erase' is specified without an offset or size, the entire flash |
| 45 | is erased. If `erase' is specified with partition but without an |
| 46 | size, the entire partition is erased. |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 47 | |
| 48 | If `clean' is specified, a JFFS2-style clean marker is written to |
Stefan Roese | 198b23e | 2006-10-28 15:55:52 +0200 | [diff] [blame] | 49 | each block after it is erased. |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 50 | |
| 51 | This command will not erase blocks that are marked bad. There is |
| 52 | a debug option in cmd_nand.c to allow bad blocks to be erased. |
| 53 | Please read the warning there before using it, as blocks marked |
| 54 | bad by the manufacturer must _NEVER_ be erased. |
| 55 | |
| 56 | nand info |
| 57 | Print information about all of the NAND devices found. |
| 58 | |
Stefan Roese | 198b23e | 2006-10-28 15:55:52 +0200 | [diff] [blame] | 59 | nand read addr ofs|partition size |
Scott Wood | 12b1c95 | 2008-06-12 13:13:23 -0500 | [diff] [blame] | 60 | Read `size' bytes from `ofs' in NAND flash to `addr'. Blocks that |
| 61 | are marked bad are skipped. If a page cannot be read because an |
| 62 | uncorrectable data error is found, the command stops with an error. |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 63 | |
Stefan Roese | 198b23e | 2006-10-28 15:55:52 +0200 | [diff] [blame] | 64 | nand read.oob addr ofs|partition size |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 65 | Read `size' bytes from the out-of-band data area corresponding to |
| 66 | `ofs' in NAND flash to `addr'. This is limited to the 16 bytes of |
| 67 | data for one 512-byte page or 2 256-byte pages. There is no check |
| 68 | for bad blocks or ECC errors. |
| 69 | |
Stefan Roese | 198b23e | 2006-10-28 15:55:52 +0200 | [diff] [blame] | 70 | nand write addr ofs|partition size |
Scott Wood | 12b1c95 | 2008-06-12 13:13:23 -0500 | [diff] [blame] | 71 | Write `size' bytes from `addr' to `ofs' in NAND flash. Blocks that |
| 72 | are marked bad are skipped. If a page cannot be read because an |
| 73 | uncorrectable data error is found, the command stops with an error. |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 74 | |
Scott Wood | 12b1c95 | 2008-06-12 13:13:23 -0500 | [diff] [blame] | 75 | As JFFS2 skips blocks similarly, this allows writing a JFFS2 image, |
| 76 | as long as the image is short enough to fit even after skipping the |
| 77 | bad blocks. Compact images, such as those produced by mkfs.jffs2 |
| 78 | should work well, but loading an image copied from another flash is |
| 79 | going to be trouble if there are any bad blocks. |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 80 | |
Ben Gardiner | bc04b4d | 2011-06-14 16:35:07 -0400 | [diff] [blame] | 81 | nand write.trimffs addr ofs|partition size |
| 82 | Enabled by the CONFIG_CMD_NAND_TRIMFFS macro. This command will write to |
| 83 | the NAND flash in a manner identical to the 'nand write' command |
| 84 | described above -- with the additional check that all pages at the end |
| 85 | of eraseblocks which contain only 0xff data will not be written to the |
| 86 | NAND flash. This behaviour is required when flashing UBI images |
| 87 | containing UBIFS volumes as per the UBI FAQ[1]. |
| 88 | |
| 89 | [1] http://www.linux-mtd.infradead.org/doc/ubi.html#L_flasher_algo |
| 90 | |
Stefan Roese | 198b23e | 2006-10-28 15:55:52 +0200 | [diff] [blame] | 91 | nand write.oob addr ofs|partition size |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 92 | Write `size' bytes from `addr' to the out-of-band data area |
| 93 | corresponding to `ofs' in NAND flash. This is limited to the 16 bytes |
| 94 | of data for one 512-byte page or 2 256-byte pages. There is no check |
| 95 | for bad blocks. |
| 96 | |
Scott Wood | 733fb02 | 2012-03-02 14:01:57 -0600 | [diff] [blame] | 97 | nand read.raw addr ofs|partition [count] |
| 98 | nand write.raw addr ofs|partition [count] |
| 99 | Read or write one or more pages at "ofs" in NAND flash, from or to |
| 100 | "addr" in memory. This is a raw access, so ECC is avoided and the |
| 101 | OOB area is transferred as well. If count is absent, it is assumed |
| 102 | to be one page. As with .yaffs2 accesses, the data is formatted as |
| 103 | a packed sequence of "data, oob, data, oob, ..." -- no alignment of |
| 104 | individual pages is maintained. |
Marek Vasut | 357b78e | 2011-09-23 15:43:10 +0200 | [diff] [blame] | 105 | |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 106 | Configuration Options: |
| 107 | |
Jon Loeliger | b8a4968 | 2007-07-09 19:10:03 -0500 | [diff] [blame] | 108 | CONFIG_CMD_NAND |
| 109 | Enables NAND support and commmands. |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 110 | |
| 111 | CONFIG_MTD_NAND_ECC_JFFS2 |
| 112 | Define this if you want the Error Correction Code information in |
| 113 | the out-of-band data to be formatted to match the JFFS2 file system. |
| 114 | CONFIG_MTD_NAND_ECC_YAFFS would be another useful choice for |
| 115 | someone to implement. |
| 116 | |
Jean-Christophe PLAGNIOL-VILLARD | 0383694 | 2008-10-16 15:01:15 +0200 | [diff] [blame] | 117 | CONFIG_SYS_MAX_NAND_DEVICE |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 118 | The maximum number of NAND devices you want to support. |
| 119 | |
Scott Wood | 12dbc24 | 2009-04-01 15:33:24 -0500 | [diff] [blame] | 120 | CONFIG_SYS_NAND_MAX_CHIPS |
| 121 | The maximum number of NAND chips per device to be supported. |
wdenk | abda5ca | 2003-05-31 18:35:21 +0000 | [diff] [blame] | 122 | |
Scott Wood | 193a0f5 | 2012-01-12 19:07:23 -0600 | [diff] [blame] | 123 | CONFIG_SYS_NAND_SELF_INIT |
| 124 | Traditionally, glue code in drivers/mtd/nand/nand.c has driven |
| 125 | the initialization process -- it provides the mtd and nand |
| 126 | structs, calls a board init function for a specific device, |
| 127 | calls nand_scan(), and registers with mtd. |
| 128 | |
| 129 | This arrangement does not provide drivers with the flexibility to |
| 130 | run code between nand_scan_ident() and nand_scan_tail(), or other |
| 131 | deviations from the "normal" flow. |
| 132 | |
| 133 | If a board defines CONFIG_SYS_NAND_SELF_INIT, drivers/mtd/nand/nand.c |
| 134 | will make one call to board_nand_init(), with no arguments. That |
| 135 | function is responsible for calling a driver init function for |
| 136 | each NAND device on the board, that performs all initialization |
| 137 | tasks except setting mtd->name, and registering with the rest of |
| 138 | U-Boot. Those last tasks are accomplished by calling nand_register() |
| 139 | on the new mtd device. |
| 140 | |
| 141 | Example of new init to be added to the end of an existing driver |
| 142 | init: |
| 143 | |
| 144 | /* |
| 145 | * devnum is the device number to be used in nand commands |
| 146 | * and in mtd->name. Must be less than |
| 147 | * CONFIG_SYS_NAND_MAX_DEVICE. |
| 148 | */ |
| 149 | mtd = &nand_info[devnum]; |
| 150 | |
| 151 | /* chip is struct nand_chip, and is now provided by the driver. */ |
| 152 | mtd->priv = &chip; |
| 153 | |
| 154 | /* |
| 155 | * Fill in appropriate values if this driver uses these fields, |
| 156 | * or uses the standard read_byte/write_buf/etc. functions from |
| 157 | * nand_base.c that use these fields. |
| 158 | */ |
| 159 | chip.IO_ADDR_R = ...; |
| 160 | chip.IO_ADDR_W = ...; |
| 161 | |
| 162 | if (nand_scan_ident(mtd, CONFIG_SYS_MAX_NAND_CHIPS, NULL)) |
| 163 | error out |
| 164 | |
| 165 | /* |
| 166 | * Insert here any code you wish to run after the chip has been |
| 167 | * identified, but before any other I/O is done. |
| 168 | */ |
| 169 | |
| 170 | if (nand_scan_tail(mtd)) |
| 171 | error out |
| 172 | |
| 173 | if (nand_register(devnum)) |
| 174 | error out |
| 175 | |
| 176 | In addition to providing more flexibility to the driver, it reduces |
| 177 | the difference between a U-Boot driver and its Linux counterpart. |
| 178 | nand_init() is now reduced to calling board_nand_init() once, and |
| 179 | printing a size summary. This should also make it easier to |
| 180 | transition to delayed NAND initialization. |
| 181 | |
| 182 | Please convert your driver even if you don't need the extra |
| 183 | flexibility, so that one day we can eliminate the old mechanism. |
| 184 | |
Wolfgang Denk | ac53b70 | 2006-03-06 11:25:22 +0100 | [diff] [blame] | 185 | NOTE: |
| 186 | ===== |
| 187 | |
Scott Wood | 12dbc24 | 2009-04-01 15:33:24 -0500 | [diff] [blame] | 188 | The current NAND implementation is based on what is in recent |
Scott Wood | 9866305 | 2009-04-01 15:02:13 -0500 | [diff] [blame] | 189 | Linux kernels. The old legacy implementation has been removed. |
Wolfgang Denk | ac53b70 | 2006-03-06 11:25:22 +0100 | [diff] [blame] | 190 | |
Scott Wood | 12dbc24 | 2009-04-01 15:33:24 -0500 | [diff] [blame] | 191 | If you have board code which used CONFIG_NAND_LEGACY, you'll need |
| 192 | to convert to the current NAND interface for it to continue to work. |
Stefan Roese | d351b2b | 2006-10-10 12:36:02 +0200 | [diff] [blame] | 193 | |
Scott Wood | 12dbc24 | 2009-04-01 15:33:24 -0500 | [diff] [blame] | 194 | The Disk On Chip driver is currently broken and has been for some time. |
| 195 | There is a driver in drivers/mtd/nand, taken from Linux, that works with |
| 196 | the current NAND system but has not yet been adapted to the u-boot |
| 197 | environment. |
Stefan Roese | d351b2b | 2006-10-10 12:36:02 +0200 | [diff] [blame] | 198 | |
Stefan Roese | d351b2b | 2006-10-10 12:36:02 +0200 | [diff] [blame] | 199 | Additional improvements to the NAND subsystem by Guido Classen, 10-10-2006 |
| 200 | |
| 201 | JFFS2 related commands: |
| 202 | |
| 203 | implement "nand erase clean" and old "nand erase" |
| 204 | using both the new code which is able to skip bad blocks |
| 205 | "nand erase clean" additionally writes JFFS2-cleanmarkers in the oob. |
| 206 | |
Stefan Roese | d351b2b | 2006-10-10 12:36:02 +0200 | [diff] [blame] | 207 | Miscellaneous and testing commands: |
| 208 | "markbad [offset]" |
| 209 | create an artificial bad block (for testing bad block handling) |
| 210 | |
| 211 | "scrub [offset length]" |
| 212 | like "erase" but don't skip bad block. Instead erase them. |
| 213 | DANGEROUS!!! Factory set bad blocks will be lost. Use only |
| 214 | to remove artificial bad blocks created with the "markbad" command. |
| 215 | |
| 216 | |
| 217 | NAND locking command (for chips with active LOCKPRE pin) |
| 218 | |
| 219 | "nand lock" |
| 220 | set NAND chip to lock state (all pages locked) |
| 221 | |
| 222 | "nand lock tight" |
| 223 | set NAND chip to lock tight state (software can't change locking anymore) |
| 224 | |
| 225 | "nand lock status" |
| 226 | displays current locking status of all pages |
| 227 | |
| 228 | "nand unlock [offset] [size]" |
| 229 | unlock consecutive area (can be called multiple times for different areas) |
| 230 | |
| 231 | |
| 232 | I have tested the code with board containing 128MiB NAND large page chips |
| 233 | and 32MiB small page chips. |