1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
|
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
<html>
<head>
<title>Memory Technology Device (MTD) Subsystem for Linux</title>
<meta name="description" content="A generic subsystem for handling memory technology devices under Linux">
<meta name="keywords" content="Disk-On-Chip, CompactFlash, Compact Flash, Disk On Chip, M-Systems, Linux, MTD, Flash, Memory Technology Device">
</head>
<body BGCOLOR="white">
<h1 ALIGN="CENTER">Memory Technology Device (MTD) Subsystem for Linux</h1>
<H2>Design aims</H2>
<P>We're working on a generic Linux subsystem for memory devices,
especially Flash devices.
</P>
<P>The aim of the system is to make it simple to provide a driver for new
hardware, by providing a generic interface between the hardware drivers
and the upper layers of the system.
</P>
<P>Hardware drivers need to know nothing about the storage formats used,
such as FTL, FFS2, etc., but will only need to provide simple routines
for <tt>read</tt>, <tt>write</tt> and <TT>erase</TT>.
Presentation of the device's contents to the user in an appropriate form
will be handled by the upper layers of the system.
</P>
<HR size=2>
<H2>Mailing list</H2>
<P>There is a majordomo-managed mailing list:
<B><I>lists-mtd@lists.infradead.org</i></B> - Full archives are available
at <A HREF="http://lists.infradead.org/pipermail/linux-mtd/">
http://lists.infradead.org/pipermail/linux-mtd/</A>. Now that I've
finally managed to get Mailman to display them properly, this includes
all the previous messages from before the changeover to Mailman.
</P><P>
To subscribe, go <A
HREF="http://lists.infradead.org/mailman/listinfo/linux-mtd/">
here</A> or send "<TT>subscribe</TT>" in the body of a
mail to <I><A HREF="mailto:linux-mtd-request@lists.infradead.org">
linux-mtd-request@lists.infradead.org</A></I>
</P>
<P><B>NOTE: DO NOT SEND YOUR SUBSCRIPTION REQUEST TO THE LIST ITSELF.<BR>
SEND IT TO <A
HREF="mailto:linux-mtd-request@lists.infradead.org">
linux-mtd-request@lists.infradead.org</A>
AS THE ABOVE SAYS.</B></P>
<HR size=2>
<H2>Download and CVS</H2>
<P>Very occasionally, I make snapshot releases. Now that the MTD code
is in the 2.4 kernel, it's become even rarer. Your best option is to
obtain the latest code from CVS, by following the instructions below.
We do break the CVS build occasionally, but we're also fairly good
at fixing it quickly when we do so. </P>
<P>Anonymous CVS access is available, and full CVS access may also be
granted to developers on request.
<PRE>
cvs -d :pserver:anoncvs@cvs.infradead.org:/home/cvs login <I>(password: anoncvs)</I>
cvs -d :pserver:anoncvs@cvs.infradead.org:/home/cvs co mtd
</PRE>
<P>If you can't access anoncvs for some reason, daily snapshots
are also available at <A
HREF="ftp://ftp.uk.linux.org/pub/people/dwmw2/mtd/cvs/">
ftp://ftp.uk.linux.org/pub/people/dwmw2/mtd/cvs/</A>
</P>
<HR size=2>
<H2>Documentation</H2>
<P>There is now some <A HREF="tech/">MTD API documentation</A> available.
It's a little out of date - the API has been evolving over the last
few months as we encounter strange pieces of hardware that we hadn't
anticipated, and occasionally when I'd just overlooked something
blatantly obvious. Volunteers to update the docs are welcome.
<br><br>13th Feb 2001: A <A HREF="ftp://ftp.uk.linux.org/pub/people/dwmw2/mtd/cvs/mtd/mtd-jffs-HOWTO.txt">mtd-jffs-HOWTO</A> is now also available under
CVS. Not all topics are covered yet, but it's a start.
</P>
<HR size=2>
<h2>Booting</H2>
<P>You can now replace the firmware on the DiskOnChip 2000, and possibly
also the DiskOnChip Millennium, with a version of <A
HREF="http://www.gnu.org/software/grub/">GNU Grub</A> which runs directly
from the flash.</P>
<P>The patches to make Grub aware of the DiskOnChip and the NFTL format
used on it, along with a first-stage loader to load Grub itself into
memory from the DiskOnChip, are in the CVS repository.</P>
<HR size=2>
<H2>TODO</H2>
<P>I'm impressed. The TODO list is getting much shorter.</P>
<uL>
<LI>Add an MTD driver for PCMCIA memory cards</LI>
<LI>Make the FFS2 code work on 2.3 kernels.</LI>
<LI>Implement XIP support in the 'map' layer and a read-only filesystem to use it.</LI>
<LI>Decrease the amount of flash space reserved by JFFS2 garbage collection
</uL>
<HR SIZE=2>
<h2>Changelog</H2>
<UL>
<LI><B>Latest CVS</B><BR>
<UL>
<LI>JFFS v2 -
<UL>
<LI>Compression</LI>
<LI>Improved wear levelling - we deal with individual erase blocks now</LI>
<LI>Improved concurrency - no longer all under one big semaphore</LI>
<LI>Smaller RAM footprint - at least theoretically. And we forget stuff under memory pressure to be nice</LI>
<LI>Hard link support</LI>
</UL>
</LI>
<LI>New 'mtdblock' driver which allows you to use any normal
filesystem on flash chips, albeit rather non-robust if you're
writing to it. See <A HREF="#mtdblock">below</A> for more
information.</LI>
<LI>New map drivers, including the Itsy (and iPAQ) driver.</LI>
<LI>Massive NFTL overhaul. ECC, bad block handling, journalling fixes, etc.</LI>
<LI>Verified DiskOnChip 2000 and Millennium support - some wonderful
person seems to have gone through the code and the docs and ensured
that we meet all the timing constraints that I ignored when I
threw the original code together.</LI>
<LI>Support all sensible geometries for AMD/Fujitsu command set CFI flash</LI>
<LI>Start (most?) of JFFS support for NAND flash</LI>
<LI>Hardware support for a handful of NAND flash chips</LI>
<LI>Beat JFFS up till it actually passes filesystem stress tests</LI>
<LI>Allow CFI reads to suspend in-progress erases.</LI>
<LI>Merged DiskOnChip 2000 and Millennium driver into one</LI>
<LI>Partitioning of flash chips into multiple devices</LI>
<LI>Map area drivers for various SBC boards</LI>
<LI>New DiskOnChip Millennium driver, contributed by Ollie Lho</LI>
<LI>Assorted JFFS fixes</LI>
<LI>CFI code for AMD/Fujitsu Standard command set, mapped in 2x16 mode, contributed by Masami Komiya</LI>
</UL>
<BR>
</LI>
<LI><B>mtd-20000704</B><BR>
Patch: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-20000704.diff">mtd-20000704.diff</A><BR>
Tarball: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-20000704.tar.gz">mtd-20000704.tar.gz</A><BR>
<UL>
<LI>Kernel patches for both 2.2.16 and 2.4.0 kernels to add MTD support, and
a script file to move the source files into the right place.</LI>
<LI>New <TT>map</TT> architecture for making mmap'd areas
available for use by generic chipset drivers</LI>
<LI>Probe code for Common Flash Interface devices</LI>
<LI>Run-time address alias detection for CFI chips</LI>
<LI>Driver code for Intel Extended Command Set devices - connected
any way you like as long as it's in 16-bit mode and not
interleaved</LI>
<LI>JFFS ported to 2.2 kernel</LI>
<LI>Remove the /dev/mtd<I>x</I> character device driver from
<TT>mtd.o</TT> and make it separately configurable as
<TT>mtdchar.o</TT></LI>
<LI>Rename <TT>mtd.o</TT> to <TT>mtdcore.o</TT> just in case I
didn't confuse enough people with the above</LI>
<LI>Split the DiskOnChip probe routine into a separate file,
allowing DiskOnChip 2000 and DiskOnChip Millennium drivers to
be developed separately.</LI>
<LI>Separate <TT>CONFIG_NFTL_RW</TT> option allowing you to
compile safe read-only drivers (by turning it off)</LI>
<LI>Imported Axis' <A
HREF="http://www.developer.axis.com/software/jffs/">Journalling
Flash File System (JFFS)</A></LI>
<LI>Updated FTL driver to the latest version from David Hinds'
<A HREF="http://pcmcia.sourceforge.org/">PCMCIA</A> code. Still
some elusive bugs which cause it to corrupt the FTL - must get
some PCMCIA flash hardware and check it worked <B>before</B> I
started playing with it.</LI>
<LI>Litter FTL driver with cpu_to_le16() et al calls in an
attempt to make it work on big-endian hardware</LI>
<LI>Some cleanups to make the code conform to the <A HREF="tech/">new
rules</A>.</LI>
<LI>Add <TT>get_mtd_device()</TT> and <TT>put_mtd_device</TT> back
again, having gone over the locking rules with a fine tooth
comb and decided that they were OK if done properly.</LI>
<LI>Added patch to GNU Grub which teaches it about the DiskOnChip.
When it can also decode NFTL, it'll be more useful.</LI>
</UL>
<BR>
</LI>
<LI><B>mtd-20000131</B><BR>
Patch: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-20000131.diff">mtd-20000131.diff</A><BR>
Tarball: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-20000131.tar.gz">mtd-20000131.tar.gz</A><BR>
Complete patch against 2.3.41 kernel: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-20000131-against-2.3.41.diff.gz">
mtd-20000131-against-2.3.41.diff.gz</A><BR>
<UL>
<LI>Updated to work with 2.3.recent kernel. Apart from the obvious
changes, this also includes turning off the block device request
merging code, which seems to be duplicating requests rather than
just merging them.</LI>
<LI>Remove the byteorder workaround - there's a patch included which
should fix the kernel.</LI>
<LI>Added Grub loader for DiskOnChip - now someone just has to add
DiskOnChip support to Grub itself.</LI>
</UL>
<BR>
</LI>
<LI><B>mtd-20000119</B><BR>
Patch: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-20000119.diff">mtd-20000119.diff</A><BR>
Tarball: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-20000119.tar.gz">mtd-20000119.tar.gz</A><BR>
<UL>
<LI>NFTL now registers the block device <I>before</I> asking the
gendisk code to scan it for partitions - it works better like
that :)</LI>
<LI>Various code cleanups, fixes for PPC and older kernels - mostly
thanks to Pavel Roskin.</LI>
<LI>No patch against 2.2 or 2.3.current kernels - I'll resume that
practice as soon as I've got more time to look after it.</LI>
</UL>
<BR>
</LI>
<LI><B>mtd-19991025</B><BR>
Patch: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19991025.diff">mtd-19991025.diff</A><BR>
Tarball: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19991025.tar.gz">mtd-19991025.tar.gz</A><BR>
Complete patch against 2.3.23 kernel: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19991025-against-2.3.23.diff.gz">
mtd-19991025-against-2.3.23.diff.gz</A><BR>
<UL>
<LI>Ported to 2.3 kernels (still compiles with 2.2)</LI>
<LI>NFTL write support seems fairly reliable and interoperates with
M-Systems' own drivers - more testing required</LI>
<LI>Various bugfixes</LI>
</UL>
<BR>
</LI>
<LI><B>mtd-19990820</B><BR>
Patch: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19990820.diff">mtd-19990820.diff</A><BR>
Tarball: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19990820.tar.gz">mtd-19990820.tar.gz</A><BR>
<UL>
<LI>Add <TT>nftl_format</TT> program
<I>(M-Systems' own drivers don't like the resulting NFTL yet)</I></LI>
<LI>First version of write support for NFTL <I>(Likewise)</I></LI>
</UL>
<BR>
</LI>
<LI><B>mtd-19990817</B><BR>
Patch: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19990817.diff">mtd-19990817.diff</A><BR>
Tarball: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19990817.tar.gz">mtd-19990817.tar.gz</A><BR>
<UL>
<LI>Support ECC for NFTL on DiskOnChip</LI>
<LI>Behave correctly with multiple DiskOnChip</LI>
<LI>Add primitive MAKEDEV script</LI>
<LI>Move include directory to make it easier to compile into kernel</LI>
<LI>Add <TT>/proc/mtd</TT> support</LI>
</UL>
<BR>
</LI>
<LI><B>mtd-19990810</B><BR>
Patch: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19990810.diff">mtd-19990810.diff</A><BR>
Tarball: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19990810.tar.gz">mtd-19990810.tar.gz</A><BR>
<UL>
<LI>Various endianness and wordsize fixes</LI>
<LI>Compile fixes for Linux/Alpha - DiskOnChip & NFTL now works</LI>
<LI>Sanity checks in NFTL driver to prevent oopsen on certain MTDs</LI>
</UL>
<BR>
</LI>
<LI><B>mtd-19990809</B><BR>
Tarball: <A HREF="ftp://ftp.linux-mtd.infradead.org/pub/mtd/mtd-19990809.tar.gz">mtd-19990809.tar.gz</A><BR>
<B>Large merge, including:</B>
<UL>
<LI>Preliminary drivers for DiskOnChip 2000 and NFTL filesystem</LI>
<LI>Microsoft FFS2 filesystem</LI>
<LI>Octagon-5066 and vmax301 drivers for SBC boards</LI>
<LI>Various other improvements from jgg.</LI>
<LI>Boot loader code</LI>
</UL>
</LI>
</UL>
<HR size=2>
<h2>MTD <I>User</I> modules</h2>
<P>These are the modules which provide interfaces that can be used
directly from userspace. The user modules currently planned include:</P>
<UL>
<LI><B>Raw character access:</B><BR>
A character device which allows direct access to the underlying
memory. Useful for creating filesystems on the devices, before
using some of the translation drivers below, or for raw storage
on infrequently-changed flash, or RAM devices.
</LI>
<LI><B><A NAME="mtdblock">Raw block access</a></B><BR>
A block device driver which allows you to pretend that the flash
is a normal device with sensible sector size. It actually works
by caching a whole flash erase block in RAM, modifying it as
requested, then erasing the whole block and writing back the modified
data.<BR>
This allows you to use normal filesystems on flash parts. Obviously
it's not particularly robust when you are writing to it - you lose
a whole erase block's worth of data if your read/modify/erase/rewrite
cycle actually goes read/modify/erase/poweroff. But for development,
and for setting up filesystems which are actually going to be
mounted read-only in production units, it should be fine.
<BR>
There is also a read-only version of this driver which doesn't have
the capacity to do the caching and erase/writeback, mainly for use
with uCLinux where the extra RAM requirement was considered too
large.
</LI>
<LI><B>Flash Translation Layer (FTL)</B></LI>
<LI><B>NFTL</B><BR>
Block device drivers which implement an FTL/NFTL filesystem on the
underlying memory device. FTL is fully functional. NFTL is currently
working for both reading and writing, but could probably do with
some more field testing before being used on production systems.
</LI>
<LI><B>Microsoft Flash Filing System</B><BR>
This provides a filesystem directly from the flash, rather than the
less efficient filesystem-on-FTL-on-flash setup. The current driver is
in alpha-testing, but should allow read-only access to such
filesystems.
</LI>
<LI><B>Journalling Flash File System</B><BR>
This was developed by <A HREF="http://www.axis.com/">Axis
Communications</A>, and provides a filesystem directly on the
flash, rather than emulating a block device. For more information,
see <A HREF="http://www.developer.axis.com/software/jffs/">Axis'
developer pages</A>.
</LI>
</UL>
<HR SIZE=2>
<h2>MTD hardware device drivers</h2>
<P>These provide physical access to memory devices, and are not used
directly - they are accessed through the <I>user</I> modules above.</P>
<UL>
<LI><B>On-board memory</B><BR>
Many PC chipsets are incapable of correctly caching system memory
above 64M or 512M. A driver exists which allows you to use this
memory with the linux-mtd system.
</LI>
<LI><B>PCMCIA devices</B><BR>
There are already a number of PCMCIA memory devices supported by Linux.
The MTD device subsystem is being designed with compatibility in mind,
so that the <A
HREF="http://hyper.stanford.edu/~dhinds/pcmcia/home.html">existing
PCMCIA drivers</A> can cleanly interface to the new MTD
subsystem. This work has not yet been undertaken, though.</LI>
<LI><B>Common Flash Interface (<A
HREF="http://www.pentium.com/design/flcomp/technote/cfi_1_1.htm">CFI</A>)
onboard memory</B><BR>
This is one of the most interesting options for small embedded systems.
With a filesystem directly on the flash, rather than through an extra
layer like FTL, we can get execute-in-place (XIP) of programs and
libraries on the flash, without needing to use up valuable system
memory as a cache for the executable code.<BR>
A patch to the Linux kernel which supports this is available, and it
should shortly be ported to the generic MTD system as a hardware
driver.</LI>
<LI><B>M-Systems' ISA Flash cards and DiskOnChip 1000</B><BR>
These devices are very simple - they seem to consist of a large array
of Intel flash memory, with a simple paging mechanism reminiscent of
EMS, which allows different pages of the flash to be mapped into the
device's 8K memory window. While working for Nortel a couple of years
ago, I developed drivers for these devices.<BR>
Unfortunately, I don't think the Disk-On-Chip 1000 is available any
more.</LI>
<LI><B>M-Systems' DiskOnChip 2000 and Millennium</B><BR>
M-Systems have now provided me with full documentation for the
DiskOnChip 2000 and Millennium. The current driver is capable of
full operation on a DiskOnChip 2000, and I have received a
sample DiskOnChip Millennium from M-Systems, so I should get round to
adding Millennium support soon. The write support for NFTL appears
to pass my simple tests for interoperability - after I mount it with
my drivers, thrash it for a while, and unmount it, the NFTL is still
recognised by M-Systems' own drivers. More testing of this would be
very much appreciated.
<BR>
<I>The <A HREF="doc2000.html">Linux and Disk-On-Chip
2000 page</A> is available here. Please note that, now that
M-Systems have provided me with complete specifications with which to
write a free driver for the device, there is no longer any reason
to avoid using it with Linux systems.</I>
<HR>
</LI>
<LI><B>CompactFlash</B> - <A HREF="http://www.compactflash.org/">http://www.compactflash.org/</A><BR>
CompactFlash emulates an IDE disk, either through the PCMCIA-ATA
standard, or by connecting directly to an IDE interface.
<BR>As such, it has no business being on this page, as to the best of
my knowledge it doesn't have any alternative method of accessing the
flash - you <I>have</I> to use the IDE emulation - I mention it here
for completeness.
</LI>
</UL>
<hr>
<address><a href="mailto:dwmw2@infradead.org">David Woodhouse</a></address>
$Id: index.html,v 1.23 2001/08/12 17:04:52 dwmw2 Exp $
</body>
</html>
|
