M-System DOC Millennium with MTD.

M-System DOC Millennium with MTD.

[Simon Wood]

Hello all,
I attempting to get a Millennium Disk on Chip eval board working under Linux with the MTD drivers (rather than the M-System's ones - which appear to work OK).

I've downloaded, applied the patches to the kernel and built the MTD stuff as modules. However I can not read the device, instead I get the error 'mount: /dev/nftla has wrong major or minor number'.

Any idea what is going wrong??

Lots of details.
1). ran 'mtd/utils/MAKEDEV' to set up devices -
	/dev/nftla is 93, 0

2). insmod'ed the modules (without any options) in the following order -
	nftlmount
	mtdcore
	nftl
	docprobe
	docecc
	doc2001

3). Tried to 'mount /dev/nftla /floppy' and got error 'mount: /dev/nftla has wrong major or minor number'.

4) tail /var/log/messages -
		Nov 29 15:57:03 leper kernel: Initializing MTD Layer
		Nov 29 15:57:23 leper kernel: M-Systems NAND Flash Translation Layer driver. (C) 1999 MVHI
		Nov 29 15:57:23 leper kernel: $Id: mtd-patch-2.2.17,v 1.3 2000/11/15 22:56:48 sjhill Exp $
		Nov 29 15:57:33 leper kernel: M-Systems DiskOnChip driver. (C) 1999 Machine Vision Holdings, Inc.
		Nov 29 15:58:09 leper kernel: NFTL_open
		Nov 29 15:58:09 leper kernel: ENODEV: thisNFTL = 0, minor = 23808, ip = c33b9a90, fp = c301f980
		Nov 29 15:58:09 leper kernel: NFTL_open
		Nov 29 15:58:09 leper kernel: ENODEV: thisNFTL = 0, minor = 23808, ip = c33b9a90, fp = c27fff64

5). Just for completion here's what the M-Systems drivers log -
		Nov 29 15:45:56 leper kernel: Copyright (C) 1998,2000 M-Systems Flash Disk Pioneers Ltd.
		Nov 29 15:45:56 leper kernel: Copyright (C) 2000 Lineo
		Nov 29 15:45:56 leper kernel: DOC device(s) found: 1
		Nov 29 15:45:56 leper kernel: Fat Filter Enabled
		Nov 29 15:45:56 leper kernel: fl_init: registered device at major: 100
		Nov 29 15:45:56 leper kernel: fl :fl_doc_mount: VOL: 0
		Nov 29 15:45:56 leper kernel: fl :debug: entering NFDC 2148 identification routine.
		Nov 29 15:45:56 leper kernel: fl :Debug: identified NFDC 2148.
		Nov 29 15:45:56 leper kernel: fl :Debug: starting NFTL mount.
		Nov 29 15:45:56 leper kernel: fl :Debug: finished NFTL mount.
		Nov 29 15:45:56 leper kernel:  fla: fla1
		Nov 29 15:45:56 leper kernel: partition: 0: start_sect: 0, nr_sects: 3e30 Fl_blk_size[]: 1f18kb
		Nov 29 15:45:56 leper kernel: partition: 1: start_sect: 1, nr_sects: 3e2f Fl_blk_size[]: 1f17kb
		Nov 29 15:46:18 leper kernel:  fla: fla1

Thanks in advance, I'm sure its something simple that I've missed but it's been a long day.......
Simon Wood



To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: M-System DOC Millennium with MTD.

[David Woodhouse]

Simon.Wood@pace.co.uk said:
> 2). insmod'ed the modules (without any options) in the following order
> -
> 	nftlmount
> 	mtdcore
> 	nftl
> 	docprobe
> 	docecc
> 	doc2001 

Load docprobe last. We are going to merge the code so you only have to load 
nftl.o and the DiskOnChip driver (we'll probably call it doc200x.o) and 
then it won't matter what order you load them in. Not quite yet though.

Simon.Wood@pace.co.uk said:
>  3). Tried to 'mount /dev/nftla /floppy' and got error 'mount: /dev/
> nftla has wrong major or minor number'.

It hasn't found the DiskOnChip. mount(8) ought to give you a better error 
than that though.

--
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: M-System DOC Millennium with MTD.

[dr john halewood]

On Thu, 30 Nov 2000, David Woodhouse wrote:
>Simon.Wood@pace.co.uk said:
>>  3). Tried to 'mount /dev/nftla /floppy' and got error 'mount: /dev/
>> nftla has wrong major or minor number'.
>
>It hasn't found the DiskOnChip. mount(8) ought to give you a better error 
>than that though.

One tiny possibility: I've managed to get something similar when playing with
some spare DoCs I've got lying around (they're 8Mb 2001s salvaged from broken
wince thin clients). On loading them with the MSys firmware image doc2.fff,
which they suggest using with Linux, I get the following, which is similar to
Simon's error (system is 2.2.17+CVS from infradead.org):
Initializing MTD Layer
Giving out device 0 to M-Systems DiskOnChip 1000
M-Systems DiskOnChip driver. (C) 1999 Machine Vision Holdings, Inc.
M-Systems NAND Flash Translation Layer driver. (C) 1999 MVHI
$Id: mtd-patch-2.2.17,v 1.3 2000/11/15 22:56:48 sjhill Exp $
NFTL_notify_add for M-Systems DiskOnChip 1000
NFTL_open
ENODEV: thisNFTL = 0, minor = 23809, ip = c38c7320, fp = c37fbf68

...which isn't really suprising as it's misdetecting the chip for something it
isn't.  Reloading the DoC firmware with the MSys bloatware fixes it however.
(No, I haven't worked out how it misidentifies it yet).
 This sort of thing probably belongs in an FAQ/mini-HOWTO. I've made quite a
lot of notes already(1) which I'll brush up and submit if anyone's interested
or thinks it's worthwhile. 

cheers
john

(1) I'll leave out the bits about how to remove a DoC with a Swiss Army knife,
but the bit about which pins you can bend or break without it mattering too
much might be useful to other hardware thugs like myself ;-)


To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: M-System DOC Millennium with MTD.

[David Woodhouse]

john@frumious.unidec.co.uk said:
>  ...which isn't really suprising as it's misdetecting the chip for
> something it isn't.  Reloading the DoC firmware with the MSys
> bloatware fixes it however. (No, I haven't worked out how it
> misidentifies it yet).

Hmmm. Turn off the doc1000 driver. They're not making the hardware any more.

john@frumious.unidec.co.uk said:
>  This sort of thing probably belongs in an FAQ/mini-HOWTO. I've made
> quite a lot of notes already(1) which I'll brush up and submit if
> anyone's interested or thinks it's worthwhile.  


Yes, please.

--
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

RE: M-System DOC Millennium with MTD.

[Simon Wood]

Hello again David,
I've updated my patch to your latest CVS (mtd-patch-2.2.17-20001130) and built that, but alas still no joy.

Just to re-iterate what I have and am doing.
I have a DoC Millennium (don't know how different this is to the 2000) eval board with a single MD-2810-D08 fitted. Jumpers are set to JP1 3,4 (0xD000) and JP2 3,4 (No connect, No wait state disabled).

Following the instructions below I get the same response as before, from /var/log/messages (I have set debug level to 3 but doesn't seem to produce anything extra):
	Nov 30 15:28:16 leper syslogd 1.3-3: restart.
	Nov 30 15:28:19 leper kernel: klogd 1.3-3, log source = /proc/kmsg started.
	Nov 30 15:28:19 leper kernel: Cannot find map file.
	Nov 30 15:28:19 leper kernel: Loaded 58 symbols from 2 modules.
	Nov 30 15:28:19 leper /usr/sbin/cron[155]: (CRON) STARTUP (fork ok) 
	Nov 30 15:34:04 leper kernel: Initializing MTD Layer
	Nov 30 15:34:23 leper kernel: M-Systems DiskOnChip driver. (C) 1999 Machine Vision Holdings, Inc.

And /proc/mtd is empty

However if I use the DoC1000 driver I get more of a response...
	Nov 30 15:35:34 leper kernel: mtd: Giving out device 0 to M-Systems DiskOnChip 1000
	Nov 30 15:35:44 leper kernel: M-Systems DiskOnChip driver. (C) 1999 Machine Vision Holdings, Inc.

And /proc/mtd reports 'mtd0: 00100000 "M-Systems DiskOnChip 1000"' - which it most definately is not !!

If nobody else can come up with a suggestion tomorrow I'll try reflashing the device's boot code from DOS. Any way it's time to wander off home now.
Simon Wood

Hardware Engineer 
Pace Micro Technology plc
Victoria Road, Saltaire, Shipley
West Yorkshire, BD18 3LF
Tel : +44(0)1274 532000  Fax: +44(0)1274 532029

This E-Mail and any attachments hereto are strictly confidential and intended solely for the addressee. If you are not the intended addressee please notify the sender by return and delete the message. You must not disclose, forward or copy this E-mail or attachments to any third party without the prior consent of the sender.

> -----Original Message-----
> From:	David Woodhouse [SMTP:dwmw2@infradead.org]
> Sent:	Thursday, November 30, 2000 11:00 AM
> To:	Simon Wood
> Subject:	Re: M-System DOC Millennium with MTD. 
> 
> 
> Build the stuff you want as modules.
> 
> Load mtdcore.o
> Load docecc.o 
> Load doc2001.o
> Load docprobe.o
> 
> You should see it say something about detecting a DiskOnChip. 
> 'cat /proc/mtd' at this stage should then produce one line of text.
> 
> 


To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: M-System DOC Millennium with MTD.

[David Woodhouse]

Simon.Wood@pace.co.uk said:
> 	Nov 30 15:34:23 leper kernel: M-Systems DiskOnChip driver. (C) 1999
> Machine Vision Holdings, Inc.

> And /proc/mtd is empty 

Try changing the jumpers so the DiskOnChip appears at a different location. 
Also try setting CONFIG_MTD_DOCPROBE_ADDRESS to the address at which you know 
the device is mapped.

Can you show me the contents of your .config? At least the CONFIG_MTD* 
parts.
--
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

RE: M-System DOC Millennium with MTD.

[dr john halewood]

On Thu, 30 Nov 2000, Simon Wood wrote:
>Hello again David,
>I've updated my patch to your latest CVS (mtd-patch-2.2.17-20001130) and built that, but alas still no joy.
>
>Just to re-iterate what I have and am doing.
>I have a DoC Millennium (don't know how different this is to the 2000) eval board with a single MD-2810-D08 fitted. Jumpers are set to JP1 3,4 (0xD000) and JP2 3,4 (No connect, No wait state disabled).
>
>And /proc/mtd is empty
>
>However if I use the DoC1000 driver I get more of a response...
>	Nov 30 15:35:34 leper kernel: mtd: Giving out device 0 to M-Systems DiskOnChip 1000
>	Nov 30 15:35:44 leper kernel: M-Systems DiskOnChip driver. (C) 1999 Machine Vision Holdings, Inc.
>
>And /proc/mtd reports 'mtd0: 00100000 "M-Systems DiskOnChip 1000"' - which it most definately is not !!
>

Yup. Been there, done that. What's happened is that when you used the M-Systems
utilities you put the doc2.fff firmware onto the flash. This is a dud firmware
which M-Systems use to get around a problem with lilo/their MTD driver (see the
MSys pdf document about using Doc with Linux for more explanation). Reflash the
DoC with another image (doc42.exb or something) and it'll reappear.
  The problem, I think (haven't gone far enough through the code) is that this
firmware doesn't have the hooks that the MTD drivers use to identify the chip,
but there is something in there that looks enough like the doc1000 signature
for it to be mis-identified.
  FWIW this is purely theoretical, but what you've done exactly follows what I
did about a month ago, and it's working here again now.

cheers
john


To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: M-System DOC Millennium with MTD.

[David Woodhouse]

john@frumious.unidec.co.uk said:
>   The problem, I think (haven't gone far enough through the code) is
> that this firmware doesn't have the hooks that the MTD drivers use to
> identify the chip, but there is something in there that looks enough
> like the doc1000 signature for it to be mis-identified.
>   FWIW this is purely theoretical, but what you've done exactly
> follows what I did about a month ago, and it's working here again now.
> 

You can also work round this by disabling CONFIG_MTD_DOCPROBE_55AA


--
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: M-System DOC Millennium with MTD.

[john halewood]

David Woodhouse wrote, re misidentification/failure to identify DoC's
 with the M-Systems doc2.fff firmware:
>You can also work round this by disabling CONFIG_MTD_DOCPROBE_55AA

Yup. you also need to disable DOCPROBE_HIMEM as well, otherwise you just
get a lot of messages of the form "found possible Doc at XXXXX" and no
mtd device.
 I've been messing with these parameters so much recently I've sometimes
felt like hacking a /proc interface to them to avoid the reconfig -
recompile modules - rmmod /insmod cycle, which can be quite painful on a slow
processor. In the long run though i think pollution of the /proc space
is probably left to those who've done it so well in the past

john


To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: M-System DOC Millennium with MTD.

[David Woodhouse]

On Sun, 3 Dec 2000, john halewood wrote:

> David Woodhouse wrote, re misidentification/failure to identify DoC's
>  with the M-Systems doc2.fff firmware:
> >You can also work round this by disabling CONFIG_MTD_DOCPROBE_55AA
>
> Yup. you also need to disable DOCPROBE_HIMEM as well, otherwise you just
> get a lot of messages of the form "found possible Doc at XXXXX" and no
> mtd device.

Yep. Although I've now removed the annoying complaints if DOCPROBE_55AA
isn't defined.

>  I've been messing with these parameters so much recently I've sometimes
> felt like hacking a /proc interface to them to avoid the reconfig -
> recompile modules - rmmod /insmod cycle, which can be quite painful on a slow
> processor.

This is what MODULE_PARM was invented for :)

-- 
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: M-System DOC Millennium with MTD.

[David Woodhouse]

On Sun, 3 Dec 2000, john halewood wrote:

>  I've been messing with these parameters so much recently I've sometimes
> felt like hacking a /proc interface to them to avoid the reconfig -
> recompile modules - rmmod /insmod cycle, which can be quite painful on a slow
> processor.

cvs up and try again:

insmod docprobe.o doc_config_location=0xd0000

-- 
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org