Has MTD production quality support for DOC2000?

Has MTD production quality support for DOC2000?

[Sigurd Urdahl]

I'm quite new at embedding my penguin, and will have to embedd it on a
DOC2000. But since I'm new at this I am quite confused... I am
considering going with MTD instead of m-sys drivers, but after
searching the archives I'm still not sure on wether MTD is stable
enough for production or not, so please help:=)

-sig

-- 
sigurd urdahl


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

Re: Has MTD production quality support for DOC2000?

[David Woodhouse]

sigurdur@ifi.uio.no said:
>  I'm quite new at embedding my penguin, and will have to embedd it on
> a DOC2000. But since I'm new at this I am quite confused... I am
> considering going with MTD instead of m-sys drivers, but after
> searching the archives I'm still not sure on wether MTD is stable
> enough for production or not, so please help:=) 

Not yet, if you want to write to the device. Nobody at the moment seems to 
be working very fast on it, and I've been busy with JFFS recently.

Linking the M-Systems drivers into your kernel will place _you_ in violation
of the GPL and you may be sued, so why not help out with the MTD drivers?

The support is mostly there - it just needs to be made a little more robust.

--
dwmw2




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

Re[2]: Has MTD production quality support for DOC2000?

[Mike Kilburn]

David Woodhouse <dwmw2@redhat.com> wrote:
> 
> Linking the M-Systems drivers into your kernel will place _you_ in violation
> of the GPL and you may be sued, so why not help out with the MTD drivers?
> 

Also, last I checked, the M-Systems driver disables interrupts for quite
a long time while writing to the device. This makes it unusable in
a real-time embedded system. Switch to a SanDisk if you can.




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

Re: Re[2]: Has MTD production quality support for DOC2000?

[Jon Burford]

I use the M-Systems driver with an RT-Linux-pacthed box and my RT threads
execute just fine.  Since the driver runs in kernel space and since the
kernel is just another RT thread, I can run higher priority threads which
preempt the kernel and therefore the driver itself.  We found the interrupt
disable quite annoying and moved all our RT stuff to RT threads and have not
had a problem.

Regards,
Jon

----- Original Message -----
From: "Mike Kilburn" <mkilburn@ftel.net>
To: <dwmw2@redhat.com>
Cc: <sigurdur@ifi.uio.no>; <mtd@infradead.org>
Sent: Thursday, August 10, 2000 4:56 PM
Subject: Re[2]: Has MTD production quality support for DOC2000?


> David Woodhouse <dwmw2@redhat.com> wrote:
> >
> > Linking the M-Systems drivers into your kernel will place _you_ in
violation
> > of the GPL and you may be sued, so why not help out with the MTD
drivers?
> >
>
> Also, last I checked, the M-Systems driver disables interrupts for quite
> a long time while writing to the device. This makes it unusable in
> a real-time embedded system. Switch to a SanDisk if you can.
>
>
>
>
> To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org



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

Re: Has MTD production quality support for DOC2000?

[Sigurd Urdahl]

David Woodhouse <dwmw2@redhat.com> writes:
> sigurdur@ifi.uio.no said:
> >  I'm quite new at embedding my penguin, and will have to embedd it on
> > a DOC2000. But since I'm new at this I am quite confused... I am
> > considering going with MTD instead of m-sys drivers, but after
> > searching the archives I'm still not sure on wether MTD is stable
> > enough for production or not, so please help:=) 
> 
> Not yet, if you want to write to the device. Nobody at the moment seems to 
> be working very fast on it, and I've been busy with JFFS recently.

:(

> Linking the M-Systems drivers into your kernel will place _you_ in violation
> of the GPL and you may be sued,

If the driver is compiled as a module and I use loadlin and initrd to
load it from a Dos-formatted DOC I thought I would be in the clear
_legally_, but I know it's not a very good idea tachnically or moraly.
And I'm probably not going to do it either - I'll rather wait for the
next prototype (there is a whole series of prototypes I'm going to
test - and no, I wasn't part of the team spec'ing them..) to arrive,
and hope it's non-DOC based:)

> so why not help out with the MTD drivers?

I would really like to if I thought I could help, but with my coding
skills and general knowledge of linux-drivers I think I would slow
down rather than speed up the development;)

-sig
-- 
sigurd urdahl


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

Re: Has MTD production quality support for DOC2000?

[Ollie Lho]

Sigurd Urdahl wrote:
> 
> 
> I would really like to if I thought I could help, but with my coding
> skills and general knowledge of linux-drivers I think I would slow
> down rather than speed up the development;)
> 

Well, actually any contribution is wellcomed. At lease you can stress test
the driver to find the exact combination of read/write which corrupt the
filesystem on the DOC.

Ollie


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

Re: Has MTD production quality support for DOC2000?

[Rogelio M. Serrano Jr.]

Sigurd Urdahl wrote:

> David Woodhouse <dwmw2@redhat.com> writes:
> > sigurdur@ifi.uio.no said:
> > >  I'm quite new at embedding my penguin, and will have to embedd it on
> > > a DOC2000. But since I'm new at this I am quite confused... I am
> > > considering going with MTD instead of m-sys drivers, but after
> > > searching the archives I'm still not sure on wether MTD is stable
> > > enough for production or not, so please help:=)
> >
> > Not yet, if you want to write to the device. Nobody at the moment seems to
> > be working very fast on it, and I've been busy with JFFS recently.
>

I have been writing to the DOC for weeks now under jffs. I have given up
on NFTL
completely. I have patches here but I will not submit them because I
have not
tested them with NFTL except for the multipage read write support.
Anyone who is
interested can email me. I got a little lost in the middle of all the
activity on
both MTD and JFFS. I have to get in sync with all the developments and
all.


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

Re: Has MTD production quality support for DOC2000?

[Stuart Lynne]

In article <ylbvgx88b7i.fsf@ganglot.ifi.uio.no>,
Sigurd Urdahl <sigurdur@ifi.uio.no> wrote:
>David Woodhouse <dwmw2@redhat.com> writes:
>> sigurdur@ifi.uio.no said:

>> Linking the M-Systems drivers into your kernel will place _you_ in violation
>> of the GPL and you may be sued,

Only if you give someone else a copy of the kernel.... you are not violating
GPL if you use it for your own use.

>If the driver is compiled as a module and I use loadlin and initrd to
>load it from a Dos-formatted DOC I thought I would be in the clear
>_legally_, but I know it's not a very good idea tachnically or moraly.

Yes, legal. Personally I think it is is no better or worse technically. And
I'm not sure how it could be bad morally.... 

There will be a new driver available from M-Systems in the near future
that makes it easier to use as a module.

-- 
                                            __O 
Fireplug - a Lineo company                _-\<,_ 
PGP Fingerprint: 28 E2 A0 15 99 62 9A 00 (_)/ (_) 88 EC A3 EE 2D 1C 15 68
Stuart Lynne <sl@fireplug.net>       www.fireplug.net        604-461-7532


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

Re: Re[2]: Has MTD production quality support for DOC2000?

[Stuart Lynne]

In article <200008102356.QAA02347@devel.ftel.com>,
Mike Kilburn <mkilburn@ftel.net> wrote:
>David Woodhouse <dwmw2@redhat.com> wrote:
>> 
>> Linking the M-Systems drivers into your kernel will place _you_ in violation
>> of the GPL and you may be sued, so why not help out with the MTD drivers?
>> 
>
>Also, last I checked, the M-Systems driver disables interrupts for quite
>a long time while writing to the device. This makes it unusable in
>a real-time embedded system. Switch to a SanDisk if you can.

The new driver addresses this issue. It is not actually the driver that 
disables interrupts, the block layer does it before calling the device
request function. And in some cases the driver can take some time writing
due to erasing. The new driver will do a schedule_timeout() instead of
busy waiting.

-- 
                                            __O 
Fireplug - a Lineo company                _-\<,_ 
PGP Fingerprint: 28 E2 A0 15 99 62 9A 00 (_)/ (_) 88 EC A3 EE 2D 1C 15 68
Stuart Lynne <sl@fireplug.net>       www.fireplug.net        604-461-7532


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