linuxppc-dev.lists.ozlabs.org archive mirror
 help / color / mirror / Atom feed
* NAND only (no NOR)
@ 2008-12-02  8:55 Norbert van Bolhuis
  2008-12-02  9:31 ` [U-Boot] " Alessandro Rubini
  2008-12-03  2:50 ` Sean MacLennan
  0 siblings, 2 replies; 7+ messages in thread
From: Norbert van Bolhuis @ 2008-12-02  8:55 UTC (permalink / raw)
  To: linuxppc-dev, u-boot


Nowadays, do many (PowerPC) embedded devices already risk omitting
NOR flash and use a NAND device solely for booting and storing images ?

I'm talking about systems with 10 years life-cycle (so no
MP3-players nor medical systems but somewhere in between).

We have a MPC8313E-RDB and I know booting from NAND is
possible. U-boot seems to support it well, I assume it can read
the kernel, ramdisk and dtb from NAND in memory and bootm this.

However, having no NOR flash means:
  - NAND should be programmable via JTAG (BDI3000 doesn't support
    this, Lauterbach/trace32 does)
  - Critical software images (u-boot, kernel, dtb) stored on
    NAND, while there's no NAND 'scrubbing', etc..

Thoughts/comments are welcome.

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [U-Boot] NAND only (no NOR)
  2008-12-02  8:55 NAND only (no NOR) Norbert van Bolhuis
@ 2008-12-02  9:31 ` Alessandro Rubini
  2008-12-03  2:50 ` Sean MacLennan
  1 sibling, 0 replies; 7+ messages in thread
From: Alessandro Rubini @ 2008-12-02  9:31 UTC (permalink / raw)
  To: nvbolhuis; +Cc: linuxppc-dev, u-boot

Disclaimer: I'm not running ppc these days

> However, having no NOR flash means:
>   - NAND should be programmable via JTAG (BDI3000 doesn't support
>     this, Lauterbach/trace32 does)

My personal preference for bringing up a new board is placing u-boot
in RAM using the available information to setup SDRAM from the jtag
tool (or, better, place the initial boot code the manufacturer give
you in static memory, and jump to is with a breakpoint in the right
place).

While my debugger is able to write flash, I've always found it easier
to do that from u-boot itself. Besides, I can fix my u-boot in RAM
without the lengthy erase-reprogram cycles.

If you work like this, it's irrelevant whether it is nor, nand or dataflash
or whatever, as long as u-boot is able to manage it.

hope this helps
/alessandro

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: NAND only (no NOR)
  2008-12-02  8:55 NAND only (no NOR) Norbert van Bolhuis
  2008-12-02  9:31 ` [U-Boot] " Alessandro Rubini
@ 2008-12-03  2:50 ` Sean MacLennan
  2008-12-03  5:48   ` [U-Boot] " Stefan Roese
  1 sibling, 1 reply; 7+ messages in thread
From: Sean MacLennan @ 2008-12-03  2:50 UTC (permalink / raw)
  To: Norbert van Bolhuis; +Cc: linuxppc-dev, u-boot

On Tue, 02 Dec 2008 09:55:24 +0100
"Norbert van Bolhuis" <nvbolhuis@aimvalley.nl> wrote:

> 
> Nowadays, do many (PowerPC) embedded devices already risk omitting
> NOR flash and use a NAND device solely for booting and storing
> images ?
> 
> I'm talking about systems with 10 years life-cycle (so no
> MP3-players nor medical systems but somewhere in between).
> 
> We have a MPC8313E-RDB and I know booting from NAND is
> possible. U-boot seems to support it well, I assume it can read
> the kernel, ramdisk and dtb from NAND in memory and bootm this.

We use a 256M NAND on the PIKA Warp appliance and where unable to boot
u-boot from the NAND. It worked on a smaller 64M NAND.

So we put our FPGA and u-boot in a small NOR. The kernel, rootfs and
other files are on the NAND.

Cheers,
   Sean

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [U-Boot] NAND only (no NOR)
  2008-12-03  2:50 ` Sean MacLennan
@ 2008-12-03  5:48   ` Stefan Roese
  2008-12-03  6:38     ` Sean MacLennan
  0 siblings, 1 reply; 7+ messages in thread
From: Stefan Roese @ 2008-12-03  5:48 UTC (permalink / raw)
  To: u-boot; +Cc: linuxppc-dev, Norbert van Bolhuis, Sean MacLennan

On Wednesday 03 December 2008, Sean MacLennan wrote:
> We use a 256M NAND on the PIKA Warp appliance and where unable to boot
> u-boot from the NAND. It worked on a smaller 64M NAND.

Currently you need to define the page-size of the NAND used for booting (512 
bytes vs. 2k). The 2k 4xx NAND booting support is was done about 1/2 a year 
ago. So perhaps you tested this when this 2k support was not available yet.

> So we put our FPGA and u-boot in a small NOR. The kernel, rootfs and
> other files are on the NAND.

Yes, I would recommend to do it this way if possible. A small NOR for U-Boot 
and environment and everything else in NAND. This makes things much easier. 
But I understand that this is sometimes a problem with space (2 FLASH chips) 
and costs.

Best regards,
Stefan

=====================================================================
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80  Email: office@denx.de
=====================================================================

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [U-Boot] NAND only (no NOR)
  2008-12-03  5:48   ` [U-Boot] " Stefan Roese
@ 2008-12-03  6:38     ` Sean MacLennan
  2008-12-03  7:40       ` Trent Piepho
  0 siblings, 1 reply; 7+ messages in thread
From: Sean MacLennan @ 2008-12-03  6:38 UTC (permalink / raw)
  To: Stefan Roese; +Cc: u-boot, Norbert van Bolhuis, linuxppc-dev

On Wed, 3 Dec 2008 06:48:57 +0100
"Stefan Roese" <sr@denx.de> wrote:


> Currently you need to define the page-size of the NAND used for
> booting (512 bytes vs. 2k). The 2k 4xx NAND booting support is was
> done about 1/2 a year ago. So perhaps you tested this when this 2k
> support was not available yet.

We are running 1.3.0, so you are probably right. It may no longer be a
problem.

> Yes, I would recommend to do it this way if possible. A small NOR for
> U-Boot and environment and everything else in NAND. This makes things
> much easier. But I understand that this is sometimes a problem with
> space (2 FLASH chips) and costs.

Mainly cost. We didn't want to pay for a second chip.

Cheers,
   Sean

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [U-Boot] NAND only (no NOR)
  2008-12-03  6:38     ` Sean MacLennan
@ 2008-12-03  7:40       ` Trent Piepho
  2008-12-03 13:57         ` Leon Woestenberg
  0 siblings, 1 reply; 7+ messages in thread
From: Trent Piepho @ 2008-12-03  7:40 UTC (permalink / raw)
  To: Sean MacLennan; +Cc: u-boot, Stefan Roese, Norbert van Bolhuis, linuxppc-dev

On Wed, 3 Dec 2008, Sean MacLennan wrote:
>> Yes, I would recommend to do it this way if possible. A small NOR for
>> U-Boot and environment and everything else in NAND. This makes things
>> much easier. But I understand that this is sometimes a problem with
>> space (2 FLASH chips) and costs.
>
> Mainly cost. We didn't want to pay for a second chip.

I think for NAND the latches necessary to de-multiplex the localbus aren't
necessary like they are for NOR?  On our board the latches might even take
more space than the flash chip.

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: [U-Boot] NAND only (no NOR)
  2008-12-03  7:40       ` Trent Piepho
@ 2008-12-03 13:57         ` Leon Woestenberg
  0 siblings, 0 replies; 7+ messages in thread
From: Leon Woestenberg @ 2008-12-03 13:57 UTC (permalink / raw)
  To: Trent Piepho
  Cc: linuxppc-dev, u-boot, Stefan Roese, Norbert van Bolhuis,
	Sean MacLennan

Hello,

On Wed, Dec 3, 2008 at 8:40 AM, Trent Piepho <tpiepho@freescale.com> wrote:
> On Wed, 3 Dec 2008, Sean MacLennan wrote:
>>> Yes, I would recommend to do it this way if possible. A small NOR for
>>> U-Boot and environment and everything else in NAND. This makes things
>>> much easier. But I understand that this is sometimes a problem with
>>> space (2 FLASH chips) and costs.
>>
>> Mainly cost. We didn't want to pay for a second chip.
>
> I think for NAND the latches necessary to de-multiplex the localbus aren't
> necessary like they are for NOR?  On our board the latches might even take
> more space than the flash chip.

The local bus can be configured to run non-multiplexed.

Regards,
-- 
Leon

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2008-12-03 13:57 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-12-02  8:55 NAND only (no NOR) Norbert van Bolhuis
2008-12-02  9:31 ` [U-Boot] " Alessandro Rubini
2008-12-03  2:50 ` Sean MacLennan
2008-12-03  5:48   ` [U-Boot] " Stefan Roese
2008-12-03  6:38     ` Sean MacLennan
2008-12-03  7:40       ` Trent Piepho
2008-12-03 13:57         ` Leon Woestenberg

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).