Re: Uniqueness of die-id.

[d...@ucore.info]

2011/9/16 Dawid Ciężarkiewicz <d...@ucore.info>:
> Hi,
>
> I'm about to use die-id to generate unique instance ID for the
> platform I'm developing that support Panda. AFAIK it should be unique
> on every processor. However it seems not.
>
> I'm geting the value:
>
> (1D620003-00000001-0612D687-01000)
>
> which seems to be quite common. Just google for 1d620003 00000001
> 0612d687 and you will see there are people getting it as well.
>
> Can someone confirm if the OMAP 4430 die-id should be unique? Any idea
> why I'm getting this non-unique value?

Hate to repost, but I'm really interested in solving this. I'm
wondering is there any documentation clarifying the uniqueness of
die-id. In Omap's TRM I was able to find the address in memory for a
die-id, not the definition of what it's supposed to store.

Any help is much appreciated,

--
Dawid Ciężarkiewicz

[Joshi, Vikas]

U-boot code can read die-id and print. Look at: http://www.mail-archive.com/u-b...@lists.denx.de/msg19890.html
Also mention how are you getting the die-id?

Vikas

[d...@ucore.info]

Yes. I'm using the same method as u-boot. It's just reading memory
from some defined physical address. Some newer versions of Linux are
reading and printing die-id as well.

I'm getting value that looks OK, however I got same value on different
board and afterward was able to google some logs of other people with
exactly same value. This makes me doubt if die-id is meant to be
unique.

On Mon, Sep 19, 2011 at 12:16 PM, Joshi, Vikas <v-jo...@ti.com> wrote:
> U-boot code can read die-id and print. Look at: http://www.mail-archive.com/u-b...@lists.denx.de/msg19890.html
> Also mention how are you getting the die-id?

--
Dawid Ciężarkiewicz

[GrueMaster]

I can confirm that on Panda A1, not all die ids are unique. Our new
build cluster at Canonical has at least two duplicates, and I had two
duplicates at home for testing. I think Linaro has also seen
duplicates. u-boot and the smsc95xx driver in the kernel use the die
id to generate the mac address for the network interface. This can
cause a lot of problems on a home network that is setup for dhcp.

I don't know the solution to this, but I just wanted to confirm what
is being seen.

Tobin Davis
Canonical QA

[Anand Gadiyar]

Tobin,

Can you share the duplicate die-ids that you see?

Also, can you tell me which registers are being read to get this - or point me to the code you're using and I'll take a look.

- Anand

[Anand Gadiyar]

As another option, you can run the following command in u-boot to read the registers:
md 0x4a002200 5

This dumps 5 registers: DIE_ID_0, ID_CODE, DIE_ID_1, DIE_ID_2, DIE_ID_3.

Let me know what you get when you read this and if this differs from the value u-boot gets. If there is a difference, then u-boot code probably needs to be updated.

- Anand