On Wed, Jul 25, 2012 at 2:03 PM, Stefan Tauner <stefan.tauner@student.tuwien.ac.at> wrote:
On Wed, 25 Jul 2012 12:14:33 -0700
David Hendricks <dhendrix@google.com> wrote:

> There's a SPI chip named "N25Q064" with 1.8V and 3V variants N25Q064A11 and
> N25Q064A13, respectively. The 1.8V has A11 suffixed, and the 3V version has
> A13 suffixed. The 'A' indicates the process technology which is likely to
> change without a change of device ID, so it's likely that we'll see
> something like N25Q064B11 or N25Q064B13 in the future.
>
> Fortunately, the two variants have different JEDEC IDs (0xBA17 and 0xBB17).
> So at least we can tell them apart via probing.
>
> stefanct and I were discussing how to name this chip couldn't really come
> up with a good solution. I am thinking that "N25Q064" is sufficient and
> that we should just re-use the same .name for two chip entries with
> different .voltage parameters.
>
> Thoughts?
>

datasheets for the devices in question:
<dhendrix> 3V version: http://www.micron.com/~/media/Documents/Products/Data%20Sheet/NOR%20Flash/Serial%20NOR/N25Q/n25q_64mb_3v_65nm.pdf
<dhendrix> 1.8V version: http://www.micron.com/~/media/Documents/Products/Data%20Sheet/NOR%20Flash/Serial%20NOR/N25Q/n25q_64mb_1_8v_65nm.pdf

david: you dont give a rationale why you think that N25Q064 is enough.

my argumnt why it is not enough: the .name is used to distinguish
what flashrom defines as individual chips when presenting them to the
user. so it is not enough that flashrom can distinguish the chips by
their IDs but also the user must be able to tell the difference hence
there must be no duplicative names.

fwiw: i think we need a place holder for "any character" in model names
and propose "."
so in this case:
N25Q064..1
N25Q064..3
rationale: easily understandable, regex semantic, not part of actual
chip names, non-distracting.

I like this. It's easy on the eyes, abstracts irrelevant details, and can be reasonably expected to scale.

We can't predict all the possible parts the vendor will come up with, but so long as they have a well-defined part numbering scheme and we can represent irrelevant letters and numbers with dots we can at least avoid being wrong.

I'll update my other patch using this schema.
 
alternatively we could just ignore that problem (it is a future problem
after all) and use N25Q064A11 and N25Q064A13.

Please let's not do that :-) It causes headaches if the information presented falls out of date or shows up incorrect.

--
David Hendricks (dhendrix)
Systems Software Engineer, Google Inc.