[ECOS] Re: eCos API to get random value or UID value?

Grant Edwards grant.b.edwards@gmail.com
Sun Mar 20 20:23:00 GMT 2011


On 2011-03-18, Stanislav Meduna <stano@meduna.org> wrote:
> On 18.03.2011 20:52, Grant Edwards wrote:
>
>> Is there any eCos API support that the HAL can use to provide either
>> random data or a unique id value to application or OS code?
>
> I don't know of any.

I've done some more grepping, but haven't found anything.

>> Both random values and UUID values are required for a lot of network
>> stuff.  It's pretty common for hardware platforms to one or both,
>> there doesn't seem to be an eCos API to access it.
>
> This is a very good request.

Decent sources for unique IDs are probably a lot more common than
decent entropy sources, but it seems like something for which we need
an API.  A UID can be pretty easily generated by hashing a few things
(CPU serial number, flash serial number, MAC addresses, etc.).
Generating entropy without hardware support is a lot harder.

> However, I don't think this belongs into the HAL - it's more a device
> driver semantics (see e.g. /dev/{u,}random on linux) and the
> cyg_io_lookup / cyg_io_read is everything one needs here.
>
> What the HAL could provide is an interface supplying a source of
> entropy

That's pretty much what I meant.

> - however, this is hard to do correctly especially on an embedded
> system that is often running in a quite deterministic environment.

Indeed.  Even if it's not cryptographically strong, something to use
for for things like pseuro-random retry-backoff would be great.

-- 
Grant Edwards               grant.b.edwards        Yow! How many retured
                                  at               bricklayers from FLORIDA
                              gmail.com            are out purchasing PENCIL
                                                   SHARPENERS right NOW??


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss



More information about the Ecos-discuss mailing list