Darren J Moffat wrote:
> Cecilia Hu wrote:
>> <Apologize for resending mail to alias, because of typo on subject....>
>>
>>
>> I am sponsoring this case for Quaker Fang. It is to provide a new
>> wireless(802.11b/g) driver, ural(7D), for Ralink RT2500USB chipset.
>> The requested release binding is micro/patch.
>>
>> The architecture and interface are clear enough, so this case is
>> marked as "closed approved automatic".
>
> It isn't clear to me why a separate driver from the existing ral(7D)
> is needed because this is attached via USB rather that PCI. Is the
> RT2500USB that different to the RT2500 that the ral(7D) driver
> supports that a new driver is needed ?
>
> However I did a tiny bit of research and found that *BSD have a
> similar ral/ural split so it seems that other operating systems chose
> this split.
For Solaris devices, the DDI for USB is *vastly* different than the PCI
DDI at some core fundamental concepts. It would be much harder to have
a unified RAL driver, than (for example) a unified pcwl (PCI/PCMCIA).
It may be possible to have a common library of shared functionality, but
that may be more work than it is worth.
Put another way, I think Quaker/Cecilia's plan here makes perfect sense.
>
> Also is this case setting any precedent that we should attempt to
> prefix USB NIC drivers with 'u' ?
I don't read it that way. It may make sense to use a 'u' prefix
sometimes (as in this case), but I think we should still do it on a case
by case basis.
Anyway, this is a fasttrack (or maybe even self-review?), so I'm not
sure we can even use it to set precedent like that. (Can you set
precedent without a written opinion?)
-- Garrett
