Eric Sultan wrote:
> This is almost correct. The project doesn't plan to deliver into the
> nevada builds, but if it becomes important to do so, your first point
> would be correct. Your second and third points are correct statements.
I think it would be best to deliver into Nevada and Indiana at the same
time, and possibly revoke pfb/nfb at the same time.
Unless there are compelling reasons not to do this?
-- Garrett
>
> -- Eric
>
>
> Alan Coopersmith wrote:
>> Eric Sultan wrote:
>>
>>> The timer having expired with no unresolved issues, I am marking this
>>> case approved.
>>>
>>
>> Sorry for the late response, but I was travelling last week and am just
>> catching up on e-mail now.
>>
>> Would the following be a correct summary of the final spec?
>>
>> - In the "Nevada" WOS builds (aka SXCE), both the existing drivers (pfb
>> & nfb) and the new driver (efb) will be delivered. Which is loaded
>> depends on the ordering in /etc/driver_aliases, which is controlled
>> by package installation order, and thus will be random depending on
>> installation method & WOS media layout for the build in question.
>> All users of XVR-100 & XVR-300 cards on Nevada will thus be responsible
>> for knowing which driver is for Xsun vs. Xorg and manually pkgrm'ing
>> the driver they don't want after each OS install/upgrade. How will
>> SXCE & internal Nevada users be informed of this flag day/installation
>> requirement?
>>
>> - In the "Indiana" WOS builds (aka OpenSolaris), only efb will be delivered,
>> and pfb & nfb will not be.
>>
>> - There are no plans to backport this to prior Solaris releases, such as
>> Solaris 10, and if those plans are ever made in the future, a new ARC
>> case will be brought forward then to figure out the additional
>> complications
>> of this model in the patch & update release world.
>>
>>
>
