MP Build... FYI

[John Watlington]

There shouldn't be any problems if 624 is used for MP.
The testing process is there for a reason, and should be strictly  
followed.

The DCON bug that 625 fixes (#4479) will only show up on a small
number of machines, and it's frequency is low enough in normal
operation of our current software (where we only support sleep,
not automatic suspend) that it will be an extremely rare complaint
(roughly one tenth of the machines will show the problem once
every 5-20K resumes --- that's a lot of button pushes!).   When the
display is powered down automatically for power conservation,
the code path should reinitialize the DCON correctly.

It is a low severity bug, the user hits any key and the problem  
clears up.
RMA requests for machines showing this problem can be answered with a
request that they upgrade their software.

John

On Nov 4, 2007, at 12:32 AM, Richard A. Smith wrote:

> Jim Gettys wrote:
>
>>>
>>>>         Ship.1 are the bits Quanta is currently putting on  
>>>> machines in
>>>>         the
>>>>         factory. It is Build 624; Firmware Q2D03.
>>>>
>>>>
>>>> We are putting build 625 on mass production machines starting  
>>>> Monday,
>>>> because this addresses a suspected bug, the failure to  
>>>> initialize the
>>>> DCON registers when power is re-applied to DCON.  Build 625 is in
>>>> testing by Quanta right now.
>
> I just spoke with Arnold about the current 624/625 testing status.
>
> 624 testing was completed by Quanta and is ready to go.  They were not
> able to start testing 625 because there was no signed image.
>
> Current plan is for 624 to go onto the machines.
>
> _If_ we claim that 625 are the new bits and sign them for use on MP
> machines then Quanta can start testing 625 on Monday morning.
>
> They _might_ be able to complete the testing in time for MP but Arnold
> says it will involve a lot of work from the .tw testing team.
>
> Gary and I will be headed to CSMC at 8pm Changshu time to look a the
> results of the > 24 hour DCON test.  I'll report back the results and
> then the team can make the call on going with 624 or trying to push  
> for
> 625.
>
> -- 
> Richard Smith  <richard at laptop.org>
> One Laptop Per Child
> _______________________________________________
> Devel mailing list
> Devel at lists.laptop.org
> http://lists.laptop.org/listinfo/devel