[Arm-netbook] SO-DIMM Upgrade Proposal: CT-PC89E with an 833mhz S5PC100 and 1gb of RAM - anyone interested?

[Luke Kenneth Casson Leighton]

On Tue, Apr 6, 2010 at 2:49 PM, Frans Pop <elendil at planet.nl> wrote:
> On Tuesday 06 April 2010, John Winters wrote:
>> Luke Kenneth Casson Leighton wrote:
>> > On Tue, Apr 6, 2010 at 12:48 PM, JLB <j at twu.net> wrote:
>> >> I'll pay $150 to get this upgrade.
>> >
>> >  fantastic!  that makes $3880.
>>
>> Put me down for $150 too.
>
> I think I'm going to bail out at this point folks. It seems to me that this
> group is abandoning the current machines we have and moving on to a fairly
> ambitious new project.

 yes and no.  it depends to some extent on what the factories dafurong
and chitech, and the gpl-violators seatron aka naijing aka mid-fun do.

> Sure, we have a somewhat working kernel, but we also have a ton of fairly
> serious unresolved issues. If nobody is going to put any effort into that,
> then I wonder seriously about the viability of the new project.

 the issues that we have are down to the fun-and-games of
reverse-engineering.  if we had the linux kernel source code, there
wouldn't _be_ any issues.  thus, by creating a new so-dimm where we
know exactly how it works, many of the issues can be solved or
avoided:

 * by working with a willing hardware engineer, we can get the
schematics for the MMC cards, and solve the interrupt issue by
entirely avoiding it.

 * likewise, the fact that the watchdog timer hardware line is
entirely missing from the CT-PC89E motherboard is _not_ a problem,
because we can put in a couple of transistors and a "compatibility"
jumper, which connects the watchdog RESET hardware line to GPG12.  so,
either GPG12 goes high or watchdog RESET goes low, machine powers off.

 about the only significant issue which does need software-only
solving is the LCD screen.


> I will continue to monitor what happens here, and I'm also willing to
> continue to release new kernels and D-I images if patches become available
> from others, but I will no longer make any effort to resolve current known
> issues myself.

 frans, you've done far more than i could have expected, already - and
managed to break new ground with d-i for ARM and d-i GUI for ARM as a
result: that's just absolutely fantastic.

 reverse-engineering isn't for everyone: it does take absolutely
masses of time.  the main thing is that you've managed to get one hell
of a lot of the necessary work done, and i'm sure that there will be
other people who will want to help out, soon enough.

> However, that may change if we do get the ChiTech kernel
> source available.

 yeh.  i'm pretty pissed about that.  important lesson there, on several levels.

 l.