By barton808 on Aug 28, 2007
About two months ago my boss Simon got tasked by Jonathan Schwartz
to work out a process to respond to and, where ever possible, deliver
on requested chipset documentation. The key ingredients of the
process, as set forth by Jonathan, were to be openness and
Who wants old docs anyway?
The value in old chipset documentation is that it allows developers, such as those in the \*BSD communities, to write drivers that enable their OS's to run on older (sometimes very much older) Sun gear. Appropriately, Simon described the start of this effort in a blog entry entitled "Hardware Archeology."
Unlike other makers of chips who create their products for external use, for most of Sun's history our chips and their docs have been produced for use in our own products and not intended for external audiences (this however has recently changed with the launch of Sun Microelectronics). In order then to make these often "ancient" documents public they need to be scrubbed by both engineering and legal to make sure there is no third party trade secrets and that they are "public-ready."
To drive this effort a cross-functional team was put together with engineering, legal and members of the FOSS group. We started by tackling seven ASICs that the OpenBSD community had previously identified as their top priorities.
A wiki is born
Two months later, thanks to the hard work of the Microelectronic engineering teams who took this on as their "night-jobs," the documentation for four of these (Psycho, Fire, Cheerio and Tomatillo) have been made public and the docs for the other three (Gem, Cassini, Schizo) are well on their way. During this process I worked closely with David Gwynne of the OpenBSD community to determine priorities and to determine the next batch of chips that OpenBSD is interested in.
the idea here is not simply to have one-on-one conversations with
select individuals but to open the process up so that anyone can make requests
for documentation and then track the progress of the request. The
vehicle we are using for this is the FOSS Open Hardware Documentation page which is part of the new wikis.sun.com site that launched at the beginning of this month and was created with the idea of enabling exactly this type of open communication.
Since the Wiki was launched on Friday, a column has already been added for implementations and David Gwynne has not only signed up for access to the wiki but has posted OpenBSD implementations for three of the four chips we've recently made docs available for!
So please join in. We cant promise that we'll be able to provide all documentation but we'll try our best and if we cant we'll explain why.