[Hardeneddrivers-discuss] RE: [cgl_discussion] Some Initial Comments on DDH-Spec-0.5h.pdf

[Randy.Dunlap]

On 24 Sep 2002, Andy Pfiffer wrote:

| > What you really want is a test-harness, both for runtime and compile-time
| > to validate that drivers are doing the right things at the right time.
|
| Once, a long time ago, I worked with a team that built a 100%
| pseudo-kernel compilation environment for driver development work
| (non-Linux OS).  The idea was to develop a driver while the hardware and
| kernel were still being developed.  All of the kernel interfaces that
| were called by the driver were dummied up and made somewhat functional,
| header files copied over, etc.

Hm, I've had that experience also.  Ethernet drivers and test harness
being developed while the silicon was also being developed, routed,
taped out, etc.

Get new silcon, drop it onto a waiting (PCI) adapter, test the drivers,
run the test harness, etc.  Good developement environment.

| In the end, the driver wound up being a 100% drop-in-and-it-just-worked.

OK, yours was a little better than mine was.  But still a good and
mostly successful development environment that allowed lots of
hardware and software co-design/development.

| Perhaps there is some way to leverage User-Mode Linux for this (he says,
| waving his hands wildly...).

-- 
~Randy