Thread: Linux 2.6.0
View Single Post
Old 09-02-03, 03:13 AM   #21
pe1chl
Registered User
 
Join Date: Aug 2003
Posts: 1,026
Default

Quote:
'Cause the ACPI in Windows and the ACPI spec are two different things. Linux ACPI got done by a bunch of people who wanted to 'do it properly' and follow the spec, not be bug for bug compatable with the Redmond implementation.
Only guess what the hardware builders target for....
But when you would ask for specs on the nVidia hardware, you would probably run into the same situation!
Of course there must be internal design docs, which may not be suitable for external publication without editing.
But then, there is probably a good informal contact between the hardware and software teams, so you would be at a definite disadvantage when writing a driver without access to inside information.

Remember a 3D video card is not something like a serial port or a USB controller. You will need a serious amount of code that drives a complex piece of hardware.
The nVidia people have a big advantage because they re-use their Windows driver code that has been debugged and optimized.

I think that a full release of hardware info and the plan to write a driver from scratch will put us back from where we are now.
There are always fewer coders willing to maintain a device driver for "today's great video card" (which will be a drag next year) than those working on generic kernel stuff and applications.
pe1chl is offline   Reply With Quote