On Monday 04 September 2006 01:11, Bari Ari wrote:
Look for laptops that have the firmware Flash write enable lines controlled only by the chipset and not also by the keyboard/system management controller. This will allow a developer to rewrite the Flash with LinuxBIOS. Most efforts to port LinuxBIOS on laptops in the past could not get past this hurdle.
Designing and manufacturing for a laptop for mass market using LinuxBIOS is simple. The bill of material and hardware design are the same as a laptop using a closed source BIOS. Laptop mainboards are typically manufactured in high volumes in single runs (much the same as desktop/platform mainboards). The LinuxBIOS would be developed for the mainboard during the design stage of the project and then programmed into the flash before manufacture or after the flash devices are installed on the boards.
There hasn't been much demand for laptops to have LinuxBIOS yet. The first major project to demand LinuxBIOS on a laptop has probably been the OLPC project. I doubt if Quanta had much experience with or knowledge of LinuxBIOS before OLPC.
I was thinking more along the way, how hardware could be simplified on one side and at the same time enhanced by nice-to-have features.
Just some ideas: +Large enough (partitioned) Flash(s) already on the board +as low as possible number of parts +get rid of stone-age hardware and replace with modern parts +new concepts to replace error prone complex initialisation sequences +combine just the best of the best (with options to scale on purpose) +open, fully documented standards +KISS =a more advanced PC/Laptop/Node...whatever (IMHO)
--- As my original post got accidently PM, here what i wrote before that mail:
The chipset docs may all be available but not docs to the power management/keyboard scan controller firmware. The docs for the micros describe the general purpose micro itself, but that won't tell you how they are using all of its gpio pins and ports.
Isn't that a pity? Those parts are not the complex parts, it are the most simple parts and some of them are just to control hardware that is available for eons as a standard.
In the age of USB i wonder if we still need a keyboard/mouse controller for other than that. (i mean: why not just USB without PS2? Why COM,LPT? )
How would a (linux) mainboard look like? Is there a parts list, e.g. one for desktop, one for laptop that would be great?
Maybe, just maybe, one of the manufacturers takes a deep breath...
Or just maybe they notice, that the selection of parts is so carefully optimized, that it would be great to do it that way for a decent windows PC as well.
- (As for the price of energy i would say, both should only use mobile-hardware)
I'am reading at that least for a while and i got the feeling, that most hardware is not support, only a few work fully - i liked to have it for my Centrino-Mini-itx but the memory controller is a problem, howe does it look now for turion based mini-itx-systems? (I use those as a desktop pc because they use so little energy to run and are fast anyway - beside such a tiny little box is not so intrusive in the living room ;-)
Part of the problem is certainly the very difficult installation and the risk of failure - i once fried a mainboard while updating the standard flash with an update from the manufacturer - it just stopped half way through. I got a replacement part, let it flash somewhere else but it didn't help. Though i work in embedded control area, flashing ECU's without trouble hundred times a day. Those ECU's are designed in a way, that they can even be reflashed when the flash-process interrupted __anywhere__ - despite the propability that this feature is actually required out there in the field is very low, maybe once in the lifetime of a car. The (missing) content of the flash can be delivered via certain serial line protocols during the flashing process from an external hardware/software (K-Line, CAN, ...) - wouldn't that be a nice thing to have for Linux as well?