As of June 8, 2005, you can download initialization and boot source code based on the IEEE-1275 Standard for Boot Initialization Configuration Firmware.
Download open source Slimline Open Firmware (SLOF)
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
So this applies ONLY to the powerPC?
hopefully the Linuxbios people can benefit.
Thing is, who still owns the patents / firmware to the PC x86 bios?
I guess you could port SLOF to x86. I don’t know why, though.
Because PC Bios suck? Heck PC motherboards are garbage as well. They are still massivly limited in what they can do, They can’t do the Target firewire disk mode apple’s can, They are limited to 15 IRQ places after 20 years.
Standard PC Bios should of been phased out shortly after Windows was required to adjust to dealing with three devices per IRQ for plug and play.
Take a look at your computer’s right now. The P11 400mhz laptop I am typing this from has 3 different , 8 total devices using IRQ 10.
Open Firmware has no such limits. Hopefully te next gen Bios that intel is working on will take care of it. Of course I was also hoping that longhorn would be a complete rewrite of XP codebase. But since they are even removing the managed code features that is now an impossiblity.
does anyone actually use a LinuxBIOS outside of the labs where it is tested and developed? I’m curious as to how friendly it is towards newer hardware.
They are limited to 15 IRQ places after 20 years.
No, they aren’t.
Even if you have an entire PCI-X bus, many times you cannot share the bus properly between two cards due to various conflicts.
The BIOS is useless for the most part to resolve conflicts. Most BIOS implementations do not even have current documentation. The whole situation is pathetic.
Wintel systems are designed by morons. They are cheap and barely work. And end up costing trillions in lost productivity.
The problem here is legacy compat. You can turn on your shiny AMD64 box and boot that piece of shit MS-DOS system from 1989. And it will work. Because closed source software needs to work on newer hardware. The BIOS and all the kludges that have been added over the years are a total joke. Even the Amiga 1000 had a more advanced autoconfig system in 1985. That’s the price you pay for Microseft’s market domination. I get sick when I imagine the damage they’ve done to the PC platform with their shitty software over the years.
The use of the word Wintel tells me that you’re probably a PPC/Mac zealot. I have no idea about PCI express so I can’t comment on that. I will comment on the rest because I had both Macs and PCs in the past. Currently I have a G4 and several PCs.
Bios is useless to resolve conflicts
This is of course not true, you can resolve most of the conflicts if you apply some common sense, and you can always find BIOS documentation in the motherboard manual if you need it.
Wintel systems aren’t designed by morons I can assure you
They are relatively cheap yes, and that’s a problem? I don’t know what you mean by “barely work” they work just fine over here, and they did so for the past 10 years. I’m also aware that your post was mainly indented to incite a pointless flamewar and I did see the “Don’t feed the trolls!” sign out at the entrance so I won’t tell you about the (bad) experience I had with OpenFirmWare’s memory management 
Either way.. Apple is going Intel get over it
The original 15 irq’s available, (0 is for the cpu itself is it not?) are all hardwired lines of interrupt. The rest of the irq’s available, and in PC’s there are at least 15 more, are all soft irq’s. Now, in all the mindless ranting going on, I doubt most of you that have been spouting off here even know what an IRQ actually is, or what it’s purpose serves. So a system designed from the ground up, long after the original x86 spec came out, is most likely going to have more than 15 irq’s available because they learned from watching the needs of the x86 world grow and mature. The limit of 15 irq’s is a physical limit that was overcome later on by rewriting the PC bios from the ground up about 8 years ago if I remember correctly. (Part of the atx std?)
A little more homework, a little less diarreah of the brain, and you sound a little more intelligent, and not like an annoying mac fan boy.
BTW, I did not include a full description of the purpose/function of an IRQ on purpose. If you confused/curious/or realize you don’t know, then a quick google will teach you all you need to know. Self-taught knowledge is something to pride yourself in.
The IRQ is an interrupt request, the original ones were hardware only and were external. Each has it’s own priority.
Their purpose is to interrupt the main CPU routine to execute code with higher priority. Without it you would have an extremely slow PC that performed very poorly (you would have a CPU that continuously wastes time by polling each device: do you have data for me to process?, even if there are no devices present.
IRQs make interacting with various cards not only possible, but fast. I’m an engineer, so don’t listen to me.
The Motorola 68000 architecture is very simple, and publicly available online, check it out for a reference.
if i’m correct, the first pc’s had a limit of 7 irq’s. (8 bits, first one (00000000) reserved). they hooked irq 2 to a new set of irq’s, so we now have a limit of 7+8=15 hardwired irq’s. they can’t add more because of compatibility problems with older hardware and software.
i don’t understand why not use firewire for internal use (sound, harddisks, …)
BIOS does some things right and many things wrong, like disk geometry! AAARGGG!!! which hard disk size limits did we experience? luckely, linux ignores the bios and communicates with the IDE harddisks directly, but I still have a windows 2000 system that can only use the first 8G of a 60G harddisk inside it… i should update th BIOS, yeah, but the user only uses 4G so who cares…
You have something like a Pentium 233 controller then like a 430TX, because starting with the Pentium 2, the BIOS supported 30 or 40GB out of the box, then 80GB on early Pentium 3 boards.
Do yourself a huge favor and buy a new IDE controller card and install it, the drive will run much faster, though your boot time will increase—nearly all retail harddrives above 120GB include one free.
You may also want to look at OpenBIOS:
http://www.openbios.org/
I just wish it was BSD-licensed instead of GPL-ditto.