Registration
The name you enter will be displayed in your entries, the email address not. After completing the fields you'll receive an email with your initial password.
|
coreboot
Created by Carl-Daniel Hailfinger, modified September 23, 2009 at 16:10:15 CET | This project has a total of 9 feature requests with 8 comments rated at 159 points | 12 unique implementers | Readiness: 9 features with at least three implementers rated at 159 points
coreboot (formerly known as LinuxBIOS) is a Free Software project aimed at replacing the proprietary BIOS/EFI (firmware) you can find in most of today's computers.
It performs just a little bit of hardware initialization (e.g. training DRAM) and then executes a so-called payload, for example a Linux kernel, NetBSD, DragonFlyBSD, FILO, GRUB2, OpenBIOS, Open Firmware, SmartFirmware, GNUFI (UEFI), Etherboot, ADLO (for booting Windows, FreeBSD and OpenBSD), Plan 9, or memtest86.
With coreboot, you can achieve boot times of 0.5 seconds from poweron to kernel and 5 seconds from poweron to working X11. List of Implementers
Bring up the K8 (Athlon64) in coreboot v3
The AMD K8 and successors are well-supported in coreboot version 2. The new coreboot v3 has a much cleaner design, cleaner code and better documentation. Port the current K8 code from v2 to v3 and clean it up.
This is the key requirement for supporting dozens of current desktop mainboards in v3 and will help validate the new design.
End user benefits: A new mainboard can be supported in a matter of hours with a few clicks instead of days/weeks of cutting and pasting code.
By the way, if we manage to implement this feature before Hackontest starts, we will work on porting the AMD 690G/SB600 (or the VIA K8T890 if the 690G/SB600 is already done) chipset support from v2 to v3.
The focus of the task, whichever it may be, is enabling the user to easily create support for a mainboard without having to look at or understand the code.
|
Bring up coreboot v3 on an Asus eeepc
The Asus eeepc is an ultraportable notebook computer with an estimated user base of ~1 million. Porting coreboot v3 would trigger interest in coreboot and provide a platform to stabilize new features on.
End user benefits: An eeepc user can remove the current xandros distro and install an alternative os (combined with gnewsense producing a full gnu system)
|
Create flashrom support for the EeePc
The Asus EeePc is one of the best-selling subnotebooks right now, but it's not possible to reflash the ROM from a free operating system with free tools.
A first step towards firmware/BIOS modification or replacement is the ability to reflash the ROM. This is mostly a reverse engineering effort and will require intimate knowledge of SPI flash chips and SPI and LPC wire protocols. With a logic analyzer, it should be possible to watch the proprietary BIOS reflashing sequence, correlate SPI and LPC traffic and write up a specification for ROM reflashing.
The 24 hours of Hackontest will be used to gather SPI/LPC traffic traces, correlate them and implement the reverse engineered flashing protocol. Depending on the availability of hardware and tools, some of this task may need to be done in advance or completed after Hackontest.
There already is a version of flashrom that supports early prototypes of the OLPC XO which has a similar flash controller. That version will be used as a basis for the new development effort.
|
Easy web interface to create coreboot ROMs
Coreboot currently has a great build system, but you still have to download the sources, configure it with the right settings, compile it and prepare the ROM.
It would be great if coreboot had a web interface where you just select a mainboard model, click on the features you want, wait a few minutes and download a perfect coreboot ROM image for your board.
Offer a few stress-tested coreboot images for paranoid people who want to be sure the ROM image is safe.
|
Add support for AMD Phenom processors
Phenom processors are the latest and greatest stuff AMD has to offer. There is already preliminary support for older Phenom variants in coreboot, but newer Phenom models need slightly different code.
Port existing code to the latest Phenom variants and test it. Make sure the new Phenom code works on all supported socket AM2 boards.
|
Offer a compat BIOS plugin
Most operating systems still need 16bit code to boot.
Offer the option to integrate the LegacyBIOS compatibility layer into a streamlined coreboot build process so people can easily create coreboot images which are able to boot DOS/Windows.
The 16bit BIOS compatibility layer plugin will be strictly optional, the coreboot code itself stays pure 32bit code.
|
Add AMD 700 chipset series support to coreboot
We want support for socket AM2+ mainboards. Most of these mainboards use AMD 700 chipsets, which are unsupported right now.
Add support for the AMD 790FX, 790X, 770, 780G, 740G, 780V, M780G chipsets and port coreboot to the most wanted boards with these chipsets.
|
Offer a compat EFI plugin
Some newer operating systems can boot with EFI instead of BIOS. While EFI is an old legacy standard, there are people who absolutely want it.
Offer the option to integrate a UEFI/GNUFI compatibility layer into a streamlined coreboot build process so people can easily create coreboot images which are able to boot operating systems requiring EFI.
The EFI compatibility layer plugin will be strictly optional, the coreboot code itself stays pure 32bit legacy-free code.
|
Add AMD 690 chipset series support to coreboot
Socket AM2 mainboards with modern AMD chipsets are not supported in coreboot right now. That restricts coreboot to very few interesting current mainboards.
Add support for the AMD 690G, 690V, M690, M690V and M690T chipsets and port coreboot to the most wanted boards with these chipsets.
|
|
Results of Hackontest
The winners of the first Hackontest event on September 24/25, 2008 at OpenExpo 2008 Zurich:
|