Developer Manual

This is work in progress!

This manual is intended for aspiring coreboot developers to help them get up to speed with the code base and the tasks required to add support for new chipsets, devices, and mainboards. It covers coreboot v4.

How to support a new board

People often ask how to support a new board. If you are willing to put in the effort to write the port, then there is a good chance that you will succeed. Supporting a new board that uses a chipset that is already supported by coreboot is much less work for obvious reasons than supporting a new board with an unsupported chipset. Supporting a new board can take from an hour to over a year of your time.

Supporting a new board with the same cpu family, chipset and superIO

To support a new board with an already supported chipset look for the most similar board in the coreboot tree to the new board that you wish to support. After you find the most similar board, look for the differences between your new board and the most similar board.

If your new board has the same cpu family, cpu socket, chipset and superIO then you can try the coreboot build for the supported board on the new board with a backup flash device and debugging turned on. Look at the debug output to determine where the boot process stops or what errors are encountered on the way. Common differences between boards with exactly the same cpu, chipset and superIO are IRQ routing, ACPI and flash write enable routines or jumpers. Make changes to the board configuration, ACPI or IRQ routing etc etc until you find the proper settings. This can take from an hour of time to a few months based upon your coding skills and hardware issues.

Supporting a new board with the same cpu family, chipset but different superIO

If your new board has the same cpu family, cpu socket, chipset but the superIO is different but it is a supported superIO then you will have to change the board config to use the different superIO. More on this later....

Common differences between boards with exactly the same cpu, chipset but a different superIO are IRQ routing, ACPI and flash write enable routines or jumpers. Make changes to the board configuration, ACPI or IRQ routing etc etc until you find the proper settings. This can take from an hour of time to a few months based upon your coding skills and hardware issues.

Supporting a new board with a unsupported cpu, chipset or superIO

If your new board uses a cpu, chipset or superIO not supported by coreboot then you will have a lot of work in front of you. You will need developer data sheets for the cpu, chipset and superIO. AMD has been releasing data sheets to the public that includes most of the information required to support a new cpu and chipset. AMD has also been releasing complete coreboot patches to many of their recent cpu's and chipsets. Many of the superIO vendors have public documents available. Intel has been closed about releasing specifications at a low enough level to support a new cpu or chipset. Specifications are generally only provided to high volume OEM's. New developers requesting data sheets might have to wait for several months after singing NDA's intil they receive the specifications.

Recommended hardware and software tools

See Developer Manual/Tools for a list of recommended tools which are useful for coreboot users and developers.

Hardware Overview

Intel Architecture

Hardware Reset

From Intel's "64 and IA-32 Architectures Software Developer’s Manual" (doc 253668-021 October 2006), Volume 3A, Section 9.1.4:

The first instruction that is fetched and executed following a hardware reset is located at physical address 0xFFFFFFF0. This address is 16 bytes below the processor’s uppermost physical address. The EPROM containing the software-initialization code must be located at this address.
The address 0xFFFFFFF0 is beyond the 1-MByte addressable range of the processor while in real-address mode. The processor is initialized to this starting address as follows. The CS register has two parts: the visible segment selector part and the hidden base address part. In real-address mode, the base address is normally formed by shifting the 16-bit segment selector value 4 bits to the left to produce a 20-bit base address. However, during a hardware reset, the segment selector in the CS register is loaded with 0xF000 and the base address is loaded with 0xFFFF0000. The starting address is thus formed by adding the base address to the value in the EIP register (that is, 0xFFFF0000 + 0xFFF0 = 0xFFFFFFF0).
The first time the CS register is loaded with a new value after a hardware reset, the processor will follow the normal rule for address translation in real-address mode (that is, [CS base address = CS segment selector * 16]). To insure that the base address in the CS register remains unchanged until the EPROM based software-initialization code is completed, the code must not contain a far jump or far call or allow an interrupt to occur (which would cause the CS selector value to be changed).

FWH/LPC Flash Memory

Modern mainboards are often equipped with Firmware Hub (FWH) or Low Pin Count (LPC) flash memory used to store the system bootloader ("BIOS"). Execution begins by fetching instructions 16 bytes below the flash memory's uppermost physical address.

coreboot Overview

View From The CPU: Intel Architecture

1. At 0xFFFFFFF0, start execution at reset_vector from src/cpu/x86/16bit/reset16.inc, which simply jumps to _start.
2. _start from src/cpu/x86/16bit/entry16.inc, invalidates the TLBs, sets up a GDT for selector 0x08 (code) and 0x10 (data), switches to protected mode, and jumps to __protected_start (setting the CS to the new selector 0x08). The selectors provide full flat access to the entire physical memory map.
3. __protected_start from src/cpu/x86/32bit/entry32.inc, sets all other segment registers to the 0x10 selector.
4. Execution continues with various mainboardinit fragments:
   1. __fpu_start from cpu/x86/fpu_enable.inc.
   2. (unlabeled) from cpu/x86/enable_sse.inc
   3. Some CPUs enable their on-chip cache to be used temporarily as a scratch RAM (stack), e.g. cpu/amd/model_lx/cache_as_ram.inc.
5. The final mainboardinit fragment is mainboard-specific, in C, called romstage.c. For non-cache-as-RAM targets, it is compiled with romcc. It includes and uses other C-code fragments for:
   1. Initializing MSRs, MTRRs, APIC.
   2. Setting up the southbridge minimally ("early setup").
   3. Setting up Super I/O serial.
   4. Initializing the console.
   5. Initializing RAM controller and RAM itself.
6. Execution continues at __main from src/arch/x86/init/crt0_romcc_epilogue.inc, where the non-romcc C coreboot code is copied (possibly decompressed) to RAM, then the RAM entry point is jumped to.
7. The RAM entry point is _start in src/arch/x86/lib/c_start.S, where new descriptor tables are set up, the stack and BSS are cleared, the IDT is initialized, and hardwaremain() is called (operation is now full 32-bit protected mode C program with stack).
8. hardwaremain() is from src/boot/hardwaremain.c, the console is initialized, devices are enumerated and initialized, configured and enabled.
9. The payload is called, either via elfboot() from boot/elfboot.c, or filo() from boot/filo.c.

Failover/Fallback/Normal images overview

See Developer Manual/Bootblock

Memory map

On x86 systems, many memory ranges are reserved for special purposes or have some other peculiar properties. The article Developer Manual/Memory map has details about this fact.

Serial output and the Super I/O

See Developer Manual/Super IO.

RAM init

See Developer Manual/RAM init.

Mainboard

devicetree.cb

The mainboard's devicetree.cb file contains many build and platform configuration settings. One of the most important items is the mainboard device list.

A device needs to be listed in the mainboard devicetree.cb if it requires more setup than standard PCI initialization (resource allocation). Typically, that includes the CPU, northbridge, southbridge, and Super I/O. These devices are usually required for system specific configuration as well as indicate the system bus structure (pci_domain).

When a device in devicetree.cb is found during the coreboot PCI/system scan process the functions to do customized initialization are called via the device_operations and the chip_operations structures. You will find these structures in the devices source files.

irq_table.c

See Creating Valid IRQ Tables.

Creating a new target

To create a new mainboard target you have to add several files.

• Multiple files in src/mainboard/vendorname/mainboardname (replace vendorname and mainboardname, of course).

Documentation and datasheets

Useful hardware/concepts documentation for developers

These external documents and slides explain fundamental concepts of hardware that coreboot supports.

Interrupts

• PCI Interrupts on x86 machines from John Baldwin
• PCI IRQ Routing on a Multiprocessor ACPI System at Microsoft's Windows Hardware Developer Central

System Managment Mode

• System Managment Mode Overview by Robery R. Collins

Specific datasheets

See Datasheets.

This work is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or any later version. This work is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.