u-boot as payload

[Rudolf Marek]

Hello,

I wrote a blog post about this:

http://blogs.coreboot.org/blog/2011/04/30/u-boot-as-coreboot-payload/

I'm attaching patches so you see what I have right now. Please consider them 
just a "test" I did not change the copyrights etc. Maybe they are just a big 
mess. Maybe you find them useful.

The current problems are:

1) I have PCI resource conflicts with QEMU and coreboot

This is just now a showstopper. Definetely no time to investigate right now/

2) I have in plan to get rid of u-boot relocation by relocating it exactly 0 
bytes ;)

3) I still don't know git despite F1 from Carebear ;)

Thanks,
Rudolf

[Rudolf Marek]

Hi again,

Getting rid of u-boot option CONFIG_PCI_PNP fixes the PCI stuff ;)

And I also did an insane hack, which just inject RSDP to right place making the 
u-boot boot system a kernel and system on AMD 64bit platform (Asrock 939A785G)!

Jumping to boot code at fc0013
POST: 0xf8
entry    = 0x00fc0013
lb_start = 0x00100000
lb_size  = 0x00094000
adjust   = 0x2fd5c000
buffer   = 0x2fdc8000
      elf_boot_notes = 0x0011ae90
adjusted_boot_notes = 0x2fe76e90
dram init
type: 16, base 0 size 1000
type: 1, base 1000 size 9f000
type: 1, base c0000 size 2fe30000
type: 16, base 2fef0000 size 110000
RSDP found at 2fef2400type: 2, base 30000000 size 10000000

U-Boot 2011.03-00214-g24b6bc2-dirty (May 03 2011 - 23:58:39)

DRAM Configuration:
Bank #0: 00001000 636 KiB
Bank #1: 000c0000 766.2 MiB
Bank #2: 00000000 0 Bytes
Bank #3: 00000000 0 Bytes
## Unknown flash on Bank 1 - Size = 0x00000000 = 0 MB
## Unknown flash on Bank 2 - Size = 0x00000000 = 0 MB
## Unknown flash on Bank 3 - Size = 0x00000000 = 0 MB
Flash: 0 Bytes
*** Warning - bad CRC, using default environment

In:    serial
Out:   serial
Err:   serial
IDE:   Bus 0: OK
   Device 0: Model: WDC WD600BB-00CAA1 Firm: 17.07W17 Ser#: WD-WMA8F113824
             Type: Hard Disk
             Capacity: 57241.8 MB = 55.9 GB (117231408 x 512)
Net:   No ethernet found.
Serck Controls eNET
boot > ext2ls ide 0:5 /boot
<DIR>       4096 .
<DIR>       4096 ..
          2063256 bzImage
<DIR>       4096 filo
<DIR>      12288 grub
            22608 memdisk
         71290880 a.tar
<DIR>       4096 a
          1225518 System.map-2.6.26-1-amd64
            85588 config-2.6.26-1-amd64
          1755088 vmlinuz-2.6.26-1-amd64
          2172192 bzImage28
          2236288 bzImage29
          7707622 initrd.img-2.6.26-1-amd64
         33554432 disk
          2414688 vmlinuz-2.6.32-5-amd64
          6802950 initrd.img-2.6.26-1-amd64.bak
           104510 config-2.6.32-5-amd64
           165084 memtest86+.bin
          1660631 System.map-2.6.32-5-amd64
          9446666 initrd.img-2.6.32-5-amd64
           124648 memtest86.bin
           167264 memtest86+_multiboot.bin
boot > setenv bootargs console=ttyS0,9600 root=/dev/sda
boot > ext2load
ext2load - load binary file from a Ext2 filesystem

Usage:
ext2load <interface> <dev[:part]> [addr] [filename] [bytes]
     - load binary file 'filename' from 'dev' on 'interface'
       to address 'addr' from ext2 filesystem
boot > ext2load 0:5 2000000 /boot/bzImage29
** No boot file defined **
boot > ext2load ide 0:5 2000000 /boot/bzImage29
Loading file "/boot/bzImage29" from ide device 0:5 (hda5)
2236288 bytes read
boot > zboot 2000000
Valid Boot Flag
Magic signature found
Setup Size = 0x00002c00
Moving Real-Mode Code to 0x00090000 (11264 bytes)
Using boot protocol version 2.09
Linux kernel version 2.6.29.2 (root@ruiktest) #1 SMP PREEMPT Sat May 2 00:24:59 
CEST 2009
Kernel command line: "console=ttyS0,9600 root=/dev/sda"
Loading bzImage at address 0x00100000 (2225024 bytes)
## Transferring control to Linux (at address 00090000) ...

Starting kernel ...

copy to 7c0 from 101e3d4 size 166
copied 9c666066
Linux version 2.6.29.2 (root@ruiktest) (gcc version 4.3.2 (Debian 4.3.2-1) ) #1 
SMP PREEMPT Sat May 2 00:24:59 CEST 2009
Command line: console=ttyS0,9600 root=/dev/sda
KERNEL supported cpus:
   Intel GenuineIntel
   AMD AuthenticAMD
   Centaur CentaurHauls
BIOS-provided physical RAM map:
  BIOS-e801: 0000000000000000 - 000000000009f000 (usable)
  BIOS-e801: 0000000000100000 - 0000000004000000 (usable)
DMI present.
AMI BIOS detected: BIOS may corrupt low RAM, working around it.
last_pfn = 0x4000 max_arch_pfn = 0x100000000
Scanning 0 areas for low memory corruption
modified physical RAM map:
  modified: 0000000000000000 - 0000000000010000 (reserved)
  modified: 0000000000010000 - 000000000009f000 (usable)
  modified: 0000000000100000 - 0000000004000000 (usable)
init_memory_mapping: 0000000000000000-0000000004000000
last_map_addr: 4000000 end: 4000000
ACPI: RSDP 000F0540, 0014 (r0 CORE  )
ACPI: RSDT 2FEF2424, 0034 (r1 CORE   COREBOOT        0 CORE        0)
ACPI: HPET 2FEF24C8, 0038 (r1 CORE   COREBOOT        0 CORE        0)
ACPI: APIC 2FEF2500, 0054 (r1 CORE   COREBOOT        0 CORE        0)
ACPI: SSDT 2FEF2554, 020C (r2 CORE   DYNADATA       2A CORE       2A)
ACPI: FACP 2FEF3B49, 00F4 (r3 CORE   COREBOOT        0 CORE        0)
...

Rest is attached.

Thanks,
Rudolf

[Rudolf Marek]

Hi

Due to popular demand here is the yet another patch on the top of those two 
previously posted. It bit talks to coreboot tables.

Known issues:

* The memory is still 64MB for linux - this can be fixed by reporting correct 
amount to uboot (bios_*.c) but because there is no e820 just need to report 
continuous memory to the first hole.

* loading stuff just in 1MB wont work because the LMA is set for bios and 
realmode trampoline there, please now load kernel from 2MB or more.

* I had to swap IDE drives because this board has only secondary IDE

Thanks,
Rudolf