July 2005 - coreboot - mail.coreboot.org

[LinuxBIOS] The Eric B. Mega-patch
by Jason Schildt July 20, 2005

July 20, 2005

Stefan, Ollie, and Ron, Has Eric's Mega-patch been incorporated into the public tree? I've just created the LNXI mirror of the freebios2 tree and want to make sure that I'm starting with the tip. Thanks, --jason-- -- Jason W. Schildt LinuxBIOS Software Engineer Linux Networx

3 4

Re: [PROPOSAL] enhanced table dumping
by YhLu July 19, 2005

July 19, 2005

About APIC and ACPI. Some chipset has problems with apic and the workaround must to be used with acpi BIOS. YH

3 2

RE: [LinuxBIOS] etherboot with LinuxBIOS
by Stephen.Kimball＠bench.com July 18, 2005

July 18, 2005

The easiest way to build Etherboot 5.2.6 is to use gcc 3.2.2. I found another machine and built an old gcc and etherboot compiled. I got errors with the new gcc, tried to fix them, but could never get it to compile. It would be good, if someone knowledgeable could make etherboot work on the current gcc. Steve -----Original Message----- From: Steve Magnani [mailto:steve@digidescorp.com] Sent: Thursday, July 14, 2005 6:39 AM To: Kimball, Stephen; linuxbios(a)openbios.org Subject: RE: [LinuxBIOS] etherboot with LinuxBIOS On Wed May 25 14:13:05 2005, Stephen.Kimball at bench.com wrote: > Does anyone know how to build etherboot with a current version of gcc? > I am using gcc version 3.4.3 and binutils 2.15. > I've tried several etherboot versions up to 5.4.0. > I know etherboot 5.2.6 with gcc 3.2.2 works, but that's an old version of > gcc. Thanks. Steve - Did you ever find a resolution to this? I didn't see any followups on the mailing list. I have the same toolchain as you, so if there's a way to avoid building another one just for Etherboot I'd like to know! Thanks, Steve Magnani

4 9

[LinuxBIOS] Moving forward: Subversion
by Stefan Reinauer July 15, 2005

July 15, 2005

Dear LinuxBIOS community, most of you might not have noticed, but we've had quite adventurous times during the last five months. The whole story started around February, when LinuxBIOS decided to go away from Sourceforge's CVS repository. CVS does not support restructuring code very well, and Sourceforge sometimes had delays of a week before changes would be visible to the world. That's why we moved on to a more modern version control system: GNU arch. While conceptionally very advanced, GNU arch also brought some issues and required hefty efforts. While concepts and architecture are important, usability and maturity is still a major concern with GNU arch. Since the Open Source/Free Software Firmware community wins nothing when we all advance to version control experts but stick with administrative issues, we decided to go to a mature system that combines modern version control techniques with a steep learning curve, making it a lot easier for everyone to have a look at LinuxBIOS development, or contribute: Subversion I've been migrating the LinuxBIOS repository to Subversion now. We have been testing it for the the last month now and it has been a good deal more reliable and usable than GNU arch. By tonight the GNU arch repository is read only and will not accept new checkins. Here's a quick introduction how the LinuxBIOS repository can be accessed from now on: Anonymous access ---------------- You can check it out as follows: $ svn co svn://openbios.org/repos/trunk/LinuxBIOSv2 or $ svn co svn://openbios.org/repos/trunk/LinuxBIOSv1 Developer Access ---------------- Access for developers is very similar to anonymous access. Just add your Subversion username as follows when checking out the repository: $ svn co svn://<username>@openbios.org/repos/trunk/LinuxBIOSv2 All people with checkin permission to the GNU arch repository will be contacted shortly to move their accounts over to Subversion (Many of them are already). If you are a LinuxBIOS developer and don't hear from me, please contact me. Source code browsing -------------------- You can also browse the LinuxBIOS Subversion repository online: http://www.openbios.org/viewcvs/trunk/ Snapshots --------- If you don't want to use Subversion, or can't use it, you can also download a tar ball of the latest repository at http://www.openbios.org/viewcvs/trunk/LinuxBIOSv2.tar.gz?view=tar More information on Subversion ------------------------------ Subversion is available as a package in all major Linux distributions. It can also be compiled from source code on most available systems. See the Subversion homepage http://subversion.tigris.org/ for more information on Subversion, including a description of available commands and the source code. There's also a very detailed online book available on Subversion: Version Control with Subversion - http://svnbook.red-bean.com/ The Subversion user interface is very similar to the CVS one, so if you have used CVS before, it will be easy for you to use it and participate in the LinuxBIOS project. Comments, flames, questions are very welcome. Best regards, Stefan Reinauer -- OpenBIOS - free your system.

1 0

[LinuxBIOS] APIC_CLUSTER and PCI 19.x?
by Li-Ta Lo July 15, 2005

July 15, 2005

YhLu, Could you tell me why by moving apci_cluster in front of other devices, we only need one chip cpu/x/y/z? And why we don't have to list 19.x neither? -- Li-Ta Lo <ollie(a)lanl.gov> Los Alamos National Lab

2 7

[LinuxBIOS] pci_rom_load() bug for non-VGA ROM
by Tao Liu July 15, 2005

July 15, 2005

Hello, pci_rom_load() returns wrong rom_header for non-VGA ROM case ? --- orig/src/devices/pci_rom.c +++ mod/src/devices/pci_rom.c @@ -95,11 +95,11 @@ return (struct rom_header *) (PCI_VGA_RAM_IMAGE_START); #endif } else { - printk_spew("%s, copying non-VGA ROM Image from %x to %x, %x bytes\n", - __func__, rom_header, pci_ram_image_start, rom_size); + printk_debug("copying non-VGA ROM Image from %x to %x, %x bytes\n", + rom_header, pci_ram_image_start, rom_size); memcpy(pci_ram_image_start, rom_header, rom_size); pci_ram_image_start += rom_size; - return (struct rom_header *) pci_ram_image_start; + return (struct rom_header *) (pci_ram_image_start - rom_size); } /* disable expansion ROM address decoding */ pci_write_config32(dev, PCI_ROM_ADDRESS, rom_address & ~PCI_ROM_ADDRESS_ENABLE); -- Regards, Liu Tao

3 3

[LinuxBIOS] Unable to determine if motherboard is supported.
by Art Fore July 15, 2005

July 15, 2005

Have an Asus K8V SE Delux motherboard with an Athelon 64 socket 754 processor with the following information from the manual Chipset VIA K8T800 VIA VT8237 1 Gig RAM Southbridge & Promise PDC20378 Raid Controller Serial ATA drive, IDE drive, IDE DVD ROM drive. Boots from ATA serial drive. Audio ADI AD1980 Marvell 88E8001 Gigabit LAN ontroller It is a tripple booth system with Suse Linux, and WinXP on the SATA serial drive, and DOS on the IDE drive. (The DOS drive is not a necessity for the final configuration of the system that will be used) Can anyone tell me if it is possible for the OpenBios or Linux Bios be used to replace ASUS AI Bios? TIA Art

2 2

[LinuxBIOS] Bios & PCI
by Art Fore July 15, 2005

July 15, 2005

Apparently, the PCI bus is setup during booth with the bios. We have a problem with this. We have a system, dual booth Suse Linux and WinXP. It is connected via a "Stargen" chip which is actually a serial PCI bridge. This controls an external instrument where each card is a PCI instrument connected via andother "Stargen" chip. If we shut down the instrument system, and change a card, and turn it back on, we then have to rebooth the computer to get it to re-instanciate the PCI bus and recognize the new card. Is it possible with OpenBios or LinuxBios to be able to perform this function from the operating system without having to do a reboot? Please note I am not a bios guru. Art

3 3

[LinuxBIOS] NUMA support for dual core Opteron
by yhlu July 14, 2005

July 14, 2005

Someone mentioned that NUMA support for dual core opteron need acpi support in LinuxBIOS. there may be some other solution for that. 1. PowerPC already support dual core and it should support NUMA, So the Open Firmware must have some NUMA entry definition. Can we make x86-64 kernel support OpenFirmware interface so we can use OpenBIOS as payload of LinuxBIOS. 2. enable acpi and add the NUMA entries into it, the Linux Kernel will be happy. 3. If we are trying to use ADLO to load Windows/Solaris/FreeBSD, We need to pass related acpi info to ADLO.... Solution 1 will be ideal one, and can make Solaris for X86-64 use OpenFirmware interface too..... which one is better? YH

4 9

RE: [discuss] Re: [LinuxBIOS] NUMA support for dual core Opteron
by Langsdorf, Mark July 14, 2005

July 14, 2005

> On Thu, 2005-07-14 at 20:48 +0200, Andi Kleen wrote: > > > AFIAK, for x86_64 kernel, it will try to read NUMA configuration > > > from HW directory. We don't have to export any ACPI table. > > > > It doesn't work for dual core or 8 sockets for some reason. > Since the > > SRAT code works fine and is in general more future proof we never > > tracked down why. Patches welcome. > > > > Does each core has its own memory controller? No. Each processor has its own memory controller. A processor may have 1 or more cores. (A processor is a physical package. A core is a complete logic execution unit). All cores in a processor have equivalent access to the processor's memory controller. -Mark Langsdorf AMD, Inc.

1 0