coreboot

coreboot@coreboot.org

12 participants
39716 discussions

[LinuxBIOS] RE: Tyan 2850 video output
by YhLu 12 Apr '05

12 Apr '05

did you get the atix.rom using dd? YH > -----Original Message----- > From: Alex [mailto:alex301@bezeqint.net] > Sent: Tuesday, April 12, 2005 2:59 PM > To: YhLu > Cc: linuxbios(a)clustermatic.org > Subject: Re: Tyan 2850 video output > > Hi YhLu, > > The binary image you sent me works fine. > Would you please help me to build LinuxBIOS properly? > > I got the latest development tree (thru tla tool). > I tried to compile it - it failed with a segment overlap error. > So I changed the LOG_LEVEL from 8 to 7. > Also I set the ROM_SIZE to 524288 instead of 475136. > The last thing I changed - was using the r8169.zelf > (etherboot 5.2.6) instead of tg3 one. > In etherboot I've set direct_vga and keyboard (I also > experimented with some other options afterwards) > > I built the linuxbios.rom and burned it onto the chip. It > works fine, prints the output to the RS232 console and > DOESN'T print anything on the screen. > > (CONFIG_OPTION_ROM_RUN and CONFIG_VGA_CONSOLE are both set to 1) > > Appreciate your help, > Alex > > PS Should I use FILO support in etherboot? I tried - > compilation failed... >

1 0

[LinuxBIOS] Where should stuff go?
by Hamish Guthrie 12 Apr '05

12 Apr '05

Hi All, and particularly Eric or Ron, I am busy doing a port of the Geode code to v2, and similar to most of the x86 SoC's the GX1 has a lot of locally memory mapped registers to configure stuff before the chip is useful at all. These registers have to be configured before RAM initialisation etc., and I was just wondering if I should do a dirty hack to include the assembler for this, or if it would be better to recode in C and create a section of code which resides in the cpu tree as opposed to the mainboard tree which is executed prior to the auto.c code. I could of course also chuck this code into the auto.c file, but with auto.c being in the mainboard section, it is not totally appropriate, because for most of these SoC's this particular code is the same for all mainboards - it is really silicon dependant, not mainboard dependant. If I were to create a new cpu init code section, how exactly would I go about incorporating this into the build structure so it will execute before auto.c? TIA Hamish -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.308 / Virus Database: 266.9.5 - Release Date: 07.04.2005

2 1

[LinuxBIOS] Speed up flash programming!
by Hamish Guthrie 12 Apr '05

12 Apr '05

Hi, Just a small note: In order to speed up flash programming, I have a small note which may be useful - when a flash device is erased, all bits are set to 0xff - when programming the device, any bit which is not a 1 is changed to a 0, therefore, it is advantageous for speed purposes to have any byte which does not need to be programmed in the flash device set to 0xff. Currently, objcopy sets all padding bytes to 0x00 which adds to the programming time considerably. If you merely insert '--gap-fill 0xff' into all lines in freebios2/src/config/Config.lb where OBJCOPY appears, the resultant binary will have 0xff's as the fill character instead of 0x00's. There are 2 ocurrences of this in the Config.lb file, and the lines should read: action "$(OBJCOPY) --gap-fill 0xff -O binary linuxbios linuxbios.strip" and action "$(OBJCOPY) --gap-fill 0xff -O binary $< $@" respectively. Regards Hamish -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.308 / Virus Database: 266.9.5 - Release Date: 07.04.2005

2 1

[LinuxBIOS] readl and read32
by YhLu 12 Apr '05

12 Apr '05

Do you think read32 arch/i386/include/arch/romcc_io.h and readl in arch/i386/include/arch/io.h should be removed one? YH

2 1

[LinuxBIOS] Geode in Linuxbios v2
by Hamish Guthrie 12 Apr '05

12 Apr '05

Hi, I have been following with interest discussions from Ramesh Bios and others relating to the Geode processor. I have a few boards here with GX1's with the CX5530 and PC97307/PC97317 superio, and am getting started with getting LinuxBIOS onto those boards. I have a board booting up to the point of attempting to load an ELF image, at which point it craps out! This is using the v1 code tree. My project would appear to be a long-term one and I was wondering if it would not be beneficial to all to port all of the GX1 related code to LinuxBIOS v2. Ramesh, I saw in one of your posts that you had mentioned willingness to start working on this - if you have started, maybe we could work together on this? I have had a brief look at the code in question to attempt to see why the elfboot code is failing, but have not found a logical reason yet. The one scarry debug message I get is the following: Welcome to elfboot, the open sourced starter. January 2002, Eric Biederman. Version 1.2 null op: 0x07 eax:0x00000000 ebx:0x00000000 ecx:0x00000000 edx:0x00000000 op: 0x08 eax:0x00000000 ebx:0x00000000 ecx:0x00000000 edx:0x00000000 op: 0x09 eax:0x00000000 ebx:0x00000000 ecx:0x00000000 edx:0x00000000 op: 0x0a eax:0x00000000 ebx:0x00000000 ecx:0x00000000 edx:0x00000000 ... ... ... It would appear as though the null is being printed out by the stream->init() call in the elfboot function in $(TOP)/lib/elfboot.c. I have had a further look into the code and do not see anywhere where the init function pointer within the stream struct in question is initialised. I have noticed in v2 that there is a completely different code interface to the stream stuff, and rather than patching broken code, maybe it is a good idea to look at moving the working GX1 code to v2. I have not looked at the v2 tree much, and am sure there is a learning curve to start porting the stuff across. Any pointers as to where there may be some docs relating to the porting process? Thanks in advance Hamish -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.308 / Virus Database: 266.9.3 - Release Date: 05.04.2005

3 2

[LinuxBIOS] a great quote
by Ronald G. Minnich 12 Apr '05

12 Apr '05

"You'd need to know the confidential information about the chips to write" a free BIOS, Insyde Software's Joseph said. Right now, "that info is only available on old hardware that nobody really cares about anymore." http://news.com.com/Battle+brews+over+unlocking+PC+secrets/2100-1016_3-5654… garsh, didn't know that opterons are so obsolete. It's amazing how much lack-of-clue there is out there masquerading as knowledge. ron

1 0

[LinuxBIOS] Re: post code "fe" means what ?
by Stefan Reinauer 12 Apr '05

12 Apr '05

* Huang-Jen Wang <huangjen.wang(a)gmail.com> [050412 10:46]: > Hi Stefan, > The message as follow: > setup_timers: CPU 2005 MHz > pci_init: Scanning PCI: found 4 devices > malloc_diag: alloc: 184 bytes (4 blocks), free: 16192 bytes (1 blocks) > pci_init: 00:18.0 1022:1100 0600 00 > pci_init: 00:18.1 1022:1101 0600 00 > pci_init: 00:18.2 1022:1102 0600 00 > pci_init: 00:18.3 1022:1103 0600 00 > Press <Enter> for default boot, or <Esc> for boot prompt... timed out > boot: hda2:/boot/vmlinuz root=/dev/hda2 console=tty0 console=ttyS0,115200 > malloc_diag: alloc: 264 bytes (5 blocks), free: 16112 bytes (1 blocks) > malloc_diag: alloc: 280 bytes (6 blocks), free: 16096 bytes (1 blocks) > file_open: dev=hda2, path=/boot/vmlinuz > find_ide_controller: PCI IDE #0 not found > IDE channel 0 not found > devopen: failed to open ide The IDE controller is not detected. Did you set PCI_BRUTE_SCAN = 1 in the filo config file? Stefan

1 0

[LinuxBIOS] Questions about romcc
by Hamish Guthrie 11 Apr '05

11 Apr '05

I am busy doing a port of the Geode code to V2, and am running into difficulty with the RAM sizing. The issue is not the RAM sizing itself, but the behaviour of romcc. I have not looked into romcc in any depth at all, but am guessing that my problem results either from the declaration of too many variables or nesting function calls too deeply. Are there any guidelines as to variable usage and function call nesting, and will romcc give any warning messages when all the registers are used up? TIA, Hamish -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.308 / Virus Database: 266.9.5 - Release Date: 07.04.2005

2 1

[LinuxBIOS] Mainboard Question
by Thomas Spear 07 Apr '05

07 Apr '05

I'm considering a mainboard that has a via C3 cpu and uses a chipset comprised of a via CLE266 and a via VT8235; it uses ddr266 and has integrated via 2 channel audio, 10/100 lan and UniChrome vga (64M shared mamory). I'm hoping that since this is almost identical to an epia-m board, the epia linux bios should work "as is". Is that correct? If I got a working image from someone, would the odds be good that it would work on this board?

1 0

Re: [LinuxBIOS] Geode in Linuxbios v2
by ramesh bios 07 Apr '05

07 Apr '05

> Ramesh, I saw in one of your posts that you had > mentioned willingness to > start working on this - if you have started, maybe > we could work together on > this? > Definitely, I'd be happy to work with you on the V2 conversion. I haven't done anything on V2 yet. Maybe I'll go do a V2 build sometime this weekend and get a feel for how things are laid out there. Ramesh __________________________________ Yahoo! Messenger Show us what our next emoticon should look like. Join the fun. http://www.advision.webevents.yahoo.com/emoticontest

2 1

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

coreboot