coreboot May 2005

coreboot@coreboot.org

36 participants
85 discussions

RE: [LinuxBIOS] CK804 suggestions
by Stephen.Kimball＠bench.com 18 May '05

18 May '05

I agree with you... I can't use a label in the romstrap file. -----Original Message----- From: YhLu [mailto:YhLu@tyan.com] Sent: Wednesday, May 18, 2005 4:08 PM To: Kimball, Stephen; yinghailu(a)gmail.com Cc: linuxbios(a)openbios.org Subject: RE: [LinuxBIOS] CK804 suggestions send out the Config.lb You need to have romstrap.inc and romstrap.lds in your MB dir. YH > -----Original Message----- > From: Stephen.Kimball(a)bench.com [mailto:Stephen.Kimball@bench.com] > Sent: Wednesday, May 18, 2005 12:43 PM > To: YhLu; yinghailu(a)gmail.com > Cc: linuxbios(a)openbios.org > Subject: RE: [LinuxBIOS] CK804 suggestions > > I get "undefined reference to `mac_address'". The > romstrap.inc is included in mainboard's Config.lb, but it's > not linked with ck804_nic.o. > So you need to use a hardcoded address. > > Steve

1 0

RE: [LinuxBIOS] CK804 suggestions
by Stephen.Kimball＠bench.com 18 May '05

18 May '05

I get "undefined reference to `mac_address'". The romstrap.inc is included in mainboard's Config.lb, but it's not linked with ck804_nic.o. So you need to use a hardcoded address. Steve -----Original Message----- From: YhLu [mailto:YhLu@tyan.com] Sent: Wednesday, May 18, 2005 2:40 PM To: Kimball, Stephen; yinghailu(a)gmail.com; Van Echaute, Peter Cc: linuxbios(a)openbios.org Subject: RE: [LinuxBIOS] CK804 suggestions don't need. Also if do that, the linker will complain the overlapping. YH > -----Original Message----- > From: Stephen.Kimball(a)bench.com [mailto:Stephen.Kimball@bench.com] > Sent: Wednesday, May 18, 2005 11:15 AM > To: yinghailu(a)gmail.com; Peter.VanEchaute(a)bench.com > Cc: linuxbios(a)openbios.org > Subject: RE: [LinuxBIOS] CK804 suggestions > > OK, so why not use a label to get the MAC address at 0xffffffd0? > > Steve > > -----Original Message----- > From: linuxbios-bounces(a)openbios.org > [mailto:linuxbios-bounces@openbios.org] On Behalf Of yhlu > Sent: Wednesday, May 18, 2005 1:24 PM > To: Van Echaute, Peter > Cc: linuxbios(a)openbios.org > Subject: Re: [LinuxBIOS] CK804 suggestions > > 1. In NV NDA doc. > 2. That is trick, in romstrap.inc, rspointers must be > 0xffffffe0, So just put MAC below that in 16bytes, and it can > be fixed. actually it is not some kind of part romstrap, just > want to make it in fixed positon. Fixed position is better > than floating..., it may make your flash util more easier to find it. > > YH > > .long 0x81543266 // 30h, MAC address > low 4 byte ---> keep it in 0xffffffd0 > .long 0x000000E0 // 34h, MAC > address high 4 byte > > .long 0x002309CE // 38h, UUID > low 4 byte > .long 0x00E08100 // 3Ch, UUID > high 4 byte > > rspointers: > .long rstables // It will be > 0xffffffe0 > .long rstables > .long rstables > .long rstables > > .globl __romstrap_end > > _______________________________________________ > LinuxBIOS mailing list > LinuxBIOS(a)openbios.org > http://www.openbios.org/mailman/listinfo/linuxbios > > _______________________________________________ > LinuxBIOS mailing list > LinuxBIOS(a)openbios.org > http://www.openbios.org/mailman/listinfo/linuxbios >

1 0

Re: [LinuxBIOS] makefile for Windows
by Toto Lebolo 18 May '05

18 May '05

Yes I am trying to compile ADLO. MiKL~ Richard Smith <smithbone(a)gmail.com> wrote: On 5/17/05, Ronald G. Minnich wrote: > uh, bcc? > > why bcc? > ADLO needs it to compile bochs bios. Are you trying to compile ADLO? -- Richard A. Smith --------------------------------- Yahoo! Mail Mobile Take Yahoo! Mail with you! Check email on your mobile phone.

3 2

RE: [LinuxBIOS] CK804 suggestions
by YhLu 18 May '05

18 May '05

don't need. Also if do that, the linker will complain the overlapping. YH > -----Original Message----- > From: Stephen.Kimball(a)bench.com [mailto:Stephen.Kimball@bench.com] > Sent: Wednesday, May 18, 2005 11:15 AM > To: yinghailu(a)gmail.com; Peter.VanEchaute(a)bench.com > Cc: linuxbios(a)openbios.org > Subject: RE: [LinuxBIOS] CK804 suggestions > > OK, so why not use a label to get the MAC address at 0xffffffd0? > > Steve > > -----Original Message----- > From: linuxbios-bounces(a)openbios.org > [mailto:linuxbios-bounces@openbios.org] On Behalf Of yhlu > Sent: Wednesday, May 18, 2005 1:24 PM > To: Van Echaute, Peter > Cc: linuxbios(a)openbios.org > Subject: Re: [LinuxBIOS] CK804 suggestions > > 1. In NV NDA doc. > 2. That is trick, in romstrap.inc, rspointers must be > 0xffffffe0, So just put MAC below that in 16bytes, and it can > be fixed. actually it is not some kind of part romstrap, just > want to make it in fixed positon. Fixed position is better > than floating..., it may make your flash util more easier to find it. > > YH > > .long 0x81543266 // 30h, MAC address > low 4 byte ---> keep it in 0xffffffd0 > .long 0x000000E0 // 34h, MAC > address high 4 byte > > .long 0x002309CE // 38h, UUID > low 4 byte > .long 0x00E08100 // 3Ch, UUID > high 4 byte > > rspointers: > .long rstables // It will be > 0xffffffe0 > .long rstables > .long rstables > .long rstables > > .globl __romstrap_end > > _______________________________________________ > LinuxBIOS mailing list > LinuxBIOS(a)openbios.org > http://www.openbios.org/mailman/listinfo/linuxbios > > _______________________________________________ > LinuxBIOS mailing list > LinuxBIOS(a)openbios.org > http://www.openbios.org/mailman/listinfo/linuxbios >

1 0

RE: [LinuxBIOS] CK804 suggestions
by Stephen.Kimball＠bench.com 18 May '05

18 May '05

OK, so why not use a label to get the MAC address at 0xffffffd0? Steve -----Original Message----- From: linuxbios-bounces(a)openbios.org [mailto:linuxbios-bounces@openbios.org] On Behalf Of yhlu Sent: Wednesday, May 18, 2005 1:24 PM To: Van Echaute, Peter Cc: linuxbios(a)openbios.org Subject: Re: [LinuxBIOS] CK804 suggestions 1. In NV NDA doc. 2. That is trick, in romstrap.inc, rspointers must be 0xffffffe0, So just put MAC below that in 16bytes, and it can be fixed. actually it is not some kind of part romstrap, just want to make it in fixed positon. Fixed position is better than floating..., it may make your flash util more easier to find it. YH .long 0x81543266 // 30h, MAC address low 4 byte ---> keep it in 0xffffffd0 .long 0x000000E0 // 34h, MAC address high 4 byte .long 0x002309CE // 38h, UUID low 4 byte .long 0x00E08100 // 3Ch, UUID high 4 byte rspointers: .long rstables // It will be 0xffffffe0 .long rstables .long rstables .long rstables .globl __romstrap_end _______________________________________________ LinuxBIOS mailing list LinuxBIOS(a)openbios.org http://www.openbios.org/mailman/listinfo/linuxbios

1 0

RE: [LinuxBIOS] CK804 suggestions
by Peter.VanEchaute＠bench.com 18 May '05

18 May '05

1. I didn't see that 0xA8 in the device config space was for the MAC address. It doesn't show it in the register spec I received. I'll give it a try when I get a chance. I only saw that offset 0xA8 in the MMIO space was for it. 2. Taking a closer look at it, I do like it. Maybe just change the if(!eeprom_valid) part to read from the ROMSTRAP the way I had it. The reason I like my way of reading the ROMSTRAPs MAC is because you have it at a fixed address. But the ROMSTRAP could actually be used to setup lots and lots of device registers, if chosen to do so. Thus pushing that address lower. // read directly from rom_strap mac_pos = read8(0xffffffe0); // refer to romstrap.inc mac_pos |= (read8(0xffffffe1)<<8); mac_pos |= (read8(0xffffffe2)<<16); mac_pos |= (read8(0xffffffe3)<<24); mac_pos += 0x30; for ( dword = nic_index ; dword > 0 ; dword-- ) { mac_pos += 8; } mac_l = read32(mac_pos); mac_h = read32(mac_pos + 4); Sorry I didn't look at it more closely. 3. yeah, I saw that. I was just using something familiar. Thanks.

2 1

RE: [LinuxBIOS] CK804 suggestions
by Peter.VanEchaute＠bench.com 18 May '05

18 May '05

I too have done a few boards where the MAC address has been stored in the flash device. The idea was to save on cost, the same as yours Ron. This matter of board design is exactly that, board specific. The CK804 has a specific place in flash that the MAC(s) should be placed anyways, in the ROMSTRAP+0x30. I noticed you knew that YH. As the CK804 area is chipset specific, it shouldn't be board specific. So what I did is placed a define for your TYAN board, at least until you take your stuff out. In the code, it reads the ROMSTRAP offset + 0x30. Each MAC address is an +0x8 from there if you have more than one. This is exactly how the CK804 BIOS Porting guides state to do it. Now do you have to listen to me or the guide, no. It's just my $0.02. On a side note YH, you have a bug in the ck804_nic.c. I hope it's OK to point this out. But the MAC address is supposed to be placed into the Memmory Mapped IO (MMIO) area at offset 0xA8, and not in the device configuration space... My recommendation... // set that into NIC MMIO write32(mac_l, base + 0xA8); write32(mac_h, base + 0xAC); your code... // set that into NIC pci_write_config32(dev, 0xa8, mac_l); pci_write_config32(dev, 0xac, mac_h); Here is the code I recommend(or at least something close to it)... #include <console/console.h> #include <device/device.h> #include <device/smbus.h> #include <device/pci.h> #include <device/pci_ids.h> #include <device/pci_ops.h> #include <arch/io.h> #include "ck804.h" #define MAC_DBG 1 #define TYANs2895 0 uint8_t read8(unsigned long addr) { return *((volatile uint8_t *)(addr)); } uint16_t read16(unsigned long addr) { return *((volatile uint16_t *)(addr)); } uint32_t read32(unsigned long addr) { return *((volatile uint32_t *)(addr)); } void write8(unsigned long addr, uint8_t value) { *((volatile uint8_t *)(addr)) = value; } void write16(unsigned long addr, uint16_t value) { *((volatile uint16_t *)(addr)) = value; } void write32(unsigned long addr, uint32_t value) { *((volatile uint32_t *)(addr)) = value; } static void nic_init(struct device *dev) { uint32_t dword, old; uint32_t mac_h, mac_l; static uint32_t nic_index = 0; unsigned long mac_pos; uint32_t base; struct resource *res; #define NvRegPhyInterface 0xC0 #define PHY_RGMII 0x10000000 res = find_resource(dev, 0x10); base = res->base; write32(base + NvRegPhyInterface, PHY_RGMII); old = dword = pci_read_config32(dev, 0x30); dword &= ~(0xf); dword |= 0xf; if(old != dword) { pci_write_config32(dev, 0x30 , dword); } #if TYAN int eeprom_valid = 0; struct southbridge_nvidia_ck804_config *conf; conf = dev->chip_info; if(conf->mac_eeprom_smbus != 0) { // read MAC address from EEPROM at first struct device *dev_eeprom; dev_eeprom = dev_find_slot_on_smbus(conf->mac_eeprom_smbus, conf->mac_eeprom_addr); if(dev_eeprom) { // if that is valid we will use that unsigned char dat[6]; int status; int i; for(i=0;i<6;i++) { status = smbus_read_byte(dev_eeprom, i); if(status < 0) break; dat[i] = status & 0xff; } if(status >= 0) { mac_l = 0; for(i=3;i>=0;i--) { mac_l <<= 8; mac_l += dat[i]; } if(mac_l != 0xffffffff) { mac_l += nic_index; mac_h = 0; for(i=5;i>=4;i--) { mac_h <<= 8; mac_h += dat[i]; } eeprom_valid = 1; } } } } // if that is invalid we will read that from romstrap if(!eeprom_valid) { mac_pos = 0xffffffd0; // refer to romstrap.inc and romstrap.lds mac_l = read32(mac_pos) + nic_index; mac_h = read32(mac_pos + 4); } #else // read directly from rom_strap mac_pos = read8(0xffffffe0); // refer to romstrap.inc mac_pos |= (read8(0xffffffe1)<<8); mac_pos |= (read8(0xffffffe2)<<16); mac_pos |= (read8(0xffffffe3)<<24); mac_pos += 0x30; for ( dword = nic_index ; dword > 0 ; dword-- ) { mac_pos += 8; } mac_l = read32(mac_pos); mac_h = read32(mac_pos + 4); #endif // set that into NIC MMIO write32(mac_l, base + 0xA8); write32(mac_h, base + 0xAC); #if MAC_DBG printk_info("MAC:\tMAC Address is: %02x:%02x:%02x:%02x:%02x:%02x\n", read8(base + 0xAD), read8(base + 0xAC), read8(base + 0xAB), read8(base + 0xAA), read8(base + 0xA9), read8(base + 0xA8)); #endif nic_index++; #if CONFIG_PCI_ROM_RUN == 1 pci_dev_init(dev);// it will init option rom #endif } static void lpci_set_subsystem(device_t dev, unsigned vendor, unsigned device) { pci_write_config32(dev, 0x40, ((device & 0xffff) << 16) | (vendor & 0xffff)); } static struct pci_operations lops_pci = { .set_subsystem = lpci_set_subsystem, }; static struct device_operations nic_ops = { .read_resources = pci_dev_read_resources, .set_resources = pci_dev_set_resources, .enable_resources = pci_dev_enable_resources, .init = nic_init, .scan_bus = 0, // .enable = ck804_enable, .ops_pci = &lops_pci, }; static struct pci_driver nic_driver __pci_driver = { .ops = &nic_ops, .vendor = PCI_VENDOR_ID_NVIDIA, .device = PCI_DEVICE_ID_NVIDIA_CK804_NIC, }; static struct pci_driver nic_bridge_driver __pci_driver = { .ops = &nic_ops, .vendor = PCI_VENDOR_ID_NVIDIA, .device = PCI_DEVICE_ID_NVIDIA_CK804_NIC_BRIDGE, };

2 1

RE: [LinuxBIOS] CK804 suggestions
by Stephen.Kimball＠bench.com 18 May '05

18 May '05

We are a contract manufacturer and don't have a design yet. We are working with the nVidia CK804 Customer Reference Board. They read the MAC address from the ROMSTRAP table. It works, but it's not a good design. Steve -----Original Message----- From: YhLu [mailto:YhLu@tyan.com] Sent: Tuesday, May 17, 2005 6:43 PM To: Kimball, Stephen; rminnich(a)lanl.gov Cc: linuxbios(a)openbios.org Subject: RE: [LinuxBIOS] CK804 suggestions what is your HW design about that? YH

1 0

RE: [LinuxBIOS] makefile for Windows
by Toto Lebolo 18 May '05

18 May '05

I am getting now an error with bcc. bcc -o rombios.s -C-c -D__i86__ -0 -S _rombios_.c Unable to execute bcc-cc1.exe I don't know why I am getting this error the file is in the path. MiKL~ Gregg C Levine <hansolofalcon(a)worldnet.att.net> wrote: Hello from Gregg C Levine Are you using the Cygwin stuff for your work? For reasons that I've never bothered to investigate, the people behind the entire Cygwin collection of GNU utilities use the bleeding edge of the source code that makes up the regular GNU collection. For example, try your make file under Linux, and report back your results. I suspect your seeing the effects of the maintainers of GCC to perfect their compiler even further. However this is good news that at least something happens. And Stefan the next time your visiting NYC, and the US, the drinks are on me. I would not of thought of suggesting it. ------------------- Gregg C Levine hansolofalcon(a)worldnet.att.net ------------------------------------------------------------ "The Force will be with you...Always." Obi-Wan Kenobi "Use the Force, Luke." Obi-Wan Kenobi -----Original Message----- From: linuxbios-bounces(a)openbios.org [mailto:linuxbios-bounces@openbios.org] On Behalf Of Toto Lebolo Sent: Thursday, May 12, 2005 5:48 PM To: Stefan Reinauer Cc: linuxbios(a)openbios.org Subject: Re: [LinuxBIOS] makefile for Windows Apparently it does not run. I started modifying the commands in the makefile, and started compiling rombios.c I am getting this error: rombios.c:2662:1: pasting "(" and ""ata_cmd_data_in : read error\n"" does not give a valid preprocessing token the source line is: BX_DEBUG_ATA("ata_cmd_data_in : read error\n"); Do you think it is due to the environment? MiKL Stefan Reinauer wrote: * Toto Lebolo [050512 20:02]: > I am trying to rebuild linux BIOS. I have installed the Windows version of the > compilers. > Does the makefile run in windows? or just in Linux? Find out and tell us, please! You'll also need python installed. Stefan Yahoo! Mail Mobile Take Yahoo! Mail with you! Check email on your mobile phone. _______________________________________________ LinuxBIOS mailing list LinuxBIOS(a)openbios.org http://www.openbios.org/mailman/listinfo/linuxbios --------------------------------- Yahoo! Mail Mobile Take Yahoo! Mail with you! Check email on your mobile phone.

3 2

RE: [LinuxBIOS] CK804 suggestions
by YhLu 18 May '05

18 May '05

In CMOS? unbelievable. Factory personal and service guys will keep kicking the design engineer and blaming their design. YH > -----Original Message----- > From: Ronald G. Minnich [mailto:rminnich@lanl.gov] > Sent: Tuesday, May 17, 2005 3:04 PM > To: YhLu > Cc: Stephen.Kimball(a)bench.com; linuxbios(a)openbios.org > Subject: RE: [LinuxBIOS] CK804 suggestions > > > > On Tue, 17 May 2005, YhLu wrote: > > > don't need, > > 1. good hw design should use one serial EEPROM to store MAC > address, > > So when flash the BIOS, the EEPROM is still there. > > yes, but there were some sis-630 based cards several years > ago that stored MAC in FLASH and CMOS. > > SO to upgrade FLASH: > - make sure address is in cmos > - write flash > - write MAC back to flash. > > > There are some stupid designs out there. > > ron >

2 1

← Newer
1
2
3
4
5
6
7
8
9
Older →

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

coreboot May 2005