coreboot February 2004

coreboot@coreboot.org

43 participants
63 discussions

src/northbridge/amdk8/coherent_ht.c
by Stefan Reinauer 17 Feb '04

17 Feb '04

Hi, can't we drop optimize_connection() from coherent_ht.c? The link optimization is done later, too, during PCI initialization and currently it assumes a fixed link setup. (ACROSS links of node 0 and 1 used to connect CPUs) Stefan -- Stefan Reinauer, SUSE LINUX AG Teamleader Architecture Development

1 0

compile error
by Fathi BOUDRA 17 Feb '04

17 Feb '04

Hi, I'm a newbie on linuxbios so i followed the howto. But when i arrived to run the make command : debfat:~/dev/freebios/util/config/winfast# make Makefile:517: attention : écrasement des commandes pour la cible « keyboard.o » Makefile:421: attention : anciennes commandes ignorées pour la cible « keyboard.o » cp /root/dev/freebios/src/arch/i386/config/crt0.base crt0.S gcc -x assembler-with-cpp -DASSEMBLY -E ... crt0.S > crt0.s /bin/sh: -c: line 1: syntax error near unexpected token `(' /bin/sh: -c: line 1: `gcc -x assembler-with-cpp -DASSEMBLY -E -I/root/dev/freebios/src/include -I/root/dev/freebios/src/arch/i386/include -I -DCC='gcc' -DCMD_LINE='"root=/dev/hda1 console=ttyS0,115200 console=tty0 video=sisfb:1024x768-32@85,font:VGA8x16"' -DCONFIGURE_L2_CACHE='1' -DCONFIG_ASSIGNIRQ='0' -DCONFIG_COMPRESS='1' -DCONFIG_LOGICAL_CPUS='1' -DCONFIG_MC146818RTC='1' -DCONFIG_PCIBIOS_IRQ='0' -DCONFIG_TSC_X86RDTSC_CALIBRATE_WITH_TIMER2='0' -DCONFIG_UNCOMPRESSED='0' -DCRT0='/root/dev/freebios/src/arch/i386/config/crt0.base' -DENABLE_FIXED_AND_VARIABLE_MTRRS='1' -DENABLE_SIS630_CPU_PIPELINE='1' -DFINAL_MAINBOARD_FIXUP='1' -DHAVE_FRAMEBUFFER='1' -DHAVE_PIRQ_TABLE='1' -DHEAP_SIZE='0x40000' -DHOSTCC='gcc' -DINTEL_PPRO_MTRR='1' -DLINUXBIOS_ASSEMBLER='Version de l'assembleur GNU 2.14.90.0.7 (i386-linux) utilisant la version BFD 2.14.90.0.7 20031029 Debian GNU/Linux' -DLINUXBIOS_BUILD='lun fév 16 16:58:23 CET 2004' -DLINUXBIOS_COMPILER='version gcc 3.3.3 20040214 (prerelease) (Debian)' -DLINUXBIOS_COMPILE_BY='root' -DLINUXBIOS_COMPILE_DOMAIN -DLINUXBIOS_COMPILE_HOST='debfat' -DLINUXBIOS_COMPILE_TIME='16:58:23' -DLINUXBIOS_LINKER='GNU ld version 2.14.90.0.7 20031029 Debian GNU/Linux' -DLINUXBIOS_VERSION='1.0.0' -DMAINBOARD_PART_NUMBER='winfast6300' -DMAINBOARD_VENDOR='leadtek' -DMAX_CPUS='1' -DMEMORY_HOLE='1' -DOBJCOPY='objcopy' -DPAYLOAD_SIZE='458752' -DROM_IMAGE_SIZE='65536' -DSERIAL_CONSOLE='1' -DSIS630='1' -DSMA_SIZE='0xA0' -DSTACK_SIZE='0x2000' -DUPDATE_MICROCODE='1' -DUSE_DOC='1' -DUSE_DOC_2000_TSOP='0' -DUSE_DOC_MIL='1' -DUSE_NEW_SUPERIO_INTERFACE='1' -D_RAMBASE='0x4000' -D_ROMBASE='0x80000' -Di586='1' -Di686='1' -I/root/dev/freebios/src crt0.S > crt0.s.new && mv crt0.s.new crt0.s' make: *** [crt0.s] Erreur 2 is there a problem with the version of software i use ? (gcc 3.3.3, python, perl, shell,...) Maybe i need to have a redhat 6.2 ? I think it's a common problem but don't find anything on the mailling list. when i change in the Makefile, the line : D_item = $(if $(subst undefined,,$(origin $1)),-D$1$(if $($1),='$($1)',),-U$1) by D_item = $(if $(subst undefined,,$(origin $1)),-D$1$(if $($1),="$($1)",),-U$1) i've got : debfat:~/dev/freebios/util/config/winfast# make Makefile:517: attention : écrasement des commandes pour la cible « keyboard.o » Makefile:421: attention : anciennes commandes ignorées pour la cible « keyboard.o » cp /root/dev/freebios/src/arch/i386/config/crt0.base crt0.S gcc -x assembler-with-cpp -DASSEMBLY -E ... crt0.S > crt0.s gcc: console=ttyS0,115200 : Aucun fichier ou répertoire de ce type gcc: console=tty0 : Aucun fichier ou répertoire de ce type gcc: video=sisfb:1024x768-32@85,font:VGA8x16 : Aucun fichier ou répertoire de ce type make: *** [crt0.s] Erreur 1 if someone have an idea ... best regards fathi

1 1

答复: Use of compact flash
by YhLu 16 Feb '04

16 Feb '04

Ron, Do you mean put kernel and rootfs on the CF? Regards YH. -----邮件原件----- 发件人: ron minnich [mailto:rminnich@lanl.gov] 发送时间: 2004年2月16日 15:58 收件人: Andy Pearce 抄送: linuxbios(a)clustermatic.org 主题: Re: Use of compact flash On Mon, 16 Feb 2004, Andy Pearce wrote: > "I think you're much better off with an adapter for compact flash. That's > an open, widely used standard which has worked well for us." > Can you explain this a bit, does it mean that its possible to > some how redirect the rom from which the PC boots ? i.e. switch > it to a CF adapter card ? you boot linuxbios from flash, but the linuxbios payload comes from CF. ron _______________________________________________ Linuxbios mailing list Linuxbios(a)clustermatic.org http://www.clustermatic.org/mailman/listinfo/linuxbios

2 1

IBM T20
by DAVID CUMMINGS 14 Feb '04

14 Feb '04

Alright, the status is IBM doesn't want it. I know this has probably been gone over hundreds of times before. What's the deal with the T20 and LinuxBIOS. Also, how would I go about searching the archives? I couldn't find a search form anywhere, and I'm on dialup so I can't d/l the entire archive and grep it or similar. Anyway, thanks -dave

2 1

ATI XL Init in kernel 2.6.2
by YhLu 14 Feb '04

14 Feb '04

Stefan, I have checked Kernel 2.6.2 for ATI XL No init in BIOS, the Makefile doesn't include the xlinit.c and it is not called. So patch is not properly merged with the kernel 2.6.2. Did you try that ever? Regards Yinghai Lu

2 1

Advice on Board Selection
by James Weir 14 Feb '04

14 Feb '04

Can anyone please recommend a board preferably small form factor ( itx? ) which the documentation is available for ( or a port of linux bios is available for ) which would be suitable for dvd playback? Thanks, James.

4 6

freebios2/util/flash_and_burn
by ron minnich 13 Feb '04

13 Feb '04

I just fixed a makefile problem that was introduced with the AMD 8111 support, so if you had trouble making it, do a cvs update. ron

1 0

linuxbios on cheaper motherboards
by Andre Dias 13 Feb '04

13 Feb '04

Hello Fellows, I have been using linuxbios for a while on some products of my company, we make pc-based jukeboxes and arcade machines for pubs and public venues and I have been using linuxbios and DOC to make a fast boot and prevent the computer to hang at boot when CMOS clears, and to store some information on DOC and prevent hard disk corruption problems. We also have busybox on DOC, so we can verify problems and provide maintenance even when HD fails. We used pcchips M810 (red motherboard) and ECS K7SEM (purple). We used this board mainly because it was the cheapest and worked fine with linuxbios. Now this motherboards are impossible to find on retailers and SDRAM modules are becoming scarce. I have seen that EPIA is being actively developed but EPIA is not really easy to find and cheap (at least in south america), also its CPU is not the best performance. I was thinking of using PCCHIPS m810D (SIS 740, closer to SIS730) but its PLCC BIOS is soldered onboard! So I had to switch to ASROCK K7VM2 (via KM266), and we could afford donnate $400 and this motherboard plus cpu and memory, to a developer that could make linuxbios work on this board (with sound, eth0, framebuffer, IDE and serial port working). Its a very small reward, but maybe it will help any developer that is already working on it. thanks for the attention

4 3

Etherboot 5.0.10 problems + patchfile
by Dave Ashley 13 Feb '04

13 Feb '04

I'm going to post this patch to the etherboot mailing list but it is relevant here as well. This is our current patch against etherboot 5.0.10 to fix various problems: 1) RTL8139 interrupt handling was wrong 2) VIA-RHINE support needs workaround for the 3065 chip version. When resetting the chip, the act of shutting down the receive machinery can cause a problem, and when the chip is reset it never gets out of reset. So etherboot would hang when trying to send out a packet, such as a bootp request. This fix is from via's linuxfet driver. Google searches tracked down other people with the same problem and the eventual solution. 3) Some issue with initializing the timer, which without it was causing a divide by zero at some point. The #2 was an especially bad problem and affected our epia-m solution. The boxes would stop sending out BOOTP requests after a while. -Dave diff -Nur etherboot-5.0.10/src/misc.c etherboot-5.0.10-new/src/misc.c --- etherboot-5.0.10/src/misc.c Sun Mar 9 03:17:15 2003 +++ etherboot-5.0.10-new/src/misc.c Wed Jul 16 09:45:26 2003 @@ -282,7 +282,7 @@ enum { Disable_A20 = 0x2400, Enable_A20 = 0x2401, Query_A20_Status = 0x2402, Query_A20_Support = 0x2403 } Int0x15Arg; -#if defined(PCBIOS) && !defined(IBM_L40) +#if defined(TAGGED_IMAGE) || (defined(PCBIOS) && !defined(IBM_L40)) static void empty_8042(void) { unsigned long time; diff -Nur etherboot-5.0.10/src/rtl8139.c etherboot-5.0.10-new/src/rtl8139.c --- etherboot-5.0.10/src/rtl8139.c Tue Jul 23 17:50:53 2002 +++ etherboot-5.0.10-new/src/rtl8139.c Wed Jul 16 09:45:42 2003 @@ -183,7 +183,6 @@ static int rtl_poll(struct nic *nic); static void rtl_disable(struct nic*); - struct nic *rtl8139_probe(struct nic *nic, unsigned short *probeaddrs, struct pci_device *pci) { @@ -252,6 +251,16 @@ #define EE_READ_CMD (6) #define EE_ERASE_CMD (7) +static unsigned short cintr=0; +static unsigned short getstatus(void) +{ + return cintr|=inw(ioaddr + IntrStatus); +} +static void setstatus(unsigned short val) +{ + cintr&=~val; +} + static int read_eeprom(int location, int addr_len) { int i; @@ -340,7 +349,7 @@ static void rtl_transmit(struct nic *nic, const char *destaddr, unsigned int type, unsigned int len, const char *data) { - unsigned int status, to, nstype; + unsigned int status, to, nstype, st; unsigned long txstatus; /* nstype assignment moved up here to avoid gcc 3.0.3 compiler bug */ @@ -365,14 +374,15 @@ outl(((TX_FIFO_THRESH<<11) & 0x003f0000) | len, ioaddr + TxStatus0 + cur_tx*4); + st= currticks(); to = currticks() + RTL_TIMEOUT; do { - status = inw(ioaddr + IntrStatus); + status = getstatus(); /* Only acknlowledge interrupt sources we can properly handle * here - the RxOverflow/RxFIFOOver MUST be handled in the * rtl_poll() function. */ - outw(status & (TxOK | TxErr | PCIErr), ioaddr + IntrStatus); + setstatus(status & (TxOK | TxErr | PCIErr)); if ((status & (TxOK | TxErr | PCIErr)) != 0) break; } while (currticks() < to); @@ -382,12 +392,12 @@ cur_tx = (cur_tx + 1) % NUM_TX_DESC; #ifdef DEBUG_TX printf("tx done (%d ticks), status %hX txstatus %X\n", - to-currticks(), status, txstatus); + currticks()-st, status, txstatus); #endif } else { #ifdef DEBUG_TX printf("tx timeout/error (%d ticks), status %hX txstatus %X\n", - currticks()-to, status, txstatus); + currticks()-st, status, txstatus); #endif rtl_reset(nic); } @@ -403,9 +413,9 @@ return 0; } - status = inw(ioaddr + IntrStatus); + status = getstatus(); /* See below for the rest of the interrupt acknowledges. */ - outw(status & ~(RxFIFOOver | RxOverflow | RxOK), ioaddr + IntrStatus); + setstatus(status & ~(RxFIFOOver | RxOverflow | RxOK)); #ifdef DEBUG_RX printf("rtl_poll: int %hX ", status); @@ -449,7 +459,7 @@ /* See RTL8139 Programming Guide V0.1 for the official handling of * Rx overflow situations. The document itself contains basically no * usable information, except for a few exception handling rules. */ - outw(status & (RxFIFOOver | RxOverflow | RxOK), ioaddr + IntrStatus); + setstatus(status & (RxFIFOOver | RxOverflow | RxOK)); return 1; } diff -Nur etherboot-5.0.10/src/timer.c etherboot-5.0.10-new/src/timer.c --- etherboot-5.0.10/src/timer.c Sat Mar 22 11:52:25 2003 +++ etherboot-5.0.10-new/src/timer.c Wed Jul 16 09:45:46 2003 @@ -138,6 +138,8 @@ { unsigned long clocks_high, clocks_low; unsigned long currticks; + +setup_timers(); /* Read the Time Stamp Counter */ rdtsc(clocks_low, clocks_high); --- etherboot-5.0.10/src/Makefile 2003-03-09 17:39:21.000000000 -0800 +++ etherboot-5.0.10/src/Makefile.new 2003-11-06 06:22:26.000000000 -0800 @@ -98,8 +98,8 @@ include Config -GCC= gcc -CPP= gcc -E -Wp,-Wall +GCC= $(CC) +CPP= $(CC) -E -Wp,-Wall STRIP= strip OBJCOPY= objcopy # MAKEROM= bin/makerom diff -Nur etherboot-5.0.10/src/via-rhine.c etherboot-5.0.10-new/src/via-rhine.c --- etherboot-5.0.10/src/via-rhine.c 2003-02-07 15:07:56.000000000 -0800 +++ etherboot-5.0.10-new/src/via-rhine.c 2004-02-06 07:24:50.000000000 -0800 @@ -106,6 +106,10 @@ #define byCFGD ioaddr + 0x7b #define wTallyCntMPA ioaddr + 0x7c #define wTallyCntCRC ioaddr + 0x7d +#define bySTICKHW ioaddr + 0x83 +#define byWOLcrClr ioaddr + 0xA4 +#define byWOLcgClr ioaddr + 0xA7 +#define byPwrcsrClr ioaddr + 0xAC /*--------------------- Exioaddr Definitions -------------------------*/ /* @@ -873,6 +877,19 @@ return nic; } +static void set_rx_mode(struct nic *nic) { + struct rhine_private *tp = (struct rhine_private *) nic->priv_data; + unsigned char rx_mode; + int ioaddr = tp->ioaddr; + + /* ! IFF_PROMISC */ + outl(0xffffffff, byMAR0); + outl(0xffffffff, byMAR4); + rx_mode = 0x0C; + + outb(0x60 /* thresh */ | rx_mode, byRCR ); +} + static struct nic * rhine_probe1 (struct nic *nic, int ioaddr, int chip_id, int options) { @@ -885,6 +902,29 @@ if (rhine_debug > 0 && did_version++ == 0) printf (version); + + /* D-Link provided reset code (with comment additions) */ + if((chip_id != 0x3043) && (chip_id != 0x6100)) { + unsigned char byOrgValue; + + if(rhine_debug > 0) + printf("Enabling Sticky Bit Workaround for Chip_id: 0x%hX\n" + , chip_id); + /* clear sticky bit before reset & read ethernet address */ + byOrgValue = inb(bySTICKHW); + byOrgValue = byOrgValue & 0xFC; + outb(byOrgValue, bySTICKHW); + + /* (bits written are cleared?) */ + /* disable force PME-enable */ + outb(0x80, byWOLcgClr); + /* disable power-event config bit */ + outb(0xFF, byWOLcrClr); + /* clear power status (undocumented in vt6102 docs?) */ + outb(0xFF, byPwrcsrClr); + + } + /* Perhaps this should be read from the EEPROM? */ for (i = 0; i < ETH_ALEN; i++) nic->node_addr[i] = inb (byPAR0 + i); @@ -980,6 +1020,9 @@ struct rhine_private *tp = (struct rhine_private *) nic->priv_data; int ioaddr = tp->ioaddr; + /* merge reset and disable */ + rhine_reset(nic); + printf ("rhine disable\n"); /* Switch to loopback mode to avoid hardware races. */ writeb(0x60 | 0x01, byTCR); @@ -1018,6 +1061,45 @@ /* Soft reset the chip. */ /*outb(CmdReset, ioaddr + ChipCmd); */ + /* if the chip is 3065, we must patch shutdown bug*/ +// if ((np->revision >= 0x40) && (np->revision <= 0x80)) + { + int ww; + //Nic Loop Back On + + outb(inb(byTCR) | 0x01, byTCR); + + //Tx Off + outb(inb(byCR0) & 0xEF, byCR0); + +#define W_MAX_TIMEOUT 0x0FFF + // W_MAX_TIMEOUT is the timeout period + for (ww = 0; ww < W_MAX_TIMEOUT; ww++) { + if((inb(byCR0) & 0x10) == 0) + break; + } + + //Rx Off + outb(inb(byCR0) & 0xF7, byCR0); + + // W_MAX_TIMEOUT is the timeout period + for (ww = 0; ww < W_MAX_TIMEOUT; ww++) { + if((inb(byCR0) & 0x08) == 0) + break; + } + + if (ww == W_MAX_TIMEOUT) { + + // Turn on fifo test + outw(inw(dwNextTDSE1) | 0x0001, dwNextTDSE1); + // Turn on fifo reject + outw(inw(dwNextTDSE1) | 0x0800, dwNextTDSE1); + // Turn off fifo test + outw(inw(dwNextTDSE1) & 0xFFFE, dwNextTDSE1); + } + } + + tx_bufs_tmp = (int) buf1; tx_ring_tmp = (int) desc1; rx_bufs_tmp = (int) buf2; @@ -1069,6 +1151,9 @@ outl (virt_to_bus (tp->rx_ring), dwCurrentRxDescAddr); outl (virt_to_bus (tp->tx_ring), dwCurrentTxDescAddr); + /* Setup Multicast */ + set_rx_mode(nic); + /* close IMR */ outw (0x0000, byIMR0);

1 0

Re: Re: LinuxBIOS....on desktop PCs?
by Hendricks David W. 11 Feb '04

11 Feb '04

I suggest the Linksys WRT54G :) On Tue, 10 Feb 2004 bjm_(a)bellsouth.net wrote: > >turns out the chaintech just added doesn't build, which I forget. Anyone > >wants to take a look, please do. > > Ah...well, maybe later then? Or maybe when I get a nice embedded linux rackmount server to replace my Linksys router, and act as a webserver I maybe could use LinuxBIOS then eh? > > -Brad > > _______________________________________________ > Linuxbios mailing list > Linuxbios(a)clustermatic.org > http://www.clustermatic.org/mailman/listinfo/linuxbios >

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

coreboot February 2004