Hi,ron
Sorry,the correct config is updated.
This is my config file and I writed 0x7c00 to registers to enable all embedded devices.
I also added the support of loading initrd.gz from DOC.(LOAD_INITRD_FROM_MIL)
Thanks.
riskin
# Sample config file for PCCHIPS m758lmr+ with DoC Millennium (as root)
# This will make a target directory of m758lmr+ target /opt/cwlinux/buildrom/m758lmr+
# PCCHIPS m758lmr+ mainboard pcchips/m758lmr+
# Enable Serial Console for debugging option SERIAL_CONSOLE=1
# enable debugging support option DEBUG=1
# enable serial post for debugging option SERIAL_POST=1
# set default consol loglevel option DEFAULT_CONSOLE_LOGLEVEL=9
# enable floppy support # option MUST_ENABLE_FLOPPY
# enable keyboard support # option NO_KEYBOARD
# use ELF boot # option USE_ELF_BOOT=1
# enable RAM test # option RAMTEST
# PIRQ tables option HAVE_PIRQ_TABLE=1 # I8259 INITITIAL option I8259=1 # don't use old kernel hack option OLD_KERNEL_HACK=1
# Enable MicroCode update and L2 Cache init for PII and PIII option UPDATE_MICROCODE=1 option CONFIGURE_L2_CACHE=1 option ENABLE_FIXED_AND_VARIABLE_MTRRS=1
# Use the internal VGA frame buffer device #option HAVE_FRAMEBUFFER=1
# Path to your kernel (vmlinux) linux /opt/cwlinux/linux
option DISABLE_INTERNAL_DEVICES=1
#support loading initrd option LOAD_INITRD_FROM_MIL=1
# Kernel command line parameters # enable serial console commandline auto BOOT_IMAGE=Image root=100 ramdisk=26000 console=ttyS0,115200 console=tty0 BOOTSTEP=0 BOOTDEVICE=fla1 #commandline root=/dev/ram0 ramdisk=14000 init=/linuxrc console=ttyS0,115200 console=tty0 #commandline root=/dev/nftla1 console=ttyS0,115200 console=tty0 video=sisfb:640x480-8@60,font:VGA8x16 # disable serial console #commandline root=/dev/nftla1 console=/dev/tty5 CONSOLE=/dev/tty5 video=sisfb:640x480-8@60,font:VGA8x16
docipl northsouthbridge/sis/630/ipl.S #option USE_DOC_2000_TSOP=1 option USE_DOC_MIL=1
this is a problem with some recent changes to the build process and I don't know what it is yet.
It also depends on version of as. version 2.11.90.0.8 gives this for crt0.o:
00000342 ? console0 00000188 ? console_test 000001c7 ? console_tx_al 00000220 ? console_tx_hex32 000001db ? console_tx_hex8 00000325 ? console_tx_string
note ? 2.13.90.0.2 replaces the ? with 't'. In either case the symbols end up undefined. I think some changes were made to some platforms but not globally. CONFIG_COMPRESS on or off does not help.
I'll keep looking.
I'm going to try to set up regression test build here because we keep getting changes that break builds, and that's inconvient.
ron
Ronald G. Minnich wrote:
this is a problem with some recent changes to the build process and I don't know what it is yet.
I have this issue as well.
If I remove SERIAL_POST from my config files then it builds. Note thats _remove_ not set =0.
./arch/i386/include/arch/intel.h seems to be the offending section for me.
It has a
#if SERIAL_POST
blah.. CALLSP(console_tx_al) blah... CALLSP(console_tx_hex8)
#else ; send to port 80 like normal. #endif
Which means it hasen't been updated to the OPTION=1 system correct? So regardless if SERIAL_POST 1 or 0 this will be built. Perhaps there is a dependancy thats not making it into some of thie builds?
On Mon, 10 Feb 2003, Richard Smith wrote:
Ronald G. Minnich wrote:
this is a problem with some recent changes to the build process and I don't know what it is yet.
If I remove SERIAL_POST from my config files then it builds. Note thats _remove_ not set =0.
./arch/i386/include/arch/intel.h seems to be the offending section for me.
It has a
#if SERIAL_POST
blah.. CALLSP(console_tx_al)
that's a good thing to know but a side-effect. I think the real problem is that the definitions of console_tx_al etc. should have a .globl with them and they don't. I'm going to try that.
ron