Re: more install problems

[Richard Lyons <richard@the-place.net>]

On Saturday 02 August 2003 11:19, Andrew McGuinness wrote:
[...]
> I really think the 8139too driver in 2.4.18 ought to be able to
> work this.
>
> The problem must be either a general PCMIA issue, or a problem with
> the 8139too driver being compiled-in instead of being a module.
[...]
I have given up trying with the 2.4.18-bf2.4 and changed to the kernel 
2.4.18-686, and would like to see if the 8139too driver can now be 
compiled.  The latest file (from the Realtek site) has this Makefile:

# Makefile for a basic kernel module

#-----------------------------------------------------------------
# This flag specifies access type to operation register on PCI \ 
ethernet chips.
# If you apply 8139/8139CP ethernet controller via cardbus interface \ 
on linux distribution Red Hat 7.3
# please turn on this macro.
#------------------------------------------------------------------
#RTL_MEM_ACCESS = USE_IO_OPS

#------------------------------------------------------------------
# Please specify the include path according to your kernel.
#------------------------------------------------------------------
NEW_INCLUDE_PATH = /usr/src/linux-2.4.18-3/include/

#------------------------------------------------------------------
# GNU C compiler and its flags
#------------------------------------------------------------------
CC=gcc
MODCFLAGS := -O6 -Wall -DCONFIG_KERNELD -DMODULE -D__KERNEL__ \
-DLINUX  -D${RTL_MEM_ACCESS}



all: 8139too

8139too:        8139too.c /usr/include/linux/version.h
                $(CC) $(MODCFLAGS) -I$(NEW_INCLUDE_PATH) -c \
8139too.c -o $@.o


'make' fails because of the '-D${RTL_MEM_ACCESS}'.  Can anybody tell 
me what this is intended to be.  Is it an item that needs editing 
(not according to the instructions in the readme), or is it an 
environment variable that Realtek expect to be available, or what.  
It is evidently interpreted as '-D'.

TIA

-- 
richard