dburr@.berkeley.edu (Donald Burr) (03/01/91)
I would like to get GCC running on my Callan Unistar 300 box. This is a
SysV R2 (I think) box that uses a MC68010 as the processor. Is this
possible, and if so, what/how much hacking would I need to do in the GCC
code, and where?
Please E-MAIL all replies to: dburr@ocf.berkeley.edu.
Many thanks!
______________________________________________________________________________
Donald Burr; Univ of California, Berkeley | America Online: DonaldBurr
INTERNET: dburr@ocf.Berkeley.EDU |_CompuServe:_72540,3071____________
or: 72540.3071@compuserve.COM | "Send flames to /dev/null."rfg@NCD.COM (Ron Guilmette) (03/03/91)
In article <1991Mar1.072541.6959@agate.berkeley.edu> dburr@.berkeley.edu (Donald Burr) writes: > >I would like to get GCC running on my Callan Unistar 300 box. This is a >SysV R2 (I think) box that uses a MC68010 as the processor. Is this >possible, and if so, what/how much hacking would I need to do in the GCC >code, and where? If you already have a C compiler on your Callan, and if it is not too brain dammaged, you should be able to get GCC up pretty easily. You will probably want to make up your own xm-callan.h and tm-callan.h files in order to do this. Fortunately, you will not have to start from scratch. Look at the tm-sun2.h file. It is correct for a 68010 processor so you should be able to make up a tm-callan.h file by just doing: tm-callan.h: ---------------------------------------------------------------- #include "tm-sun2.h" /* ... overriding definitions go in here ... */ ---------------------------------------------------------------- You may have to override a few of the definitions that get included from the tm-sun2.h file, but it should not be too many. One thing that I *know* you will have to override is the following: #undef DBX_DEBUGGING_INFO #define SDB_DEBUGGING_INFO That's because your target system (i.e. Callan) uses System V. For your xm-callan.h file you probably want to start from one of the xm-*.h files that are used for System V hosts. Any one of the xm-*.h files that define the symbol `USG' will probably be a good starting point for your new xm-callan.h file. If you just do: cd config grep USG xm-*.h You will see which of the existing xm-*.h files are already used with System V based systems. (Fortuanately, they are all almost identical and they are all pretty short, so if you look at the contents of a few of these files it should be obvious what your xm-callan.h file should look like. Once you get these two configuration files set up properly, you will need to complete the "configuration" step by doing: ln config/xm-callan.h config.h ln config/tm-callan.h tm.h ln config/m68k.md md ln config/out-m68k.c aux-output.c Then just type `make' and keep your fingers crossed. Summary: You should have to `hack' very little (if at all). -- // Ron Guilmette - C++ Entomologist // Internet: rfg@ncd.com uucp: ...uunet!lupine!rfg // New motto: If it ain't broke, try using a bigger hammer.
jtc@motcad.portal.com (J.T. Conklin) (03/04/91)
In article <1991Mar1.072541.6959@agate.berkeley.edu> dburr@.berkeley.edu (Donald Burr) writes: >I would like to get GCC running on my Callan Unistar 300 box. This is a >SysV R2 (I think) box that uses a MC68010 as the processor. Is this >possible, and if so, what/how much hacking would I need to do in the GCC >code, and where? I was able to port GCC to a old motorola 68k box with a wierd assembler syntax in about a week. The main problem seems to be having a compiler that is "smart enough" to compile the first pass of GCC. AT&T UNIX PC owners had a rough time getting gcc to compile, as the stock compiler couldn't do it --- eventually someone got it to work and binary copies were distributed. If work at it, you should be able to port gcc to your callan. If you do get it to work, send your fixes into the FSF. Not only does it make it easier for you to port gcc when the next version comes out, but also you get something widely distributed with your name on it. :-) --jtc -- J.T. Conklin jtc@motcad.portal.com, ...!portal!motcad!jtc