This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: retargeting gcc...
- To: "wayne c. deprince jr." <wcd2 at lehigh dot edu>
- Subject: Re: retargeting gcc...
- From: Jeffrey A Law <law at cygnus dot com>
- Date: Sat, 21 Nov 1998 17:01:20 -0700
- cc: egcs at cygnus dot com
- Reply-To: law at cygnus dot com
In message <3656349D.E4BE0197@lehigh.edu>you write:
> hey,
> i am planning on creating a new target architecture for gcc, which is
> an architecture im developing. i really dont know that much about gcc,
> and would appreciate any help on what this takes.. i have been browsing
> the source and i figured out that at least a machine definition file
> must be written in RTL, along with a bunch of support files.. if
> somebody could at summarize the steps, share experiences, or at least
> point me in the write direction, i would really appreciate it. also,
> there is a book called using and porting gnu CC.. is it worth it, or is
> it no good... thanks and many blessings a million times in advance,
> wayne
>
> im not on the list yet, so please reply to me.. thanks
The bulk of a retarget effort is building the machine description and header
files. The other files are usually trivial.
Yes, using and porting gnu CC is worth the time to print and read, especially
if you are going to retarget gcc for a new architecture.
In general, the easiest way I've found to retarget gcc is to find a machine
with similar characteristics to the new target that is already supported by
gcc to use as a starting point for the machine description and header file.
jeff