Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!julius.cs.uiuc.edu!apple!vsi1!daver!bungi.com!news From: jvh@spiff.hut.fi (Johannes Helander) Newsgroups: comp.sys.nsc.32k Subject: GNU cross development tools. Message-ID: <9012130436.AA01452@spiff.hut.fi> Date: 13 Dec 90 04:36:27 GMT Sender: news@daver.bungi.com Reply-To: jvh@sauna.hut.fi Organization: Helsinki University of Technology, CS Lab. Lines: 38 Approved: news@daver.bungi.com Gary Jennejohn writes: > I'm in the process of implementing the GDB remote debugging stuff in the > debugger. It works fairly well, I can set breakpoints, single step, > disassemble, look at stack frames, continue, etc. > I took the path of getting gcc, gas and gld to work so that I can use > the dbx symbol table stuff. This minimized the amount of hacking I had > to do to gdb. > If anyone else is interested i can provide diffs at a later date (both > to gdb and the debugger). This is all based on gdb3.5. Jordan and I both > plan to use gdb in the future. This sounds very useful indeed. Diffs would certainly be appreciated. In fact I had plans on implementing the remote gdb myself, but haven't got to it yet, and now luckily do not need to. We now have a working gcc 1.37.94/95 + gas 1.37 + gld, ar, nm, strip + Bruce's disassembler working as cross compiling tools on a 386 box running Mach 2.5 / 3.0. They are mostly a combination of various patches (including those for Mach). Some parts (like print_operand_address) have been completely rewritten. The programs are probably not suited for Minix and might not work on a big endian machine. If somebody is, however, interested we could produce some clean diffs. We also built a number display. It uses the eight led outputs and +5V to drive 10 hexdigits with dots and some additional leds. The chips used on it are not very practical but if somebody wants to build one or redesign it, we have schemes available + assembly routines to drive them. I have one question regarding the pc532 board: Is the debug (/DBG) pin of the CPU connected _directly_ to the power layer or is it adequate to just cut the trace leading from it to be able to connect a debug switch? Nobody here seemed to have the layouts in printable form. Johannes