From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Jacobowitz To: Andrew Cagney Cc: Eli Zaretskii , gdb-patches@sources.redhat.com Subject: Re: gdbserver/{,,}.c?; Was: [rfa] gdbserver overhaul Date: Thu, 18 Oct 2001 12:08:00 -0000 Message-id: <20011018150810.A21022@nevyn.them.org> References: <3BCEFBF4.3000508@cygnus.com> <7999-Thu18Oct2001202904+0200-eliz@is.elta.co.il> <3BCF2661.3020304@cygnus.com> X-SW-Source: 2001-10/msg00257.html On Thu, Oct 18, 2001 at 02:58:41PM -0400, Andrew Cagney wrote: > >>Date: Thu, 18 Oct 2001 11:57:40 -0400 > >>From: Andrew Cagney > > > >>> > > > >>>> or if it did it couldn't use any of these low-* files right? > > > >>> > >>> > >>> Sorry, I don't understand this. How is the DJGPP port different from > >>> any other port in this respect? > > > >> > >>I don't think DJGPP's low/nat interface looks like UNIX's ptrace() or > >>/procfs so much of the existing low-* code wouldn't be used. > > > > > >That's true, but the low-* files are supposed to be compiled on the > >target system, not on the host, right? If so, all but one of these > >low-* files is irrelevant for other platforms as well, right? > > Yes. My understanding of DanielJ's cleanup is that it better structures > (separate os/cpu specific and generic files) the unix low code. Yes, that's right. Well, more specifically: - I create a file for each CPU supported, ideally - there may be multiple variants of this for different OS targets, and I haven't checked any besides Linux, but I don't think there should be... - I create a new file, broken out of the monstrous linux target file, for each supported linux target. There is little common Unix support code, but there looks as if there should be more. After some more factoring of architecture-specific bits, perhaps there will be. -- Daniel Jacobowitz Carnegie Mellon University MontaVista Software Debian GNU/Linux Developer