Date: Thu, 26 Nov 1998 15:07:08 -0500 From: "Gary Palmer" <gpalmer@FreeBSD.ORG> To: Erik Cameron <ecameron@bsd.uchicago.edu> Cc: Eivind Eklund <eivind@yes.no>, eddy@ISI.EDU, Paolo Di Francesco <paipai@tin.it>, freebsd-sparc@FreeBSD.ORG Subject: Re: What we need Message-ID: <28778.912110828@gjp.erols.com> In-Reply-To: Your message of "Thu, 26 Nov 1998 13:14:54 CST." <19981126131454.A5740@bsd.uchicago.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
Erik Cameron wrote in message ID <19981126131454.A5740@bsd.uchicago.edu>: > > The real problem I've found with this is is the (relatively) large > amount of assembly code involved in building libc; libc being the > first thing to work on, IMHO. >From memory, half the assembly stuff in libc is optimized versions of speed-critical components (ntohl,bcopy, etc). I seem to remember their being C versions as well. This isn't true for everything (such as the syscall interface), but it cuts down on the ammount of work that needs to be done. A bunch of the .S files you see being compiled/assembled when you make libc are syscall veneers that are machine generated from a single copy of the file. So, yes, assembly is required to build libc, but not as much as you might think. And to be perfectly honest, the biggest hurdle is going to be the kernel, not libc, and I think thats where your focus should be. You can boot and debug a kernel without ever needing a libc. Gary -- Gary Palmer FreeBSD Core Team Member FreeBSD: Turning PC's into workstations. See http://www.FreeBSD.ORG/ for info To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-sparc" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?28778.912110828>