Date: Tue, 24 Nov 1998 15:11:20 +0000 From: "Paolo Di Francesco" <paipai@tin.it> To: freebsd-sparc@FreeBSD.ORG Subject: Re: [Code writers ONLY] Developer Kit Message-ID: <19981124140921.JPEV22050.fep03-svc@winworkstation> In-Reply-To: <Pine.BSF.4.05.9811232337160.10312-100000@bright.fx.genx.net> References: <Pine.BSF.4.05.9811231847190.1845-100000@ender.sf.scient.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> Is it unrealistic to try to build it under solaris? I'm a bit > confused i had three different paths i thought about taking: Teorically, no. > a) compile under solaris (i have a spare disk in the same box and > it would make it very easy if this is a feasable option) > > b) cross compile under freebsd, hard to test but i could use > netbooting > > c) netbsd (this one is kinda not going to work as they don't have > bootblocks and they have been quiet about developemnt) > > If i choose option 'a' do i need to make a cross compiler anyway > being that solaris 2.6 is 32 bit? if so, i think i might as well > do option 'b'. > > C seems a bit farfetched as i'd be porting netbsd instead of freebsd > :) > > *cringing* > > d) under ultrapenguine? > > Only problem it seems is that for all the hype ultra penguine got > it supports almost no hardware, not even serial ports, just gx > graphics ( i have a creator ) and some odd ethernet chipset (i have > hme, they don't) > Ok, we are talking of Kernel, right? Kernel is an app, a particoular one, but it is an app. How to write a basic kernel or something similar to a kernel? The answer is: mainly using assembly code, then C code could be usefull. The real difference is you have not to link the code. You must use the same compiler you use for your applications, but with different flags. Nothing more nothing less. On intel, to boot something from a floppy you must put the object code "raw" on the floppy. On Sparc I think something similar. So you have to read a bootable disk (for example the Ultralinux) and then you'll know if it is pure assembly traduced in 010101 code or something like this. About paths. We can use A or D. Solaris is stable enough if we want to do serious works. D is good because we can test many things. I hope to have a bootable UltraLinux to see how it works. For example we can start reading the floppy image byte-after-byte and trying to reverse decode it. If the fist 100 byte (for example) are Sparc istructions we can assume that it works as I think. More infos, docs, welcome! About path B. This could be interesting for me, because I have no Ultra. So I could use crosscompiling on my Intel. But we have to compare the object codes what we have from the solaris gcc and what we have from the gcc/Intel/Crosscomp. Actually I think this could be and interesting issue.... I have downloaded gcc2.8. Today I'll see where it is different from 2.7.2 and if it's better than 2.7.2 for our purpose. 8) P.S. Tonight I'll download the Ultralinux boot disk (if I will find one) Ciao Ciao Paolo Di Francesco _ ->B<- All Recycled Bytes Message ... ~ 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?19981124140921.JPEV22050.fep03-svc>