From mboxrd@z Thu Jan  1 00:00:00 1970
Message-ID: <1e2cac07e8e7f1908cef95c5b6da0613@mail.gmx.net>
To: 9fans@cse.psu.edu
From: "Sascha Retzki" <sretzki@gmx.de>
Date: Sat, 26 Aug 2006 16:40:00 +0200
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Subject: [9fans] sparc64 port
Topicbox-Message-UUID: a78c3602-ead1-11e9-9d60-3106f5b1d025


Hi friends,


I tried out Newshams sparc64 ports today (the latest tarball), and I experienced two and a half problems.

First of, there is a file missing in /sys/src/ape/lib/9/sparc64: getcallerpc.s.
As it is a oneliner which just returns 'something' ( I failed to parse it, honestly my asm is bad ), I just copied over the sparc/getcallerpc.s. (Hit me with your cluesticks later)

That again made ghostscript (and only ghostscript) not compile, syntax errors in gdevddrw.c arround an int64_t c; .. anyway. I won't use gs on sparci right now, anyway ;)


The other major problem is that the kernel does not build. I don't know how it is supposed to work, anyway. main.c modulglobally declares FPsave initfp which is used in main.c/miscinit(): savefpregs(&initfp); but also in trap.c/fpenable(): restorefpregs(&initfp);, which is called by trapinit() (which makes me think trap.c is a 'young' module which has never been tested).


Furthermore, I had to copy pc/pcf to sparc64/sparc64f, including a lot of tweaks of course - a working default-config may have been cool :P


That's how far I got. Tim, are you still here? Any comments? Anybody looked into this more deeply? Made it work?



Mfg, Sascha