Date: Mon, 4 Mar 2002 12:34:46 -0500 From: Jake Burkholder <jake@locore.ca> To: Jesper Skriver <jesper@FreeBSD.ORG> Cc: freebsd-sparc@FreeBSD.ORG Subject: Re: freebsd/sparc64 bootable iso available Message-ID: <20020304123446.L75158@locore.ca> In-Reply-To: <20020303232153.B90067@skriver.dk>; from jesper@FreeBSD.ORG on Sun, Mar 03, 2002 at 11:21:53PM %2B0100 References: <20020303060453.D75158@locore.ca> <20020303230653.C76077@FreeBSD.org> <20020303171500.I75158@locore.ca> <20020303232153.B90067@skriver.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
Apparently, On Sun, Mar 03, 2002 at 11:21:53PM +0100, Jesper Skriver said words to the effect of; > On Sun, Mar 03, 2002 at 05:15:00PM -0500, Jake Burkholder wrote: > > Apparently, On Sun, Mar 03, 2002 at 11:06:53PM +0100, > > Jesper Skriver said words to the effect of; > > > > > On Sun, Mar 03, 2002 at 06:04:53AM -0500, Jake Burkholder wrote: > > > > > > > > I've put up a bootable ISO image, which can be used to install a system > > > > relatively easily. I've tested it on an ultra 10 with the onboard ide > > > > and an ultra 60 with sym scsi, should also work on ultra 5s and blade 100s > > > > at least. The cd boots multi user and has a binary distribution tarball > > > > on it; you can use system on the cd to install on disk. The root password > > > > is abc123. > > > > > > > > Use the tools on the cd to label, install the bootblock, and newfs your > > > > drive, then mount your partitions and untar the tarball in root's home > > > > directory on them. I put up a transcript of an install on an ultra 60. > > > > When labeling your drive be sure that the partitions are on cylinder > > > > boundaries, the disklabel format requires it. Insert usual warnings > > > > about experimental software and destruction of data. > > > > > > > > http://people.freebsd.org/~jake/sparc64 > > > > > > Nice, it works perfect on my 400 MHz Ultra5 > > > > > > But it panics instantly when trying to start ntpd > > > > It should drop into ddb in that case, right? Can you provide a traceback? > > Yes, that should have been in that email, but I by mistake sent > it too quickly, I was trying to figure out how to load the > symbol file. > > bar# ntpd > bar# Mar 3 23:20:17 bar ntpd[205]: ntpd 4.1.0-a Sat Mar 2 15:49:31 EST 2002 (1) > Debugger("trapsig") > Stopped at Debugger+0x1c: ta %xcc, 1 > db> trace > trap() at trap+0x4d8 > -- fast data access mmu miss tar=0x2f14 -- > userland() at 0x40510d5c > user trace: trap %o7 (leaf return) is 0x10e5bc > pc 0x40510d5c, sp 0x7fdffffee01 > pc 0x10d388, sp 0x7fdffffeed1 > pc 0x11c858, sp 0x7fdffffefa1 > pc 0x11c520, sp 0x7fdfffff181 > pc 0x1022f0, sp 0x7fdfffff241 > pc 0x4023bda4, sp 0x7fdfffff301 > done This is actually ntpd that crashed, not the kernel. That kernel has code that I forgot to remove which drops into the debugger when a process gets a fatal signal; push c to continue. You can try using the objdump binary in the distribution to disassemble ntpd and find what these addresses refer to, might give some clues. > > An other thing, is this build from the code present in -current, or it > it p4 I need to track ? This is built from p4. You can get the sparc64 branch from cvsup10 using this supfile: http://people.freebsd.org/~tmm/sparc64-supfile. The patches to make ata work are the biggest difference between that and cvs. Jake 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?20020304123446.L75158>