From owner-freebsd-current Sun Oct 18 09:55:12 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id JAA12612 for freebsd-current-outgoing; Sun, 18 Oct 1998 09:55:12 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from silverback.gorilla.net (silverback.gorilla.net [208.128.8.1]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id JAA12600 for ; Sun, 18 Oct 1998 09:55:07 -0700 (PDT) (envelope-from tom@gorilla.net) Received: from peeper.TOJ.org [208.143.84.62] by silverback.gorilla.net with ESMTP (SMTPD32-4.07) id AC4214F00CE; Sun, 18 Oct 1998 11:50:10 CDT Received: (from tom@localhost) by peeper.TOJ.org (8.9.1/8.8.8) id LAA04691; Sun, 18 Oct 1998 11:54:39 -0500 (CDT) (envelope-from tom) Message-ID: <19981018115408.A4680@TOJ.org> Date: Sun, 18 Oct 1998 11:54:08 -0500 From: Tom Jackson To: Peter Wemm , Mike Smith Cc: Karl Pielorz , Chuck Robey , freebsd-current@FreeBSD.ORG Subject: Re: panic References: <199810180538.WAA09998@dingo.cdrom.com> <199810180655.OAA17728@spinner.netplex.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.93.2i In-Reply-To: <199810180655.OAA17728@spinner.netplex.com.au>; from Peter Wemm on Sun, Oct 18, 1998 at 02:55:56PM +0800 Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Peter, With the patch, I get the printf MATH_EMULATE activated. All the faults are the same. I missed one: recording kernel -c changes additional daemons: syslogdOct 18 11:15:20 slim /kernel.math: pid 10 (fsck), uid 0:exited on signal 10 When I am finally running, top faults with signal 10, ps functions okay. I also noticed a lockup of the boot loader at the loader prompt for the first time. Had to use reset to free it. On Sun, Oct 18, 1998 at 02:55:56PM +0800, Peter Wemm wrote: > Mike Smith wrote: > > > *Bingo* no ahh shit! That fixed me here. > > > > > > I'm began running gpl_math_emulate so that the rc5des client would run on > > > the Thinkpad. For your amusement, here's what I get if I change back to the > > > straight math_emulate: > > > > > > -normal probes and swapon- > > > /dev/wd0s1a: clean, 10064 [pid 10 (fsck),uid 0: exited on signal 10] > > > Bus error > > > > Actually, that almost looks like no math emulator at all; there's an FP > > calculation performed to print those stats, but it's just simple > > division and the old emulator used to get that more or less right. > > Hmm.. I don't have any machines that don't have a FPU.. > > However, in trap.c, if there was no math emulator loaded, I'd have > expected a SIGFPE, not a SIGBUS: > if (!pmath_emulate) { > i = SIGFPE; > ucode = FPE_FPU_NP_TRAP; > break; > } > > Just to satisfy my curiosity, I'd like somebody to try this: > Index: math_emulate.c > =================================================================== > RCS file: /home/ncvs/src/sys/i386/i386/math_emulate.c,v > retrieving revision 1.28 > diff -c -3 -r1.28 math_emulate.c > *** math_emulate.c 1998/10/16 03:54:59 1.28 > --- math_emulate.c 1998/10/18 06:45:32 > *************** > *** 1578,1585 **** > case MOD_LOAD: > if (pmath_emulate) > printf("Another Math emulator already present\n"); > ! else > pmath_emulate = math_emulate; > break; > case MOD_UNLOAD: > if (pmath_emulate != math_emulate) { > --- 1578,1587 ---- > case MOD_LOAD: > if (pmath_emulate) > printf("Another Math emulator already present\n"); > ! else { > pmath_emulate = math_emulate; > + printf("MATH_EMULATE activated\n"); > + } > break; > case MOD_UNLOAD: > if (pmath_emulate != math_emulate) { > > And make sure that it's actually being initialized... > > Cheers, > -Peter > > > -- Tom To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message