From owner-freebsd-questions Tue May 21 02:06:41 1996 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id CAA11021 for questions-outgoing; Tue, 21 May 1996 02:06:41 -0700 (PDT) Received: from shiva.jussieu.fr (shiva.jussieu.fr [134.157.0.129]) by freefall.freebsd.org (8.7.3/8.7.3) with ESMTP id CAA10998; Tue, 21 May 1996 02:06:35 -0700 (PDT) From: af@biomath.jussieu.fr Received: from mekong.biomath.jussieu.fr (mekong.biomath.jussieu.fr [134.157.72.87]) by shiva.jussieu.fr (8.7.5/jtpda-5.2) with SMTP id LAA18055 ; Tue, 21 May 1996 11:06:28 +0200 (METDST) Received: from garfield.biomath.jussieu.fr (garfield) by mekong.biomath.jussieu.fr (5.67b/jn930126+af960507(mailhost)) at Tue, 21 May 1996 11:06:25 +0100 Received: from (af@localhost) by garfield.biomath.jussieu.fr (8.6.12/jtpda-5.2) id LAA17953 ; Tue, 21 May 1996 11:06:24 +0100 Message-Id: <199605211006.LAA17953@garfield.biomath.jussieu.fr> Subject: Endeavour 133Mhz+Adaptec 2940=cc1 got signal 11 To: questions@FreeBSD.org, hardware@FreeBSD.org Date: Tue, 21 May 1996 11:06:24 +0100 (GMT-1) X-Mailer: ELM [version 2.4 PL24] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-questions@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk Hello, I've installed FreeBSD Release 2.1 on the new 133Mhz Pentium (Intel Endeavour m/b, Triton chipset) that will replace my old Sparc II which needs retirement. It has 64Mb of main (70ns) memory, 512Kb cache, a 2Gb IDE boot disk, an Adaptec 2940 PCI SCSI adapter which connects a DEC RZ24 200Mb disk only for now. I chose FreeBSD after testing Linux, because of its clear superiority in networking, especially NFS. And then... the dreaded "cc1 got signal 11" symptom :-( It happens only during large compiles (gcc 2.7.2 typically) when sources are on the SCSI disk *or* during kernel compiles (on the IDE disk) while there are I/O on the SCSI bus. I've clocked down the motherboard (to 66Mhz I guess, the AMIBIOS setup doesn't say): same. I've completely invalidated the secondary cache: same. I've swapped the AHA 2940 controller: same. Invalidating PCI burst and playing with PCI latency clock counts didn't change anything either. The hardware vendor claims that the machine has run a three-day burn-in cycle with no problems. When I mentioned that I had read articles saying that 60ns memory was required, he replied that 70ns memory is OK on that kind of motherboard because the BIOS automatically sets wait cycles (???) If it were a memory access time problem, though, I would have expected the downclocking to make it disappear. So what ? my gut feeling is that the DMA cycles from the AHA2940 somehow end up corrupting memory... it's most probably a motherboard problem. Is there any way to configure the AHA2940 driver in PIO mode only ? From checking the sources, I'm afraid there isn't... Or any way to "throttle down" the DMA ? (if that means anything) Any hints please ? the vendor seems a bit reluctant to swapping the motherboard, and even more reluctant to changing memory SIMMs (he says 60ns memory is virtually impossible to find nowadays). Thanks in advance, _Alain_ Excerpts from /var/log/messages: =============================== May 20 13:05:02 garfield /kernel: FreeBSD 2.1.0-RELEASE #0: Fri May 3 18:20:59 MET DST 1996 May 20 13:05:03 garfield /kernel: root@garfield.biomath.jussieu.fr:/usr/src/ sys/compile/GARFIELD May 20 13:05:03 garfield /kernel: CPU: 133-MHz Pentium 735\90 or 815\100 (Pentiu m-class CPU) May 20 13:05:03 garfield /kernel: Origin = "GenuineIntel" Id = 0x52b Steppin g=11 May 20 13:05:03 garfield /kernel: Features=0x1bf May 20 13:05:03 garfield /kernel: real memory = 67108864 (65536K bytes) May 20 13:05:03 garfield /kernel: avail memory = 63057920 (61580K bytes) [...] May 20 13:05:04 garfield /kernel: wdc0 at 0x1f0-0x1f7 irq 14 on isa May 20 13:05:04 garfield /kernel: wdc0: unit 0 (wd0): May 20 13:05:05 garfield /kernel: wd0: 1916MB (3924144 sectors), 3893 cyls, 16 h eads, 63 S/T, 512 B/S May 20 13:05:05 garfield /kernel: 1 3C5x9 board(s) on ISA found at 0x300 May 20 13:05:05 garfield /kernel: ep0 at 0x300-0x30f irq 10 on isa May 20 13:05:05 garfield /kernel: ep0: aui/bnc/utp[*AUI*] address 00:a0:24:7a:5a :e8 irq 10 [...] May 20 13:05:05 garfield /kernel: sb0 at 0x220 irq 5 drq 1 on isa May 20 13:05:05 garfield /kernel: sb0: May 20 13:05:05 garfield /kernel: sbxvi0 at 0x0 drq 5 on isa May 20 13:05:05 garfield /kernel: sbxvo0: May 20 13:05:05 garfield /kernel: sbmidi0 at 0x330 on isa May 20 13:05:05 garfield /kernel: May 20 13:05:05 garfield /kernel: opl0 at 0x388 on isa May 20 13:05:05 garfield /kernel: opl0: May 20 13:05:05 garfield /kernel: Probing for devices on the PCI bus: May 20 13:05:06 garfield /kernel: chip0 rev 1 on pci0:0 May 20 13:05:06 garfield /kernel: chip1 rev 2 on pci0:7 May 20 13:05:06 garfield /kernel: ahc0 rev 0 in t a irq 9 on pci0:14 May 20 13:05:06 garfield /kernel: ahc0: 2940 Single Channel, SCSI Id=7, aic7870, 16 SCBs May 20 13:05:06 garfield /kernel: ahc0 waiting for scsi devices to settle May 20 13:05:06 garfield /kernel: (ahc0:0:0): "DEC RZ24 (C) DEC 4041" type 0 fixed SCSI 1 May 20 13:05:07 garfield /kernel: sd0(ahc0:0:0): Direct-Access 200MB (409792 512 byte sectors) -- Alain FAUCONNET Ingenieur systeme - System Manager AP-HP/SIM Public Health 91 bld de l'Hopital 75013 PARIS FRANCE Medical Computing Research Labs Mail: af@biomath.jussieu.fr Tel: (+33) 1-40-77-96-19 Fax: (+33) 1-45-86-80-68 I've RTFMed. It says: "Refer to your system administrator" But... I *am* the system administrator :-]