From owner-freebsd-bugs Sun Jan 6 17:20: 6 2002 Delivered-To: freebsd-bugs@hub.freebsd.org Received: from freefall.freebsd.org (freefall.FreeBSD.org [216.136.204.21]) by hub.freebsd.org (Postfix) with ESMTP id ECE8D37B41D for ; Sun, 6 Jan 2002 17:20:01 -0800 (PST) Received: (from gnats@localhost) by freefall.freebsd.org (8.11.6/8.11.6) id g071K1v69814; Sun, 6 Jan 2002 17:20:01 -0800 (PST) (envelope-from gnats) Date: Sun, 6 Jan 2002 17:20:01 -0800 (PST) Message-Id: <200201070120.g071K1v69814@freefall.freebsd.org> To: freebsd-bugs@FreeBSD.org Cc: From: "Robin Breathe" Subject: Re: kern/33397: "panic: unknown/reserved trap" with moderate network usage and recent -STABLE Reply-To: "Robin Breathe" Sender: owner-freebsd-bugs@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org The following reply was made to PR kern/33397; it has been noted by GNATS. From: "Robin Breathe" To: Cc: Subject: Re: kern/33397: "panic: unknown/reserved trap" with moderate network usage and recent -STABLE Date: Mon, 7 Jan 2002 01:19:25 -0000 I think the problem is related to some recent change in the SMP code. Having been fiddling with this for the last week, I've found that a more common lock is of the following form: Fatal trap 12: page fault while in kernel mode mp_lock = 01000002; cpuid = 1; lapic.id = 01000000 fault virtual address = 0x30 fault code = supervisor read, page not present instruction pointer = 0x8:0xc01f0fe5 # nm -n /kernel | grep c01f0f c01f0f9c t acquire_lock c01f0ffc t free_lock stack pointer = 0x10:0xff80fe28 frame pointer = 0x10:0xff80fe30 code segment = base 0x0, limit 0xfffff, type 0x1b = DPL 0, pres 1, def32 1, gran 1 processor eflags = interrupt enabled, resume, IOPL = 0 current process = Idle interrupt mask = bio <- SMP: XXX trap number = 12 panic: page fault mp_lock = 01000002; cpuid = 1; lapic.id = 01000000 boot() called on cpu#1 Uptime: 23m13s Automatic reboot in 15 seconds - press a key on the console to abort Am I right in thinking this is something SMP-related? Kernel conf is as above, as is dmesg, though this is now going from a cvsup at roughly 9pm GMT 06/01/2002. I've been seeing a lot of locks over the past few days pointing to this same section of the kernel. Another particularly mysterious symptom is that the network would lock-up, unless I kept a ping -f going on the serial console. On an SMP kernel, this would cause a kernel panic within seconds, but on a UP kernel it was stable (so long as I kept the ping flood going)... after cvsup'ing this evening, that seems to have cleared up, but the kernel panics continue on overdrive. Either ping-f, or a fast samba copy seem to spark it. I'd really love to get some input on this: is it a kernel bug, is it hardware(!?), is it OBL? Robin Breathe To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message