From owner-freebsd-current@FreeBSD.ORG Mon Sep 20 17:55:07 2004 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 45E4116A4CE; Mon, 20 Sep 2004 17:55:07 +0000 (GMT) Received: from exchange.sandvine.com (sandvine.com [199.243.201.138]) by mx1.FreeBSD.org (Postfix) with ESMTP id D155A43D45; Mon, 20 Sep 2004 17:55:06 +0000 (GMT) (envelope-from ahoff@sandvine.com) X-MimeOLE: Produced By Microsoft Exchange V6.0.6556.0 content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Date: Mon, 20 Sep 2004 13:55:06 -0400 Message-ID: X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: panic: pmap_enter: attempted pmap_enter on 4MB page Thread-Index: AcSfOfhFjxClP4FQTBWlOdmXrjhYlgAADKPg From: "Alex Hoff" To: cc: alc@freebsd.org cc: Gerrit Nagelhout cc: tegge@FreeBSD.org Subject: RESEND: panic: pmap_enter: attempted pmap_enter on 4MB page X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 20 Sep 2004 17:55:07 -0000 Sorry, it sent by accident. Here is the complete email: > -----Original Message----- > From: Alex Hoff =20 > Sent: Monday, September 20, 2004 13:48 > To: 'current@freebsd.org' > Cc: 'alc@freebsd.org'; 'tegge@FreeBSD.org'; Gerrit Nagelhout > Subject: panic: pmap_enter: attempted pmap_enter on 4MB page >=20 > Hello,=20 >=20 > I've talked to a few people about this crash already, but I might as = well get a discussion going. Running 5.3 beta 4 (as of Sunday) with = Postgress 7.4.2 running on it under heavy load will run into "panic: = pmap_enter: attempted pmap_enter on 4MB page". The machine we are = running on is a dual xeon 2.8 ghz w/ HyperThreading off, 4 gigs of ram = and a 980G RAID5 setup. Using a small database with a total of 100 = clients, 80 doing updates and 20 doing queries via tcl (db + clients on = the same machine), it runs until fills up 4gigs of ram, then 4 gigs of = swap, then we kill it cleanly rc.d script stop). If I repeat this = process X number of times (it varies), the last time will bring down the = kernel as postgress is exiting. So far (from 2 dumps) the curpte seems = invalid Crash 1: swap_pager_getswapspace(16): failed swap_pager_getswapspace(16): failed swap_pager_getswapspace(16): failed <3>pid 3327 (postgres), uid 70, was killed: out of swap space pmap_enter: va 0x29853000, pte 0xbfca614c, origpte 0x1999 panic: pmap_enter: attempted pmap_enter on 4MB page and Crash 2: swap_pager_getswapspace(3): failed swap_pager_getswapspace(3): failed <3>pid 9822 (postgres), uid 70, was killed: out of swap space <3>pid 9816 (postgres), uid 70, was killed: out of swap space pmap_enter: va 0x2c846000, pte 0xbfcb2118, origpte 0x29f1 panic: pmap_enter: attempted pmap_enter on 4MB page Anyone have any ideas/input? Here is the backtrace if it helps: (kgdb) bt #0 doadump () at pcpu.h:159 #1 0xc065623e in boot (howto=3D260) at /d2/ahoff/cvs/src/sys/kern/kern_shutdown.c:397 #2 0xc06565ec in poweroff_wait (junk=3D0xc08ae384, howto=3D696594432) at /d2/ahoff/cvs/src/sys/kern/kern_shutdown.c:553 #3 0xc082d3cd in pmap_enter (pmap=3D0xc5c9ac78, va=3D3230327684, = m=3D0xc37bb130, prot=3D7 '\a', wired=3D0) at = /d2/ahoff/cvs/src/sys/i386/i386/pmap.c:1959 #4 0xc07db6b9 in vm_fault (map=3D0xc5c9abb8, vaddr=3D696594432, fault_type=3D1 '\001', fault_flags=3D0) at /d2/ahoff/cvs/src/sys/vm/vm_fault.c:871 #5 0xc083188f in trap_pfault (frame=3D0xf01c1d48, usermode=3D1, = eva=3D696595024) at /d2/ahoff/cvs/src/sys/i386/i386/trap.c:704 #6 0xc083143d in trap (frame=3D {tf_fs =3D 47, tf_es =3D 47, tf_ds =3D -1078001617, tf_edi =3D = 156606464, tf_esi =3D 696595024, tf_ebp =3D -1077945656, tf_isp =3D -266592908, tf_ebx =3D = 137957328, tf_ed x =3D 156606432, tf_ecx =3D 23, tf_eax =3D 100, tf_trapno =3D 12, tf_err = =3D 4, tf_eip =3D 1 34671476, tf_cs =3D 31, tf_eflags =3D 66118, tf_esp =3D -1077945680, = tf_ss =3D 47}) at /d2/ahoff/cvs/src/sys/i386/i386/trap.c:316 #7 0x0806ec74 in ?? ()