From owner-freebsd-smp Tue Jul 20 20:46:14 1999 Delivered-To: freebsd-smp@freebsd.org Received: from wwdg.com (mail.wwdg.com [209.181.65.217]) by hub.freebsd.org (Postfix) with ESMTP id B1AE414D6E for ; Tue, 20 Jul 1999 20:46:10 -0700 (PDT) (envelope-from dvwd@wwdg.com) Received: (from web@localhost) by wwdg.com (8.8.5/8.8.0) id VAA24286 for freebsd-smp@freebsd.org; Tue, 20 Jul 1999 21:37:09 -0600 Date: Tue, 20 Jul 1999 21:37:09 -0600 Message-Id: <199907210337.VAA24286@wwdg.com> From: dvwd@wwdg.com To: freebsd-smp@freebsd.org Full-Name: Dave Wood Subject: mptable output hosed Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hello, I have SMP running on my SuperMicro P6DBE mother board, however when I run mptable, I only get some of the normal output. It ends with: MP Config Extended Table Entries: Extended Table HOSED! I am running FreeBSD 3.2-RELEASE. When I rebuilt the kernel, I turned on SMP, and forced memory size to 256megs. Any ideas? Dave -------- Full output of mptable: =============================================================================== MPTable, version 2.0.15 ------------------------------------------------------------------------------- MP Floating Pointer Structure: location: BIOS physical address: 0x000fb4f0 signature: '_MP_' length: 16 bytes version: 1.4 checksum: 0xdd mode: Virtual Wire ------------------------------------------------------------------------------- MP Config Table Header: physical address: 0x000f2490 signature: 'PCMP' base table length: 268 version: 1.4 checksum: 0xd7 OEM ID: 'INTEL ' Product ID: '440BX ' OEM table pointer: 0x00000000 OEM table size: 0 entry count: 25 local APIC address: 0xfee00000 extended table length: 16 extended table checksum: 234 ------------------------------------------------------------------------------- MP Config Base Table Entries: -- Processors: APIC ID Version State Family Model Step Flags 0 0x11 BSP, usable 6 5 2 0x183fbff 1 0x11 AP, usable 6 5 2 0x183fbff -- Bus: Bus ID Type 0 PCI 1 PCI 2 ISA -- I/O APICs: APIC ID Version State Address 2 0x11 usable 0xfec00000 -- I/O Ints: Type Polarity Trigger Bus ID IRQ APIC ID PIN# ExtINT conforms conforms 2 0 2 0 INT conforms conforms 2 1 2 1 INT conforms conforms 2 0 2 2 INT conforms conforms 2 3 2 3 INT conforms conforms 2 4 2 4 INT conforms conforms 2 5 2 5 INT conforms conforms 2 6 2 6 INT conforms conforms 2 7 2 7 INT active-hi edge 2 8 2 8 INT conforms conforms 2 9 2 9 INT conforms conforms 2 10 2 10 INT conforms conforms 2 12 2 12 INT conforms conforms 2 13 2 13 INT conforms conforms 2 14 2 14 INT conforms conforms 2 15 2 15 INT active-lo level 2 11 2 18 SMI conforms conforms 2 0 2 23 -- Local Ints: Type Polarity Trigger Bus ID IRQ APIC ID PIN# ExtINT conforms conforms 0 0:A 255 0 NMI conforms conforms 0 0:A 255 1 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Jul 20 21: 1:37 1999 Delivered-To: freebsd-smp@freebsd.org Received: from darkstar.cis.ohio-state.edu (darkstar.cis.ohio-state.edu [164.107.114.52]) by hub.freebsd.org (Postfix) with ESMTP id 2B52F14FA5 for ; Tue, 20 Jul 1999 21:01:35 -0700 (PDT) (envelope-from rowland@darkstar.cis.ohio-state.edu) Received: (from rowland@localhost) by darkstar.cis.ohio-state.edu (8.9.3/8.9.3) id XAA34048; Tue, 20 Jul 1999 23:58:56 -0400 (EDT) (envelope-from rowland) To: dvwd@wwdg.com Cc: freebsd-smp@FreeBSD.ORG Subject: Re: mptable output hosed References: <199907210337.VAA24286@wwdg.com> X-Face: "<|%>c@Vfv/8}+Av1z:R5sDzf!3QVer!n\,.&+&h\k1,BHAoyw8Gp10<-SqZ<*"|!U!a#xg6ls?1,Vj$m@r?uHcfB,'i:LLgtyb;~}O8v7zZThuB`X~#IE{v*"PhI]cl/>&ys(MGa%y:6~TuHFw&~|V?!9HZ_R"<}dC5D:%igP2Q6ZJex,P0M From: Shaun Rowland Date: 20 Jul 1999 23:58:55 -0400 In-Reply-To: dvwd@wwdg.com's message of "Tue, 20 Jul 1999 21:37:09 -0600" Message-ID: <87lncaslfk.fsf@darkstar.cis.ohio-state.edu> Lines: 29 X-Mailer: Gnus v5.6.45/XEmacs 21.1 - "Arches" Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org dvwd@wwdg.com writes: > Hello, > > I have SMP running on my SuperMicro P6DBE mother board, however when > I run mptable, I only get some of the normal output. It ends with: > > MP Config Extended Table Entries: > > Extended Table HOSED! > > I am running FreeBSD 3.2-RELEASE. When I rebuilt the kernel, I > turned on SMP, and forced memory size to 256megs. Any ideas? > > Dave I have had the same problem with a similar motherboard (I am not sure if it is exactly the same number... I am at work and it is at home :-). It was a SuperMicro motherboard though. I was able to change the MP version from 1.4 to 1.1 and all was well after that. Can you change this: > version: 1.4 to version 1.1 in the BIOS and see if that fixes your mptable? -- Shaun Rowland rowland@cis.ohio-state.edu http://www.cis.ohio-state.edu/~rowland/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Tue Jul 20 22:44:44 1999 Delivered-To: freebsd-smp@freebsd.org Received: from wwdg.com (mail.wwdg.com [209.181.65.217]) by hub.freebsd.org (Postfix) with ESMTP id 38BCC15473 for ; Tue, 20 Jul 1999 22:44:39 -0700 (PDT) (envelope-from dvwd@wwdg.com) Received: (from web@localhost) by wwdg.com (8.8.5/8.8.0) id XAA27120 for freebsd-smp@FreeBSD.ORG; Tue, 20 Jul 1999 23:37:47 -0600 Date: Tue, 20 Jul 1999 23:37:47 -0600 Message-Id: <199907210537.XAA27120@wwdg.com> From: dvwd@wwdg.com To: freebsd-smp@FreeBSD.ORG Full-Name: Dave Wood Subject: Re: mptable output hosed Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > Hello, > > > > I have SMP running on my SuperMicro P6DBE mother board, however when > > I run mptable, I only get some of the normal output. It ends with: > > > > MP Config Extended Table Entries: > > > > Extended Table HOSED! > > > > I am running FreeBSD 3.2-RELEASE. When I rebuilt the kernel, I > > turned on SMP, and forced memory size to 256megs. Any ideas? > > > > Dave > > I have had the same problem with a similar motherboard (I am not sure > if it is exactly the same number... I am at work and it is at home > :-). It was a SuperMicro motherboard though. I was able to change > the MP version from 1.4 to 1.1 and all was well after that. Can you > change this: > > > version: 1.4 > > to version 1.1 in the BIOS and see if that fixes your mptable? I'll give that a try. The system is running as a server and is co-located (where they've got a better connection :), but will try that next time I'm there. Any idea what I might loose going from 1.4 to 1.1? To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Wed Jul 21 5:40:23 1999 Delivered-To: freebsd-smp@freebsd.org Received: from dhcp9545042.columbus.rr.com (dhcp9545042.columbus.rr.com [24.95.45.42]) by hub.freebsd.org (Postfix) with ESMTP id 56603154A7 for ; Wed, 21 Jul 1999 05:40:08 -0700 (PDT) (envelope-from rowland@dhcp9545042.columbus.rr.com) Received: (from rowland@localhost) by dhcp9545042.columbus.rr.com (8.9.3/8.9.3) id IAA54265; Wed, 21 Jul 1999 08:39:45 -0400 (EDT) (envelope-from rowland) To: dvwd@wwdg.com Cc: freebsd-smp@FreeBSD.ORG Subject: Re: mptable output hosed References: <199907210537.XAA27120@wwdg.com> X-Face: "<|%>c@Vfv/8}+Av1z:R5sDzf!3QVer!n\,.&+&h\k1,BHAoyw8Gp10<-SqZ<*"|!U!a#xg6ls?1,Vj$m@r?uHcfB,'i:LLgtyb;~}O8v7zZThuB`X~#IE{v*"PhI]cl/>&ys(MGa%y:6~TuHFw&~|V?!9HZ_R"<}dC5D:%igP2Q6ZJex,P0M From: Shaun Rowland Date: 21 Jul 1999 08:39:45 -0400 In-Reply-To: dvwd@wwdg.com's message of "Tue, 20 Jul 1999 23:37:47 -0600" Message-ID: <87r9m2tbvy.fsf@dhcp9545042.columbus.rr.com> Lines: 8 X-Mailer: Gnus v5.6.45/XEmacs 21.1 - "Arches" Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org dvwd@wwdg.com writes: > Any idea what I might loose going from 1.4 to 1.1? I am not 100% certain, but I doubt you will actually loose anything. -- Shaun Rowland rowland@cis.ohio-state.edu http://www.cis.ohio-state.edu/~rowland/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 5:33:57 1999 Delivered-To: freebsd-smp@freebsd.org Received: from beast.toad.net (beast.toad.net [205.197.182.100]) by hub.freebsd.org (Postfix) with ESMTP id 79F9C14D81 for ; Thu, 22 Jul 1999 05:33:34 -0700 (PDT) (envelope-from scheper@beast.toad.net) Received: from localhost (scheper@localhost) by beast.toad.net (8.9.3/8.9.3) with ESMTP id IAA27945 for ; Thu, 22 Jul 1999 08:31:47 -0400 Date: Thu, 22 Jul 1999 08:31:47 -0400 (EDT) From: Richard Scheper To: freebsd-smp@FreeBSD.ORG Subject: SMP + XDM = keyboard lockup Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I'm having a slight problem with my current setup, and the solution is eluding me. I've got a supermicro P6DLS with 2 PII-266 processors. I'm running 3.1-RELEASE. I configured the kernel for SMP according to the mptable output. I found no problems in the dmesg output upon bootup. SMP reports the second CPU activated just fine and the boot sequence completes without a hitch. Heres my problem. I have an rc.local which starts up xdm, i.e. /usr/X11R6/bin/xdm which starts up the xdm login screen. However, at this point the keyboard input is locked out. Absolutely no key sequences have any effect. All else appears to work, i.e. the mouse still functions. I've experimented with this and have gathered the following clues: 1. xdm works fine if SMP is disabled in the kernel 2. SMP works fine if xdm is not started in rc.local or in /etc/ttys 3. SMP and xdm work fine if xdm is started under a root login session. 4. the keyboard lockup only occurs after the rc.local script is finished executing. That is, if I put a sleep 60 command after the xdm command in rc.local, the login screen accepts my input and starts X, and everything is fine for exactly one minute at which time the keyboard locks. 5. sleep commands before the xdm command in rc.local have no effect - lockup still occurs. Any insight into this problem would be greatly appreciated. Thanks -Rich To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 5:42:58 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id 4B0D814D4E for ; Thu, 22 Jul 1999 05:42:50 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:t7WM11fL5Z3QqDsydJ4lWARBOGq04FU8@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id VAA11228; Thu, 22 Jul 1999 21:41:44 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id VAA18075; Thu, 22 Jul 1999 21:46:01 +0900 (JST) Message-Id: <199907221246.VAA18075@zodiac.mech.utsunomiya-u.ac.jp> To: Richard Scheper Cc: freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Thu, 22 Jul 1999 08:31:47 -0400." References: Date: Thu, 22 Jul 1999 21:46:01 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >I'm having a slight problem with my current setup, and the solution is >eluding me. I've got a supermicro P6DLS with 2 PII-266 processors. I'm >running 3.1-RELEASE. [...] >Heres my problem. I have an rc.local which starts up xdm, i.e. > >/usr/X11R6/bin/xdm > >which starts up the xdm login screen. However, at this point the keyboard >input is locked out. Absolutely no key sequences have any effect. All else >appears to work, i.e. the mouse still functions. I've experimented with >this and have gathered the following clues: > >1. xdm works fine if SMP is disabled in the kernel > >2. SMP works fine if xdm is not started in rc.local or in /etc/ttys > >3. SMP and xdm work fine if xdm is started under a root login session. > >4. the keyboard lockup only occurs after the rc.local script is finished > executing. That is, if I put a sleep 60 command after the xdm command > in rc.local, the login screen accepts my input and starts X, and > everything is fine for exactly one minute at which time the keyboard > locks. > >5. sleep commands before the xdm command in rc.local have no effect - > lockup still occurs. It appears that xdm and getty is fighting for keyboard input. Try (sleep 60; /usr/X11R6/bin/xdm) & in /etc/rc.local. Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 6: 5:18 1999 Delivered-To: freebsd-smp@freebsd.org Received: from beast.toad.net (beast.toad.net [205.197.182.100]) by hub.freebsd.org (Postfix) with ESMTP id A80211525F for ; Thu, 22 Jul 1999 06:05:10 -0700 (PDT) (envelope-from scheper@beast.toad.net) Received: from localhost (scheper@localhost) by beast.toad.net (8.9.3/8.9.3) with ESMTP id JAA01046 for ; Thu, 22 Jul 1999 09:05:04 -0400 Date: Thu, 22 Jul 1999 09:05:04 -0400 (EDT) From: Richard Scheper To: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221246.VAA18075@zodiac.mech.utsunomiya-u.ac.jp> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, 22 Jul 1999, Kazutaka YOKOTA wrote: > > It appears that xdm and getty is fighting for keyboard input. Try > > (sleep 60; /usr/X11R6/bin/xdm) & > > in /etc/rc.local. > > Kazu > Thanks for the suggestion. I'll try it. I wonder why this problem is peculiar to a SMP enabled kernel? Is it because xdm and getty are possibly started on different processors? What difference sould running xdm on the background make? The problem seems to occur when /etc/rc.local completes execution whether or not xdm is executed from rc.local. That is, I tried setting up /etc/ttys as tyv3 "/usr/X11R6/bin/xdm -nodaemon" xterm on secure with an empty /etc/rc.local and still got the lockup. Wouldn't that setup avoid the getty conflict? I am admittedly confused.. -Rich To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 7:18:36 1999 Delivered-To: freebsd-smp@freebsd.org Received: from beast.toad.net (beast.toad.net [205.197.182.100]) by hub.freebsd.org (Postfix) with ESMTP id 624C414D61 for ; Thu, 22 Jul 1999 07:18:32 -0700 (PDT) (envelope-from scheper@beast.toad.net) Received: from localhost (scheper@localhost) by beast.toad.net (8.9.3/8.9.3) with ESMTP id KAA15996 for ; Thu, 22 Jul 1999 10:18:00 -0400 Date: Thu, 22 Jul 1999 10:18:00 -0400 (EDT) From: Richard Scheper To: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221405.XAA19695@zodiac.mech.utsunomiya-u.ac.jp> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Wow. Thanks for the explanation. That makes sense. So if I throw the sleep and xdm commands into the background in /etc/rc.local, the script will complete before xdm starts running, giving init enough time to set itself up. Then getty grabs tyv1 before xdm can.. if I understand this right. I thought I had it set up so that the X server could only grab tyv3, but I guess I was wrong.. -Rich On Thu, 22 Jul 1999, Kazutaka YOKOTA wrote: > > >Heres my problem. I have an rc.local which starts up xdm, i.e. > > > >/usr/X11R6/bin/xdm > > > >which starts up the xdm login screen. However, at this point the keyboard > >input is locked out. Absolutely no key sequences have any effect. All else > >appears to work, i.e. the mouse still functions. > > Here is a bit of explanation. > > The XFree86 server tries to locate an empty vty to use for the X > sesssion. If xdm is started in rc.local, the X server will run BEFORE > init(8) starts invoking getty(8), and will find /dev/ttyv1 empty and > tries to occupy that vty. When rc.init is finished, init(8) will run > a copy of getty(8) in /dev/ttyv1, then, *Bomb* The X server and getty > both tries to read from the keyboard. > > >I've experimented with > >this and have gathered the following clues: > > > >1. xdm works fine if SMP is disabled in the kernel > > This means that the X server has not started before getty(8). > Probably xdm is still busy parsing and processing its configuration > file when init is staring getty in vtys. > > This is just a timing problem, and you are just lucky. > > >2. SMP works fine if xdm is not started in rc.local or in /etc/ttys > > It should :-) > > >3. SMP and xdm work fine if xdm is started under a root login session. > > It should too. > > >4. the keyboard lockup only occurs after the rc.local script is finished > > executing. That is, if I put a sleep 60 command after the xdm command > > in rc.local, the login screen accepts my input and starts X, and > > everything is fine for exactly one minute at which time the keyboard > > locks. > > This sleep command only delays the completion of rc.local. The X > server will be able to read from the keyboard only until getty starts > running. This won't solve the problem at all ;-< > > >5. sleep commands before the xdm command in rc.local have no effect - > > lockup still occurs. > > It won't work either. > > > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 9:36:44 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ix.netcom.com (sil-wa19-10.ix.netcom.com [207.93.145.10]) by hub.freebsd.org (Postfix) with ESMTP id 100C715314 for ; Thu, 22 Jul 1999 09:36:40 -0700 (PDT) (envelope-from tomdean@ix.netcom.com) Received: (from tomdean@localhost) by ix.netcom.com (8.9.3/8.8.8) id JAA51919; Thu, 22 Jul 1999 09:34:59 -0700 (PDT) (envelope-from tomdean@ix.netcom.com) Date: Thu, 22 Jul 1999 09:34:59 -0700 (PDT) Message-Id: <199907221634.JAA51919@ix.netcom.com> X-Authentication-Warning: celebris.tddhome: tomdean set sender to tomdean@ix.netcom.com using -f From: Thomas Dean To: scheper@beast.toad.net Cc: freebsd-smp@FreeBSD.ORG In-reply-to: (message from Richard Scheper on Thu, 22 Jul 1999 10:18:00 -0400 (EDT)) Subject: Re: SMP + XDM = keyboard lockup References: Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I have xdm started from /usr/local/etc/rc.d/xdm.sh. It was stated earlier that ./rc.d/*.sh is the wave of the future. This works fine. No sleep, no '&'. /usr/local/etc/rc.d/xdm.sh is run before rc.local! /etc/rc looks for scripts in the location contained in the local_startup variable, which is set in /etc/defaults/rc.defaults to /usr/local/etc/rc.d. The script is executed with '(set -T ; trap 'exit 1' 2 ; ${script} start)' My /etc/ttys contains ... ttyv0 "/usr/libexec/getty Pc" cons25 on secure ttyv1 "/usr/libexec/getty Pc" cons25 on secure ttyv2 "/usr/libexec/getty Pc" cons25 on secure ttyv3 "/usr/libexec/getty Pc" cons25 on secure ttyv4 "/usr/libexec/getty Pc" cons25 on secure ttyv5 "/usr/libexec/getty Pc" cons25 on secure ttyv6 "/usr/libexec/getty Pc" cons25 on secure ttyv7 "/usr/libexec/getty Pc" cons25 off secure ... my /usr/local/etc/rc.d/xdm.sh: #! /bin/sh # $Id: xdm.sh,v 1.1 1999/02/14 06:05:30 tomdean Exp $ # start xdm # 19990115 tomdean - initial version if [ -x /usr/X11R6/bin/xdm ] ; then # need to cleanup first. if [ -f /usr/X11R6/lib/X11/xdm/xdm-pid ] ; then rm -f /usr/X11R6/lib/X11/xdm/xdm-pid fi if [ -f /tmp/.X0-lock ] ; then rm -f /tmp/.X0-lock fi # now, we can start it. echo " xdm"; /usr/X11R6/bin/xdm fi tomdean To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 10:11:13 1999 Delivered-To: freebsd-smp@freebsd.org Received: from beast.toad.net (beast.toad.net [205.197.182.100]) by hub.freebsd.org (Postfix) with ESMTP id ACE2615480 for ; Thu, 22 Jul 1999 10:11:09 -0700 (PDT) (envelope-from scheper@beast.toad.net) Received: from localhost (scheper@localhost) by beast.toad.net (8.9.3/8.9.3) with ESMTP id NAA18858 for ; Thu, 22 Jul 1999 13:10:18 -0400 Date: Thu, 22 Jul 1999 13:10:18 -0400 (EDT) From: Richard Scheper To: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221634.JAA51919@ix.netcom.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Thanks for the response.. but i must admit I'm still confused. I thought the problem was that the speed increase of SMP was allowing the X server to grab ttyv0 before init started getty on it. If rc.d runs before rc.local, then I don't see how this could help. Also, I don't understand why initializing xdm in /etc/ttys doesn't work since by then init should have executed getty on the appropriate ttvs. Could anyone explain this to me? Thanks -Rich On Thu, 22 Jul 1999, Thomas Dean wrote: > I have xdm started from /usr/local/etc/rc.d/xdm.sh. It was stated > earlier that ./rc.d/*.sh is the wave of the future. This works fine. > No sleep, no '&'. /usr/local/etc/rc.d/xdm.sh is run before rc.local! > > /etc/rc looks for scripts in the location contained in the > local_startup variable, which is set in /etc/defaults/rc.defaults to > /usr/local/etc/rc.d. > > The script is executed with '(set -T ; trap 'exit 1' 2 ; ${script} start)' > > My /etc/ttys contains > ... > ttyv0 "/usr/libexec/getty Pc" cons25 on secure > ttyv1 "/usr/libexec/getty Pc" cons25 on secure > ttyv2 "/usr/libexec/getty Pc" cons25 on secure > ttyv3 "/usr/libexec/getty Pc" cons25 on secure > ttyv4 "/usr/libexec/getty Pc" cons25 on secure > ttyv5 "/usr/libexec/getty Pc" cons25 on secure > ttyv6 "/usr/libexec/getty Pc" cons25 on secure > ttyv7 "/usr/libexec/getty Pc" cons25 off secure > ... > > > my /usr/local/etc/rc.d/xdm.sh: > > #! /bin/sh > # $Id: xdm.sh,v 1.1 1999/02/14 06:05:30 tomdean Exp $ > > # start xdm > > # 19990115 tomdean - initial version > > if [ -x /usr/X11R6/bin/xdm ] ; then > # need to cleanup first. > if [ -f /usr/X11R6/lib/X11/xdm/xdm-pid ] ; then > rm -f /usr/X11R6/lib/X11/xdm/xdm-pid > fi > if [ -f /tmp/.X0-lock ] ; then > rm -f /tmp/.X0-lock > fi > # now, we can start it. > echo " xdm"; /usr/X11R6/bin/xdm > fi > > tomdean > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 10:25:36 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ns.mt.sri.com (ns.mt.sri.com [206.127.79.91]) by hub.freebsd.org (Postfix) with ESMTP id D9EDE14DFB for ; Thu, 22 Jul 1999 10:25:22 -0700 (PDT) (envelope-from nate@mt.sri.com) Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id LAA22967; Thu, 22 Jul 1999 11:25:21 -0600 (MDT) (envelope-from nate@rocky.mt.sri.com) Received: by mt.sri.com (SMI-8.6/SMI-SVR4) id LAA02645; Thu, 22 Jul 1999 11:25:21 -0600 Date: Thu, 22 Jul 1999 11:25:21 -0600 Message-Id: <199907221725.LAA02645@mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Richard Scheper Cc: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: References: <199907221634.JAA51919@ix.netcom.com> X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Also, I don't understand why initializing xdm in /etc/ttys doesn't work > since by then init should have executed getty on the appropriate ttvs. It *should* work. I'm not sure why it's not working, unless there are no free ttys for it to work on. Nate To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11: 0:55 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp01.primenet.com (smtp01.primenet.com [206.165.6.131]) by hub.freebsd.org (Postfix) with ESMTP id 4E9AD1528B for ; Thu, 22 Jul 1999 11:00:45 -0700 (PDT) (envelope-from tlambert@usr05.primenet.com) Received: (from daemon@localhost) by smtp01.primenet.com (8.8.8/8.8.8) id LAA16499; Thu, 22 Jul 1999 11:00:41 -0700 (MST) Received: from usr05.primenet.com(206.165.6.205) via SMTP by smtp01.primenet.com, id smtpd016441; Thu Jul 22 11:00:33 1999 Received: (from tlambert@localhost) by usr05.primenet.com (8.8.5/8.8.5) id LAA08972; Thu, 22 Jul 1999 11:00:29 -0700 (MST) From: Terry Lambert Message-Id: <199907221800.LAA08972@usr05.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: scheper@beast.toad.net (Richard Scheper) Date: Thu, 22 Jul 1999 18:00:29 +0000 (GMT) Cc: freebsd-smp@FreeBSD.ORG In-Reply-To: from "Richard Scheper" at Jul 22, 99 08:31:47 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Heres my problem. I have an rc.local which starts up xdm, i.e. > > /usr/X11R6/bin/xdm > > which starts up the xdm login screen. However, at this point the keyboard > input is locked out. Absolutely no key sequences have any effect. All else > appears to work, i.e. the mouse still functions. I've experimented with > this and have gathered the following clues: The xdm program, among other things, starts X. I believe the issue is one CPU running X doing inb/outb's while the other CPU is in the kernel -- doing the same. In order to fix this properly, you would need to do one of the following: o modify X to use a system-call based mechanism to do the requisite inb/outb calls o modify X to use a device-I/O based mechanism to do the requisite inb/outb calls o map a read-only page into user space as the I/O space, and have X write to it, doing the actual I/O in the fault handler in the kernel o Implement GGI for FreeBSD (the GGI people are willing to let the kernel components be BSD licensed for FreeBSD, and seem very interested in a FreeBSD version, but they lack a FreeBSD volunteer for their project) o Externalize access to the "Big Giant Lock", and let the X server hold the lock while doing its inb/outb calls, taking special care about involuntary context switches while the lock is held (Note: this approach sucks for many reasons, including that user space access to the lock seriously damages the concurrency model) o Do what Linux does, and virtualize the I/O bus space through traps (this would also let you limit the I/O space exposure you allow, since it would prevent access to all but designated I/O areas; in all, it's a security win, as well as a method of serializing I/O space access) o Don't have more than one CPU running while the X server has its cold, wet nose in the kernel's crotch All of these have to do with preventing reentrancy of the I/O bus space by more than one processor, and all but the last of them have to do with holding the BGL while doing the I/O. The last keeps the processors from hitting the BGL until after X is up (e.g. they spin on the AP startup spinlock waiting for the BP to get their environment fully set up). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11: 5:52 1999 Delivered-To: freebsd-smp@freebsd.org Received: from beast.toad.net (beast.toad.net [205.197.182.100]) by hub.freebsd.org (Postfix) with ESMTP id 88C4814E41 for ; Thu, 22 Jul 1999 11:05:50 -0700 (PDT) (envelope-from scheper@beast.toad.net) Received: from localhost (scheper@localhost) by beast.toad.net (8.9.3/8.9.3) with ESMTP id OAA30792 for ; Thu, 22 Jul 1999 14:05:26 -0400 Date: Thu, 22 Jul 1999 14:05:25 -0400 (EDT) From: Richard Scheper To: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Thanks to all who have responded thus far. I still don't think we have an explanation as to what is happening. In my mind, the following is evidence against a getty conflict. 1. the problem is delayed if rc.local is put to sleep. During that time I can alt-ctrl-f1 to ttyv1. I can stay there all day if I want with no probs., so it doesn't seem that the X server is trying to grab it. However, as soon as I switch back to ttyv3, the keyboard is locked. 2. If I abandon the rc script method of starting xdm and instead declare ttyv3 "/usr/X11R6/bin/xdm -nodaemon" xterm on secure I still get the problem. In that case, I don't see how getty could cause a conflict. I'm really clueless here! (but learning alot at least) -Rich To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:13:17 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ns.mt.sri.com (ns.mt.sri.com [206.127.79.91]) by hub.freebsd.org (Postfix) with ESMTP id 3F2D214DE0 for ; Thu, 22 Jul 1999 11:13:15 -0700 (PDT) (envelope-from nate@mt.sri.com) Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id MAA23607; Thu, 22 Jul 1999 12:12:07 -0600 (MDT) (envelope-from nate@rocky.mt.sri.com) Received: by mt.sri.com (SMI-8.6/SMI-SVR4) id MAA02907; Thu, 22 Jul 1999 12:12:02 -0600 Date: Thu, 22 Jul 1999 12:12:02 -0600 Message-Id: <199907221812.MAA02907@mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Terry Lambert Cc: scheper@beast.toad.net (Richard Scheper), freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221800.LAA08972@usr05.primenet.com> References: <199907221800.LAA08972@usr05.primenet.com> X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > Heres my problem. I have an rc.local which starts up xdm, i.e. > > > > /usr/X11R6/bin/xdm > > > > which starts up the xdm login screen. However, at this point the keyboard > > input is locked out. Absolutely no key sequences have any effect. All else > > appears to work, i.e. the mouse still functions. I've experimented with > > this and have gathered the following clues: > > The xdm program, among other things, starts X. > > I believe the issue is one CPU running X doing inb/outb's while > the other CPU is in the kernel -- doing the same. Why would the other CPU doing the same inb/outb's as X? By the time init is run, the kernel is done 'setting up' the kernel. Nate To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:18:48 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ns.mt.sri.com (ns.mt.sri.com [206.127.79.91]) by hub.freebsd.org (Postfix) with ESMTP id EE28314DE0 for ; Thu, 22 Jul 1999 11:18:46 -0700 (PDT) (envelope-from nate@mt.sri.com) Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id MAA23669; Thu, 22 Jul 1999 12:16:57 -0600 (MDT) (envelope-from nate@rocky.mt.sri.com) Received: by mt.sri.com (SMI-8.6/SMI-SVR4) id MAA02941; Thu, 22 Jul 1999 12:16:56 -0600 Date: Thu, 22 Jul 1999 12:16:56 -0600 Message-Id: <199907221816.MAA02941@mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Nate Williams Cc: Terry Lambert , scheper@beast.toad.net (Richard Scheper), freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221812.MAA02907@mt.sri.com> References: <199907221800.LAA08972@usr05.primenet.com> <199907221812.MAA02907@mt.sri.com> X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > I believe the issue is one CPU running X doing inb/outb's while > > the other CPU is in the kernel -- doing the same. > > Why would the other CPU doing the same inb/outb's as X? By the time > init is run, the kernel is done 'setting up' the kernel. That should be 'the kernel is done setting up the ttys in the kernel'. Nate To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:18:53 1999 Delivered-To: freebsd-smp@freebsd.org Received: from beast.toad.net (beast.toad.net [205.197.182.100]) by hub.freebsd.org (Postfix) with ESMTP id 1AF621558B for ; Thu, 22 Jul 1999 11:18:47 -0700 (PDT) (envelope-from scheper@beast.toad.net) Received: from localhost (scheper@localhost) by beast.toad.net (8.9.3/8.9.3) with ESMTP id OAA00917 for ; Thu, 22 Jul 1999 14:17:07 -0400 Date: Thu, 22 Jul 1999 14:17:07 -0400 (EDT) From: Richard Scheper To: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221800.LAA08972@usr05.primenet.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Wow! That's a lot to chew and mostly over my head, but a couple of questions.. If what you say is true, then wouldn't I have toruble running xdm even from a root login after the boot process is over? I don't. How could I delay the startup of one of the processors, as you suggest in your last point? thanks for the detailed response! -Rich On Thu, 22 Jul 1999, Terry Lambert wrote: > > Heres my problem. I have an rc.local which starts up xdm, i.e. > > > > /usr/X11R6/bin/xdm > > > > which starts up the xdm login screen. However, at this point the keyboard > > input is locked out. Absolutely no key sequences have any effect. All else > > appears to work, i.e. the mouse still functions. I've experimented with > > this and have gathered the following clues: > > The xdm program, among other things, starts X. > > I believe the issue is one CPU running X doing inb/outb's while > the other CPU is in the kernel -- doing the same. > > In order to fix this properly, you would need to do one of the > following: > > o modify X to use a system-call based mechanism to do the > requisite inb/outb calls > > o modify X to use a device-I/O based mechanism to do the > requisite inb/outb calls > > o map a read-only page into user space as the I/O space, > and have X write to it, doing the actual I/O in the fault > handler in the kernel > > o Implement GGI for FreeBSD (the GGI people are willing to > let the kernel components be BSD licensed for FreeBSD, > and seem very interested in a FreeBSD version, but they > lack a FreeBSD volunteer for their project) > > o Externalize access to the "Big Giant Lock", and let the > X server hold the lock while doing its inb/outb calls, > taking special care about involuntary context switches > while the lock is held (Note: this approach sucks for > many reasons, including that user space access to the > lock seriously damages the concurrency model) > > o Do what Linux does, and virtualize the I/O bus space > through traps (this would also let you limit the I/O > space exposure you allow, since it would prevent access > to all but designated I/O areas; in all, it's a security > win, as well as a method of serializing I/O space access) > > o Don't have more than one CPU running while the X server > has its cold, wet nose in the kernel's crotch > > All of these have to do with preventing reentrancy of the I/O bus > space by more than one processor, and all but the last of them > have to do with holding the BGL while doing the I/O. The last > keeps the processors from hitting the BGL until after X is up > (e.g. they spin on the AP startup spinlock waiting for the BP > to get their environment fully set up). > > > Terry Lambert > terry@lambert.org > --- > Any opinions in this posting are my own and not those of my present > or previous employers. > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:19: 3 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp02.primenet.com (smtp02.primenet.com [206.165.6.132]) by hub.freebsd.org (Postfix) with ESMTP id CAC8215608 for ; Thu, 22 Jul 1999 11:18:55 -0700 (PDT) (envelope-from tlambert@usr05.primenet.com) Received: (from daemon@localhost) by smtp02.primenet.com (8.8.8/8.8.8) id LAA19240; Thu, 22 Jul 1999 11:17:27 -0700 (MST) Received: from usr05.primenet.com(206.165.6.205) via SMTP by smtp02.primenet.com, id smtpd019217; Thu Jul 22 11:17:23 1999 Received: (from tlambert@localhost) by usr05.primenet.com (8.8.5/8.8.5) id LAA09698; Thu, 22 Jul 1999 11:17:22 -0700 (MST) From: Terry Lambert Message-Id: <199907221817.LAA09698@usr05.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: scheper@beast.toad.net (Richard Scheper) Date: Thu, 22 Jul 1999 18:17:19 +0000 (GMT) Cc: freebsd-smp@FreeBSD.ORG In-Reply-To: from "Richard Scheper" at Jul 22, 99 10:18:00 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > The XFree86 server tries to locate an empty vty to use for the X > > sesssion. If xdm is started in rc.local, the X server will run BEFORE > > init(8) starts invoking getty(8), and will find /dev/ttyv1 empty and > > tries to occupy that vty. When rc.init is finished, init(8) will run > > a copy of getty(8) in /dev/ttyv1, then, *Bomb* The X server and getty > > both tries to read from the keyboard. I don't believe this is the case, because of the SMP factor. If SMP is disabled, he does not see the problem. I believe that if what you are saying is correct, merely following the xdm documentation and starting the X server in /etc/ttys should resolve his problem. I still believe, however, that there is a real problem with user space code accessing the I/O bus directly under SMP, without it holding the Big Giant Lock. There is a real reason that the lock exists; it's not just there to piss all over concurrency -- that's just a side effect. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:41:38 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp04.primenet.com (smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (Postfix) with ESMTP id 487C314CD0 for ; Thu, 22 Jul 1999 11:41:32 -0700 (PDT) (envelope-from tlambert@usr05.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id MAA05220; Thu, 22 Jul 1999 12:03:45 -0700 (MST) Received: from usr05.primenet.com(206.165.6.205) via SMTP by smtp04.primenet.com, id smtpd005164; Thu Jul 22 12:03:35 1999 Received: (from tlambert@localhost) by usr05.primenet.com (8.8.5/8.8.5) id LAA10550; Thu, 22 Jul 1999 11:39:10 -0700 (MST) From: Terry Lambert Message-Id: <199907221839.LAA10550@usr05.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: nate@mt.sri.com (Nate Williams) Date: Thu, 22 Jul 1999 18:39:10 +0000 (GMT) Cc: tlambert@primenet.com, scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG In-Reply-To: <199907221812.MAA02907@mt.sri.com> from "Nate Williams" at Jul 22, 99 12:12:02 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > The xdm program, among other things, starts X. > > > > I believe the issue is one CPU running X doing inb/outb's while > > the other CPU is in the kernel -- doing the same. > > Why would the other CPU doing the same inb/outb's as X? By the time > init is run, the kernel is done 'setting up' the kernel. It wouldn't be. It would merely be trying to latch different addresses on the I/O bus and/or write data to the previously latched address, at the same time. Basically, user space code should keep its cold, wet nose out of the kernels crotch. I believe the issue is the system is still starting up, and thus has the kernel poking at things. You could guess that the init and the X server might be running on different CPU's, but then you would expect the /etc/ttys file based xdm startup to work. This jibes with his statement that manually starting the xdm from a root shell prompt on a relatively quiescent system "works". It is likely that it "works" only because there is not simultaneous activity. If I am correct, I suspect that on an active SMP system, console switching in and out of X would result in similar lockups, so this should be verifiable. Since he appears to have a way in, other than the console, he should: 1) Cause the lockup 2) Go in and do a 'ps' to see if conflicting processes are running 3) Force a console switch via escape sequence to a console vty not currently in use. This should result in the X server disengaging, and the input devices being detached and reattached to the new vty, if the problem is merely a race. Also, if both terminals are reading input, as has been posited, a ctrl-alt-F2 should still result in a console switch, regardless of who takes the keypress event (ctrl and alt are modifiers, so only the F2 keycode should be relevent). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:46: 8 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ns.mt.sri.com (ns.mt.sri.com [206.127.79.91]) by hub.freebsd.org (Postfix) with ESMTP id E1E6614BF8 for ; Thu, 22 Jul 1999 11:46:05 -0700 (PDT) (envelope-from nate@mt.sri.com) Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id MAA24056; Thu, 22 Jul 1999 12:45:34 -0600 (MDT) (envelope-from nate@rocky.mt.sri.com) Received: by mt.sri.com (SMI-8.6/SMI-SVR4) id MAA03196; Thu, 22 Jul 1999 12:45:33 -0600 Date: Thu, 22 Jul 1999 12:45:33 -0600 Message-Id: <199907221845.MAA03196@mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Terry Lambert Cc: nate@mt.sri.com (Nate Williams), scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221839.LAA10550@usr05.primenet.com> References: <199907221812.MAA02907@mt.sri.com> <199907221839.LAA10550@usr05.primenet.com> X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > The xdm program, among other things, starts X. > > > > > > I believe the issue is one CPU running X doing inb/outb's while > > > the other CPU is in the kernel -- doing the same. > > > > Why would the other CPU doing the same inb/outb's as X? By the time > > init is run, the kernel is done 'setting up' the kernel. > > It wouldn't be. > > It would merely be trying to latch different addresses on the I/O > bus and/or write data to the previously latched address, at the > same time. > > Basically, user space code should keep its cold, wet nose out of > the kernels crotch. > > I believe the issue is the system is still starting up, and thus > has the kernel poking at things. It shouldn't be. The kernel is completely done 'starting up', else all the /etc/rc.* files wouldn't have been run and init wouldn't have started spawning off gettys. > You could guess that the init and the X server might be running > on different CPU's, but then you would expect the /etc/ttys file > based xdm startup to work. Right. > This jibes with his statement that manually starting the xdm from > a root shell prompt on a relatively quiescent system "works". It > is likely that it "works" only because there is not simultaneous > activity. Or there is a a race in the video/keyboard driver that is causing problems. X calls into the driver, so it's possible a missing spl() or somesuch is causing problem. I doubt very much the problem is related to inb/outb, and is instead a just a 'bug'. :) Nate To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:46:37 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp03.primenet.com (smtp03.primenet.com [206.165.6.133]) by hub.freebsd.org (Postfix) with ESMTP id 4795814BF8 for ; Thu, 22 Jul 1999 11:46:34 -0700 (PDT) (envelope-from tlambert@usr05.primenet.com) Received: (from daemon@localhost) by smtp03.primenet.com (8.8.8/8.8.8) id LAA23663; Thu, 22 Jul 1999 11:45:49 -0700 (MST) Received: from usr05.primenet.com(206.165.6.205) via SMTP by smtp03.primenet.com, id smtpd023636; Thu Jul 22 11:45:44 1999 Received: (from tlambert@localhost) by usr05.primenet.com (8.8.5/8.8.5) id LAA10738; Thu, 22 Jul 1999 11:45:41 -0700 (MST) From: Terry Lambert Message-Id: <199907221845.LAA10738@usr05.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: scheper@beast.toad.net (Richard Scheper) Date: Thu, 22 Jul 1999 18:45:41 +0000 (GMT) Cc: freebsd-smp@FreeBSD.ORG In-Reply-To: from "Richard Scheper" at Jul 22, 99 02:17:07 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > Wow! That's a lot to chew and mostly over my head, but a couple of > questions.. > > If what you say is true, then wouldn't I have toruble running xdm even > from a root login after the boot process is over? I don't. No, not if the system is relatively quiescent. As I told Nate, you should be able to easily test this by loading the I/O system (preferrably *not* your disks) and bouncing the console in and out of X using ctrl-alt-F1/[ctrl-]alt-F4. If it locks up, the problem has been identified. > How could I delay the startup of one of the processors, as you suggest in > your last point? There is a sysctl which controlls the startup of the auxillary processors. If you were to delay starting the processors until after X was running, you would be able to see if this were related to the SMP code path differences if it still locks up without the other processors running (I don't believe it will) or to an I/O space conflict (It may lock up following the AP's starting up, but it probably won't, since when you do the manual start, the system will be relatively quiescent). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:49: 8 1999 Delivered-To: freebsd-smp@freebsd.org Received: from dingo.cdrom.com (dingo.cdrom.com [204.216.28.145]) by hub.freebsd.org (Postfix) with ESMTP id 8DDA814BF8 for ; Thu, 22 Jul 1999 11:49:06 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost.cdrom.com [127.0.0.1]) by dingo.cdrom.com (8.9.3/8.8.8) with ESMTP id LAA01257; Thu, 22 Jul 1999 11:42:37 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Message-Id: <199907221842.LAA01257@dingo.cdrom.com> X-Mailer: exmh version 2.0.2 2/24/98 To: Terry Lambert Cc: nate@mt.sri.com (Nate Williams), scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Thu, 22 Jul 1999 18:39:10 -0000." <199907221839.LAA10550@usr05.primenet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 22 Jul 1999 11:42:37 -0700 From: Mike Smith Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > The xdm program, among other things, starts X. > > > > > > I believe the issue is one CPU running X doing inb/outb's while > > > the other CPU is in the kernel -- doing the same. > > > > Why would the other CPU doing the same inb/outb's as X? By the time > > init is run, the kernel is done 'setting up' the kernel. > > It wouldn't be. > > It would merely be trying to latch different addresses on the I/O > bus and/or write data to the previously latched address, at the > same time. This comment demonstrates that you're out of your depth. Back to the other end of the pool, and take your squeaky toys with you... (It doesn't work like that.) -- \\ The mind's the standard \\ Mike Smith \\ of the man. \\ msmith@freebsd.org \\ -- Joseph Merrick \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:51:42 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp04.primenet.com (smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (Postfix) with ESMTP id 3D5A114BF8 for ; Thu, 22 Jul 1999 11:51:35 -0700 (PDT) (envelope-from tlambert@usr05.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id MAA10102; Thu, 22 Jul 1999 12:15:04 -0700 (MST) Received: from usr05.primenet.com(206.165.6.205) via SMTP by smtp04.primenet.com, id smtpd009968; Thu Jul 22 12:14:47 1999 Received: (from tlambert@localhost) by usr05.primenet.com (8.8.5/8.8.5) id LAA10852; Thu, 22 Jul 1999 11:50:24 -0700 (MST) From: Terry Lambert Message-Id: <199907221850.LAA10852@usr05.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: nate@mt.sri.com (Nate Williams) Date: Thu, 22 Jul 1999 18:50:24 +0000 (GMT) Cc: tlambert@primenet.com, nate@mt.sri.com, scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG In-Reply-To: <199907221845.MAA03196@mt.sri.com> from "Nate Williams" at Jul 22, 99 12:45:33 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > I believe the issue is the system is still starting up, and thus > > has the kernel poking at things. > > It shouldn't be. The kernel is completely done 'starting up', else all > the /etc/rc.* files wouldn't have been run and init wouldn't have > started spawning off gettys. I'm referring to things that result in a lot of I/O activity which are started in the rc files (NFS, DNS, sendmail, etc.). The system isn't quiescent until a while _after_ the rc files are run (long enough to log in). > > This jibes with his statement that manually starting the xdm from > > a root shell prompt on a relatively quiescent system "works". It > > is likely that it "works" only because there is not simultaneous > > activity. > > Or there is a a race in the video/keyboard driver that is causing > problems. X calls into the driver, so it's possible a missing spl() or > somesuch is causing problem. > > I doubt very much the problem is related to inb/outb, and is instead a > just a 'bug'. :) Maybe. I think we should hold off investing too much of ourselves into our respective theories until after he has done some of the tests I suggested. You should feel free to suggest tests, too, of course, since I could be wrong, no matter how "right" it feels to me. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:55: 6 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp02.primenet.com (smtp02.primenet.com [206.165.6.132]) by hub.freebsd.org (Postfix) with ESMTP id A9EB414D82 for ; Thu, 22 Jul 1999 11:55:02 -0700 (PDT) (envelope-from tlambert@usr05.primenet.com) Received: (from daemon@localhost) by smtp02.primenet.com (8.8.8/8.8.8) id LAA05454; Thu, 22 Jul 1999 11:53:13 -0700 (MST) Received: from usr05.primenet.com(206.165.6.205) via SMTP by smtp02.primenet.com, id smtpd005423; Thu Jul 22 11:53:06 1999 Received: (from tlambert@localhost) by usr05.primenet.com (8.8.5/8.8.5) id LAA10953; Thu, 22 Jul 1999 11:53:05 -0700 (MST) From: Terry Lambert Message-Id: <199907221853.LAA10953@usr05.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: mike@smith.net.au (Mike Smith) Date: Thu, 22 Jul 1999 18:53:05 +0000 (GMT) Cc: tlambert@primenet.com, nate@mt.sri.com, scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG In-Reply-To: <199907221842.LAA01257@dingo.cdrom.com> from "Mike Smith" at Jul 22, 99 11:42:37 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > Why would the other CPU doing the same inb/outb's as X? By the time > > > init is run, the kernel is done 'setting up' the kernel. > > > > It wouldn't be. > > > > It would merely be trying to latch different addresses on the I/O > > bus and/or write data to the previously latched address, at the > > same time. > > This comment demonstrates that you're out of your depth. Back to the > other end of the pool, and take your squeaky toys with you... > > (It doesn't work like that.) Cool. So we don't need the BGL for interrupt processing... Note: "...trying to..." Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 11:56: 8 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ns.mt.sri.com (ns.mt.sri.com [206.127.79.91]) by hub.freebsd.org (Postfix) with ESMTP id E296814DEE for ; Thu, 22 Jul 1999 11:55:59 -0700 (PDT) (envelope-from nate@mt.sri.com) Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id MAA24264; Thu, 22 Jul 1999 12:55:51 -0600 (MDT) (envelope-from nate@rocky.mt.sri.com) Received: by mt.sri.com (SMI-8.6/SMI-SVR4) id MAA03429; Thu, 22 Jul 1999 12:55:50 -0600 Date: Thu, 22 Jul 1999 12:55:50 -0600 Message-Id: <199907221855.MAA03429@mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Terry Lambert Cc: nate@mt.sri.com (Nate Williams), scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221850.LAA10852@usr05.primenet.com> References: <199907221845.MAA03196@mt.sri.com> <199907221850.LAA10852@usr05.primenet.com> X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > I believe the issue is the system is still starting up, and thus > > > has the kernel poking at things. > > > > It shouldn't be. The kernel is completely done 'starting up', else all > > the /etc/rc.* files wouldn't have been run and init wouldn't have > > started spawning off gettys. > > I'm referring to things that result in a lot of I/O activity > which are started in the rc files (NFS, DNS, sendmail, etc.). Sure, but none of those are going to compete with X for resources that 'conflict' or would cause lockups. > > > This jibes with his statement that manually starting the xdm from > > > a root shell prompt on a relatively quiescent system "works". It > > > is likely that it "works" only because there is not simultaneous > > > activity. > > > > Or there is a a race in the video/keyboard driver that is causing > > problems. X calls into the driver, so it's possible a missing spl() or > > somesuch is causing problem. > > > > I doubt very much the problem is related to inb/outb, and is instead a > > just a 'bug'. :) > > Maybe. I think we should hold off investing too much of ourselves > into our respective theories until after he has done some of the > tests I suggested. Your tests won't prove anything one way or the other. > You should feel free to suggest tests, too, of course, since I could > be wrong, no matter how "right" it feels to me. I don't think we can 'test' this one to test any particular theory. Any test complex enough to point to a particular problem would be almost as complex as fixing the bug. Instead, we should find someone with the knowledge (Kazu and/or Soren) and hardware to recreate and fix the problem. Nate To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 12: 0:50 1999 Delivered-To: freebsd-smp@freebsd.org Received: from dingo.cdrom.com (dingo.cdrom.com [204.216.28.145]) by hub.freebsd.org (Postfix) with ESMTP id D472B14D82 for ; Thu, 22 Jul 1999 12:00:48 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Received: from dingo.cdrom.com (localhost.cdrom.com [127.0.0.1]) by dingo.cdrom.com (8.9.3/8.8.8) with ESMTP id LAA01328; Thu, 22 Jul 1999 11:54:03 -0700 (PDT) (envelope-from mike@dingo.cdrom.com) Message-Id: <199907221854.LAA01328@dingo.cdrom.com> X-Mailer: exmh version 2.0.2 2/24/98 To: Terry Lambert Cc: mike@smith.net.au (Mike Smith), nate@mt.sri.com, scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Thu, 22 Jul 1999 18:53:05 -0000." <199907221853.LAA10953@usr05.primenet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 22 Jul 1999 11:54:03 -0700 From: Mike Smith Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > > Why would the other CPU doing the same inb/outb's as X? By the time > > > > init is run, the kernel is done 'setting up' the kernel. > > > > > > It wouldn't be. > > > > > > It would merely be trying to latch different addresses on the I/O > > > bus and/or write data to the previously latched address, at the > > > same time. > > > > This comment demonstrates that you're out of your depth. Back to the > > other end of the pool, and take your squeaky toys with you... > > > > (It doesn't work like that.) > > Cool. So we don't need the BGL for interrupt processing... At the moment we do, but with finer-grained locking on a few things, no. -- \\ The mind's the standard \\ Mike Smith \\ of the man. \\ msmith@freebsd.org \\ -- Joseph Merrick \\ msmith@cdrom.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 12:17:39 1999 Delivered-To: freebsd-smp@freebsd.org Received: from beast.toad.net (beast.toad.net [205.197.182.100]) by hub.freebsd.org (Postfix) with ESMTP id A512D14D82 for ; Thu, 22 Jul 1999 12:17:31 -0700 (PDT) (envelope-from scheper@beast.toad.net) Received: from localhost (scheper@localhost) by beast.toad.net (8.9.3/8.9.3) with ESMTP id PAA14623 for ; Thu, 22 Jul 1999 15:16:20 -0400 Date: Thu, 22 Jul 1999 15:16:20 -0400 (EDT) From: Richard Scheper To: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907221839.LAA10550@usr05.primenet.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Actually, I don't have another way in, just the console. If I cause the lockup I have no recourse but to turn the computer off. It makes it tough to get info, I know. I never see any error messages in /var/log/messages -Rich On Thu, 22 Jul 1999, Terry Lambert wrote: > > Since he appears to have a way in, other than the console, he > should: > > 1) Cause the lockup > > 2) Go in and do a 'ps' to see if conflicting processes are > running > > 3) Force a console switch via escape sequence to a console > vty not currently in use. This should result in the X > server disengaging, and the input devices being detached > and reattached to the new vty, if the problem is merely > a race. > > Also, if both terminals are reading input, as has been posited, > a ctrl-alt-F2 should still result in a console switch, regardless > of who takes the keypress event (ctrl and alt are modifiers, so > only the F2 keycode should be relevent). > > > Terry Lambert > terry@lambert.org > --- > Any opinions in this posting are my own and not those of my present > or previous employers. > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 13:25:45 1999 Delivered-To: freebsd-smp@freebsd.org Received: from verdi.nethelp.no (verdi.nethelp.no [158.36.41.162]) by hub.freebsd.org (Postfix) with SMTP id 13D2414C08 for ; Thu, 22 Jul 1999 13:25:38 -0700 (PDT) (envelope-from sthaug@nethelp.no) Received: (qmail 57421 invoked by uid 1001); 22 Jul 1999 20:24:57 +0000 (GMT) To: freebsd-hackers@freebsd.org, freebsd-smp@freebsd.org Subject: rndcontrol and SMP From: sthaug@nethelp.no X-Mailer: Mew version 1.05+ on Emacs 19.34.2 Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Date: Thu, 22 Jul 1999 22:24:57 +0200 Message-ID: <57419.932675097@verdi.nethelp.no> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org rndcontrol doesn't work very well for SMP systems. I have a system here with IRQs 16 and 18 for Ethernet and SCSI: fxp0: rev 0x05 int a irq 18 on pci0.10.0 ahc0: rev 0x00 int a irq 16 on pci0.12.0 and I'd like to use these with rndcontrol. However, the ioctl chokes on IRQ >= 16. From i386/i386/mem.c: /* * XXX the data is 16-bit due to a historical botch, so we use * magic 16's instead of ICU_LEN and can't support 24 interrupts * under SMP. */ intr = *(int16_t *)data; if (cmd != MEM_RETURNIRQ && (intr < 0 || intr >= 16)) return (EINVAL); What is needed to make this support a more sensible number of IRQs? Also, rndcontrol naturally returns an error message, which could have been better: rndcontrol: rndcontrol: Invalid argument rndcontrol uses warn() with argv[0] as the argument - but warn() is documented to always print the program name. So it gets doubled. Below is a patch against 3.2-STABLE to make it slightly more intelligent, so we get an error message like this instead: rndcontrol: setting irq 16: Invalid argument Steinar Haug, Nethelp consulting, sthaug@nethelp.no ---------------------------------------------------------------------- *** rndcontrol.c.orig Mon Oct 13 13:08:47 1997 --- rndcontrol.c Thu Jul 22 22:06:52 1999 *************** *** 76,82 **** printf("%s: setting irq %d\n", argv[0], irq); result = ioctl(fd, MEM_SETIRQ, (char *)&irq); if (result == -1) { ! warn("%s", argv[0]); return (1); } break; --- 76,82 ---- printf("%s: setting irq %d\n", argv[0], irq); result = ioctl(fd, MEM_SETIRQ, (char *)&irq); if (result == -1) { ! warn("setting irq %d", irq); return (1); } break; *************** *** 86,92 **** printf("%s: clearing irq %d\n", argv[0], irq); result = ioctl(fd, MEM_CLEARIRQ, (char *)&irq); if (result == -1) { ! warn("%s", argv[0]); return (1); } break; --- 86,92 ---- printf("%s: clearing irq %d\n", argv[0], irq); result = ioctl(fd, MEM_CLEARIRQ, (char *)&irq); if (result == -1) { ! warn("clearing irq %d", irq); return (1); } break; *************** *** 98,104 **** if (verbose) { result = ioctl(fd, MEM_RETURNIRQ, (char *)&irq); if (result == -1) { ! warn("%s", argv[0]); return (1); } printf("%s: interrupts in use:", argv[0]); --- 98,104 ---- if (verbose) { result = ioctl(fd, MEM_RETURNIRQ, (char *)&irq); if (result == -1) { ! warn(""); return (1); } printf("%s: interrupts in use:", argv[0]); To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 14: 0: 4 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mail1.its.rpi.edu (mail1.its.rpi.edu [128.113.100.7]) by hub.freebsd.org (Postfix) with ESMTP id 418E714BD6 for ; Thu, 22 Jul 1999 13:59:02 -0700 (PDT) (envelope-from drosih@rpi.edu) Received: from [128.113.24.47] (gilead.acs.rpi.edu [128.113.24.47]) by mail1.its.rpi.edu (8.8.8/8.8.8) with ESMTP id QAA29752; Thu, 22 Jul 1999 16:57:59 -0400 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Sender: drosih@mail.rpi.edu Message-Id: In-Reply-To: <199907221725.LAA02645@mt.sri.com> References: <199907221634.JAA51919@ix.netcom.com> Date: Thu, 22 Jul 1999 16:57:59 -0400 To: Nate Williams , Richard Scheper From: Garance A Drosihn Subject: Re: SMP + XDM = keyboard lockup Cc: freebsd-smp@FreeBSD.ORG Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 11:25 AM -0600 7/22/99, Nate Williams wrote: >> Also, I don't understand why initializing xdm in /etc/ttys >> doesn't work since by then init should have executed getty >> on the appropriate ttvs. > > It *should* work. I'm not sure why it's not working, unless > there are no free ttys for it to work on. This probably doesn't help much, but I'm running xdm on my SMP system without any apparent trouble. I just turned the entry for XDM in '/etc/ttys' from off to on. Dual-PPro system, running 3.stable as of a little earlier this week. I built it JUST before the new XFree86 was released, so I don't have the latest XFree. I've run freebsd quite awhile without installing X. When I first installed and configured it, I tried to start it up by typing "X" instead of xdm. *That* locked up my system in a way that sounds somewhat like what is being suggested here. Typing at the keyboard did not do anything. I had to slogin from another machine, kill that X process, and then put my brain in gear to remember how I was SUPPOSED to startup X... I haven't done a lot with X on this system, but everything I've done has worked fine. I've rebooted it a few dozen times without it ever locking up. --- Garance Alistair Drosehn = gad@eclipse.acs.rpi.edu Senior Systems Programmer or drosih@rpi.edu Rensselaer Polytechnic Institute To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 14: 1:22 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mail1.its.rpi.edu (mail1.its.rpi.edu [128.113.100.7]) by hub.freebsd.org (Postfix) with ESMTP id DB032155AF for ; Thu, 22 Jul 1999 14:01:11 -0700 (PDT) (envelope-from drosih@rpi.edu) Received: from [128.113.24.47] (gilead.acs.rpi.edu [128.113.24.47]) by mail1.its.rpi.edu (8.8.8/8.8.8) with ESMTP id RAA24728; Thu, 22 Jul 1999 17:00:55 -0400 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Sender: drosih@mail.rpi.edu Message-Id: In-Reply-To: References: <199907221839.LAA10550@usr05.primenet.com> Date: Thu, 22 Jul 1999 17:00:55 -0400 To: Richard Scheper , freebsd-smp@FreeBSD.ORG From: Garance A Drosihn Subject: Re: SMP + XDM = keyboard lockup Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 3:16 PM -0400 7/22/99, Richard Scheper wrote: > Actually, I don't have another way in, just the console. If I > cause the lockup I have no recourse but to turn the computer > off. It makes it tough to get info, I know. I never see any > error messages in /var/log/messages Early in the login process, start some dumb script which just loops around dumping the output of 'ps' into some file (with a 'sleep' in the loop, of course!). If you're a bit more daring, you could also have this other script kill off your xdm task after a few minutes, which should give you control back. --- Garance Alistair Drosehn = gad@eclipse.acs.rpi.edu Senior Systems Programmer or drosih@rpi.edu Rensselaer Polytechnic Institute To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 15: 4:59 1999 Delivered-To: freebsd-smp@freebsd.org Received: from postoffice.imsa.edu (cappio.imsa.edu [143.195.1.6]) by hub.freebsd.org (Postfix) with ESMTP id 401DB15609 for ; Thu, 22 Jul 1999 15:04:38 -0700 (PDT) (envelope-from unixboy@imsa.edu) Received: from pepsi.imsa.edu (pepsi [143.195.1.5]) by postoffice.imsa.edu (8.9.1a/8.9.1) with ESMTP id RAA29545 for ; Thu, 22 Jul 1999 17:03:05 -0500 (CDT) Received: from localhost (unixboy@localhost) by pepsi.imsa.edu (8.8.8+Sun/8.8.5) with SMTP id RAA00069 for ; Thu, 22 Jul 1999 17:03:22 -0500 (CDT) X-Authentication-Warning: pepsi.imsa.edu: unixboy owned process doing -bs Date: Thu, 22 Jul 1999 17:03:22 -0500 (CDT) From: "Christopher W. Banek" X-Sender: unixboy@pepsi To: freebsd-smp@freebsd.org Subject: Page Fault Problems with SMP kernel in 3.2-RELEASE Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi. I am a a bit new to FreeBSD, came over from linux, but here is what is happening. I installed 3.2-RELEASE and everything went fine. I compilied an SMP kernel and now it page faults when ever I am doing something system intensive. It happens during configure scripts, but not when I do a make. But if I do a make -j2, it crashes quickly. Oh yeah, I have a GA-686LX2 with Dual PII/333 and 128MB of Ram. Here is the exact error message. Fatal trap 12: page fault while in kernel mode mp_lock = 01000002; cpuid=1; lapic.id=01000000 fault virtual address = 0x2b fault code = supervisor read, page not present instruction pointer = 0x8:0xc01cb3ab stack pointer = 0x10:0xc4a41e40 frame pointer = 0x10:0xc4a41e4c code segment = base 0x0, limit 0xfffff, type ox1b = DPL 0, pres1, def32, gran 1 processor eflags - interrupt enabled, resume,IOPL = 0 current process = 2800 (sh) inturrupt mask = net tty bio cam <- SMP: XXX trap number = 12 panic: page fault mp_lock = 01000002; cpuid = 1; lapic.id = 01000000 boot() called on cpu#1 Thanks for any help you guys can provide. Christopher Banek To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 15:21:49 1999 Delivered-To: freebsd-smp@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2]) by hub.freebsd.org (Postfix) with ESMTP id 71BA915616 for ; Thu, 22 Jul 1999 15:21:25 -0700 (PDT) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.9.3/8.9.1) id PAA88210; Thu, 22 Jul 1999 15:20:22 -0700 (PDT) (envelope-from dillon) Date: Thu, 22 Jul 1999 15:20:22 -0700 (PDT) From: Matthew Dillon Message-Id: <199907222220.PAA88210@apollo.backplane.com> To: "Christopher W. Banek" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE References: Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org :Hi. I am a a bit new to FreeBSD, came over from linux, but here is what :is happening. I installed 3.2-RELEASE and everything went fine. I :compilied an SMP kernel and now it page faults when ever I am doing :something system intensive. It happens during configure scripts, but not :when I do a make. But if I do a make -j2, it crashes quickly. : :Oh yeah, I have a GA-686LX2 with Dual PII/333 and 128MB of Ram. There's a good chance the problem has been fixed in FreeBSD-STABLE (i.e. changes made since the 3.2 release), but it's hard to tell from the information you provided. There are two ways to get additional information: * Try to get the kernel to generate a kernel core dump This involves having a swap area that is at least as large as main memory. Setting the dumpdev in /etc/rc.conf correctly. If you have IDE disks it might be something like I show below. Note that you have to be careful to specify the correct partition name. Swap partitions almost always end with 'b'. dumpdev="/dev/wd0b" Ensuring that /var/crash exists and has enough free space to hold the dump (at least as much space as you have main memory, plus a couple of megabytes more). You can make /var/crash a softlink to some other partition if /var doesn't have enough space. For example you can create a /usr/crash and then remove /var/crash and create a softlink 'ln -s /usr/crash /var/crash'. Once these changes are made and you reboot, the next time the machine crashes it should be able to generate a kernel core. You can then gdb the core like this: cd /var/crash ls -la gdb -k kernel.0 vmcore.0 (or whatever the names wind up being) back (to get a backtrace) * Recompile the kernel with the DDB option turned on. Now when the kernel crashes it will drop into a DDB> prompt and you should be able to do a 'trace' to get a call stack backtrace from which we may be able to get more information. (Then type 'panic' and hit return a couple of times to panic the system). Recompiling the kernel for someone who has never done it before can take a bit of work. If the problem has already been solved you will have to learn how to download the latest source dist in order to recompile a new kernel. Take a look at the FreeBSD documentation, especially the handbook. See http://www.freebsd.org/ -Matt Matthew Dillon :Here is the exact error message. : :Fatal trap 12: page fault while in kernel mode :mp_lock = 01000002; cpuid=1; lapic.id=01000000 :fault virtual address = 0x2b :fault code = supervisor read, page not present :instruction pointer = 0x8:0xc01cb3ab :stack pointer = 0x10:0xc4a41e40 :frame pointer = 0x10:0xc4a41e4c :code segment = base 0x0, limit 0xfffff, type ox1b : = DPL 0, pres1, def32, gran 1 :processor eflags - interrupt enabled, resume,IOPL = 0 :current process = 2800 (sh) :inturrupt mask = net tty bio cam <- SMP: XXX :trap number = 12 :panic: page fault :mp_lock = 01000002; cpuid = 1; lapic.id = 01000000 :boot() called on cpu#1 : : :Thanks for any help you guys can provide. :Christopher Banek : : : : :To Unsubscribe: send mail to majordomo@FreeBSD.org :with "unsubscribe freebsd-smp" in the body of the message : To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 16: 0:22 1999 Delivered-To: freebsd-smp@freebsd.org Received: from postoffice.imsa.edu (cappio.imsa.edu [143.195.1.6]) by hub.freebsd.org (Postfix) with ESMTP id D212A14D19 for ; Thu, 22 Jul 1999 16:00:01 -0700 (PDT) (envelope-from unixboy@imsa.edu) Received: from coke.imsa.edu (coke [143.195.1.4]) by postoffice.imsa.edu (8.9.1a/8.9.1) with ESMTP id RAA02720; Thu, 22 Jul 1999 17:57:07 -0500 (CDT) Received: from localhost (unixboy@localhost) by coke.imsa.edu (8.8.8+Sun/8.8.5) with SMTP id RAA28499; Thu, 22 Jul 1999 17:57:24 -0500 (CDT) X-Authentication-Warning: coke.imsa.edu: unixboy owned process doing -bs Date: Thu, 22 Jul 1999 17:57:23 -0500 (CDT) From: "Christopher W. Banek" X-Sender: unixboy@coke To: Matthew Dillon Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE In-Reply-To: <199907222220.PAA88210@apollo.backplane.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, 22 Jul 1999, Matthew Dillon wrote: > :Hi. I am a a bit new to FreeBSD, came over from linux, but here is what > :is happening. I installed 3.2-RELEASE and everything went fine. I > :compilied an SMP kernel and now it page faults when ever I am doing > :something system intensive. It happens during configure scripts, but not > :when I do a make. But if I do a make -j2, it crashes quickly. > : > :Oh yeah, I have a GA-686LX2 with Dual PII/333 and 128MB of Ram. > > There's a good chance the problem has been fixed in FreeBSD-STABLE > (i.e. changes made since the 3.2 release), but it's hard to tell > from the information you provided. > > There are two ways to get additional information: > > * Try to get the kernel to generate a kernel core dump > > This involves having a swap area that is at least as large as > main memory. > > Setting the dumpdev in /etc/rc.conf correctly. If you have IDE > disks it might be something like I show below. Note that you > have to be careful to specify the correct partition name. Swap > partitions almost always end with 'b'. > > dumpdev="/dev/wd0b" > > Ensuring that /var/crash exists and has enough free space to hold > the dump (at least as much space as you have main memory, plus a > couple of megabytes more). You can make /var/crash a softlink to > some other partition if /var doesn't have enough space. For example > you can create a /usr/crash and then remove /var/crash and create > a softlink 'ln -s /usr/crash /var/crash'. > > Once these changes are made and you reboot, the next time the machine > crashes it should be able to generate a kernel core. You can then > gdb the core like this: > > cd /var/crash > ls -la > gdb -k kernel.0 vmcore.0 (or whatever > the names wind up being) > back (to get a backtrace) > > * Recompile the kernel with the DDB option turned on. Now when the > kernel crashes it will drop into a DDB> prompt and you should be > able to do a 'trace' to get a call stack backtrace from which we > may be able to get more information. (Then type 'panic' and hit > return a couple of times to panic the system). > > Recompiling the kernel for someone who has never done it before > can take a bit of work. > > > If the problem has already been solved you will have to learn how to > download the latest source dist in order to recompile a new kernel. > > Take a look at the FreeBSD documentation, especially the handbook. > See http://www.freebsd.org/ > > -Matt > Matthew Dillon > > > :Here is the exact error message. > : > :Fatal trap 12: page fault while in kernel mode > :mp_lock = 01000002; cpuid=1; lapic.id=01000000 > :fault virtual address = 0x2b > :fault code = supervisor read, page not present > :instruction pointer = 0x8:0xc01cb3ab > :stack pointer = 0x10:0xc4a41e40 > :frame pointer = 0x10:0xc4a41e4c > :code segment = base 0x0, limit 0xfffff, type ox1b > : = DPL 0, pres1, def32, gran 1 > :processor eflags - interrupt enabled, resume,IOPL = 0 > :current process = 2800 (sh) > :inturrupt mask = net tty bio cam <- SMP: XXX > :trap number = 12 > :panic: page fault > :mp_lock = 01000002; cpuid = 1; lapic.id = 01000000 > :boot() called on cpu#1 > : > : > :Thanks for any help you guys can provide. > :Christopher Banek > : > : > : > : > :To Unsubscribe: send mail to majordomo@FreeBSD.org > :with "unsubscribe freebsd-smp" in the body of the message > : > I have done all of the things that you said, but when I recompiled the kernel with debugger support enabled, now it doesn't page fault, it just freezes with no messages at all. And no core dumps are being made. I am totally clueless about this one. Christopher Banek To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 16:18:17 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mail.rdc1.sfba.home.com (ha1.rdc1.sfba.home.com [24.0.0.66]) by hub.freebsd.org (Postfix) with ESMTP id 81DCB14C9D for ; Thu, 22 Jul 1999 16:18:14 -0700 (PDT) (envelope-from adsharma@c62443-a.frmt1.sfba.home.com) Received: from c62443-a.frmt1.sfba.home.com ([24.0.69.165]) by mail.rdc1.sfba.home.com (InterMail v4.01.01.00 201-229-111) with ESMTP id <19990722231739.QTKM8807.mail.rdc1.sfba.home.com@c62443-a.frmt1.sfba.home.com>; Thu, 22 Jul 1999 16:17:39 -0700 Received: (from adsharma@localhost) by c62443-a.frmt1.sfba.home.com (8.8.7/8.8.7) id QAA21737; Thu, 22 Jul 1999 16:17:39 -0700 Date: Thu, 22 Jul 1999 16:17:39 -0700 From: Arun Sharma To: "Christopher W. Banek" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE Message-ID: <19990722161739.A21710@home.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.5i In-Reply-To: ; from Christopher W. Banek on Thu, Jul 22, 1999 at 05:03:22PM -0500 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, Jul 22, 1999 at 05:03:22PM -0500, Christopher W. Banek wrote: > fault virtual address = 0x2b > fault code = supervisor read, page not present > instruction pointer = 0x8:0xc01cb3ab Can you just disassemble a few instructions at the above IP and convert it to a func+offset using the symbol table and post it here ? Just curious and I don't have access to a 3.2 system. -Arun To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 16:25:13 1999 Delivered-To: freebsd-smp@freebsd.org Received: from postoffice.imsa.edu (cappio.imsa.edu [143.195.1.6]) by hub.freebsd.org (Postfix) with ESMTP id E1ADC1563B for ; Thu, 22 Jul 1999 16:25:09 -0700 (PDT) (envelope-from unixboy@imsa.edu) Received: from coke.imsa.edu (coke [143.195.1.4]) by postoffice.imsa.edu (8.9.1a/8.9.1) with ESMTP id SAA04364; Thu, 22 Jul 1999 18:24:49 -0500 (CDT) Received: from localhost (unixboy@localhost) by coke.imsa.edu (8.8.8+Sun/8.8.5) with SMTP id SAA29967; Thu, 22 Jul 1999 18:25:08 -0500 (CDT) X-Authentication-Warning: coke.imsa.edu: unixboy owned process doing -bs Date: Thu, 22 Jul 1999 18:25:07 -0500 (CDT) From: "Christopher W. Banek" X-Sender: unixboy@coke To: Arun Sharma Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE In-Reply-To: <19990722161739.A21710@home.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I am confused about that one, I don't really know how to do that... Some instructions would be helpful. On Thu, 22 Jul 1999, Arun Sharma wrote: > On Thu, Jul 22, 1999 at 05:03:22PM -0500, Christopher W. Banek wrote: > > fault virtual address = 0x2b > > fault code = supervisor read, page not present > > instruction pointer = 0x8:0xc01cb3ab > > Can you just disassemble a few instructions at the above IP and convert > it to a func+offset using the symbol table and post it here ? > > Just curious and I don't have access to a 3.2 system. > > -Arun > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 16:35:11 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mail.rdc1.sfba.home.com (ha1.rdc1.sfba.home.com [24.0.0.66]) by hub.freebsd.org (Postfix) with ESMTP id 0F4E51563B for ; Thu, 22 Jul 1999 16:35:06 -0700 (PDT) (envelope-from adsharma@c62443-a.frmt1.sfba.home.com) Received: from c62443-a.frmt1.sfba.home.com ([24.0.69.165]) by mail.rdc1.sfba.home.com (InterMail v4.01.01.00 201-229-111) with ESMTP id <19990722233244.QYFA8807.mail.rdc1.sfba.home.com@c62443-a.frmt1.sfba.home.com>; Thu, 22 Jul 1999 16:32:44 -0700 Received: (from adsharma@localhost) by c62443-a.frmt1.sfba.home.com (8.8.7/8.8.7) id QAA21804; Thu, 22 Jul 1999 16:32:44 -0700 Date: Thu, 22 Jul 1999 16:32:44 -0700 From: Arun Sharma To: "Christopher W. Banek" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE Message-ID: <19990722163244.A21793@home.com> References: <19990722161739.A21710@home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.5i In-Reply-To: ; from Christopher W. Banek on Thu, Jul 22, 1999 at 06:25:07PM -0500 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, Jul 22, 1999 at 06:25:07PM -0500, Christopher W. Banek wrote: > I am confused about that one, I don't really know how to do that... Some > instructions would be helpful. gdb /kernel (gdb) disassemble 0xc01cb3ab If your kernel was different from /kernel, use the appropriate name. -Arun To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 16:36: 7 1999 Delivered-To: freebsd-smp@freebsd.org Received: from postoffice.imsa.edu (cappio.imsa.edu [143.195.1.6]) by hub.freebsd.org (Postfix) with ESMTP id 509CF15660 for ; Thu, 22 Jul 1999 16:36:04 -0700 (PDT) (envelope-from unixboy@imsa.edu) Received: from coke.imsa.edu (coke [143.195.1.4]) by postoffice.imsa.edu (8.9.1a/8.9.1) with ESMTP id SAA04967; Thu, 22 Jul 1999 18:35:01 -0500 (CDT) Received: from localhost (unixboy@localhost) by coke.imsa.edu (8.8.8+Sun/8.8.5) with SMTP id SAA00630; Thu, 22 Jul 1999 18:35:20 -0500 (CDT) X-Authentication-Warning: coke.imsa.edu: unixboy owned process doing -bs Date: Thu, 22 Jul 1999 18:35:19 -0500 (CDT) From: "Christopher W. Banek" X-Sender: unixboy@coke To: Arun Sharma Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE In-Reply-To: <19990722163244.A21793@home.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org The problem with that is, I am not getting any core dumps since I compiled with ddb support in the kernel, it just freezes the entire machine. Christopher Banek To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 16:42:14 1999 Delivered-To: freebsd-smp@freebsd.org Received: from pop3-3.enteract.com (pop3-3.enteract.com [207.229.143.32]) by hub.freebsd.org (Postfix) with SMTP id C954414E3C for ; Thu, 22 Jul 1999 16:42:10 -0700 (PDT) (envelope-from dscheidt@tumbolia.com) Received: (qmail 97505 invoked from network); 22 Jul 1999 23:42:00 -0000 Received: from shell-3.enteract.com (dscheidt@207.229.143.42) by pop3-3.enteract.com with SMTP; 22 Jul 1999 23:42:00 -0000 Date: Thu, 22 Jul 1999 18:41:59 -0500 (CDT) From: David Scheidt X-Sender: dscheidt@shell-3.enteract.com To: "Christopher W. Banek" Cc: Arun Sharma , freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, 22 Jul 1999, Christopher W. Banek wrote: > The problem with that is, I am not getting any core dumps since I compiled > with ddb support in the kernel, it just freezes the entire machine. > Are you running X? DDB doesn't play well with X. (At least you can't drop into it from X. I have 't ever had a panic while running both.) Don't start X, and you should get the db> prompt when it panics. If you aren't running X, and the machien is just hanging, try ctrl-alt-esc to break into the debugger. David To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 16:47:27 1999 Delivered-To: freebsd-smp@freebsd.org Received: from postoffice.imsa.edu (cappio.imsa.edu [143.195.1.6]) by hub.freebsd.org (Postfix) with ESMTP id B162F14E3C for ; Thu, 22 Jul 1999 16:47:21 -0700 (PDT) (envelope-from unixboy@imsa.edu) Received: from coke.imsa.edu (coke [143.195.1.4]) by postoffice.imsa.edu (8.9.1a/8.9.1) with ESMTP id SAA05600 for ; Thu, 22 Jul 1999 18:44:37 -0500 (CDT) Received: from localhost (unixboy@localhost) by coke.imsa.edu (8.8.8+Sun/8.8.5) with SMTP id SAA01123 for ; Thu, 22 Jul 1999 18:44:56 -0500 (CDT) X-Authentication-Warning: coke.imsa.edu: unixboy owned process doing -bs Date: Thu, 22 Jul 1999 18:44:56 -0500 (CDT) From: "Christopher W. Banek" X-Sender: unixboy@coke To: freebsd-smp@freebsd.org Subject: page fault in SMP Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org When I changed the MP comp. setting in the bios from 1.4 to 1.1, I get page fault error messages again. I was able to drop into the debugger and trace it. Here is the output: pmap_remove_pages(c5e06924,0,bfbfe000,c5e02520,c022b5d0) at pmap_remove_pages+0xc8 exit1(c5e02520,0,c5e86fb4,c01ef3d7,c5e02520) at exit1+0x1d3 exit1(c5e02520,c5386f94,80a6a88,1,80a5000) at exit1 syscall(27,27,80a5000,1,bfbfd8e8) at syscall+0x187 Xint0x80_syscall() at Xint0x80syscall+0x4c Anything else I should do in the debugger? To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 17: 5:33 1999 Delivered-To: freebsd-smp@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2]) by hub.freebsd.org (Postfix) with ESMTP id 304E414FEF for ; Thu, 22 Jul 1999 17:05:31 -0700 (PDT) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.9.3/8.9.1) id RAA32201; Thu, 22 Jul 1999 17:03:29 -0700 (PDT) (envelope-from dillon) Date: Thu, 22 Jul 1999 17:03:29 -0700 (PDT) From: Matthew Dillon Message-Id: <199907230003.RAA32201@apollo.backplane.com> To: "Christopher W. Banek" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: page fault in SMP References: Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org :When I changed the MP comp. setting in the bios from 1.4 to 1.1, I get :page fault error messages again. I was able to drop into the debugger and :trace it. : :Here is the output: : :pmap_remove_pages(c5e06924,0,bfbfe000,c5e02520,c022b5d0) at :pmap_remove_pages+0xc8 :exit1(c5e02520,0,c5e86fb4,c01ef3d7,c5e02520) at exit1+0x1d3 :exit1(c5e02520,c5386f94,80a6a88,1,80a5000) at exit1 :syscall(27,27,80a5000,1,bfbfd8e8) at syscall+0x187 :Xint0x80_syscall() at Xint0x80syscall+0x4c : :Anything else I should do in the debugger? I was hoping it was set at 1.1 and you changed it to 1.4, but if it was set at 1.4 then that is the correct MP setting. There is something very basic going wrong for it to crash where it is crashing. I think your best bet may be to try to upgrade the system from 3.2-RELEASE to 3.2-STABLE and see where you stand after that. -Matt Matthew Dillon To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 17:12:29 1999 Delivered-To: freebsd-smp@freebsd.org Received: from alpo.whistle.com (alpo.whistle.com [207.76.204.38]) by hub.freebsd.org (Postfix) with ESMTP id 65A5414FC9 for ; Thu, 22 Jul 1999 17:12:16 -0700 (PDT) (envelope-from julian@whistle.com) Received: from current1.whistle.com (current1.whistle.com [207.76.205.22]) by alpo.whistle.com (8.9.1a/8.9.1) with SMTP id RAA19379; Thu, 22 Jul 1999 17:11:25 -0700 (PDT) Date: Thu, 22 Jul 1999 17:11:24 -0700 (PDT) From: Julian Elischer To: "Christopher W. Banek" Cc: Arun Sharma , freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org If you are in X then it's on the console and you are not...... On Thu, 22 Jul 1999, Christopher W. Banek wrote: > The problem with that is, I am not getting any core dumps since I compiled > with ddb support in the kernel, it just freezes the entire machine. > > Christopher Banek > > > > > > > To Unsubscribe: send mail to majordomo@FreeBSD.org > with "unsubscribe freebsd-smp" in the body of the message > To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 17:23: 5 1999 Delivered-To: freebsd-smp@freebsd.org Received: from postoffice.imsa.edu (cappio.imsa.edu [143.195.1.6]) by hub.freebsd.org (Postfix) with ESMTP id 0929C14C27 for ; Thu, 22 Jul 1999 17:23:02 -0700 (PDT) (envelope-from unixboy@imsa.edu) Received: from coke.imsa.edu (coke [143.195.1.4]) by postoffice.imsa.edu (8.9.1a/8.9.1) with ESMTP id TAA07312; Thu, 22 Jul 1999 19:19:39 -0500 (CDT) Received: from localhost (unixboy@localhost) by coke.imsa.edu (8.8.8+Sun/8.8.5) with SMTP id TAA02741; Thu, 22 Jul 1999 19:19:58 -0500 (CDT) X-Authentication-Warning: coke.imsa.edu: unixboy owned process doing -bs Date: Thu, 22 Jul 1999 19:19:57 -0500 (CDT) From: "Christopher W. Banek" X-Sender: unixboy@coke To: Julian Elischer Cc: Arun Sharma , freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I was never in X. Chris Bane To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 17:37:40 1999 Delivered-To: freebsd-smp@freebsd.org Received: from postoffice.imsa.edu (cappio.imsa.edu [143.195.1.6]) by hub.freebsd.org (Postfix) with ESMTP id 4F82614E88 for ; Thu, 22 Jul 1999 17:37:33 -0700 (PDT) (envelope-from unixboy@imsa.edu) Received: from coke.imsa.edu (coke [143.195.1.4]) by postoffice.imsa.edu (8.9.1a/8.9.1) with ESMTP id TAA08103 for ; Thu, 22 Jul 1999 19:37:01 -0500 (CDT) Received: from localhost (unixboy@localhost) by coke.imsa.edu (8.8.8+Sun/8.8.5) with SMTP id TAA03543 for ; Thu, 22 Jul 1999 19:37:20 -0500 (CDT) X-Authentication-Warning: coke.imsa.edu: unixboy owned process doing -bs Date: Thu, 22 Jul 1999 19:37:19 -0500 (CDT) From: "Christopher W. Banek" X-Sender: unixboy@coke To: freebsd-smp@freebsd.org Subject: More stuff on Page faults. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Well, I downloaded the 3.2-STABLE kernel sources and compiled it. It compiled fine, but I am stiff having the same problems that I did before. The error is the same. It is stopping at pmap_remove_pages+0xc8 : decl 0(%ecx) when I trace it here is the output: pmap_remove_pages(c5e069e4,0,bfbfe000,2,c5423034) at pmap_remove_pages+0xc8 exec_new_vmspace(c5e82e94,2,c5e82,9,c5e02680,0) at exec_new_vmspace+0x40 exec_elf_imgact(c5e82e94,c5e02680,c022b7a0,0,2 at exec_elf_imgact+0xe8 execve(c5e02680,c5e82f94,80a68d8,ffffffff,80a6934) at execve+0x21f syscall(27,27,80a6934,ffffffff,bfbfd964) at syscall +0x187 Xint0x80_syscall() at Xint0x80_syscall+0x4c Any ideas now? Chris Banke To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 18:30:30 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mail.rdc1.sfba.home.com (ha1.rdc1.sfba.home.com [24.0.0.66]) by hub.freebsd.org (Postfix) with ESMTP id 4928514C57 for ; Thu, 22 Jul 1999 18:30:26 -0700 (PDT) (envelope-from adsharma@c62443-a.frmt1.sfba.home.com) Received: from c62443-a.frmt1.sfba.home.com ([24.0.69.165]) by mail.rdc1.sfba.home.com (InterMail v4.01.01.00 201-229-111) with ESMTP id <19990723012737.SEHI8807.mail.rdc1.sfba.home.com@c62443-a.frmt1.sfba.home.com>; Thu, 22 Jul 1999 18:27:37 -0700 Received: (from adsharma@localhost) by c62443-a.frmt1.sfba.home.com (8.8.7/8.8.7) id SAA21993; Thu, 22 Jul 1999 18:27:35 -0700 Date: Thu, 22 Jul 1999 18:27:35 -0700 From: Arun Sharma To: "Christopher W. Banek" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Page Fault Problems with SMP kernel in 3.2-RELEASE Message-ID: <19990722182735.A21983@home.com> References: <19990722163244.A21793@home.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.5i In-Reply-To: ; from Christopher W. Banek on Thu, Jul 22, 1999 at 06:35:19PM -0500 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Thu, Jul 22, 1999 at 06:35:19PM -0500, Christopher W. Banek wrote: > The problem with that is, I am not getting any core dumps since I compiled > with ddb support in the kernel, it just freezes the entire machine. You don't need a core dump. Just use /kernel. -Arun To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 18:37:17 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp04.primenet.com (smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (Postfix) with ESMTP id 2961B156A8 for ; Thu, 22 Jul 1999 18:36:34 -0700 (PDT) (envelope-from tlambert@usr09.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id SAA16856; Thu, 22 Jul 1999 18:35:23 -0700 (MST) Received: from usr09.primenet.com(206.165.6.209) via SMTP by smtp04.primenet.com, id smtpd016813; Thu Jul 22 18:35:16 1999 Received: (from tlambert@localhost) by usr09.primenet.com (8.8.5/8.8.5) id SAA07594; Thu, 22 Jul 1999 18:35:14 -0700 (MST) From: Terry Lambert Message-Id: <199907230135.SAA07594@usr09.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: drosih@rpi.edu (Garance A Drosihn) Date: Fri, 23 Jul 1999 01:35:13 +0000 (GMT) Cc: nate@mt.sri.com, scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG In-Reply-To: from "Garance A Drosihn" at Jul 22, 99 04:57:59 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > At 11:25 AM -0600 7/22/99, Nate Williams wrote: > >> Also, I don't understand why initializing xdm in /etc/ttys > >> doesn't work since by then init should have executed getty > >> on the appropriate ttvs. > > > > It *should* work. I'm not sure why it's not working, unless > > there are no free ttys for it to work on. > > This probably doesn't help much, but I'm running xdm on my > SMP system without any apparent trouble. I just turned the > entry for XDM in '/etc/ttys' from off to on. Dual-PPro > system, running 3.stable as of a little earlier this week. > I built it JUST before the new XFree86 was released, so I > don't have the latest XFree. Me too. I really, really think this is a hardware specific problem, at the I/O bus contention level. > I've run freebsd quite awhile without installing X. When > I first installed and configured it, I tried to start it > up by typing "X" instead of xdm. *That* locked up my > system in a way that sounds somewhat like what is being > suggested here. Typing at the keyboard did not do anything. > I had to slogin from another machine, kill that X process, > and then put my brain in gear to remember how I was SUPPOSED > to startup X... I never get lockups, but I am running an old Neptune-based dual P90 (it was quite something at the time I bought it to do FreeBSD SMP hacking using Jack Vogel's original SMP work from October of 1995). > I haven't done a lot with X on this system, but everything > I've done has worked fine. I've rebooted it a few dozen > times without it ever locking up. As an experiment, really, really I/O load your system (I think that serial I/O might be the modet I/O bus intensive), and then bounce back and forth between the console in X and a text console, and see if things lock up. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 18:37:18 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp01.primenet.com (smtp01.primenet.com [206.165.6.131]) by hub.freebsd.org (Postfix) with ESMTP id E352414C8D for ; Thu, 22 Jul 1999 18:36:51 -0700 (PDT) (envelope-from tlambert@usr09.primenet.com) Received: (from daemon@localhost) by smtp01.primenet.com (8.8.8/8.8.8) id SAA23066; Thu, 22 Jul 1999 18:36:46 -0700 (MST) Received: from usr09.primenet.com(206.165.6.209) via SMTP by smtp01.primenet.com, id smtpd023041; Thu Jul 22 18:36:44 1999 Received: (from tlambert@localhost) by usr09.primenet.com (8.8.5/8.8.5) id SAA07637; Thu, 22 Jul 1999 18:36:44 -0700 (MST) From: Terry Lambert Message-Id: <199907230136.SAA07637@usr09.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: drosih@rpi.edu (Garance A Drosihn) Date: Fri, 23 Jul 1999 01:36:43 +0000 (GMT) Cc: scheper@beast.toad.net, freebsd-smp@FreeBSD.ORG In-Reply-To: from "Garance A Drosihn" at Jul 22, 99 05:00:55 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > At 3:16 PM -0400 7/22/99, Richard Scheper wrote: > > Actually, I don't have another way in, just the console. If I > > cause the lockup I have no recourse but to turn the computer > > off. It makes it tough to get info, I know. I never see any > > error messages in /var/log/messages > > Early in the login process, start some dumb script which just > loops around dumping the output of 'ps' into some file (with > a 'sleep' in the loop, of course!). And a "sync" or three before the sleep, in case you have to BRS it. > If you're a bit more > daring, you could also have this other script kill off your > xdm task after a few minutes, which should give you control > back. Depending on the actual source of the lockup... Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 18:37:21 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mercury.toad.net (mercury.toad.net [205.197.182.105]) by hub.freebsd.org (Postfix) with ESMTP id 97845156DA for ; Thu, 22 Jul 1999 18:36:53 -0700 (PDT) (envelope-from scheper@toad.net) Received: from toad.net (core10d30.dynamic-dialup.toad.net [209.150.114.30]) by mercury.toad.net (8.8.8/8.8.8) with ESMTP id VAA20936 for ; Thu, 22 Jul 1999 21:36:53 -0400 Message-ID: <3797C66E.B9382583@toad.net> Date: Thu, 22 Jul 1999 21:33:34 -0400 From: "Richard A. Scheper" X-Mailer: Mozilla 4.08 [en] (X11; I; FreeBSD 3.1-RELEASE i386) MIME-Version: 1.0 To: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup References: <199907221845.LAA10738@usr05.primenet.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org It looks like your suspiscions are correct. After further attempts at starting xdm from a root seesion, it does in fact hang more often than not. Perhaps I was just "lucky" and it only worked with a quiet system. I'll try the ps test and see what I get -Rich Terry Lambert wrote: > > Wow! That's a lot to chew and mostly over my head, but a couple of > > questions.. > > > > If what you say is true, then wouldn't I have toruble running xdm even > > from a root login after the boot process is over? I don't. > > No, not if the system is relatively quiescent. > > As I told Nate, you should be able to easily test this by loading > the I/O system (preferrably *not* your disks) and bouncing the > console in and out of X using ctrl-alt-F1/[ctrl-]alt-F4. If it > locks up, the problem has been identified. > > > How could I delay the startup of one of the processors, as you suggest in > > your last point? > > There is a sysctl which controlls the startup of the auxillary > processors. If you were to delay starting the processors until > after X was running, you would be able to see if this were related > to the SMP code path differences if it still locks up without the > other processors running (I don't believe it will) or to an I/O > space conflict (It may lock up following the AP's starting up, > but it probably won't, since when you do the manual start, the > system will be relatively quiescent). > > Terry Lambert > terry@lambert.org > --- > Any opinions in this posting are my own and not those of my present > or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 18:48:44 1999 Delivered-To: freebsd-smp@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2]) by hub.freebsd.org (Postfix) with ESMTP id 1BEC014CF3 for ; Thu, 22 Jul 1999 18:48:42 -0700 (PDT) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.9.3/8.9.1) id SAA94924; Thu, 22 Jul 1999 18:45:53 -0700 (PDT) (envelope-from dillon) Date: Thu, 22 Jul 1999 18:45:53 -0700 (PDT) From: Matthew Dillon Message-Id: <199907230145.SAA94924@apollo.backplane.com> To: "Richard A. Scheper" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup References: <199907221845.LAA10738@usr05.primenet.com> <3797C66E.B9382583@toad.net> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Going back the original posting in this thread, I think I've had something similar happen to me. In my case the problem was a *conflict* between getty and X. Specifically, I was running xdm from rc.local and it wound up using a vty that init would later try to start a getty on. When init started the getty, BANG! The getty steals the keyboard and X would freeze up. My solution was to leave ttyv0 and ttyv1 turned off in /etc/ttys. I later found a second solution which also worked quite well, and that is to start xdm *FROM* init, via an /etc/ttys entry: ttyv0 "/usr/X11R6/bin/xdm -nodaemon cons25 on secure These two solutions solved my keyboard lockup problems. -Matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 18:49:27 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ix.netcom.com (sil-wa17-26.ix.netcom.com [207.93.156.26]) by hub.freebsd.org (Postfix) with ESMTP id 52C9D15689 for ; Thu, 22 Jul 1999 18:49:07 -0700 (PDT) (envelope-from tomdean@ix.netcom.com) Received: (from tomdean@localhost) by ix.netcom.com (8.9.3/8.8.8) id SAA52651; Thu, 22 Jul 1999 18:48:00 -0700 (PDT) (envelope-from tomdean@ix.netcom.com) Date: Thu, 22 Jul 1999 18:48:00 -0700 (PDT) Message-Id: <199907230148.SAA52651@ix.netcom.com> X-Authentication-Warning: celebris.tddhome: tomdean set sender to tomdean@ix.netcom.com using -f From: Thomas Dean To: scheper@beast.toad.net Cc: freebsd-smp@FreeBSD.ORG In-reply-to: (message from Richard Scheper on Thu, 22 Jul 1999 13:10:18 -0400 (EDT)) Subject: Re: SMP + XDM = keyboard lockup References: Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I had a problem with SMP + xdm some time ago. I don't remember the details. The resolution was to start xdm "the correct way", in /usr/local/etc/rc.d/xdm.sh, with one ttyvx "off", as I stated. Have you tried creating /usr/local/etc/rc.d/xdm.sh and turning off one ttyv(x)? tomdean To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 20: 7:22 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id 93EAA15673 for ; Thu, 22 Jul 1999 20:07:15 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:6f0qZg4ZS4M71g6VTCb/1zIVesMqi6+x@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id MAA14115; Fri, 23 Jul 1999 12:06:58 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id MAA00450; Fri, 23 Jul 1999 12:11:16 +0900 (JST) Message-Id: <199907230311.MAA00450@zodiac.mech.utsunomiya-u.ac.jp> To: Richard Scheper Cc: freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Thu, 22 Jul 1999 15:16:20 -0400." References: Date: Fri, 23 Jul 1999 12:11:15 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >Actually, I don't have another way in, just the console. If I cause the >lockup I have no recourse but to turn the computer off. It makes it tough >to get info, I know. I never see any error messages in /var/log/messages Ok. Even if you don't have network connection to your system, it is possible to find clues as to what is happening. After you experience keyboard lockup, reboot your system, give the `-s' option at the boot loader prompt. You will then be in the single user mode. Examine the file /usr/X11R6/lib/X11/xdm/xdm-errors which records output from the X server and clients in the previous xdm invocation. Locate the lines: Using syscons driver with X support (version 2.0) (using VT number 4) ~~~~~~~~~~~ As you are trying to run xdm in /dev/ttyv3 by /etc/ttys, you should see "VT number 4" (notice the offset by 1; this is because the X server gives the first vty (/dev/ttyv0) the number 1), when the X server is running without the keyboard lockup. If the nubmer is 1 though 3, both the X server and getty are fighting for keyboard input in the same vty. When you run xdm from rc.local (or /usr/local/etc/rc/something) in one way or another, and have the keyboard lockup, you can veryfy xdm-getty conflicts in the same way by looking at this log file. As I stated before, I believe this is a timing problem. It may not always be reproducable by every SMP system; my SMP box has no trouble when I start xdm from /etc/ttys. Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 20:42:26 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id D3A801567C for ; Thu, 22 Jul 1999 20:42:17 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:JxF8kQ0cKmvXiA+PbMDtxqplZatpYfU2@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id MAA14917; Fri, 23 Jul 1999 12:41:13 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id MAA01470; Fri, 23 Jul 1999 12:45:29 +0900 (JST) Message-Id: <199907230345.MAA01470@zodiac.mech.utsunomiya-u.ac.jp> To: "Richard A. Scheper" Cc: freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Thu, 22 Jul 1999 21:33:34 -0400." <3797C66E.B9382583@toad.net> References: <199907221845.LAA10738@usr05.primenet.com> <3797C66E.B9382583@toad.net> Date: Fri, 23 Jul 1999 12:45:28 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org It is true that switching between X and text vtys fast by hitting CTL-ALT-Fx repeatedly may cause keyboard lockup. You can lock up the keyboard if you try to switch between them at the *wrong* moment. For example, many people have reported the keyboard lockup when they tried to switch vtys while the X server is starting up. VT switching is a tricky business and the XFree86 servers and syscons, from time to time, do not see each other in eye to eye... This is not I/O contention. It is an algorithm problem. # This, I believe, is mostly fixed in 4.0-CURRENT. But, I do not believe this is the cause of your problem. Because it is the xdm/X server alone that is switching vtys when your keyboard lockup happens. You, the user, are not switching vtys when this is happening. Nor over-loaded I/O is the cause; the X server alone is exercising inb()/outb() when your lockup happens. There is no I/O contention. Kazu >It looks like your suspiscions are correct. After further attempts at >starting >xdm from a root seesion, it does in fact hang more often than not. Perhaps >I was just "lucky" and it only worked with a quiet system. > >I'll try the ps test and see what I get > >-Rich > >Terry Lambert wrote: > >> > Wow! That's a lot to chew and mostly over my head, but a couple of >> > questions.. >> > >> > If what you say is true, then wouldn't I have toruble running xdm even >> > from a root login after the boot process is over? I don't. >> >> No, not if the system is relatively quiescent. >> >> As I told Nate, you should be able to easily test this by loading >> the I/O system (preferrably *not* your disks) and bouncing the >> console in and out of X using ctrl-alt-F1/[ctrl-]alt-F4. If it >> locks up, the problem has been identified. [...] To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 20:59:44 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id CE50914C28 for ; Thu, 22 Jul 1999 20:59:40 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:10aSkJKPL4tlcjGfvuT49ewJpNa9lFy6@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id MAA15512 for ; Fri, 23 Jul 1999 12:57:51 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id NAA02136; Fri, 23 Jul 1999 13:02:08 +0900 (JST) Message-Id: <199907230402.NAA02136@zodiac.mech.utsunomiya-u.ac.jp> To: freebsd-smp@freebsd.org Cc: yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Thu, 22 Jul 1999 18:17:19 GMT." <199907221817.LAA09698@usr05.primenet.com> References: <199907221817.LAA09698@usr05.primenet.com> Date: Fri, 23 Jul 1999 13:02:02 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >> > The XFree86 server tries to locate an empty vty to use for the X >> > sesssion. If xdm is started in rc.local, the X server will run BEFORE >> > init(8) starts invoking getty(8), and will find /dev/ttyv1 empty and >> > tries to occupy that vty. When rc.init is finished, init(8) will run >> > a copy of getty(8) in /dev/ttyv1, then, *Bomb* The X server and getty >> > both tries to read from the keyboard. > >I don't believe this is the case, because of the SMP factor. If >SMP is disabled, he does not see the problem. > >I believe that if what you are saying is correct, merely following >the xdm documentation and starting the X server in /etc/ttys should >resolve his problem. I guess it may not. After init forks and execs getty and xdm, we don't know the exact order in which these processes will be scheduled to run on which CPU, do we? /etc/ttys does not ensure the order of execution. If xdm happens to run before other getty processes, it may grab a vty which will be eventually opened by a getty later. Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 21:21:13 1999 Delivered-To: freebsd-smp@freebsd.org Received: from pawn.primelocation.net (pawn.primelocation.net [205.161.238.235]) by hub.freebsd.org (Postfix) with ESMTP id D52E614C0B for ; Thu, 22 Jul 1999 21:21:02 -0700 (PDT) (envelope-from jedgar@fxp.org) Received: by pawn.primelocation.net (Postfix, from userid 1003) id BD0E0F818; Fri, 23 Jul 1999 00:17:44 -0400 (EDT) Received: from localhost (localhost [127.0.0.1]) by pawn.primelocation.net (Postfix) with ESMTP id B0B999B15; Fri, 23 Jul 1999 00:17:44 -0400 (EDT) Date: Fri, 23 Jul 1999 00:17:44 -0400 (EDT) From: "Chris D. Faulhaber" X-Sender: jedgar@pawn.primelocation.net To: Kazutaka YOKOTA Cc: freebsd-smp@freebsd.org Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907230402.NAA02136@zodiac.mech.utsunomiya-u.ac.jp> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Fri, 23 Jul 1999, Kazutaka YOKOTA wrote: > >I believe that if what you are saying is correct, merely following > >the xdm documentation and starting the X server in /etc/ttys should > >resolve his problem. > > I guess it may not. After init forks and execs getty and xdm, we > don't know the exact order in which these processes will be scheduled > to run on which CPU, do we? /etc/ttys does not ensure the order of > execution. > > If xdm happens to run before other getty processes, it may grab a vty > which will be eventually opened by a getty later. > vty's are assigned in /etc/ttys, though: ttyv4 "/usr/X11R6/bin/xdm -nodaemon" xterm off secure ^^^^^ so it the starting order shouldn't matter, unless you have a getty and xdm assigned the same vty, and then it _should_ lock up whether using SMP or not. ----- Chris D. Faulhaber | All the true gurus I've met never System/Network Administrator, | claimed they were one, and always Reality Check Information, Inc. | pointed to someone better. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 21:33:32 1999 Delivered-To: freebsd-smp@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2]) by hub.freebsd.org (Postfix) with ESMTP id 68F7314CCB for ; Thu, 22 Jul 1999 21:33:30 -0700 (PDT) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.9.3/8.9.1) id VAA21465; Thu, 22 Jul 1999 21:32:09 -0700 (PDT) (envelope-from dillon) Date: Thu, 22 Jul 1999 21:32:09 -0700 (PDT) From: Matthew Dillon Message-Id: <199907230432.VAA21465@apollo.backplane.com> To: "Chris D. Faulhaber" Cc: Kazutaka YOKOTA , freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup References: Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org :> execution. :> :> If xdm happens to run before other getty processes, it may grab a vty :> which will be eventually opened by a getty later. :> : :vty's are assigned in /etc/ttys, though: : :ttyv4 "/usr/X11R6/bin/xdm -nodaemon" xterm off secure :^^^^^ :so it the starting order shouldn't matter, unless you have a getty and xdm :assigned the same vty, and then it _should_ lock up whether using SMP or :not. : :----- :Chris D. Faulhaber | All the true gurus I've met never That's what I thought too... and it's correct for getty's. But xdm ignores the argument init passes to it. So here is what happens: init opens the vty and runs xdm. xdm runs X which ignores the vty and locates a new one by scanning the vty's sequentially looking for one that is unused. In this case a race is possible. X will not, in fact, run in vty 0 like we expected. I tested this. My /etc/ttys runs xdm from ttyv0, but the X sessions winds up on ttyv4. However X does search vty's sequentially, so all you need to do is make sure the first two slots in /etc/ttys for ttyv0 and ttyv1 are turned off. So this is my recommendation now: Run xdm on ttyv0, leave ttyv1 turned off. Run your getty's starting with ttyv2. # ttyv0 "/usr/X11R6/bin/xdm -nodaemon cons25 on secure ttyv1 "/usr/libexec/getty Pc" cons25 off secure ttyv2 "/usr/libexec/getty Pc" cons25 on secure ttyv3 "/usr/libexec/getty Pc" cons25 on secure Or, if you want to run xdm from your rc.local, leave both ttyv0 AND ttyv1 turned off. Why? Because ttyv0 will be allocated while the rc files are being run so you still have to leave at least ttyv1 turned off in init to avoid a race. If you do this I believe there is no possibility of a race. It would be nice if we could specify the vty that X is to allocate as an option to X. I couldn't find anything in the man pages right offhand to accomplish this, though. -Matt Matthew Dillon To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 21:41:24 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id 8D5B714C4A for ; Thu, 22 Jul 1999 21:41:21 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:9J7kKKAMKsPDXp2OS67yXOKUZNwGDGHH@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id NAA14270; Fri, 23 Jul 1999 13:41:11 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id NAA03665; Fri, 23 Jul 1999 13:45:28 +0900 (JST) Message-Id: <199907230445.NAA03665@zodiac.mech.utsunomiya-u.ac.jp> To: "Chris D. Faulhaber" Cc: freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Fri, 23 Jul 1999 00:17:44 -0400." References: Date: Fri, 23 Jul 1999 13:45:22 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >> I guess it may not. After init forks and execs getty and xdm, we >> don't know the exact order in which these processes will be scheduled >> to run on which CPU, do we? /etc/ttys does not ensure the order of >> execution. >> >> If xdm happens to run before other getty processes, it may grab a vty >> which will be eventually opened by a getty later. >> > >vty's are assigned in /etc/ttys, though: > >ttyv4 "/usr/X11R6/bin/xdm -nodaemon" xterm off secure > >^^^^^ >so it the starting order shouldn't matter, unless you have a getty and xdm >assigned the same vty, and then it _should_ lock up whether using SMP or >not. This does not necessarily mean the X server will run in the same ttyv4 as xdm is. It will still look for an empty vty. I just experimented with the following /etc/ttys: ttyv1 "/usr/libexec/getty Pc" cons25 on secure ttyv2 "/usr/libexec/getty Pc" cons25 on secure ttyv3 "/usr/libexec/getty Pc" cons25 on secure ttyv4 "/usr/libexec/getty Pc" cons25 on secure ttyv5 "/usr/libexec/getty Pc" cons25 on secure ttyv6 "/usr/libexec/getty Pc" cons25 on secure ttyv7 "/usr/libexec/getty Pc" cons25 off secure ttyv8 "/usr/libexec/getty Pc" cons25 off secure ttyv9 "/usr/X11R6/bin/xdm -nodaemon" xterm on secure Notice two empty vtys: ttyv7 and ttyv8. Did `kill -HUP 1' and I got the X server in ttyv7 rather than ttyv9. Here is lines from /usr/X11R6/lib/X11/xdm/xdm-errors: XFree86 Version 3.3.3.1 / X Window System (protocol Version 11, revision 0, vendor release 6300) Release Date: December 29 1998 If the server is older than 6-12 months, or if your card is newer than the above date, look for a newer version before reportiing problems. (see http://www.XFree86.Org/FAQ) Operating System: FreeBSD 3.0-CURRENT i386 [ELF] Configured drivers: SVGA: server for SVGA graphics adaptors (Patchlevel 0): NV1, STG2000, RIVA128, RIVATNT, ET4000, ET4000W32, ET4000W32i, [...] ct64300, mediagx, V1000, V2x00, p9100, spc8110, generic Using syscons driver with X support (version 2.0) (using VT number 8) ~~~~~~~~~~~~ XF86Config: /etc/XF86Config [...] "VT number 8" means /dev/ttyv7. This means that if xdm if run before any getty when the system is staring up, it may grab a vty which will be subsequently used by a getty process. Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 21:44:47 1999 Delivered-To: freebsd-smp@freebsd.org Received: from scam.xcf.berkeley.edu (scam.XCF.Berkeley.EDU [128.32.43.201]) by hub.freebsd.org (Postfix) with SMTP id 2C90A14FAE for ; Thu, 22 Jul 1999 21:44:45 -0700 (PDT) (envelope-from grady@scam.XCF.Berkeley.EDU) Received: (qmail 1626 invoked by uid 348); 23 Jul 1999 04:17:40 -0000 Received: from localhost (HELO scam.XCF.Berkeley.EDU) (sendmail-bs@127.0.0.1) by localhost with SMTP; 23 Jul 1999 04:17:40 -0000 To: freebsd-smp@freebsd.org Subject: slow network reading with SMP From: grady@xcf.berkeley.edu (Steven Grady) MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <1622.932703459.1@scam.XCF.Berkeley.EDU> Date: Thu, 22 Jul 1999 21:17:39 -0700 Message-Id: <19990723044445.2C90A14FAE@hub.freebsd.org> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org We've got a SuperMicro P6DGU system (other specs: 2 450 MHz Pentium IIIs, 256M mem, 18G SCSI, 25G UDMA) running 3.2-RELEASE. The network card is a NetGear FA310TX Rev D1 (pn driver). With an SMP kernel, TCP-based file transfers to the machine are very slow (i.e. 50k/sec over a 100Mbit ethernet). Transfers from the machine are fine -- a couple megs per second. After spending a lot of time trying different things (swapping network cards, changing between a switch and a hub, transferring to Windows vs. another BSD machine, transferring via Samba vs. FTP), the one thing that made a difference was to replace the kernel with a non-SMP version. With the non-SMP kernel, writes went up to 6 M/sec (depending on the protocol). The only difference between the kernels was the deletion of the following lines: options SMP options APIC_IO Why is this happening? And more importantly, what can we do to use both processors while still keeping full network speed? (Others in the company have mentioned Linux -- I'd like to avoid that.) Steven grady@xcf.berkeley.edu "When will I learn? The answers to life's problems aren't at the bottom of a bottle. They're on TV!" To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 21:52: 6 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ix.netcom.com (sil-wa17-26.ix.netcom.com [207.93.156.26]) by hub.freebsd.org (Postfix) with ESMTP id 7DE1D14C4A for ; Thu, 22 Jul 1999 21:52:04 -0700 (PDT) (envelope-from tomdean@ix.netcom.com) Received: (from tomdean@localhost) by ix.netcom.com (8.9.3/8.8.8) id VAA52884; Thu, 22 Jul 1999 21:51:41 -0700 (PDT) (envelope-from tomdean@ix.netcom.com) Date: Thu, 22 Jul 1999 21:51:41 -0700 (PDT) Message-Id: <199907230451.VAA52884@ix.netcom.com> X-Authentication-Warning: celebris.tddhome: tomdean set sender to tomdean@ix.netcom.com using -f From: Thomas Dean To: dillon@apollo.backplane.com Cc: jedgar@fxp.org, yokota@zodiac.mech.utsunomiya-u.ac.jp, freebsd-smp@FreeBSD.ORG In-reply-to: <199907230432.VAA21465@apollo.backplane.com> (message from Matthew Dillon on Thu, 22 Jul 1999 21:32:09 -0700 (PDT)) Subject: Re: SMP + XDM = keyboard lockup References: <199907230432.VAA21465@apollo.backplane.com> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I am runing 4.0 current SMP. Some time back, we had the discussion on this list about how to start X. The CTL-ALT-Fx problem was discussed and no solution was produced. This is still an open issue. If you shutdown an X session and use CTL-ALT-Fx before X completes resetting the screen and presents the login prompt, the keyboard will lock. This happens every time. After a long sequence, the answer to the question of how to start X was to set /etc/ttys as I said earlier and start xdm in /usr/local/etc/rc.d/.sh. I switched to this mechanism and have not had a problem since. Has the "correct" way to start X changed back to using rc.local or /etc/ttys? tomdean To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 22: 2:38 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id 1065014E3C for ; Thu, 22 Jul 1999 22:02:35 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:ySbSL0fyUEgjepbLqH33yPtH21QWHNez@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id OAA14136; Fri, 23 Jul 1999 14:02:28 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id OAA04583; Fri, 23 Jul 1999 14:06:46 +0900 (JST) Message-Id: <199907230506.OAA04583@zodiac.mech.utsunomiya-u.ac.jp> To: Matthew Dillon Cc: freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Thu, 22 Jul 1999 21:32:09 MST." <199907230432.VAA21465@apollo.backplane.com> References: <199907230432.VAA21465@apollo.backplane.com> Date: Fri, 23 Jul 1999 14:06:45 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > That's what I thought too... and it's correct for getty's. But xdm > ignores the argument init passes to it. > > So here is what happens: > > init opens the vty and runs xdm. > > xdm runs X which ignores the vty and locates a new one by scanning > the vty's sequentially looking for one that is unused. That's correct. > It would be nice if we could specify the vty that X is to allocate > as an option to X. I couldn't find anything in the man pages right > offhand to accomplish this, though. There is. You can tell the XFree86 server to run in a specific vty by starting it as /usr/X11R6/bin/X vt4 (Note that "vt4" means /dev/ttyv3. The number is off by one.) This is XFree86 extention. See XFree86(1). When we run xdm, this option must be entered in /usr/X11R6/lib/X11/xdm/Xservers. :0 local /usr/X11R6/bin/X vt4 The trouble is that you must update both /usr/X11R6/lib/X11/xdm/Xservers and /etc/ttys, if you want to change the vty for the X session later. Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 22:17:44 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id AB9C114D48 for ; Thu, 22 Jul 1999 22:17:41 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:5LD07e7+oQ/uf4W8z4U4TMrbkcNZzTz1@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id OAA15872; Fri, 23 Jul 1999 14:15:48 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id OAA05070; Fri, 23 Jul 1999 14:20:05 +0900 (JST) Message-Id: <199907230520.OAA05070@zodiac.mech.utsunomiya-u.ac.jp> To: Thomas Dean Cc: freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Thu, 22 Jul 1999 21:51:41 MST." <199907230451.VAA52884@ix.netcom.com> References: <199907230432.VAA21465@apollo.backplane.com> <199907230451.VAA52884@ix.netcom.com> Date: Fri, 23 Jul 1999 14:19:59 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >I am runing 4.0 current SMP. How recent is it? >The CTL-ALT-Fx problem was discussed and no solution was produced. >This is still an open issue. If you shutdown an X session and use >CTL-ALT-Fx before X completes resetting the screen and presents the >login prompt, the keyboard will lock. This happens every time. Please describe the exact sequence and symptom. How did you start X? By xdm? How did you "shutdown" the X sesssion? When the keyboard is looked up, which vty are you in? Do NumLock, CapsLock and ScrollLock change LED state? If you try to switch to another vty by hitting ALT-Fx, do you here beep? Would you hit ALT-Fx several times to see if it makes any difference? >After a long sequence, the answer to the question of how to start X >was to set /etc/ttys as I said earlier and start xdm in >/usr/local/etc/rc.d/.sh. > >I switched to this mechanism and have not had a problem since. > >Has the "correct" way to start X changed back to using rc.local or >/etc/ttys? Some people like /etc/ttys. Other people prefer rc.local or /usr/local/etc/rc*. Me? I don't care which is "correct", so long as it works. Well, maybe I should think harder to decide which is more "correct"... Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 22:26:54 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id 634DB14D48 for ; Thu, 22 Jul 1999 22:26:51 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:VBwDj8x/t2swYPKK4l9nG55TaLAC5rCt@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id OAA15823; Fri, 23 Jul 1999 14:25:32 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id OAA05573; Fri, 23 Jul 1999 14:29:50 +0900 (JST) Message-Id: <199907230529.OAA05573@zodiac.mech.utsunomiya-u.ac.jp> To: Thomas Dean Cc: freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Fri, 23 Jul 1999 14:19:59 JST." <199907230520.OAA05070@zodiac.mech.utsunomiya-u.ac.jp> References: <199907230432.VAA21465@apollo.backplane.com> <199907230451.VAA52884@ix.netcom.com> <199907230520.OAA05070@zodiac.mech.utsunomiya-u.ac.jp> Date: Fri, 23 Jul 1999 14:29:47 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I wrote: >Please describe the exact sequence and symptom. How did you start X? >By xdm? How did you "shutdown" the X sesssion? When the keyboard is >looked up, which vty are you in? Do NumLock, CapsLock and ScrollLock >change LED state? If you try to switch to another vty by hitting >ALT-Fx, do you here beep? Would you hit ALT-Fx several times to see >if it makes any difference? One more thing. If you have network connection, login to the system via network, then run 'kbd_mode -a'. (You may need to be root to do this.) This command will bring the keyboard back to the translation mode and might revive the keyboard in the situation you described. Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Thu Jul 22 23: 5:56 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ix.netcom.com (sil-wa17-26.ix.netcom.com [207.93.156.26]) by hub.freebsd.org (Postfix) with ESMTP id AFD4E156BF for ; Thu, 22 Jul 1999 23:05:50 -0700 (PDT) (envelope-from tomdean@ix.netcom.com) Received: (from tomdean@localhost) by ix.netcom.com (8.9.3/8.8.8) id XAA53320; Thu, 22 Jul 1999 23:03:36 -0700 (PDT) (envelope-from tomdean@ix.netcom.com) Date: Thu, 22 Jul 1999 23:03:36 -0700 (PDT) Message-Id: <199907230603.XAA53320@ix.netcom.com> X-Authentication-Warning: celebris.tddhome: tomdean set sender to tomdean@ix.netcom.com using -f From: Thomas Dean To: yokota@zodiac.mech.utsunomiya-u.ac.jp Cc: freebsd-smp@freebsd.org In-reply-to: <199907230520.OAA05070@zodiac.mech.utsunomiya-u.ac.jp> (message from Kazutaka YOKOTA on Fri, 23 Jul 1999 14:19:59 +0900) Subject: Re: SMP + XDM = keyboard lockup References: <199907230432.VAA21465@apollo.backplane.com> <199907230451.VAA52884@ix.netcom.com> <199907230520.OAA05070@zodiac.mech.utsunomiya-u.ac.jp> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I have not tried this for some time. The symptoms are somewhat different. I have rebuilt my complete system since the last time I caused a hang. This was because of the elf and egcs changes. I rebuilt all ports, including X. I learned to not touch the keyboard during the end of session transition. Now, the X session will start. But, CTL-ALT-Fx just beeps. Everything with the X session seems OK. I start X from /usr/local/etc/rc.d/xdm.sh, included below. I use the xsm session manager. If I select xsm->shutdown->immediately with the mouse, the session ends. The video mode changes: 1. The screen blanks. 2. A seemingly low-resolution color display is presented with some bars across the top 1 cm of the screen. 3. The login prompt is presented. Previously, during the video mode changing, during #1 or #2 above, if I used CTL-ALT-Fx, the keyboard would hang. I would have to telnet into the system and restart things manually. I killed xdm, restarted getty, and, restarted xdm. This fixed things. I can telnet in to the system. 'kbd_mode -a' does not change the keyboard action. The NumLock, CapsLock and ScrollLock led's do not change state. I can not switch out of the X mode using CTL-ALT-Fx, it just beeps. I will try to restart getty to see if it recovers. I will send the results of this in a few minutes. tomdean === /usr/local/etc/rc.d/xdm.sh ======= #! /bin/sh # $Id: xdm.sh,v 1.1 1999/02/14 06:05:30 tomdean Exp $ # start xdm # 19990115 tomdean - initial version if [ -x /usr/X11R6/bin/xdm ] ; then # need to cleanup first. if [ -f /usr/X11R6/lib/X11/xdm/xdm-pid ] ; then rm -f /usr/X11R6/lib/X11/xdm/xdm-pid fi if [ -f /tmp/.X0-lock ] ; then rm -f /tmp/.X0-lock fi # now, we can start it. echo " xdm"; /usr/X11R6/bin/xdm fi === from /etc/ttys ============ ... ttyv0 "/usr/libexec/getty Pc" cons25 on secure ttyv1 "/usr/libexec/getty Pc" cons25 on secure ttyv2 "/usr/libexec/getty Pc" cons25 on secure ttyv3 "/usr/libexec/getty Pc" cons25 on secure ttyv4 "/usr/libexec/getty Pc" cons25 on secure ttyv5 "/usr/libexec/getty Pc" cons25 on secure ttyv6 "/usr/libexec/getty Pc" cons25 on secure ttyv7 "/usr/libexec/getty Pc" cons25 off secure ... === from top ============= last pid: 53300; load averages: 0.00, 0.01, 0.01 up 24+06:34:06 22:50:30 51 processes: 1 running, 50 sleeping CPU states: 0.6% user, 0.0% nice, 0.8% system, 0.4% interrupt, 98.3% idle Mem: 40M Active, 22M Inact, 18M Wired, 6236K Cache, 8024K Buf, 6100K Free Swap: 227M Total, 2224K Used, 225M Free PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU COMMAND 53300 root 28 0 1596K 800K CPU1 0 0:01 1.43% 0.98% top 53269 tomdean 2 0 6184K 5560K select 1 0:14 0.63% 0.63% emacs 7367 root 2 0 10192K 7836K select 1 84:43 0.05% 0.05% XF86_SVGA 32229 tomdean 2 0 19296K 12380K select 1 22:25 0.00% 0.00% netscape 32295 root 2 0 1592K 880K select 0 17:45 0.00% 0.00% ppp 252 root 2 0 1460K 516K select 0 1:34 0.00% 0.00% nmbd 171 root 2 0 1300K 568K select 0 0:58 0.00% 0.00% sendmail 164 root 10 0 1012K 340K nanslp 1 0:54 0.00% 0.00% cron 123 root 2 0 300K 20K nfsd 0 0:52 0.00% 0.00% nfsd 98 root 2 0 840K 372K select 0 0:37 0.00% 0.00% syslogd 47506 tomdean 2 0 1096K 688K select 0 0:13 0.00% 0.00% fetchmail 161 root 2 0 924K 380K select 0 0:05 0.00% 0.00% inetd 125 root 2 0 300K 20K nfsd 0 0:04 0.00% 0.00% nfsd 53258 tomdean 2 0 2328K 1752K select 1 0:01 0.00% 0.00% twm 1 root 10 0 504K 72K wait 1 0:01 0.00% 0.00% init 53263 tomdean 2 0 2780K 2236K select 1 0:01 0.00% 0.00% xterm 53248 tomdean 3 0 1392K 1044K ttyin 0 0:01 0.00% 0.00% tcsh 53243 tomdean 2 0 2544K 2008K select 0 0:01 0.00% 0.00% xterm 53226 tomdean 2 0 2332K 1776K select 1 0:01 0.00% 0.00% xsm 53260 tomdean 2 0 1956K 1364K select 1 0:00 0.00% 0.00% smproxy 53241 tomdean 2 0 2464K 1924K select 0 0:00 0.00% 0.00% xterm 53244 tomdean 2 0 2544K 2004K select 0 0:00 0.00% 0.00% xterm 53274 tomdean 18 0 1320K 1004K pause 1 0:00 0.00% 0.00% tcsh 53249 tomdean 3 0 1320K 968K ttyin 0 0:00 0.00% 0.00% tcsh 53250 tomdean 3 0 1320K 968K ttyin 0 0:00 0.00% 0.00% tcsh 286 root 3 0 476K 8K ttyin 0 0:00 0.00% 0.00% csh 53264 root 18 0 480K 312K pause 0 0:00 0.00% 0.00% csh 118 root 2 0 500K 8K select 0 0:00 0.00% 0.00% mountd 53215 root 10 0 2200K 1452K wait 1 0:00 0.00% 0.00% xdm 250 root 2 0 1824K 356K select 0 0:00 0.00% 0.00% smbd 53273 root 2 0 908K 584K select 0 0:00 0.00% 0.00% telnetd 53282 root 3 0 472K 324K ttyin 1 0:00 0.00% 0.00% csh 107 daemon 2 0 848K 288K select 1 0:00 0.00% 0.00% portmap 7364 root 2 0 2148K 956K select 1 0:00 0.00% 0.00% xdm 168 root 2 0 852K 316K select 1 0:00 0.00% 0.00% lpd 289 root 3 0 852K 248K ttyin 1 0:00 0.00% 0.00% getty 287 root 3 0 852K 248K ttyin 0 0:00 0.00% 0.00% getty 290 root 3 0 852K 248K ttyin 1 0:00 0.00% 0.00% getty 288 root 3 0 852K 248K ttyin 1 0:00 0.00% 0.00% getty 291 root 3 0 852K 248K ttyin 0 0:00 0.00% 0.00% getty 53270 tomdean 2 0 820K 396K select 0 0:00 0.00% 0.00% emacsserv === from ps -al ============ 0 286 1 1 3 0 476 8 ttyin Is+ v0 0:00.27 -csh (csh) 0 287 1 0 3 0 852 248 ttyin Is+ v1 0:00.06 /usr/libexec/getty Pc ttyv1 0 288 1 0 3 0 852 248 ttyin Is+ v2 0:00.05 /usr/libexec/getty Pc ttyv2 0 289 1 0 3 0 852 248 ttyin Is+ v3 0:00.06 /usr/libexec/getty Pc ttyv3 0 290 1 0 3 0 852 248 ttyin Is+ v4 0:00.05 /usr/libexec/getty Pc ttyv4 0 291 1 0 3 0 852 248 ttyin Is+ v5 0:00.05 /usr/libexec/getty Pc ttyv5 0 292 1 1 3 0 852 248 ttyin Is+ v6 0:00.04 /usr/libexec/getty Pc ttyv6 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 0:46: 0 1999 Delivered-To: freebsd-smp@freebsd.org Received: from news.IAEhv.nl (news.IAE.nl [194.151.64.4]) by hub.freebsd.org (Postfix) with ESMTP id E54C115776 for ; Fri, 23 Jul 1999 00:45:30 -0700 (PDT) (envelope-from marc@bowtie.nl) Received: (from uucp@localhost) by news.IAEhv.nl (8.9.1/8.9.1) with IAEhv.nl id JAA00794; Fri, 23 Jul 1999 09:45:04 +0200 (MET DST) Received: from localhost (localhost [127.0.0.1]) by bowtie.nl (8.8.8/8.8.8) with ESMTP id JAA11980; Fri, 23 Jul 1999 09:43:07 +0200 (CEST) (envelope-from marc@bowtie.nl) Message-Id: <199907230743.JAA11980@bowtie.nl> X-Mailer: exmh version 2.0.2 2/24/98 To: Kazutaka YOKOTA Cc: Richard Scheper , freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-reply-to: yokota's message of Fri, 23 Jul 1999 12:11:15 +0900. <199907230311.MAA00450@zodiac.mech.utsunomiya-u.ac.jp> Date: Fri, 23 Jul 1999 09:43:07 +0200 From: Marc van Kempen Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > >Actually, I don't have another way in, just the console. If I cause the > >lockup I have no recourse but to turn the computer off. It makes it tough > >to get info, I know. I never see any error messages in /var/log/messages > > Ok. Even if you don't have network connection to your system, it is > possible to find clues as to what is happening. > > After you experience keyboard lockup, reboot your system, give the > `-s' option at the boot loader prompt. You will then be in the single > user mode. > > Examine the file /usr/X11R6/lib/X11/xdm/xdm-errors which records > output from the X server and clients in the previous xdm invocation. > Locate the lines: > > Using syscons driver with X support (version 2.0) > (using VT number 4) > ~~~~~~~~~~~ > > As you are trying to run xdm in /dev/ttyv3 by /etc/ttys, you should > see "VT number 4" (notice the offset by 1; this is because the X > server gives the first vty (/dev/ttyv0) the number 1), when the X server > is running without the keyboard lockup. > > If the nubmer is 1 though 3, both the X server and getty are > fighting for keyboard input in the same vty. > > When you run xdm from rc.local (or /usr/local/etc/rc/something) in one > way or another, and have the keyboard lockup, you can veryfy xdm-getty > conflicts in the same way by looking at this log file. > > As I stated before, I believe this is a timing problem. It may not > always be reproducable by every SMP system; my SMP box has no trouble > when I start xdm from /etc/ttys. > I have had this happen to me on several systems (UNI processor), and the problem would not occur all the time which suggested to me too that it was some kind of timing problem between init and X trying to grab the keyboard. I could always solve it by starting xdm from a script which would sleep for several seconds (5-10) (giving init time to complete) and then start xdm. So it seems it isn't even SMP related. Regards, Marc. ---------------------------------------------------- Marc van Kempen BowTie Technology Email: marc@bowtie.nl WWW & Databases tel. +31 40 2 43 20 65 fax. +31 40 2 44 21 86 http://www.bowtie.nl ---------------------------------------------------- To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 1:44:13 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mbox.net96.it (mbox.net96.it [194.244.102.152]) by hub.freebsd.org (Postfix) with ESMTP id B43BF14BCF for ; Fri, 23 Jul 1999 01:44:08 -0700 (PDT) (envelope-from ccau@itsyn.it) Received: by mbox.net96.it from localhost (router,SLMail V2.6); Fri, 23 Jul 1999 10:40:32 +0200 Received: by mbox.net96.it from 0 (151.4.19.5::mail daemon; unverified,SLMail V2.6); Fri, 23 Jul 1999 10:40:31 +0200 Message-ID: <001201bed4e7$190d79c0$050aa8c0@0> Reply-To: "Corrado Cau" From: "Corrado Cau" To: Subject: Another SMP FreeBSD was born... Date: Fri, 23 Jul 1999 10:40:57 +0200 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 4.72.3110.5 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3 Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi all, thanks to your help I suceeded setting up FreeBSD 3.2 SMP on my machine (Asus P2B-D m/b). If anybody is interested in the "hardware compatibility list" I can provide the full HW list, but it seemed an off-topic here and now. Corrado To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 8:28:51 1999 Delivered-To: freebsd-smp@freebsd.org Received: from sasami.jurai.net (sasami.jurai.net [63.67.141.99]) by hub.freebsd.org (Postfix) with ESMTP id EEF5114D21 for ; Fri, 23 Jul 1999 08:28:34 -0700 (PDT) (envelope-from winter@jurai.net) Received: from localhost (winter@localhost) by sasami.jurai.net (8.8.8/8.8.7) with ESMTP id LAA15878 for ; Fri, 23 Jul 1999 11:28:15 -0400 (EDT) Date: Fri, 23 Jul 1999 11:28:15 -0400 (EDT) From: "Matthew N. Dodd" To: freebsd-smp@FreeBSD.ORG Subject: Best PPro stepping? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org I'm starting to bring together hardware for a system that will be using an ALR Revoloution 6x6 board and am wondering if there is any particular stepping of PPro I should be looking for. Obviously I'll need all 6 processors to be of the same stepping though its likely that I can get away with slightly different steppings on each CPU board (ie: 2 groups of 3 CPUs each; each group having the same stepping.) should finding 6 of the same steppng prove to be difficult. Thanks. -- | Matthew N. Dodd | '78 Datsun 280Z | '75 Volvo 164E | FreeBSD/NetBSD | | winter@jurai.net | 2 x '84 Volvo 245DL | ix86,sparc,pmax | | http://www.jurai.net/~winter | This Space For Rent | ISO8802.5 4ever | To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 9: 2:12 1999 Delivered-To: freebsd-smp@freebsd.org Received: from mercury.toad.net (mercury.toad.net [205.197.182.105]) by hub.freebsd.org (Postfix) with ESMTP id 2458F14BD6 for ; Fri, 23 Jul 1999 09:01:53 -0700 (PDT) (envelope-from scheper@toad.net) Received: from toad.net (core10d9.dynamic-dialup.toad.net [209.150.114.9]) by mercury.toad.net (8.8.8/8.8.8) with ESMTP id MAA01321 for ; Fri, 23 Jul 1999 12:01:16 -0400 Message-ID: <37989106.C7758997@toad.net> Date: Fri, 23 Jul 1999 11:57:58 -0400 From: "Richard A. Scheper" X-Mailer: Mozilla 4.08 [en] (X11; I; FreeBSD 3.1-RELEASE i386) MIME-Version: 1.0 To: freebsd-smp@FreeBSD.ORG Subject: SMP + XDM problem resolved! Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org OK. Problem solved. Yes it was something stupid and I feel rather sheepish.. As someone pointed out, the termonal setting in /usr/X11R6/lib/X11/xdm/Xservers /usr/X11R6/bin/X vt4 actually refers to vtty3. I had mine set to vt3, which is vtty2. Thus, X must have tried to grab vtty2 which caused a conflict. Once I corrected this the problem went away. I can strt xdm in /etc/rc.local with no fancy tricks, as before. I guess the fact that it worked without SMP in its incorrect setup had to do with timing issues after all. SMP just revealed the mistake. Thanks to all who spent so much time troubleshooting this problem. I sincerely appreciate your time. -Rich To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 9:20:24 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ns.mt.sri.com (ns.mt.sri.com [206.127.79.91]) by hub.freebsd.org (Postfix) with ESMTP id EA38C14BFF for ; Fri, 23 Jul 1999 09:20:20 -0700 (PDT) (envelope-from nate@mt.sri.com) Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id KAA09735; Fri, 23 Jul 1999 10:20:00 -0600 (MDT) (envelope-from nate@rocky.mt.sri.com) Received: by mt.sri.com (SMI-8.6/SMI-SVR4) id KAA07705; Fri, 23 Jul 1999 10:19:59 -0600 Date: Fri, 23 Jul 1999 10:19:59 -0600 Message-Id: <199907231619.KAA07705@mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Matthew Dillon Cc: "Richard A. Scheper" , freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907230145.SAA94924@apollo.backplane.com> References: <199907221845.LAA10738@usr05.primenet.com> <3797C66E.B9382583@toad.net> <199907230145.SAA94924@apollo.backplane.com> X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > I later found a second solution which also worked quite well, and > that is to start xdm *FROM* init, via an /etc/ttys entry: > > ttyv0 "/usr/X11R6/bin/xdm -nodaemon cons25 on secure The author tried this, and it didn't work. ;( Nate To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 9:24:46 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ns.mt.sri.com (ns.mt.sri.com [206.127.79.91]) by hub.freebsd.org (Postfix) with ESMTP id 666D514BFF for ; Fri, 23 Jul 1999 09:24:40 -0700 (PDT) (envelope-from nate@mt.sri.com) Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id KAA09806; Fri, 23 Jul 1999 10:24:11 -0600 (MDT) (envelope-from nate@rocky.mt.sri.com) Received: by mt.sri.com (SMI-8.6/SMI-SVR4) id KAA07741; Fri, 23 Jul 1999 10:24:10 -0600 Date: Fri, 23 Jul 1999 10:24:10 -0600 Message-Id: <199907231624.KAA07741@mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Kazutaka YOKOTA Cc: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907230402.NAA02136@zodiac.mech.utsunomiya-u.ac.jp> References: <199907221817.LAA09698@usr05.primenet.com> <199907230402.NAA02136@zodiac.mech.utsunomiya-u.ac.jp> X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > I guess it may not. After init forks and execs getty and xdm, we > don't know the exact order in which these processes will be scheduled > to run on which CPU, do we? /etc/ttys does not ensure the order of > execution. > > If xdm happens to run before other getty processes, it may grab a vty > which will be eventually opened by a getty later. Actually, I had forgotten about this. This is true because the line in /etc/ttys doesn't specify *which* tty for X to use. If we could easily force X to use that specific tty, it would avoid this problem. Nate To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 9:28:35 1999 Delivered-To: freebsd-smp@freebsd.org Received: from ns.mt.sri.com (ns.mt.sri.com [206.127.79.91]) by hub.freebsd.org (Postfix) with ESMTP id B639E14BFF for ; Fri, 23 Jul 1999 09:28:33 -0700 (PDT) (envelope-from nate@mt.sri.com) Received: from mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by ns.mt.sri.com (8.8.8/8.8.8) with SMTP id KAA09845; Fri, 23 Jul 1999 10:27:48 -0600 (MDT) (envelope-from nate@rocky.mt.sri.com) Received: by mt.sri.com (SMI-8.6/SMI-SVR4) id KAA07766; Fri, 23 Jul 1999 10:27:46 -0600 Date: Fri, 23 Jul 1999 10:27:46 -0600 Message-Id: <199907231627.KAA07766@mt.sri.com> From: Nate Williams MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Thomas Dean Cc: dillon@apollo.backplane.com, jedgar@fxp.org, yokota@zodiac.mech.utsunomiya-u.ac.jp, freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup In-Reply-To: <199907230451.VAA52884@ix.netcom.com> References: <199907230432.VAA21465@apollo.backplane.com> <199907230451.VAA52884@ix.netcom.com> X-Mailer: VM 6.34 under 19.16 "Lille" XEmacs Lucid Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > After a long sequence, the answer to the question of how to start X > was to set /etc/ttys as I said earlier and start xdm in > /usr/local/etc/rc.d/.sh. Actually, if this was the discussion that involved J'org and I, the result was that we agreed it should be started out of /etc/ttys, which is why I committed the change to the default /etc/ttys in FreeBSD. > I switched to this mechanism and have not had a problem since. Starting out of rc.d is *NO* safer than starting out of /etc/ttys, and has the chance of causing problems. By putting it in /etc/ttys, you can immediately see where conflicts would arise since everything else is started out of /etc/ttys. Nate To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 11:49:43 1999 Delivered-To: freebsd-smp@freebsd.org Received: from shell.futuresouth.com (shell.futuresouth.com [198.78.58.28]) by hub.freebsd.org (Postfix) with ESMTP id 2223815728 for ; Fri, 23 Jul 1999 11:48:41 -0700 (PDT) (envelope-from fullermd@futuresouth.com) Received: (from fullermd@localhost) by shell.futuresouth.com (8.9.3/8.9.3) id NAA06126; Fri, 23 Jul 1999 13:47:13 -0500 (CDT) Date: Fri, 23 Jul 1999 13:47:12 -0500 From: "Matthew D. Fuller" To: "Matthew N. Dodd" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Best PPro stepping? Message-ID: <19990723134712.M12369@futuresouth.com> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.3i In-Reply-To: ; from Matthew N. Dodd on Fri, Jul 23, 1999 at 11:28:15AM -0400 X-OS: FreeBSD Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Fri, Jul 23, 1999 at 11:28:15AM -0400, a little birdie told me that Matthew N. Dodd remarked > I'm starting to bring together hardware for a system that will be using an > ALR Revoloution 6x6 board and am wondering if there is any particular > stepping of PPro I should be looking for. Obviously I'll need all 6 > processors to be of the same stepping though its likely that I can get > away with slightly different steppings on each CPU board (ie: 2 groups of > 3 CPUs each; each group having the same stepping.) should finding 6 of the > same steppng prove to be difficult. I'd be fascinated to know where you got this hardware and what it cost. -- Matthew Fuller (MF4839) | fullermd@over-yonder.net Unix Systems Administrator | fullermd@futuresouth.com Specializing in FreeBSD | http://www.over-yonder.net/ FutureSouth Communications | ISPHelp ISP Consulting "The only reason I'm burning my candle at both ends, is because I haven't figured out how to light the middle yet" To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 11:56:25 1999 Delivered-To: freebsd-smp@freebsd.org Received: from sasami.jurai.net (sasami.jurai.net [63.67.141.99]) by hub.freebsd.org (Postfix) with ESMTP id C39F01568D for ; Fri, 23 Jul 1999 11:54:54 -0700 (PDT) (envelope-from winter@jurai.net) Received: from localhost (winter@localhost) by sasami.jurai.net (8.8.8/8.8.7) with ESMTP id OAA18809; Fri, 23 Jul 1999 14:53:17 -0400 (EDT) Date: Fri, 23 Jul 1999 14:53:17 -0400 (EDT) From: "Matthew N. Dodd" To: "Matthew D. Fuller" Cc: freebsd-smp@FreeBSD.ORG Subject: Re: Best PPro stepping? In-Reply-To: <19990723134712.M12369@futuresouth.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Fri, 23 Jul 1999, Matthew D. Fuller wrote: > On Fri, Jul 23, 1999 at 11:28:15AM -0400, a little birdie told me > that Matthew N. Dodd remarked > > I'm starting to bring together hardware for a system that will be using an > > ALR Revoloution 6x6 board and am wondering if there is any particular > > stepping of PPro I should be looking for. Obviously I'll need all 6 > > processors to be of the same stepping though its likely that I can get > > away with slightly different steppings on each CPU board (ie: 2 groups of > > 3 CPUs each; each group having the same stepping.) should finding 6 of the > > same steppng prove to be difficult. > > I'd be fascinated to know where you got this hardware and what it cost. I do run a mailing list to minimize the effort I have to expend to tell people about all the weird things I find. majordomo@jurai.net -> subscribe matts-hardware === begin quoted message === From winter@jurai.net Fri Jul 23 11:52:52 1999 -0400 Status: R X-Status: X-Keywords: Received: (from majordom@localhost) by sasami.jurai.net (8.8.8/8.8.7) id LAA16345 for matts-hardware-outgoing; Fri, 23 Jul 1999 11:52:49 -0400 (EDT) Date: Fri, 23 Jul 1999 11:52:47 -0400 (EDT) From: "Matthew N. Dodd" To: matts-hardware@jurai.net Subject: ALR Revoloution 6x6 motherboards. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-matts-hardware Precedence: bulk Reply-To: "Matthew N. Dodd" http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?ViewItem&item=135126274 More info on this board: http://www.alr.com/service/support/MEMORY/rev_6x6.htm This is a 6 way Pentium Pro motherboard with 8 PCI, 5 EISA slots, builtin IDE and an Adaptec 7880 controller. The seller is selling the system with the DIMM card and 2 TRI-6 CPU boards. While he is offering them at auction, I emailed him and asked him how much he would sell boards for outside of auction. I'm getting one at $104.00 (s/h included). I'm pretty sure that PC100 ECC DIMMs will work on the DIMM board and I've found places that are selling PPro200/256k for $41.00 each (pulls, matched quads). The only other thing I"m unsure of is the power supply requirements and just how big the case is going to be. Anyhow, if you've been looking for a monster SMP box, I don't think you'll be able to put together one with as many CPUs for as little as this setup. === end quoted message === I've also discovered that an 'I/O board' is required for this machine since it doesn't have the keyboard port on the motherboard. ALR/Gateway probably has more. I'll be tracking down the info and posting it to matts-hardware if anyone is interested. -- | Matthew N. Dodd | '78 Datsun 280Z | '75 Volvo 164E | FreeBSD/NetBSD | | winter@jurai.net | 2 x '84 Volvo 245DL | ix86,sparc,pmax | | http://www.jurai.net/~winter | This Space For Rent | ISO8802.5 4ever | To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 13:42:53 1999 Delivered-To: freebsd-smp@freebsd.org Received: from thomson.iqm.unicamp.br (thomson.iqm.unicamp.br [143.106.13.161]) by hub.freebsd.org (Postfix) with ESMTP id 634A814E50 for ; Fri, 23 Jul 1999 13:42:46 -0700 (PDT) (envelope-from fabio@thomson.iqm.unicamp.br) Received: (from fabio@localhost) by thomson.iqm.unicamp.br (8.9.3/8.9.2) id RAA20231; Fri, 23 Jul 1999 17:39:32 -0300 (EST) (envelope-from fabio) From: Fabio Cesar Gozzo Message-Id: <199907232039.RAA20231@thomson.iqm.unicamp.br> Subject: Dual PIII motherboard To: freebsd-smp@freebsd.org Date: Fri, 23 Jul 1999 17:39:32 -0300 (EST) Cc: fabio@iqm.unicamp.br X-Mailer: ELM [version 2.4ME+ PL43 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Hi, I'm planning to buy a dual PIII 550 Mhz and I am looking for advices on manufacturer/model. I've been using ASUS already, but I was browsing their home pages and didn't find a dual board supporting 550 Mhz processors. I found an Intel board, but I have seen people complaining about the on board NCR SCSI. Any suggestions ? Thank you, -- ************************************************** Fabio Gozzo fabio@iqm.unicamp.br State University of Campinas UNICAMP Chemistry Institute http://thomson.iqm.unicamp.br ************************************************** To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 16:22: 9 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp03.primenet.com (smtp03.primenet.com [206.165.6.133]) by hub.freebsd.org (Postfix) with ESMTP id 88A0A14F11 for ; Fri, 23 Jul 1999 16:22:04 -0700 (PDT) (envelope-from tlambert@usr09.primenet.com) Received: (from daemon@localhost) by smtp03.primenet.com (8.8.8/8.8.8) id QAA19706 for ; Fri, 23 Jul 1999 16:19:58 -0700 (MST) Received: from usr09.primenet.com(206.165.6.209) via SMTP by smtp03.primenet.com, id smtpd019692; Fri Jul 23 16:19:49 1999 Received: (from tlambert@localhost) by usr09.primenet.com (8.8.5/8.8.5) id QAA14623; Fri, 23 Jul 1999 16:18:28 -0700 (MST) From: Terry Lambert Message-Id: <199907232318.QAA14623@usr09.primenet.com> Subject: Re: slow network reading with SMP To: grady@xcf.berkeley.edu (Steven Grady) Date: Fri, 23 Jul 1999 23:18:27 +0000 (GMT) Cc: freebsd-smp@FreeBSD.ORG In-Reply-To: <19990723044445.2C90A14FAE@hub.freebsd.org> from "Steven Grady" at Jul 22, 99 09:17:39 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org [ ... 50k/S vs. 6M/S ... ] > The only difference between the kernels was the deletion of the following > lines: > options SMP > options APIC_IO > > Why is this happening? And more importantly, what can we do to use > both processors while still keeping full network speed? (Others in the > company have mentioned Linux -- I'd like to avoid that.) You should diff the dmesg output for both kernels. Let us know it it is complaining about the clock not being routed via APIC. There is another configuration option (see LINT) that will let you fix this, if this is the case. There are also a number of BIOS settings that can effect the performance. Specifically, there are BIOS' where you can select an OS by name; I believe "UnixWare" works best on most of these, but I can't be specific as to what you should select to make the BIOS happy without detailed knowledge of your BIOS (if you obtain that, your in a better positition to decide, anyway). Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 16:22:38 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp02.primenet.com (smtp02.primenet.com [206.165.6.132]) by hub.freebsd.org (Postfix) with ESMTP id AB6C915703 for ; Fri, 23 Jul 1999 16:22:32 -0700 (PDT) (envelope-from tlambert@usr09.primenet.com) Received: (from daemon@localhost) by smtp02.primenet.com (8.8.8/8.8.8) id QAA02612; Fri, 23 Jul 1999 16:22:06 -0700 (MST) Received: from usr09.primenet.com(206.165.6.209) via SMTP by smtp02.primenet.com, id smtpd002558; Fri Jul 23 16:21:59 1999 Received: (from tlambert@localhost) by usr09.primenet.com (8.8.5/8.8.5) id QAA14822; Fri, 23 Jul 1999 16:21:54 -0700 (MST) From: Terry Lambert Message-Id: <199907232321.QAA14822@usr09.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: yokota@zodiac.mech.utsunomiya-u.ac.jp (Kazutaka YOKOTA) Date: Fri, 23 Jul 1999 23:21:54 +0000 (GMT) Cc: jedgar@fxp.org, freebsd-smp@FreeBSD.ORG, yokota@zodiac.mech.utsunomiya-u.ac.jp In-Reply-To: <199907230445.NAA03665@zodiac.mech.utsunomiya-u.ac.jp> from "Kazutaka YOKOTA" at Jul 23, 99 01:45:22 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > "VT number 8" means /dev/ttyv7. > > This means that if xdm if run before any getty when the system is > staring up, it may grab a vty which will be subsequently used by a > getty process. 1) Did it lock up your console when this happened? 2) Why doesn't getty use O_EXCL on the open, such that if it can't get an exclusive lock, it fails? Worst case for this would be that you get a "respawning too rapidly" message, but the system would not lock. 3) Why does it only happen in SMP? Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 16:44:52 1999 Delivered-To: freebsd-smp@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2]) by hub.freebsd.org (Postfix) with ESMTP id DB46514F11 for ; Fri, 23 Jul 1999 16:44:50 -0700 (PDT) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.9.3/8.9.1) id QAA28422; Fri, 23 Jul 1999 16:43:35 -0700 (PDT) (envelope-from dillon) Date: Fri, 23 Jul 1999 16:43:35 -0700 (PDT) From: Matthew Dillon Message-Id: <199907232343.QAA28422@apollo.backplane.com> To: Terry Lambert Cc: yokota@zodiac.mech.utsunomiya-u.ac.jp (Kazutaka YOKOTA), jedgar@fxp.org, freebsd-smp@FreeBSD.ORG Subject: Re: SMP + XDM = keyboard lockup References: <199907232321.QAA14822@usr09.primenet.com> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org :3) Why does it only happen in SMP? : : : Terry Lambert : terry@lambert.org I'm sure the answer to #3 is simply that the startup order winds up being different w/SMP due to there being two cpu's. I've had the X/keyboard/getty problem occur on UP machines and, in fact one of my first UP machines had this race condition unknowingly and *never* had a failure, but the same configuration on a faster UP box a little later led to the getty winning the race and I spent a few hours scratching my head trying to figure out what was wrong with X :-) -Matt Matthew Dillon To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 17: 0:31 1999 Delivered-To: freebsd-smp@freebsd.org Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.26.10.9]) by hub.freebsd.org (Postfix) with ESMTP id 2E73D14F11 for ; Fri, 23 Jul 1999 17:00:27 -0700 (PDT) (envelope-from bde@godzilla.zeta.org.au) Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.7/8.8.7) id JAA29461; Sat, 24 Jul 1999 09:59:58 +1000 Date: Sat, 24 Jul 1999 09:59:58 +1000 From: Bruce Evans Message-Id: <199907232359.JAA29461@godzilla.zeta.org.au> To: tlambert@primenet.com, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup Cc: freebsd-smp@FreeBSD.ORG, jedgar@fxp.org Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >2) Why doesn't getty use O_EXCL on the open, such that if > it can't get an exclusive lock, it fails? Worst case Perhaps because O_EXCL has no effect for devices under BSD (its effect is implementation-defined except for regular files). Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 17:13:44 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp04.primenet.com (smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (Postfix) with ESMTP id 12DAF1570B for ; Fri, 23 Jul 1999 17:13:27 -0700 (PDT) (envelope-from tlambert@usr09.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id RAA10128; Fri, 23 Jul 1999 17:13:23 -0700 (MST) Received: from usr09.primenet.com(206.165.6.209) via SMTP by smtp04.primenet.com, id smtpd010104; Fri Jul 23 17:13:16 1999 Received: (from tlambert@localhost) by usr09.primenet.com (8.8.5/8.8.5) id RAA16496; Fri, 23 Jul 1999 17:13:15 -0700 (MST) From: Terry Lambert Message-Id: <199907240013.RAA16496@usr09.primenet.com> Subject: Re: SMP + XDM problem resolved! To: scheper@toad.net (Richard A. Scheper) Date: Sat, 24 Jul 1999 00:13:15 +0000 (GMT) Cc: freebsd-smp@FreeBSD.ORG In-Reply-To: <37989106.C7758997@toad.net> from "Richard A. Scheper" at Jul 23, 99 11:57:58 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > I guess the fact that it worked without SMP in its incorrect setup had > to do with timing issues after all. SMP just revealed the mistake. Thank god! Dodged that bullet, at least until we try to port FreeBSD/SMP to an MEI architecture... Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 17:16:39 1999 Delivered-To: freebsd-smp@freebsd.org Received: from smtp01.primenet.com (smtp01.primenet.com [206.165.6.131]) by hub.freebsd.org (Postfix) with ESMTP id 93A8615712 for ; Fri, 23 Jul 1999 17:16:29 -0700 (PDT) (envelope-from tlambert@usr09.primenet.com) Received: (from daemon@localhost) by smtp01.primenet.com (8.8.8/8.8.8) id RAA11668; Fri, 23 Jul 1999 17:14:35 -0700 (MST) Received: from usr09.primenet.com(206.165.6.209) via SMTP by smtp01.primenet.com, id smtpd011650; Fri Jul 23 17:14:32 1999 Received: (from tlambert@localhost) by usr09.primenet.com (8.8.5/8.8.5) id RAA16548; Fri, 23 Jul 1999 17:14:32 -0700 (MST) From: Terry Lambert Message-Id: <199907240014.RAA16548@usr09.primenet.com> Subject: Re: SMP + XDM = keyboard lockup To: bde@zeta.org.au (Bruce Evans) Date: Sat, 24 Jul 1999 00:14:31 +0000 (GMT) Cc: tlambert@primenet.com, yokota@zodiac.mech.utsunomiya-u.ac.jp, freebsd-smp@FreeBSD.ORG, jedgar@fxp.org In-Reply-To: <199907232359.JAA29461@godzilla.zeta.org.au> from "Bruce Evans" at Jul 24, 99 09:59:58 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > >2) Why doesn't getty use O_EXCL on the open, such that if > > it can't get an exclusive lock, it fails? Worst case > > Perhaps because O_EXCL has no effect for devices under BSD (its effect > is implementation-defined except for regular files). Hmmm... would you commit patches? This seems to be a bug, to me. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 17:43:23 1999 Delivered-To: freebsd-smp@freebsd.org Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.26.10.9]) by hub.freebsd.org (Postfix) with ESMTP id 965C714F01 for ; Fri, 23 Jul 1999 17:43:18 -0700 (PDT) (envelope-from bde@godzilla.zeta.org.au) Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.7/8.8.7) id KAA32109; Sat, 24 Jul 1999 10:41:41 +1000 Date: Sat, 24 Jul 1999 10:41:41 +1000 From: Bruce Evans Message-Id: <199907240041.KAA32109@godzilla.zeta.org.au> To: bde@zeta.org.au, tlambert@primenet.com Subject: Re: SMP + XDM = keyboard lockup Cc: freebsd-smp@FreeBSD.ORG, jedgar@fxp.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >> >2) Why doesn't getty use O_EXCL on the open, such that if >> > it can't get an exclusive lock, it fails? Worst case >> >> Perhaps because O_EXCL has no effect for devices under BSD (its effect >> is implementation-defined except for regular files). > >Hmmm... would you commit patches? This seems to be a bug, to me. Perhaps if the patches are well written and complete with documentation :-). Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 18:43:33 1999 Delivered-To: freebsd-smp@freebsd.org Received: from scientia.demon.co.uk (scientia.demon.co.uk [212.228.14.13]) by hub.freebsd.org (Postfix) with ESMTP id DC0C015760; Fri, 23 Jul 1999 18:43:24 -0700 (PDT) (envelope-from ben@scientia.demon.co.uk) Received: from lithium.scientia.demon.co.uk ([192.168.1.2] ident=exim) by scientia.demon.co.uk with esmtp (Exim 3.02 #1) id 117hjB-0003vb-00; Fri, 23 Jul 1999 16:59:17 +0100 (envelope-from ben@lithium.scientia.demon.co.uk) Received: from lithium.scientia.demon.co.uk (ident=ben) by lithium.scientia.demon.co.uk with local (Exim 3.02 #1) id 117hjB-0000lA-00; Fri, 23 Jul 1999 16:59:17 +0100 (envelope-from ben@lithium.scientia.demon.co.uk) Date: Fri, 23 Jul 1999 16:59:17 +0100 From: Ben Smithurst To: sthaug@nethelp.no Cc: freebsd-hackers@freebsd.org, freebsd-smp@freebsd.org Subject: Re: rndcontrol and SMP Message-ID: <19990723165917.A2862@lithium.scientia.demon.co.uk> References: <57419.932675097@verdi.nethelp.no> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.95.4i In-Reply-To: <57419.932675097@verdi.nethelp.no> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org sthaug@nethelp.no wrote: > ! warn(""); That should (probably) be `warn(NULL);', otherwise you'd get something like `rndcontrol: : error' rather than the (probably) desired `rndcontrol: error'. (Or so my simple test showed, at least...) -- Ben Smithurst | PGP: 0x99392F7D ben@scientia.demon.co.uk | key available from keyservers and | ben+pgp@scientia.demon.co.uk To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Fri Jul 23 21:43:56 1999 Delivered-To: freebsd-smp@freebsd.org Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.26.10.9]) by hub.freebsd.org (Postfix) with ESMTP id BF4F614FA6; Fri, 23 Jul 1999 21:43:50 -0700 (PDT) (envelope-from bde@godzilla.zeta.org.au) Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.7/8.8.7) id OAA14734; Sat, 24 Jul 1999 14:43:09 +1000 Date: Sat, 24 Jul 1999 14:43:09 +1000 From: Bruce Evans Message-Id: <199907240443.OAA14734@godzilla.zeta.org.au> To: freebsd-hackers@FreeBSD.ORG, freebsd-smp@FreeBSD.ORG, sthaug@nethelp.no Subject: Re: rndcontrol and SMP Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > /* > * XXX the data is 16-bit due to a historical botch, so we use > * magic 16's instead of ICU_LEN and can't support 24 interrupts > * under SMP. > */ > intr = *(int16_t *)data; > if (cmd != MEM_RETURNIRQ && (intr < 0 || intr >= 16)) > return (EINVAL); > >What is needed to make this support a more sensible number of IRQs? Mainly changing the ioctl and its clients (rndcontrol only?) to supply more bits. Bruce To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Jul 24 0: 8:13 1999 Delivered-To: freebsd-smp@freebsd.org Received: from scam.xcf.berkeley.edu (scam.XCF.Berkeley.EDU [128.32.43.201]) by hub.freebsd.org (Postfix) with SMTP id 8728514E0B for ; Sat, 24 Jul 1999 00:08:09 -0700 (PDT) (envelope-from grady@scam.XCF.Berkeley.EDU) Received: (qmail 6141 invoked by uid 348); 24 Jul 1999 06:41:10 -0000 Received: from localhost (HELO scam.XCF.Berkeley.EDU) (sendmail-bs@127.0.0.1) by localhost with SMTP; 24 Jul 1999 06:41:10 -0000 To: freebsd-smp@FreeBSD.ORG Subject: Re: slow network reading with SMP From: grady@xcf.berkeley.edu (Steven Grady) In-reply-to: Your message of Fri, 23 Jul 1999 23:18:27 +0000 (GMT) <199907232318.QAA14623@usr09.primenet.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-ID: <6137.932798469.1@scam.XCF.Berkeley.EDU> Date: Fri, 23 Jul 1999 23:41:09 -0700 Message-Id: <19990724070810.8728514E0B@hub.freebsd.org> Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Thanks for the reply. > You should diff the dmesg output for both kernels. The non-trivial diffs are approximately (for various reasons, this is hand-typed...): diff SMP.dmesg NOSMP.dmesg: < CPU: Pentium III (686-class CPU) > CPU: Pentium III (451.02-MHz 686-class CPU) < Programming 24 pins in IOAPIC #0 < FreeBSD/SMP: Multiprocesor motherboard < cpu0 (BSP): apic id: 0, version: 0x00040011, at 0xfee00000 < cpu1 (AP): apic id: 1, version: 0x00040011, at 0xfee00000 < io0 (APIC): apic id: 2, version: 0x00170011, at 0xfec00000 < pn0: <82c169 PNIC 10/100BaseTX> rev 0x21 int a irq 18 on pci0.13.0 > pn0: <82c169 PNIC 10/100BaseTX> rev 0x21 int a irq 9 on pci0.13.0 < ahc0: rev 0x00 int a irq 16 on pci0.14.0 > ahc0: rev 0x00 int a irq 10 on pci0.14.0 < vga0: rev 0x01 int a irq 19 on pci0.20.0 > vga0: rev 0x01 int a irq 11 on pci0.20.0 < APIC_IO: Testing 8254 interrupt delivery < APIC_IO: routing 8254 via pin 2 < SMP: AP CPU #1 Launched! Sadly, I know very little about SMP, so I don't understand why different irq's are being used, but I assume that the difference is expected. > Let us know it it is complaining about the clock not being > routed via APIC. There is another configuration option (see > LINT) that will let you fix this, if this is the case. Doesn't look like it to me.. > There are also a number of BIOS settings that can effect the > performance. Specifically, there are BIOS' where you can > select an OS by name; I believe "UnixWare" works best on most > of these, but I can't be specific as to what you should select > to make the BIOS happy without detailed knowledge of your > BIOS (if you obtain that, your in a better positition to decide, > anyway). It was hard to find the BIOS version. It's AMIBIOS, but I couldn't find a version number -- the best I could find was the setup version, which was R2.0. There are a _lot_ of options (like > 100), most of them with cryptic abbreviations that seemed irrelevant. There was no option for selecting an OS -- the closest were options about whether the OS is Plug-And-Play compatible (I said yes), and whether it should boot into OS/2 (I said no). If there are particular things I should look for, I'd be more than glad to provide further information. -- Steven grady@xcf.berkeley.edu "It's for you, McGruff!" "Did you hear what he called me? I HATE that! Let's sneak up to his room later and drain all the liquid out of his body." To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Jul 24 1:13:31 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id 17ABA15082 for ; Sat, 24 Jul 1999 01:13:23 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:kAbke/o+MiYTacAkGQVF982g81ttK2JF@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id RAA18834; Sat, 24 Jul 1999 17:13:12 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id RAA18855; Sat, 24 Jul 1999 17:17:28 +0900 (JST) Message-Id: <199907240817.RAA18855@zodiac.mech.utsunomiya-u.ac.jp> To: Bruce Evans Cc: tlambert@primenet.com, freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Sat, 24 Jul 1999 09:59:58 +1000." <199907232359.JAA29461@godzilla.zeta.org.au> References: <199907232359.JAA29461@godzilla.zeta.org.au> Date: Sat, 24 Jul 1999 17:17:28 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >>2) Why doesn't getty use O_EXCL on the open, such that if >> it can't get an exclusive lock, it fails? Worst case > >Perhaps because O_EXCL has no effect for devices under BSD (its effect >is implementation-defined except for regular files). Isn't this O_EXLOCK rather than O_EXCL? (open(2)) In any case, we don't seem to have exclusive lock for non-regular files. (flock(2)) syscons observes TS_XCLUDE just like tty(4), but it is not useful for getty and xdm (and the X server) because they are run by root. TIOCEXCL void Set exclusive use on the terminal. No further opens are per- mitted except by root. Of course, this means that programs that are run by root (or setuid) will not obey the exclusive setting - which limits the usefulness of this feature. Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Jul 24 1:18:31 1999 Delivered-To: freebsd-smp@freebsd.org Received: from outmail.utsunomiya-u.ac.jp (outmail.utsunomiya-u.ac.jp [160.12.196.3]) by hub.freebsd.org (Postfix) with ESMTP id 538B41501A for ; Sat, 24 Jul 1999 01:18:25 -0700 (PDT) (envelope-from yokota@zodiac.mech.utsunomiya-u.ac.jp) Received: from zodiac.mech.utsunomiya-u.ac.jp (IDENT:LJZtsgeGLezTuEk9W6MG/SIv1SAke9e/@zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by outmail.utsunomiya-u.ac.jp (8.9.3/3.7Wpl2) with ESMTP id RAA18906; Sat, 24 Jul 1999 17:17:19 +0900 (JST) Received: from zodiac.mech.utsunomiya-u.ac.jp (zodiac.mech.utsunomiya-u.ac.jp [160.12.42.1]) by zodiac.mech.utsunomiya-u.ac.jp (8.7.6+2.6Wbeta7/3.4W/zodiac-May96) with ESMTP id RAA18933; Sat, 24 Jul 1999 17:21:37 +0900 (JST) Message-Id: <199907240821.RAA18933@zodiac.mech.utsunomiya-u.ac.jp> To: Terry Lambert Cc: freebsd-smp@freebsd.org, yokota@zodiac.mech.utsunomiya-u.ac.jp Subject: Re: SMP + XDM = keyboard lockup In-reply-to: Your message of "Fri, 23 Jul 1999 23:21:54 GMT." <199907232321.QAA14822@usr09.primenet.com> References: <199907232321.QAA14822@usr09.primenet.com> Date: Sat, 24 Jul 1999 17:21:36 +0900 From: Kazutaka YOKOTA Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >> "VT number 8" means /dev/ttyv7. >> >> This means that if xdm if run before any getty when the system is >> staring up, it may grab a vty which will be subsequently used by a >> getty process. > >1) Did it lock up your console when this happened? No. And it wasn't my intention of the experiment. I wanted to demonstrate that the X server still searches for an empty vty even if xdm is run in a particular vty specified in /etc/ttys. >3) Why does it only happen in SMP? I explained several times. This is a timing problem which exists in both UP and SMP environment. It is just that the problem reporter observed it with the SMP kernel in his set up. Kazu To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Jul 24 6:47:33 1999 Delivered-To: freebsd-smp@freebsd.org Received: from gratis.grondar.za (gratis.grondar.za [196.7.18.65]) by hub.freebsd.org (Postfix) with ESMTP id F2CFB14D40; Sat, 24 Jul 1999 06:47:21 -0700 (PDT) (envelope-from mark@grondar.za) Received: from grondar.za (localhost [127.0.0.1]) by gratis.grondar.za (8.9.3/8.9.3) with ESMTP id PAA58793; Sat, 24 Jul 1999 15:46:26 +0200 (SAST) (envelope-from mark@grondar.za) Message-Id: <199907241346.PAA58793@gratis.grondar.za> To: Bruce Evans Cc: freebsd-hackers@FreeBSD.ORG, freebsd-smp@FreeBSD.ORG, sthaug@nethelp.no Subject: Re: rndcontrol and SMP Date: Sat, 24 Jul 1999 15:46:25 +0200 From: Mark Murray Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > >What is needed to make this support a more sensible number of IRQs? > > Mainly changing the ioctl and its clients (rndcontrol only?) to supply > more bits. I am currently rewriting /dev/random (and rndcontrol). M -- Mark Murray Join the anti-SPAM movement: http://www.cauce.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Jul 24 8:28:36 1999 Delivered-To: freebsd-smp@freebsd.org Received: from rebel.net.au (rebel.rebel.net.au [203.20.69.66]) by hub.freebsd.org (Postfix) with ESMTP id C95D114A13 for ; Sat, 24 Jul 1999 08:28:16 -0700 (PDT) (envelope-from kkenn@rebel.net.au) Received: from 203.20.69.78 (dialup-8.rebel.net.au [203.20.69.78]) by rebel.net.au (8.8.5/8.8.4) with ESMTP id AAA25754 for ; Sun, 25 Jul 1999 00:56:58 +0930 Received: (qmail 49866 invoked from network); 24 Jul 1999 15:26:29 -0000 Received: from localhost (kkenn@127.0.0.1) by localhost with SMTP; 24 Jul 1999 15:26:29 -0000 Date: Sun, 25 Jul 1999 00:56:28 +0930 (CST) From: Kris Kennaway Reply-To: kkenn@rebel.net.au To: Mark Murray Cc: Bruce Evans , freebsd-hackers@FreeBSD.ORG, freebsd-smp@FreeBSD.ORG, sthaug@nethelp.no Subject: Re: rndcontrol and SMP In-Reply-To: <199907241346.PAA58793@gratis.grondar.za> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org On Sat, 24 Jul 1999, Mark Murray wrote: > > >What is needed to make this support a more sensible number of IRQs? > > > > Mainly changing the ioctl and its clients (rndcontrol only?) to supply > > more bits. > > I am currently rewriting /dev/random (and rndcontrol). When you say rewriting, do you mean syncing with the version of the code in Linux (1.04, instead of our 0.95) or actually rewriting? If the latter, I'm curious as to what your aims are. Kris To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message From owner-freebsd-smp Sat Jul 24 9: 9:41 1999 Delivered-To: freebsd-smp@freebsd.org Received: from gratis.grondar.za (gratis.grondar.za [196.7.18.65]) by hub.freebsd.org (Postfix) with ESMTP id B139B150DB; Sat, 24 Jul 1999 09:09:30 -0700 (PDT) (envelope-from mark@grondar.za) Received: from grondar.za (localhost [127.0.0.1]) by gratis.grondar.za (8.9.3/8.9.3) with ESMTP id SAA59761; Sat, 24 Jul 1999 18:07:36 +0200 (SAST) (envelope-from mark@grondar.za) Message-Id: <199907241607.SAA59761@gratis.grondar.za> To: kkenn@rebel.net.au Cc: Bruce Evans , freebsd-hackers@FreeBSD.ORG, freebsd-smp@FreeBSD.ORG, sthaug@nethelp.no Subject: Re: rndcontrol and SMP Date: Sat, 24 Jul 1999 18:07:36 +0200 From: Mark Murray Sender: owner-freebsd-smp@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > When you say rewriting, do you mean syncing with the version of the code > in Linux (1.04, instead of our 0.95) or actually rewriting? If the latter, > I'm curious as to what your aims are. I want to implement Bruce Schneier's Yarrow. M -- Mark Murray Join the anti-SPAM movement: http://www.cauce.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-smp" in the body of the message