Re: 4.5 on Thinkpad 600x issue
On Wed, May 27, 2009 at 3:04 PM, Johan Beisser wrote: > On Wed, May 27, 2009 at 10:26 AM, STeve Andre' wrote: > >> I've never tried installing OpenBSD on a 600x but I'm a little surprised > that >> it isn't working fine. > > You're in for a few surprises when you do then. It should work fine, > but there's some ACPI issues that have never been addressed. > >> Since you are new to OpenBSD, how did you get OpenBSD, and also how >> (where) did you get the packages? You MUST get the packages that >> match the version of OpenBSD. More than one person has gotten a >> release CD and then gotten the packages in snapshots/packages/i386 >> which is "-current", the wip stuff that will be a part of the next release. > > The 600x has a CDRom/DVD drive in it. It comes standard. > >> Also, it would be good to post the contents of /var/run/dmesg.boot, to >> see what the kernel thinks of the hardware. Thats a start. > > I'll include something I sent to Donald Allen, edited to make things a > little more contextually relevant: > > "The key problem would keep happening [the freezing/slowdown]. Mostly > due to IRQ 11 being shared between USB, keyboard and PCMCIA. Large > amounts of traffic through that IRQ would cause locking issues in the > kernel. It really > is a hardware issue with that specific model of laptop; I had them > with FreeBSD [5.2], OpenBSD [4.1, 4.2, and 4.3], and Linux [2.6.10]." > > It's a problem I presumed was just with my 600x, but some of my > research has shown it's a model issue, related to IRQ assignment in > kernel. The only OS that hasn't had a problem with the hardware is > Windows XP. Whether that's due to the OS masking it or knowing > something more intimately about the odd hybrid of ACPI and APM the > BIOS presents, I can't say. I ran Windows 2000 on my 600x for a time and that also worked well (at least with respect to the issue at hand). Having said that, I don't think I've ever run this machine in the configuration I was subjecting OpenBSD to -- wired to the network via pcmcia and moving a lot of data through that wire (I've usually used this machine wireless, thus lower bandwidth and interrupt frequency). So I can't say that I've ever subjected any other OS to the conditions that produced trouble with OpenBSD. I will be restoring the Arch Linux disk to this machine shortly. If I have a chance, I'll test it with the ethernet card and see if I can kill it. One way or another, I'll post what I find. As for OpenBSD, I've installed it on my TP G41 and have run across a couple of small problems (the Sawfish window manager does not work correctly; I've gone back to TWM -- I'm a minimalist; RCS does not work correctly with emacs VC -- I haven't figured that one out yet, though I suspect it's some small incompatibility in the OpenBSD reimplementation of RCS). Other than that, the system looks great thus far, and I'm probably going to proceed with replacing FreeBSD with it on my two main machines (a Thinkpad X61 and a Lenovo workstation). FreeBSD has serious problems with its USB support (which is being completely reimplemented in the upcoming version 8 release) and I do backups on sata drives in USB shoeboxes. This has forced me into a complicated system for backing up, involving an Arch Linux LiveCD and rsyncing my home directory to the backup disk with a Linux system, because I don't trust FreeBSD writing to those drives. I've done some preliminary testing with OpenBSD and have had no problems talking to those drives. If more testing turns up no problems, I'm sold. Thank you again for your help with the 600x issue. /Don Allen > > I'm just not surprised the problem still exists in 4.5.
Re: 4.5 on Thinkpad 600x issue
On Wed, May 27, 2009 at 10:26 AM, STeve Andre' wrote: > I've never tried installing OpenBSD on a 600x but I'm a little surprised that > it isn't working fine. You're in for a few surprises when you do then. It should work fine, but there's some ACPI issues that have never been addressed. > Since you are new to OpenBSD, how did you get OpenBSD, and also how > (where) did you get the packages? You MUST get the packages that > match the version of OpenBSD. More than one person has gotten a > release CD and then gotten the packages in snapshots/packages/i386 > which is "-current", the wip stuff that will be a part of the next release. The 600x has a CDRom/DVD drive in it. It comes standard. > Also, it would be good to post the contents of /var/run/dmesg.boot, to > see what the kernel thinks of the hardware. Thats a start. I'll include something I sent to Donald Allen, edited to make things a little more contextually relevant: "The key problem would keep happening [the freezing/slowdown]. Mostly due to IRQ 11 being shared between USB, keyboard and PCMCIA. Large amounts of traffic through that IRQ would cause locking issues in the kernel. It really is a hardware issue with that specific model of laptop; I had them with FreeBSD [5.2], OpenBSD [4.1, 4.2, and 4.3], and Linux [2.6.10]." It's a problem I presumed was just with my 600x, but some of my research has shown it's a model issue, related to IRQ assignment in kernel. The only OS that hasn't had a problem with the hardware is Windows XP. Whether that's due to the OS masking it or knowing something more intimately about the odd hybrid of ACPI and APM the BIOS presents, I can't say. I'm just not surprised the problem still exists in 4.5.
Re: 4.5 on Thinkpad 600x issue
On Wed, May 27, 2009 at 1:26 PM, STeve Andre' wrote: > On Wednesday 27 May 2009 13:12:26 you wrote: >> Update: rsync completed. I brought up X, Firefox, emacs and was downloading >> packages when the curse struck again. Little or no response to the mouse. >> ctrl-alt F2 got me to a fresh login prompt, but every character I type is >> repeated 7 times, so login is impossible. No response to ping and, not >> surprisingly, I can't ssh in. It occurs to me that I'm using a pcmcia 3com >> ethernet card that I haven't used in years and that I don't use when I run >> Linux on this machine (I use a wireless card in that case, but wasn't ready >> to tackle wireless vs. OpenBSD just yet), so that's another hardware >> difference. I suspect that this is just crufty old hardware acting up. I >> think to debug this I will install OpenBSD on another machine I have that >> is >> >> - newer >> - there will be no hardware variation and it is all known to be good (Linux >> and FreeBSD have both run reliably on that machine). >> >> /Don > > I've never tried installing OpenBSD on a 600x but I'm a little surprised that > it isn't working fine. > > Since you are new to OpenBSD, how did you get OpenBSD, and also how > (where) did you get the packages? ftp://ftp.cse.buffalo.edu/pub/OpenBSD/4.5/i386/ Downloaded cd45.iso and burned that to a cd. Downloaded bsd plus *45.tgz (except for the games) and wrote them to a cd (4.5/i386 directory). I probably should have included INSTALL.i386, since the installer noticed that it was missing from the cd, but other than making it grumpy, it seemed to do no harm. You MUST get the packages that > match the version of OpenBSD. More than one person has gotten a > release CD and then gotten the packages in snapshots/packages/i386 > which is "-current", the wip stuff that will be a part of the next release. > > Also, it would be good to post the contents of /var/run/dmesg.boot, to > see what the kernel thinks of the hardware. Thats a start. Ok, will do in a separate msg. /Don > > --STeve Andre'
Re: 4.5 on Thinkpad 600x issue
On Wednesday 27 May 2009 13:12:26 you wrote: > Update: rsync completed. I brought up X, Firefox, emacs and was downloading > packages when the curse struck again. Little or no response to the mouse. > ctrl-alt F2 got me to a fresh login prompt, but every character I type is > repeated 7 times, so login is impossible. No response to ping and, not > surprisingly, I can't ssh in. It occurs to me that I'm using a pcmcia 3com > ethernet card that I haven't used in years and that I don't use when I run > Linux on this machine (I use a wireless card in that case, but wasn't ready > to tackle wireless vs. OpenBSD just yet), so that's another hardware > difference. I suspect that this is just crufty old hardware acting up. I > think to debug this I will install OpenBSD on another machine I have that > is > > - newer > - there will be no hardware variation and it is all known to be good (Linux > and FreeBSD have both run reliably on that machine). > > /Don I've never tried installing OpenBSD on a 600x but I'm a little surprised that it isn't working fine. Since you are new to OpenBSD, how did you get OpenBSD, and also how (where) did you get the packages? You MUST get the packages that match the version of OpenBSD. More than one person has gotten a release CD and then gotten the packages in snapshots/packages/i386 which is "-current", the wip stuff that will be a part of the next release. Also, it would be good to post the contents of /var/run/dmesg.boot, to see what the kernel thinks of the hardware. Thats a start. --STeve Andre'
Re: 4.5 on Thinkpad 600x issue
Update: rsync completed. I brought up X, Firefox, emacs and was downloading packages when the curse struck again. Little or no response to the mouse. ctrl-alt F2 got me to a fresh login prompt, but every character I type is repeated 7 times, so login is impossible. No response to ping and, not surprisingly, I can't ssh in. It occurs to me that I'm using a pcmcia 3com ethernet card that I haven't used in years and that I don't use when I run Linux on this machine (I use a wireless card in that case, but wasn't ready to tackle wireless vs. OpenBSD just yet), so that's another hardware difference. I suspect that this is just crufty old hardware acting up. I think to debug this I will install OpenBSD on another machine I have that is - newer - there will be no hardware variation and it is all known to be good (Linux and FreeBSD have both run reliably on that machine). /Don On Wed, May 27, 2009 at 12:27 PM, Donald Allen wrote: > I'm an experienced hand with Linux (Gentoo, more recently Arch) and with > FreeBSD. I've recently become interested in OpenBSD and have just done a > test install of 4.5 on an old Thinkpad 600x (650 mhz, .5 Gb, 20 Gb 5400 rpm > disk, 3com Megahertz pcmcia ethernet adapter) for purposes of evaluation. > > Using the system to download and install packages and doing general setup > tasks, it behaves normally, no problems. But today, I am attempting to rsync > (I've arranged for the rsync daemon to be started at boot time) the contents > of my home directory from a FreeBSD system (something I do all the time with > other targets, for backup purposes, and to allow me to use different > machines as appropriate). I've twice had the rsync fail, with the client > complaining that it could not write to its output pipe. The OpenBSD system > was sitting at its login prompt, and attempting to login proved impossible. > Characters got echoed extremely slowly, if at all, and when they did, they > got echoed multiple times. I could not ping the system, though it was up, > but obviously in distress. As an a very experienced systems programmer > (though I haven't done any OS-level work in years), I'd offer the guess, and > its only a guess, that the system was being flooded with interrupts. Unable > to ssh in, I finally just turned the power off and rebooted. After the > fscks, the system came up normally. I checked /var/log/message and found > nothing unusual. I resumed the rsync and ran into the same problem again > after a relatively short time. I am now on my third attempt, this time > running 'top' on the OpenBSD machine, and in the spirit of Heisenberg, the > rsync is proceeding normally, almost finished. > > I normally run Arch Linux on this machine (different disk) and have had no > problems with it (I did the same rsync from the same source machine > uneventfully), so I'm not too inclined to suspect the hardware, old as it > is, except perhaps the disk, which is different hardware than when I run > Linux. > > Here's my question: should I be able to provoke this problem again, can the > collective you suggest things I should be doing, log files I ought to be > looking at, perhaps running with a kernel debugger available, etc., to have > a chance of debugging this problem? It's possible that this old machine or > the disk that's been gathering dust for some time has decided to > malfunction. But since I'm evaluating OpenBSD, I'd like to either exonerate > it or confirm that it's a bug in the system. Any help would be appreciated. > > /Don Allen
4.5 on Thinkpad 600x issue
I'm an experienced hand with Linux (Gentoo, more recently Arch) and with FreeBSD. I've recently become interested in OpenBSD and have just done a test install of 4.5 on an old Thinkpad 600x (650 mhz, .5 Gb, 20 Gb 5400 rpm disk, 3com Megahertz pcmcia ethernet adapter) for purposes of evaluation. Using the system to download and install packages and doing general setup tasks, it behaves normally, no problems. But today, I am attempting to rsync (I've arranged for the rsync daemon to be started at boot time) the contents of my home directory from a FreeBSD system (something I do all the time with other targets, for backup purposes, and to allow me to use different machines as appropriate). I've twice had the rsync fail, with the client complaining that it could not write to its output pipe. The OpenBSD system was sitting at its login prompt, and attempting to login proved impossible. Characters got echoed extremely slowly, if at all, and when they did, they got echoed multiple times. I could not ping the system, though it was up, but obviously in distress. As an a very experienced systems programmer (though I haven't done any OS-level work in years), I'd offer the guess, and its only a guess, that the system was being flooded with interrupts. Unable to ssh in, I finally just turned the power off and rebooted. After the fscks, the system came up normally. I checked /var/log/message and found nothing unusual. I resumed the rsync and ran into the same problem again after a relatively short time. I am now on my third attempt, this time running 'top' on the OpenBSD machine, and in the spirit of Heisenberg, the rsync is proceeding normally, almost finished. I normally run Arch Linux on this machine (different disk) and have had no problems with it (I did the same rsync from the same source machine uneventfully), so I'm not too inclined to suspect the hardware, old as it is, except perhaps the disk, which is different hardware than when I run Linux. Here's my question: should I be able to provoke this problem again, can the collective you suggest things I should be doing, log files I ought to be looking at, perhaps running with a kernel debugger available, etc., to have a chance of debugging this problem? It's possible that this old machine or the disk that's been gathering dust for some time has decided to malfunction. But since I'm evaluating OpenBSD, I'd like to either exonerate it or confirm that it's a bug in the system. Any help would be appreciated. /Don Allen
