many lost ticks, clock drifts problems also with debian smp kernel?
Hi, I have a machine with an AMD64 x2 4800+ mounted on an asus A8V Deluxe. I have experienced massive clock drift problems with smp kernels from the fedora core 4 x86_64 distribution, also with the newer 2.6.14.3 kernel from kernel.org: -- dmesg : Losing some ticks... checking if CPU frequency changed. warning: many lost ticks. Your time source seems to be instable or some driver is hogging interupts rip 0x435036 --- I have tried the pci=routeirq, acpi=off, noapic, nolapic, no_timer_check, no_tsc... bootparams without any real succes, my system clock keeps on drifting away faster and faster after the system has been up for some time. My question to the list is : Are the debian amd64-k8-smp kernels also subject to this problem? I have browsed the entirety of the 2005 amd64 lists and have found various posts about clock problems, but they were all laptop related, with the non smp kernel. So far I haven't found any debian user reporting the same problme I face now, but who knows, I'd like to make this sure before I break my current fedora install and start a rather time consuming debian install (time consuming because I have never used debian before). Thanks for your help, /rémi -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: many lost ticks, clock drifts problems also with debian smp kernel?
On 10:36 Fri 09 Dec , [EMAIL PROTECTED] wrote: Hi, I have a machine with an AMD64 x2 4800+ mounted on an asus A8V Deluxe. I have experienced massive clock drift problems with smp kernels from the fedora core 4 x86_64 distribution, also with the newer 2.6.14.3 kernel from kernel.org: -- dmesg : Losing some ticks... checking if CPU frequency changed. warning: many lost ticks. Your time source seems to be instable or some driver is hogging interupts rip 0x435036 --- I have tried the pci=routeirq, acpi=off, noapic, nolapic, no_timer_check, no_tsc... bootparams without any real succes, my system clock keeps on drifting away faster and faster after the system has been up for some time. My question to the list is : Are the debian amd64-k8-smp kernels also subject to this problem? This is a kernel problem with SMP having unsynced TSC's running at different freqs, supposedly fixed in kernel 2.6.14. Before 2.6.14 i needed to boot with pci=noacpi and clock=pmtmr to solve the 'lost 'ticks'. A Debian kernel won't change anything. assuming you have a similar config (e.g. HPET_TIMER=y PM_TIMER=y). You may want to post this on the LKML. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: DualCore Dual-Opteron Board - suggestion?
we planned to have dual core dual opteron machine with tyan mobo. however tyans are not available at the moment (and this situation will last at least till the end of the year). we were offered supermicro h8da8 mobo. does anyone have any experience with it, or should we wait for tyan? regards, -- Lubos [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: AMD64 mirror
Fielder George Dowding [EMAIL PROTECTED] writes: Ha! I was just about to rattle your chain about this problem - serious lack of joy in Mudville as a result of errors in the mirroring processes - Release and Release.gpg disappearing. Yes, pool seemed to be populated (40G or so), trying to update from my mirror gave no joy. I changed from debian.csail.mit.edu to ftp.de.debian.org with the result of even less joy. Then I took a look at the debmirror man page and added --ignore-small-errors. I went back to debian.csail.mit.edu and now joy. Here is the error from my original setup using debian.csail.mit.edu: Errors: Release signature does not verify. You should add the archive key to your keyring and verify it. Release signature does not verify. Release signature does not verify. Download of pool/main/libx/libxalan2-java/libxalan2-java_2.6.0.orig.tar.gz failed: Failed to open file. Download of pool/main/t/tomcat5/tomcat5_5.0.30.orig.tar.gz failed: Failed to open file. Failed to download files (2 errors)! The man page pointed out not downloading a file or two is considered a small error. fgd. Those are actualy broken packages. They have to do a new sourcefull upload to get the orig.tar.gz files to be moved from contrib to main. Feel free to report bugs on them if not already present. MfG Goswin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Full Mirror vs apt-proxy
Fielder George Dowding [EMAIL PROTECTED] writes: Greetings all and thank you for the suggestions regarding mirroring the amd64 distribution. Having given some thought to the situation, I have concluded I need to have a full mirror. I have an external USB hard drive on which I now mirror both the i386 (Woody through Sid) and amd64 (Sarge through Sid). I have found that maintaining i386 installations even with a cable or DSL connection is all too often too time consuming which is the reason for the portable mirror in the first place. Also, there are two such installations with dial-up access which alone justifies the portable mirror. At the moment I have only my laptop to maintain, but in the coming year, I should have at least one amd64 server which may not be connected to the Internet at all. Thus, a full, portable, mirror is what I really need. Cheerio! A full mirror is easy enough to setup. I (despite being the debmirror maintainer) prefer reprepro for this as it allows building a single new archive out of multiple sources (debian, debian-amd64, security). apt-get install reprepro cd /mnt/mirror mkdir conf Create config: == conf/distributions == Origin: Debian Label: Debian Suite: stable Codename: sarge Version: 3.1 Architectures: source i386 amd64 Components: main contrib non-free UDebComponents: main Description: Debian sarge Update: - debian-security debian debian-amd64 SignWith: yes Origin: Debian Label: Debian Suite: testing Codename: etch Version: 3.2 Architectures: source i386 amd64 Components: main contrib non-free UDebComponents: main Description: Debian etch Update: - debian debian-amd64 SignWith: yes Origin: Debian Label: Debian Suite: unstable Codename: sid Version: 3.3 Architectures: source i386 amd64 Components: main contrib non-free UDebComponents: main Description: Debian sid Update: - debian debian-amd64 SignWith: yes == conf/updates == Architectures: source i386 mips mipsel m68k Method: http://ftp.de.debian.org/debian VerifyRelease: F1D53D8C4F368D5D Name: debian-amd64 Architectures: source amd64 Components: main contrib non-free Method: http://amd64.debian.net/debian ListHook: /mnt/mirror/debian/conf/filter-amd64.sh VerifyRelease: FBC60EA91B67D3C0 Name: debian-security Architectures: source i386 amd64 mips mipsel m68k Method: http://security.debian.org Suite: */updates UDebComponents: none VerifyRelease: F1D53D8C4F368D5D EOF == conf/filter-amd64.sh == #!/bin/sh echo /dev/tty Filter-amd64: '$1' - '$2' # Screwed up security updates: # apache2 crips gtkdiskfree gpsdrive graphviz lynx weex # Screwed up sources: # gasql gnome-db zcat $1 \ | grep-dctrl -v -F MD5sum b60c968b05bc16bfc15ca06ac41a \ | grep-dctrl -v -F MD5sum bba3d95da8ba3daa757f37beb938beab \ | grep-dctrl -v -F MD5sum 1b0e4ddcacacdfc60b72c3771823dbde \ | grep-dctrl -v -F MD5sum ae61d1b0484a26c877b688dff5158b9e \ | grep-dctrl -v -F MD5sum 97e89c81ef111c4b0f8c1d60e7f6559b \ | grep-dctrl -v -F MD5sum 268d7abc24919e605b7f77b25366cf11 \ | grep-dctrl -v -F MD5sum 01e38a9a43c9c3c3edb0fb2389a634e7 \ | grep-dctrl -v -F MD5sum 8e3e20fef441fe0da1bc5a981b7b8bec \ | grep-dctrl -v -F MD5sum 6bd9495ae8f18fd7a0c81a66fd0f5f38 \ | grep-dctrl -v -F MD5sum b7c8512911f9516788672732575bd127 \ | grep-dctrl -v -F MD5sum e02608ec367dd1bae87199100e090fae \ | grep-dctrl -v -F MD5sum dee2258699f4bae1029e98fd9fcec3e1 \ | grep-dctrl -v -F MD5sum 9fd3190f61ad1d80f1c13aae86f6ed92 \ | grep-dctrl -v -F MD5sum 6a5af608e0007cc026c635c6c4b576be \ | grep-dctrl -v -F MD5sum 938334e84c183b8938d90b80f08812e0 \ | grep-dctrl -v -F MD5sum ca9dac731128e8f09fa43b87935c79f0 \ | grep-dctrl -v -F MD5sum bbb95485471d5ed0eb067e0eff975d8c \ | grep-dctrl -v -F MD5sum 7e36967187ff099ea46a5bb7ca4c85a9 \ | grep-dctrl -v -F MD5sum 597d905716f7579a9a5d0114594c5810 \ | grep-dctrl -v a7402b9e6575d8ebe74e30e0f7f0bd36 \ | grep-dctrl -v eb4bf3b5476abfd9188d0218c372d6ea \ $2 exit 0 chmod a+x conf/filter-amd64.sh reprepro -V -b . update *make coffee, sleep, watch a movie* The filer-amd64.sh removes the duplicate packages between amd64.debian.net and security.debian.net. MfG Goswin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Re: many lost ticks, clock drifts problems also with debian smp kernel?
On Fri, Dec 09, 2005 at 03:06:16PM +0100, [EMAIL PROTECTED] wrote: what is a TSC? Time Stamp Counter or something like that. It is a counter in the cpu that can theoretically be used as a fast way to keep track of time. Most modern CPUs have one. Some count at full cpu clock rate, some count slower. Depends on the design. On multi cpu systems, they are often not in sync, but initially it was thought that dual core AMDs would have them in sync, but that isn't true apparently due to frequency scaling and power management stuff changing the counter speed. what does the clock=pmtmr bootparam do? I think it changes what device the kernel uses to keep track of time. The is the power management timer, the hpet (high performance event timer), the TSC, and a few other choices. Some are faster to read, some are more accurate, and it all depends on the type of system which ones you can even rely on. what is HPET_TIMER=y ? what is PM_TIMER=y ? That would be the two I mentioned above. and what is LKML? Linux Kernel Mailing List. See vger.kernel.org. Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Firewire 800 Cards, PCIe/PCI-X under debian amd64?
On Fri, Dec 09, 2005 at 12:28:58AM +0100, Lars Schimmer wrote: After the board suggestion now the special one: I need 4 Firewire 800 cards on ONE PC. How many ports do you need though? Most cards have multiple ports and the port count varies with the card model. Because of this special situation, I thought about 2 PCIe and two PCI-X firewire 800 cards. Has anyone a suggestion which card works under amd64 flawless at full speed? I have only used onboard firewire 400 personally, which seems to work just fine. Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Using VMWare Player
There is a failed message here. You should ask VMware why it doesn't work, or, if you are longing for some action, open the VMware startup script (/etc/init.d/vmware) and see if you can spot the problem. Hi! Thanks for your hint. I found out I had to install ia32-libs to sucessfully install. Now it's running just fine. Helge -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
OpenOffice-2.0.
Greetings all, I'm looking to install OpenOffice 2.0, anyone using it? Thanks for any replies, Chris W. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: DualCore Dual-Opteron Board - suggestion?
No clue personally, just seems amusing given supermicro I believe is the company that doesn't even want to admit to making AMD compatible boards. :) maybe it changed (see below) You can go through their website looking at models and they only mention intel cpus. once in a while you can click on something, like say the SC812S-420C in the chassis section, clearly labeled as DP Xeon / EM64T in the chart, but once you go to the details of it, you discover that amazingly it apparently can also be had with opterons installed. About the only way to find anything non intel, is by random clicking, or doing a site search for opteron. They sure aren't trying to get you to buy their opteron offerings. Maybe that would make intel mad at them or something. I wonder if they have any experience dealing with opterons yet? I wonder what the service would be like if there is a problem in the bios. well i think that there is some information: http://www.supermicro.com/Aplus/motherboard/ the truth is, that tyan is our favorite :) it's classics after all... and has sata... http://www.tyan.com/products/html/thunderk8sdpro.html however, the question is when it will be available - not possible to get it in czech republic, nor in germany... regards, -- Lubos [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: locked out.
Hi Chris, Just did a fresh install. Now I'm locked out of my machine Do you mean your keyboard is permanently disabled? Is it USB or PS/2? What happens if you try to use the keyboard during the BIOS messages, eg to access the BIOS setup? Or does it only go dead after the Linux kernel has booted? Cheers Daniel -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OpenOffice-2.0.
:o) This one comes up every few days. Basically it goes like this... OOo 2.0 isn't likely to ever find it's way into the AMD64 repository at least not as a 64 bit app. It's not 64 bit clean and, if you do get it compiled, won't run correctly. Your best bet is to install it in a 32 bit chroot. It's not a terribly nice solution IMHO but it does work pretty well. Graham On Friday 09 December 2005 10:42, Chris Wakefield wrote: Greetings all, I'm looking to install OpenOffice 2.0, anyone using it? Thanks for any replies, Chris W. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OpenOffice-2.0.
I guess I could have made myself a little more clear. I understand the 64 bit OO2 progress, thanks. I noticed that OO2 is not offered in any linux binary except those wrapped in rpm's I was thinking that OO2 would come with the same installer as 1.2 didI think it was a _bin_ ? Anyways, I tried converting the OO2 rpm's to debs, but ran into a few problems and was hoping there was a simpler way. By-the-way, I still don't understand why one would need to set up a 32 bit chroot just to run OO when we have 32 bit emulation? Thanks for the ideas and opinions, Chris w. On Friday 09 December 2005 20:10, Graham Smith wrote: :o) This one comes up every few days. Basically it goes like this... OOo 2.0 isn't likely to ever find it's way into the AMD64 repository at least not as a 64 bit app. It's not 64 bit clean and, if you do get it compiled, won't run correctly. Your best bet is to install it in a 32 bit chroot. It's not a terribly nice solution IMHO but it does work pretty well. Graham On Friday 09 December 2005 10:42, Chris Wakefield wrote: Greetings all, I'm looking to install OpenOffice 2.0, anyone using it? Thanks for any replies, Chris W. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: sound-driver
On Fri, Dec 09, 2005 at 08:57:49PM +0100, Rob van Kraanen wrote: Hi i have a Asus a7d laptop. After some searching i found i have a dound-driver with a acore chipset, at least thats what i think. I'm a newby as it comes to compiling drivers, but i got no errors and even in mij config i can see a device: Midi through Midi Through Port-0 ALSE device But i hear nothing, anyone can help me? Midi is for musical instrument connections. It is not for playing sound on your computer. That would be done by the wave support part of the card instead. Of course given it is an ati chipset laptop, it is probably going to take a while to get proper support for the system under linux. In general, ati chips are just not well supported. I would expect the driver to use, if any work yet, would be snd-atiixp alsa driver. The newer the kernel the better the chance of it working is. Try adding that to /etc/modules, so that it loads early in the boot process. alsamixer should work for unmuting master and wave and turning up the volumes if it detects the chip. Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OpenOffice-2.0.
On Fri, Dec 09, 2005 at 12:32:35PM +, Chris Wakefield wrote: I guess I could have made myself a little more clear. I understand the 64 bit OO2 progress, thanks. I noticed that OO2 is not offered in any linux binary except those wrapped in rpm's I was thinking that OO2 would come with the same installer as 1.2 didI think it was a _bin_ ? Anyways, I tried converting the OO2 rpm's to debs, but ran into a few problems and was hoping there was a simpler way. By-the-way, I still don't understand why one would need to set up a 32 bit chroot just to run OO when we have 32 bit emulation? There are 32bit OO2 binaries in sid on i386. No reason to use rpms. Len Sorensen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OpenOffice-2.0.
Chris Wakefield ha scritto: By-the-way, I still don't understand why one would need to set up a 32 bit chroot just to run OO when we have 32 bit emulation? lol I'm trying since three months ago to make this emulation work, but I miss something...I'm stuck on an error saying: javaldx: Could not find a Java Runtime Environment! no suitable windowing system found, exiting. But java is not a strict dep I think! However libgcj6 in 32 bit form is installed!! This fd openoffice...I'm running even Firefox with Flash without chroot!!! -- Alessandro Dal Grande Student In The University Of Padua - Computer Science Linux Registered User #359258 System: GNU/Linux Debian Sid Pure64 on [EMAIL PROTECTED] Chat: Psi ICQ) 150487234 Jabber) [EMAIL PROTECTED] Home: http://v0n0.altervista.org signature.asc Description: OpenPGP digital signature
Re: OpenOffice-2.0.
v0n0 wrote: lol I'm trying since three months ago to make this emulation work, but I miss something...I'm stuck on an error saying: javaldx: Could not find a Java Runtime Environment! no suitable windowing system found, exiting. But java is not a strict dep I think! However libgcj6 in 32 bit form is installed!! This fd openoffice...I'm running even Firefox with Flash without chroot!!! I think the two errors above are unrelated. 10 minutes ago I upgraded openoffice to 2.0 (on my i386 laptop), and I get that java error too but openoffice runs fine otherwise. I haven't investigated why -- ironically, a java program is running on that system at this very moment. Sorry I can't help you with the error, though. I use openoffice chrooted on my amd64 system. -Corey -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OpenOffice-2.0.
On 12/9/05, Chris Wakefield [EMAIL PROTECTED] wrote: ... By-the-way, I still don't understand why one would need to set up a 32 bit chroot just to run OO when we have 32 bit emulation? Sorry, but this is a hot-button of mine. When you're using a AMD64 processor, you have full native 32-bit support. The chip can run 32-bit binaries with NO speed penalties, whereas other architectures require some funky architecture emulators or they slow down. So, in short, you don't have 32-bit emulation, you have native 32-bit support. There's a difference - assuming you're using an AMD processor.
