December 2003 - openSUSE AMD64 - openSUSE Mailing Lists

fibre channel hba
by Kris Ongbongan 13 Dec '03

13 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 can anyone recommend a fibre channel HBA to use with SuSE Enterprise 8 on Opteron? Any experiences positive/negative with them? Looking to possible hook into an IBM FasT SAN. Kris. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3 (FreeBSD) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQE/2iFkuBLfyXibQuYRAh6wAKCXBQUWqdGAx//mXuzrqoxbcCuvuACdG55s YG6NbaI7eNZogXfblX06pJk= =5+1/ -----END PGP SIGNATURE-----

3 2

4GB RAM with Tyan Thunder K8W
by Reimar Bauer 12 Dec '03

12 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Dear all, I have a problem to get a kernel working with 4GB of RAM. The main board of the dual opteron system is a Tyan Thunder K8W. I tried a lot of kernels. Now the 2.4.23 kernel is running. But it runs only if I set a memory limitation to 2GB as boot option. Is this already a known problem? What should I try next? regards Reimar - -- Reimar Bauer Institut fuer Stratosphaerische Chemie (ICG-I) Forschungszentrum Juelich email: R.Bauer(a)fz-juelich.de - ------------------------------------------------------------------- ~ a IDL library at ForschungsZentrum Juelich ~ http://www.fz-juelich.de/icg/icg-i/idl_icglib/idl_lib_intro.html =================================================================== -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2-rc1-SuSE (GNU/Linux) iD8DBQE/2dvk5aOc3Q9hk/kRAoLEAJ0VhxRNs7EST30mWHVetnRS6GYpYACfevVr z4IRwOLsP5QRLuN+h09fqDc= =zbYO -----END PGP SIGNATURE-----

3 3

Re: [suse-amd64] 4GB RAM with Tyan Thunder K8W
by Andi Kleen 12 Dec '03

12 Dec '03

On Fri, 12 Dec 2003 15:41:35 +0000 (GMT) Andrew Halliwell <ah(a)gnd.com> wrote: [cc'ing back to suse-amd64] > > I've managed to find 4 good DIMMS out of 8 (still checking the last 2, but > with 6 it crashes and 8 it crashes. (kernel panics usually, but also had a > libc corruption which prevented most programs from starting with a seg fault > (fortunately isolated to RAM, not disk)) > > Memtest86 won't work on this board. It starts, then reboots. When you have a two bit error in RAM the system will reboot. That's unavoidable. There are some BIOS who can diagnose this after the reboot ("event log"), but that is normally only available in highend boards. I don't know if the Tyan K8W supports that. > (I note it's not offering memtest at bootup, didn't check the install cd) > > Is there anything that can be done to make this board less sensitive to very > slight variations in RAM? I'm arguing with someone here who just won't > believe it's "just the ram". He's convinced it's linux that's at fault, even > though we've had it running the tests on 2 gig, switched a dimm and it > failed. You would need to ask this question to the designers of the motherboard. It is possible that the BIOS is driving the memory controller too aggressively. In theory it would be possible to change the configuration of the memory controller early in bootup to more conservative values, but that is getting very hardware specific and will depend a lot on the board (= requires knowledge that the kernel doesn't have). You can try if a BIOS update helps. A safe route is to stick only to DIMMs that are approved by the vendor. If these don't work neither you could RMA it. > I only subscribed to this list yesterday, because of this very problem but > as it was hardware related (in my mind at least) I didn't bother posting > about it. Thanks. -Andi

1 0

lm-sensors fix for Tyan K8W
by Emil Briggs 11 Dec '03

11 Dec '03

I sent to a message to the lm_sensors folks but in the meantime this should help anyone who is having trouble getting lm_sensors to work on the Tyan K8W. Emil -------------------------------------------------------------------------- lm_sensors was not working on this motherboard. It uses an AMD8111 chipset which provides both smbus 1.0 and 2.0 controllers. The 2.0 controller module (i2c-amd8111) would load but the 1.0 controller which is compatible with the i2c-amd756 controller would not. According to the Tyan documents the sensor chip is an ADT7463 connected to the 1.0 controller. After poking around through the AMD docs I found the problem. The PCI device ID was not present in the pci_device_id structure. The following patch fixes this and the i2c-amd756 driver loads. --- i2c-amd756.c.orig 2003-12-11 09:38:56.000000000 -0500 +++ i2c-amd756.c 2003-12-11 09:40:11.000000000 -0500 @@ -318,6 +318,7 @@ {PCI_VENDOR_ID_AMD, 0x740B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AMD756 }, {PCI_VENDOR_ID_AMD, 0x7413, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AMD766 }, {PCI_VENDOR_ID_AMD, 0x7443, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AMD768 }, + {PCI_VENDOR_ID_AMD, 0x746b, PCI_ANY_ID, PCI_ANY_ID, 0, 0, AMD768 }, {PCI_VENDOR_ID_NVIDIA, 0x01B4, PCI_ANY_ID, PCI_ANY_ID, 0, 0, NFORCE }, { 0, } }; @@ -403,6 +404,7 @@ static int __init i2c_amd756_init(void) { printk(KERN_INFO "i2c-amd756.o version %s (%s)\n", LM_VERSION, LM_DATE); + return pci_module_init(&amd756_driver); }

1 0

[suse-amd64] DMA mode booting problems with Suse 9.0 64 bit
by Lothar Hansche 10 Dec '03

10 Dec '03

On Tue, 9 Dec 2003, Jeremy Sonander wrote: > Hi Lothar, > > This sounds interesting. Can you tell me how to control when in the boot > process DMA is turned on? On running multiuser mode you can switch on/off the DMA via hdparm every time. /sbin/hdparm -d /dev/hde # get the actual mode. /sbin/hdparm -d 1 -X udma2 /dev/hde # set to UDMA mode 2 (ON) hdparm -d 0 /dev/hdem # unset DMA mode (OFF) This to have in the boot process is comfort. There is the booting script /etc/init.d/boot.idedma what has an effect if the variable in /etc/sysconfig/hardware is set to a value. If this one is empty, then the script is exiting without an action for switching on/off DMA mode. SuSE default is empty value I did see. Please look in the file hardware: # # DEVICES_FORCE_IDE_DMA="/dev/hdc:off /dev/hdd:off /dev/hde:udma2" DEVICES_FORCE_IDE_DMA="" # I created a new variable for idedma boot script DEVICES_FORCE_IDE_DMA_NEW="/dev/hde:udma2" The value can you change in YAST. I take the ascii variant from yast only for the because in GUI the computer hangs an destroy the Reisefs on SATA-Harddisk. SATA = serial ATA. And if the boot procedure switch on DMA before hwscan then my maschine hangs and I get many harddisk errors. And on the other hand my laptop ASUS L3800C works via boot.idedma before hwscan. I have no idea what is wrong by SuSE-Software or Linux-Kernel 2.2.21-149 default or Shuttle-BIOS or SATA-HD! !!!!!!!!!!!!!!!!!!!!WARNING!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! !!!! Here is a warning by SuSE company also. And my warnings are: Please do experiments only, if you had saved your data before and when you can handle the CD-Rom rescue system to switch off the DMA in the file /etc/sysconfig/hardware, if the computer hangs. By using Reiserfs I have to install SuSE new always. If there is one root partition only (SuSE-default) then you lost all user data also!!! I use from now ext2 and ext3 file system. There I don't lose any data. !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Under root you can do experiments via "hwparm" to see wether disk can DMA generally. See man page for hwparm and aministration handbook for init scripts and runlevels. By me runs principally the boot.idedma after hwscan. I did change the boot procedure manually by editing the dependend scripts how hwscan, xdm and new boot.idedma, by me named as idedma. Here I did insert the new sysconfig variable. After edit the scripts I did execute the command "insserv boot_script" for the changed boot scripts to make the right links automatically for you. Have a good success! Gruss Lothar Hansche, Germany, Berlin

1 0

Adaptec RAID dpt_i2o driver now available
by Derek Liu 10 Dec '03

10 Dec '03

Thanks to a kind individual I was supplied with this pre-released driver which should compile on 64bit systems. It's sitting on my server at: http://www.syste.ms/test/dpt_i2o-2.5.0-2301.tgz Work had been stacking up and I don't know how soon I can try it out. Perhaps someone who has the kernel source handy can verify on the driver? Good luck. :) __________________________________ Do you Yahoo!? Protect your identity with Yahoo! Mail AddressGuard http://antispam.yahoo.com/whatsnewfree

1 0

Installation problems with Suse 9.0 64 bit
by Jeremy Sonander 09 Dec '03

09 Dec '03

Hi Guys, I have a Biostar ideq 200p computer containing an Athlon 64 3200+ CPU, Nvidia Nforce3 chipset and a Nvidia Geforce 4 440 mx graphics card. Booting from the DVD and selecting "Installation" causes the machine to hang after the message "starting hardware detection" Picking "Safe Settings" has the same result Using text mode instead causes it to lock up at "searching for Adding a floppy disk gets past this problem, and I can choose language etc. 5 OKs later the machine again locks up, this time saying starting yast... I have tried turning off all the on-board peripherals appart from the usb controller that I need for my mouse. It hasn't made any difference. I also tried the switch hdc=cdrom suggested in the manual, since my CDROM is a master on the second ide interface, and there is only 1 hard disk on the primary ide channel. However that made no noticeable difference. Any suggestions would be appreciated. Jeremy Sonander. Saros Technology Limited. The Spirella Building, Bridge Road, Letchworth, SG6 4ET. Tel 01462 476111, Fax 01462 476112. For Support email support(a)saros.co.uk

3 4

Ace's Server Guide: Dual Xeon, Dual Opteron and Quad Opteron
by Lenz Grimmer 09 Dec '03

09 Dec '03

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 An interesting read: http://www.aceshardware.com/read.jsp?id=60000275 "In our last article about x86 server CPUs, the Opteron 244 made a quite an impression when compared to the Xeon DP. In many of our real-world server benchmarks, the 1.8 GHz Opteron bested the 2.8 GHz Xeon DP. Of course, neither Intel nor AMD have been sitting still since that time, and to make the Xeon DP more competitive, Intel equipped its midrange server CPU with a 1 MB L3-cache. Correspondingly, 2.2 GHz Opteron 248 and 848 processors are available this month. This article serves as an update on how these server platforms compare, but it's also much more. We've benchmarked AMD's Quartet, a four-headed Opteron hydra with up to 8.8 GHz (4 x 2.2 GHz) of processing power. Unfortunately, the quad Xeon MP was not yet able to join the party, but rest assured that this is not our last server CPU article." - -- - ------------------------------------------------------------------ Lenz Grimmer <lenz(a)grimmer.com> -o) [ICQ: 160767607 | Jabber: LenZGr(a)jabber.org] /\\ http://www.lenzg.org/ V_V -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.2 (GNU/Linux) Comment: For info see http://quantumlab.net/pine_privacy_guard/ iD8DBQE/1Rp9SVDhKrJykfIRApm6AJ0bu/XlnDeXHexH3DD6IEkz/zOzpgCfZrbi 00A6fxRrQL8adGTcE6xX/iM= =57k3 -----END PGP SIGNATURE-----

1 0

Re: [suse-amd64] SuSE 9.0 AMD64 - Hardware detection issues with Shuttle SN85G4?
by dreadnought 07 Dec '03

07 Dec '03

Hi Vojtech, > ------------Original Message------------ > From: Vojtech Pavlik <vojtech(a)suse.cz> > To: Andreas Jaeger <aj(a)suse.de> > Cc: dreadnought <dreadnought(a)arsenal.net>, suse-amd64(a)suse.com > Date: Sun, Dec-7-2003 11:48 AM > Subject: Re: [suse-amd64] SuSE 9.0 AMD64 - Hardware detection issues with Shuttle SN85G4? > > On Sun, Dec 07, 2003 at 08:20:28PM +0100, Andreas Jaeger wrote: > > "dreadnought" <dreadnought(a)arsenal.net> writes: > > > > > While working through the SuSE 9.0 AMD64 installation on my Shuttle SN85G4, I noticed the following oddities: > > > > > > No network card was detected > > > > The nvidia network card is binary only, you need to download it from > > nvidia... > > Actually the nVidia Ethernet has been reverse engineered and now is > available under the name "forcedeth". Unfortunately this was after 9.0 > was released, so we could not include it. Interesting! I installed (or at least think I installed) the nforce driver rpm for SuSE 9.0 off nvidia's website. As far as I can tell, nothing happened. I didn't get any errors, but didn't get any success messages. Maybe this is because the driver is 32-bit only? Any idea if I did anything bad by installing this rpm? Can I simply remove it and any damage I may have done with the package manager? > I don't expect the nVidia binary only ethernet drivers to work on a > 64-bit system, since last time I checked, they were 32-bit only. > > So, if you need to get the ethernet to work, go on the 'net (possibly > with another network card, they're damn cheap anyway), search for > 'forcedeth' patch for a 2.4 kernel, apply it to the SuSE kernel sources, > compile, install ... Excellent .. I will try this! I've never recompiled a kernel before. Are there any SuSE specific docs for this process I could check out? > Regarding the 3D, have you tried just ignoring the "unavailable" message > and enabling it in SaX2 (YaST2's X setup tool)? It worked for me at > least in one case. No, haven't tried this. I'm very new to SuSE. I will try it it! Thanks for your help, Mark

1 0

Fw: Re: [suse-amd64] SuSE 9.0 AMD64 - Hardware detection issues with Shuttle SN85G4?
by dreadnought 07 Dec '03

07 Dec '03

(Re-posted to group.) Hi Andreas, > > ------------Original Message------------ > > From: Andreas Jaeger <aj(a)suse.de> > > To: "dreadnought" <dreadnought(a)arsenal.net> > > Cc: suse-amd64(a)suse.com > > Date: Sun, Dec-7-2003 11:22 AM > > Subject: Re: [suse-amd64] SuSE 9.0 AMD64 - Hardware detection issues with Shuttle SN85G4? > > > > "dreadnought" <dreadnought(a)arsenal.net> writes: > > > > > While working through the SuSE 9.0 AMD64 installation on my Shuttle SN85G4, I noticed the following oddities: > > > > > > No network card was detected > > > > The nvidia network card is binary only, you need to download it from > > nvidia... I got the nforce rpm for SuSE 9.0 from the nvidia website and ran it, but I can't tell if anything actually happened. There was no success message and I don't notice anything different. Kind of frustrating .. I know sometimes the philosophy in Linux is that if you don't get an error message that everything must have gone ok, but I find in practice it's more like a gamble. I rebooted the machine and still didn't notice anything different, no network card detected. Mark

1 0