December 2003 - openSUSE AMD64 - openSUSE Mailing Lists

Tyan Tomcat S2850
by Thomas Hofer 25 Dec '03

25 Dec '03

I plan to set up a (rather low cost) fileserver that is able saturate a gigabit ethernet link. I have in mind a single Opteron 240 on a Tyan Tomcat S2850 (G2NR) with two disks on the (P)ATA channels. I hope someone likes to share his experience before I order it. I'm not sure if the Tomcat is perfectly suited for my demands, because the two integrated gigabit ethernet NICs are 32bit/33MHz PCI devices and the board has no PCI-X slot. The bandwidth of GbE and PCI are roughly the same. So if I use only one NIC and have no other device on the bus, then that would be sufficient (I hope). So the main question I'm asking is whether the two (P-)ATA channels are accessed via the PCI bus. If this were the case, then the NIC had to share half of the bus with the disks. Looking at the block diagram of the board, it seems that the ATA channels are connected directly to the IO-Hub and not using the PCI-bus - is this true? (A lspci -vv of a Tomcat S2850 would be nice!) Are there any other potential bottlenecks? What file transfer performance can I expect from this board? I'm aware that it isn't that easy to gain maximum throughput with GbE - I was hoping to get top file transfers speeds in the range of 50-80 MB/s, but I'm not sure if this is realistic. How much power does the Tomcat need (equipped with a 240 or 140)? It is important for me that the whole system takes less than 100 Watt idle. Are there compatibility problems or instabilities with Linux? I think that SATA isn't supported yet, but this would be no problem - I could wait for that a year or two. The most important things are the NICs and the PATA channels. Thanks, Thomas.

1 0

New siimage patches in lates 2.6.0-mm1 patch
by Steven A. DuChene 23 Dec '03

23 Dec '03

I was just looking at the LinuxToday headlines and I see Andrew Morton has posted a mm1 patch to 2.6.0 In that patch are three new patches listed which seem to relate to the Silicon Omage SATA controller. ide-siimage-seagate.patch ide-siimage-stack-fix.patch ide-siimage-sil3114.patch My problem is I still cannot get a 2.6.X kernel to work with any install media. Still looking for suggestions on my initrd problems with 2.6.X I have not tried this latest mm1 patch yet so hopefully I will get a chance in the next few days. -- Steven A. DuChene

1 0

mplayer does not work on 9.0pro
by Yoriaki FUJIMORI 23 Dec '03

23 Dec '03

Dear list, By chance I tried to compile MPlayer-1.0pre3 on an AMD64 based box running SuSE 9.0 (Pro, commercial one). I noticed a funny behaviour of mplayer on the desktop, that three or four small windows are opened at the top of the desktop without colours; sound seems to be fair. I picked up MPlayer-0.9 or something like that from SuSE8.2beta9 together with libdvdread etc. But, the symthom is the same. Is this a known problem? Where should I look for causes of this problem? Thanks for your attention. yoriaki fujimori

1 1

Booting from SATA Drives - Forced Module Load on boot
by Joe Busch 23 Dec '03

23 Dec '03

Hallo, I have an Shuttle SN85G4 AMD64 with SIL3512 Sata Raid ship. One Drive is connectet (No Raid). On Sil Support Site (ww.cmd.com) i found Linux Driver for Suse 8.2 (32 bit) (64 Bit only aviable as beta for 3114 Chip). I burned this on CD and installed Suse 8.2 by pressing F3 (insert Driverdisk/floppy)on boot. Copyed sii6512.o to modules directory and added to Grub goes well. But first reboot says me, that the module is compiled for other kernel. The difficult is, the module is compiled for 2.2.4.18-GB and the installed kernel is 2.2.4.18-4GB-athlon. I have found the -f option for insmod, but how can i give this option on bootime or in Grub, so it loads the module on boot (if the option works). Other Question is, anyone nows about a 64bit for sil3512? joe -- +++ GMX - die erste Adresse für Mail, Message, More +++ Neu: Preissenkung für MMS und FreeMMS! http://www.gmx.net -- +++ GMX - die erste Adresse für Mail, Message, More +++ Neu: Preissenkung für MMS und FreeMMS! http://www.gmx.net -- +++ GMX - die erste Adresse für Mail, Message, More +++ Neu: Preissenkung für MMS und FreeMMS! http://www.gmx.net

1 0

RE: [suse-amd64] Tyan 2885 and 2.6.0-test11-mm1 "no init found" on boot.iso
by Alan Gray 22 Dec '03

22 Dec '03

Looks like exactly the same config as I had. The wierd thing was that 8.2 Beta for Suse had support from Silicon Image for the Sata drives. I did experience wierd pauses when writing large files to disk. Most of these disappeared when I ran Raid (striped). I emailed Silicon Image and they pretty well ignored any request for a ETA on drivers. I'd say (IMHO) that whey are waiting for the 2.6 kernel to have support. I also had some problems under Windows XP 64 bit for AMD, where I'd get pauses on the system for up to 1 minute and then the computer continued as if nothing happened. I have to say the 3/6 months is my opionion/guestimate. I could be proven wrong (happens quite offen ;-). But I've burned 3 weeks on trying to get the sata to work with help from lists.suse.com but I've real work to do so I'm going back to tried/trusted/expensive SCSI (all my previous computers were SCSI) Alan > -----Original Message----- > From: Steven A. DuChene [mailto:linux-clusters@mindspring.com] > Sent: Tuesday, 23 December 2003 8:27 AM > To: Alan Gray > Cc: suse-amd64(a)suse.com > Subject: Re: [suse-amd64] Tyan 2885 and 2.6.0-test11-mm1 "no > init found" > on boot.iso > > > On Tue, Dec 23, 2003 at 07:15:09AM +1100, Alan Gray wrote: > > All I can say at the moment is I feel for you. > > I gave up yesterday and purchased 2 SCSI disks. (you can > see all of my horrors in previous messages) > > > > It is probably 3/6 months too early for Sata(n) and linux. > > One of the posters, Frank did get his Sata disks running, > but I have Seagate drives > > and he did not. There where a number of patches in Bart's > siimage.c for seagates. > > > > Not sure this helps. > > Hmm, that does not speak well of the technology that it > hasn't settled out enough yet > that there would still be implementation differences in > different manufacturer's > drives. This system I am working with has two Seagate > ST380013AS drives. > > > > > > -----Original Message----- > > > From: Steven A. DuChene [mailto:linux-clusters@mindspring.com] > > > Sent: Tuesday, 23 December 2003 6:24 AM > > > To: suse-amd64(a)suse.com > > > Subject: [suse-amd64] Tyan 2885 and 2.6.0-test11-mm1 "no > init found" on boot.iso > > > > > > I have quite a bit of Linux experience (since 1993) and I > am running other 2.6.0-test > > > kernel on my other x386 system but this is the first > Opteron system I have goofed with. > > > > > BTW, what I meant to say here is that I am running various > 2.6.0-test kernels on many > different X86 Intel/AMD systems with no problems. > -- > Steven A. DuChene linux-clusters(a)mindspring.com > sduchene(a)mindspring.com >

2 1

RE: [suse-amd64] Tyan 2885 and 2.6.0-test11-mm1 "no init found" on boot.iso
by Alan Gray 22 Dec '03

22 Dec '03

All I can say at the moment is I feel for you. I gave up yesterday and purchased 2 SCSI disks. (you can see all of my horrors in previous messages) It is probably 3/6 months too early for Sata(n) and linux. One of the posters, Frank did get his Sata disks running, but I have Seagate drives and he did not. There where a number of patches in Bart's siimage.c for seagates. Not sure this helps. > -----Original Message----- > From: Steven A. DuChene [mailto:linux-clusters@mindspring.com] > Sent: Tuesday, 23 December 2003 6:24 AM > To: suse-amd64(a)suse.com > Subject: [suse-amd64] Tyan 2885 and 2.6.0-test11-mm1 "no init > found" on > boot.iso > > > I am attempting to modify the SuSE9-AMD64 boot.iso so I can > attempt to install SuSE9-X86_64 > to a Tyan 2885 system with two SATA drives hung off of the > Silicon Image 3114 controller. > The system has 2Gb of memory (two 512Kb sticks stuck in each > processor memory bank). > I have the controller configured in IDE mode in the BIOS and > I was able to build a 2.6 > kernel (2.6.0-test11-mm1) that has the siimage driver in it > that seems to recognize the > drives during boot up. The SATA drives show up as hde and > hdg. The CD/DVD drive hung off > of the built in AMD8111 IDE controller shows up as hda. I > have been unable to get this > kernel placed on an altered SuSE9 boot.iso install image > without causing various boot > problems. > > I have built the kernel with all of the appropriate drivers > built into the kernel image > so I shouldn't require any modules. I built in the following drivers: > > Processor family (Opteron/Athlon64/Hammer/K8) > Symmetric multi-processing support > Allocate 3rd-level pagetables from highmem > MTRR (Memory Type Range Register) support > PCI support > PCI access mode (Any) > Vector-based interrupt indexing > Legacy /proc/pci interface > PCI device name database > ISA support > Plug and Play support > Plug and Play BIOS support > Plug and Play BIOS /proc interface > Normal floppy disk support > Loopback device support > Broadcom Tigon3 support <<<<<<<<<< tg3 > Silicon Image chipset support <<<<<<<<<< siimage > Initial RAM disk (initrd) support > Support for Large Block Devices > ELF > a.out > MISC (executable format) > Enable loadable module support > Automatic kernel module loading > RAM disk support > (65536) Default RAM disk size > ATA/ATAPI/MFM/RLL support > AMD and nVidia IDE support (CD/DVD drive conntected to the > onboard IDE controller) > IDE/ATAPI CDROM support > IDE/ATA-2 DISK support > Enhanced IDE/MFM/RLL disk/cdrom/tape/floppy support > IDE Taskfile IO (EXPERIMENTAL) > PCI IDE chipset support > Sharing PCI IDE interrupts support > Generic PCI IDE Chipset Support > Generic PCI bus-master DMA support > Use PCI DMA by default when available > Second extended fs support > Ext3 journalling file system support > ROM file system support > ISO 9660 CDROM file system support > Microsoft Joliet CDROM extensions > UDF file system support > DOS FAT fs support > MSDOS fs support > VFAT (Windows-95) fs support > /proc file system support > /dev/pts file system for Unix98 PTYs > Virtual memory file system support (former shm fs) > Compressed ROM file system support > NFS file system support > SCSI device support > legacy /proc/scsi/ support > SCSI disk support > > > Now I did build in other things like input device and USB > support and etc. but the above > is a pretty complete list (I can supply my kernel config file > if necessary). I did not > include any SCSI controller drivers or SCSI emulation support > in the IDE section. > I also did not include Network block device support or Minix > fs support. Since I did not > include Minix fs support I altered the initrd file to be > gzipped ext2 instead of Minix. > (created a ext2 fs on a file and then tar'ed all the stuff > over from the uncompressed and > loopback mounted Minix initrd file). Since I have what I hope > to be a 2.6 kernel that > should not require any modules I removed all of the > SuSE9-X86_64 2.4 kernel module files > from the initrd file that was on the original boot.iso > > I have quite a bit of Linux experience (since 1993) and I am > running other 2.6.0-test > kernel on my other x386 system but this is the first Opteron > system I have goofed with. > > During kernel bootup it does say > > RAMDISK: Compressed image found at block 0 > VFS: Mounted root (ext2 filesystem) read only > Freeing unused kernel memory: 168k freed > Kernel panic: No init found. Try passing init= option to the kerme. > > The mkisofs command I am using to build my test boot.iso images is: > > mkisofs -b boot/loader/isolinux.bin -c > boot/loader/isolinux.cat -no-emul-boot\ > -boot-load-size 4 -boot-info-table -l -R -r -o > boot_k26sminitrd.iso newboot/ > > I tried passing init=/linuxrc or init=linuxrc but I still get > the "No init found" > I also tried a 2.4 kernel built which a friend in CZ built > with a sata_sil > driver in it but that did not find the 3114 device. I did not > have a "No init found" > problem with that kernel however. > > Can anyone see from my above list of drivers I compiled into > the kernel anything > obvious I would be missing? Or any other problems that might > cause this? > > BTW, I originally was working with the default 4Mb > ramdisk_size as compiled into the > 2.6 kernel but then I thought I might try bumping that up to > 64Mb since that is > what is needed for the rescue image (I have been trying to > boot to either that or > the standard install choice). Still same thing. > > I tried also booting this kernel and initrd via PXE across > the network but I had no > luck with that either. ARGH!!! > -- > Steven A. DuChene linux-clusters(a)mindspring.com > sduchene(a)mindspring.com > > -- > Check the List-Unsubscribe header to unsubscribe > For additional commands, email: suse-amd64-help(a)suse.com > >

2 1

Tyan 2885 and 2.6.0-test11-mm1 "no init found" on boot.iso
by Steven A. DuChene 22 Dec '03

22 Dec '03

I am attempting to modify the SuSE9-AMD64 boot.iso so I can attempt to install SuSE9-X86_64 to a Tyan 2885 system with two SATA drives hung off of the Silicon Image 3114 controller. The system has 2Gb of memory (two 512Kb sticks stuck in each processor memory bank). I have the controller configured in IDE mode in the BIOS and I was able to build a 2.6 kernel (2.6.0-test11-mm1) that has the siimage driver in it that seems to recognize the drives during boot up. The SATA drives show up as hde and hdg. The CD/DVD drive hung off of the built in AMD8111 IDE controller shows up as hda. I have been unable to get this kernel placed on an altered SuSE9 boot.iso install image without causing various boot problems. I have built the kernel with all of the appropriate drivers built into the kernel image so I shouldn't require any modules. I built in the following drivers: Processor family (Opteron/Athlon64/Hammer/K8) Symmetric multi-processing support Allocate 3rd-level pagetables from highmem MTRR (Memory Type Range Register) support PCI support PCI access mode (Any) Vector-based interrupt indexing Legacy /proc/pci interface PCI device name database ISA support Plug and Play support Plug and Play BIOS support Plug and Play BIOS /proc interface Normal floppy disk support Loopback device support Broadcom Tigon3 support <<<<<<<<<< tg3 Silicon Image chipset support <<<<<<<<<< siimage Initial RAM disk (initrd) support Support for Large Block Devices ELF a.out MISC (executable format) Enable loadable module support Automatic kernel module loading RAM disk support (65536) Default RAM disk size ATA/ATAPI/MFM/RLL support AMD and nVidia IDE support (CD/DVD drive conntected to the onboard IDE controller) IDE/ATAPI CDROM support IDE/ATA-2 DISK support Enhanced IDE/MFM/RLL disk/cdrom/tape/floppy support IDE Taskfile IO (EXPERIMENTAL) PCI IDE chipset support Sharing PCI IDE interrupts support Generic PCI IDE Chipset Support Generic PCI bus-master DMA support Use PCI DMA by default when available Second extended fs support Ext3 journalling file system support ROM file system support ISO 9660 CDROM file system support Microsoft Joliet CDROM extensions UDF file system support DOS FAT fs support MSDOS fs support VFAT (Windows-95) fs support /proc file system support /dev/pts file system for Unix98 PTYs Virtual memory file system support (former shm fs) Compressed ROM file system support NFS file system support SCSI device support legacy /proc/scsi/ support SCSI disk support Now I did build in other things like input device and USB support and etc. but the above is a pretty complete list (I can supply my kernel config file if necessary). I did not include any SCSI controller drivers or SCSI emulation support in the IDE section. I also did not include Network block device support or Minix fs support. Since I did not include Minix fs support I altered the initrd file to be gzipped ext2 instead of Minix. (created a ext2 fs on a file and then tar'ed all the stuff over from the uncompressed and loopback mounted Minix initrd file). Since I have what I hope to be a 2.6 kernel that should not require any modules I removed all of the SuSE9-X86_64 2.4 kernel module files from the initrd file that was on the original boot.iso I have quite a bit of Linux experience (since 1993) and I am running other 2.6.0-test kernel on my other x386 system but this is the first Opteron system I have goofed with. During kernel bootup it does say RAMDISK: Compressed image found at block 0 VFS: Mounted root (ext2 filesystem) read only Freeing unused kernel memory: 168k freed Kernel panic: No init found. Try passing init= option to the kerme. The mkisofs command I am using to build my test boot.iso images is: mkisofs -b boot/loader/isolinux.bin -c boot/loader/isolinux.cat -no-emul-boot\ -boot-load-size 4 -boot-info-table -l -R -r -o boot_k26sminitrd.iso newboot/ I tried passing init=/linuxrc or init=linuxrc but I still get the "No init found" I also tried a 2.4 kernel built which a friend in CZ built with a sata_sil driver in it but that did not find the 3114 device. I did not have a "No init found" problem with that kernel however. Can anyone see from my above list of drivers I compiled into the kernel anything obvious I would be missing? Or any other problems that might cause this? BTW, I originally was working with the default 4Mb ramdisk_size as compiled into the 2.6 kernel but then I thought I might try bumping that up to 64Mb since that is what is needed for the rescue image (I have been trying to boot to either that or the standard install choice). Still same thing. I tried also booting this kernel and initrd via PXE across the network but I had no luck with that either. ARGH!!! -- Steven A. DuChene linux-clusters(a)mindspring.com sduchene(a)mindspring.com

1 0

RE: [suse-amd64] 8.2 Beta
by Alan Gray 21 Dec '03

21 Dec '03

Thanks for that. Now that I have found a spare Adaptec 29160, I'm totally ditching using Sata(n) and will get a couple of 15K SCSI drives. Regards Alan > -----Original Message----- > From: Andi Kleen [mailto:ak@suse.de] > Sent: Monday, 22 December 2003 6:38 AM > To: Alan Gray > Cc: suse-amd64(a)suse.com > Subject: Re: [suse-amd64] 8.2 Beta > > > On Sun, 21 Dec 2003 17:53:35 +1100 > "Alan Gray" <alan.gray(a)12d.com> wrote: > > > Was the 8.2 Beta a full 64 bit OS? I've just had a thought > that if it > > was, I probably can do 80% of my work on 8.2, so long as g++ gcc > > produce 64 bit code. > > Yes, it was. > > -Andi > >

1 0

RE: [suse-amd64] Adaptec 29320 & 39320 Support
by Alan Gray 21 Dec '03

21 Dec '03

Yep, you guys are 100% correct. I took a 29160 out of a server and 9.0 booted ok. Do you think they will notice the tape drive is now a book end ? ;-) I'll hook the tape up on the spare channel of another SCSI card in the server and hope that it will not interfere with the other devices. This Sony AIT SDX-400C does not like being on the same SCSI chain as any other device. When the SCSI bios loaded, either the AIT was in visible or all other devices where. I'm hoping that being on Channel B, this problem will not arise. (All this occurred on Windows 2000, and I'm not trying to get it running on Suse so you can rest easy) Regards Alan > -----Original Message----- > From: Vojtech Pavlik [mailto:vojtech@suse.cz] > Sent: Wednesday, 17 December 2003 8:44 AM > To: Andreas Jaeger > Cc: Alan Gray; suse-amd64(a)suse.com > Subject: Re: [suse-amd64] Adaptec 29320 & 39320 Support > > > On Tue, Dec 16, 2003 at 10:41:10PM +0100, Andreas Jaeger wrote: > > "Alan Gray" <alan.gray(a)12d.com> writes: > > > > > Hi, > > > > > > Danger, danger Will Roberson. Knowing my luck at the moment, I'll > > > buy the one that does not work. > > > > ;-( > > > > > Any of these definately yes/no. > > > > 29160 gets a definite yes, I just checked a system that has it ;-) > > I've used 29160 with many AMD64 systems and different SuSE > Linux versions and it always worked great. > > > Note it's the same source driver for all of these... > > > -- > Vojtech Pavlik > SuSE Labs, SuSE CR >

1 0

8.2 Beta
by Alan Gray 21 Dec '03

21 Dec '03

Andreas, Was the 8.2 Beta a full 64 bit OS? I've just had a thought that if it was, I probably can do 80% of my work on 8.2, so long as g++ gcc produce 64 bit code. I am a little gun shy about getting an Adaptec 29320 or 39320 lest it does not work. If you have any way of finding out, I'd be very appreciative. Regards Alan Have a Merry XMAS

2 1