March 2006 - openSUSE AMD64 - openSUSE Mailing Lists

32-bit code under 64-bit kernel: OpenOffice crashes
by Constantine 'Gus' Fantanas 09 Mar '06

09 Mar '06

Hello all. I have 32-bit Novell OpenOffice installed on by 64-bit SuSE 10. When I open the tools page and click on Java (where I should be able to see the installed Java environments), OpenOffice hangs, 100% of processor power is sapped, but 'kill -TERM' stops OpenOffice. The curious thing is that while Open Office is hung, 'ps auxw' reveals that 64-bit Java is running. Somehow, therefore, OpenOffice tries to invoke the 64-bit Java (Sun, 1.5 version), not the 32-bit Java. My 32-bit Firefox works with the 32-bit Java just fine. So, how can I force OpenOffice to look only at the 32-bit Java, so that it doesn't hang? CF -- Running 64-bit Linux on AMD64

2 1

AW: [suse-amd64] Xine and 64bit windows codecs
by Dittmann, Werner 08 Mar '06

08 Mar '06

Just an idea: put a linux32 in front of the configure call. linux32 switches the system identify to a 32-bit 686 system for the commands. linux32 ./configure This in addition to set the CFLAGS and LDFLAGS to the -m32 architecture. Doing this helped in some of my config problems. Another idea is to recreate the configure file if the files configure.ac (or configure.in) Makefile.am (these also in the subdirectories) are available. Then you are able to run linux32 ./autogen.sh this shell script calls a number of programms to recereate the configure, makefile, etc. Regards, Werner > -----Ursprüngliche Nachricht----- > Von: Tim Janssen [mailto:timtjanssen@yahoo.com] > Gesendet: Samstag, 4. März 2006 19:01 > An: Manuel Soriano; suse-amd64(a)suse.com > Betreff: Re: [suse-amd64] Xine and 64bit windows codecs > > Thanks you all, for the abundant info on this. I am fine with > sticking to 32 bit for now. However, > how do I compile Mplayer in 32bit mode? Manu suggested > > > And change, in the configure, the search order of the libraries : > > > > /usr/lib:/usr/lib64 instead of /usr/lib64:/usr/lib > > but I have been unable to find these lines in the CVS version > configure file. When I ignore that > bit and simply set the flags (as Manu also suggested) > > > export CFLAGS=-m32 > > export LDFLAGS=-m32 > > doing a configure/make results in a mess: > > wrapper.S: Assembler messages: > wrapper.S:25: Error: `pusha' is not supported in 64-bit mode > wrapper.S:28: Error: suffix or operands invalid for `push' > wrapper.S:32: Error: suffix or operands invalid for `push' > wrapper.S:34: Error: suffix or operands invalid for `push' > wrapper.S:39: Error: suffix or operands invalid for `push' > wrapper.S:40: Error: suffix or operands invalid for `push' > wrapper.S:49: Error: `popa' is not supported in 64-bit mode > wrapper.S:51: Error: suffix or operands invalid for `pop' > wrapper.S:52: Error: suffix or operands invalid for `push' > wrapper.S:58: Error: suffix or operands invalid for `push' > wrapper.S:59: Error: `pusha' is not supported in 64-bit mode > wrapper.S:62: Error: suffix or operands invalid for `push' > wrapper.S:66: Error: suffix or operands invalid for `push' > wrapper.S:68: Error: suffix or operands invalid for `push' > wrapper.S:73: Error: suffix or operands invalid for `push' > wrapper.S:74: Error: suffix or operands invalid for `push' > wrapper.S:81: Error: `popa' is not supported in 64-bit mode > make[1]: *** [wrapper.o] Error 1 > > Thus: how to compile mplayer as a 32bit application? > > Thanks! Tim > > > > > > > --- Manuel Soriano <manu(a)valux.org> wrote: > > > El Sábado 04 Marzo 2006 13:50, Tom Corr escribió: > > > On Friday 03 March 2006 09:04, Manuel Soriano wrote: > > > > For a while i don't use other viewer than mplayer : > > > > > > > > http://www.mplayerhq.hu/design7/dload.html > > > > > > > > very easy to configure, compile & using... > > > > mplayer have some problems in 64bit & with the compiler > included in the > > Suse-10.0, you have some solutions : > > > > compile in 32bit mode with the stock compiler > > compile in 32bit mode with the old gcc compiler > > compile in 64bit mode with the stock compiler,but with some > restrictions, like > > the windows codes > > cimpile in 64bit mode with the old gcc compiler, with the > same restrictions. > > > > The old gcc compiler gives you more stability when you copy > "autorised" > > pictures... I use the version 3.3.6 of the gcc. > > > > To compile in 32bit mode, you need to declare two variables : > > > > export CFLAGS=-m32 > > export LDFLAGS=-m32 > > > > And change, in the configure, the search order of the libraries : > > > > /usr/lib:/usr/lib64 instead of /usr/lib64:/usr/lib > > > > Enjoy... > > > > > > > > -- > > > > Bye, Manu. > > > > Break the rules! Break the windows! Use Linux! > > manu at valux.org http://www.valux.org > > Linux is here to stay. FUD won't make it go away. > > ************************************************* > > * Manuel Soriano * Lo Perello/Valencia/Spain * ( AbM Compliant ) > > Linux Counter Nº: 294597 > > Para que funcione la alianza de civilizaciones, todos deben > ser civilizados. > > > > -- > > Check the List-Unsubscribe header to unsubscribe > > For additional commands, email: suse-amd64-help(a)suse.com > > > > > > > -- > Check the List-Unsubscribe header to unsubscribe > For additional commands, email: suse-amd64-help(a)suse.com > >

4 3

Re: [suse-amd64] Hardware Raid in SuSe10
by Jaume Fortuny 08 Mar '06

08 Mar '06

Thanks for the answers and sorry for not beeing very specific. It's a real hardware raid. The machine is a Supermicro AS-1020A-8B 2 with Bios H8DA8/H8DAR and 2 AMD Opteron Processors and 2Gb of RAM. It has an SCSI controller Adaptec AIC-7902 with two channels and 3 SCSI hard disks Fujitsu MAT3147NC of 147Gb. Through Adaptec HostRAID Bios v4.30 I've created a RAID 1 (Mirror) with 2 of the 3 disks, with the idea of having the operating system in the first drive and using the mirror for data. At this moment, with the provided bios SCSI utility there exist the following configuration: Hard Drive AIC-7902A 0 Fujitsu 147Gb Mirrored(R1) AIC-7902A 1 Adaptec 147Gb The problem that I have is that SuSe 10 can not see the RAID as one single drive. From the operating system I see 2 drives. Can anyone direct me towards a solutions? Thanks again, Jaume Fortuny En/na Rainer Koenig ha escrit: >Jaume Fortuny <j.fortuny(a)inform.es> writes: > > > >>Problem in SuSe 10: >>A RAID 1 array created and recognized by BIOS is not recognized by >>SuSe10 installation. It appears as 2 separate disks. >> >>Question: >>Is a driver needed? >>Or is SuSe 10 not able to recognize a Hardware array? >> >> > >Can you be a bit more specific about your "hardware RAID"? Is this a >real hardware RAID controller or just a thing that pretends to be one? >Usually the mainboard "hardware RAID" controllers are just software >RAID and need a driver that can read the settings that the BIOS stored >somewehere in the NVRAM or on the disks. > >The problem with those solutions is that drivers for Linux are very >rarely available and they are often just binaries that require a >specific kernel version. That means, if you need to upgrade your kernel >you will run into problems if you have no binaries for the new kernel >available. > >Linux itself has a very good software RAID driver in the kernel that you >can use and configure with the tools that your Linux distribution brings >with it. This software RAID is even capable of doing a RAID 5 with hot >spare disks, so you will have a real redundant system that allows you to >do a hotplug of the disks and rebuild the array while running the system. >The BIOS based "hardware RAID" solutions often just detect that you changed >a disk and then do a copy of the remaining disk to the new one, all in your >BIOS, you can imagine how long it takes to rebuild a mirror with 80GB. :-) > >My personal experience is that Linux software RAID is always the best >solution unless you have a real hardware RAID, that means a controller >that has the "intelligence" on its silicon and that is just showing one >logical block device to the system. But I'm afraid that your controller >doesn't have that. > >So you can chose if your CPU spends the cycles in a proprietary software >RAID driver (if available) or in an open source kernel module that is >always available. > >HTH >Rainer > > -- Check the List-Unsubscribe header to unsubscribe For additional commands, email: suse-amd64-help(a)suse.com

2 1

Hardware Raid in SuSe10
by Jaume Fortuny 07 Mar '06

07 Mar '06

Hello *, Problem in SuSe 10: A RAID 1 array created and recognized by BIOS is not recognized by SuSe10 installation. It appears as 2 separate disks. Question: Is a driver needed? Or is SuSe 10 not able to recognize a Hardware array? Thanks, Jaume -- Check the List-Unsubscribe header to unsubscribe For additional commands, email: suse-amd64-help(a)suse.com

4 3

Suse 9.3 Eval And Package Gcc-g77 V3.3.5-5.1 : where to find this package ?
by Pascal THEVENET 07 Mar '06

07 Mar '06

Hello It appears that there is a very small difference between SuSE 9.3 Pro and SuSE 9.3 Eval. This is the case for the package gcc : version 3.3.5-5 for 9.3 Pro, version 3.3.5-5.1 for 9.3 Eval. I need the package gcc-g77. But this one coming from 9.3 Pro can't be installed on 9.3 Eval. I can't find the right package on internet ... or I am not searching very well :-). I have tried to recompile the source of gcc 3.3.5-5 coming from 9.3 Pro but it requires gnat, gnat-runtime and ncurses-devel whose version 5.4-68 from 9.3 Pro (not availabe in 9.3 Eval) is not in accordance with ncurses of 9.3 Eval (version 5.4-68.1). And idea to get a package gcc-g77 version 3.3.5-5.1 ? Regards Pascal

8 10

Re: [suse-amd64] compile problems for Nat'l Instruments software - help, anyone?
by Jonathan Puthoff 06 Mar '06

06 Mar '06

Yes, I did those things, but it sounds like the NIKAL kernel module just won't work on my 64-bit system. I was hoping to get around it, but I may be stuck. . . ----- Original Message ----- From: Markus Koßmann <mkossmann_nospam(a)schlundmail.de> Date: Thursday, February 9, 2006 0:14 am Subject: Re: [suse-amd64] compile problems for Nat'l Instruments software - help, anyone? To: suse-amd64(a)suse.com Cc: Jonathan Puthoff <jputhoff(a)wisc.edu> > Am Mittwoch, 8. Februar 2006 17:11 schrieb Jonathan Puthoff: > > Hello all! I am new here (and pretty new to Linux, for that > matter) and > > was referred here for some assistance. > > > > Recently I built an AMD Sempron machine for my lab and installed > SuSE> 10.0 x86_64. I want to install and run National Instruments > software> (LabVIEW, NIDAQmx) to control my DAQ PCI card. There is > a problem, > > however, with the compiling of the NI-KAL kernel module, upon > which the > > rest of the installation seems to depend. The "./configure" > script that > > I need to run prior to my "make" checks for a file called > > /usr/src/linux/include/asm/asm_offsets.h and returns an "kernel not > > configured correctly" error. > asm_offsets.h is a generated file. > Did you do a > > cd /usr/src/linux > make cloneconfig > make prepare-all > > to prepare the kernel sources for external module compilation ? >

3 3

Re: [suse-amd64] upgrade problem
by chu＠tes-mail.jpl.nasa.gov 04 Mar '06

04 Mar '06

I wrote: > ... Some strange problems encoutnered while trying to upgrade my > system with YOU ... Many thanks to everyone who helped my find my problem. I was able to fix it, and retrieve and install the upgrades. During the install process, I got many notices about things I should change, including rebuilding of the latest downloaded kernaded kernel (2.4.21-306). (Why didn't YOU just build and install it automatically like before?) So I followed the instructions on configuring the kernel options that I wanted, but it does not get very far when I tried to build it. I get some errors pretty early on in make dep: make dep ... make[1]: Entering directory `/usr/src/linux-2.4.21-306' make -C arch/x86_64/tools fastdep make[2]: Entering directory `/usr/src/linux-2.4.21-306/arch/x86_64/tools' gcc -D__KERNEL__ -I/usr/src/linux-2.4.21-306/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -mno-red-zone -mcmodel=kernel -msoft-float -pipe -fno-reorder-blocks -finline-limit=2000 -fno-strength-reduce -Wno-sign-compare -fno-asynchronous-unwind-tables -S -o offset.tmp offset.c In file included from /usr/src/linux-2.4.21-306/include/linux/spinlock.h:56, from /usr/src/linux-2.4.21-306/include/linux/wait.h:16, from /usr/src/linux-2.4.21-306/include/linux/fs.h:12, from /usr/src/linux-2.4.21-306/include/linux/capability.h:17, from /usr/src/linux-2.4.21-306/include/linux/binfmts.h:5, from /usr/src/linux-2.4.21-306/include/linux/sched.h:11, from offset.c:9: /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: parse error before '{' token /usr/src/linux-2.4.21-306/include/asm/spinlock.h: In function `__foo': /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: `__file' undeclared (first use in this function) /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: (Each undeclared identifier is reported only once /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: for each function it appears in.) /usr/src/linux-2.4.21-306/include/asm/spinlock.h: At top level: /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: warning: type defaults to `int' in declaration of `__foo' /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: `__foo' redeclared as different kind of symbol /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: previous declaration of `__foo' /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: warning: data definition has no type or storage class /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: parse error before '}' token /usr/src/linux-2.4.21-306/include/asm/spinlock.h:10: warning: `__foo' defined but not used make[2]: *** [offset.h] Error 1 make[2]: Leaving directory `/usr/src/linux-2.4.21-306/arch/x86_64/tools' make[1]: *** [_sfdep_arch/x86_64/tools] Error 2 make[1]: Leaving directory `/usr/src/linux-2.4.21-306' make: *** [dep-files] Error 2 What could be wrong now? I see another directory has been deposited in /usr/src by YOU: linux-2.4.21-306-include Does this have to some how be integrated with the linux-2.4.21-306 directory tree? There were no notices to that effect that I could see. thanks in advance for any information. eyc

1 0

tcp checksum errors
by Matt Hurd 02 Mar '06

02 Mar '06

Dear List, We've had a problem with a tcp app communicating with an external system. On diagnosis tcpdump is showing a chksum error on the packets that go out that don't get received. This is Suse 10.0 - 2.6.13-15.8-smp x86_64 The socket layer in the app, ::send and ::write, report no errors. We've confirmed such errors as reported by tcpdump are occurring reasonably frequently without too much consequence on intel e1000, intel e100, broadcom, nvidia ck804. (tyan, shuttle and iwill mainboards). Problem has been hidden we think as the apps seem to recover or not loose the pkts. The only discernable difference is there is a firewall on the route between the systems where the packets don't get through, but we're grasping at straws. We can duplicate the error conditions with scp and telnet (if the chksum error is real). We are seeing no problems with a couple of systems on their pci card based realtek 8169 cards. Same systems have issues with their mainboard adaptors. Smells like a problem. It could be tcpdump / libpcap misreporting but the lost packets indicate a problem. We don't lose packets that have a good chksum but we do occaisonally see packets survive that are marked as bad chksum... On similar hardware CentOS 4.2 with 2.6.9-22.ELsmp doesn't report similar errors. Just installed the same CentOS on a system reporting the errors and it no longer has such errors. We saw some notes about issues with APIC with various adaptors, tried booting with no apic and same errors. Any thoughts? Matt. matthurd(a)acm.org

2 2

ASUS K8V-SE motherborad / Promise SATA-PATA / SuSE 10.0
by SuSE eMail Collector 01 Mar '06

01 Mar '06

SuSE 10.0 does not recognize PATA drives attached to a PROMISE PDC20378 SATA/PATA controller. No SATA drives are installed in the system. SuSE 9.3 has no difficulty with these drives in IDE mode. How do I convince the sata_promise driver on SuSE 10.0 to recognize my drives ???

1 0

Segmentation faults on Opteron System
by Christian Hopfensitz 01 Mar '06

01 Mar '06

I have a strange problem with my smp-system. Hardware config: Tyan s2895 mobo, 2x Opteron 270 Dual-Core CPU, 16GB ECC Reg. RAM We use this Workstation for CFD-Analyses. When I start 4 Jobs on 4 CPUs, at least 2 jobs die (segmentation fault). One job works fine. A parallel run using mpich as MPI aborts (Error message says that it has recieved a SIGABORT signal form a CPU / SIGSEGV). The strage thing: If I remove 8GB of RAM, the system works stable. Neither an update from Suse 9 to Suse 10 nor a manual compiled kernel with the current stable version helped. A memtest86+ showed no errors (running 5 days), so I think that a bad RAM module is not hte reason. Furthermore, a cpu-stress test with cpuburn-in worked well (24 hours running without any errors). Might be a problem with the memory managment? Any hints how I can resolve my problem? Christian

2 1