Linux SMP HOWTO David Mentré, David.Mentre@irisa.fr v1.6, 21 October 1999 This HOWTO reviews main issues (and I hope solutions) related to SMP configuration under Linux. ______________________________________________________________________ Table of Contents 1. Introduction 2. What's new ? 3. Questions related to any architectures 3.1 Kernel Side 3.2 User Side 3.3 SMP Programming 3.3.1 Parallelization methods 3.3.2 The C Library 3.3.3 Languages, Compilers and debuggers 3.3.4 Other libraries 3.3.5 Other points about SMP Programming 4. x86 architecture specific questions 4.1 Why it doesn't work on my machine? 4.2 Possible causes of crash 4.3 Motherboard specific information 4.3.1 Motherboards with known problems 4.4 Low cost SMP Linux box (dual Celeron box) 4.4.1 Is it possible to run a dual Intel Celeron box ? 4.4.2 How does Linux behave on a dual Celeron system ? 4.4.3 Celeron processors are known to be easily overclockable. And dual Celeron system ? 4.4.4 And making a quad Celeron system ? 4.4.5 What about mixing Celeron and Pentium II processor ? 5. Sparc architecture specific questions 5.1 Which Sparc machines are supported ? 5.2 Specific problem related to Sparc SMP support 5.3 SMP specific limit with current kernel (2.2) 6. PowerPC architecture specific questions 6.1 Which PPC machines are supported ? 6.2 Specific problem related to PPC SMP support 7. Alpha architecture specific questions 7.1 Which Alpha machines are supported ? 7.2 Specific problem related to Alpha SMP support 8. Useful pointers 8.1 Various 8.2 Multithreaded programs and library 8.3 SMP specific patches 8.4 Parallelizing/Optimizing Compilers for 586/686 machines ( 9. Glossary 10. List of contributors ______________________________________________________________________ 1. Introduction Linux can work on SMP (Symmetric Multi-Processors) machines. SMP support started with the 2.0 family and has been improved in the 2.2 version. While this version is not fully SMP aware, current 2.3 (next 2.4) fills the remaining gaps. The kernel locking granularity is much finer in 2.2.x than in 2.0.x, which should result in better performance when processes are accessing the kernel! HOWTO maintained by David Mentré (David.Mentre@irisa.fr). The latest edition of this HOWTO can be found at · http://www.irisa.fr/prive/mentre/smp-howto/ (France) · http://www.phy.duke.edu/brahma/smp-faq/ (USA) If you want to contribute to this HOWTO, I would prefer a diff against the SGML version of this document, but any remarks (in plain text) will be greatly appreciated. If you send me an email about this HOWTO, please include a tag like [Linux SMP HOWTO] in the Subject: field of your e- mail. It helps me to automatically sort mails (and you will have a faster reply ;)). This HOWTO is an improvement of a first draft made by Chris Pirih. All information contained in this HOWTO is provided "as is." All warranties, expressed, implied or statutory, concerning the accuracy of the information of the suitability for any particular use are hereby specifically disclaimed. While every effort has been taken to ensure the accuracy of the information contained in this HOWTO, the authors assume no responsibility for errors or omissions, or for damages resulting from the use of the information contained herein. 2. What's new ? v1.6, 21 october 1999 · added an information on xosview scheduling perturbation · added "APIC error interrupt on CPU#n" message information · added information on hard lockup · deleted section "How to optain maximum performance" (was obsolete) · added info on dual systems with different x86 procs (a Celeron and a P-II) v1.5, 4 october 1999 · more precision in PSET description v1.4, 30 september 1999 · precize to enable MTRR support for an x86 SMP kernel (me) v1.3, 29 september 1999 · many many grammar and typographical fixes (Wade Hampton aka hww) · added info in short introduction related to 2.2/2.4/2.0 diffs (hww) · added step by step things to do to recompile a kernel (hww and me) · added info related to SMP/UP modules problems (hww) · added precision in Posix Threads section related to user (hww) vs. kernel threads (hww) · new item about NFS and kernel lock (hww) · new item about kernel lock without message (hww) · new item about debugging lockup problems (hww) · added info about heating problems (hww) · miscellaneous updates I've forget about (hww) · new item about floppy access and sound (hww) v1.2, 27 septembre 1999 · name change: this document is now a HOWTO. TWD, and fast! (Guylhem Aznar) v1.1, 26 septembre 1999 · added a link to first Chris Pirih FAQ draft · expanted an IRQ related problems v1.00, 25 septembre 1999 · first upgrade since a long long time! · reprocessed the whole FAQ: 2.2 is here and 2.4 soon · added kernel locking information from Ingo Molnar · deleted item "How will my application perform under SMP?": outdated · deleted item "My SMP system is locking up all the time.": outdated · deleted item "You are running 2.0.35 aren't you ?": outdated · deleted item "Some hardware is also known to cause problems.": outdated · blanked section "Motherboards with known problems". We should restart from scratch · deleted section "Motherboards with NO known problems": outdated · updated dual celeron section (numerous people) · added "SPARC sun4m SMP machines" to supported SMP sparc machines (Anton Blanchard) · added a "During boot machine hang signaling an IOAPIC problem" item in "Why it doesn't work on my machine?" section · added a "What about SMP performances?" item · updated "Why doesn't my old Compaq work?" item · fixed an outdated pointer · added a pointer to Ingo test SMP patches v0.54, 13 march 1999 · Added a section about SMP Alpha systems v0.53, 08 march 1999 · Added a section about SMP PowerPC systems v0.52, 07 march 1999 · Added a section about SMP Sparc systems v0.51, 06 march 1999 · Added a dual-celeron section · Deleted Adaptec section · Updated procps link · Updated xosview link · Added an answer for quad Xeon boot hang · Updated item about glibc patch for gd: should be included in RH 5.2 v0.50, 03 february 1999 · Updated "Multithreaded programs on linux" link v0.49, 13 january 1999 · Update about CONFIG_SMP. Added .txt to Documentation/smp. (Michael Elizabeth Chastain) v0.48, 10 december 1998 · Mispelled corrected. Email address corrected. v0.47, 20 november 1998 · Added that 2.0.36 as the MTRR patch (related to the BogoMips problem) v0.46, 10 november 1998 · Update about Epox KP6-LS motherboards v0.45, 25 october 1998 · Corrected an error regarding /proc/stat file · Added a pointer to CESDIS Ethernet Linux Drivers site v0.44, 14 october 1998 · Updated the link to the web page: Motherboards rumored to run Linux SMP · Added Jakob explanation how to time SMP systems with 2.0 kernels v0.43, 9 september 1998 · Updated first question in section 3.1 · Updated mt-Mesa link: multi-threaded is now included as experimental in the Mesa distribution v0.42, 2 september 1998 · Minor cosmetic update in sect 3.3 · Two links (multithreaded Mesa and SMP performance) marked outdated · Updated the item about threads and exceptions in C++ (sect 3.3) v0.41, 1 september 1998 · Added a major section: "3.3 SMP Programming" written by Jakob Østergaard · moved some item of section "3.2 User side" in sect 3.3 v0.40, 27 august 1998 · Updated section 3.1, item 7: processor affinity v0.39, 27 august 1998 · Updated needed Award BIOS version for Tyan motherboards (hASCII) · Added an item on IRQ in the crash section (me and hASCII) · Added good support of Asus P2B-DS (Ulf Rompe) · Added another smp-list archive in pointer section (Hank Leininger) v0.38, 8 august 1998 · Added a pointer to the Linux Threads FAQ v0.37, 30 July 1998 · Emil Briggs is working on parallel plugins for Gimp (see "Is there any threaded programs or library?", sect. "User side") v0.36, 26 July 1998 · Thanks to Jakob Østergaard, two changes in "Possible causes of Crash" · Changed 2.0.33 to 2.0.35 (latest stable) · Added a "BIOS related causes of failure" v0.35, 14 July 1998 · Added N440BX Server Board in Motherboards with NO problems · Added a succes story for GigaByte motherboard with BIOS upgrade · Added a "How to obtain maximum performance ?" section (waiting for your contributions ;) v0.34, 10 june 1998 · Added a "Parallelizing/Optimizing Compilers for 586/686 machines" section in section "Useful Pointers", thanks to Sumit Roy · Corrected a mispelling, "Asus P/I-UP5" is in fact "Asus P/I- P65UP5" v0.33, 3 june 1998 · Yet another success story for a GigaByte DLX Motherboard. · A tip for Tyan motherboards, disable the "DRAM Fast Leadoff" BIOS option v0.32, 27 may 1998 · Asus P/I-UP5 added in the motherboard-with-NO-problem section v0.31, 18 may 1998 · Elitegroup P6LX2-A works with 2.1.100 and 101 · Bugs should be reported to linux-smp@vger.rutgers.edu v0.30, 12 may 1998 · SuperMicro is now in the motherboard-with-NO-problem section v0.29, 11 may 1998 · A success story for a GigaByte 686 motherboard with 2.1.101 · Added a new item in the "User Side" section: "Is there any threaded programs or library?" · OpenGL Mesa library is beeing multithreaded. Cool! See the new section for details. v0.28, 09 may 1998 · A US mirror of this FAQ is now available (see Introduction) · Merge of the two confusing Gigabyte 686 entries v0.27, 05 may 1998 · New info for the Adaptec and TekRam drivers · Micronics W6-LI motherboard works under SMP 3. Questions related to any architectures 3.1. Kernel Side 1. Does Linux support multi-threading? If I start two or more processes, will they be distributed among the available CPUs? Yes. Processes and kernel-threads are distributed among processors. User-space threads are not. 2. What kind of architectures are supported in SMP? From Alan Cox: SMP is supported in 2.0 on the hypersparc (SS20, etc.) systems and Intel 486, Pentium or higher machines which are Intel MP1.1/1.4 compliant. Richard Jelinek adds: right now, systems have been tested up to 4 CPUs and the MP standard (and so Linux) theoretically allows up to 16 CPUs. SMP support for UltraSparc, SparcServer, Alpha and PowerPC machines is in available in 2.2.x. From Ralf Bächle: MIPS, m68k and ARM does not support SMP; the latter two probly won't ever. That is, I'm going to hack on MIPS-SMP as soon as I get a SMP box ... 3. How do I make a Linux SMP kernel? In the 2.0 series (and before 2.1.132), uncomment the SMP=1 line in the main Makefile (/usr/src/linux/Makefile). In the 2.2 version, configure the kernel and answer "yes" to the question "Symmetric multi-processing support" (Michael Elizabeth Chastain). AND enable "RTC support" (from Robert G. Brown). Note that inserting RTC support actually doesn't afaik prevent drift, but according to a discussion [Robert G. Brown] remember from a year ago or so it can prevent lockup when the clock is read at boot time. A note from Richard Jelinek says also that activating the Enhanced RTC is necessary to get the second CPU working (identified) on some original Intel Mainboards. AND (x86 kernel) do NOT enable APM! APM and SMP are not compatible, and your system will almost certainly (or at least probably ;)) crash under boot if APM is enabled (Jakob Oestergaard). Alan Cox confirms this : 2.1.x turns APM off for SMP boxes. Basically APM is undefined in the presence of SMP systems, and anything could occur. AND (x86 kernel) enable "MTRR (Memory Type Range Register) support". Some BIOS are buggy as they do not activate cache for the second processor. The MTRR support contains code that solve such processor misconfiguration. You must rebuild all your kernel and kernel modules when changing to and from SMP mode. Remember to make modules and make modules_install (from Alan Cox). If you get module load errors, you probably did not rebuild and/or re-install your modules. Also with some 2.2.x kernels people have reported problems when changing the compile from SMP to UP. To fix this, save your .config file, do make mrproper, restore your .config file, then remake your kernel (make dep, etc.) (Wade Hampton). Do not forget to run lilo after having copied your brand new kernel. To say it shortly: ___________________________________________________________________ make config # or menuconfig or xconfig make dep make clean make bzImage # or whatever you want # copy the kernel image manually then RUN LILO # or make lilo make modules make modules_install ___________________________________________________________________ 4. How do I make a Linux non-SMP kernel? In the 2.0 series, comment the SMP=1 line in the main Makefile (/usr/src/linux/Makefile). In the 2.2 series, configure the kernel and answer "no" to the question "Symmetric multi-processing support" (Michael Elizabeth Chastain). You must rebuild all your kernel and kernel modules when changing to and from SMP mode. Remember to make modules and make modules_install and to run lilo. See notes above about possible configuration problems. 5. How can I tell if it worked? cat /proc/cpuinfo Typical output (dual PentiumII): ______________________________________________________________________ processor : 0 cpu : 686 model : 3 vendor_id : GenuineIntel [...] bogomips : 267.06 processor : 1 cpu : 686 model : 3 vendor_id : GenuineIntel [...] bogomips : 267.06 ______________________________________________________________________ 6. What is the status of converting the kernel toward finer grained locking and multithreading? 2.2 has signal handling, interrupts and some I/O stuff fine grain locked. The rest is gradually migrating. All the scheduling is SMP safe. 2.3 (next 2.4) is having really fine grained locking. In the 2.3 kernels the usage of the big kernel lock has basically disappeared, all major Linux kernel subsystems are fully threaded: networking, VFS, VM, IO, block/page caches, scheduling, interrupts, signals, etc. (Ingo Molnar) 7. Does Linux SMP support processor affinity? Standard kernel No and Yes. There is no way to force a process onto specific CPU's but the linux scheduler has a processor bias for each process, which tends to keep processes tied to a specific CPU. Patch Yes. Look at PSET - Processor Sets for the Linux kernel : The goal of this project is to make a source compatible and functionally equivalent version of pset (as defined by SGI - partially removed from their IRIX 6.4 kernel) for Linux. This enables users to determine which proces­ sor or set of processors a process may run on. Possible uses include forcing threads to seperate processors, tim­ ings, security (a `root' only CPU?) and probably more. It is focused around the syscall sysmp(). This function takes a number of parameters that determine which function is requested. Functions include: · binding a process/thread to a specific CPU · restricting a CPU's ability to execute some processes · restricting a CPU from running at all · forcing a cpu to run _only_ one process (and its children) · getting information about a CPU's state · creating/destroying sets of processors, to which processes may be bound 8. Where should one report SMP bugs to? Please report bugs to linux-smp@vger.rutgers.edu. 9. What about SMP performances? If you want to try your SMP system, you could run some tests made by Cameron MacKinnon and available at http://www.phy.duke.edu/brahma/benchmarks.smp. 3.2. User Side 1. Do I really need SMP? If you have to ask, you probably don't. :) 2. Do I get the same performance from 2-300 MHz processors as from one 600 MHz processor? This depends on the application, but most likely not. SMP adds some overhead that a faster uniprocessor box would not incur (Wade Hampton). :) 3. How does one display mutiple cpu performance? Thanks to Samuel S. Chessman, here is some useful utilities: Character based: http://www.cs.inf.ethz.ch/~rauch/procps.html Basically, it's procps v1.12.2 (top, ps, et. al.) and some patches to support SMP. For 2.2.x, Gregory R. Warnes as made a patch available at http://queenbee.fhcrc.org/~warnes/procps Graphic: xosview-1.5.1 supports SMP. And kernels above 2.1.85 (included) the cpuX entry in /proc/stat file. The official homepage for xosview is: http://lore.ece.utexas.edu/~bgrayson/xosview.html You'll find a version patched for 2.2.x kernels by Kumsup Lee : http://www.ima.umn.edu/~klee/linux/xosview-1.6.1-5a1.tgz The various forissier's kernel patches are at: http://www- isia.cma.fr/~forissie/smp_kernel_patch/ By the way, you cannot monitor finely processor scheduling with xosview, as xosview itself will cause a scheduling perturbation. (H. Peter Anvin) 4. How can I enable more than 1 process for my kernel compile? use: ___________________________________________________________________ # make [modules|zImage|bzImages] MAKE="make -jX" where X=max number of processes. WARNING: This won't work for "make dep". ___________________________________________________________________ With a 2.2 like kernel, see also the file /usr/src/linux/Documentation/smp.txt for specific instruction. BTW, since running multiple compilers allows a machine with sufficient memory to use use the otherwise wasted CPU time during I/O caused delays, make MAKE="make -j 2" -j 2 actually even helps on uniprocessor boxes (from Ralf Bächle). 5. Why the time given by the time command is false ? (from Joel Marchand) In the 2.0 series, the result given by the time command is false. The sum user+system is right *but* the spreading between user and system time is false. More precisely: "The explanation is, that all time spent in processors other than the boot cpu is accounted as system time. If you time a program, add the user time and the system time, then you timing will be almost right, except for also including the system time that is correctly accounted for" (Jakob Østergaard). This bug is corrected in 2.2 kernels. 3.3. SMP Programming Section made by Jakob Østergaard. This section is intended to outline what works, and what doesn't when it comes to programming multi-threaded software for SMP Linux. 3.3.1. Parallelization methods 1. POSIX Threads 2. PVM / MPI Message Passing Libraries 3. fork() -- Multiple processes Since both fork() and PVM/MPI processes usually do not share memory, but either communicate by means of IPC or a messaging API, they will not be described further in this section. They are not very specific to SMP, since they are used just as much - or more - on uniprocessor computers, and clusters thereof. Only POSIX Threads provide us with multiple threads sharing ressources like - especially - memory. This is the thing that makes a SMP machine special, allowing many processors to share their memory. To use both (or more ;) processors of an SMP, use a kernel-thread library. A good library is the LinuxThreads, a pthread library made by Xavier Leroy which is now integrated with glibc2 (aka libc6). Newer Linux distributions include this library by default, hence you do not have to obtain a separate package to use kernel threads. There are implementations of threads (and POSIX threads) that are application-level, and do not take advantage of the kernel-threading. These thread packages keep the threading in a single process, hence do not take advantage of SMP. However, they are good for many applications and tend to actually run faster than kernel-threads on single processor systems. Multi-threading has never been really popular in the UN*X world though. For some reason, applications requiring multiple processes or threads, have mostly been written using fork(). Therefore, when using the thread approach, one runs into problems of incompatible (not thread-ready) libraries, compilers, and debuggers. GNU/Linux is no exception to this. Hopefully the next few sections will sched a little light over what is currently possible, and what is not. 3.3.2. The C Library Older C libraries are not thread-safe. It is very important that you use GNU LibC (glibc), also known as libc6. Earlier versions are, of course possible to use, but it will cause you much more trouble than upgrading your system will, well probably :) If you want to use GDB to debug your programs, see below. 3.3.3. Languages, Compilers and debuggers There is a wealth of programming languages available for GNU/Linux, and many of them can be made to use threads one way or the other (some languages like Ada and Java even have threads as primitives in the language). This section will, however, currently only describe C and C++. If you have experience in SMP Programming with other languages, please enlighten us. GNU C and C++, as well as the EGCS C and C++ compilers work with the thread support from the standard C library (glibc). There are however a few issues: 1. When compiling C or C++, use the -D_REENTRANT define in the compiler command line. This is necessary to make certain error- handling functions work like the errno variable. 2. When using C++, If two threads throw exceptions concurrently, the program will segfault. The compiler does not generate thread-safe exception code. The workaround is to put a pthread_mutex_lock(&global_exception_lock) in the constructor(s) of every class you throw(), and to put the corresponding pthread_mutex_unlock(...) in the destructor. It's ugly, but it works. This solution was given by Markus Ferch. The GNU Debugger GDB as of version 4.18, should handle threads correctly. Most Linux distribution offer a patched, thread-aware gdb. It is not necessary to patch glibc in any way just to make it work with threads. If you do not need to debug the software (this could be true for all machines that are not development workstations), there is no need to patch glibc. Note that core-dumps are of no use when using multiple threads. Somehow, the core dump is attached to one of the currently running threads, and not to the program as a whole. Therefore, whenever you are debugging anything, run it from the debugger. Hint: If you have a thread running haywire, like eating 100% CPU time, and you cannot seem to figure out why, here is a nice way to find out what's going on: Run the program straight from the shell, no GDB. Make the thread go haywire. Use top to get the PID of the process. Run GDB like gdb program pid. This will make GDB attach itself to the process with the PID you specified, and stop the thead. Now you have a GDB session with the offending thread, and can use bt and the like to see what is happening. 3.3.4. Other libraries ElectricFence: This library is not thread safe. It should be possible, however, to make it work in SMP environments by inserting mutex locks in the ElectricFence code. 3.3.5. Other points about SMP Programming 1. Where can I found more information about parallel programming? Look at the Linux Parallel Processing HOWTO Lots of useful information can be found at Parallel Processing using Linux Look also at the Linux Threads FAQ 2. Are there any threaded programs or libraries? Yes. For programs, you should look at: Multithreaded programs on linux (I love hyperlinks, did you know that ? ;)) As far as library are concerned, there are: OpenGL Mesa library Thanks to David Buccarelli, Andreas Schiffler and Emil Briggs, it exists in a multithreaded version (right now [1998-05-11], there is a working version that provides speedups of 5-30% on some OpenGL benchmarks). The multithreaded stuff is now included in the regular Mesa distribution as an experimental option. For more information, look at the Mesa library BLAS Pentium Pro Optimized BLAS and FFTs for Intel Linux Multithreaded BLAS routines are not available right now, but a dual proc library is planned for 1998-05-27, see Blas News for details. The GIMP Emil Briggs, the same guy who is involved in multithreaded Mesa, is also working on multithreaded The GIMP plugins. Look at http://nemo.physics.ncsu.edu/~briggs/gimp/index.html for more info. 4. x86 architecture specific questions 4.1. Why it doesn't work on my machine? 1. Can I use my Cyrix/AMD/non-Intel CPU in SMP? Short answer: no. Long answer: Intel claims ownership to the APIC SMP scheme, and unless a company licenses it from Intel they may not use it. There are currently no companies that have done so. (This of course can change in the future) FYI - Both Cyrix and AMD support the non- proprietary OpenPIC SMP standard but currently there are no motherboards that use it. 2. Why doesn't my old Compaq work? Put it into MP1.1/1.4 compliant mode. check "Configure Hardware" -> "View / Edit details" -> "Advanced mode" (F7 I think) for a configuration option "APIC mode" and set this to "full Table mode". This is an official Compaq recommandation, although I was never able to find this option in the BIOS... (Daniel Roesen) 3. Why doesnt my ALR work? From Robert Hyatt : ALR Revolution quad-6 seems quite safe, while some older revolution quad machines without P6 processors seem "iffy"... 4. Why does SMP go so slowly? or Why does one CPU show a very low bogomips value while the first one is normal? From Alan Cox: If one of your CPU's is reporting a very low bogomips value the cache is not enabled on it. Your vendor probably provides a buggy BIOS. Get the patch to work around this or better yet send it back and buy a board from a competent supplier. A 2.0 kernel (> 2.0.36) contains the MTRR patch which should solve this problem (select option "Handle buggy SMP BIOSes with bad MTRR setup" in the "General setup" menu). I think buggy SMP BIOS handling is automatic in latest 2.2 kernels. 5. I've heard IBM machines have problems Some IBM machines have the MP1.4 bios block in the EBDA, allowed but not supported below 2.2 kernels. There is an old 486SLC based IBM SMP box. Linux/SMP requires hardware FPU support. 6. Is there any advantage of Intel MP 1.4 over 1.1 specification? Nope (according to Alan :) ), 1.4 is just a stricker specs of 1.1. 7. Why does the clock drift so rapidly when I run linux SMP? This is known problem with IRQ handling and long kernel locks in the 2.0 series kernels. Consider upgrading to a later 2.2 kernel. From Jakob Oestergaard: Or, consider running xntpd. That should keep your clock right on time. (I think that I've heard that enabling RTC in the kernel also fixes the clock drift. It works for me! but I'm not sure whether that's general or I'm just being lucky) There are some kernel fixes in the later 2.2.x series that may fix this. 8. Why are my CPU's numbered 0 and 2 instead of 0 and 1 (or some other odd numbering)? The CPU number is assigned by the MB manufacturer and doesn't mean anything. Ignore it. 9. My quad-Xeon system hangs as soon as it has decompressed the kernel (Doug Ledford) Try recompiling LILO with LARGE_EBDA support and then making sure to always use make bzImage when compiling the kernel. That appears to have fixed the SMP boot hangs here on Intel multi-Xeon boards. However, please note that this also appears to break LILO in that the root= option no longer works, so make sure you rdev your kernel image at the same time you run lilo to make sure that the kernel loads the correct root filesystem at boot. (Robert M. Hyatt) With 3 cpus, do you have a terminator in the 4th slot? 10. During boot machine hang signaling an IOAPIC problem Try boot options "noapic" (John Aldrich) and/or "reboot=bios" (Terry Shull). 11. My system locks up during heavy NFS traffic Try the later 2.2.x kernels and the knfsd patches. This is currently under investigation. (Wade Hampton) 12. My system locks up with no oops messages If you are using kernels 2.2.11 or 2.2.12, get the latest kernel. For example 2.2.13 has a number of SMP fixes. Several people have reported these kernels to be unstable for SMP. These same kernels may have NFS problems that can cause lockups. Also, use a serial console to capture your oops messages. (Wade Hampton) If the problem remains (and the other suggestions on this list didn't help either), then you could try the latest 2.3 kernels. They have more verbose (and more robust) SMP/APIC code, and automatic hard-lockup-prevention code which will produce meaningful oopses instead of a silent hang. (Ingo Molnar) 13. Debugging lockups (item by Wade Hampton) A good means of debugging lockups is to get the ikd patch from Andrea Arcangeli: ftp://ftp.suse.com/pub/people/andrea/kernel- patches There are several of debug options, but do NOT use the soft lockup option! For newer SMP boxes, turn kernel debugging then turn on the NMI oopser. To verify that the NMI oopser is working, after booting the new kernel, /cat /proc/interrupts and verify that you are getting NMIs. When the box locks up, you should get an OOPS. You may also try the %eip option. This allows the kernel to print on the console the %eip address every time a kernel function is called. When the box locks up, write down the first column ordered by the second column then lookup the addresses in the System.map file. This works only in console mode. Also note that the use of a serial console can greatly facilitate debugging kernel lockups, not just SMP kernel lockups! 14. "APIC error interrupt on CPU#n, should never happen" messages in logs A message like: ___________________________________________________________________ APIC error interrupt on CPU#0, should never happen. ... APIC ESR0: 00000002 ... APIC ESR1: 00000000 ___________________________________________________________________ indicates a 'receive checksum error'. This cannot be caused by Linux as the APIC message checksumming part is completely in hardware. It might be marginal hardware. As long as you dont see any instability, they are not a problem - APIC messages are retried until delivered. (Ingo Molnar) 4.2. Possible causes of crash In this section you'll find some possible reasons for a crash of an SMP machine (credits are due to Jakob Østergaard for this part). As far as I (David) know, theses problems are Intel specific. · Cooling problems From Ralf Bächle: [Related to case size and fans] It's important that the air is flowing. It of course can't where cables etc. are preventing this like in too small cases. On the other side I've seen oversized cases causing big problems. There are some tower cases on the market that actually are worse for cooling than desktops. In short, the right thing is thinking about aerodynamics in the case. Extra cases for hot peripherals are usefull as well. Of course you can always go to Radio Shack (or similar) and get another fan. You can use the lm_sensors to monitor the CPU temperature of newer PII and PIII processors. This might help you to determine if heat is a problem. (Wade Hampton) · Bad memory Don't buy cheap RAM and don't use mixed RAM modules on a motherboard that is picky about it. Especially Tyan motherboards are known to be picky about RAM speed (see the Tyan paragraph below for a possible solution). There have been some report of 10ns PC100 RAM being sold with motherboards where the CPU really needs 8ns RAM. (Wade Hampton) · Bad combination of different stepping CPUs Check /proc/cpuinfo to see that your CPUs are same stepping. · If your system is unstable, then DON'T overclock it! ...and even if it is stable, DON'T overclock. From Ralf Bächle: Overclocking causes very subtle problems. I have a nice example, one of my overclocked old machines misscomputes a couple of pixels of a 640 x 400 fractal. The problem is only visible when comparing them using tools. So better say never, nuncas, jamais, niemals overclock. · 2.0.x kernel and fast ethernet (from Robert G. Brown) 2.0.x kernels on high performance fast ethernet systems have significant (and known) problems with a race/deadlock condition in the networking interrupt handler. The solution is to get the latest 100BT development drivers from CESDIS Linux Ethernet device drivers site (ones that define SMPCHECK). · A bug in the 440FX chipset (from Emil Briggs) If you had a system using the 440FX chipset then your problem with the lockups was possibly due to a documented errata in the chipset. Here is a reference References: Intel 440FX PCIset 82441FX (PMC) and 82442FX (DBX) Specification Update. pg. 13 http://www.intel.com/design/pcisets/specupdt/297654.htm The problem can be fixed with a BIOS workaround (Or a kernel patch) and in fact David Wragg wrote a patch that's included with Richard Gooch's MTTR patch. For more information and a fix look here: http://nemo.physics.ncsu.edu/~briggs/vfix.html · DONT run emm386.exe before booting linux SMP From Mark Duguid, dumb rule #1 with W6LI motherboards. ;) · If the machine reboots/freezes after a while, there can be two good BIOS + memory related reasons (from Jakob Østergaard) · If the BIOS has settings like "memory hole at 16M" and/or "OS/2 memory > 64MB", try disabling them both. Linux does not always react well with theese options. · If you have more than 64 MB of memory in the machine, and you specified the exact number manually in the LILO configuration, you should specify one MB less than you actually have in the machine. If you have 128 MB, you lilo.conf line looks like: append="mem=127M" · Be aware of IRQ related problems Sometime, some cards are not recognized or can trigger IRQ conflicts. Try shuffling cards on slots in different ways and possibly moving them to different IRQs. Contributed by hASCII : removing an " append="hisax=9,2,3"" line in lilo.conf allowed using a kernel from the 2.1.xx series with activated ISDN + Hisax support. Kernels from the 2.0.xx series doesn't make problems like this. Try also to set BIOS setup option like "MP 1.4 mode" or "route PCI interrupts through IOAPIC", or "OS Type" not set to DOS neither Novell (Ingo Molnar). · Floppy access while sound is active If you lockup when trying to access the floppy (for example while sound is playing) you may have to edit drivers/pci/quirks.c and set /int isa_dma_bridge_buggy = 1; This is a problem with my Dell WS400 dual PII/300, 2.2.x, SMP (Wade Hampton). 4.3. Motherboard specific information Please note: Some more specific information can be found with the list of Motherboards rumored to run Linux SMP 4.3.1. Motherboards with known problems · none right now 4.4. Low cost SMP Linux box (dual Celeron box) (Stéphane Écolivet) The lowest cost SMP Linux boxes with nowadays buyable processors are dual Celeron systems. Such a system is not officially possible according to Intel. Better think about the second generation of Celeron, those with 128 Kb L2 cache. 4.4.1. Is it possible to run a dual Intel Celeron box ? Official answer from Intel: no, Celeron cannot work in SMP mode. Practical answer: it is possible, but requires hardware alteration for Slot 1 processors. Alteration is described by Tomohiro Kawada on his Dual Celeron System page. Of course, this kind of modification removes warranties... Some versions of Celeron processor are also available in Socket 370 format. In that case, alteration may just be done on the Socket 370 to Slot 1 adapter or may even be sold pre-wired for SMP use. (Andy Poling, Hans - Erik Skyttberg, James Beard) There is also a motherboard (ABIT BP6) allowing two Celerons in Socket 370 format to be inserted (Martijn Kruithof, Ryan McCue). ABIT Computer BP6 verified tested and native to linux with dual ppga socket 370 (Andre Hedrick). 4.4.2. How does Linux behave on a dual Celeron system ? Fine, thank you. 4.4.3. Celeron system ? Celeron processors are known to be easily overclockable. And dual It may work. However, overclocking this kind of system is not as easy as overclocking a mono-processor one. It is definitly not a good idea for a production system. For personal use, dual Celeron 300A systems running rock-solid at 450 MHz have been reported. (numerous people) 4.4.4. And making a quad Celeron system ? It is impossible. Celeron processors have nearly the same features as basic Pentium II chips. If you want more than 2 processors in your system, you'll have to look at Pentium Pro, Pentium II Xeon or Pentium III (?) boxes. 4.4.5. What about mixing Celeron and Pentium II processor ? A system using a "re-enable" Celeron processor and a Pentium II processor with the same steppings may theorically work. Alexandre Charbey as made such a system: · Asus P2B-D motherboard, proc 1: Celeron 366, proc 2: Pentium II 400@266 · 66Mhz and 75Mhz bus frenquencies where functionnal · the fastest processor (in this case the Celeron) should be put on the second slot. Swapping processors (fatest first) leads to quick failure. 5. Sparc architecture specific questions 5.1. Which Sparc machines are supported ? Quoting the UltraLinux web page (only SMP systems): · UltraSPARC PCI based workstations: Ultra60, Ultra450 · UltraSPARC SBUS based servers: Enterprise 1, 2, 150 · UltraSPARC SBUS based large servers: Enterprise 3000, 4000, 5000, 6000, 10000 · UltraSPARC PCI based servers: Enterprise 250, 450 · SPARC sun4m SMP machines (Anton Blanchard) UltraLinux has ran on a 14 CPUs machine (see the dmesg output ). 5.2. Specific problem related to Sparc SMP support (David Miller) There should not be any worries. The only known problem, and one we don't intend to fix, is that if you build an SMP kernel for 32-bit (ie. non-ultrasparc) systems, this kernel will not work on sun4c systems. 5.3. SMP specific limit with current kernel (2.2) (David Miller) There is a bug in the include/linux/tasks.h header file, it needs to define NR_CPUS to 64 on UltraSparc as this is the upper limit for the hardware we support :-) 6. PowerPC architecture specific questions 6.1. Which PPC machines are supported ? · PowerSurge boards (including UMAX s900) · PowerMac · Motorola MTX: support under developement. Patches not yet integrated into the main kernel (Troy Benjegerdes) (Cort Dougan) Not supported: PPC RS/6000 systems 6.2. Specific problem related to PPC SMP support Nothing. Usual SMP compiling (see above). As usual, be aware, modules are specific either for UP or SMP. Recompile them. (Paul Mackerras) 7. Alpha architecture specific questions 7.1. Which Alpha machines are supported ? (Geerten Kuiper) SMP works for most, if not all, AXP servers. (Jay A Estabrook) SMP does seem to work on most of our [Compaq] boxes with 2 or more CPUs. That includes : · AS2000/2100 (SABLE) · AS4000/4100 (RAWHIDE) · DS20 (DP264) It does not include : · AS2100A (LYNX) · TurboLaser bigboys (8200/8400) 7.2. Specific problem related to Alpha SMP support None (really ? :-) 8. Useful pointers 8.1. Various · Parallel Processing using Linux · Linux Parallel Processing HOWTO · (outdated) Linux SMP home page · linux-smp mailing list To subscribe, send subscribe linux-smp in the message body at majordomo@vger.rutgers.edu To unsubscribe, send unsubscribe linux-smp in the message body at majordomo@vger.rutgers.edu Linux SMP archives Linux SMP archives at progressive-comp.com · pthread library made by Xavier Leroy · Motherboards rumored to run Linux SMP · procps · procps patch for 2.2.x · xosview · xosview for 2.2.x · SMP Performance of Linux · CESDIS Linux Ethernet device drivers site · Dual Celeron System 8.2. Multithreaded programs and library · Linux Threads FAQ · Multithreaded programs on linux · Pentium Pro Optimized BLAS and FFTs for Intel Linux (not available right now, but a dual proc library is planned for 5/27/98, see Blas News for details) · Mesa library (with experimental multi-threading) · Parallel plugins for The GIMP 8.3. SMP specific patches · Forissier kernel patches · Patch for a bug in the 440FX chipset · MTRR patch (latest version: 1.9) · PSET - Processor Sets for the Linux kernel · Ingo Molnar SMP patches (for experts only, please read linux-smp@vger.rutgers.edu) 8.4. ( Sumit Roy ) Parallelizing/Optimizing Compilers for 586/686 machines · Pentium Compiler Group creators of pgcc · Absoft , Fortran 90 and Fortran 77 compilers · The Portland Group, Inc. , supports the OpenMP standard for Fortran parallelization on Linux · Pacific-Sierra Research Corporation , has a free F90 compiler for Linux, as well as parallelizing compilers for SMP Linux · Applied Parallel Research , currently have parallelizing compilers for WinNT 9. Glossary · SMP Symmetric Multi-Processors · APIC Advanced Programmable Interrupt Controler · thread A thread is a processor activity in a process. The same process can have multiple threads. Those threads share the process address space and can therefore share data. · pthread Posix thread, threads defined by the Posix standard. · APM Advanced Power Managment 10. List of contributors Many thanks to those who help me to maintain this HOWTO: 1. Tigran A. Aivazian 2. John Aldrich 3. Niels Ammerlaan 4. H. Peter Anvin 5. Guylhem Aznar 6. Ralf Bächle 7. James Beard 8. Troy Benjegerdes 9. Anton Blanchard 10. Emil Briggs 11. Robert G. Brown 12. Alexandre Charbey 13. Michael Elizabeth Chastain 14. Samuel S. Chessman 15. Alan Cox 16. Andrew Crane 17. Cort Dougan 18. Mark Duguid 19. Stéphane Écolivet 20. Jocelyne Erhel 21. Jay A Estabrook 22. Byron Faber 23. Mark Garlanger 24. hASCII 25. Wade Hampton 26. Andre Hedrick 27. Claus-Justus Heine 28. Benedikt Heinen 29. Florian Hinzmann 30. Moni Hollmann 31. Robert M. Hyatt 32. Richard Jelinek 33. Tony Kocurko 34. Geerten Kuiper 35. Martijn Kruithof 36. Doug Ledford 37. Kumsup Lee 38. Hank Leininger 39. Ryan McCue 40. Paul Mackerras 41. Cameron MacKinnon 42. Joel Marchand 43. David Maslen 44. Chris Mauritz 45. Jean-Francois Micouleau 46. David Miller 47. Ingo Molnar 48. Ulf Nielsen 49. Jakob Oestergaard 50. Matt Ranney 51. Daniel Roesen 52. Ulf Rompe 53. Jean-Michel Rouet 54. Volker Reichelt 55. Sean Reifschneider 56. Sumit Roy 57. Thomas Schenk 58. Terry Shull 59. Chris K. Skinner 60. Hans - Erik Skyttberg 61. Szakacsits Szabolcs 62. Jukka Tainio 63. Simen Timian Thoresen 64. El Warren 65. Gregory R. Warnes 66. Christopher Allen Wing 67. Leonard N. Zubkoff