Re: Problem with LOCK_PROFILING stats in 7.2
On 05.22.2009, at 14:21 , Nikola Knežević wrote: I'm using LOCK_PROFILING to, well, profile some locks in the module I'm working on. I also have INVARIANTS turned on. Numbers I'm getting look correct, however, name of the lock is strange (for some of the locks I'm using): 25725 51777294 10979957 7331938 7 1 408303 896814 /home/knl/work/mx/click/bsdmodule/..//include/ click/sync.hh:395 (sleep mutex:?? ?Yf;??? ??:?) Why am I getting this instead of lock's real name: MasterLock[10060]? To reply to myself :) mtx_init & co. expects the name to be persistent during the whole run. So, never pass values from the stack to it, as I did. Cheers, Nikola___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Problem with LOCK_PROFILING stats in 7.2
Hi, I'm using LOCK_PROFILING to, well, profile some locks in the module I'm working on. I also have INVARIANTS turned on. Numbers I'm getting look correct, however, name of the lock is strange (for some of the locks I'm using): 25725 51777294 10979957 7331938 7 1 408303 896814 /home/knl/work/mx/click/bsdmodule/..//include/ click/sync.hh:395 (sleep mutex:?? ?Yf;??? ??:?) Why am I getting this instead of lock's real name: MasterLock[10060]? Cheers, Nikola___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Failing to get dc-misses with pmc
Hi, I'm trying to get number of L2 cache misses on a system level via: sudo pmcstat -S dc-misses -O /tmp/sample.out However, I'm getting this error: pmcstat: ERROR: Cannot allocate system-mode pmc with specification "dc- misses": Invalid argument I'm running 7.1-RELEASE-p3 on: CPU: Intel(R) Xeon(R) CPU X5472 @ 3.00GHz (3004.49-MHz K8- class CPU) Origin = "GenuineIntel" Id = 0x10676 Stepping = 6 Features = 0xbfebfbff < FPU ,VME ,DE ,PSE ,TSC ,MSR ,PAE ,MCE ,CX8 ,APIC ,SEP ,MTRR ,PGE ,MCA,CMOV,PAT,PSE36,CLFLUSH,DTS,ACPI,MMX,FXSR,SSE,SSE2,SS,HTT,TM,PBE> Features2 = 0xce3bd > AMD Features=0x20100800 AMD Features2=0x1 Cores per package: 4 usable memory = 8570097664 (8173 MB) avail memory = 8243294208 (7861 MB) ACPI APIC Table: What am I doing wrong? Thanks, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Cannot access memory at address 0xb
Hi, my 8-core amd64 machine crashes when I'm trying to run my module. The problem is that I can't debug it, since core dump is broken. ---8<--- GNU gdb 6.1.1 [FreeBSD] Copyright 2004 Free Software Foundation, Inc. GDB is free software, covered by the GNU General Public License, and you are welcome to change it and/or distribute copies of it under certain conditions. Type "show copying" to see the conditions. There is absolutely no warranty for GDB. Type "show warranty" for details. This GDB was configured as "amd64-marcel-freebsd"... Cannot access memory at address 0xb (kgdb) bt #0 0x in ?? () Cannot access memory at address 0x0 (kgdb) --->8--- I can obtain correct coredumps when I'm running everything inside QEMU, but not when I'm running on a real machine. Also, panic doesn't occur inside QEMU :( Any hints? Cheers, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Re: Setting per processor (/core) affinity from within FreeBSD
16 Dec 2008, at 06:40 , Garrett Cooper wrote: I was wondering if anyone has written a utility for FreeBSD to tie a particular process group to a processor / core, similar to what Linux has done with taskset, so that affinity can be properly set with FreeBSD and the ULE scheduler. I believe cpuset(2) will do what you want. It is available starting with 7.1-RELEASE (which isn't released yet, but you can grab 7.1-RC1 to test it out). Hi, what I read from the documentation is that cpuset works on processes. Is it possible to pin a kthread to a particular core? Cheers, N. ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Cluster-wide installation of FreeBSD
Hi, I have a machine running FreeBSD 7.1, and now I would like to "replicate" it to other machines in our cluster. Other machines have smaller disks, and a different processor, but all are amd64. Is there a way to do this kind of installation (either by copying content, or installing and then copying /etc) in automated fashion? I can set up the netboot environment (since I'm using FAI to install linux on the same machines). Cheers, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "freebsd-questions-unsubscr...@freebsd.org"
Compiling C++ modules
Hi, what are the recommended CXXFLAGS for C++ code which should go in kernel module? Yes, I know C++ in kernel is a bad idea, but those are the requirements... Best, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Strange hang with dummynet
On 16 Sep 2008, at 14:38 , Nikola Kne?evi? wrote: I'm running FreeBSD 7.1-PRERELEASE with custom kernel on my box, and when I did: kldload dummynet && kldload ipfw the machine just "hanged" - I couldn't access it over ssh, and current sessions were blocked. This happens also with GENERIC. I didn't touch anything related to ipfw or dummynet's configuration (that I know). Any thoughts? How can I debug this? To reply to myself, for future reference :) dummynet loads ipfw, which has the default rule of drop all. Thus machines shows as inaccessible/unresponsive. Cheers, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: /: write failed, filesystem is full
On 5 Sep 2008, at 04:34 , Gautham Ganapathy wrote: As soon as I start copying files from FTP, this is the error I get: /: write failed, filesystem is full I used to get the same error on freebsd 7.0-release if the bsd partition was not aligned to cylinder boundaries. did sysinstall give any warnings about this when you partitioned your harddrive? No, there was no warning about alignment. Only warning was about disk geometry. I tried to install FreeBSD couple of times, and third time it worked like charm :) No warning, no errors. Since I have two machines which are the same, I will try the second one today. Cheers, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
error: symbol ether_poll_register undefined
Hi guys, I'm trying to port Click modular router to FreeBSD 7.0. Along the way, I changed it quite a bit, but I managed to get it to compile. Well, when I try to load it (via sudo make load), this is what I get: link_elf_obj: symbol ether_poll_register undefined kldload: /usr/home/knl/work/mx-dfly/core/module/mxcore.ko: Unsupported file type What am I missing? I assume that ether_poll_register should be in the kernel. objdump -x -d mxcore.ko gives me a bunch of '*UND*', like _vn_lock, but they exist in the kernel. Cheers, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
/: write failed, filesystem is full
Hi, I'm trying to install FreeBSD 7.0 on our new server (Intel Xeon 2x4- core machine), which has a 250GB SATAII disk. I used -bootonly CD image, and assigned whole disk to FreeBSD. Breakdown is: / - 2GB swap - 4GB /var - 9GB /tmp - 4GB /usr - 213GB As soon as I start copying files from FTP, this is the error I get: /: write failed, filesystem is full This shouldn't happen, since sysinstall did newfs on these partitions. What to do? I'll now try to run livefs disk, to see what is happening there. Cheers, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"
SMP aware programming
Hi, I'm about to port ModelNet (a large-scale internet emulator) from FreeBSD 4 to FreeBSD 7. It is implemented as a module. Well, changing couple of headers and structs will probably solve the thing, but I would like to make it SMP aware. Since it emulates a lot of nodes, each of which is doing some processing, I would like to "group" them, so ones are executing on one core, other on second and so on... Can someone give me some pointers to some good tutorial, text, documentation? I know about multiprogramming, but I never did anything in a kernel. Thank, Nikola ___ freebsd-questions@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-questions To unsubscribe, send any mail to "[EMAIL PROTECTED]"