Hugo Braga wrote: > Hi, > > I desabled xeno_irqbench as Jan Kiszka told me to but the system keeps > without response. Sometimes when I run xeno-test, the system stops > working after this message has been shown: > ... > running: cat /proc/xenomai/rtdm/named_devices > Ter Nov 6 15:52:15 BRT 2007 > running: cat /proc/xenomai/rtdm/protocol_devices > Hash ProtocolFamily:SocketType Driver /proc > > Ter Nov 6 15:52:15 BRT 2007 > running: top -bn1c > top - 15:52:16 up 3:59, 3 users, load average: 4.61, 4.08, 2.87 > Tasks: 80 total, 6 running, 74 sleeping, 0 stopped, 0 zombie > Cpu(s): 16.0%us, 11.9%sy, 1.0%ni, 70.3%id, 0.9%wa, 0.0%hi, 0.0%si, 0.0%st > Mem: 507600k total, 500820k used, 6780k free, 13272k buffers > Swap: 979924k total, 40k used, 979884k free, 330728k cached > > PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND > 8715 root 25 0 2880 560 460 R 19.7 0.1 10:17.42 dd if /dev/zero > of > 9220 root 25 0 2880 556 460 R 19.7 0.1 6:37.86 dd if /dev/zero > of > 9695 root 25 0 2880 556 460 R 19.7 0.1 3:50.25 dd if /dev/zero > of > > Ter Nov 6 15:52:16 BRT 2007 > running: ./run -- -sh -T 120 -t0 # latency > * > * > * Type ^C to stop this application. > * > * > ... > > Sometimes it completes latency test 1 and 2 but halts at latency test > 3. In both cases I have to reboot the machine. Even when it completes > latency test 1 or 2, it takes a long time to start showing this > message: > ... > == Sampling period: 100 us > == Test mode: periodic user-mode task > == All results in microseconds > warming up... > RTT| 00:00:01 (periodic user-mode task, 100 us period, priority 99) > ... > > When I disabled "Processor type and features->Symmetric > multi-processing support" and "Processor type and features->Multi-core > scheduler support" in the kernel's configuration file, the system > didn't stop, althought the latency has increased.
By what did it increase precisely? Hundreds or more of microseconds? Do you have SMI disabled in your box (see kernel messages)? > > My machine is: > Intel Core 2 Duo 2.00GHz > Intel Deskop Board D946GZIS > I built linux 2.6.19.7 linux kernel with xenomai 2.3.4 and > adeos-ipipe-2.6.19-i386-1.7-03.patch Can you retry your .config over 2.6.23? Recent fixes for I-pipe did not make it into 2.6.19 patches any more (2.6.20 and 2.6.22 should be OK as well) - just to exclude that we are hunting a fixed bug here. Then, we also need your .config to scan for correlations with "usual problems". Thanks, Jan -- Siemens AG, Corporate Technology, CT SE 2 Corporate Competence Center Embedded Linux _______________________________________________ Xenomai-help mailing list [email protected] https://mail.gna.org/listinfo/xenomai-help
