Karl Reichert wrote: > Jan Kiszka wrote: >> Nadym Salem wrote: >>> Hello, >>> >>> working on the same project as Karl Reichert, I'm using the same >>> configuration and start script. As soon as I configure the tdma cycle >>> (tdmacfg rteth0 master 5000) the keyboard stops working. Cursor keeps >>> blinking and the script also runs to the end after that command, so I >>> guess it's only the keyboard and the machine keeps running. >>> >>> Logfiles only say "RTmac/TDMA: init time division multiple access >> control >>> mechanism" >>> >>> Do you have any idea ? >> Often, such plain lock-ups point to IRQ conflicts of the RT subsystem >> with Linux (though Xenomai 2.3 should tell you about this). Please check >> your hardware setup (/proc/interrupts, /proc/xenomai/irq). > > Hi Jan and all others. Will give you further information concering this > problem. > > /var/log/messages: > [...] > Jul 12 13:40:01 rtnet kernel: [15372.838147] *** RTnet 0.9.9 - built on Jun > 7 2007 12:09:59 *** > Jul 12 13:40:01 rtnet kernel: [15372.838148] > Jul 12 13:40:01 rtnet kernel: [15372.838152] RTnet: initialising real-time > networking > Jul 12 13:40:01 rtnet kernel: [15372.869741] Intel(R) PRO/1000 Network Driver > - version 7.1.9 > Jul 12 13:40:01 rtnet kernel: [15372.869746] Copyright (c) 1999-2006 Intel > Corporation. > Jul 12 13:40:02 rtnet kernel: [15373.136666] e1000: 0000:07:04.0: > e1000_probe: (PCI:33MHz:32-bit) 00:0e:0c:d7:d4:e4 > Jul 12 13:40:02 rtnet kernel: [15373.282579] RTnet: registered rteth0 > Jul 12 13:40:02 rtnet kernel: [15373.282583] e1000: rteth0: e1000_probe: > Intel(R) PRO/1000 Network Connection > Jul 12 13:40:02 rtnet kernel: [15373.287434] initializing loopback... > Jul 12 13:40:02 rtnet kernel: [15373.287535] RTnet: registered rtlo > Jul 12 13:40:02 rtnet kernel: [15373.294205] RTcap: real-time capturing > interface > Jul 12 13:40:02 rtnet kernel: [15373.319586] RTmac: init realtime media > access control > Jul 12 13:40:02 rtnet kernel: [15373.330788] RTmac/TDMA: init time division > multiple access control mechanism > [...] > > /proc/xenomai/irq (after reset; I don't have any possibility to access logs > before resetting, due to freezing issue; so most probably not very helpfull): > IRQ CPU0 CPU1 > 213: 0 0 > 216: 1028 1027 [timer] > 217: 0 0 > 226: 2 0 [virtual]
Indeed, that doesn't help much (but it triggered some tiny patch to give IRQ 213 and 217 names). > > /proc/interrupts (same reset issue like above): > CPU0 CPU1 > 0: 959992 0 local-APIC-edge-fasteio timer, > rthal_broadcast_timer > 1: 1296 0 IO-APIC-edge i8042 > 2: 0 0 XT-PIC-XT cascade > 8: 3 0 IO-APIC-edge rtc > 12: 15084 0 IO-APIC-edge i8042 > 14: 5148 0 IO-APIC-edge libata > 15: 10511 0 IO-APIC-edge libata > 18: 0 0 IO-APIC-fasteoi libata > 20: 26586 0 IO-APIC-fasteoi uhci_hcd:usb1, ehci_hcd:usb4, > uhci_hcd:usb5 > 21: 104 0 IO-APIC-fasteoi uhci_hcd:usb2, uhci_hcd:usb6, > HDA Intel > 22: 0 0 IO-APIC-fasteoi ehci_hcd:usb3, uhci_hcd:usb7 > NMI: 0 0 > LOC: 959830 959828 > ERR: 0 > MIS: 0 > > rtnet_start: > #!/bin/bash > mknod /dev/rtnet c 10 240 > ifconfig eth0 down > rmmod e1000 > insmod /usr/local/rtnet/modules/rtnet.ko || exit 1 > insmod /usr/local/rtnet/modules/rtpacket.ko || exit 1 > insmod /usr/local/rtnet/modules/rt_e1000.ko cards=1,0,0,0 || exit 1 > insmod /usr/local/rtnet/modules/rt_loopback.ko || exit 1 > insmod /usr/local/rtnet/modules/rtcap.ko || exit 1 > /usr/local/rtnet/sbin/rtifconfig rtlo up 127.0.0.1 > ifconfig rteth0 up > ifconfig rteth0-mac up > ifconfig rtlo up > insmod /usr/local/rtnet/modules/rtmac.ko > insmod /usr/local/rtnet/modules/tdma.ko > /usr/local/rtnet/sbin/tdmacfg rteth0 master 5000 > /usr/local/rtnet/sbin/tdmacfg rteth0 slot 0 0 > > After configuring tdma cycle length (5000 ms), the strange things happen: > - machine isn't freezing, but you can't use keyboard > - keyboard's leds still responding (e.g. scroll lock turns on/off led) If you can play with scroll lock/num lock, your _are_ using the keyboard. This bug is weird. > - last line is also called, so machine isn't freezing at all, only keyboard > doesn't works anymore > > So, seems to me like some IRQ problems. But I can't access /proc without > reseting to check this. Any chance to - attach a second NIC and log on the box remotely or - attach a null modem and redirect kernel console + a login session to that link? > > Xenomai 2.3.1 (SMP) on linux 2.6.20-11 (from kernel.org) with RTnet 0.9.9 What I-pipe patch? It would be worth try all the latest versions (over 2.6.20) here as well. Just to avoid hunting ghosts! > For configs please see attached files Does booting with maxcpus=1 change the picture? Maybe we are hunting some race that shows up only/more easily on SMP. > >> Does the >> lock-up also happen when you configure RTnet manually for TDMA-less >> operation and call "rtroute solicite ..." e.g.? Typo: s/solicite/solicit/ >> > > How to do that? I guess first make menuconfig and configuring without TDMA > and make install, but how to use this rtroute solicite tool? No man is > available, where can I found further information? README.routing, rtroute --help And no, you can leave TDMA enabled in the config. Anyway, that test is not of highest interest right now (I'm quite sure it will show the same effect). > >> If the issue remains unexplainable, please repeat your precise current >> setup (used kernel/Xenomai/RTnet versions, kernel config, RTnet setup). >> >> Jan >> > > Thanks a lot in advance > Karl > > Jan
signature.asc
Description: OpenPGP digital signature
------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/
_______________________________________________ RTnet-users mailing list RTnet-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/rtnet-users
