Breno wrote: > Wolfgang Grandegger wrote: >> Breno wrote: >> >>> Wolfgang Grandegger wrote: >>> >>>> Wolfgang Grandegger wrote: >>>> >>>>> [EMAIL PROTECTED] wrote: >>>>> >>>>>>> Breno wrote: >>>>>>> >>>>>>>> Wolfgang Grandegger wrote: >>>>>>>> >>>>>>>>> Breno wrote: >>>>>>>>> >>>>>>>>> >>>>>>>>>> Wolfgang Grandegger wrote: >>>>>>>>>> >>>>>>>>>> >>>>>>>>>>> Breno wrote: >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>>> Hi Wolfgang, I got the cannetdriver26 as you suggested some >>>>>>>>>>>> e-mail >>>>>>>>>>>> ago. Sorry my ignorance but how can I get it working o my >>>>>>>>>>>> kernel. I >>>>>>>>>>>> mentioned before that you have a script do get it build. >>>>>>>>>>>> Could you >>>>>>>>>>>> help me again on this issue? >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> For what kernel? >>>>>>>>>>> >>>>>>>>>>> Wolfgang. >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>> Hi Wolfgang, I got a DENX Linux 2.6 Kernel Tree from DENX >>>>>>>>>> project. >>>>>>>>>> The last >>>>>>>>>> commit I have is: >>>>>>>>>> >>>>>>>>>> commit 76e8fbcc9e3e928a6898a93e6077264f5cb8f432 >>>>>>>>>> Author: Wolfgang Grandegger <[EMAIL PROTECTED]> >>>>>>>>>> Date: Thu Mar 27 17:41:09 2008 +0100 >>>>>>>>>> >>>>>>>>>> Socket-CAN: merge with SVN trunk revision 690 >>>>>>>>>> >>>>>>>>>> >>>>>>>>> Can you use linux-2.6.25.4 from kernel.org for your hardware? >>>>>>>>> >>>>>>>>> Wolfgang. >>>>>>>>> >>>>>>>>> >>>>>>>> But it doesn't have support for Xenomai, rigth? If so, which >>>>>>>> version >>>>>>>> should I use? >>>>>>>> >>>>>>> You are right, an ADEOS i-Pipe patch is not yet avaliable for >>>>>>> 2.6.25. >>>>>>> Just to understand your requirements. You want to use the Xenomai >>>>>>> real-time extension because you have hard real-time requirements, >>>>>>> right? >>>>>>> Do you want to access the CAN bus from a Xenomai real-time task? >>>>>>> If yes, >>>>>>> then you should use RT-Socket-CAN. >>>>>>> >>>>>>> >>>>>>>> Anyway,I got the Denx-2.6.25 and I wanna configure it on my system >>>>>>>> regardless of this >>>>>>>> version doesn't have Xenomai support. To get an correct kernel >>>>>>>> image on >>>>>>>> my >>>>>>>> present tree(2.6.24-commit >>>>>>>> 76e8fbcc9e3e928a6898a93e6077264f5cb8f432 ) I >>>>>>>> executed make lite5200_defconfig, at the kernel option and I set >>>>>>>> lite5200.dts >>>>>>>> and then do make zImage. That new tree seems have some >>>>>>>> differences, do >>>>>>>> you >>>>>>>> know what commands do I have to execute? I'm using lite5200 powerpc >>>>>>>> board. >>>>>>>> >>>>>>> It does not make sense to switch to 2.6.25 now if you have real-time >>>>>>> requirements. Please stop jumping from one setup to another. If >>>>>>> something is not working, I'm willing to help. >>>>>>> >>>>>>> Wolfgang. >>>>>>> >>>>>>> >>>>>> Hi Wolfgang, thanks for your support. I'm intending to use CAN >>>>>> along with >>>>>> rt-tasks. I had dropped the rt-socket CAN because of the error it >>>>>> appears >>>>>> when I send messages as explained some e-mails ago. For example: >>>>>> >>>>>> ./rtcanconfig rtcan0 -b100000 start ---> works properly! >>>>>> ./rtcanrecv rtcan1 & >>>>>> ./rtcansend rtcan0 -i0x123 1 2 3 4 5 6 7 8 >>>>>> #0 [8] <0x123> 00 00 00 04 05 06 07 08 ---> it seems to lose the >>>>>> threes >>>>>> first data. >>>>>> >>>>> Then let's try to fix that. Have you built and ran rtcansend and >>>>> rtcanrecv as I suggested (the one created by "$ make" in the >>>>> Xenomai top >>>>> directory)? Did it make a difference? >>>>> >>>>> Do you see errors when you use the rtcanrecv option "-e 0xffff"? >>>>> >>>>> ./rtcanrecv -e 0xffff rtcan1 >>>>> >>>>> >>>>>> My socket-can applications works perfectly!!! >>>>>> >>>>> Fine, but it's not useful to call Socket-CAN from an Xenomai real-time >>>>> application. You will loose determinism. >>>>> >>>>> And we should continue this thread with CC to the Xenomai-help mailing >>>>> list because it might be useful for others as well. >>>>> >>>> Maybe there is something wrong with your MSCAN clock settings. >>>> >>>> What frequency have you specified for >>>> CONFIG_XENO_DRIVERS_CAN_MSCAN_CLOCK (check your kernel .config)? >>>> >>>> And what MPC5200 CPU do you have? Could you show use the initial output >>>> of your bootloader? >>>> >>>> Wolfgang. >>>> >>>> >>>> >>>> >>>> >>>> >>> Hi Wolfgang, I have a stick on my board with lite5200 v2.0b. And at the >>> top of my processor chip I have MPC5200b. The clock frequency is >>> 400MHz. >>> >> >> Well, this information does not match the boot output below. >> >> >>> My first messages on boot process is: >>> >>> U-Boot 1.1.4 (Feb 10 2006 - 11:06:33) >>> CPU: MPC5200 v1.2 at 462 MHz >>> >> >> Seems to be a Rev. A CPU at 462 MHz. 1.1.4 is very old, though. >> >> >>> Bus 132 MHz, IPB 66 MHz, PCI 33 MHz >>> >>> Board: Motorola MPC5200 (IceCube) >>> >> >> BTW: do you have one or two PCI connectors on the board? >> >> >>> I2C: 85 kHz, ready >>> DRAM: 64 MB >>> FLASH: 16 MB >>> PCI: Bus Dev VenId DevId Class Int >>> 00 1a 1057 5803 0680 00 >>> In: serial >>> Out: serial >>> Err: serial >>> Net: FEC ETHERNET >>> IDE: Bus 0: not available Autostarting. Press any key to >>> abort... >>> Hit any key to stop autoboot: 5 >>> => tftp >>> Using FEC ETHERNET device >>> TFTP from server 192.168.1.100; our IP address is 192.168.1.101 >>> Filename '5200/cuImage.52xx'. >>> Load address: 0x100000 >>> Loading: >>> *################################################################# >>> ################################################################# >>> ################################################################# >>> ################################################################# >>> done >>> Bytes transferred = 1329135 (1447ef hex) >>> => bootm >>> ## Booting image at 00100000 ... >>> Image Name: Linux-2.6.24.4-g5e9901e9-dirty >>> Created: 2008-04-15 16:35:09 UTC >>> Image Type: PowerPC Linux Kernel Image (gzip compressed) >>> Data Size: 1329071 Bytes = 1.3 MB >>> Load Address: 00400000 >>> Entry Point: 0040055c >>> Verifying Checksum ... OK >>> Uncompressing Kernel Image ... OK >>> >>> >>> The frequency specified on CONFIG_XENO_DRIVERS_CAN_MSCAN_CLOCK is >>> 66000000Hz. Is that correct? >>> >> >> For MPC5200 Rev. B CPU, it should be 33000000, but you have an old Rev. >> A CPU and in that case 66000000 is correct. >> >> There is some help for corresponding Kconfig option: >> >> config XENO_DRIVERS_CAN_MSCAN_CLOCK >> depends on XENO_DRIVERS_CAN_MSCAN >> int "Clock Frequency in Hz" >> default 66000000 >> help >> >> The MSCAN driver selects the oscillator clock (SYS_XTAL_IN) as >> clock source for MSCAN, which is typically 33 MHz. Due to a >> hardware bug on the MPC5200 Rev. A chips, the IP bus clock >> (IP_CLK) is used instead, which is typically 66 or 132 MHz. >> >> How does your CAN cabling look like. Is the CAN bus properly terminated >> on both ends? >> >> Wolfgang. >> > Hi Wolfgang, it's not a professional cable!!! I just use a 10 pins flat > cable to connect CAN0 and CAN1 on board interfaces for testings, I > mean, I don't have a bus exactly but just a simple cable. Could you > pass me a good reference of cables for a property use?
Handmade cables are OK but you need proper termination with 120 Ohm at both ends of the cable. Either there is on-board termination, sometimes selectable via jumpers, or you have to add resistors at the end of your cable. Take an Ohmmeter and measure the resistance between pin 2 and 7 of your connector. Wolfgang. _______________________________________________ Xenomai-help mailing list [email protected] https://mail.gna.org/listinfo/xenomai-help
