#linuxcnc-devel | Logs for 2013-09-23

[10:23:35] <skunkworks> micges, the motherboard is a asus FiA55-m LX plus
[10:23:51] <skunkworks> (but I have tried it on a newer one with the same results.
[10:24:20] <micges> I see thanks
[10:24:48] <micges> skunkworks: I'm working on problems, I'll notice you
[10:25:17] <skunkworks> Both the wiki install and mhaberler try
[10:25:22] <skunkworks> Ok
[11:05:03] <norbert> Did someone find out, why editing WIKI Pages is not possible? How is maintaining the WIKI?
[11:06:36] <pcw_home> Its fixed I think
[11:07:37] <norbert> yepp, now I can edit!
[11:08:16] <pcw_home> I think Alex Joni figured it out
[11:13:13] <alex_joni> norbert: should work
[11:13:46] <jepler> alex_joni: thank you
[11:14:50] <norbert> Thanks Alex, I am edditing gmoccapy WIKI at this moment to get it up to date.
[12:29:55] <skunkworks> pcw_home, http://imagebin.org/271719
[12:44:32] <pcw_home> looks like my desk except no assortment of odd encoders
[12:46:30] <skunkworks> heh
[13:02:45] <skunkworks> darn - the nic on the motherboard is a 8168
[13:04:08] <pcw_home> yeah I wish the RT-net people would fix the 8168 support (the 8168 is more common on newer MBs)
[13:04:53] <pcw_home> (8168 is newer the 8169)
[13:05:52] <skunkworks> yes
[13:06:09] <skunkworks> hmm - I don
[13:06:21] <skunkworks> t think I have a intel 1000 here
[13:07:09] <pcw_home> I think there are even bugs in the xenomai standard network support for the 8168
[13:13:34] <skunkworks> hmm - I do have a dell here that may be gigabit. it is intel but don't know if it is pro 1000
[13:15:09] <micges> skunkworks: try this: http://www.mail-archive.com/rtnet-users@lists.sourceforge.net/msg03271.html
[13:32:14] <skunkworks> micges, there doesn't seem to be any RT_ directories
[13:35:13] <micges> I just found it on rtnet ml
[13:36:37] <skunkworks> looks promising
[13:41:51] <micges> there is also partial driver
[13:42:02] <micges> for 8168
[13:58:28] <pcw_home> Yes, if you search for 8168 there is someone who made a 8168 RT-Net driver but
[13:58:29] <pcw_home> it does not appear to have been incorporated in the RT-Net source tree
[14:02:14] <skunkworks> hey - look at this
[14:02:15] <skunkworks> http://pastebin.ca/2457470
[14:02:51] <skunkworks> no encoder errors - but it still generates them (and pwm)
[14:03:05] <skunkworks> and the watchdog didn't bite yet
[14:03:11] <skunkworks> this is using a newer pro/100
[14:15:18] <micges> cool!
[14:15:20] <skunkworks> is this config line right?
[14:15:21] <skunkworks> loadrt hostmot2 loadrt hm2_eth ip="" config="num_encoders=0 num_pwmgens=0 num_stepgens=0
[14:15:30] <skunkworks> well .127
[14:16:08] <skunkworks> loadrt hostmot2 loadrt hm2_eth ip="" config="num_encoders=0 num_pwmgens=0 num_stepgens=0"
[14:16:28] <micges> yes second line is ok
[14:17:10] <skunkworks> odd - so every computer I have tried this on - the config line doesn't seem to work
[14:17:24] <skunkworks> (it always creates the encoders and pwm)
[14:17:43] <micges> oh
[14:19:07] <pcw_home> I wonder if the config line parsing is busted by the IP address setting somehow?
[14:20:07] <jepler> what branch is this in? I can take a look at the parsing and see if anything jumps out
[14:21:40] <micges> parsing config is in hostmot2 module
[14:21:47] <skunkworks> I did this loadrt hostmot2 loadrt hm2_eth config="num_encoders=0 num_pwmgens=0 num_stepgens=0" ip="" with no change in results
[14:22:06] <micges> jepler: no branch only wiki to follow
[14:22:12] <skunkworks> I am currently using the one mhableber did..
[14:22:22] <pcw_home> ( I say that because the default is to enable all hardware so it looks like the config string was ignored)
[14:22:23] <skunkworks> let me find the list email
[14:22:44] <skunkworks> http://article.gmane.org/gmane.linux.distributions.emc.devel/10871/match=7i80
[14:22:55] <skunkworks> jepler, ^
[14:25:08] <jepler> nothing obvious is wrong in the argument parsing
[14:25:16] <jepler> it just passes config on to something else
[14:25:26] <jepler> the usual hm2_register
[14:26:30] <skunkworks> pcw_home, does it not do stepgens by default?
[14:27:26] <pcw_home> the default is to enable all hardware
[14:28:58] <skunkworks> heh
[14:29:21] <pcw_home> (if not mentioned in the config string)
[14:29:22] <pcw_home> I would be tempted to ask for something thats not there to see if it notices
[14:29:24] <pcw_home> num_encoders=24 or some such
[14:29:43] <micges> make typo
[14:29:47] <pcw_home> num 3pwmgens=4
[14:30:01] <jepler> HM2_DBG("parsing config string \"%s\"\n", config_string);
[14:30:08] <jepler> do you have this debug information in your dmesg?
[14:30:39] <jepler> .. I don't think you get HM2_DBG by default
[14:34:22] <pcw_home> there are 5 debug flags, might as well set them all
[14:34:45] <pcw_home> oops just 4
[14:35:54] <pcw_home> debug_idrom,debug_module_descriptors,debug_pin_descriptors,debug_modules
[14:37:16] <jepler> skunkworks: try taking the quotes off ip=
[14:40:06] <jepler> I wonder what halcmd is *supposed* to do with quotes
[14:40:13] <jepler> it's certainly not very shell-like
[14:40:28] <jepler> $ printf "x %s x\n" "hello world" "bye dave"
[14:40:28] <jepler> x hello world x
[14:40:28] <jepler> x bye dave x
[14:40:34] <jepler> halcmd: loadusr -w printf "x %s x\n" "hello world" "bye dave"
[14:40:34] <jepler> "x "hello world" x
[14:40:34] <jepler> ""x "bye dave" x
[14:40:34] <jepler> "halcmd:
[14:40:56] <jepler> halcmd word splitting understands quotation marks, but it leaves them in the argument...
[14:43:34] <jepler> very old versions of insmod(1) say that insmod has its own handling of double-quoted strings, noting that they may require extra shell quoting
[14:45:34] <skunkworks> sorry - this is with the new messaging
[14:45:41] <skunkworks> so this is with quotes
[14:45:43] <skunkworks> http://pastebin.ca/2457478
[14:45:49] <skunkworks> I will try without quotes on the ip
[14:46:04] <skunkworks> Sep 23 14:29:00 empire-System-Product-Name msgd:0: hal_lib:2853:user config="num_encoders=0 num_pwmgens=0 num_stepgens=0 num_3pwmgen=4"
[14:48:12] <skunkworks> jepler, taking the quotes off of the ip address didn't change anything.
[14:49:37] <skunkworks> no watchdog bites yet and no encoder errors... That is a start!
[14:50:06] <jepler> you still don't have the HM2_DBG messages in there. I wish I knew offhand how to turn them on.
[14:50:27] <andypugh> echo 5 > /proc/rtapi/debug
[14:50:38] <jepler> andypugh: also true for xenomai in ubc3 branch?
[14:50:50] <andypugh> Ah! Good question.
[14:50:59] <jepler> this change in hostmot2.h (+ rebuild) would "fix" it
[14:50:59] <jepler> -#define HM2_DBG(fmt, args...) rtapi_print_msg(RTAPI_MSG_DBG, HM2_NAME "/%s: " fmt, hm2->llio->name, ## args)
[14:51:02] <jepler> +#define HM2_DBG(fmt, args...) rtapi_print_msg(RTAPI_MSG_ERR, HM2_NAME "/%s: " fmt, hm2->llio->name, ## args)
[14:51:13] <pcw_home> so its ignoring the config line altogether
[14:52:10] <jepler> hmmm
[14:52:22] <jepler> "HAL: initializing component" is printed at level RTAPI_MSG_DBG
[14:53:24] <skunkworks> hmm - I am doing a DEBUG=5 halrun yata yata
[14:53:24] <micges> seems executin didn't reach there
[14:54:58] <jepler> r = hm2_register(&board->llio, config[num_boards]);
[14:54:58] <jepler> if (r != 0) {
[14:54:58] <jepler> rtapi_print("board fails HM2 registration\n");
[14:54:58] <jepler> return;
[14:54:58] <jepler> }
[14:55:00] <jepler> rtapi_print("board 7i80 registred succesfully\n");
[14:55:03] <jepler> it doesn't print either one of these messages
[14:56:26] <jepler> if (config_string == NULL) {
[14:56:26] <jepler> HM2_PRINT_NO_LL("no firmware specified in config modparam! the board had better have firmware configured already, or this won't work\n");
[14:56:29] <jepler> }
[14:56:36] <jepler> OK, here's *a* problem
[14:56:40] <jepler> num_boards++;
[14:56:40] <jepler> r = hm2_register(&board->llio, config[num_boards]);
[14:56:53] <jepler> here, the config for the first board will be read from config[1], but it's only config[0] that has something
[14:57:14] <micges> ewww
[14:57:17] <jepler> the increment of num_boards should be after the call to hm2_register
[14:58:04] <micges> thanks jepler
[15:00:33] <jepler> skunkworks: did you cut off that last paste?
[15:01:01] <skunkworks> I don't think so...
[15:01:04] <skunkworks> let me look
[15:03:10] <jepler> skunkworks: anyway, this is the edit to try making: http://pastebin.com/0kg1jRgC
[15:03:42] <skunkworks> jepler, I am doing a tail -f of the linuxcnc.log file. when starting halrunning - that is where the terminal stops outputing
[15:05:21] <jepler> skunkworks: It may just be too much debug information generated in an RT context
[15:05:31] <jepler> anyway try with that patch and let us know
[15:06:12] * jepler wanders off
[15:07:19] <pcw_home> thanks for looking into this!
[15:09:09] <jepler> welcome
[15:10:45] <skunkworks> jepler, that worked
[15:10:50] <skunkworks> Thank you!
[15:11:38] <skunkworks> now that config line just makes gpio
[15:11:45] <jepler> nice
[15:11:56] <alex_joni> who want
[15:11:59] <alex_joni> grr
[15:12:06] <alex_joni> who wants to take a look at a schematic?
[15:13:21] <alex_joni> BBB/parport cape
[15:15:05] <andypugh> I don't tkink I have the skiils to look at a schmeatic and make any judgement.
[15:15:30] <skunkworks> the num_3pwmgens=4 doesn't generate any new pins - is that normal?
[15:16:07] <andypugh> You probably don't have any to generate
[15:16:29] <skunkworks> ok - so it would assume some hardware attached to the 7i80?
[15:17:00] <andypugh> They would be part of the firmware.
[15:17:12] <skunkworks> it doesn't error - I just get gpio pins
[15:18:06] <andypugh> If you add encoders do you get encoders?
[15:19:05] <andypugh> It should error if you ask for things that are not there: http://git.linuxcnc.org/gitweb?p=linuxcnc.git;a=blob;f=src/hal/drivers/mesa-hostmot2/tp_pwmgen.c;h=6bcc4da497e9f6a99c88b4d594a12ba39080055d;hb=refs/heads/unified-build-candidate-3#l222
[15:20:51] <skunkworks> Sep 23 14:29:00 empire-System-Product-Name msgd:0: hal_lib:2853:user config="num_encoders=1 num_pwmgens=1 num_stepgens=1 num_3pwmgen=4"
[15:21:03] <skunkworks> gets me 1 pwmgen and 1 encoder
[15:21:23] <skunkworks> let me grab a log
[15:25:56] <pcw_home> try num_3pwmgens=4
[15:26:11] <skunkworks> http://pastebin.com/3nm3Bwue
[15:26:35] <skunkworks> pcw_home, sorry - I corrected that.
[15:30:24] <skunkworks> this is what I get http://pastebin.com/jLjFFrgK
[15:30:31] <skunkworks> bbl
[20:07:54] <micges> ok people I have most of 7i80 driver up and running with rtnet under rtai on 10.04!
[20:10:22] <andypugh> Grand!
[20:11:00] <andypugh> I have the hardware to test if that helps.
[20:11:18] <micges> sure