#emc-devel | Logs for 2006-01-12

[18:35:56] <rayh> Shutting down and cleaning up EMC...
[18:35:56] <rayh> scripts/emc: line 592: 14783 Killed sudo $EMC2_BIN_DIR/$EMCIO -ini $INIFILE
[18:36:25] <rayh> Got this after an attempt to get a reply from halcmd save comp.
[18:36:31] <rayh> any clues?
[18:38:17] <rayh> and "sudo rmmod hal_m5i20" fails to remove the module.
[18:38:46] <rayh> * rayh reboots
[19:08:09] <SWP_Away> hiya Ray
[19:09:08] <rayh> Hi steve.
[19:09:15] <SWP_Away> when you got that error, had you loaded any modules with long parameter strings?
[19:09:46] <rayh> It was a m5i20
[19:10:00] <rayh> let me look for the longest param string.
[19:10:06] <SWP_Away> I haven't loaded the driver (ever :) ), though I finally installed the card in the machine
[19:10:48] <SWP_Away> there is a bug in halcmd when saving the parameters, so that may be the cause of the problem if there are long lines
[19:11:36] <rayh> 5i20.0.digital-out-write.tmax looks like the longest
[19:11:46] <SWP_Away> I mean the loadrt line
[19:12:02] <rayh> oh.
[19:12:05] <SWP_Away> like loadrt m5i20 foo=123;bar=thisisalongparameter
[19:12:37] <SWP_Away> heh - funny how you see ambiguity in retrospect
[19:12:55] <SWP_Away> (I wouldn't have though of that interpretation, but now that you mention it ... ;) )
[19:13:03] <SWP_Away> thought
[19:14:45] <rayh> loadrt classicladder_rt numRungs=4 numBits=16 numWords=4 numTimers=2 numMonostables=2 numPhysInputs=16 numPhysOutputs=8 numArithmExpr=4 numSections=1
[19:15:01] <rayh> would probably qualify as a long loadrt string
[19:15:10] <SWP_Away> ok - the same one that was truncated in jmk's initial output
[19:15:11] <SWP_Away> yep
[19:15:15] <rayh> but it isn't started by default
[19:15:24] <SWP_Away> before he increased the max string length
[19:16:03] <SWP_Away> I can't fix the bug now, but I'll describe it so it's in the logs
[19:16:56] <SWP_Away> in the parameter save code, the arguments are copied into the arg_string, using strncpy
[19:17:20] <SWP_Away> the max length is MAX_CMD_LEN each time strncpy is called
[19:17:25] <SWP_Away> strncat, I mean
[19:17:57] <SWP_Away> this allows the string to grow without bound, since up to MAX_CMD_LEN bytes can be added for each arg on the command line
[19:18:37] <SWP_Away> this may be OK, since the data is hal_malloc'd later, but it probably isn't what was intended (or MAX_CMD_LEN ishould MAX_TOKEN_LENGTH or something else)
[19:18:42] <rayh> ok I can see that
[19:19:16] <SWP_Away> ok
[19:19:47] <SWP_Away> ah - OK, it is a potential problem, since arg_string is MAX_CMD_LEN+1 bytes long
[19:20:51] <rayh> No hurry. I've got plenty to work on with watching and editing.
[19:21:06] <SWP_Away> heh - if you overwrite the end of arg_string, you screw up n, the token index in that loop (oops) :)
[19:21:15] <SWP_Away> yep
[19:21:30] <SWP_Away> one day soon, I'll fire up hte emc machine again - right now there's too much other stuff to do
[19:22:34] <SWP_Away> speaking of too much time on our hands, are there any plans for a Fest'06 yet? :)
[19:23:59] <rayh> jmk was talking about having it along with -- before or after cnc-workshop in may.
[19:24:23] <SWP_Away> OK - that's in Indiana or something?
[19:24:25] <rayh> Roland of Cardinal said he would be happy to host the extra session.
[19:24:59] <rayh> though he prefered after. In western IL not far from Davinport iowa
[19:25:05] <SWP_Away> ok
[19:25:16] <cradek> yay IL
[19:25:27] <SWP_Away> after probably makes sense, since we may learn some new features at the workshop
[19:25:38] <SWP_Away> features / bugs / ...
[19:25:39] <rayh> Not quite so far.
[19:26:07] <rayh> And we have immediate access to test machines.
[19:26:25] <rayh> so it need not all be code
[19:27:18] <rayh> We should talk about it -- perhaps this sunday
[19:27:20] <SWP_Away> NO RETROFITS! :)
[19:27:46] <rayh> swp slaps rayh for last year?
[19:27:51] <SWP_Away> but I'm sure you know that better than I
[19:27:53] <SWP_Away> nope
[19:27:59] <SWP_Away> I wasn;t there, remember? ;)
[19:28:14] <SWP_Away> (at least at MazakFest)
[19:28:25] <rayh> The Mazak will be there, right across the parking lot.
[19:28:53] <SWP_Away> as long as the retrofit is done, that's OK
[19:29:14] <SWP_Away> it might be fun to try a "non-standard" retrofit, like a bandsaw or lathe
[19:29:17] <rayh> Yes it will be.
[19:29:30] <rayh> Feature add time by then.
[19:31:14] <SWP_Away> it would be good to have access to appropriate machines for certain features, like threading
[19:32:37] <rayh> I like the idea quite a bit.
[19:32:49] <rayh> Plenty of classroom space. Plenty of internet.
[19:33:02] <SWP_Away> sounds good to me
[19:33:06] <rayh> No limit to open hours
[19:33:19] <SWP_Away> no security cards/guards ...
[19:33:33] <rayh> Inexpensive hotels and food not to far away.
[19:34:04] <rayh> Reminds me of my recent conversation with a sherline users at Los Alamos National Labs.
[19:34:25] <rayh> talk about guarded to the point of paranoia
[19:34:44] <SWP_Away> indeed, especially after the nuclear secrets problems (they had some of those, right?)
[19:34:46] <cradek> yay, Davenport IA is only 350 miles from me
[19:35:09] <cradek> ohh IL, not IA
[19:35:11] <cradek> duh
[19:35:52] <SWP_Away> 1120 miles for me - hmmm
[19:35:54] <cradek> I'm so confused, forget it
[19:36:03] <rayh> cradek: is closer than rayh -- 500 even from here
[19:36:28] <cradek> SWP_Away: that's a long drive even split to two days
[19:36:33] <SWP_Away> Galesburg, IL
[19:36:43] <SWP_Away> ok - 1130 miles to Galesburg ;)
[19:36:47] <rayh> SWP_Away: Need me to pick you up at an airport?
[19:37:04] <SWP_Away> possibly. I had hoped to be able to drive so I can bring computer stuff along
[19:37:21] <cradek> flying into chicago would be cheap from anywhere
[19:37:23] <SWP_Away> the laptop doesn't cut it for m5i20 and the like
[19:37:41] <SWP_Away> I wish I were 13 miles closer to Chicago - it would then be only 15000 miles for a free flight
[19:37:57] <SWP_Away> (I'm 763 away, the limit is 750)
[19:38:19] <rayh> Ah. I could probably bring several extra tower boxes.
[19:38:34] <cradek> I would want to take my mill and everything, which means driving
[19:38:38] <cradek> no way would I ship it
[19:38:44] <rayh> hang them off a network with ssh -X setup.
[19:38:45] <SWP_Away> I was thinking of my dual Opteron, for SMP and 64-bit testing, plus my kiosk machine for m5i20 and other PCI tests
[19:39:04] <rayh> Better drive then.
[19:39:10] <SWP_Away> heh
[19:39:19] <cradek> 1100 miles is possible over two days
[19:39:30] <cradek> but what a drag
[19:39:31] <SWP_Away> we have two vehicles now, so I can bring lots of stuff in the minivan
[19:39:49] <SWP_Away> I've driven to Austin TX from here in two days - driving doesn't bother me
[19:40:35] <SWP_Away> (a little over 2000 miles, and I made it by 8:00 PM the second day)
[19:40:40] <rayh> There you go.
[19:41:06] <SWP_Away> actually, I think I went 2200 miles - I took the southern route
[19:41:08] <cradek> for me the harder part is driving home afterward
[19:41:15] <SWP_Away> heh
[19:41:29] <cradek> staying up late, sleeping in motels, etc
[19:41:43] <SWP_Away> I have a friend in Pittsburgh - not that I'd want to leave $6k worth of computers in the car overnight
[19:41:55] <rayh> Yep.
[19:42:01] <SWP_Away> I'm not sure I'd want to leave the car overnight, actually ;)
[19:42:34] <SWP_Away> though he does, so it probably isn't too bad
[19:53:55] <rayh> just started up a scripts/realtime added two ddt's and tried save comp
[19:54:07] <rayh> # components
[19:54:08] <rayh> loadrt blocks ‰D$‹A‰$èàúÿ¸/
[19:54:20] <SWP_Away> oops - that looks incorrect
[19:55:15] <rayh> seems so but what do I know.
[19:55:39] <rayh> I did just make clean and reinstall so the binaries should be okay.
[19:55:44] <SWP_Away> hmmm
[19:56:10] <SWP_Away> that was "bin/halcmd loadrt blocks ddt=1" or similar?
[19:56:14] <SWP_Away> ddt=2
[23:58:53] <SWP_Away> SWP_Away is now known as SWPadnos