2011-04-18T00:02:22 *** xiangfu has quit IRC 2011-04-18T00:08:58 *** xiangfu has joined #rtems 2011-04-18T00:53:54 *** cloud_ has joined #rtems 2011-04-18T00:54:13 hello 2011-04-18T00:54:49 is there a way to organize file system in eeprom using rtems? 2011-04-18T00:54:59 *** arvind_khadri has quit IRC 2011-04-18T01:50:46 *** arvind_khadri has joined #rtems 2011-04-18T02:10:47 *** peerst has left #rtems 2011-04-18T02:16:26 *** dr__house has quit IRC 2011-04-18T02:16:31 *** dr__house has joined #rtems 2011-04-18T02:16:31 *** dr__house has joined #rtems 2011-04-18T02:20:26 *** dr__house has quit IRC 2011-04-18T02:20:37 *** dr__house has joined #rtems 2011-04-18T02:20:37 *** dr__house has joined #rtems 2011-04-18T02:26:26 *** dr__house has quit IRC 2011-04-18T02:26:37 *** dr__house has joined #rtems 2011-04-18T02:30:26 *** dr__house has quit IRC 2011-04-18T02:30:37 *** dr__house has joined #rtems 2011-04-18T02:40:26 *** dr__house has quit IRC 2011-04-18T02:40:37 *** dr__house has joined #rtems 2011-04-18T02:40:38 *** dr__house has joined #rtems 2011-04-18T02:42:26 *** dr__house has quit IRC 2011-04-18T02:42:37 *** dr__house has joined #rtems 2011-04-18T02:52:26 *** dr__house has quit IRC 2011-04-18T02:52:37 *** dr__house has joined #rtems 2011-04-18T02:54:26 *** dr__house has quit IRC 2011-04-18T02:54:37 *** dr__house has joined #rtems 2011-04-18T03:28:26 *** dr__house has quit IRC 2011-04-18T03:28:37 *** dr__house has joined #rtems 2011-04-18T03:32:26 *** dr__house has quit IRC 2011-04-18T03:32:37 *** dr__house has joined #rtems 2011-04-18T03:58:26 *** dr__house has quit IRC 2011-04-18T03:58:32 *** dr__house has joined #rtems 2011-04-18T03:58:32 *** dr__house has joined #rtems 2011-04-18T04:00:26 *** dr__house has quit IRC 2011-04-18T04:00:32 *** dr__house has joined #rtems 2011-04-18T04:00:32 *** dr__house has joined #rtems 2011-04-18T04:06:48 is there a way to organize file system in eeprom using rtems? 2011-04-18T04:18:23 *** xiangfu has quit IRC 2011-04-18T04:30:18 *** petben has joined #rtems 2011-04-18T05:00:47 *** arvind_khadri has quit IRC 2011-04-18T05:12:55 *** petben has quit IRC 2011-04-18T05:13:46 *** arvind_khadri has joined #rtems 2011-04-18T06:16:06 *** methril_work has joined #rtems 2011-04-18T06:47:21 *** lj8175 has joined #rtems 2011-04-18T07:07:54 *** petben has joined #rtems 2011-04-18T07:15:19 *** dr__house has quit IRC 2011-04-18T07:53:23 *** cloud_ has quit IRC 2011-04-18T08:36:59 *** zwj has joined #rtems 2011-04-18T08:52:28 *** kristianpaul has quit IRC 2011-04-18T08:52:44 *** kristianpaul has joined #rtems 2011-04-18T08:52:44 *** kristianpaul has joined #rtems 2011-04-18T08:52:44 *** kristianpaul has joined #rtems 2011-04-18T09:20:39 *** DrJoel has joined #rtems 2011-04-18T09:20:39 *** ChanServ sets mode: +o DrJoel 2011-04-18T09:20:45 *** xiangfu has joined #rtems 2011-04-18T09:25:52 * DrJoel waves hello 2011-04-18T09:42:53 DrJoel:hello 2011-04-18T09:44:32 zwj howdy.. how was your weekend? 2011-04-18T09:45:46 it is wonderful. how about the GSOC meeting 2011-04-18T09:46:11 i am now look deep into rtems components 2011-04-18T09:47:12 and think about the method that whether can make rtems configurable to native or hypervisor 2011-04-18T09:48:26 that's the biggest hurdle in the merge I think 2011-04-18T09:50:27 *** antgreen has joined #rtems 2011-04-18T09:50:38 yeah, now all the rtems build is controlled by makefile, and whether there is gui configure for rtmes develop? 2011-04-18T09:50:45 like linux make menuconfig 2011-04-18T09:51:21 no.. no gui.. simple autoconf 2011-04-18T09:54:22 i should be familar with the components build, know more about how to controll which components should be linked which not 2011-04-18T09:55:51 *** petben has quit IRC 2011-04-18T10:11:25 *** zwj has quit IRC 2011-04-18T10:11:34 zwj: good 2011-04-18T10:27:28 *** gedare has joined #rtems 2011-04-18T10:31:51 hi drjoel. how was the renfair 2011-04-18T10:33:58 marvelous!! If you saw the news about the weather in the south this weekend, you will realize we got LUCKY! Friday afternoon and night sucked but it was a beautiful clear day with no clouds Saturday. never got over about 75 2011-04-18T10:34:10 i did, that's great 2011-04-18T10:34:13 for those on Facebook, pictures are plentiful 2011-04-18T10:35:03 We are considering going to the TN renn faire in a few weeks. :-D 2011-04-18T10:35:48 cool 2011-04-18T10:37:35 Huntsville class next week .. and Munich class next month is a go! 2011-04-18T10:38:02 very nice! 2011-04-18T10:38:18 *** xiangfu has quit IRC 2011-04-18T10:38:51 booked tickets last night... 2011-04-18T10:39:21 i have a mess of a proposal that I need to organize and probably rewrite :) 2011-04-18T10:39:28 need to make sure GSoC students know web access to class next week is available and they can attend in Munich or here. But this is a bad time for students to physically come to Huntsville 2011-04-18T10:39:42 If you need a review, I am happy to try to get to it 2011-04-18T10:40:11 Thanks, but I think I'll be OK. I just need to figure out exactly what I plan to do and make sure I know how I will do it. :) 2011-04-18T10:40:43 ok 2011-04-18T10:40:56 *** core-ix has joined #rtems 2011-04-18T10:46:26 interesting discussion about how rtems evolves. time for you to make a plea for support contracts :P 2011-04-18T10:49:28 which thread? 2011-04-18T10:49:46 It evolves as people either pay for things or implement them themselves. :-D 2011-04-18T10:52:15 the bit words.h one 2011-04-18T10:56:58 *** antgreen has quit IRC 2011-04-18T11:04:32 *** lj8175 has quit IRC 2011-04-18T11:18:50 yeah.. just emailed response 2011-04-18T11:19:52 *** arvind_khadri has quit IRC 2011-04-18T12:38:06 *** Shisui has joined #rtems 2011-04-18T12:38:52 Hi 2011-04-18T12:41:05 hey Shisui 2011-04-18T12:42:28 hey DrJoel 2011-04-18T12:43:00 I was just looking for someone who could help me :) 2011-04-18T12:44:15 could you tell me if there is any configuration needed to bring msdos mount option in rtems shell ? 2011-04-18T12:44:55 even with the CONFIGURE_SHELL_MOUNT_MSDOS define the msdos FS is not listed in the list of FS that "mount" can understand 2011-04-18T12:45:41 however I successfully managed to format a ramdisk with this fs. 2011-04-18T12:45:47 I'm a bit confused! 2011-04-18T12:55:32 kiwichris (Chris Johns) is the best source but let's see if I can help 2011-04-18T12:56:11 #define CONFIGURE_SHELL_MOUNT_MSDOS 2011-04-18T12:56:17 is there .. ok 2011-04-18T12:56:56 CONFIGURE_USE_IMFS_AS_BASE_FILESYSTEM? 2011-04-18T12:57:58 defined too. 2011-04-18T12:58:26 CONFIGURE_FILESYSTEM_DOSFS? 2011-04-18T12:59:33 *** arvind_khadri has joined #rtems 2011-04-18T13:00:32 hm.. with this define mount -L shows the "dosfs" filesystem, but i'm not able to mount my msdos formatted ramdisk 2011-04-18T13:02:00 It's strange to see the differences between the examples from examples-v2 CVS branch and the actual system. 2011-04-18T13:02:08 are these examples really old ? 2011-04-18T13:03:06 ohhh.. sorry.. the filesystem examples are new and the dos one doesn't work yet. I wrote it and kiwichris hasn't gotten around to telling me what is broken. Sorry 2011-04-18T13:03:21 I wanted a set of the simplest possible examples of using a RAM disk with various file systems 2011-04-18T13:03:26 but they don't work yet 2011-04-18T13:04:20 eeeeh, good to know ! I tried it for a while without success. 2011-04-18T13:07:07 email chrisj@rtems.org and ask for some help.. cc me on it. 2011-04-18T13:08:27 fwiw I worked on an example of what I think the fstests should more or less look like. I should probably email that to you and Chris for review. I did a simple test that creates a file, writes it, and then reads it back and verifies it. But the code can be run on IMFS, mounted IMFS, etc. So the same test on multiple file systems. Structure is more or less what I had in mind 2011-04-18T13:11:03 hm.. this one is only initializing a ramdisk, formatting it with FAT and running a shell, is this what you're talking about ? 2011-04-18T13:21:58 *** arvind_khadri has quit IRC 2011-04-18T13:22:07 *** arvind_khadri has joined #rtems 2011-04-18T13:35:04 *** panzon has joined #rtems 2011-04-18T13:41:44 yep.. that's the one 2011-04-18T13:42:00 examples-v2/filesystem/fat_ramdisk/ 2011-04-18T14:09:49 *** core-ix has quit IRC 2011-04-18T14:35:24 *** Shisui has quit IRC 2011-04-18T14:47:54 *** core-ix has joined #rtems 2011-04-18T15:45:45 *** gedare has quit IRC 2011-04-18T16:06:30 im trying to output data with vga 2011-04-18T16:06:55 with this xilinx spartan 3 fpga 2011-04-18T16:08:01 first directly. later through leon3 & rtems 2011-04-18T16:10:23 but all i get is black screen :) 2011-04-18T16:10:51 at least monitor detectes 640x480@ 31kHz/59Hz mode 2011-04-18T16:23:56 :-D 2011-04-18T16:24:11 will you be using this for a console display or graphics? 2011-04-18T16:30:19 both maybe 2011-04-18T16:31:52 graphics would be cool.. realtime indicators for some sensor values 2011-04-18T16:33:18 i need better monitor, this one does not give me enought debug data about whats wrong with my signal :D 2011-04-18T16:40:03 I was asking because the vga console code is tied to the pc386 now.. should be easy to abstract it to work on multiple environments.. graphics solution is currently microwindows and that would layer on top of a framebuffer driver for it 2011-04-18T16:42:08 ok, i will check it if i can get colors :D 2011-04-18T16:42:27 and of course i have to get leon3 running, and rtems running 2011-04-18T16:42:29 :> 2011-04-18T16:44:26 grlib has vga.. 2011-04-18T16:44:28 hmm 2011-04-18T16:45:09 640x480 80x37 textmode 2011-04-18T16:46:14 ahh then you have enough for console.. microwindows would just add windows/graphics package 2011-04-18T16:47:37 yep.. for console i can just modify rtems to support leon3 vga textmode 2011-04-18T16:48:11 for graphics i need to modify or add new leon3 vga graphics module 2011-04-18T16:48:57 or not.. there is already svgactrl 2011-04-18T16:48:59 :D 2011-04-18T18:07:50 *** DrJoel has quit IRC 2011-04-18T19:54:46 *** dr__house has joined #rtems 2011-04-18T19:54:46 *** dr__house has joined #rtems 2011-04-18T20:31:27 *** dr__house has quit IRC 2011-04-18T20:57:47 *** kristianpaul has quit IRC 2011-04-18T20:57:57 *** kristianpaul has joined #rtems 2011-04-18T20:57:57 *** kristianpaul has joined #rtems 2011-04-18T21:06:23 *** xiangfu has joined #rtems 2011-04-18T21:12:15 *** xiangfu has quit IRC 2011-04-18T21:18:47 *** core-ix has quit IRC 2011-04-18T21:26:45 *** core-ix has joined #rtems 2011-04-18T22:27:12 *** arvind_khadri has quit IRC 2011-04-18T22:27:35 *** arvind_khadri has joined #rtems 2011-04-19T00:24:21 *** arvind_khadri has quit IRC 2011-04-19T01:38:14 *** arvind_khadri has joined #rtems 2011-04-19T01:44:42 *** xiangfu has joined #rtems 2011-04-19T01:49:57 *** xiangfu has quit IRC 2011-04-19T02:02:10 *** xiangfu has joined #rtems 2011-04-19T02:06:52 *** projectgus has quit IRC 2011-04-19T02:42:58 *** core-ix has quit IRC 2011-04-19T03:27:36 *** core-ix has joined #rtems 2011-04-19T04:46:30 *** core-ix has quit IRC 2011-04-19T04:56:15 *** xiangfu has quit IRC 2011-04-19T05:01:50 *** arvind_khadri has quit IRC 2011-04-19T05:14:00 *** arvind_khadri has joined #rtems 2011-04-19T06:02:03 *** lj8175 has joined #rtems 2011-04-19T06:54:06 *** rokka has quit IRC 2011-04-19T06:56:17 *** rokka has joined #rtems 2011-04-19T07:21:08 *** DrJoel has joined #rtems 2011-04-19T07:21:08 *** DrJoel has joined #rtems 2011-04-19T07:21:08 *** ChanServ sets mode: +o DrJoel 2011-04-19T07:21:29 *** DrJoel has quit IRC 2011-04-19T07:55:35 *** xiangfu has joined #rtems 2011-04-19T08:01:30 *** xiangfu has quit IRC 2011-04-19T08:06:33 *** DrJoel has joined #rtems 2011-04-19T08:06:33 *** ChanServ sets mode: +o DrJoel 2011-04-19T08:13:22 *** xiangfu has joined #rtems 2011-04-19T08:30:05 *** zwj has joined #rtems 2011-04-19T08:34:24 *** authors has joined #rtems 2011-04-19T08:42:31 *** authors has quit IRC 2011-04-19T08:44:28 *** xiangfu has quit IRC 2011-04-19T08:46:18 *** arvind_khadri has quit IRC 2011-04-19T09:41:00 *** peerst has joined #rtems 2011-04-19T09:41:09 hi 2011-04-19T09:46:59 hey peerst 2011-04-19T09:47:27 Hopefully we can have dinner again together when I am in Munich.. (May 30-June 2) 2011-04-19T09:53:11 *** arvind_khadri has joined #rtems 2011-04-19T09:53:11 *** arvind_khadri has joined #rtems 2011-04-19T10:12:07 *** gedare has joined #rtems 2011-04-19T10:17:09 hi DrJoel, ah the training in Munich filled up? 2011-04-19T10:17:37 enough to have it. :-D 2011-04-19T10:17:43 great 2011-04-19T10:18:14 if I'm in country I will be looking forward to the dinner 2011-04-19T10:18:16 this is the first time I have had enough during the summer months and can take the family. We will be there for the weekend before and the weekend after.. 2011-04-19T10:18:49 I don't know if we have any gsoc students in the area.. aanjhan is still reasonably close 2011-04-19T10:19:28 from where is aanjhan? 2011-04-19T10:20:29 btw on saturday I shipped the first 5 units of the gadget I was developing since fall of 2009 2011-04-19T10:20:32 switzerland I think.. we crossed wires in 2009 when I was there and he didn't make it in 2011-04-19T10:21:07 having two powerpc 5517 gateways powered with rtems 2011-04-19T10:21:54 yeah switzerland is reasonably close 2011-04-19T10:22:48 cool! Is this something that could be on one of the application flyers? 2011-04-19T10:23:11 odd question.. do I understand the "Bayern Ticket" correctly? Up to a party of 5 on one ticket? 2011-04-19T10:23:15 so if you want this on some list of "things built with RTEMS" on the website ... 2011-04-19T10:23:22 yeah 2011-04-19T10:23:25 sure 2011-04-19T10:23:29 indeed yep on the website 2011-04-19T10:24:00 everything under my control with this project ... don't have to ask a customer or such 2011-04-19T10:24:31 So for about 30EU we could take a day trip to another place in Bavaria? I found this mentioned as a way to visit Salzburg from Munich. Any other suggested destinations? 2011-04-19T10:24:38 The thing is described here: http://www.stritzinger.com/Peer-Stritzinger.GmbH/Hydraprog-3_files/hydraprog_Broschuere_Web.pdf 2011-04-19T10:25:01 bayern ticket is only valid on some trains 2011-04-19T10:25:24 I saw that.. the REs.. 2011-04-19T10:25:33 yep 2011-04-19T10:26:21 so this is a device to program multiple devices at the same time? 2011-04-19T10:31:46 sorry was on the phonw 2011-04-19T10:31:49 yeah 2011-04-19T10:32:21 its when the whole device stock is to be reprogrammed before it gets built into the car 2011-04-19T10:32:56 because reprogramming it when it is in the car is the most expensive way (whole cars standing around) 2011-04-19T10:34:21 DrJoel: here there are tips what to do with the bayern ticket.http://www.bahnland-bayern.de/freizeittipps 2011-04-19T10:34:52 but I can give you lots of tips myself if you tell me what you like to do? 2011-04-19T10:36:42 nature? culture? beer drinking? 2011-04-19T10:37:28 culture = beer_drinking+nature 2011-04-19T10:37:44 sounds good to be :-) 2011-04-19T10:38:00 cool brochure. I don't understand half of it, but it looks nice 2011-04-19T10:38:14 thanks 2011-04-19T10:38:41 a friend of mind is graphics designer, hired her for that 2011-04-19T10:38:52 in print it even looks nicer 2011-04-19T10:38:57 cool. it has a nice mix of words and pics 2011-04-19T10:39:18 I don't have a good feeling for real-time applications yet... mostly RT theory and RTOS design 2011-04-19T10:39:43 so it is nice to see how things are put together 2011-04-19T10:39:52 in the box there is a interesting mix of realtime techniques implemented 2011-04-19T10:40:17 are the channels isolated? 2011-04-19T10:40:30 on the gateways (the modules that poke out left an right of the box) I'm using RTEMS 2011-04-19T10:40:49 insolated in which way? electrical? software? 2011-04-19T10:40:56 software sorry 2011-04-19T10:41:11 the behave as if they are isolated 2011-04-19T10:41:24 each channel has its own gateway (mpc5517) 2011-04-19T10:41:38 that talks USB-2 on the inside end 2011-04-19T10:41:51 and the automotive protocols on the outside 2011-04-19T10:42:01 like CAN, FlexRay, MOST, LIN etc. 2011-04-19T10:42:21 makes sense 2011-04-19T10:42:29 in the box there is also a mini ITX motherboard running the central software 2011-04-19T10:42:53 Intel Atom, with LVDS LCD screen that you can see 2011-04-19T10:42:56 that is what handles the input of flashing and global timing? 2011-04-19T10:42:59 ah 2011-04-19T10:43:29 the gateways also handle switching and detecting the power for the devices to be programmed 2011-04-19T10:43:30 neat. 2011-04-19T10:43:41 and the big LED's on the front 2011-04-19T10:43:58 on the ITX board I run FreeBSD from a CF-Card 2011-04-19T10:44:06 i was just about to ask :) 2011-04-19T10:44:09 in a NanoBSD setup 2011-04-19T10:44:23 which is a pretty nice tool for embedded use of FreeBSD 2011-04-19T10:44:50 http://www.freebsd.org/doc/en_US.ISO8859-1/articles/nanobsd/howto.html 2011-04-19T10:44:51 Did you roll your own rtems USB and automotive stacks? 2011-04-19T10:44:55 yup 2011-04-19T10:45:28 i used something like nanobsd before, for usb booting a minimal os 2011-04-19T10:45:34 i forget what it was though. 2011-04-19T10:45:38 under FreeBSD I use Erlang for all highlevel and also quite a bit of lowlevel stuff 2011-04-19T10:45:49 interesting 2011-04-19T10:46:28 and the comibnation of Erlang for soft-realtime and RTEMS for hard-realtime and close to the metal stuff is really great 2011-04-19T10:46:46 yeah, I bet 2011-04-19T10:46:50 Erlang gives me at least a factor of 10 productivity for implementing this 2011-04-19T10:47:28 its part of the secret sauce that makes me not fear my competitors 2011-04-19T10:48:17 so don't tell them ;-) 2011-04-19T10:48:49 the device is the third generation of these devices 2011-04-19T10:49:15 haha ok 2011-04-19T10:49:19 the first 2 generations I build with somebody else ... but this did'nt work out 2011-04-19T10:49:31 why not 2011-04-19T10:49:51 (was only joking ... it wouldn't help them if they knew) 2011-04-19T10:50:12 the short reason: he was an A.....e 2011-04-19T10:50:30 the long: also inept 2011-04-19T10:50:56 the third gen I did alone with hardware developed by embedded-brains 2011-04-19T10:51:43 Thomas built the gateway and a cental USB-Hub/Ethernet-Swich/Powerdistribution board 2011-04-19T10:51:55 the housing is custom built 2011-04-19T10:52:19 worked over 1,5 years on it 2011-04-19T10:53:14 Ah, so personnel and not necessarily technical reasons :) 2011-04-19T10:53:23 both reasons 2011-04-19T10:53:48 he insisted doing the hardware and was not able to do it professionally enough 2011-04-19T10:53:57 ahhh 2011-04-19T10:54:17 the embedded-brains folks seem very professional. 2011-04-19T10:54:22 beside hindering the potential of the software 2011-04-19T10:54:40 Thomas is by far the best hardware developer that I ever met 2011-04-19T10:55:08 Sebastian is really good with software and RTEMS internals 2011-04-19T10:55:28 the others you don't find here on IRC 2011-04-19T10:55:45 i don't think i've interacted with thomas 2011-04-19T10:56:30 the main problem with the non professional hardware was that our customers are very professional hardware developers (building control units for cars) 2011-04-19T10:56:44 and they want to look inside the box before buying it 2011-04-19T10:57:09 makes sense 2011-04-19T10:57:14 and they are big companies that have often problems buying from small companies 2011-04-19T10:58:12 so the hardware really has to send the message: this is serious -- we will do this also in 10-20 years and not just tinkering a bit here 2011-04-19T10:58:25 and this message the old hardware didn't send 2011-04-19T10:58:29 at all 2011-04-19T10:58:42 *** arvind_khadri has quit IRC 2011-04-19T10:59:11 yeah 2011-04-19T10:59:36 very cool, congratulations on shipping :) 2011-04-19T10:59:37 regarding your question of the automotive and usb stacks on RTEMS 2011-04-19T10:59:45 they are pretty lightweight 2011-04-19T11:00:00 only some simple queues built out of chains 2011-04-19T11:00:14 BTW chains are great in RTEMS 2011-04-19T11:00:30 yeah they make simple queuing easy 2011-04-19T11:00:47 just picking some packets from the USB and shoving them out on the CAN e.g. 2011-04-19T11:00:50 and back 2011-04-19T11:01:07 and all the rest runs in Erlang at the moment 2011-04-19T11:01:22 (with the factor 10 productivity) 2011-04-19T11:01:23 yeah i could see that, FIFO queues are very useful to decouple producers/consumers. 2011-04-19T11:02:10 nice 2011-04-19T11:02:13 for performance reasons (programming time is the main feature the customer is interested in) 2011-04-19T11:02:31 what does the customer provide for the configurations 2011-04-19T11:03:00 I will shove everything that needs a short turn around (flow control on the CAN) down to the gateway 2011-04-19T11:03:19 black-box programs? 2011-04-19T11:03:25 there is a rule based scripting language on top 2011-04-19T11:03:29 ahhh 2011-04-19T11:03:32 black-box? 2011-04-19T11:04:07 i'm not aware of what the use case is yet... I think the customer wants to reprogram control units in cars? 2011-04-19T11:04:26 not in cars but before they go into the cars 2011-04-19T11:04:30 ahh 2011-04-19T11:04:37 *** lj8175 has quit IRC 2011-04-19T11:04:38 right, meant for cars. 2011-04-19T11:04:51 So, the programs are control unit dependent? 2011-04-19T11:04:53 car manufacturers dont want to get delivered control units with old software 2011-04-19T11:05:19 because programming the car is expensive (needs lot of room) 2011-04-19T11:06:14 *** zwj has quit IRC 2011-04-19T11:06:14 ok. so your tool hooks up with control units directly so that new programs can be loaded 2011-04-19T11:06:18 an in a global economy, my customery buy CPU's which are already preprogrammed (about 6-8 week lead time when you want to change the software) 2011-04-19T11:06:44 ok that makes sense... 2011-04-19T11:06:59 then some board is built in country A shipped to country B where it is assembled into a control unit then going to coutry C where the cars a assembled 2011-04-19T11:07:33 and in country C close to the assembly line of the car you need some stock with no-risk transportation 2011-04-19T11:07:45 about 2 weeks assembly line capacity 2011-04-19T11:08:22 and there is the main use of Hydraprog: programming this stock before it goes to the assembly line 2011-04-19T11:08:34 OK I see it now 2011-04-19T11:09:15 so you need safe handling, just plugging autostarting, easy readably result display red-bad , green-good 2011-04-19T11:09:17 so a customer provides a program, tells the hydraprog where/how to send it (with script), and hydraprog deals with the control unit interface? 2011-04-19T11:09:37 the programm files are put on a USB stick 2011-04-19T11:10:10 ok. and the script pulls programs off the USB and ships them down to the modules, which know the automotive protocol to use to load the program onto the control unit? 2011-04-19T11:10:55 *** arvind_khadri has joined #rtems 2011-04-19T11:11:26 together with a script that does things like this: if the bootloader version on the device is such and such first update to bootloader version bla, then rematch the rules 2011-04-19T11:11:50 right, the script is the "program" for the hydraprog 2011-04-19T11:12:05 and the control unit program is just a payload 2011-04-19T11:12:10 if I understood :) 2011-04-19T11:12:11 its rule based and has pattern matching so you can write one script that picks up where it left off without keeping state 2011-04-19T11:12:24 yes its only payload 2011-04-19T11:12:30 okay 2011-04-19T11:12:40 cool, i learned something new today 2011-04-19T11:13:09 because sometimes you have chains of bootloader versions you first have to go through until you can get to the right version 2011-04-19T11:13:11 :-) 2011-04-19T11:13:26 real world system and product building 2011-04-19T11:14:05 if you are interested in all kinds of realtime stuff I can only strongly recommend looking into Erlang 2011-04-19T11:15:02 it has great features that are very useful also outside the Telecom business (where it matured for > 20 years) 2011-04-19T11:15:20 and you learn functional programming and pattern languages at the same time 2011-04-19T11:20:03 I'll have to check it out 2011-04-19T11:22:21 i'm more interested in hw/rtos integration, but languages are always fun 2011-04-19T11:23:17 yeah there Erlang is more on the top end of the stack, more farther from the hardware 2011-04-19T11:23:35 thats why I need RTEMS on the gateways :-) 2011-04-19T11:35:59 * DrJoel always like a variety of programming languages.. that's why I keep pushing RTEMS to support more.. Ada, Go.. working in background on Objective-C 2011-04-19T11:36:55 peerst: we really don't know where we would like to go on RE in Bavaria for a day trip. Open to suggestions 2011-04-19T11:37:35 hiking is out.. probably a visit to an interesting town.. visit what we can on foot.. 2011-04-19T11:44:54 DrJoel: have you bee to Königsee yet? 2011-04-19T11:45:07 been 2011-04-19T11:47:57 DrJoel: regarding Objective-C on RTEMS: I have production code (a embedded webservice) using Objective-C on RTEMS ... also very nice combination of Objective-C beeing simply C (full mix of rtems and objects) and also very dynamic and late binding. 2011-04-19T11:49:20 I went with Thomas on my first trip to Munich. It is lovely. The family hasn't been though. 2011-04-19T11:49:46 I know exactly why Apple likes using Objective-C on their mobile devices, much better performance if you need it and being more flexible and dynamic when you need that, as e.g. VM based things like Java which can do a bit of both but nothing well. 2011-04-19T11:49:49 Objective-C appears to be just a matter of getting the objc threading primitive adapters implemented. Only about 20 routines I think. 2011-04-19T11:49:57 yes 2011-04-19T11:50:06 mostly it just works 2011-04-19T11:50:20 did it on 4.6 or so 2011-04-19T11:50:23 All non-threading tests appear to pass on sparc 2011-04-19T11:50:41 will probably port it to current this year 2011-04-19T11:51:00 you have objective-c test cases? 2011-04-19T11:51:43 the ones in GCC 2011-04-19T11:52:00 Do you have implementations of the _objc_ routines in gthr-rtems.h? 2011-04-19T11:52:06 Königssee is very nice, you can go there with train and go over with a ship to st.bartholomä which is totally nice 2011-04-19T11:52:28 The area is really beautiful 2011-04-19T11:52:43 I forgot the state of the stuff, didn't touch it for a few years 2011-04-19T11:53:20 or you can visit the some of the castles? 2011-04-19T11:53:25 if you can dig out that file I will be happy to see if it helps me. This is pure volunteer hobby time for me on this and if it helped, that would be awesom 2011-04-19T11:54:31 That's an idea. I think you know how Thomas got us to Neuschwanstein too late for the last tour. LOL 2011-04-19T11:54:39 I have to check this out anyway soon, because at the moment we are only using objective-c in one thread but we will need multiple threads in the next project step (which just got funded so i know it will happen) 2011-04-19T11:54:58 oh so you did only see it from the outside 2011-04-19T11:55:29 yep.. saw the castle across the valley from it. is that a possibility by train? I don't remember any rail near it 2011-04-19T11:55:52 there is lots of stuff that can be done on short notice without a lot of preplanning, so we can work it out just before you come 2011-04-19T11:56:24 its not so easy i think, more a train/bus combination 2011-04-19T11:57:38 if I am here maybe I can also take you some place filling in for thomas 2011-04-19T11:57:48 ok.. thanks. Our German sucks but we are adventurous 2011-04-19T11:57:52 have you been to andechs? 2011-04-19T11:57:57 that would be appreciated if we need it 2011-04-19T11:58:01 no.. 2011-04-19T11:58:05 :-) 2011-04-19T11:58:25 another place to go ... if you like beer 2011-04-19T11:58:58 its a monastery on top of a hill that brews monk beer 2011-04-19T11:59:08 and there is a nice beer garden there 2011-04-19T12:00:29 cool.. I will have to run these by Michele and the family. this is the most site-seeing time I have had on a trip. I have done a lot in the city but not ventured out as much 2011-04-19T12:00:33 the main trouble is picking some place for the short time 2011-04-19T12:01:23 do you mean that would be OK for just a day trip? 2011-04-19T12:01:31 andechs? 2011-04-19T12:01:41 yes its even ok for half day 2011-04-19T12:01:45 its pretty close 2011-04-19T12:02:02 can be combined with other things 2011-04-19T12:02:26 königssee is defintely a whole day trip 2011-04-19T12:03:10 Is there a weekend or two day version of the Bayern ticket? Or do we just buy day versions .. if we do two thing 2011-04-19T12:03:35 we usually do this area as a 2-3 day trip but we hike around in the national park and that takes time 2011-04-19T12:03:57 i think its only day passes 2011-04-19T12:04:04 but i'm not sure 2011-04-19T12:04:23 Michele has some joint trouble so trying to stick to somewhat lighter activities 2011-04-19T12:04:42 *** gedare has quit IRC 2011-04-19T12:04:54 I see yes, all of what I suggested is almost no walk 2011-04-19T12:13:09 sorry phone.. 2011-04-19T12:13:26 FWIW my Simple SMP scheduler is looking pretty good on the scheduler simulator.. 2011-04-19T12:13:36 and Jennifer is working through the "Big Kernel Lock" 2011-04-19T12:13:52 goal is simple functional SMP.. then to fine-tune it 2011-04-19T12:14:08 yeah way to go 2011-04-19T12:14:26 first get it running then optmize it 2011-04-19T12:15:11 if there was LTE support for powercp as well I could start my second cores on the gateways (not that I needed them yet) 2011-04-19T12:22:13 That's the Thumb-ish mode ? 2011-04-19T12:22:19 Yes 2011-04-19T12:22:55 is there a compiler? gcc doesn't support it AFAIK 2011-04-19T12:22:58 the second core on MPC5516 only runs in LTE 2011-04-19T12:23:15 i think the missing gcc support is the main hurdle 2011-04-19T12:23:51 but for me there is no hurry, just don't need the performance now 2011-04-19T12:24:22 and i might also just programm some assembler routines for it if there is a need 2011-04-19T12:24:47 thats how it is designed, as a bit of superintelligent dma enginge for example 2011-04-19T12:25:06 and there is hardware semaphores for syncing it 2011-04-19T12:25:49 and you can have interrupts in the other cpu 2011-04-19T12:26:16 it just would be a nice test case for SMP if it would run on it. 2011-04-19T12:26:34 but SMP assumes homogeneous instruction sets :( 2011-04-19T12:26:58 assumes you can move threads between cores .. 2011-04-19T12:27:01 yeal well thats the "S" in it 2011-04-19T12:27:17 you can run the main core both in LTE and normal 2011-04-19T12:27:27 this is probably a good case for the current MP support in RTEMS 2011-04-19T12:27:46 LTE on both ==> SMP ;) 2011-04-19T12:27:48 MP? 2011-04-19T12:28:03 yes that would be one way. 2011-04-19T12:28:24 I never heard of MP support in rtems= 2011-04-19T12:28:27 RTEMS has multiprocessing now.. processors run own copies of RTEMS and communicate over say shared memory 2011-04-19T12:28:41 would this be a own node on the scond cpu? 2011-04-19T12:28:55 yes a 2 node system.. 2011-04-19T12:29:00 ah yes that would be possible 2011-04-19T12:29:08 even in mixed mode 2011-04-19T12:29:45 but LTE support would be nice in its own way ... practically twice the flash size in the cpu 2011-04-19T12:29:46 yep.. the MP support now is effectively a transparent RPC on objects that are GLOBAL and remote 2011-04-19T12:29:54 and I think it even runs faster 2011-04-19T12:30:13 e200 doesnt have a cache 2011-04-19T12:30:36 but e.g. mpc5516 has one cycle word fetch from flash 2011-04-19T12:30:50 that would be 2 instructions in the pipeline in case of LTE 2011-04-19T12:30:57 and with smaller instructions, it could hold a larger program 2011-04-19T12:31:10 yeah thats what I meant with "twice size" 2011-04-19T12:31:31 don't know why gcc project isnt interested in LTE 2011-04-19T12:31:34 on powerpc 2011-04-19T12:32:15 not a case of not interested in general.. just no one has had enough interest to implement it (or pay for it) 2011-04-19T12:32:45 yeah thats what I meant with "interested" like practially interested 2011-04-19T12:33:18 gcc backend is also a bit messy and no fun to work for 2011-04-19T12:34:31 how about the LLVM backend? maybe it already has LTE? 2011-04-19T12:38:17 I have no idea on that one. 2011-04-19T12:38:48 did somebody use a llvm integrated gcc with rtems yet? 2011-04-19T12:39:05 I haven't heard of it 2011-04-19T12:51:08 might work, probably more easily than clang 2011-04-19T13:17:32 DrJoel: llvm doesnt have LTE support either 2011-04-19T13:17:45 just asked in their irc channel 2011-04-19T13:19:44 ahhh.. then no free compiler.. is there a closed source one? 2011-04-19T13:21:55 probably 2011-04-19T13:22:02 something like Diab 2011-04-19T13:24:14 *** methril has quit IRC 2011-04-19T13:24:34 Is Diab still independent or have they been bought over the years? 2011-04-19T13:25:37 I'm not sure to whom they belong ... just a compile I at some time heard mentioned at a car manufacturer 2011-04-19T13:26:04 They are WindRiver now 2011-04-19T13:26:43 LOL nobody is solo anymore 2011-04-19T13:28:34 CodeWarrior 2011-04-19T13:29:07 == Freescale itself ... :-/ 2011-04-19T13:29:42 Codesourcery.. Mentor 2011-04-19T13:31:27 BTW LTE was the wrong acronym its calle VLE!! 2011-04-19T13:32:43 VLE = variable length encoding 2011-04-19T13:35:37 I knew that didn't sound right but couldn't rememebr 2011-04-19T13:38:51 Here is somebody planning to do VLE http://www.mail-archive.com/gcc@gcc.gnu.org/msg48210.html 2011-04-19T13:40:38 *** methril has joined #rtems 2011-04-19T13:41:27 and not much since then. :( 2011-04-19T13:42:22 I sent him email.. maybe we will find out if something happened 2011-04-19T13:44:35 *** arvind_khadri has quit IRC 2011-04-19T14:17:43 *** dr__house has joined #rtems 2011-04-19T14:17:44 *** dr__house has joined #rtems 2011-04-19T14:22:42 *** gedare has joined #rtems 2011-04-19T15:30:11 *** gedare has quit IRC 2011-04-19T16:12:48 * DrJoel waves .. 2011-04-19T16:20:22 * shayden waves back 2011-04-19T16:30:53 how are things going for you? 2011-04-19T17:18:33 *** dr__house has quit IRC 2011-04-19T18:41:59 *** projectgus has joined #rtems 2011-04-19T19:16:09 hi DrJoel 2011-04-19T19:16:35 DrJoel, recently i discover OSC, i'll give a try before use the example you gave me 2011-04-19T19:17:21 i heard about it time ago, but tought it was just for sound specif stuff, but seems reusable for telemtry.. well i need confirm that :-) 2011-04-19T20:04:58 *** xiangfu has joined #rtems 2011-04-19T20:09:36 *** xiangfu_ has joined #rtems 2011-04-19T20:09:50 *** xiangfu has quit IRC 2011-04-19T20:11:22 *** xiangfu_ is now known as xiangfu 2011-04-19T20:34:49 *** xiangfu has quit IRC 2011-04-19T20:35:16 *** xiangfu has joined #rtems 2011-04-19T20:51:14 *** xiangfu has quit IRC 2011-04-19T20:52:12 *** xiangfu has joined #rtems 2011-04-19T23:17:36 *** arvind_khadri has joined #rtems 2011-04-20T00:36:07 *** cloud has joined #rtems 2011-04-20T01:19:20 *** kiwichris has joined #rtems 2011-04-20T01:23:20 *** arvind_khadri has quit IRC 2011-04-20T02:15:54 *** projectgus has quit IRC 2011-04-20T02:29:20 *** arvind_khadri has joined #rtems 2011-04-20T02:29:30 *** arvind_khadri has joined #rtems 2011-04-20T03:00:40 *** xiangfu has quit IRC 2011-04-20T03:07:26 *** xiangfu has joined #rtems 2011-04-20T03:11:34 *** Hyper-Core has joined #rtems 2011-04-20T03:13:17 *** xiangfu has quit IRC 2011-04-20T03:18:04 *** dr__house has joined #rtems 2011-04-20T03:18:15 *** Hyper-Core has left #rtems 2011-04-20T03:27:14 *** xiangfu has joined #rtems 2011-04-20T03:47:13 *** xiangfu has quit IRC 2011-04-20T03:55:16 *** kiwichris has quit IRC 2011-04-20T04:02:56 hello 2011-04-20T04:02:58 anybody? 2011-04-20T04:33:27 *** cloud has quit IRC 2011-04-20T04:53:09 *** sqm has joined #rtems 2011-04-20T05:01:25 *** arvind_khadri has quit IRC 2011-04-20T05:14:28 *** arvind_khadri has joined #rtems 2011-04-20T05:56:23 *** dr__house has quit IRC 2011-04-20T06:23:05 *** lj8175 has joined #rtems 2011-04-20T07:53:42 *** xiangfu has joined #rtems 2011-04-20T08:01:19 * DrJoel apparently didn't log off when I left yesterday... 2011-04-20T08:22:23 *** petben has joined #rtems 2011-04-20T08:27:13 DrJoel: just read your blog pos about PDF's 2011-04-20T08:27:51 totally forgot how bothersome PDF's are outside the mac 2011-04-20T08:35:22 This is really only a problem because there are so many. And there is the job of automating the PowerPoint -> PDF . The merging PDFs part was to make it simpler to print without screwing up. Thomas can now drop it to a copy shop. I have done that also but for smaller classes, I just use our printer 2011-04-20T08:35:50 *** antgreen has joined #rtems 2011-04-20T08:38:10 *** zwj has joined #rtems 2011-04-20T08:39:08 Yeah but on a mac there wouldn't be any problem at all, you'd just dragging the whole bunch of PDF's into preiview and hit save 2011-04-20T08:39:42 And I got so used to OS builtin PDF stuff that I really didn't remember how it had been 2011-04-20T08:53:44 wow .. that's nice. I might be able to do it with PDFCreator but since this gets done repeatedly, I do like the trustworthiness and hands off approach of a script 2011-04-20T09:00:53 *** arvind_khadri has quit IRC 2011-04-20T09:02:41 *** antgreen has quit IRC 2011-04-20T09:04:50 I usually like scripts for this too, but have been quite spoilt by some of the Mac stuff since I switched to a mac .. as X11 server with a nice display first but drawn in more and more 2011-04-20T09:05:24 *** arvind_khadri has joined #rtems 2011-04-20T09:05:52 nowadays I still develop via X11 xterm/emacs "IDE" on FreeBSD boxes via the mac but use all the daily doc/mail/PDF andwhatnot stuff directly on the mac 2011-04-20T09:06:41 and if I use the mac via bash ... its just another BSD for me 2011-04-20T09:07:20 *** antgreen has joined #rtems 2011-04-20T09:07:52 the only thing I don't like at the moment: my development system is a Intel Atom board ... no joy building RTEMS or FreeBSD world on it. 2011-04-20T09:08:52 But its the board that also is built into the boxes I sell so I get a quick testing cycle (except if it includes rebuilding of RTEMS of course ;-) 2011-04-20T09:09:08 *** antgreen has quit IRC 2011-04-20T09:09:27 BTW: is it stable to build RTEMS with a parallel make? 2011-04-20T09:10:25 *** arvind_khadri has quit IRC 2011-04-20T09:17:18 Atom is slow. My laptop is a dual core CPU that is a couple of years old but still zippy. 2011-04-20T09:17:49 It should be OK to build with parallel make. My scripts check out the number of CPUs and do something like -j with 1.5 that. 2011-04-20T09:19:09 I am very interested in ccdist with RTEMS. The build farm here is currently 2 quad core machines, a couple of dual core ones, and a couple of single core P4s. We will probably get a 6-core AMD machine in the next couple of months. Using distcc would be awesome for tool builds and testing 2011-04-20T09:19:09 My current tool test switch is like 3 days on a quad-core machine 2011-04-20T09:22:03 *** gedare has joined #rtems 2011-04-20T09:27:53 we have a student conflict gedare ... FFmpeg and RTEMS both look to be happy with the same student 2011-04-20T09:29:33 DrJoel: pm 2011-04-20T09:55:18 distcc sounds very interesting, it has to be combined with make -j to help? 2011-04-20T09:56:39 "distcc is designed to be used with the -j parallel-build feature in GNU Make or SCons, or other build tools." 2011-04-20T09:56:59 looks like it will relocate build jobs across a network 2011-04-20T09:57:31 yeah and without make -j this wouldn't help 2011-04-20T09:57:36 you'd need the -j option so that the make system knows it is being parallel, otherwise dependencies might not be tracked 2011-04-20T09:58:01 * gedare always uses make -j anyway 2011-04-20T09:58:21 except when i have to build things with archaic gcc:) 2011-04-20T09:58:32 no it will be tracked but it will wait until each distcc is finished before starting the next one, not speeding up anything 2011-04-20T09:58:41 right 2011-04-20T09:59:16 without parallelism distcc is just a extra slow cc 2011-04-20T10:02:16 and you don't need it for SMP/CMP machines, just "cluster"-like configurations 2011-04-20T10:07:06 gedare: right.. and in the RTEMS build farm, there are currently 2 quad core machines, a couple of dual core ones, and a couple of single core P4s. We will probably get a 6-core AMD machine in the next couple of months. 2011-04-20T10:07:18 So this is a good candidate for distcc.. 2011-04-20T10:09:21 *** antgreen has joined #rtems 2011-04-20T10:22:43 *** xiangfu has quit IRC 2011-04-20T10:25:34 distcc is good if you have a lot of machines and you use the latest version with it's include server 2011-04-20T10:26:00 previous versions would preprocess the file locally and send that file upstream, using the include server distributes the required headers as well and does remote preprocessing 2011-04-20T10:26:25 ccache can help even more than distcc by storing objects 2011-04-20T10:27:19 another (relativly) cheap speedup is getting an SSD to build on with machines that are 4 core and greater then running make -j 7 with ccache and a 1 line file change you can do full builds in seconds 2011-04-20T10:28:53 hi, a problem: on linux STACKBOTTOM is usually 0xc0000000 , what for RTEMS/I386 ? In the starting code, Init thread also use _Heap_Allocate to get memory for stack from workspace. 2011-04-20T10:30:52 lj8175: stack addresses vary depending on where heap is started. 2011-04-20T10:31:57 verm__: I have considered an SSD.. but am leaning to 16GB RAM in next machine and letting 4-8GB be RAM disk. The tool binary builds rely heavily on mock and having a local distro copy 2011-04-20T10:33:44 DrJoel: so if I want to get the address, is there a way(a methord) doing that ? 2011-04-20T10:35:37 verm__: is ccache only helping in the make clean; make scenario or is building RTEMS doing lots of redundant builds? 2011-04-20T10:35:47 lj8175: most boot-up code has to set the stack, so check the libbsp/ARCH/BSP/start/start.S file 2011-04-20T10:36:10 for i386, i guess you want libbsp/i386/pc386/start/start.S file... see how it sets up the stack 2011-04-20T10:36:40 gedare: Thank you. 2011-04-20T10:37:11 DrJoel: that won't help as much as you think it would, unless you plan on copying every single dependency to memory and chrooting it 2011-04-20T10:37:14 that will be the base of the stack, although I think thread stacks are pulled somewhere else... 2011-04-20T10:37:38 Whats "mock"? 2011-04-20T10:38:49 gedare: yes , I think thread stacks are pulled in workspace, and i can get the address through Configuration.work_space_start . :) 2011-04-20T10:39:09 DrJoel: mdisks aren't useful for two reasons: they tend to invalidate system cache quickly which is where most of the speedup happens during a build anyway, it's IOPS that's the most important and while you get a lot of that with memory it'll be far faster than your cpu can handle anyway 2011-04-20T10:39:11 lj8175: that sounds right to me. 2011-04-20T10:39:47 on a dual 6 core machine with 24 threads we use a 80gb ssd at NTP and the drive barely hits capacity due to the system cache, using an mdisk was no faster (24gb of memory in this machine) 2011-04-20T10:39:54 *** petben has quit IRC 2011-04-20T10:39:56 lj8175: some of the libbsp/shared/*.c files use linkcmds-provided variables that you might be able to also access to compute the stack address 2011-04-20T10:40:02 lj8175: workspace has more than stacks in it. There is no guarantee from app to app where the stack will be except in general terms 2011-04-20T10:40:53 peerst: ccache is useful if you're only updating a few files at a time as long as your source doesn't permutate you can get massive speedups. in a typical project each commit only touches a small percentage of files which leaves ccache to copy the objects over if required 2011-04-20T10:40:54 gedare: no.. stack is allocated at task creation time. It will be from the workspace (unless using stack allocation hooks) .. and will vary.. 2011-04-20T10:41:37 DrJoel: is the initially set-up stack only used for initialization? 2011-04-20T10:42:01 does ccache help if compiling the same source with different options? That's what we are doing.. same source over and over to different targets. 2011-04-20T10:42:12 gedare: yep 2011-04-20T10:42:25 oh, sorry lj8175, I seem to have misled you a bit. 2011-04-20T10:42:45 DrJoel: yes and no, ccache will store the objects on a per-flag and per-file basis 2011-04-20T10:43:00 if you have 5 files and 12 targets it'll store those 60 objects 2011-04-20T10:43:28 ccache shouldn't be necessary if the make system works, it shouldn't recompile stuff that didn't change 2011-04-20T10:43:38 so likely not to help us much at all. we would never recompile those 60 in a short time frame. 2011-04-20T10:43:51 peerst: ccache helps after a make clean; make distclean or new builds 2011-04-20T10:44:10 ok 2011-04-20T10:44:26 one of my projects takes about 5 mins to compile from scratch, if i do a make distclean and use ccache a new build takes about 4 seconds 2011-04-20T10:44:30 Thank you DrJoel and gedare :) 2011-04-20T10:45:00 DrJoel: ccache is good for repetative building, which is what you're doing isn't it? 2011-04-20T10:45:11 rebuild after every commit? unless your commits touch every file it'll help quite a lot 2011-04-20T10:45:25 they build with a new configure each time I would guess. 2011-04-20T10:45:53 that's fine ccache checks sthe file + includes + flags to see if it it needs to rebuild 2011-04-20T10:46:02 it stores the objects outside of the build tree 2011-04-20T10:47:01 verm__: I am trying to speed up builds of the binutils,gcc/newlib, gdb toolchain binaries right now. so it is basically a straight through build of each one, install, run tests. When something updates, do it again. 2011-04-20T10:47:40 it is really preferred to build from scratch each time.. so lots of RAM for disk caching might help.. moving source to RAM disk might help 2011-04-20T10:50:09 that's the exact situation ccache was designed for... 2011-04-20T10:50:45 try a cursory test CC="ccache cc" 2011-04-20T10:50:50 *** qmor has joined #rtems 2011-04-20T10:50:58 on all your builds, build once then try a second build 2011-04-20T10:50:59 ok.. will do on next build 2011-04-20T10:51:22 where does it keep .o's? 2011-04-20T10:51:24 if this gets you a good speedup you can integreate it better there are some tricks you can do make it even faster and more useful 2011-04-20T10:51:30 by default it will go into ~/.ccache 2011-04-20T10:51:59 hello everyone 2011-04-20T10:52:03 you can modify this with CCACHE_DIR 2011-04-20T10:52:35 can anybody help me with code what initiliasing nvram-disk? 2011-04-20T10:53:16 DrJoel: when you install ccache do a ccache -M 4GB just in case i think the default may be too small for your usage 2011-04-20T10:56:21 size of my last set of build trees was 56GB.. so it would need to be near that to keep anything around from one cycle to the next 2011-04-20T10:57:01 yeah, then 100GB is around where you would want it 2011-04-20T10:58:22 DrJoel: do you always build each target? 2011-04-20T11:01:54 qmor: What do you mean / what target? 2011-04-20T11:03:35 *** sqm has quit IRC 2011-04-20T11:03:40 *** lj8175 has quit IRC 2011-04-20T11:04:25 i have the crash after sc = rtems_io_register_driver (0,(rtems_driver_address_table *) &nv_io_ops, &major); this comand 2011-04-20T11:05:20 i want to use the part of eeprom as a disc for dosfs 2011-04-20T11:05:26 qmor: what version of rtems, what BSP, and how did you configure it, and what are your CONFIGURE_ options? 2011-04-20T11:08:22 rtems 4.10 2011-04-20T11:08:48 gedare, maybe you could see all of code? 2011-04-20T11:09:08 is it not many 2011-04-20T11:09:10 i doubt I can help that much... I don't know much about the filesystems 2011-04-20T11:09:46 does rtems_io_register_driver return successfully? 2011-04-20T11:09:52 or is the crash while you register 2011-04-20T11:11:39 http://pastebin.com/dGx1iuEa 2011-04-20T11:12:11 it's crashed when i calling it 2011-04-20T11:14:24 *** zwj has quit IRC 2011-04-20T11:16:31 qmor: i think you're missing some initialize hook 2011-04-20T11:17:04 nv_iops is not of type rtems_driver_address_table and you forced it. That's why it blew up. 2011-04-20T11:17:23 well, there you go :) 2011-04-20T11:17:46 better to ask for this type of help on the mailing list. More people can answer. I don't see anyone here who has done much with non-volatile file systems. I am working on some examples now and have a pending request for help myself 2011-04-20T11:18:07 i've tryed without forced 2011-04-20T11:18:15 okay 2011-04-20T11:18:28 i'm try to send it to maillist 2011-04-20T11:19:20 but maybe/// 2011-04-20T11:20:03 sorry i say about this forced typecast "rtems_io_register_driver (0,(rtems_driver_address_table *) &nv_io_ops, &major)" 2011-04-20T11:22:03 what i should use instead of nv_io_ops? 2011-04-20T11:22:23 there are no good examples 2011-04-20T11:22:34 this code is the only part of some example 2011-04-20T11:25:33 qmor: people are working on examples, send your question to the mailing list 2011-04-20T11:29:23 jkay 2011-04-20T11:29:29 thank you 2011-04-20T11:42:47 *** qmor has quit IRC 2011-04-20T12:00:12 *** methril_work has quit IRC 2011-04-20T12:07:03 *** arvind_khadri has joined #rtems 2011-04-20T12:10:04 *** gedare has quit IRC 2011-04-20T12:31:53 *** gedare has joined #rtems 2011-04-20T13:12:31 *** gedare has quit IRC 2011-04-20T13:24:40 *** antgreen has quit IRC 2011-04-20T13:32:35 *** kiwichris has joined #rtems 2011-04-20T13:40:47 *** arvind_khadri has quit IRC 2011-04-20T13:45:52 *** methril_work has joined #rtems 2011-04-20T13:51:14 Hi Joel 2011-04-20T13:51:28 hey chris .. 2011-04-20T13:51:37 Is my fileio patch ok for 4.10 or should I just put it in 4.11 ? 2011-04-20T13:52:24 I can't find the email quickly but I think it is Ok for 4.10. It is an example, not a critical test 2011-04-20T13:52:48 Great. It adds making a RAM disk as a shell command. 2011-04-20T13:52:59 With the filesystem test we cannot have shell support ? 2011-04-20T13:54:00 I would prefer not to but if rfs_support or fat_support need to run a series of commands, that would be OK. If the tests turn out to be too large, then it is just the test support for those file systems 2011-04-20T13:54:35 let's avoid usign the tar trick for initial contents if possible.. malloc RAM, mk a ram disk, format it, and then mount it 2011-04-20T13:55:52 Ok. I took a look at the freebsd file system tests and they are mostly shell scripts. I know we cannot run /bin/sh scripts but this was the basis for the idea. 2011-04-20T13:56:07 Our shell would need logic 2011-04-20T13:57:00 I will take a look over the long weekend. Today will be busy. 2011-04-20T13:57:17 right.. these are to test the section 2 and 3 POSIX APIs. So tests themselves are in C. I suppose the XXX_support/ could invoke commands 2011-04-20T13:57:20 extern void test_initialize_filesystem(void); 2011-04-20T13:57:20 extern void test_shutdown_filesystem(void); 2011-04-20T13:58:07 Those are the two methods an instance of XXX_support has to provide right now. They are in a file by themselves. So could do anything if it can make sure tests know how to configure extra resources 2011-04-20T13:58:23 If it is in C there is not need for the shell. The fileio addition is nice because I can now run tests from users for various file system issues. 2011-04-20T13:59:22 I will take a look and get back to you. 2011-04-20T14:00:19 Thanks.. this is the foundation for transitioning the psx file tests from IMFS only to all file systems. 2011-04-20T14:00:45 I don't see anything wrong with having more tests in this suite when we get through the existing C tests that are shell driven. 2011-04-20T14:00:57 But the goal is automated coverage testing on all platforms 2011-04-20T14:01:59 On a subject related to testing. I would like to chat to Ralf about a way we can remove the small BSP meory hack I added. 2011-04-20T14:02:14 It was a temporary thing and well pass its used by date. 2011-04-20T14:02:42 we need a per-BSP list of tests which do not fit into the memory of the BSP. 2011-04-20T14:03:20 That is one way. 2011-04-20T14:03:55 Does this mean a change requires checking all BSPs ? 2011-04-20T14:04:42 yeah.. I suppose it would to be proper but your autotester would be sufficient 2011-04-20T14:04:45 is it back online? 2011-04-20T14:05:13 No. I am still looking into. I have just ticked a timed build. 2011-04-20T14:06:25 The cause of the problem is the linkcmd files that declare regions. I understand the purpose but could this be done some other way ? 2011-04-20T14:06:43 ok.. but how else would you know if you pushed one byte too many in it 2011-04-20T14:07:08 size + script + BSP data ? 2011-04-20T14:07:39 How do you know the size until you compile it and link it? 2011-04-20T14:08:00 If the linker script has any idea how much RAM there is, then you will overflow it no matter what 2011-04-20T14:08:07 You do not. This is issue. 2011-04-20T14:08:50 The issue is a change in a test allows it to run. The auto-builder only removes tests 2011-04-20T14:08:53 *** antgreen has joined #rtems 2011-04-20T14:09:31 ok I don't see it but trust you 2011-04-20T14:10:16 The auto builder reports a problem because a test does not fit. It is removed from the build for a BSP. The test is changed and now it does fit but we never know. 2011-04-20T14:11:15 ahhh.. I see. could have option to rebuild list.. where do the lists live? 2011-04-20T14:11:52 As a sort of BSP maintenance thing ? 2011-04-20T14:12:25 yes.. rebuild as part of a maintenance. 2011-04-20T14:12:27 I do not know. I would like to ask Ralf to comment and get his feed back because this is all autotool related. 2011-04-20T14:13:35 yep.. but he has to accept that some tests do not fit on some targets.. and we want to build (and run) all we can 2011-04-20T14:14:37 Sure. 2011-04-20T14:15:46 brb 2011-04-20T14:15:47 If we have a means to remove tests from a BSP then a configure option to disable this we could build all tests. Then we use make -k to build 2011-04-20T14:23:52 and use the failures to rebuild the list? 2011-04-20T14:24:27 Yes if the error is a linker memory overflow. 2011-04-20T14:24:40 On the fstester. I have taken a look. 2011-04-20T14:24:46 just typing that you have detect the error type 2011-04-20T14:25:44 The complication is the test is not normally built and may have rotted in other ways. 2011-04-20T14:26:39 WIth the file system tester, are we only to consider IMFS, MIMFS and libblock file systems ? 2011-04-20T14:29:28 At this point, I think so. NFS and FTPFS would require somethign local and smart to talk to. :-D 2011-04-20T14:30:11 Agreed. This means we can break down the driver and file system into different parts. 2011-04-20T14:30:42 The libblock cache and driver can be tested by direct writes to the /dev/ entry 2011-04-20T14:31:00 This is not the focus of the task but the tester should handle this. 2011-04-20T14:31:16 ahhh.. that would be specific tests beyond this initial set of section 2 and 3 tests, right? 2011-04-20T14:32:02 Yeah and very important ones because you can control how the cache handles calls. With a file system in the way you cannot do this without a great deal of magic and luck 2011-04-20T14:32:38 By this I mean you could get something that has the cache do something but a change and this all changes. 2011-04-20T14:33:04 got it. 2011-04-20T14:33:10 We need to have a way of configuring and managing the cache set up. 2011-04-20T14:33:15 Does the goal of fstest sound right? 2011-04-20T14:34:06 Yes but we need a few more parts than just file system init. We also need a series of functions that the test can use to perform specific operations. 2011-04-20T14:34:14 I was thinking of some type of adaptive scheme that tried to malloc memory in known sizes.. free'ing and working down in size on failure. So on a PC, it might test GB's but on SIS would only be MBs 2011-04-20T14:34:19 For example a test on a file system is to call unmount while in use. 2011-04-20T14:34:54 This does not control the test. I would go for the SIS size and build all tests around tha. 2011-04-20T14:34:57 yes.. just focusing on the basics now.. i knew I was leaving stuff out.. just didn't think of it. 2011-04-20T14:35:21 assume no more than 1-4 MB available for the tests to use? 2011-04-20T14:35:25 If the test works on the SIS with a small amount of memory the same test will do the same thing on a PC 2011-04-20T14:36:01 My fileio has a workable size for SIS. That is what I used to fix the IMFS bug eval on 4.10 2011-04-20T14:36:46 On the cache. We have read ahead sizes, write sizes, a swap out timer and more than one swap out thread. 2011-04-20T14:37:04 ok.. that will set a reasonable baseline 2011-04-20T14:37:14 This means we need more than RAM disk active. 2011-04-20T14:37:31 Or the ability of a test to create more than one. 2011-04-20T14:37:59 I will think about this a little and put it in an email. 2011-04-20T14:38:25 Work ftp area has a buf tester for the cache. 2011-04-20T14:38:35 I cannot remember what it did. 2011-04-20T14:39:30 are we getting old? 2011-04-20T14:39:42 daily ;) 2011-04-20T15:28:04 *** gedare has joined #rtems 2011-04-20T16:10:33 *** antgreen has quit IRC 2011-04-20T16:31:26 *** methril_work has quit IRC 2011-04-20T16:50:35 night all 2011-04-20T16:50:38 *** DrJoel has quit IRC 2011-04-20T16:55:14 *** gedare has quit IRC 2011-04-20T17:05:39 *** DrJoel has joined #rtems 2011-04-20T17:05:39 *** ChanServ sets mode: +o DrJoel 2011-04-20T17:35:20 *** DrJoel has quit IRC 2011-04-20T18:48:21 *** projectgus has joined #rtems 2011-04-20T18:49:59 *** methril_ has joined #rtems 2011-04-20T18:50:49 *** methril has quit IRC 2011-04-20T19:12:35 *** methril__ has joined #rtems 2011-04-20T19:13:15 *** methril_ has quit IRC 2011-04-20T19:35:43 *** methril__ has quit IRC 2011-04-20T19:47:38 *** methril__ has joined #rtems 2011-04-20T20:19:33 *** xiangfu has joined #rtems 2011-04-20T20:42:34 *** xiangfu has quit IRC 2011-04-20T20:50:54 *** xiangfu has joined #rtems 2011-04-20T23:01:32 *** arvind_khadri has joined #rtems 2011-04-20T23:01:32 *** arvind_khadri has joined #rtems 2011-04-20T23:41:44 *** kiwichris has quit IRC 2011-04-21T00:36:32 *** arvind_khadri has quit IRC 2011-04-21T00:55:20 *** dr__house has joined #rtems 2011-04-21T00:55:20 *** dr__house has joined #rtems 2011-04-21T01:01:55 *** projectgus has quit IRC 2011-04-21T01:15:29 *** dr__house has quit IRC 2011-04-21T01:15:46 *** peerst has quit IRC 2011-04-21T01:15:47 *** peerst_ has joined #rtems 2011-04-21T02:06:59 *** monstr2 has joined #rtems 2011-04-21T02:30:41 *** xiangfu has quit IRC 2011-04-21T02:56:26 *** Fallenou has quit IRC 2011-04-21T03:40:26 *** arvind_khadri has joined #rtems 2011-04-21T03:54:08 *** authors has joined #rtems 2011-04-21T03:58:43 Good morning everyone. I am completely new to rtems and Cexp and i wonder if someone can help me learning how to use Cexp. More accurately how to build a symbol table from the object. 2011-04-21T03:59:23 s/object/executable/ 2011-04-21T04:42:58 *** arvind_khadri has quit IRC 2011-04-21T04:43:24 *** authors has quit IRC 2011-04-21T04:55:54 *** arvind_khadri has joined #rtems 2011-04-21T05:18:55 *** Fallenou has joined #rtems 2011-04-21T06:35:57 *** dimonov has quit IRC 2011-04-21T06:44:45 *** lj8175 has joined #rtems 2011-04-21T06:46:09 *** dimonov has joined #rtems 2011-04-21T07:47:40 *** arvind_khadri has quit IRC 2011-04-21T08:32:42 *** monstr2 has quit IRC 2011-04-21T08:32:58 *** DrJoel has joined #rtems 2011-04-21T08:32:59 *** ChanServ sets mode: +o DrJoel 2011-04-21T08:34:27 Someone here familiar with Spartan-3 & Xilinx ISE / PlanAhead & Partitions? 2011-04-21T08:39:11 *** zwj has joined #rtems 2011-04-21T08:51:15 *** methril__ has quit IRC 2011-04-21T09:11:20 *** xiangfu has joined #rtems 2011-04-21T09:18:33 *** dr__house has joined #rtems 2011-04-21T09:37:38 *** gedare has joined #rtems 2011-04-21T09:59:44 drjoel: http://www.contrib.andrew.cmu.edu/~shinpei/papers/techrep09.pdf 2011-04-21T09:59:59 i just ran across this paper, it talks about pluggable schedulers in linux, kind of interesting 2011-04-21T10:00:14 some of the API / design decisions might make sense as rtems adds Smp 2011-04-21T10:16:57 looks like a good read.. (when I don't have a headache) :-D 2011-04-21T10:17:34 aww 2011-04-21T10:18:02 In our working tree, my simple SMP scheduler is merged and is scheduling. Have RTEMS make dispatching requests on secondary cores and OS locking are not complete 2011-04-21T10:18:08 So it is getting closer and closer 2011-04-21T10:18:56 yeah 2011-04-21T10:59:55 *** lj8175 has quit IRC 2011-04-21T11:03:11 *** xiangfu has quit IRC 2011-04-21T11:23:21 *** zwj has quit IRC 2011-04-21T11:55:24 *** arvind_khadri has joined #rtems 2011-04-21T11:55:24 *** arvind_khadri has joined #rtems 2011-04-21T12:24:21 *** qmor has joined #rtems 2011-04-21T12:29:39 *** arvind_khadri has quit IRC 2011-04-21T12:30:02 *** arvind_khadri has joined #rtems 2011-04-21T12:30:03 *** arvind_khadri has joined #rtems 2011-04-21T12:39:16 hello 2011-04-21T12:39:32 yesterday i'v send the question to maillist 2011-04-21T12:39:48 but not have any approve answer 2011-04-21T12:40:07 is it normal to wait some days? 2011-04-21T12:43:37 did you subscribe to the list before you send it? You have to be subscribed for messages to get distributed.. spam control 2011-04-21T12:44:16 ohh 2011-04-21T12:44:20 no i have not 2011-04-21T12:52:09 do that and resend 2011-04-21T12:57:51 *** panzon has joined #rtems 2011-04-21T12:58:42 I've done 2011-04-21T12:59:29 but i'h receive an email Your mail to 'rtems-users' with the subject 2011-04-21T12:59:31 nvdisk initialisation trouble 2011-04-21T12:59:33 Is being held until the list moderator can review it for approval. 2011-04-21T12:59:35 The reason it is being held: 2011-04-21T12:59:37 Post by non-member to a members-only list 2011-04-21T12:59:49 right.. did you subscribe from the address you sent it? 2011-04-21T13:00:03 http://www.rtems.org/mailman/listinfo/rtems-users 2011-04-21T13:00:19 absolutely yes 2011-04-21T13:03:07 now i'm on the page that called rtems-users mailing list membership configuration for qmor.qmor at gmail.com, qmor 2011-04-21T13:03:30 i think it mean what i' done registration 2011-04-21T13:04:17 then you resend 2011-04-21T13:11:03 i'v did it 2011-04-21T13:11:20 but receive an email with text below 2011-04-21T13:24:04 *** monstr2 has joined #rtems 2011-04-21T13:24:35 qmor: emailing you an address to check with privately 2011-04-21T13:25:11 *** qmor has quit IRC 2011-04-21T13:33:16 *** qmor has joined #rtems 2011-04-21T13:46:17 *** qmor has quit IRC 2011-04-21T13:46:21 *** panzon has quit IRC 2011-04-21T14:14:29 *** methril has joined #rtems 2011-04-21T14:21:47 *** monstr2 has quit IRC 2011-04-21T14:36:40 * DrJoel waves bye.. calling it a day early 2011-04-21T14:36:43 *** DrJoel has quit IRC 2011-04-21T15:09:03 *** arvind_khadri has quit IRC 2011-04-21T15:09:36 *** arvind_khadri has joined #rtems 2011-04-21T15:34:58 *** dr__house has quit IRC 2011-04-21T16:11:50 *** arvind_khadri has quit IRC 2011-04-21T16:22:06 *** arvind_khadri has joined #rtems 2011-04-21T18:12:28 *** gedare has quit IRC 2011-04-21T18:45:17 *** arvind_khadri has quit IRC 2011-04-21T19:00:53 *** arvind_khadri has joined #rtems 2011-04-21T19:41:02 *** xiangfu has joined #rtems 2011-04-21T21:26:47 *** arvind_khadri has quit IRC 2011-04-21T21:39:42 *** arvind_khadri has joined #rtems 2011-04-21T21:47:17 *** arvind_khadri has quit IRC 2011-04-22T00:14:03 *** monstr2 has joined #rtems 2011-04-22T00:16:01 *** arvind_khadri has joined #rtems 2011-04-22T00:16:02 *** arvind_khadri has joined #rtems 2011-04-22T00:22:06 *** arvind_khadri has quit IRC 2011-04-22T00:44:29 *** lc0474696 has joined #rtems 2011-04-22T00:44:45 hello everyone 2011-04-22T00:45:08 i have a trouble with mailing list 2011-04-22T00:45:44 *** arvind_khadri has joined #rtems 2011-04-22T00:46:20 i'v registered and receive a messages from this list 2011-04-22T00:48:27 but i can's send a letter 2011-04-22T01:48:57 *** lc0474696 has quit IRC 2011-04-22T04:55:43 *** xiangfu has quit IRC 2011-04-22T05:47:06 *** panzon has joined #rtems 2011-04-22T06:25:12 *** petben has joined #rtems 2011-04-22T06:54:29 *** arvind_khadri has quit IRC 2011-04-22T07:00:12 *** arvind_khadri has joined #rtems 2011-04-22T07:08:46 *** xiangfu has joined #rtems 2011-04-22T07:14:23 *** arvind_khadri has quit IRC 2011-04-22T07:29:58 *** arvind_khadri has joined #rtems 2011-04-22T07:33:47 *** DrJoel has joined #rtems 2011-04-22T07:33:48 *** ChanServ sets mode: +o DrJoel 2011-04-22T07:34:19 peerst_: Got a few minutes to answer a question? 2011-04-22T07:40:39 *** gedare has joined #rtems 2011-04-22T08:18:28 *** arvind_khadri has quit IRC 2011-04-22T08:24:26 hi 2011-04-22T08:28:40 drjoel: are you still here? 2011-04-22T08:28:49 yep.. 2011-04-22T08:28:54 ok shoot 2011-04-22T08:29:26 our cell phones are unlocked.. in the past we have just gotten t-mobile prepaid but this time my wife and I would like to have Internet access on the phones. 2011-04-22T08:29:45 I think I see that O2 and Tchibo have EU10 per month flat rate pre-paid SIMs. 2011-04-22T08:30:39 http://www.tchibo.de/Mobiles-Internet-bei-Tchibo-mobil-c400000221.html?cs=1 2011-04-22T08:31:06 you know what tschibo is? 2011-04-22T08:31:11 If I am reading that right, you get 500MB then it throttles 2011-04-22T08:31:21 the coffee shop.. but this is mobile 2011-04-22T08:31:22 *** arvind_khadri has joined #rtems 2011-04-22T08:31:31 its the same company 2011-04-22T08:31:42 the sell coffee and everything else 2011-04-22T08:32:14 :D 2011-04-22T08:32:41 don't care what they sell as long as the mobile works OK and is not ghastly expensive.. T-Mobile is 5EU per day 2011-04-22T08:33:07 t-mobile has no attractive prepaid data as far as i know 2011-04-22T08:33:20 on that plan, how much are SMS and calls? 2011-04-22T08:33:58 the 9,95 per month is only for data 2011-04-22T08:33:59 that's what I figured out from the website .. cell phone stuff is hard enough to read but when translating the descriptions, it is worse 2011-04-22T08:34:42 I thought so.. calls more is OK .. looking at 2 SIMs with data/voice/SMS and 1-3 with voice/SMS (3 kids coming this time) 2011-04-22T08:34:54 so you probably need a normal tschibo phone tarif extra 2011-04-22T08:35:00 http://www.tchibo.de/Mobiles-Tchibofonieren-bei-Tchibo-mobil-c400000270.html?cs=2 2011-04-22T08:35:56 from tschibo to tschibo 5ct/min all other german numbers no matter if mobile or not 15ct /min 2011-04-22T08:36:00 sms 15ct 2011-04-22T08:36:15 and its billed 60/1 2011-04-22T08:36:25 60/1? 2011-04-22T08:36:43 *** arvind_khadri has quit IRC 2011-04-22T08:36:49 -> allways 1 minute will be charged per call after this its second precision 2011-04-22T08:37:08 aaahhh.. ok .. 2011-04-22T08:37:11 20sec call -> 1min, 1min30 -> 1min30 2011-04-22T08:37:56 they have a flat to all landlines in germany and between tchibo 12,95 2011-04-22T08:38:04 can that plan include the data Internet-Flatrate L? And am I reading that correctly that after 500MB you are throttled, not cut off 2011-04-22T08:38:36 WOW! That would be awesome for us! Then no real worries.. does that include SMS? 2011-04-22T08:42:15 no voice or video allowed over data 2011-04-22T08:42:58 yeah when the 500 MB are gone you will be throttled to max 64 kbit/s 2011-04-22T08:43:23 *** arvind_khadri has joined #rtems 2011-04-22T08:43:32 I checked that that's more than we have used on our unlimited plans here in a month 2011-04-22T08:43:48 if you use their day flat for intenet its 1 GB per da for 2,95? 2011-04-22T08:44:17 same basic limitation on our US plans.. no voip officially but we can watch YouTUbe.. and video chat (although we don't) 2011-04-22T08:44:29 we will be there 10 days so buying a month makes more sense 2011-04-22T08:44:32 hi peerst_ 2011-04-22T08:44:37 and its running on the O2 network 2011-04-22T08:44:41 hi gedare 2011-04-22T08:45:02 Cool! I checked and you can get Facebook notifications with SMS on O2 in germany.. don't tell my wife LOL 2011-04-22T08:45:14 we have unlimited texting on our plan here 2011-04-22T08:45:33 peerst_: I was wondering if you might know some sample rt apps that use priority queues, like message queues with priority or similar.. 2011-04-22T08:46:18 you mean the builting queues? 2011-04-22T08:46:20 since you were talking about queues the other day for transferring from USB to automotive stacks, I thought you might 2011-04-22T08:46:22 builtin 2011-04-22T08:46:35 right, but ones that use priority to sort the entries instead of e.g. FIFO 2011-04-22T08:46:51 network routers do this 2011-04-22T08:47:10 you can get texting packages for cheaper sms with the thschibo 2011-04-22T08:47:32 are there priority queues in rtems? 2011-04-22T08:47:34 peerst_: some time before we will get there can you email me links to what we probably should get? And a place near Marienplatz to buy them 2011-04-22T08:47:38 for POSIX.. yes 2011-04-22T08:47:41 posix queues have priorities 2011-04-22T08:47:44 don't like the normal messages queus much 2011-04-22T08:47:51 posix queues are better 2011-04-22T08:48:11 well do you know use cases that benefit from them? 2011-04-22T08:48:17 off-hand 2011-04-22T08:48:31 drjoel are you staying near marienplatz? 2011-04-22T08:49:03 hmm, i mean if you need a priority queue you need one 2011-04-22T08:49:41 yeah... I just want some reasonably compelling example, like sensor data processing maybe 2011-04-22T08:49:49 but i don't know some realistic uses 2011-04-22T08:50:11 one advantage is if you need a very simple sequential queue with urgent messages that jump the queue you can easily build it out of prio queues 2011-04-22T08:50:14 (other than routing in packet-switch / prioritized networks) 2011-04-22T08:50:18 just use 2 prios 2011-04-22T08:50:47 with sensor data you usually have fixed rate pipes 2011-04-22T08:51:10 dedicated pipes? 2011-04-22T08:51:16 (as opposed to shared) 2011-04-22T08:51:28 how you build it depends on the application 2011-04-22T08:52:06 so .. if i said an application uses multiple sensors and filters the data from all sensors through a priority queue before processing... 2011-04-22T08:52:09 most sensor stuff is built around syncronous designs 2011-04-22T08:52:11 that might make sense? 2011-04-22T08:52:16 no 2011-04-22T08:52:16 mm, I see 2011-04-22T08:52:30 priority queue only makes sense in a decoupled design 2011-04-22T08:52:34 cant think of a use for this 2011-04-22T08:52:47 inter-process communication is the only thing I found 2011-04-22T08:53:01 in a quick search 2011-04-22T08:53:07 as a "general" use case 2011-04-22T08:53:17 * gedare has a hammer in search of a nail 2011-04-22T08:53:45 priority queues are useful data structures in many algorithms 2011-04-22T08:53:47 alright, thanks for your thoughts/input 2011-04-22T08:53:49 yeah 2011-04-22T08:53:56 I have a dijkstra's algorithm implementation using one 2011-04-22T08:54:07 and will adapt an A* algorithm, basically the same thing 2011-04-22T08:54:13 maybe if you want to have such algorithm distributed between processes 2011-04-22T08:54:15 but I am looking for something inherently "real-time" 2011-04-22T08:54:28 hmmm, 2011-04-22T08:54:31 real-time dijkstra ;-) 2011-04-22T08:54:33 I see, that might work 2011-04-22T08:54:53 I have dijkstra implemented with periodic loops, so it has bounded behavior 2011-04-22T08:54:56 for what do you need an application ... 2011-04-22T08:55:00 of course, termination is not bounded 2011-04-22T08:55:05 my phd thesis:) 2011-04-22T08:55:10 ah ok 2011-04-22T08:55:22 just make stuff up to justify research 2011-04-22T08:55:29 :-D 2011-04-22T08:55:30 but I want it to be realistic made-up stuff 2011-04-22T08:55:51 are you working with rtems for your phd? 2011-04-22T08:55:54 yeah 2011-04-22T08:56:36 I presented a poster and short paper at a conference that gives a teaser about it... 2011-04-22T08:57:02 http://cse.unl.edu/rtss2008/archive/rtss2010/WIP2010/12.pdf 2011-04-22T08:57:08 whats the topic? 2011-04-22T08:57:22 ah ih see 2011-04-22T08:58:00 it turns out that scheduling isn't an interesting use case, because most rt systems have light load / small # of tasks 2011-04-22T08:58:02 I throw it on my ipad and read it some time 2011-04-22T08:58:10 :) 2011-04-22T08:58:43 but... maybe I can think on parallel dijkstra with a shared PQ 2011-04-22T08:58:54 or some elaborate IPC application 2011-04-22T08:59:10 btw maybe something interesting from the Erlang world related to this 2011-04-22T08:59:18 anyhow, thanks for your input, it was in fact helpful; now I have an argument not to look at sensors 2011-04-22T08:59:20 sure 2011-04-22T08:59:59 message passing in erlang does not use priorities 2011-04-22T09:00:14 its ordered if between the same processes 2011-04-22T09:00:29 but you can receive them out of order 2011-04-22T09:00:38 witch is quite powerful 2011-04-22T09:00:59 erlang has pattern matching (prolog like) built into the language 2011-04-22T09:01:21 and when i have a receive statement i have to specify the patterns i want to see 2011-04-22T09:01:29 peerst_: no actually at the Holiday Inn Schwabing .. so something at Muncher Freiheit would be even better. 2011-04-22T09:01:44 i see...so the message queue will do pattern matching 2011-04-22T09:01:46 there are tschibos everywhere 2011-04-22T09:02:00 yes when receiving 2011-04-22T09:02:08 and pushes in fifo 2011-04-22T09:02:16 by default 2011-04-22T09:02:31 allways 2011-04-22T09:02:44 okay 2011-04-22T09:02:47 but unmatched messages stay in the mailbox (input queue) 2011-04-22T09:03:00 and this is really powerful 2011-04-22T09:03:12 the sender can remove messages or not? 2011-04-22T09:03:17 because you can avoid fsm explosion 2011-04-22T09:03:30 no, once its sent its out of its control 2011-04-22T09:03:43 these are processes, no shared state at all 2011-04-22T09:03:46 fsm explosion... because you only build an fsm for the pattern specified? 2011-04-22T09:03:55 and then throw it away? 2011-04-22T09:04:34 often when you wnat to model real world stuff with a fsm it tends to grow exponentially if you add such real worldly this as error handling and stuff 2011-04-22T09:05:02 because every event can occur in every state and you have to handle it there 2011-04-22T09:05:27 * gedare nods 2011-04-22T09:05:29 with selective recieve i can ignore messages i can't handle right now 2011-04-22T09:06:01 and on occasion pick out these rare things when I'm able to do something about it 2011-04-22T09:06:35 which leads to very clean structured code for protocols and similar stuff 2011-04-22T09:06:47 this seems interesting, I would probably need a look to get a better clue on the power though 2011-04-22T09:07:05 * gedare jots in his research journal 2011-04-22T09:07:09 usually once in each server loop you have a catch all pattern to clean out the clutter from you mailbox 2011-04-22T09:07:36 i see.. pulling whatever didnt' match / you don't care about 2011-04-22T09:08:00 http://learnyousomeerlang.com/the-hitchhikers-guide-to-concurrency#dont-panic 2011-04-22T09:08:06 haha 2011-04-22T09:08:27 scroll down until you see some code that contains the keyword "receive" 2011-04-22T09:08:52 like the dolphin1 function 2011-04-22T09:09:13 or more realistic dolphin2 2011-04-22T09:11:35 for a low number of fixed priorities you can have a prio queue easily with selective receive 2011-04-22T09:11:54 yeah i found that 2011-04-22T09:12:06 just match the higher prios further up in a receive statement 2011-04-22T09:12:33 or further out in a nested receive 2011-04-22T09:12:38 or whatever 2011-04-22T09:12:39 yeah 2011-04-22T09:12:48 this gives me some other idea 2011-04-22T09:12:52 sooo maybe you can conoct a use case for prio queues 2011-04-22T09:12:58 out of this 2011-04-22T09:13:11 given the other advantage of selective recieve 2011-04-22T09:13:20 thanks; it is something like cascading case statements 2011-04-22T09:13:22 *** monstr2 has quit IRC 2011-04-22T09:13:26 the one i mentioned above 2011-04-22T09:13:30 yeah 2011-04-22T09:13:42 no there is definitely an interesting idea in there 2011-04-22T09:13:43 *** dr__house has joined #rtems 2011-04-22T09:13:43 *** dr__house has joined #rtems 2011-04-22T09:13:54 I wrote it down. I don't know if it is clean enough to pursue now though 2011-04-22T09:14:02 but definitely gave me some idea :) 2011-04-22T09:14:20 so having more clearly structure code by handling normal (high) prio normally and having (low) prio stuff only done when you can do something about 2011-04-22T09:14:38 :-) thats what i intended ... no be creative 2011-04-22T09:14:45 s/no/now/ 2011-04-22T09:15:04 opposite meaning 2011-04-22T09:15:07 hehe 2011-04-22T09:15:16 i sent you a pm to let you know where my mind went 2011-04-22T09:17:35 * gedare tries to get back on track 2011-04-22T09:30:04 peerst_: should this location be OK? http://nachbarschaft.immobilienscout24.de/adressen/bayern,muenchen,schwabing-west/essen/tchibo,329041.html 2011-04-22T09:30:38 It is a Tchibo which is very close to the u-bahn an muncher freiheit 2011-04-22T09:32:03 looks perfec 2011-04-22T09:32:47 If you can send me some email with what to ask for, I will print that out and we will be in business.. 2011-04-22T09:33:13 *** zwj has joined #rtems 2011-04-22T09:33:55 *** arvind_khadri has quit IRC 2011-04-22T09:36:38 *** arvind_khadri has joined #rtems 2011-04-22T09:37:08 ok 2011-04-22T09:38:01 *** dr__house has quit IRC 2011-04-22T09:38:08 *** dr__house has joined #rtems 2011-04-22T09:38:08 *** dr__house has joined #rtems 2011-04-22T09:40:01 *** dr__house has quit IRC 2011-04-22T09:40:07 *** dr__house has joined #rtems 2011-04-22T09:43:33 DrJoel: i think all you need to get a SIM card charged with enough Euros, the rest is done by sending sms to activate the stuff 2011-04-22T09:47:08 I don't think there are peoply to ask in the shop 2011-04-22T09:47:49 I mean its probably just prepacked precharged cards 2011-04-22T09:48:36 I found a phone shop near Viktualianmark (sp?) that looked possible. I figured Marienplatz area was more likely to be tourist friendly 2011-04-22T09:50:01 hi 2011-04-22T09:50:10 panzon: hi 2011-04-22T09:50:34 do you know where can I found the log 2011-04-22T09:50:56 ? 2011-04-22T09:51:06 http://www.rtems.org/irclogs/index.html 2011-04-22T09:51:18 There is a big O2 shop right on Marienplatz 2011-04-22T09:51:19 weekly logs i think 2011-04-22T09:51:22 but long lines 2011-04-22T09:51:43 :) 2011-04-22T09:52:00 no, I'm not talking about that logs 2011-04-22T09:52:24 what logs? 2011-04-22T09:52:26 then 2011-04-22T09:52:35 DrJoel: just emailed a friend that runs on Tschibo prepaid how it works in the Shop 2011-04-22T09:52:50 I'm running the tmtests 2011-04-22T09:52:54 thanks.. 2011-04-22T09:52:59 panzon: cool! which target? 2011-04-22T09:53:22 if you don't hear from be please bug me ... I'm forgetting lots of stuff because of the other lots of stuff 2011-04-22T09:53:31 sparc-sis 2011-04-22T09:54:17 oh.. if you're using the rtems-testing script, it creates a log subdirectory from wherever you run it 2011-04-22T09:54:37 * gedare keeps guessing 2011-04-22T09:58:04 mmm, maybe I'm doing something wrong 2011-04-22T09:58:23 Do you have executables in your tree with names like "tm01.exe" 2011-04-22T09:58:27 find . -name "tm*.exe" 2011-04-22T09:58:31 yes 2011-04-22T09:58:48 Did you check out rtems-testing? Is "sis" in your PATH? 2011-04-22T09:58:49 type sis 2011-04-22T10:00:00 no 2011-04-22T10:01:19 actually I though that this line was wrong: "find . -name "tm*.exe" | xargs ../rtems-testing/sim-scripts/sis" so I change it to "find . -name "tm*.exe" | xargs ../rtems-testing/sim-scripts/sis.in" 2011-04-22T10:01:43 no.. sis.in is not executable 2011-04-22T10:01:59 there should be a rtems-testing/bin directory.. you want that in your PATH 2011-04-22T10:02:03 and I ass execution permission to sis.in 2011-04-22T10:02:19 sis.in is a fragment -- not an entire script 2011-04-22T10:02:26 ok 2011-04-22T10:02:55 *** arvind_khadri has quit IRC 2011-04-22T10:11:16 panzon == Ricardo? 2011-04-22T10:11:29 si 2011-04-22T10:11:32 yes 2011-04-22T10:11:38 ok.. just making sure 2011-04-22T10:14:16 I think that I was making something wrong, I'm compiling using this line 2011-04-22T10:14:31 ../rtems/configure --target=sparc-rtems4.11 --enable-rtemsbsp=sis --enable-tests --disable-posix --disable-itron 2011-04-22T10:15:11 that looks right 2011-04-22T10:15:14 type sis 2011-04-22T10:15:32 doesn't works 2011-04-22T10:15:39 It isn't in your PATH 2011-04-22T10:16:16 *** dr__house has quit IRC 2011-04-22T10:16:24 *** dr__house has joined #rtems 2011-04-22T10:16:24 *** dr__house has joined #rtems 2011-04-22T10:16:56 yea I know, but I don't know where is this "sis" 2011-04-22T10:17:03 do you know? 2011-04-22T10:17:42 i don't think sis has to be... but sparc-rtems4.11-run should be 2011-04-22T10:19:08 if you used the sis simulator for the hello world example..it should be available 2011-04-22T10:19:51 *** arvind_khadri has joined #rtems 2011-04-22T10:19:51 *** arvind_khadri has joined #rtems 2011-04-22T10:24:50 mmm, but then I have to change this execution line : "find . -name "tm*.exe" | xargs ../rtems-testing/sim-scripts/sis" 2011-04-22T10:25:17 the sis script looks for sparc-rtems4.11-run 2011-04-22T10:25:49 to this: "find . -name "tm*.exe" | xargs sparc-rtems4.11-run" 2011-04-22T10:26:26 sparc-rtems4.11-run works 2011-04-22T10:27:00 but I don't know how to use rtems-testing/sim-scripts 2011-04-22T10:27:52 you have rtems-testing/sim-scripts, right? 2011-04-22T10:28:11 go into that directory, and manually try running the sis script there 2011-04-22T10:30:07 do you mean sis.in? 2011-04-22T10:30:20 no, i mean sis 2011-04-22T10:30:31 ./sis 2011-04-22T10:30:44 *** xiangfu has quit IRC 2011-04-22T10:31:20 no, this file doesn't exist 2011-04-22T10:31:38 *** arvind_khadri has quit IRC 2011-04-22T10:32:16 I have a doubt. 2011-04-22T10:33:02 first, I check rtems-testing using cvs 2011-04-22T10:33:17 then 2011-04-22T10:33:30 I have to compile 2011-04-22T10:34:39 does anybody know if it is possible to upload rtems image to leon board with a free software? i see that grmon-rpc has only 21 day evaluation license 2011-04-22T10:35:09 but I don't know if I have to make a new directory, like helloWorld and compile using "../rtems/configure --target=sparc-rtems4.11 --enable-rtemsbsp=sis --enable-tests --disable-posix --disable-itron" 2011-04-22T10:35:38 or I have to compile inside the rtems-testing directory 2011-04-22T10:36:24 if I do it inside new directory I have no problems it compile, when I exec "make" 2011-04-22T10:36:49 rokka: I think you can get it with an academic license 2011-04-22T10:36:53 if you are an academic 2011-04-22T10:37:13 but if I compile inside "rtems-testing" I get an error 2011-04-22T10:37:40 gedare: ok, great 2011-04-22T10:38:02 do you know how much commercial license costS? 2011-04-22T10:38:05 gmake[2]: *** No rule to make target `all'. Stop. 2011-04-22T10:38:20 panzon: I think you messed up your rtems-testing directory 2011-04-22T10:38:50 panzon: you shouldn't run configure inside rtems-testing 2011-04-22T10:39:17 ok 2011-04-22T10:39:29 panzon: you should just check out rtems-testing, and then 'make' directly 2011-04-22T10:39:32 I will check another time 2011-04-22T10:40:02 this is why you do not have a 'sis' in sim-scripts 2011-04-22T10:41:02 panzon: to add a test.. you will be in the psxtmtests directory.. 2011-04-22T10:41:51 and use mktest in rtems-testing/bin. There are templates for various types of test starting points in rtems-testing/rtems-test-templates 2011-04-22T10:42:07 it will put a starting point in the directory you provide on the command line. 2011-04-22T10:42:34 Then add the test directory name to configure.ac and Makefile.am in psxtmtests .. and bootstrap in that directory 2011-04-22T10:42:59 when you are adding tests and editing makefiles, you want to configure with "--enable-maintainer-mode" so they get rebuilt 2011-04-22T10:48:26 ok, now everything is working 2011-04-22T10:50:01 thank you 2011-04-22T10:53:51 *** gedare has quit IRC 2011-04-22T11:01:18 :D 2011-04-22T11:01:25 bbl .. need to run errand 2011-04-22T11:16:35 *** lj8175 has joined #rtems 2011-04-22T11:30:18 *** gedare has joined #rtems 2011-04-22T11:30:29 *** zwj has quit IRC 2011-04-22T11:30:47 *** zwj has joined #rtems 2011-04-22T11:45:41 *** arvind_khadri has joined #rtems 2011-04-22T11:50:01 *** dr__house has quit IRC 2011-04-22T11:50:07 *** dr__house has joined #rtems 2011-04-22T11:50:07 *** dr__house has joined #rtems 2011-04-22T11:54:02 *** dr__house has quit IRC 2011-04-22T11:54:08 *** dr__house has joined #rtems 2011-04-22T11:54:08 *** dr__house has joined #rtems 2011-04-22T11:54:39 *** arvind_khadri has quit IRC 2011-04-22T11:58:01 *** dr__house has quit IRC 2011-04-22T11:58:07 *** dr__house has joined #rtems 2011-04-22T11:58:08 *** dr__house has joined #rtems 2011-04-22T11:58:18 *** arvind_khadri has joined #rtems 2011-04-22T11:58:32 wow.. net connectivity looks poor today 2011-04-22T12:00:02 *** dr__house has quit IRC 2011-04-22T12:00:07 *** dr__house has joined #rtems 2011-04-22T12:00:07 *** dr__house has joined #rtems 2011-04-22T12:00:21 for someone at least. 2011-04-22T12:01:29 :D 2011-04-22T12:04:34 DrJoel: hi 2011-04-22T12:05:09 in the Ml thread about lua, you mean i the goal is to migrate the current rtem shell comands to lua shell, right? 2011-04-22T12:05:22 but what about /dev/* 2011-04-22T12:05:30 *** lj8175 has quit IRC 2011-04-22T12:05:39 actually is this also implemented in the shell? 2011-04-22T12:06:45 /dev/* is in the file system. Independent of the shell 2011-04-22T12:07:23 what is in the shell is commands like cat, dd, stack usage, cpu use, date, etc. I just want all RTEMS commands like UNIX commands to be available for Lua scripts 2011-04-22T12:08:24 ah ok 2011-04-22T12:08:54 so lua need to support rtem file system at first in order to acess /dev/* ? 2011-04-22T12:09:26 ah, avaliable for scripts 2011-04-22T12:09:30 okay i got my answer 2011-04-22T12:09:59 if you think in terms of "pieces and parts".. we have telnet.. login.. password authentication.. command line entry.. shell.. commands 2011-04-22T12:09:59 instead of telnet, you can also use console directly.. 2011-04-22T12:10:03 for each of those, we could have alternatives 2011-04-22T12:10:36 if lua can use regular open, read, write, .. section 2 and 3 POSIX calls, then it is OK to go. 2011-04-22T12:13:26 lua have an ansi c mode, thats how i got boot it on the mm1 board, just compiled statically against rtems, and standard in/out just worked :-) 2011-04-22T12:13:56 but there is also a posix compilation option, 2011-04-22T12:15:10 anyway..i still not understand the topic about OSs very well... 2011-04-22T12:18:01 *** dr__house has quit IRC 2011-04-22T12:18:08 *** dr__house has joined #rtems 2011-04-22T12:18:08 *** dr__house has joined #rtems 2011-04-22T12:18:46 we have the posix version compiling outside rtems. if the gsoc project is accepted, the student will integrate this into rtems, and then add support for shell commands 2011-04-22T12:19:08 i anticipate it will always be used with posix 2011-04-22T12:20:02 *** dr__house has quit IRC 2011-04-22T12:20:07 *** dr__house has joined #rtems 2011-04-22T12:20:07 *** dr__house has joined #rtems 2011-04-22T12:31:54 *** arvind_khadri has quit IRC 2011-04-22T12:44:57 *** arvind_khadri has joined #rtems 2011-04-22T12:44:57 *** arvind_khadri has joined #rtems 2011-04-22T12:55:52 *** petben has quit IRC 2011-04-22T12:57:37 *** arvind_khadri has quit IRC 2011-04-22T12:58:01 *** dr__house has quit IRC 2011-04-22T12:58:07 *** dr__house has joined #rtems 2011-04-22T12:58:07 *** dr__house has joined #rtems 2011-04-22T13:00:01 *** dr__house has quit IRC 2011-04-22T13:00:07 *** dr__house has joined #rtems 2011-04-22T13:00:07 *** dr__house has joined #rtems 2011-04-22T13:12:49 *** dimonov has quit IRC 2011-04-22T13:26:18 Do you know where does psxtm* should be? 2011-04-22T13:26:56 *** dr__house has quit IRC 2011-04-22T13:27:49 *** dimonov has joined #rtems 2011-04-22T13:28:23 panzon: are you still using --disable-posix in your configure flags? 2011-04-22T13:29:03 ok, I understand 2011-04-22T14:00:11 drjoel: meeting's started 2011-04-22T14:02:45 *** DrJoel is now known as rtems|drjoel 2011-04-22T14:03:05 sorry .. missing an IPI on LEON3 and lost track 2011-04-22T14:03:16 heh 2011-04-22T14:04:36 quick when Hugin doesn't have a person on channel 2011-04-22T14:04:58 yeah 2011-04-22T14:05:30 not many conflicts 2011-04-22T14:12:05 panzon: can you post your results (times.txt, psxtimes.txt) somewhere, like pastebin 2011-04-22T14:38:03 leon3 highlight :) 2011-04-22T14:47:40 *** rtems|drjoel is now known as DrJoel 2011-04-22T15:27:34 *** gedare has quit IRC 2011-04-22T15:28:34 *** DrJoel has quit IRC 2011-04-22T17:49:07 *** kiwichris has joined #rtems 2011-04-22T18:54:34 *** dr__house has joined #rtems 2011-04-22T18:54:34 *** dr__house has joined #rtems 2011-04-22T19:41:01 *** methril has quit IRC 2011-04-22T19:43:15 *** dr__house has quit IRC 2011-04-22T19:43:17 *** dr__house` has joined #rtems 2011-04-22T19:48:56 *** dr__house` has quit IRC 2011-04-22T19:49:02 *** dr__house` has joined #rtems 2011-04-22T19:53:32 *** methril has joined #rtems 2011-04-22T19:55:24 *** dr__house` is now known as dr__house 2011-04-22T20:03:53 *** panzon has quit IRC 2011-04-22T20:27:47 *** dr__house has quit IRC 2011-04-22T21:13:45 *** zwj has quit IRC 2011-04-22T21:19:39 *** zwj has joined #rtems 2011-04-22T21:34:09 *** kiwichris has quit IRC 2011-04-22T22:01:54 *** methril has quit IRC 2011-04-22T22:08:23 *** zwj has quit IRC 2011-04-23T03:16:30 *** lj8175 has joined #rtems 2011-04-23T03:29:17 *** panzon has joined #rtems 2011-04-23T03:42:22 *** panzon has quit IRC 2011-04-23T03:42:45 *** panzon has joined #rtems 2011-04-23T04:00:53 gedare: here are the results: http://pastebin.com/xNwQZGHB 2011-04-23T07:15:17 *** dr__house has joined #rtems 2011-04-23T07:15:17 *** dr__house has joined #rtems 2011-04-23T09:24:24 *** xiangfu has joined #rtems 2011-04-23T09:59:41 *** dr__house has quit IRC 2011-04-23T09:59:55 *** dr__house has joined #rtems 2011-04-23T09:59:56 *** dr__house has joined #rtems 2011-04-23T10:25:53 *** xiangfu has quit IRC 2011-04-23T10:31:39 *** dr__house has quit IRC 2011-04-23T10:32:41 *** dr__house has joined #rtems 2011-04-23T10:38:44 *** sqm has joined #rtems 2011-04-23T10:48:43 *** sqm has quit IRC 2011-04-23T10:50:01 *** sqm has joined #rtems 2011-04-23T11:56:21 *** dr__house has quit IRC 2011-04-23T11:56:41 *** dr__house has joined #rtems 2011-04-23T11:56:42 *** dr__house has joined #rtems 2011-04-23T12:01:07 *** dr__house has quit IRC 2011-04-23T12:03:05 *** lj8175 has quit IRC 2011-04-23T12:03:51 *** sqm has quit IRC 2011-04-23T12:42:29 *** gedare has joined #rtems 2011-04-23T12:48:29 *** gedare has quit IRC 2011-04-23T15:09:23 *** gedare has joined #rtems 2011-04-23T15:10:08 *** gedare has joined #rtems 2011-04-23T15:11:57 *** gedare has quit IRC 2011-04-23T17:23:07 *** dr__house has joined #rtems 2011-04-23T17:37:35 *** panzon has quit IRC 2011-04-23T18:15:51 *** methril has joined #rtems 2011-04-23T18:19:19 *** dr__house has quit IRC 2011-04-23T20:27:39 *** kristian1aul has joined #rtems 2011-04-23T20:31:38 *** kristian1aul has quit IRC 2011-04-23T20:31:46 *** kristian1aul has joined #rtems 2011-04-23T20:31:55 *** kristianpaul has quit IRC 2011-04-23T20:31:58 *** kristian1aul has quit IRC 2011-04-23T20:32:03 *** kristianpaul has joined #rtems 2011-04-23T20:32:03 *** kristianpaul has joined #rtems 2011-04-24T01:29:19 *** lj8175 has joined #rtems 2011-04-24T04:39:04 *** xiangfu has joined #rtems 2011-04-24T05:31:05 *** xiangfu has quit IRC 2011-04-24T07:07:44 *** sqm has joined #rtems 2011-04-24T08:03:36 *** dr__house has joined #rtems 2011-04-24T08:03:36 *** dr__house has joined #rtems 2011-04-24T09:27:43 *** lj8175 has quit IRC 2011-04-24T09:31:36 *** lj8175 has joined #rtems 2011-04-24T11:03:44 *** sqm has quit IRC 2011-04-24T11:04:03 *** lj8175 has quit IRC 2011-04-24T11:28:55 *** peerst_ has left #rtems 2011-04-24T12:00:16 *** dr__house has quit IRC 2011-04-24T12:04:28 *** kristianpaul has quit IRC 2011-04-24T12:05:16 *** kristianpaul has joined #rtems 2011-04-24T12:05:16 *** kristianpaul has joined #rtems 2011-04-24T12:05:34 *** dr__house has joined #rtems 2011-04-24T12:07:05 *** dr__house` has joined #rtems 2011-04-24T12:07:06 *** dr__house has quit IRC 2011-04-24T12:07:25 *** dr__house` is now known as dr__house 2011-04-24T12:10:33 *** dr__house has quit IRC 2011-04-24T12:10:48 *** dr__house has joined #rtems 2011-04-24T14:08:01 *** dr__house has quit IRC 2011-04-24T16:01:03 *** dr__house has joined #rtems 2011-04-24T16:02:10 *** dr__house has quit IRC 2011-04-24T16:02:44 *** dr__house has joined #rtems 2011-04-24T16:04:03 *** dr__house has joined #rtems 2011-04-24T16:04:03 *** dr__house has joined #rtems 2011-04-24T16:08:47 *** dr__house has quit IRC 2011-04-24T17:00:21 *** core-ix has joined #rtems 2011-04-24T17:29:46 *** dr__house has joined #rtems 2011-04-24T17:29:47 *** dr__house has joined #rtems 2011-04-24T19:40:57 *** methril has quit IRC 2011-04-24T19:53:30 *** methril has joined #rtems 2011-04-24T20:25:40 *** xiangfu has joined #rtems 2011-04-24T20:42:59 *** dr__house has quit IRC 2011-04-24T22:40:47 *** xiangfu has quit IRC