2010-06-28T00:01:36 *** arvind_khadri has quit IRC 2010-06-28T00:03:55 *** dr__house has quit IRC 2010-06-28T00:23:32 *** dr__house has joined #rtems 2010-06-28T02:04:41 *** sebhub has joined #rtems 2010-06-28T02:05:13 good morning 2010-06-28T02:19:54 *** madrazr has joined #rtems 2010-06-28T02:22:05 *** alinrus has quit IRC 2010-06-28T02:42:32 *** madrazr has quit IRC 2010-06-28T02:45:18 *** madrazr has joined #rtems 2010-06-28T02:48:57 *** dr__house has quit IRC 2010-06-28T03:04:48 *** madrazr has quit IRC 2010-06-28T03:05:06 *** lekernel has quit IRC 2010-06-28T03:05:50 *** lekernel has joined #rtems 2010-06-28T03:05:54 *** groleo has joined #rtems 2010-06-28T04:07:27 *** dr__house has joined #rtems 2010-06-28T04:21:18 *** madrazr has joined #rtems 2010-06-28T04:45:50 *** alinrus has joined #rtems 2010-06-28T04:50:05 *** lekernel has quit IRC 2010-06-28T04:51:11 *** lekernel has joined #rtems 2010-06-28T05:36:59 *** exception13 has joined #rtems 2010-06-28T06:40:17 *** zwj has joined #rtems 2010-06-28T06:40:46 sebhub:hi 2010-06-28T06:40:49 *** methril_ has quit IRC 2010-06-28T06:40:51 good evening 2010-06-28T06:41:17 *** methril_ has joined #rtems 2010-06-28T06:57:33 hi 2010-06-28T06:57:37 *** madrazr has quit IRC 2010-06-28T06:58:42 zwj: can you add the patch as an attachement to https://www.rtems.org/bugzilla/show_bug.cgi?id=1593 2010-06-28T07:02:00 *** madrazr has joined #rtems 2010-06-28T07:19:28 *** tuxmaniac has joined #rtems 2010-06-28T07:23:41 *** madrazr has quit IRC 2010-06-28T07:26:12 *** madrazr has joined #rtems 2010-06-28T07:35:51 *** madrazr has quit IRC 2010-06-28T07:38:53 *** madrazr has joined #rtems 2010-06-28T07:40:56 sebhub:ok, i am now compiling the sysinit, until it is ok i will make a patch attached to it 2010-06-28T07:41:08 ok 2010-06-28T08:04:35 *** madrazr has quit IRC 2010-06-28T08:05:10 *** exception13 has left #rtems 2010-06-28T08:06:18 sebhub:about test case of the sysinit, how to test the two functions? modify the linkcmd? 2010-06-28T08:06:52 yes, you need to modify the linkcmds 2010-06-28T08:07:17 *** madrazr has joined #rtems 2010-06-28T08:08:04 and using the macor SYSINIT_ENTRY define some init functions in test files? 2010-06-28T08:08:24 yes 2010-06-28T08:08:52 ok, and based on which branch of rtems we make a patch? 2010-06-28T08:09:03 please use always CVS head 2010-06-28T08:09:56 but yestoday i used CVS update and ./bootstape it says my autoconf need 2.65 or high 2010-06-28T08:10:23 you should update your tools in this case 2010-06-28T08:10:24 i have already install rtems_autoconf2.65 2010-06-28T08:10:41 are the tools in front of your $PATH? 2010-06-28T08:11:14 what is the output of autoconf --version? 2010-06-28T08:11:32 yes, i can bootstrape rtems in the early cvs head successfully 2010-06-28T08:12:13 2.63 2010-06-28T08:12:39 what is the output of: which autoconf 2010-06-28T08:13:15 *** madrazr has quit IRC 2010-06-28T08:14:51 it is /usr/bin/autoconf 2010-06-28T08:15:37 you should put the RTEMS tools in front of you $PATH for RTEMS development 2010-06-28T08:16:01 PATH="/opt/rtems-4.10/bin:$PATH" 2010-06-28T08:16:48 *** madrazr has joined #rtems 2010-06-28T08:19:44 but i echo $PATH , the opt/rtems-4.11/bin is in path 2010-06-28T08:20:10 yes, but the order in $PATH is importent 2010-06-28T08:21:17 oh, ok. i see 2010-06-28T08:24:51 *** tuxmaniac has quit IRC 2010-06-28T08:25:00 *** madrazr has quit IRC 2010-06-28T08:29:21 *** madrazr has joined #rtems 2010-06-28T08:33:29 *** madrazr has quit IRC 2010-06-28T08:33:35 *** DrJoel has joined #rtems 2010-06-28T08:33:35 *** ChanServ sets mode: +o DrJoel 2010-06-28T08:35:52 hi joel, did you ever work with the boost unit test framework or the google c++ testing framework? 2010-06-28T08:36:33 no.. afraid not.. 2010-06-28T08:37:32 i do currently some research to find a good unit testing framework which my be also suitable for rtems tests 2010-06-28T08:43:31 Two requirements are to run tests easily on simulators or real hardware under control and to be able to diff screens with some acceptable level of variation 2010-06-28T08:45:49 I first tried the boost unit test framework, but it seems that this one is not thread-safe 2010-06-28T08:45:52 *** madrazr has joined #rtems 2010-06-28T08:46:03 now i have a look at the google c++ testing framework 2010-06-28T08:46:23 if you have time, you may have a look at http://code.google.com/p/googletest/ 2010-06-28T08:47:12 it may be nice to get rtems on the front page of this ;-) 2010-06-28T08:52:07 Any spot like that is good. I will try to look at it soon. Does it meet your first review? 2010-06-28T08:53:10 yes, it is promising 2010-06-28T08:53:12 FWIW RTEMS is goign to be in the Coverity open source annual-ish report 2010-06-28T08:53:25 that is great 2010-06-28T08:56:50 *** bubaflub has joined #rtems 2010-06-28T09:00:48 *** madrazr has quit IRC 2010-06-28T09:05:28 *** madrazr has joined #rtems 2010-06-28T09:12:37 *** madrazr has quit IRC 2010-06-28T09:16:29 *** madrazr has joined #rtems 2010-06-28T09:18:23 *** methril_work has joined #rtems 2010-06-28T09:28:03 *** madrazr has quit IRC 2010-06-28T09:32:27 *** madrazr has joined #rtems 2010-06-28T09:32:28 *** dr__house` has joined #rtems 2010-06-28T09:32:28 *** dr__house` has joined #rtems 2010-06-28T09:33:23 *** dr__house has quit IRC 2010-06-28T09:34:50 does rtems support pthread_key_create? 2010-06-28T09:36:06 ok, a config value failed 2010-06-28T09:36:40 yes.. pthread_key's are supported.. there is an issue when you use keys with unlimited threads/tasks but they are there 2010-06-28T09:37:30 ok, i can run the sample test of the google c++ testing framework, it was very easy 2010-06-28T09:37:38 on SIS 2010-06-28T09:38:20 cool.. how much work to use it on the existing tests though 2010-06-28T09:39:06 hm, first it requires C++ and POSIX api 2010-06-28T09:39:16 *** madrazr has quit IRC 2010-06-28T09:40:37 the good thing is that it is considerable easier to write unit test with such a framework 2010-06-28T09:41:55 *** madrazr has joined #rtems 2010-06-28T09:46:40 *** madrazr has quit IRC 2010-06-28T09:51:08 *** madrazr has joined #rtems 2010-06-28T09:57:10 *** madrazr has quit IRC 2010-06-28T09:59:11 *** madrazr has joined #rtems 2010-06-28T10:00:22 *** groleo has quit IRC 2010-06-28T10:02:58 *** dr__house` is now known as dr__house 2010-06-28T10:07:35 *** madrazr has quit IRC 2010-06-28T10:15:28 *** madrazr has joined #rtems 2010-06-28T10:20:13 *** madrazr has quit IRC 2010-06-28T10:21:24 *** tuxmaniac has joined #rtems 2010-06-28T10:45:02 *** bubaflub has quit IRC 2010-06-28T10:53:28 *** gedare has joined #rtems 2010-06-28T11:00:56 drjoel: are you trying to reduce the amount of inline code? 2010-06-28T11:01:22 it seems some comments related to coverage and inline code indicates this 2010-06-28T11:03:07 yes and no. Inlining code that has multiple compares in it has turned out to be bad if it is used a lot. 2010-06-28T11:03:42 ok so maybe a rule of thumb to avoid inlining multiple if'else constructs and switch blocks? 2010-06-28T11:04:31 the tradeoff may end up being larger functions, where we don't want to sacrifice performance 2010-06-28T11:04:34 yes. if used in more places than you can easily test. There was one filesystem one that had 2 compares and was used ~80 times. That was 160 extra test cases (2 bad and 1 good) 2010-06-28T11:04:55 oh ok i see 2010-06-28T11:05:04 most of the inline code I deal with gets used in <5 places 2010-06-28T11:05:15 some stuff isn't really as time critical as you might think. But most of what you are working on is. 2010-06-28T11:05:20 yeah 2010-06-28T11:05:27 i think i'm in a "special" box 2010-06-28T11:05:30 Yes and is simple -- initialize, one test, one math operation 2010-06-28T11:05:37 :) 2010-06-28T11:05:40 *** sebhub has quit IRC 2010-06-28T11:06:00 ok .. well I'm mostly concerned with getting things working, but if I can make decisions now that make it easier to merge later, it is better 2010-06-28T11:06:50 You should see that coverage runs this morning .. aniceberg adn I both worked on tests this weekend. I worked on some termios test framework and he did "rtems/error.h". went from 89.9% to 91.57% 2010-06-28T11:07:25 i think i managed to run coverage myself once, but couldn't make heads or tails out of what happened: ) 2010-06-28T11:07:35 that's a nice jump though 2010-06-28T11:08:10 i saw his PRs, seems to be making good headway 2010-06-28T11:09:21 I'm thinking of how I might be able to make it easier to "plug" code in to the scheduler, without having to touch the core files, for example I use switch blocks to vector to specific readyq implementations 2010-06-28T11:09:40 I wonder if I can use some better magic to make it easier to register a new implementation 2010-06-28T11:09:45 something I have to think on further. :) 2010-06-28T11:10:21 the one way I can think of right now is to have a readyq operations table that can be overriden 2010-06-28T11:10:33 so that function pointers do all the work 2010-06-28T11:10:47 but i'm not a huge fan of function pointers :P 2010-06-28T11:11:04 better than hard-coded jump tables. :) 2010-06-28T11:11:34 yeah and that's more or less what is developing now, so I'll probably re-work that part, maybe while i'm merging the EDF code 2010-06-28T11:12:15 Or while I do the confdefs part, since that is how users will be able to select the implementations anyway.. 2010-06-28T11:12:27 At any rate, I'm happy for now. :) 2010-06-28T11:13:36 looks like percpu will get in tonight / tomorrow, huh? 2010-06-28T11:16:17 I am swamped this Monday and haven't ready that one. :) 2010-06-28T11:24:14 fair enough 2010-06-28T12:07:59 *** aanjhan has joined #rtems 2010-06-28T12:08:38 *** aanjhan has joined #rtems 2010-06-28T12:11:53 *** tuxmaniac has quit IRC 2010-06-28T12:16:59 *** bubaflub has joined #rtems 2010-06-28T12:20:14 DrJoel:hi 2010-06-28T12:21:27 i have creat a PR1593 https://www.rtems.org/bugzilla/show_bug.cgi?id=1593 for sequenced initilization. please help me comment it .thank you . it is time to bed . byb 2010-06-28T12:22:20 night.. you are in the queue :-D 2010-06-28T12:22:52 *** zwj has quit IRC 2010-06-28T12:30:27 *** arvind_khadri has joined #rtems 2010-06-28T12:30:46 *** madrazr has joined #rtems 2010-06-28T12:31:16 DrJoel, hi 2010-06-28T12:31:22 hey 2010-06-28T12:31:37 how are you :) 2010-06-28T12:32:26 swamped on a Monday.. took off early Friday to go to a wedding out of town 2010-06-28T12:32:40 have you had any luck? 2010-06-28T12:33:29 nothing great :) waiting for my results to come, will come by tomorrow .... 2010-06-28T12:33:37 *** mpanetta has quit IRC 2010-06-28T12:34:04 good luck.. 2010-06-28T12:34:16 thanks a lot :) 2010-06-28T12:34:24 Wanted to know what to do next ? 2010-06-28T12:34:44 and sorry if i was rude by sending the mail back to you which you had sent to me 2010-06-28T12:35:23 no.. I actually get so much email, it isn't a problem to remind me. 2010-06-28T12:35:37 *** mpanetta has joined #rtems 2010-06-28T12:35:47 So can you run say the telnetd network demo on qemu? 2010-06-28T12:38:25 *** madrazr has quit IRC 2010-06-28T12:41:19 *** madrazr has joined #rtems 2010-06-28T12:43:05 DrJoel, do you mean running qemu-nic from rtems-testing ?? 2010-06-28T12:46:03 DrJoel, do you mean running qemu-nic from rtems-testing ?? 2010-06-28T12:48:15 *** madrazr has quit IRC 2010-06-28T12:50:02 arvind_khadri: yes .. and successfully running the telnetd demo 2010-06-28T12:50:59 DrJoel, oh ok, is there any tutorial about the telnetd demo, like how to run it and stuff 2010-06-28T12:53:33 DrJoel, i could run qemu-nic successfully now, with output on the stdout... 2010-06-28T12:53:42 Other than whatever readme, not really .. 2010-06-28T12:54:09 if it is really running .. you can ping 10.0.2.2 and get a reply from qemu 2010-06-28T12:54:10 then telnet 10.0.2.5 (should be the RTEMS target) 2010-06-28T12:55:26 README.connex ?? 2010-06-28T12:55:56 no.. that is for the gumstix board 2010-06-28T12:56:07 I just gave you the main part of a readme 2010-06-28T12:56:21 the readme would be in network-demos/telnetd 2010-06-28T12:57:50 DrJoel, oh ok, would network-demos be another module that i need to pull from CVS ? 2010-06-28T12:58:00 yes .. 2010-06-28T12:58:30 oh ok :) 2010-06-28T12:59:18 and when you get it checked out, there is a set of network-config*h files in the top. You want network-qemu.h to be the main one. By default, it is built for loopback only 2010-06-28T12:59:27 *** exception13 has joined #rtems 2010-06-28T12:59:45 and there are some instructions in rtems-testing/qemu-support on installing the host OS network intefface supprot for tunneling/tap driver 2010-06-28T12:59:48 *** alinrus has quit IRC 2010-06-28T12:59:50 *** methril_work has quit IRC 2010-06-28T12:59:50 *** gedare has quit IRC 2010-06-28T12:59:51 *** dr__house has quit IRC 2010-06-28T12:59:52 *** methril_ has quit IRC 2010-06-28T12:59:52 *** ChanServ has quit IRC 2010-06-28T13:01:02 *** dr__house has joined #rtems 2010-06-28T13:01:02 *** methril_work has joined #rtems 2010-06-28T13:01:02 *** methril_ has joined #rtems 2010-06-28T13:01:02 *** ChanServ has joined #rtems 2010-06-28T13:01:02 *** pratchett.freenode.net sets mode: +o ChanServ 2010-06-28T13:01:26 *** gedare has joined #rtems 2010-06-28T13:03:15 DrJoel, I am able to digest something, would get back to you when once am done getting the module 2010-06-28T13:06:23 ok.. I need to move to the Centos machine to test something for a whlie and will be awk 2010-06-28T13:07:15 *** alinrus has joined #rtems 2010-06-28T13:12:28 *** madrazr has joined #rtems 2010-06-28T13:28:27 *** madrazr1 has joined #rtems 2010-06-28T13:29:43 *** madrazr has quit IRC 2010-06-28T13:29:49 dr__house: check out the coverage progress over the weekend 2010-06-28T13:29:49 http://www.rtems.org/ftp/pub/rtems/people/joel/coverage/erc32.html#OsPD 2010-06-28T13:29:53 4% on developmental in 1 week :-D 2010-06-28T13:32:27 DrJoel: wow!! that's super cool! :D 2010-06-28T13:33:13 *** madrazr1 has quit IRC 2010-06-28T13:33:32 aniceberg got through the libmisc error.c file and I figured out how to structure test drivers to support testing termios! 2010-06-28T13:37:07 *** madrazr has joined #rtems 2010-06-28T13:40:11 DrJoel: haven't been able to help out much with the mentoring. Life's been in a big mess :P 2010-06-28T13:40:59 it's ok.. 2010-06-28T13:43:16 *** madrazr has quit IRC 2010-06-28T13:44:05 DrJoel: I wanted to ask you a few things. Can you suggest a set of seminal papers on basics of embedded systems, something that every embedded systems programmer shoud read? Also a book for the same? 2010-06-28T13:45:10 *** madrazr has joined #rtems 2010-06-28T13:47:04 Wow! That's a big question. I personally like a lot of the articles at embedded.com. They have a LOT of practical columns and much of it is pretty solid theorectically. 2010-06-28T13:47:29 Much of the research paper is on scheduling theory. 2010-06-28T13:57:09 alinrus: replying to your email 2010-06-28T13:57:34 DrJoel: thanks a lot 2010-06-28T13:58:04 DrJoel: ok. Any preparation that I need to do before I join in? these are the courses I shall be taking http://www.cis.upenn.edu/grad/embedded-curriculum.shtml 2010-06-28T14:01:23 The obvious thing is to figure out which books they are using for the Core courses and buy the cheaper Indian edition used if it is available and start reading. 2010-06-28T14:01:53 US and UK editions tend to be more expensive. 2010-06-28T14:02:12 You might be able to borrow them from uni library there via a friend 2010-06-28T14:02:34 then look at their references 2010-06-28T14:03:19 another approach is to play with the software they recommend if it is freely available 2010-06-28T14:03:31 ok 2010-06-28T14:03:57 the Cis 540 web page has their books and sw links 2010-06-28T14:04:14 that way (just like gsoc) you won't be stuck fighting basic stuff 2010-06-28T14:05:00 DrJoel: yeah :) 2010-06-28T14:05:01 are you an ACM or IEEE member? 2010-06-28T14:05:48 *** madrazr has quit IRC 2010-06-28T14:06:12 DrJoel: not anymore. The ACM account given by google expired last month 2010-06-28T14:06:21 they have a lot of free stuff for members.. skillware type online courses and a number of classic CS books 2010-06-28T14:06:21 DrJoel: would you suggest becoming one? 2010-06-28T14:07:17 I am in both.. my life insurance is through IEEE. As a student, I would do it since it is cheap and the resources are good. As a professional, IMO it is part of being serious about your professino and keeping up. 2010-06-28T14:07:57 I read CACM, IEEE Computer, and IEEE Spectrum every month (even if I don't read ever article.. some don't interest me) 2010-06-28T14:08:47 *** madrazr has joined #rtems 2010-06-28T14:09:52 I don't know if they have the full text but Spectrum is online also http://spectrum.ieee.org/magazine/ 2010-06-28T14:15:35 *** madrazr has quit IRC 2010-06-28T14:17:14 DrJoel: is embedded.com open only to professionals? 2010-06-28T14:18:33 *** madrazr has joined #rtems 2010-06-28T14:20:58 *** madrazr has quit IRC 2010-06-28T14:47:29 *** cdcs has joined #rtems 2010-06-28T15:08:10 dr__house: no its for everyone 2010-06-28T15:08:17 there are some nice articles at times on that 2010-06-28T15:21:49 DrJoel, i got the module and read the README, i need to edit a few files according to it 2010-06-28T15:30:03 *** arvind_khadri has quit IRC 2010-06-28T15:35:02 *** arvind_khadri has joined #rtems 2010-06-28T15:36:14 DrJoel, i got the module and read the README, i need to edit a few files according to it 2010-06-28T15:39:56 shouldn't have to edit a lot 2010-06-28T15:45:24 but am not sure what to edit :( 2010-06-28T15:47:04 DrJoel, ^ 2010-06-28T15:47:23 you have to copy networkconfig-qemu.h to networkconfig.h.. 2010-06-28T15:47:26 *** bubaflub has quit IRC 2010-06-28T15:48:00 place qemu-ifup and qemu-ifdown in /etc and do a modprobe .. see the readme in qemu-support. 2010-06-28T15:48:19 can you run heloo on qemu? 2010-06-28T15:48:47 then copy the telnetd.exe ~/qemu/hd/test.exe and run qemu-nic 2010-06-28T15:49:17 * dr__house heads off for the day. 2010-06-28T15:49:33 DrJoel: thanks for some very valuable information :) 2010-06-28T15:49:33 DrJoel, /etc of fedora ? 2010-06-28T15:49:39 DrJoel: good day 2010-06-28T15:49:44 arvind_khadri: good night 2010-06-28T15:49:50 dr__house, gn :) tc 2010-06-28T15:49:51 aanjhan: good day 2010-06-28T15:50:02 *** dr__house has quit IRC 2010-06-28T15:50:43 arvind_khadri: yes 2010-06-28T15:51:34 DrJoel, ok, i could run qemu-nic , doesnt that mean that I ran hello. sorry for being so dumb 2010-06-28T15:54:16 what program run when you did qemu-nic? 2010-06-28T15:54:54 qemu-nic is setup to run the RTEMS executable that has been copied to ~/qemu/hd/test.exe. For networking examples, you have to copy them by hand to that folder 2010-06-28T15:55:40 aanjhan: can you help arvind_khadri with qemu for a few minutes? have telecon and won't be able to pay attention 2010-06-28T15:56:59 I just ran qemu-nic, with no parameter. and then it showed me the output of hello world 2010-06-28T15:58:41 then you need to build the network-demos, copy telnetd.exe ~/qemu/hd/test.exe and repeat 2010-06-28T16:00:16 DrJoel, oh ok, i just copied qemu-if* to /etc 2010-06-28T16:00:36 there are instructions about doing a modprobe 2010-06-28T16:00:45 the copy to /etc/ is one time 2010-06-28T16:01:00 the modprobe lasts until you reboot 2010-06-28T16:01:14 the copy -> ~/qemu/hd/test.exe happens for each executable. 2010-06-28T16:02:02 yeah read it. understood thanks a lot 2010-06-28T16:02:12 its late here, would like to call a day 2010-06-28T16:02:26 will come back to you tomorrow with all this done :) 2010-06-28T16:02:37 thanks a lot for helping :) and bearing with me 2010-06-28T16:02:52 have a good day :) DrJoel 2010-06-28T16:03:06 night.. there is a readme. :) 2010-06-28T16:03:41 DrJoel, yeah i meant that I read the readme 2010-06-28T16:04:45 night :-D 2010-06-28T16:44:41 *** methril__ has joined #rtems 2010-06-28T16:45:36 *** methril_ has quit IRC 2010-06-28T16:45:44 *** methril_work has quit IRC 2010-06-28T16:56:51 *** methril__ has quit IRC 2010-06-28T16:57:47 *** cdcs has quit IRC 2010-06-28T17:09:04 arvind_khadri: around still? 2010-06-28T17:09:23 arvind_khadri: sorry. Was away getting all my toolchains fixed on the new ubuntu 2010-06-28T17:09:45 arvind_khadri: need to get started on RTEMS work again. Now that I am having little work :-) 2010-06-28T17:09:49 He is logged in but I don't think so. 2010-06-28T17:10:03 DrJoel: aah ok. 2010-06-28T17:10:13 alinrus: any specific issue I can help you with? 2010-06-28T17:10:49 DrJoel: let me know if I can be of any help to any one. I am relatively free from now for 2 months or even more :) 2010-06-28T17:13:30 *** methril__ has joined #rtems 2010-06-28T17:22:31 * DrJoel waves.. time to head home 2010-06-28T17:24:04 *** DrJoel has quit IRC 2010-06-28T17:25:14 *** exception13 has left #rtems 2010-06-28T19:25:24 *** DrJoel has joined #rtems 2010-06-28T19:25:24 *** ChanServ sets mode: +o DrJoel 2010-06-28T19:30:02 DrJoel: I'm using the POSIX api and I call aio_enqueue from POSIX_Init. Now aio_enqueue tries to create a detached thread to work on the given request but unlike on Linux it only runs after aio_enqueue finishes, so having a while (1); at the end of aio_enqueue means the create thread will never run. any thoughts on this 2010-06-28T19:30:33 do you set the attributes on the thread explicitly? 2010-06-28T19:31:18 pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_DETACHED); 2010-06-28T19:31:18 int ret = pthread_create (threadp, &attr, tf, arg); 2010-06-28T19:31:36 and you initialize the attributes with pthread_attr_initialize? 2010-06-28T19:31:46 yes 2010-06-28T19:32:10 it's weird that it works after aio_enqueue finishes 2010-06-28T19:32:28 *** mpanetta has quit IRC 2010-06-28T19:32:42 I suspect you are creating the aio_enqueue from the initialization task which runs at the highest priority and non-preemptible. The default pthread attributes are different between RTEMS and Linux. 2010-06-28T19:33:11 DrJoel: yes i'm running it from the initialization task 2010-06-28T19:33:31 should I use another task ? 2010-06-28T19:33:45 The easiest thing would be to let the Init thread sleep a little while to let it do its work. 2010-06-28T19:33:52 a lot of the tests have to do that. 2010-06-28T19:34:18 *** mpanetta has joined #rtems 2010-06-28T19:34:24 but since I'll never have to call aio_enqueue from my application but aio_write and aio_read call this 2010-06-28T19:34:37 remember to turn on the clock driver. You don't have to sleep for very long 2010-06-28T19:35:04 I think a way to do this could also be to start a thread from aio_write 2010-06-28T19:35:48 Doesn't matter where you start it.. if the set of other threads in the system don't let it run, no work will get done. The network stack and disk IO have the same issue. 2010-06-28T19:36:28 That's why the network stack lets you specify the priority of its daemons 2010-06-28T19:36:39 doesn't solve the problem but gives the user a little more control 2010-06-28T19:46:20 For testing purposes, you might want to lower the priority of the Init thread and make it preemptible. Or just sleep and check 2010-06-28T20:07:08 it works if I add a sleep(1) to the init task :D 2010-06-28T20:07:36 great. that is important but it is more important that you understand why that makes it work. :) 2010-06-28T20:08:00 and why if you delete the CONFIGURE_APPLICATION_NEEDS_CLOCK_DRIVER it will break (for a different and interesting) reason 2010-06-28T20:19:26 * DrJoel is going awk .. 2010-06-28T21:14:25 *** arvind_khadri has quit IRC 2010-06-28T21:26:42 *** arvind_khadri has joined #rtems 2010-06-28T22:12:47 *** dr__house has joined #rtems 2010-06-28T22:42:36 *** alinrus1 has joined #rtems 2010-06-28T22:42:53 *** alinrus has quit IRC 2010-06-28T23:45:11 *** alinrus1 has quit IRC 2010-06-28T23:58:33 *** n3oo3n_ is now known as n3oo3n 2010-06-29T00:00:43 *** alinrus has joined #rtems 2010-06-29T00:04:58 *** arvind_khadri has quit IRC 2010-06-29T00:14:31 *** arvind_khadri has joined #rtems 2010-06-29T00:53:21 *** arvind_khadri has quit IRC 2010-06-29T01:05:29 *** arvind_khadri has joined #rtems 2010-06-29T01:59:24 *** sebhub has joined #rtems 2010-06-29T01:59:35 good morning 2010-06-29T02:17:15 *** madrazr has joined #rtems 2010-06-29T02:27:17 *** lekernel has quit IRC 2010-06-29T02:48:57 hi, does anyone know how do i compile network-demos ?? 2010-06-29T03:09:57 *** madrazr has quit IRC 2010-06-29T03:13:56 look in the README and Makefile 2010-06-29T03:40:40 *** mpanetta has quit IRC 2010-06-29T03:42:05 *** mpanetta has joined #rtems 2010-06-29T04:03:09 *** Raj_ has joined #rtems 2010-06-29T04:06:47 *** tuxmaniac has joined #rtems 2010-06-29T04:20:40 *** madrazr has joined #rtems 2010-06-29T04:23:08 *** madrazr has quit IRC 2010-06-29T04:23:11 *** madrazr has joined #rtems 2010-06-29T04:35:35 DrJoel: hello 2010-06-29T04:36:39 *** madrazr has quit IRC 2010-06-29T04:36:40 *** madrazr1 has joined #rtems 2010-06-29T04:43:00 *** madrazr1 has quit IRC 2010-06-29T04:49:38 *** lekernel has joined #rtems 2010-06-29T04:54:25 *** exception13 has joined #rtems 2010-06-29T05:08:53 *** Raj_ has quit IRC 2010-06-29T05:20:12 *** madrazr has joined #rtems 2010-06-29T05:56:54 *** groleo has joined #rtems 2010-06-29T06:01:56 hi 2010-06-29T06:02:32 DrJoel: i tried to compile the POSIX api and i got the too old newlib error :p i will try to compile a gcc myself with a newer newlib 2010-06-29T06:02:44 *** groleo has joined #rtems 2010-06-29T06:02:44 cause i got problemes with sched_ss_* stuff 2010-06-29T06:12:55 *** lekernel has quit IRC 2010-06-29T06:13:08 *** lekernel has joined #rtems 2010-06-29T06:23:47 *** arvind_khadri has quit IRC 2010-06-29T06:31:53 *** methril_work has joined #rtems 2010-06-29T07:20:28 Fallenou: what host are you on? 2010-06-29T07:20:53 alinrus: moving along again? 2010-06-29T07:21:14 Fallenou doing the framebuffer driver 2010-06-29T07:21:21 DrJoel* 2010-06-29T07:21:25 * DrJoel is planning a GSOC mid-term report on all projects. Will send out request for info later 2010-06-29T07:21:39 *** tuxmaniac has quit IRC 2010-06-29T07:21:46 Fallenou: Host OS.... Ubuntu? Fedora? Windows? 2010-06-29T07:21:51 Ubuntu 2010-06-29T07:22:10 i've seen a wiki page about how to compile a toolchain on ubuntu 2010-06-29T07:22:17 on rtems wiki 2010-06-29T07:22:25 actually alinrus gave me the link =) 2010-06-29T07:22:32 If you are on the frame buffer driver then you definitely should submit some core part of the BSP and let it get merged. Let us get through zwj's first GSOC submission and then do it 2010-06-29T07:23:12 Fallenou: in rtems-testing/simple-build-script is a very simple build script which might help also 2010-06-29T07:23:23 *** tuxmaniac has joined #rtems 2010-06-29T07:23:29 what is this script doing ? 2010-06-29T07:23:53 building autoconf, make, binutils, gcc/newlib and gdb :-D 2010-06-29T07:23:54 is there some doc about how to submit code ? 2010-06-29T07:24:03 oh ok :) exactly what i'm doing :p 2010-06-29T07:24:04 it assumes you have patched the source trees appropriately 2010-06-29T07:24:15 yes.. in wiki 2010-06-29T07:24:33 i still am going to do it manually, at least for now, just to have done it once, to know how to do it, to learn 2010-06-29T07:24:52 but thanks for the intel :) 2010-06-29T07:25:19 you can look at the script for configure commands at least 2010-06-29T07:25:28 yes 2010-06-29T07:27:29 sebhub: simple question if you are there 2010-06-29T07:29:59 Fallenou, did you do an asynchronous UART driver already? 2010-06-29T07:30:36 maybe this should go before the framebuffer 2010-06-29T07:30:42 no i don't 2010-06-29T07:31:45 sebhub: need to run.. can you submit a patch to fix tftptest.. network-demos/tftpTest/init.c:39: undefined reference to `rtems_bsdnet_initialize_tftp_filesystem' 2010-06-29T07:31:50 since the uart already works i wanted to add more support 2010-06-29T07:32:01 I assume this needs to be replaced with a proper mount call 2010-06-29T07:32:08 *** DrJoel has quit IRC 2010-06-29T08:05:20 *** exception13 has left #rtems 2010-06-29T08:26:46 *** tsm_ice9 has joined #rtems 2010-06-29T08:27:01 *** tsm_ice9 has left #rtems 2010-06-29T08:27:38 *** zwj has joined #rtems 2010-06-29T08:48:00 *** aniceberg has joined #rtems 2010-06-29T08:54:26 ~seen ChrisJohns 2010-06-29T08:54:40 ot 2010-06-29T08:54:43 ~bot 2010-06-29T08:55:44 i would like to get in contact with some openbdm devs, ChrisJohns is the RTEMS Coldfire guy, isn't it? 2010-06-29T08:57:38 or some other Coldfire guys that could me give a hand with it? 2010-06-29T08:59:25 *** bubaflub has joined #rtems 2010-06-29T09:00:27 you can meet chris normally four hours earlier 2010-06-29T09:01:18 well i could try and help out 2010-06-29T09:01:40 methril_work, ? 2010-06-29T09:02:28 yes 2010-06-29T09:02:31 thank you 2010-06-29T09:02:39 i'm looking at the openbdm code 2010-06-29T09:02:51 you mean bdm-tools ? 2010-06-29T09:02:57 and i would like to understand the driver and the library options 2010-06-29T09:03:06 http://openbdm.sourceforge.net 2010-06-29T09:03:58 I think you can get the code from here : http://sourceforge.net/projects/openbdm/develop 2010-06-29T09:04:07 i've the code 2010-06-29T09:04:22 i would like to understand if i could extend to support the USBDM devices 2010-06-29T09:05:14 oh, well, I don't want to speak in Chris name. drop him an email 2010-06-29T09:05:36 ok groleo, thank you 2010-06-29T09:05:52 np 2010-06-29T09:05:54 yw 2010-06-29T09:06:08 i'm trying to sve time understanding the full source code :) 2010-06-29T09:06:29 *** arvind_khadri has joined #rtems 2010-06-29T09:25:49 *** arvind_khadri has quit IRC 2010-06-29T09:32:56 *** aniceberg has quit IRC 2010-06-29T09:44:47 *** aniceberg has joined #rtems 2010-06-29T09:49:33 sebhub: hi, i have write a test case, but donot have test it. now i try to modify the linkcmd for sis platform. which cpu should i modify Erc32, leon2 or leon3? 2010-06-29T09:51:06 oh, i have seen it . linkcmd.base in shared dir 2010-06-29T10:10:56 yes, exactly 2010-06-29T10:11:35 *** aniceberg has quit IRC 2010-06-29T10:12:07 *** DrJoel has joined #rtems 2010-06-29T10:12:08 *** ChanServ sets mode: +o DrJoel 2010-06-29T10:16:40 now compiling , if it is ok i will send to you 2010-06-29T10:17:40 *** madrazr has quit IRC 2010-06-29T10:18:08 *** peerst has joined #rtems 2010-06-29T10:18:16 hi 2010-06-29T10:18:54 hi 2010-06-29T10:20:03 does anyone here know if the gdb "script" file that shows rtems objects still exists? 2010-06-29T10:20:44 yes, one moment 2010-06-29T10:21:32 http://wiki.rtems.org/wiki/index.php/GDB_Python 2010-06-29T10:22:22 oh its in python now!! 2010-06-29T10:23:06 peerst: it is in the cvs module rtems-testing .. but Chris has started some Python pretty printers 2010-06-29T10:24:04 the one in the rtems-testing module is the old one in gdb "script"? 2010-06-29T10:25:12 yes 2010-06-29T10:25:22 ah thanks! 2010-06-29T10:26:03 I'll start with the old one (no time for experiments now) and try out the pyhthon thing later 2010-06-29T10:41:39 ok my new freshly compiled lm32-gcc-rtems4.11 seems to be working :) 2010-06-29T10:43:16 *** groleo has quit IRC 2010-06-29T10:44:18 rtems-gdb-macros already solved the mystery :-) 2010-06-29T10:56:33 sebhub:hi, i find that i havenot configure the sptest compiled. if i want to compiled the sptests into rtems, should i use --enable-tests=sptests configure? 2010-06-29T10:57:36 DrJoel < a user using the framebuffer api wanting to change a pixel value, what should he do ? 2010-06-29T10:57:52 a fd = open("/dev/fb0", mode); and write(fd); ? 2010-06-29T11:00:15 Fallenou: that's a damn good question. I think so. I have never used that. An example for a driver is pc386/console/fb_vga.c for the basic 16 color vga 2010-06-29T11:01:01 i took the framebuffer doc as example 2010-06-29T11:01:16 http://rtems.org/onlinedocs/doc-current/share/rtems/html/bsp_howto/bsp_howto00136.html 2010-06-29T11:01:25 thanks for the exampel 2010-06-29T11:01:42 DrJoel: is there some testsuit program that is using vga ? 2010-06-29T11:01:57 it would be clearer in my mind if i can see how the end user would use the API 2010-06-29T11:05:55 it should be close to the linux driver.. microwindows in the rtems-graphics-toolkit is probably the best example and that's large.. 2010-06-29T11:06:06 it would be nice to have something in examples-v2 that used it. 2010-06-29T11:06:12 *** sebhub has quit IRC 2010-06-29T11:14:03 *** DrJoel has quit IRC 2010-06-29T11:35:45 *** madrazr has joined #rtems 2010-06-29T11:39:53 *** tuxmaniac has quit IRC 2010-06-29T11:43:34 *** DrJoel has joined #rtems 2010-06-29T11:43:35 *** ChanServ sets mode: +o DrJoel 2010-06-29T11:46:40 DrJoel: hi 2010-06-29T11:47:19 hi zwj .. 2010-06-29T11:48:44 in the sptests i want to make a lib for test case to link, for example in the sptests/spsysinit/ i have sysinittask.c and sysinitfirst.c but i donnot link all them, i just want to link them as a lib for test to link. how to do it? 2010-06-29T11:52:51 DrJoel: at the moment milkymist applications directly modify the backbuffer and do a swap_buffers(); when buffer is ready to be showed on the screen 2010-06-29T11:52:53 look at cpukit/libmd. It is a simple library Makefile. 2010-06-29T11:53:06 i guess that's not what rtems want 2010-06-29T11:53:26 (milkymist uses triple buffering) 2010-06-29T11:53:38 Fallenou: this is more akin to how linux fb console apps operate 2010-06-29T11:53:49 akin ? 2010-06-29T11:53:55 (don't know what that means) 2010-06-29T11:54:34 like.. similar ...akin(p): similar in quality or character; 2010-06-29T11:54:36 *** gedare has joined #rtems 2010-06-29T11:54:53 ok :) 2010-06-29T11:57:43 but i guess we can do like this : setpix(buff, x, y, colour) { lseek(fd, x + y*hsize, SEEK_SET); write(fd, colour, sizeof(colour)); } 2010-06-29T11:57:55 it would allow to set a pix to a coordinate 2010-06-29T11:57:55 DrJoel: i can use directly noinst_LIBRARIES = libsysinit.a libsysinit_a_SOURCES=sysinittask.c libsysinit_a_SOURCES=sysinitfirst.c. but how my test find this lib? 2010-06-29T11:58:05 and use the common open close read write posix calls 2010-06-29T11:58:34 supposing that the file descriptor gives access to the framebuffer array 2010-06-29T12:01:25 * DrJoel hopes that is the case.. check for a linux example.. once you get it working, we need to try it on pc386 2010-06-29T12:08:28 DrJoel: you are telling this to me ? 2010-06-29T12:10:09 seems pretty clear to me now that the read and write functions directly read from/write to the framebuffer array 2010-06-29T12:10:19 with the current offset 2010-06-29T12:10:45 so it's optimized for sequential access, not random direct access 2010-06-29T12:11:50 I guess so. You would have to seek to change the position 2010-06-29T12:11:57 afternoon 2010-06-29T12:12:46 that's strange to me to consider the framebuffer array as a file descriptor but if rtems's api is like this, i guess i have no choice 2010-06-29T12:14:39 but who does the swap_buffers(); ? the user application ? to control the display more precisly ? 2010-06-29T12:14:49 with ioctl or sync() flush() functions 2010-06-29T12:14:52 we copied the linux one that is used by mincrowindows.. 2010-06-29T12:15:00 hey gedare 2010-06-29T12:15:35 ok, but it seems to be for simple buffered framebuffers am i right ? 2010-06-29T12:15:50 yes 2010-06-29T12:16:06 so i'm wondering where to put the swap_buffers(); calls 2010-06-29T12:16:11 maybe let the user do them ... 2010-06-29T12:17:33 for now that would be OK 2010-06-29T12:19:50 ok ! 2010-06-29T12:19:57 i will try to make that work for now 2010-06-29T12:20:30 there will be time to think about a better implementation when it will be working with a simple one :p 2010-06-29T12:20:37 * Fallenou gets back to work 2010-06-29T12:21:08 DrJoel: hi, did you see my message? 2010-06-29T12:21:59 zwj: sorry missed it.. on phone.. let me answer it after off 2010-06-29T12:22:18 ok, thank you 2010-06-29T12:26:49 Error: undefined symbol `SIZEOF_VOID_P' in operation 2010-06-29T12:26:53 for crying out loud! :p 2010-06-29T12:29:55 mmmm, looks like it didn't get commited for i386 / pc386 BSP 2010-06-29T12:32:02 needs that autoconf macro for the i386 BSPs in configure.ac... Ralf and I are going back and forth over that now :) 2010-06-29T12:33:31 yeah i saw 2010-06-29T12:33:44 would it be more palatable to have a define in cpu.h? 2010-06-29T12:34:53 I don't know.. he needs to make a positive suggestion .. so far I haven't seen one.. just criticism 2010-06-29T12:34:54 *** Raj_ has joined #rtems 2010-06-29T12:34:59 yah 2010-06-29T12:35:08 well i get the feeling he hasn't really bothered to read what you are doing 2010-06-29T12:35:18 zwj: add libXXX.a as a dependency to the executable .. 2010-06-29T12:35:28 hard to make a suggestion on how to do it when you don't know why it is done :p 2010-06-29T12:35:48 Hello, DrJoel 2010-06-29T12:35:51 I think adding it everywhere sucks .. 2010-06-29T12:35:52 i'm going to reply to the thread and ask about using a define in cpu.h 2010-06-29T12:35:57 hey Raj_ 2010-06-29T12:36:14 I would rather it be autoconf'ed and placed in cpuopts.h and then the problem would be fixed. 2010-06-29T12:36:37 Use autoconf to generate it. But instead of into config.h .. cpuopts.h and then it is available automatically 2010-06-29T12:36:39 as long as it can get somewhere and only has to be defined once per architecture it would be easier 2010-06-29T12:36:49 DrJoel: regarding writing tests for utimes and lchown 2010-06-29T12:36:52 yes but this is something autoconf already kows about 2010-06-29T12:36:54 that's magic that I don't know :) 2010-06-29T12:36:56 yeah 2010-06-29T12:37:14 Where do I put the tests ? I don't see specific tests for utime or chown 2010-06-29T12:37:16 i think the solution you have is decent, but it does touch a lot of the autoconf files 2010-06-29T12:37:37 Raj_: are those part of posix specification? 2010-06-29T12:37:53 Yes 2010-06-29T12:38:11 add them to psxtests but add them to the part in Makefile.am that is built all the time not conditional on POSIX being enabled. 2010-06-29T12:38:12 Raj_: probably make posix tests (psxtest) 2010-06-29T12:38:19 I see psxstat tests contain chown 2010-06-29T12:38:42 well, if they're going to be built all the time, why not use sptests? 2010-06-29T12:38:51 Also use "mktest" from rtems-testing for instantiating a new test. It will save you trouble 2010-06-29T12:39:28 building "some" psxtests when posix is disabled is a little confusing to me 2010-06-29T12:40:24 gedare: posix disable is really only about threading related operations.. many posix routines are assumed to always be there since they are also defined in the C library standards. But mostly it is threading you are disabling 2010-06-29T12:40:25 Shall I send in the patches for utimes and lchown now and the tests later, or it would be better together ? 2010-06-29T12:40:45 i see. 2010-06-29T12:41:20 ../../../../../pc386/lib/librtemsbsp.a(irq_asm.o):../../../../../../../../rtems/c/src/lib/libbsp/i386/pc386/../../i386/shared/irq/irq_asm.S:144: undefined reference to `rtems_per_cpu_info' 2010-06-29T12:41:45 DrJoel: After I updated my git repo of rtems, I see many files modified to versions older than May 2010, is it because v 4.10 has been fixed ? 2010-06-29T12:41:57 Renamed to _Per_CPU_Information 2010-06-29T12:42:02 Raj_: yes.. branched about then 2010-06-29T12:42:15 then cvs is out-of-date 2010-06-29T12:42:19 4.11 is moving fast.. develop against head 2010-06-29T12:42:28 4.10 is forked 2010-06-29T12:42:47 Ohh ok. 2010-06-29T12:43:02 Any idea how I can do that without creating a new git clone ? 2010-06-29T12:43:21 * DrJoel is not good with git 2010-06-29T12:43:30 DrJoel: it looks like the i386 bsp percpu code is out-of-date. 2010-06-29T12:44:23 DrJoel: It's ok. Will figure that out. 2010-06-29T12:45:03 Yea, I had sent a mail regarding some doubts about configuration 2010-06-29T12:45:18 DrJoel: I just had to add the sizeof check and fix the rename to _Per_CPU_Information to get pc386 to build. 2010-06-29T12:45:30 do you want a patch or do you have that code somewhere 2010-06-29T12:45:30 gedare: committing that now 2010-06-29T12:45:33 OK 2010-06-29T12:46:47 hmmm..ticker isn't running, not sure if it's the percpu or my changes though. rebuilding from a clean CVS. 2010-06-29T12:47:16 I will be building that shortly.. it was running 2010-06-29T12:48:05 DrJoel: I hope you got my mail regarding configuration options ? 2010-06-29T12:49:23 I saw a lot of mail to the go lists .. is this separate? what was the subject? 2010-06-29T12:50:10 gedare: starting a pc386 build 2010-06-29T12:50:18 Ok 2010-06-29T12:50:24 No it was not sent to the go list, it was just sent to you. 2010-06-29T12:51:07 DrJoel: Doubts about testing and patches - this is the subject line 2010-06-29T12:51:26 ok.. 2010-06-29T12:52:56 *** arvind_khadri has joined #rtems 2010-06-29T13:01:12 lekernel: milkymist is still using rgb565 pixels ? 2010-06-29T13:02:27 DrJoel, hi 2010-06-29T13:04:00 Fallenou: when you have something working at all, we can try the microwindows demos. You can try playing minesweeper or displyaing the rtems logo 2010-06-29T13:04:11 hey arvind.. any luck on the net demos 2010-06-29T13:04:24 DrJoel: ok good idea :) 2010-06-29T13:07:03 yes 2010-06-29T13:07:38 DrJoel: looks like the sizeof check still didn't get in to the pc386 BSP 2010-06-29T13:08:06 I am about to commit adding __RTEMS_SIZE_VOIDP__ to cpuopts.h and avoid this crap 2010-06-29T13:09:09 oh ok 2010-06-29T13:11:25 DrJoel, no, i tried to figure out how to build it, but couldnt get any luck 2010-06-29T13:12:40 gedare: help arvind for a sec so I can fix the code :) 2010-06-29T13:13:04 let's see... think i built net demos once, but i don't think i ever got them to run 2010-06-29T13:13:10 arvind_khadri: how far have you gotten? 2010-06-29T13:13:35 * gedare isn't sure what arvind's trying to do either 2010-06-29T13:14:14 gedare, i just have the module from the cvs, thats it ... i am just trying to build demos , so that i get a exe for telenetd which then i can run with qemu-nic :) 2010-06-29T13:14:34 oh ok. 2010-06-29T13:14:47 let me update my module, make sure nothing has changed much 2010-06-29T13:15:04 IIRC, the first trick is to point the makefile to a directory in which you have already built RTEMS 2010-06-29T13:16:40 so the first thing you want to do is to define at the top of Makefile two variables 2010-06-29T13:16:54 RTEMS_ROOT=/opt/rtems-4.11 2010-06-29T13:17:11 (if that is where you install rtems-4.11 (assuming you are using 4.11!) 2010-06-29T13:17:26 am on 4.10 iirc ... 2010-06-29T13:17:56 will you be here for a while ? i need to fire up the vm 2010-06-29T13:17:59 ok, well that should work fine 2010-06-29T13:18:07 yeah, and maybe drj will be back to help by then :p 2010-06-29T13:19:14 nice :) 2010-06-29T13:21:04 although i might disappear to watch soccer .:) 2010-06-29T13:22:00 ticker works off CVS (after correcting the sizeof thing) ... guess it is my error 2010-06-29T13:22:04 *sigh* 2010-06-29T13:25:15 DrJoel: Should I work with 4.11 or 4.10 ? 2010-06-29T13:27:46 raj_: I think for the gsoc projects, they prefer us to work against 4.11 2010-06-29T13:27:55 so that our work will be easier to merge with the "state-of-the-art" 2010-06-29T13:28:52 gedare: Thanks. Yea, that seems reasonable. 2010-06-29T13:30:45 awk - soccer:) 2010-06-29T13:31:51 Raj_: 4.11 2010-06-29T13:32:49 DrJoel: Ok, 2010-06-29T13:37:27 gedare, which is the second variable i need to define 2010-06-29T13:43:07 RTEMS_MAKEFILE_PATH=PREFIX_ON_RTEMS_CONFIGURE/TARGET/BSP/ 2010-06-29T13:48:34 DrJoel, erm, just paste that or they will have some values :S 2010-06-29T13:49:15 PREFIX_ON_RTEMS_CONFIGURE is the --prefix option you set during rtems configure 2010-06-29T13:49:33 TARGET is target cpu , eg:- i386 2010-06-29T13:49:58 BSP is board support package eg:- pc386 2010-06-29T13:50:35 gedare: sis, psim and pc386 all run ticker now for me.. about to start committing 2010-06-29T13:50:36 All the actual values depend on your configuration options during rtems install 2010-06-29T13:50:50 back. Ok DrJoel; pretty sure the ticker issue is my fault. 2010-06-29T13:51:44 or at least a combination of percpu changes and my gsoc changes. 2010-06-29T13:52:17 I know I have screwed up once or twice with these mods though :) 2010-06-29T13:52:45 *** exception13 has joined #rtems 2010-06-29T13:53:11 *** dr__house has quit IRC 2010-06-29T13:53:55 all committed 2010-06-29T13:55:09 Raj_: which email again :-D 2010-06-29T13:56:00 DrJoel: The one with subject line - Doubts about testing and patches :) 2010-06-29T13:58:05 ohh.. June 24.. that was like before yesterday :> 2010-06-29T14:01:25 mmm, can't use _Thread_Executing directly in gdb anymore :) 2010-06-29T14:03:13 Raj_: hope the reply helps. 2010-06-29T14:03:23 gedare: that was the one big downside to this that I saw. LOL 2010-06-29T14:05:43 DrJoel: The problem with Go is the goroutine/thread stack 2010-06-29T14:06:33 Some tests have around 300 - 400 goroutines and we need to setup stack for each posix thread, currently, I set a min of 8KB (2 * 4 KB) 2010-06-29T14:06:53 So the total memory reaches to 3 - 4 MB easily just by setting up the stack. 2010-06-29T14:07:56 Let's set the baseline for testing at 8K and see how many fail with it that small. For Ada, I know we -DXXX the stack size on a few tests to make it larger 2010-06-29T14:10:35 DrJoel: Yes, I ran the tests again with 8K stack and 400 posix threads max, only around 20 tests fail out of around 1500 2010-06-29T14:10:54 So that is good. I am currently going through the failures individually 2010-06-29T14:11:05 oops, I'm not allocating space for the ready queue, and wind up using whatever lives at 0x0 :p 2010-06-29T14:11:48 Raj_: That's the best thing.. find a value that runs nearly everything .. then tweak it up or down for the few that can't run at that. 2010-06-29T14:12:09 Yes, exactly what I am doing :) 2010-06-29T14:40:38 yay bug fixed 2010-06-29T14:41:04 watch the VC mailing list for the smoke 2010-06-29T14:41:21 I think my change to cpukit/configure.ac was pretty mild. 2010-06-29T14:41:23 haha yeah. 2010-06-29T14:41:46 it seems Ralf has just an issue with using #define'd values in the source for this kind of information 2010-06-29T14:42:01 regardless of how those values are generated 2010-06-29T14:43:24 now... what was i planning to do today ... 2010-06-29T14:43:33 :) 2010-06-29T14:45:54 DrJoel: Any good reason that percpu.h doesn't include thread.h? 2010-06-29T14:46:21 brings us back to the question of why put the Thread_Control forward define in :) 2010-06-29T14:46:26 gedare: yes.. thread is a "high" .h ... percpu.h is VERY VERY low and we don't want cycles 2010-06-29T14:46:55 ah, i see. 2010-06-29T15:10:23 DrJoel: when linking with my test application , i got a undefined reference to Init() 2010-06-29T15:10:36 in the init.o of my application 2010-06-29T15:11:09 i'm using POSIX_Init 2010-06-29T15:13:55 Look at the CONFIGURE_XXX defines.. 2010-06-29T15:14:25 #define CONFIGURE_POSIX_INIT_THREAD_TABLE 2010-06-29T15:14:25 #define CONFIGURE_MAXIMUM_POSIX_THREADS 1 2010-06-29T15:14:27 i've looked at the defines used by the othersposix samples 2010-06-29T15:14:30 for POSIX init thread 2010-06-29T15:14:54 i have it 2010-06-29T15:14:57 but at the end 2010-06-29T15:15:03 should i put this at the beginning ? 2010-06-29T15:15:13 Before or after CONFIGURE_INIT? 2010-06-29T15:15:31 i don't know 2010-06-29T15:15:49 i havn't quite understood yet the importance of the defines locations 2010-06-29T15:15:53 The simplest example is in examples-v2/hello/posix_hello_world 2010-06-29T15:16:01 ok thanks :) 2010-06-29T15:16:52 rtems/confdefs.h instantiates a number of tables.. it DEFaults the CONFiguration parameters you don't specify. So you override the ones you care about by defining variables, then define CONFIGURE_INIT to say instantiate the tables, then include the .h 2010-06-29T15:18:08 ok my CONFIGURE_INIT was at the beginning 2010-06-29T15:18:11 it's maybe the problem 2010-06-29T15:18:29 well ... maybe not 2010-06-29T15:18:44 pastebin the stuff 2010-06-29T15:20:04 http://pastebin.com/RAWmSduP 2010-06-29T15:21:36 it actually does nothing, but as soon as this runs (it should, as soon as it compiles) i will try to write some colours and it should work :) 2010-06-29T15:22:58 it doesn't compile for me .. hold on.. 2010-06-29T15:23:22 it doesn't compile for me neither 2010-06-29T15:23:24 that's my problem :) 2010-06-29T15:23:39 init.o:(.data+0x2c): undefined reference to `Init' 2010-06-29T15:23:54 everything else is ok 2010-06-29T15:24:05 i'm sure it's just a problem of include or define or function name 2010-06-29T15:24:27 got it.. just seeing how it goes :-D 2010-06-29T15:24:35 but since i'm newbie ... 2010-06-29T15:24:39 hard to spot 2010-06-29T15:24:50 it seems like all the others posix samples ... 2010-06-29T15:25:42 close but there is a subtle difference.. let me paste it back to you 2010-06-29T15:25:58 ok thanks :) 2010-06-29T15:27:06 now runs on pc386 2010-06-29T15:27:29 great ! :) 2010-06-29T15:27:38 * Fallenou excited to print some red colour 2010-06-29T15:27:40 *** tuxmaniac has joined #rtems 2010-06-29T15:28:14 http://pastebin.com/Wb5B2udL 2010-06-29T15:28:28 I screwed up the comment block at the top but ... 2010-06-29T15:28:37 + added an assert if the open doesn't work 2010-06-29T15:29:09 + deleted CONFIGURE_MAXIMUM_TASKS and Classic API Tasks Init thread table being configured (source of error) 2010-06-29T15:29:25 + added CONFIGURE_MAXIMUM_FILE_DESCRIPTORS 4 since you need one more to open the FB. 2010-06-29T15:29:28 With that open works 2010-06-29T15:29:32 on the pc386 2010-06-29T15:30:01 oh ok 2010-06-29T15:30:10 but the init error ? :o 2010-06-29T15:30:41 that was having defines to have the class api init task table.. and no "Init" thread. You can actually have both.. see the "both hello world" example 2010-06-29T15:31:31 ok good i understand now for the init :) 2010-06-29T15:31:50 i have an error for assert now :p i surely missed one .h 2010-06-29T15:33:25 assert.h 2010-06-29T15:33:45 sorry.. I had some undo's accidentally at the last minute apparently 2010-06-29T15:34:53 no no that's my fault :p 2010-06-29T15:35:12 it compiles all well now thanks to you ! 2010-06-29T15:36:48 no problem.. when you have it showing color on your bsp, try it on pc386. Do you have that built? 2010-06-29T15:37:51 it runs fine with -nographic and without -nographic 2010-06-29T15:37:57 i will add a write() to put some color 2010-06-29T15:37:59 and test again 2010-06-29T15:39:51 would there be a reason why accessing data from a structure sent as an argument to a function works fine and accessing that structure sent as an argument to a thread via pthread_create fails ? 2010-06-29T15:41:09 if the structure is on the stack, then it may go away before the thread accesses it 2010-06-29T15:41:28 remember the lectures in college on variable scoping? :-D 2010-06-29T15:42:13 the function from which i create the thread sleeps for a few seconds 2010-06-29T15:42:19 so that shouldn't be an issue 2010-06-29T15:42:37 *** tuxmaniac has quit IRC 2010-06-29T15:43:27 did you cast it back to the original type? 2010-06-29T15:43:41 yep 2010-06-29T15:44:55 http://pastebin.com/ixWfkRHD 2010-06-29T15:45:09 from add_request_to_runlist works 2010-06-29T15:45:22 handle_io no 2010-06-29T15:47:15 ok nothing is happening =) 2010-06-29T15:47:42 Fallenou: on the bright side, nothing wrong is happening :)) 2010-06-29T15:47:53 where is chn declared? 2010-06-29T15:48:39 hum yes it is actually :p 2010-06-29T15:48:41 hi alinrus :) 2010-06-29T15:48:43 that's a piece of code from aio_enqueue_request 2010-06-29T15:49:20 http://pastebin.com/fJ8Wm8pN 2010-06-29T15:49:46 http://pastebin.com/TsaW1d2d 2010-06-29T15:50:21 damn seeing sched_param makes me vomit :p 2010-06-29T15:50:22 Fallenou: does it display anything? 2010-06-29T15:50:26 no nothing at all 2010-06-29T15:50:47 Fallenou: I don't know on that.. qemu-lm32 is for mwalle and lekernel to help with 2010-06-29T15:50:49 mwalle: ping 2010-06-29T15:50:51 lekernel: ping 2010-06-29T15:51:13 i'm gonna check my code first :p 2010-06-29T15:51:28 but the error message is strange 2010-06-29T15:51:31 i don't understand it 2010-06-29T15:51:43 aanjhan: and the first code segment is from the same method 2010-06-29T15:52:27 Fallenou: limitations in simulator.. the questino is .. is it you or the simulator? does a graphics window for vga pop up and you don't change it or does it not pop up 2010-06-29T15:53:23 it's me, the simulator works :) 2010-06-29T15:53:33 yes a vga pops up 2010-06-29T15:53:41 but nothing changes in it 2010-06-29T15:53:44 http://pastebin.com/xmN0yF6M it's my code 2010-06-29T15:56:13 you declared a huge array on the stack. 2 x640x480 == 614400 bytes. The stack is probably 8k .. move it outside the task. or just write the memory a smaller amount at a time in a loop (preferable) stack is precious in embedded systems and you can't do stuff like this 2010-06-29T15:56:55 ok i can do 640x480 writes in a loop :) 2010-06-29T15:57:18 but yes it looks like an overflow somewhere 2010-06-29T15:58:46 600K is a bit more than the 4-8k most init tasks have. 2010-06-29T15:59:31 ^^ 2010-06-29T15:59:56 ok now i don't have the error messages anymore 2010-06-29T16:01:02 ok I left out only the relevant parts http://pastebin.com/53Qpqjww would you mind taking a look at this 2010-06-29T16:02:23 is chn ever set? I only see it dereferenced. 2010-06-29T16:02:31 Fallenou: and do you have read? 2010-06-29T16:02:35 red I mean 2010-06-29T16:03:09 yes it is set in aio_enqueue_request and I checked the values before calling add_to_runlinst and handle_io 2010-06-29T16:03:19 the values match add_to_runlist 2010-06-29T16:03:34 but the file descriptor is set to 0 in handle_io 2010-06-29T16:04:48 I would have to see a complete running example to debug it. :( 2010-06-29T16:05:20 no, no red 2010-06-29T16:05:23 all is black 2010-06-29T16:06:27 did you fill in the buffer with the right values? 2010-06-29T16:07:17 did you look at the write in the pc386 version. If uses the file descriptor offset field to know where to write 2010-06-29T16:08:36 hum no i didn't look at the pc386 for the write 2010-06-29T16:08:50 i'm trying some more tests and i'm gonna have a look 2010-06-29T16:11:48 it calculates an offset into the buffer.. every write advances in the "file" 2010-06-29T16:12:01 * DrJoel needs to logoff and wrap up so I can get home... see you all alter 2010-06-29T16:12:03 later 2010-06-29T16:12:21 alinrus: if you need to, bundle up me a buildable example and send it to me.. I will try to figur eit out 2010-06-29T16:12:25 *** DrJoel has quit IRC 2010-06-29T16:12:37 *** madrazr has quit IRC 2010-06-29T16:26:08 *** Raj_ has quit IRC 2010-06-29T16:29:18 Fallenou, i think you wrote to the flash address space 2010-06-29T16:32:30 your framebuffer shouldn't be on the stack, make a static global variable, use the heap (no idea if this is a good idea) or use a dedicated ram space (you would have to modify the linker script) 2010-06-29T16:34:10 ah forget the last sentence, you are writing to the fb from within your application, or so you should write small pieces :) 2010-06-29T16:34:57 open(fb), for (..) write(fb, 0xff, 1) 2010-06-29T16:39:53 Fallenou, why is your driver writing to the flash address space? null pointer dereference? 2010-06-29T16:47:51 no just a stack oveflow 2010-06-29T16:47:55 i corrected it 2010-06-29T16:47:59 now it just does nothing 2010-06-29T16:48:18 where do you write? 2010-06-29T16:48:20 backbuffer? 2010-06-29T16:48:24 lastbuffer 2010-06-29T16:48:50 well, write to the frontbuffer then 2010-06-29T16:49:02 or to the backbuffer if you have triple-buffering enabled 2010-06-29T16:49:10 but you never should write to lastbuffer 2010-06-29T16:49:54 ok i will first try to write to the frontbuffer 2010-06-29T16:50:03 i'm pasting the code anyway 2010-06-29T16:50:59 http://pastebin.com/JKp2u0hE 2010-06-29T16:51:28 huh? 2010-06-29T16:51:48 vga_swap_buffers() in the app? 2010-06-29T16:51:59 use ioctl 2010-06-29T16:52:13 and if you write to the frontbuffer you should not swap buffers 2010-06-29T16:52:13 to call it through an ioctl ? 2010-06-29T16:52:18 well of course 2010-06-29T16:52:23 ok 2010-06-29T16:52:34 but for the test i call it directly 2010-06-29T16:52:40 but begin with a driver without triple buffering 2010-06-29T16:52:45 ok 2010-06-29T16:52:59 well you'll need to get that ioctl done... so do it 2010-06-29T16:52:59 i am changing to write directly to the frontbuffer 2010-06-29T16:54:02 and this is assuming RTEMS does not already have a standard API to handle multiple buffering, in this case make your driver compatible with it 2010-06-29T16:55:04 i do not think so 2010-06-29T16:55:14 the fb_fix just has a pointer to the frontbuffer 2010-06-29T16:55:21 but i will check 2010-06-29T16:55:29 look for it... it's part of the job :) 2010-06-29T16:56:42 sure 2010-06-29T16:59:58 i tried with only a frontbuffer it doesn't work 2010-06-29T17:00:11 do you want me to commit & push ? 2010-06-29T17:02:09 Fallenou, have you already tried to write directly to the frontbuffer? (just to make sure something happens) 2010-06-29T17:02:53 not directly, but the code should do it 2010-06-29T17:03:54 for some reason the open() function i wrote is never called 2010-06-29T17:04:19 so i put the write to the MM_VGA_RESET register in the initialization function 2010-06-29T17:04:23 which is called 2010-06-29T17:04:36 i am wondering if the write is really called, i'm gonna check 2010-06-29T17:04:57 Fallenou, btw you could enable the debug flag in milkymist-vgafb.c 2010-06-29T17:05:18 ok write is never called :) 2010-06-29T17:05:55 i'm checking with pc386 bsp which has vga 2010-06-29T17:05:56 i have to go to bed now, you can reach me on mail 2010-06-29T17:06:03 ok thank you :) gn8 ! 2010-06-29T17:11:25 *** exception13 has left #rtems 2010-06-29T18:24:35 gn8 2010-06-29T19:29:20 *** gedare has quit IRC 2010-06-29T19:40:32 *** alinrus has left #rtems 2010-06-29T20:02:41 *** methril__ has quit IRC 2010-06-29T20:09:39 *** kiwichris has joined #rtems 2010-06-29T20:19:23 *** methril__ has joined #rtems 2010-06-29T20:19:41 Rafael ? 2010-06-29T21:04:19 hi 2010-06-29T21:04:29 hi 2010-06-29T21:04:39 nice to see you :) 2010-06-29T21:04:49 i was going to go to sleep 2010-06-29T21:05:48 Ok. I am looking at the project set up now 2010-06-29T21:05:54 nice 2010-06-29T21:06:12 i'm doing some work on this at work 2010-06-29T21:06:27 and i hope to have a USBDM in the next weeks 2010-06-29T21:06:55 Which protocol ? 2010-06-29T21:07:37 did you refer to USBDM? 2010-06-29T21:37:08 *** bubaflub has quit IRC 2010-06-29T22:26:24 *** dr__house has joined #rtems 2010-06-30T01:41:49 *** groleo has joined #rtems 2010-06-30T01:46:30 *** _Lucretia_ has quit IRC 2010-06-30T01:50:43 *** arvind_khadri has quit IRC 2010-06-30T02:05:12 *** sebhub has joined #rtems 2010-06-30T02:06:56 good morning 2010-06-30T02:19:31 *** alinrus has joined #rtems 2010-06-30T02:20:48 rtems failed to build after a cvs update 2010-06-30T02:20:49 fpaste.org/X6fw/ 2010-06-30T02:21:10 testsuites/samples/capture fails 2010-06-30T02:28:20 *** madrazr has joined #rtems 2010-06-30T02:29:35 *** tuxmaniac has joined #rtems 2010-06-30T02:35:59 kiwichris: please add assert.h to rtems/cpukit/libfs/src/imfs/memfile.c 2010-06-30T03:00:55 *** tuxmaniac has quit IRC 2010-06-30T03:05:58 *** madrazr has quit IRC 2010-06-30T03:52:50 *** alinrus has quit IRC 2010-06-30T04:11:29 hi 2010-06-30T04:21:32 *** _Lucretia_ has joined #rtems 2010-06-30T04:22:02 *** exception13 has joined #rtems 2010-06-30T05:46:46 *** tonylijo has joined #rtems 2010-06-30T05:49:01 *** madrazr has joined #rtems 2010-06-30T05:50:15 *** methril_work has quit IRC 2010-06-30T05:53:55 *** methril_work has joined #rtems 2010-06-30T05:55:01 *** tonylijo has quit IRC 2010-06-30T06:23:38 *** madrazr has quit IRC 2010-06-30T07:24:59 *** madrazr has joined #rtems 2010-06-30T07:29:09 *** kiwichris has quit IRC 2010-06-30T07:45:25 *** alinrus has joined #rtems 2010-06-30T07:47:54 *** groleo has quit IRC 2010-06-30T08:33:50 sebhub: hi 2010-06-30T08:43:17 alinrus: how are things at your end? project is going good? 2010-06-30T08:46:29 I think so, I'm waiting for a review any day now :) 2010-06-30T08:46:30 sebhub: i have write a test case for sysinit, and test the sysinit api and find a error in api code. now i have correct it . send it to your email and help me comment it 2010-06-30T08:50:03 *** bubaflub has joined #rtems 2010-06-30T08:50:46 aanjhan: I've added you to the google code repo if you want to take a look :) 2010-06-30T08:58:02 alinrus: oh ok. Will take a look 2010-06-30T09:01:04 hi 2010-06-30T09:01:30 ok, i will look at the mail 2010-06-30T09:16:37 *** lekernel has quit IRC 2010-06-30T09:26:50 *** DrJoel has joined #rtems 2010-06-30T09:26:50 *** ChanServ sets mode: +o DrJoel 2010-06-30T09:32:59 *** exception13 has left #rtems 2010-06-30T09:33:18 *** lekernel has joined #rtems 2010-06-30T09:43:29 *** Raj_ has joined #rtems 2010-06-30T09:49:31 *** lekernel has quit IRC 2010-06-30T09:52:48 DrJoel: hi 2010-06-30T09:53:23 Hey Raj_ 2010-06-30T09:53:45 So, I was thinking of adding the utimes test to psx13 2010-06-30T09:53:52 in psxtests 2010-06-30T09:54:18 and lchown to psxstat 2010-06-30T09:54:59 This psx13 has a test for utime, so basically the utimes test will be based on that 2010-06-30T09:55:11 that makes sense for psx13 2010-06-30T09:55:14 why psxstat? 2010-06-30T09:56:38 psxstat because I see it uses chown and creates symlinks 2010-06-30T09:56:54 so I was thinking of doing lchown on those symlinks and testing 2010-06-30T09:57:52 very logical. Just submit the code and tests please 2010-06-30T09:58:16 DrJoel: Have some doubts though 2010-06-30T09:58:26 what are they? 2010-06-30T09:58:29 These scn files are I guess the expected output ? 2010-06-30T09:59:22 So I just run my test and copy the output (if all is well) to the .scn file ? 2010-06-30T10:00:30 yes. remove simulator and BSP specific output.. make sure it is a UNIX format file without ^M at the end of the line. 2010-06-30T10:00:38 If there is a .doc file, update it as needed 2010-06-30T10:01:15 Ok. And there is no lstat in psxstat 2010-06-30T10:01:17 *** lekernel has joined #rtems 2010-06-30T10:02:00 Shall I add lstat to the psxstat test, for lchown ? 2010-06-30T10:02:24 please.. it is in the list of symbols with no tests at the moment 2010-06-30T10:02:31 as is lstat_r 2010-06-30T10:02:42 sebhub: hi, i have send it to you. now i must go to bed. byb 2010-06-30T10:02:47 cu 2010-06-30T10:03:04 i look at it tomorrow morning 2010-06-30T10:04:15 DrJoel: So, you are ok with lstat test going into psxstat. No need of creating psxlstat right ? 2010-06-30T10:05:44 *** zwj has quit IRC 2010-06-30T10:05:56 hmmm... it won't add much code will it? That is a large test already. Shouldn't require more than a couple of new test cases, right? 2010-06-30T10:07:19 Yea, I can reduce duplication by just making common functions with an option to call lstat/stat and similarly for lchown/chown 2010-06-30T10:07:31 so shouldn't add much code 2010-06-30T10:11:05 DrJoel: another doubt is regarding the lchown code. 2010-06-30T10:11:49 As you advised, I added a common function with an option to follow the link or not. 2010-06-30T10:11:59 ok 2010-06-30T10:12:16 Since, this function will be a global default symbol, I have used __attribute__ visibility hidden 2010-06-30T10:12:23 is that ok ? 2010-06-30T10:12:56 We dont want the function exposed right ? 2010-06-30T10:13:35 We don't have dll's .. does it matter otherwise? 2010-06-30T10:13:49 Just make it start with _ so like _chown_helper 2010-06-30T10:14:49 If it's a global default, it can be called from user code right ? 2010-06-30T10:15:36 Well, it doesn't matter much for this, I just thought I would discuss that. Sure I can start it with a _ to indicate it's for internal use. 2010-06-30T10:16:06 yes but by convention everyone knows 2010-06-30T10:16:19 rtems is all static linkage for now so... and we avoid attributes :-D 2010-06-30T10:16:20 Ok. Good. 2010-06-30T10:16:39 Another thing is I remember you mentioning something about mktest 2010-06-30T10:16:54 I couldn't see a file by that name though 2010-06-30T10:17:39 it is in rtems-testing cvs module.. fairly recent.. and a directory named rtems-test-templates to instantiate a test from. But that doesn't matter since you are adding to existing tests 2010-06-30T10:19:14 DrJoel: By the way, I don't see rtems-testing in the private git repo 2010-06-30T10:22:28 It isn't mirrored yet... :( 2010-06-30T10:23:23 Ok. :) 2010-06-30T10:26:08 *** sebhub has quit IRC 2010-06-30T10:40:44 *** gedare has joined #rtems 2010-06-30T10:42:25 * gedare wants to make a bugzilla account named Janitor 2010-06-30T10:42:35 *ducks* 2010-06-30T10:47:29 DrJoel: I'm having trouble using an inline function from one .inl file in another .inl file ... 2010-06-30T10:59:13 stupid header file dependencies 2010-06-30T11:00:17 from where to where? and what dependencies 2010-06-30T11:00:44 it's in my new scheduler code. I have a readyq.inl with a first() inline routine, and scheduler.h that tries to use first() 2010-06-30T11:00:55 er, scheduler.inl that tries to use first() 2010-06-30T11:01:43 readyq.h includes readq.inl.. and scheduler.inl (not scheduler.h) needs to include readyq.h? 2010-06-30T11:01:55 but readyq.inl includes scheduler.h, which I think is causing problesm maybe 2010-06-30T11:02:08 yep.. can you avoid that? :-D 2010-06-30T11:02:11 yes, but scheduler.h was including readyq.h too 2010-06-30T11:02:14 ok 2010-06-30T11:02:33 yeah, it was for a variable definition, but i can move that to readyq.h to isolate it more 2010-06-30T11:02:51 circular dependencies need to be avoided .. RTEMS tries to be strictly hierarchical 2010-06-30T11:02:55 like i said... " stupid header file dependencies" 2010-06-30T11:02:55 :) 2010-06-30T11:02:57 yeah 2010-06-30T11:03:02 i'm seeeeing that 2010-06-30T11:03:56 It is hard at times to find a way to maintain that structure but it is easier overall than battling things 2010-06-30T11:04:24 yeah and once it is set up, it is easier to understand 2010-06-30T11:04:32 except you wind up with weird things like tqdata.h :) 2010-06-30T11:05:07 which actually isn't that bad... separting data structure definitions from function prototypes can be quite handy 2010-06-30T11:05:30 anywho, back to trying to get my code to compile 2010-06-30T11:05:32 That's the Ada legacy on tqdata 2010-06-30T11:06:21 ah, i assumed it was because of thread.h not wanting to include files that need Thread_Control defined. :) 2010-06-30T11:06:46 That is part of it also but the package had to be split in Ada 2010-06-30T11:06:51 It is better to be clean.. 2010-06-30T11:06:59 That's what momma always said 2010-06-30T11:20:05 DrJoel: I define _Thread_Ready_queue in readyq.h, but am getting undefined references to it when I try to pass it as parameter to readyq.inl functions from some .c files (which include readyq.h) -- any thoughts? 2010-06-30T11:22:06 SCORE_EXTERN Ready_queue_Control _Thread_Ready_queue; 2010-06-30T11:22:23 If you did a "SCORE_EXTERN", then it has to be instantiated by being included like in exinit.c 2010-06-30T11:22:42 not just like in exinit.c.. in exinit.c 2010-06-30T11:22:48 oh. I was trying to find where that was done before for the thread.h ones and didn't see, so I thought there was some magic :) 2010-06-30T11:22:53 haha ok 2010-06-30T11:23:15 got it. 2010-06-30T11:25:58 gotta run .. see you all later 2010-06-30T11:25:59 *** DrJoel has quit IRC 2010-06-30T12:06:02 *** gedare has quit IRC 2010-06-30T12:24:08 *** gedare has joined #rtems 2010-06-30T12:29:54 *** lekernel has quit IRC 2010-06-30T12:32:46 *** madrazr has quit IRC 2010-06-30T12:32:57 *** madrazr has joined #rtems 2010-06-30T12:42:22 *** lekernel has joined #rtems 2010-06-30T12:42:45 *** dr__house has quit IRC 2010-06-30T12:59:17 *** Raj_ has quit IRC 2010-06-30T13:02:31 *** gedare has quit IRC 2010-06-30T13:05:17 *** lekernel has quit IRC 2010-06-30T13:17:54 *** lekernel has joined #rtems 2010-06-30T13:24:36 *** methril_work has quit IRC 2010-06-30T14:04:32 *** aniceberg has joined #rtems 2010-06-30T14:41:00 *** madrazr has quit IRC 2010-06-30T15:03:09 *** rtemsLogger has joined #rtems 2010-06-30T15:09:36 *** lekernel has quit IRC 2010-06-30T15:11:57 *** lekernel has joined #rtems 2010-06-30T15:14:34 *** exception13 has joined #rtems 2010-06-30T15:21:07 *** bubaflub has quit IRC 2010-06-30T15:29:19 alinrus: you have a google code repo ? 2010-06-30T15:29:35 where can i set up one ? is it one dedicated for gsoc ? 2010-06-30T15:29:51 code.google.com 2010-06-30T15:30:00 no it's not a gsoc thing 2010-06-30T15:30:07 it has been available for years 2010-06-30T15:30:12 cvs, svn and hg 2010-06-30T15:30:38 code.google.com/projecthosting/ 2010-06-30T15:31:59 oh ok i know this kind of repo 2010-06-30T15:32:08 i thought it was a repo for gsoc 2010-06-30T15:32:43 it's faster than gitorious and github 2010-06-30T15:32:44 :) 2010-06-30T15:33:43 i'm using github :p 2010-06-30T15:34:01 it has many fancy stuff 2010-06-30T15:34:07 like publicating on irc via CIA 2010-06-30T15:34:07 rtems is too big for github 2010-06-30T15:34:08 and on twitter etc 2010-06-30T15:34:17 too big ? :p 2010-06-30T15:34:22 github is nice for small projects 2010-06-30T15:34:32 there are linux repos on github ! 2010-06-30T15:34:44 on paid accounts i guess 2010-06-30T15:34:55 anyway my rtems is submitted on github 2010-06-30T15:35:04 but only my work, not all cvs history 2010-06-30T15:35:23 i don't think it's that important where do you host the code 2010-06-30T15:35:31 i guess you were talking about the entire history when you said it was too big for github 2010-06-30T15:35:40 no obviously not :p 2010-06-30T15:35:49 but it's interesting to compare those tools :) 2010-06-30T15:36:16 i was under the impression that we must use google code that's why i hosted the project there 2010-06-30T15:36:18 :) 2010-06-30T15:36:28 huhu 2010-06-30T15:36:36 you become corporate ! :p 2010-06-30T15:37:00 you got google stickers, google notebook, google pen, google repo :p 2010-06-30T15:37:03 * Fallenou kidding =) 2010-06-30T15:37:14 google homepage :)) 2010-06-30T15:37:19 aouch ! (me too =)) 2010-06-30T15:39:16 Fallenou: did you see this docs.google.com/Doc?docid=0AXgKwYy0qUGZZGhyOHpmdmRfNDdkZmt6M3BnOQ&hl=en&authkey=CIn-9okO 2010-06-30T15:39:21 ? 2010-06-30T15:40:39 yes i saw it this afternoon while buying some jeans :p 2010-06-30T15:40:53 (it's the first day of the prices cut in France) 2010-06-30T15:41:12 i'm doing a report for a project for school right now 2010-06-30T15:41:15 well it's the first day of a +5% VAT in romania 2010-06-30T15:41:16 :)) 2010-06-30T15:41:20 it's for tonight 2010-06-30T15:41:28 but right after that i will submit my progression status :) 2010-06-30T15:41:50 arg +5% VAT ... happy christmas =) 2010-06-30T15:47:39 alinrus, I'm also hosting large repositories on github without problems 2010-06-30T15:47:54 actually they don't enforce strictly the quota for free accounts 2010-06-30T15:48:07 if you have a good reason for using it, they say it's ok 2010-06-30T15:48:27 lekernel: i tried github some time ago, maybe things have changed by now 2010-06-30T15:49:03 if you hate system administration github is nice :) 2010-06-30T15:49:09 a lot of things are push-button 2010-06-30T15:49:23 i have my own git hosting server 2010-06-30T15:49:24 :) 2010-06-30T15:49:38 no fancy buttons 2010-06-30T15:53:00 hehe 2010-06-30T15:59:23 *** cdcs has joined #rtems 2010-06-30T16:04:01 *** lekernel has quit IRC 2010-06-30T16:14:00 *** methril_work has joined #rtems 2010-06-30T16:43:28 *** lekernel has joined #rtems 2010-06-30T16:48:28 *** methril__ has quit IRC 2010-06-30T17:00:46 *** methril__ has joined #rtems 2010-06-30T17:01:44 *** cdcs has quit IRC 2010-06-30T17:30:53 *** exception13 has left #rtems 2010-06-30T18:10:02 *** aniceberg has quit IRC 2010-06-30T18:11:34 *** mwalle has quit IRC 2010-06-30T18:15:23 *** mwalle has joined #rtems 2010-06-30T18:21:33 *** _Lucretia_ has quit IRC 2010-06-30T18:29:00 *** _Lucretia_ has joined #rtems 2010-06-30T18:30:16 *** mwalle has quit IRC 2010-06-30T18:38:56 *** mwalle has joined #rtems 2010-06-30T18:47:40 gn8 2010-06-30T21:19:05 *** dr__house has joined #rtems 2010-07-01T00:45:51 *** dr__house has quit IRC 2010-07-01T00:45:51 *** dr__house has joined #rtems 2010-07-01T01:05:25 *** madrazr has joined #rtems 2010-07-01T01:21:34 *** arvind_khadri has joined #rtems 2010-07-01T01:24:55 *** alinrus has quit IRC 2010-07-01T02:05:58 *** alinrus has joined #rtems 2010-07-01T02:11:25 *** arvind_khadri has quit IRC 2010-07-01T02:29:09 *** sebhub has joined #rtems 2010-07-01T02:38:29 *** madrazr has quit IRC 2010-07-01T02:49:41 good morning 2010-07-01T03:34:05 *** groleo has joined #rtems 2010-07-01T05:08:42 *** rtemsLogger has quit IRC 2010-07-01T05:08:48 *** rtemsLogger has joined #rtems 2010-07-01T05:22:21 *** madrazr has joined #rtems 2010-07-01T05:23:07 Great and thank Sebastian. 2010-07-01T05:23:28 I am heading away on the weekend for a couple of week and this has been distracting. 2010-07-01T05:24:01 i will also back port the mount changes/fixes to the rtems 4.10 branch 2010-07-01T05:24:13 Great. 2010-07-01T05:39:07 kiwichris, it seems that the BDM usb is not working as it should, regarding the /dev/ symlink 2010-07-01T05:39:40 have you set up udev ? 2010-07-01T05:40:54 i used a ln -sf 2010-07-01T05:41:16 Oh. Have you disconnected and reconnected ? 2010-07-01T05:41:21 *** methril_work has quit IRC 2010-07-01T05:41:29 nop 2010-07-01T05:41:49 the device has the same product/vendor if 2010-07-01T05:41:52 id* 2010-07-01T05:41:59 If you have for any reason the link will be wrong which is why you need udev 2010-07-01T05:42:37 i checked with a ls 2010-07-01T05:43:13 I do not know because I am not sure what you have done. 2010-07-01T05:43:19 lrwxrwxrwx 1 root root 20 Jul 1 12:09 /dev/bdmusb0 -> /dev/bus/usb/002/007 2010-07-01T05:43:29 ls -al /dev/bus/usb/002/007 2010-07-01T05:43:29 crw-rw-r-- 1 root root 189, 134 Jul 1 10:11 /dev/bus/usb/002/007 2010-07-01T05:43:41 *** methril_work has joined #rtems 2010-07-01T05:44:04 sudo ./test/bdm-chk /dev/bdmusb0 2010-07-01T05:44:08 this should work 2010-07-01T05:44:09 I have .... 2010-07-01T05:44:27 [chris@kaka sf]$ ls -las /dev/tblcf1 2010-07-01T05:44:28 0 lrwxrwxrwx 1 root root 15 2010-05-08 14:12 /dev/tblcf1 -> bus/usb/003/009 2010-07-01T05:45:37 have you looked in dmesg 2010-07-01T05:46:55 nothing in dmesg 2010-07-01T05:48:12 i think is some hardcoded scheme for the device name 2010-07-01T05:48:28 Sorry ? 2010-07-01T05:49:13 if I try bdm-chk /dev/bdmcf0 it will try LPT 2010-07-01T05:49:22 Yes. 2010-07-01T05:49:33 even though /dev/bdmcf0 is a symlink 2010-07-01T05:50:30 Oh. The whole device name things is hack for driver compatibility with kernel drivers. 2010-07-01T05:50:37 There is no magic 2010-07-01T05:51:18 which part of the bdmcf0; the bdmcf or the 0 2010-07-01T05:51:39 The whole thing. The access is via ioperm 2010-07-01T05:52:06 so one should not use bdmcf0 no matter what 2010-07-01T05:52:11 for people still wondering, I've found my problem :) 2010-07-01T05:52:20 tiemen, great 2010-07-01T05:52:20 unless it's used for LPT 2010-07-01T05:52:41 the custom firmware part we have that's hooked up to the amba-bus uses so much bus-time that the leon doesn't get through 2010-07-01T05:52:57 or not enough anyhow 2010-07-01T05:52:59 That is correct. You need a new name that makes the pod. 2010-07-01T05:53:18 i think this should be documented 2010-07-01T05:53:20 You are using /dev/bdmusb0 so I am lost following what you are doing 2010-07-01T05:53:37 i'll email you 2010-07-01T05:53:52 Please mail the list 2010-07-01T05:54:23 yep 2010-07-01T05:56:32 sebhub, do any test need to be fixed for the NFS change ? 2010-07-01T06:12:41 kiwichris, i found the problem 2010-07-01T06:12:48 great 2010-07-01T06:13:00 the thing is, any device name, matching a pattern like: 2010-07-01T06:13:31 /dev/bdm*, will be tested against LPT 2010-07-01T06:13:54 since you use /dev/tblcf, it didnt matched the patter 2010-07-01T06:13:55 n 2010-07-01T06:14:01 Yeap. It is a historical hack. I suggest /dev/pebdmusb0 2010-07-01T06:14:05 nfs is only used in the netdemos as far as i know 2010-07-01T06:14:30 and this is likely broken due to the recent mount changes 2010-07-01T06:14:37 sebhub, I would grep the tests. I know a broke a few that Joel kindly fixed. 2010-07-01T06:15:12 i run the test suite reguarlry 2010-07-01T06:15:18 Ok 2010-07-01T06:15:42 the boost unit testing framework turned out to be not thread-safe 2010-07-01T06:15:51 Oh 2010-07-01T06:15:55 now i evaluate the google c++ testing framework 2010-07-01T06:20:46 kiwichris, patch submited 2010-07-01T06:20:53 Ok 2010-07-01T06:33:16 *** exception13 has joined #rtems 2010-07-01T06:50:48 *** madrazr has left #rtems 2010-07-01T06:59:10 *** aanjhan has quit IRC 2010-07-01T07:31:35 *** DrJoel has joined #rtems 2010-07-01T07:31:36 *** ChanServ sets mode: +o DrJoel 2010-07-01T07:34:28 joel: did you run the test suite on the mips gdb simulator recently? 2010-07-01T07:34:47 no.. is something failing? 2010-07-01T07:35:04 Chris and I have talked about the next stage in his autobuilder is to run tests and report results 2010-07-01T07:35:12 *** FILE I/O SAMPLE AND TEST *** 2010-07-01T07:35:13 ========================= 2010-07-01T07:35:15 RTEMS FILE I/O Test Menu 2010-07-01T07:35:16 ========================= 2010-07-01T07:35:18 p -> part_table_initialize 2010-07-01T07:35:19 f -> mount all disks in fs_table 2010-07-01T07:35:21 l -> list file 2010-07-01T07:35:22 r -> read file 2010-07-01T07:35:24 w -> write file 2010-07-01T07:35:25 s -> start shell 2010-07-01T07:35:27 Enter your selection ==>mips-core: 4 byte read to unaligned address 0x88071c8d at 0x8800ca6c 2010-07-01T07:35:28 Program received signal SIGBUS, Bus error. 2010-07-01T07:35:30 0x8800ca6c in _Thread_Clear_state (the_thread=0x88071ee0, state=) 2010-07-01T07:35:31 at ../../../../../../../../../rtems-sync/c/src/../../cpukit/score/src/threadclearstate.c:88 2010-07-01T07:35:33 88 if ( the_thread->current_priority < _Thread_Heir->current_priority ) { 2010-07-01T07:35:34 (gdb) where 2010-07-01T07:35:36 #0 0x8800ca6c in _Thread_Clear_state (the_thread=0x88071ee0, state=) 2010-07-01T07:35:37 at ../../../../../../../../../rtems-sync/c/src/../../cpukit/score/src/threadclearstate.c:88 2010-07-01T07:35:39 #1 0x8800ccec in _Thread_Delay_ended (id=, ignored=) 2010-07-01T07:35:41 at ../../../../../../../../../rtems-sync/c/src/../../cpukit/score/src/threaddelayended.c:62 2010-07-01T07:35:42 #2 0x8800eb3c in _Watchdog_Tickle (header=0x8806f294) at ../../../../../../../../../rtems-sync/c/src/../../cpukit/score/src/watchdogtickle.c:95 2010-07-01T07:35:43 #3 0x8800901c in _Watchdog_Tickle_ticks () at ../../cpukit/../../../jmr3904/lib/include/rtems/score/watchdog.inl:100 2010-07-01T07:35:45 #4 rtems_clock_tick () at ../../../../../../../../../rtems-sync/c/src/../../cpukit/rtems/src/clocktick.c:46 2010-07-01T07:35:46 #5 0x880017e0 in Clock_isr (vector=2282168032) 2010-07-01T07:35:48 at ../../../../../../../../../../../rtems-sync/c/src/lib/libbsp/mips/jmr3904/clock/../../../shared/clockdrv_shell.h:66 2010-07-01T07:35:49 #6 0x8805db74 in mips_vector_isr_handlers (frame=0x880777ac) at ../../../../../../../../../../rtems-sync/c/src/lib/libcpu/mips/tx39/vectorisrs/vectorisrs.c:39 2010-07-01T07:35:51 #7 0x8800f120 in _ISR_Handler_1 () at ../../../../../../../../../../../rtems-sync/c/src/../../cpukit/score/cpu/mips/cpu_asm.S:916 2010-07-01T07:35:52 Backtrace stopped: frame did not save the PC 2010-07-01T07:35:54 (gdb) p _Thread_Heir 2010-07-01T07:35:56 No symbol "_Thread_Heir" in current context. 2010-07-01T07:35:57 (gdb) p _Per_CPU_Information 2010-07-01T07:35:58 $1 = {isr_nest_level = 0, executing = 0x88071c78, heir = 0x88071c79, idle = 0x88071c78, dispatch_needed = false} 2010-07-01T07:36:00 the boost unit testing framework turned out to be not thread-safe, now i evaluate the google c++ testing framework 2010-07-01T07:36:19 stack overflow? 2010-07-01T07:36:39 I am not sure I have ever run that particular test on the mips simulator 2010-07-01T07:37:05 all tests with more than one thread (except idle) do not work 2010-07-01T07:37:35 Ahh... I wonder if I broke it with some of the last changes to the per cpu structure 2010-07-01T07:37:38 *** CLOCK TICK TEST *** 2010-07-01T07:37:40 TA1 - rtems_clock_get_tod - 09:00:00 12/31/1988 2010-07-01T07:37:41 TA2 - rtems_clock_get_tod - 09:00:00 12/31/1988 2010-07-01T07:37:43 TA3 - rtems_clock_get_tod - 09:00:00 12/31/1988 2010-07-01T07:37:44 mips-core: 4 byte read to unaligned address 0x8801d6dd at 0x8800957c 2010-07-01T07:37:46 Program received signal SIGBUS, Bus error. 2010-07-01T07:37:47 0x8800957c in _Thread_Clear_state (the_thread=0x8801d930, state=) 2010-07-01T07:37:49 at ../../../../../../../../../rtems-sync/c/src/../../cpukit/score/src/threadclearstate.c:88 2010-07-01T07:37:50 88 if ( the_thread->current_priority < _Thread_Heir->current_priority ) { 2010-07-01T07:37:52 (gdb) p _Per_CPU_Information 2010-07-01T07:37:53 $1 = {isr_nest_level = 0, executing = 0x8801d6c8, heir = 0x8801d6c9, idle = 0x8801d6c8, dispatch_needed = false} 2010-07-01T07:38:03 or.. is the stack misaligned? 2010-07-01T07:38:27 i have no idea, discovered this 5 minutes ago 2010-07-01T07:38:48 i did some libio changes and wanted to run the test suite before the commit 2010-07-01T07:38:52 I have some new tests to commit and then will look at it when I get to the office. It can't be much 2010-07-01T07:39:06 ok, thanks 2010-07-01T07:39:18 if they run on sis or psim, that's OK 2010-07-01T07:39:54 the mips is nice due to its NULL pointer read error and the missaligned exception 2010-07-01T07:40:39 I guess I need to run it more. I have been working on coverage so tend to switch between sis on one machine and erc32 (requires tsim license) on another to get the results 2010-07-01T07:41:03 FWIW I now have termios tests which cover a LOT of termios.. this will make it safer to do some clean up 2010-07-01T07:42:25 ok 2010-07-01T07:50:27 * DrJoel is heading to the office 2010-07-01T07:50:30 *** DrJoel has quit IRC 2010-07-01T08:10:39 *** madrazr has joined #rtems 2010-07-01T08:11:13 *** exception13 has left #rtems 2010-07-01T08:25:36 *** zwj has joined #rtems 2010-07-01T09:02:25 *** bubaflub has joined #rtems 2010-07-01T09:06:56 *** exception13 has joined #rtems 2010-07-01T09:09:28 *** DrJoel has joined #rtems 2010-07-01T09:09:28 *** ChanServ sets mode: +o DrJoel 2010-07-01T09:09:56 *** kiwichris has quit IRC 2010-07-01T09:11:40 *** groleo has quit IRC 2010-07-01T09:14:03 *** bubaflub_ has joined #rtems 2010-07-01T09:14:25 *** bubaflub_ has joined #rtems 2010-07-01T09:14:28 *** bubaflub has quit IRC 2010-07-01T09:14:28 *** bubaflub_ is now known as bubaflub 2010-07-01T09:19:05 *** madrazr has quit IRC 2010-07-01T09:43:10 *** gedare has joined #rtems 2010-07-01T09:45:55 DrJoel: hi back 2010-07-01T09:46:05 thanks you for your answers to my mails 2010-07-01T09:46:48 but unfortunately i think that i've done everything you stated in your last mail, and the code was linked in my mail 2010-07-01T09:46:54 if you want to check 2010-07-01T09:47:26 zwj: i saw you are planning to do itron initialization, but itron was pruned from CVS 2010-07-01T09:47:37 you might want to discuss with your mentor if you haven't already 2010-07-01T09:47:59 that work might not be needed? 2010-07-01T09:48:08 * gedare tries to form a coherent thought 2010-07-01T09:48:24 we are not that far 2010-07-01T09:48:41 the actual initialization changes will be some time in the future 2010-07-01T09:49:00 sebhub: ok, I was just reading his "next steps" from his status update and noticed that 2010-07-01T09:49:19 just thought i'd bring it to attention 2010-07-01T09:49:30 ok, thanks 2010-07-01T10:02:39 zwj: freebie.. itron is gone.. no work for you. :-D 2010-07-01T10:08:03 merging in CVS is a nightmare 2010-07-01T10:08:16 agree 2010-07-01T10:08:54 i'd like to see rtems move to a distributed vcs, i haven't had much experience with them, but they seem to handle merging a lot better 2010-07-01T10:15:32 i started developing with svn and then moved to git and have never looked back 2010-07-01T10:16:42 in svn merging sucks too 2010-07-01T10:18:21 yeah svn isn't much different than cvs 2010-07-01T10:18:38 i hear really good things about git / mercurial 2010-07-01T10:19:22 i only scratched the surface of git 2010-07-01T10:20:03 thank you for your attention, i am working on writing test case. so sorry for delay reponse 2010-07-01T10:20:58 I think the big advantage of distributed VCS is that you can commit locally so you can get local revisions without breaking the repo and make a lot of small changes, which makes merging easier because of smaller changesets 2010-07-01T10:21:04 itron initialization is now not doing, why itron pruned? 2010-07-01T10:21:33 zwj: I think the steering committee decided no one was using it... I'm not sure though, I didn't see an announcement on the users list 2010-07-01T10:21:41 yeah, now i am using git locally to maintain my code 2010-07-01T10:22:20 a second big advantage is that it is easier to check-out another developer's branch 2010-07-01T10:22:38 so for example, I could see what DrJoel is brewing in his SMP work ;) 2010-07-01T10:22:47 if he made his branch available that is. 2010-07-01T10:24:14 in the git, it is easy to make a new branch and it has no cost 2010-07-01T10:28:56 drjoel: cvs is broke -- rtems_malloc_boundary_helpers is used without declared 2010-07-01T10:30:04 should be completely out of the tree. let me grep again 2010-07-01T10:30:21 where is it? 2010-07-01T10:30:26 check cpukit/libcsupport/src/rtems_memalign.c: In function ‘rtems_memalign’: 2010-07-01T10:31:24 well at least not broken in the "doesn't compile" way 2010-07-01T10:31:24 i'm re-cvs'upping, but didn't see that file touched so probably still broke 2010-07-01T10:31:28 yeah sorry 2010-07-01T10:31:35 er, well, it doesn't compile for me :) 2010-07-01T10:31:54 Hmmm.. that shouldn't be defined 2010-07-01T10:32:05 where is that define enabled? 2010-07-01T10:32:11 good question! 2010-07-01T10:32:27 cpukit/libcsupport/src/rtems_memalign.c: 2010-07-01T10:32:30 nope 2010-07-01T10:32:33 cpukit/score/include/rtems/score/heap.h: 2010-07-01T10:32:37 #define RTEMS_MALLOC_BOUNDARY_HELPERS 2010-07-01T10:32:59 That should never have been defined in there... grrre.. it is malloc not heap.. will catch that also 2010-07-01T10:33:05 heh 2010-07-01T10:33:29 ahhh.. only if debug enabled.. makes sense then 2010-07-01T10:34:12 yeah. good thing i build with that~! catch problems early ;) 2010-07-01T10:34:36 DrJoel: https://www.rtems.org/bugzilla/show_bug.cgi?id=1594 2010-07-01T10:35:36 mm, that is also with debug enabled 2010-07-01T10:35:48 yes 2010-07-01T10:36:03 hmmm.. sounds like we need a test build with debug enabled in the autobuilder 2010-07-01T10:36:12 hold on.. will fix that also 2010-07-01T10:40:04 ok.. update now :-D 2010-07-01T10:59:44 is the doxygen documentation on the web site up to date? 2010-07-01T11:07:08 *** sebhub has quit IRC 2010-07-01T11:26:12 DrJoel, do you know about the status of the USB stack? 2010-07-01T11:26:41 according to the wiki: "The basic stack functionality works well." ... "Contact: Thomas Doerfler and Sebastian Huber" 2010-07-01T11:27:46 it is supposed to be working.. sebhub and th_d ported it. They should have USB disks working now. But they are the ones who know. Thomas' wife has been sick and I assume once he returns to work full time, it will get merged. FOr details, contact sebhub 2010-07-01T11:28:33 cool :) 2010-07-01T11:29:33 is the code available somewhere already? 2010-07-01T11:29:46 I don't think so. :( 2010-07-01T11:29:54 ok, I'm mailing them 2010-07-01T11:30:37 we'll be developing a USB Host core in Milkymist shortly, so I need to know how it could interface nicely with their drivers 2010-07-01T11:32:10 *** madrazr has joined #rtems 2010-07-01T11:32:10 DrJoel: i mailed you back 2010-07-01T11:37:48 mkChangeLogList is sloooooow 2010-07-01T11:41:37 *** arvind_khadri has joined #rtems 2010-07-01T11:42:44 rtems.org was bogged down badly earlier but it is doing multiple cvs diff's to figure things out. It would be better to do the cvs output to a file and process it multiple times 2010-07-01T11:43:08 Fallenou: ok.. in queue 2010-07-01T11:43:25 yeah. or to take a cvs diff result file as input, since i generate that anyway :) 2010-07-01T11:43:41 would make it easier to apply against submitted patches too! 2010-07-01T11:44:02 thanks :) 2010-07-01T11:44:13 you seem to be overbooked 2010-07-01T11:44:23 anyway, i'm generating changelogs for PR 1599 that i just created for _Dispatch_needed. DrJoel always seems overbooked ;0 2010-07-01T11:44:38 hehe 2010-07-01T11:44:54 i've been overbooked too recently, but now i'm totally on my gsoc :) 2010-07-01T11:44:57 but the coverage numbers are moving up. :-D 2010-07-01T11:48:02 *** bubaflub has quit IRC 2010-07-01T11:50:45 *** zwj has quit IRC 2010-07-01T11:51:23 *** gedare has quit IRC 2010-07-01T11:55:34 *** arvind_khadri has quit IRC 2010-07-01T12:15:55 *** aanjhan has joined #rtems 2010-07-01T12:22:22 *** aniceberg has joined #rtems 2010-07-01T12:30:12 *** dr__house has quit IRC 2010-07-01T12:30:13 Hello DrJoel 2010-07-01T12:31:13 *** dr__house has joined #rtems 2010-07-01T12:35:50 hey! New test results coming shortly. 2010-07-01T12:36:28 If you update, you should notice I have committed a lot and Jennifer committed her work 2010-07-01T12:37:03 ah. ok. i was a bit worried after seeing the test results. :D 2010-07-01T12:40:50 I realized I have missed the ChangeLog. I am updating the PR now, with the ChangeLog 2010-07-01T12:40:52 I forgot to commit them when I said I did. I did verify it ran on my laptop but didn't commit it. So the server coverage run didn't get it 2010-07-01T12:41:02 I wrote your changelog I think. Update 2010-07-01T12:45:26 Starting another coverage run now.. 2010-07-01T12:45:34 So what makes sense to work on next? 2010-07-01T12:45:40 * Fallenou is cvs uping 2010-07-01T12:47:39 and waiting I assume.. rtems.org is bogged down :( 2010-07-01T12:49:16 Added the ChangeLog to the PR 2010-07-01T12:51:38 update finished =) 2010-07-01T12:53:10 Now looking at the report to pick the next one... 2010-07-01T12:56:03 and I managed to attach the ChangeLog to the wrong PR. Wow! 2010-07-01T13:00:19 aniceberg: what next? There are still a number of pipe.c cases 2010-07-01T13:00:51 I was looking at those after filtering using Bharath Suri 2010-07-01T13:00:58 there they are.. 2010-07-01T13:02:35 and filtering on file name for fifo.c, I see a lot of cases 2010-07-01T13:02:58 looks to be 22 cases in fifo.c 2010-07-01T13:04:54 okay then. fifo.c it is... 2010-07-01T13:04:55 and 764 bytes on erc32.. ~10% of the 8k left in the report 2010-07-01T13:05:13 that may also get the imfs_fifo cases 2010-07-01T13:05:33 Did you notice that I added these lines to the summary? 2010-07-01T13:05:43 Symbols unreferenced: 37 2010-07-01T13:05:43 Bytes in unreferenced symbols : 3544 2010-07-01T13:06:23 Before yesterday, it was ~60 symbols and 6K. I hacked together some tests for the smallest ones and gxx_wrappers.c 2010-07-01T13:06:56 *** exception13 has left #rtems 2010-07-01T13:07:01 Plus Vinu added lchown, utimes, refactored out a help routine and added test code for those. He needed them for Go. :) 2010-07-01T13:07:18 * DrJoel is going awk for a while.. will post when run is on website 2010-07-01T13:07:43 i noticed those lines now. But I usually have a look at the unreferenced symbols themselves... 2010-07-01T13:16:27 *** aniceberg has quit IRC 2010-07-01T13:24:19 *** gedare has joined #rtems 2010-07-01T13:30:46 *** dr__house has quit IRC 2010-07-01T13:55:47 *** madrazr has quit IRC 2010-07-01T14:02:56 *** bubaflub has joined #rtems 2010-07-01T15:58:58 *** bubaflub has quit IRC 2010-07-01T16:00:10 DrJoel: i posted my report, sorry for being late 2010-07-01T16:00:16 and thanks for your answer :) 2010-07-01T16:00:40 Fallenou: no problem. I don't directly see the problem. :( 2010-07-01T16:01:04 anyway i added something in bsp.h 2010-07-01T16:01:08 But apparently I have caused plenty of my own lately. Still try to get all the tests to pass again. Checking mips, powerpc, and sparc 2010-07-01T16:01:12 Did that fix it? 2010-07-01T16:01:19 which made the "make all" compile some more files 2010-07-01T16:01:33 it didn't fix anything but i think the line i added was missing 2010-07-01T16:01:40 it's just a little less wrong i guess 2010-07-01T16:01:43 LOL 2010-07-01T16:02:06 i added 2010-07-01T16:02:14 #define BSP_HAS_FRAME_BUFFER 1 2010-07-01T16:03:09 i like that. "a little less wrong" 2010-07-01T16:03:29 :p 2010-07-01T16:03:53 yea, since it's still not working but i think this line was maybe necessary 2010-07-01T16:03:58 to me it's a little less wrong ^^ 2010-07-01T16:10:33 assertion "fb != -1" failed: file "../../../../../../../rtems/c/src/../../testsuites/samples/milkymist_vga/init.c", line 66, function: POSIX_Init 2010-07-01T16:10:33 hehe =) 2010-07-01T16:11:04 is there an errno ? 2010-07-01T16:11:49 i test 2010-07-01T16:12:06 open:: Too many open files in system 2010-07-01T16:12:09 hum ! 2010-07-01T16:12:27 #define CONFIGURE_MAXIMUM_FILE_DESCRIPTORS 5 2010-07-01T16:12:30 i have this 2010-07-01T16:12:34 do i need to increase it ? :o 2010-07-01T16:12:36 i'm trying 2010-07-01T16:14:08 * mwalle beliefs, that Fallenou now never forgets to validate and check for errors after function calls :) 2010-07-01T16:14:23 sure ... 2010-07-01T16:14:39 i'm used to test when i program for x86 2010-07-01T16:14:45 but not on embedded ^^ 2010-07-01T16:15:12 now perror() is my friend ! 2010-07-01T16:17:08 errors happen and it is good to check.. see the testsuites for paranoia in action 2010-07-01T16:17:17 yes increase it one at a time until the open passes 2010-07-01T16:17:40 hum i put 10 directly, and it still does not pass 2010-07-01T16:18:20 maybe my #define is incorrect 2010-07-01T16:18:23 is it the correct spelling ? 2010-07-01T16:18:33 * Fallenou is grepping 2010-07-01T16:18:36 your app needs framebuffer? 2010-07-01T16:18:45 is right, i checked that 2010-07-01T16:20:35 seems its CONFIGURE_LIBIO_MAXIMUM_FILE_DESCRIPTORS :) 2010-07-01T16:20:42 forgot LIBIO ... 2010-07-01T16:20:49 i'm trying again 2010-07-01T16:21:22 I forgot that myself yesterday 2010-07-01T16:22:01 it's my fault i just copy pasted without veryfying 2010-07-01T16:22:12 i'm so releaved that i found the origin of my problem ! 2010-07-01T16:22:23 now the open() from my driver is called ! 2010-07-01T16:27:31 seems that pthread_mutex_trylock(&mutex) fails 2010-07-01T16:27:36 it does not return 0 2010-07-01T16:31:25 Why do you have pthread mutexes? 2010-07-01T16:31:49 You didn't configure any... CONFIGURE_MAXIMUM_POSIX_MUTEXES 2010-07-01T16:32:19 the posix threading API is optional. If you need a mutex in your BSP, please use the classic api rtems_semaphore... 2010-07-01T16:33:16 oh ok 2010-07-01T16:33:22 i was doing like the pc386 2010-07-01T16:33:36 they use a posix mutex in their open() and close() in their framebuffer driver 2010-07-01T16:33:47 but yes i needed a #define CONFIGURE_MAXIMUM_POSIX_MUTEXES 1 2010-07-01T16:34:24 DrJoel: i think maybe posix is mandatory to use the framebuffer anyway? 2010-07-01T16:35:38 hmm.. must be.. ok.. I will file it away as somethign to think about. Really using a framebuffer and considering graphics support a core part of RTEMS is relatively new. 2010-07-01T16:35:41 i don't think if the rtems framebuffer api needs really posix 2010-07-01T16:35:46 The Microwindows support has been around almost 10 years but it wasn't part of the standard "we test that" set. 2010-07-01T16:35:47 maybe i can do without it 2010-07-01T16:36:17 I would leave it out for now. If it is needed, it shuold be a layer above 2010-07-01T16:36:25 but for now , just to step forward quickly, i will keep the mutex, to get the driver work 2010-07-01T16:36:50 but anyway i should warn you, pc386 uses a mutex too ! 2010-07-01T16:37:02 yes.. agreed... and that was meant ..."it needs to be thought about" 2010-07-01T16:37:13 ok good :) 2010-07-01T16:37:19 thinking is always good ! 2010-07-01T16:38:21 ok open and close are working good now 2010-07-01T16:40:07 mwalle: when in graphic mode in qemu, can we still print in the console the uart tx ? 2010-07-01T16:41:32 DrJoel: got it, i have a nice red screen :) 2010-07-01T16:41:36 framebuffer works 2010-07-01T16:41:43 Fallenou, you can either switch with strg+alt 1,2,3 2010-07-01T16:41:48 or redirect serial 2010-07-01T16:41:55 via -serial stdio 2010-07-01T16:42:00 oh ok 2010-07-01T16:42:15 -serial stdio works perfectly, thanks 2010-07-01T16:42:54 the actual driver is single buffered and without anti-flicker, since it directly writes to the frontbuffer 2010-07-01T16:43:12 next move : double or triple buffering 2010-07-01T16:43:52 Fallenou: do you have data going to the buffer now? 2010-07-01T16:44:22 DrJoel: the vga screen is all red, i think it means the write() is doing ok :) 2010-07-01T16:44:35 since i write red pixels on the entire framebuffer in a for loop 2010-07-01T16:44:42 * gedare claps 2010-07-01T16:44:47 YEAH! 2010-07-01T16:44:50 he thanks :) 2010-07-01T16:44:59 so do you need to update your paragraph of progress. :-D 2010-07-01T16:45:03 sure ! ;) 2010-07-01T16:45:06 DrJoel, then you should update your docu, have a look at framebuffer.t 2010-07-01T16:45:09 doing it right now 2010-07-01T16:45:22 in the example, they are using pthread_mutex 2010-07-01T16:47:08 doc updated 2010-07-01T16:47:31 Fallenou: can you file a PR about this .. mention the pc386 and framebuffer.t. I think it is AT LEAST violating the rule that drivers shuold only use Classic API so they can work when POSIX is disabled. And it may be that we can assume a FB driver is used in support of another layer (windowing manager) which shuold be managing the update access 2010-07-01T16:48:47 i just send a mail ? or there is a web interface somewhere to fill a report ? 2010-07-01T16:49:17 http://www.rtems.com/wiki/index.php/RTEMSBugReporting 2010-07-01T16:49:21 thanks :) 2010-07-01T16:49:40 * Fallenou cannot count the number of "thanks" he said today 2010-07-01T16:49:48 * Fallenou newbie 2010-07-01T16:50:11 no problem.. I am just glad all of the GSOC students are making such great progress. 2010-07-01T16:51:03 :) yes it's really great 2010-07-01T16:52:03 * DrJoel thinks it is time to head home. :-D 2010-07-01T16:52:07 * DrJoel waves night 2010-07-01T16:52:10 *** DrJoel has quit IRC 2010-07-01T17:14:42 gn8 ! 2010-07-01T17:19:36 *** madrazr has joined #rtems 2010-07-01T17:30:09 *** madrazr has quit IRC 2010-07-01T17:51:49 *** gedare has quit IRC 2010-07-01T18:07:25 *** DrJoel has joined #rtems 2010-07-01T18:07:25 *** ChanServ sets mode: +o DrJoel 2010-07-01T18:15:23 *** mwalle has quit IRC 2010-07-01T20:54:01 *** methril_ has joined #rtems 2010-07-01T20:55:19 *** methril__ has quit IRC 2010-07-01T21:27:31 *** kiwichris has joined #rtems 2010-07-01T21:37:07 *** dr__house has joined #rtems 2010-07-01T21:40:31 *** DrJoel has quit IRC 2010-07-02T08:31:09 *** rtemsLogger has joined #rtems 2010-07-02T08:32:35 sebhub:hi 2010-07-02T08:33:46 i have modified the test case. now it can detect whether it is called and called more than once 2010-07-02T08:33:48 hi 2010-07-02T08:33:59 excellent 2010-07-02T08:34:02 now i send to you. 2010-07-02T08:34:08 ok 2010-07-02T08:41:04 *** DrJoel has joined #rtems 2010-07-02T08:41:04 *** ChanServ sets mode: +o DrJoel 2010-07-02T08:47:49 sebhub:have you recived ? 2010-07-02T08:52:34 not yet 2010-07-02T08:54:13 *** madrazr has joined #rtems 2010-07-02T09:00:29 i send it again, now leave a monment 2010-07-02T09:01:10 *** bubaflub has joined #rtems 2010-07-02T09:02:17 ok, now i have two ;-) 2010-07-02T09:05:01 the test case is now a little bit complicated 2010-07-02T09:05:58 *** zwj has quit IRC 2010-07-02T09:38:44 *** gedare has joined #rtems 2010-07-02T09:56:30 *** exception13 has joined #rtems 2010-07-02T10:08:28 joel: can i add the USB stack somewhere at the RTEMS CVS? 2010-07-02T10:17:00 I am concerned about dumping it in the CVS tree and not doing it in the right place. Can I just put a diff or a tarball on the ftp site? 2010-07-02T10:17:43 yes, i can put it on a ftp server 2010-07-02T10:18:22 I can put it in the RTEMS ftp server if you like... when Thomas gets back (hopefully Birgitta is better) we need to merge it 2010-07-02T10:18:26 (And LWIP) 2010-07-02T10:18:50 before we merge it, it should be reviewed 2010-07-02T10:19:44 yes.. that is the issue.. we can put it on the RTEMS ftp server, file a PR to track comments and have that point to the ftp server for code 2010-07-02T10:19:56 and announce it to the list for open review 2010-07-02T10:20:12 ok, i try to prepare the tar files 2010-07-02T10:20:21 with Chris on holiday and Thomas otherwise engaged, it is going to be slow 2010-07-02T10:20:44 yes, but the milymist people may have a look at it 2010-07-02T10:22:45 right.. and you can deal with issues they have 2010-07-02T10:23:06 the first issue will be the documentation ;-) 2010-07-02T10:26:02 Yes.. it has to be able to be built by another human 2010-07-02T10:46:57 *** sebhub has quit IRC 2010-07-02T10:52:05 have a nice weekend drjoel! 2010-07-02T10:52:27 the project is on target, and i'm off to be in a wedding ;) 2010-07-02T10:52:57 *** gedare has quit IRC 2010-07-02T10:55:58 *** zwj has joined #rtems 2010-07-02T11:00:13 *** madrazr has quit IRC 2010-07-02T11:01:38 *** madrazr has joined #rtems 2010-07-02T11:07:04 *** zwj has quit IRC 2010-07-02T11:20:45 *** madrazr has quit IRC 2010-07-02T11:38:38 *** bubaflub has quit IRC 2010-07-02T11:47:09 *** backhappy has joined #rtems 2010-07-02T12:01:06 *** madrazr has joined #rtems 2010-07-02T12:42:22 *** aniceberg has joined #rtems 2010-07-02T12:45:15 hello DrJoel. Warm friday afternoon, is it? 2010-07-02T12:48:29 You mean your keyboard or the weather? You are smoking on the test cases :-D 2010-07-02T12:50:09 :D. Getting hold of things now... And the presently focused cases seem easy. 2010-07-02T12:50:18 and I meant the weather there 2010-07-02T12:50:20 :) 2010-07-02T12:53:45 It is hot here .. 2010-07-02T12:54:45 oh. google summer :) 2010-07-02T12:56:08 and DrJoel, I have created a new test instead of adding these to the spfifo02. I felt spfifo02 would become big with these and i also intend to use this for a few more routines. 2010-07-02T12:56:41 that's ok.. I like smaller tests when they break and you have to debug them :-D 2010-07-02T12:56:48 starting to merge it now. 2010-07-02T12:57:12 and changes await spfifo03... i will get a few more on read/write with those. Or do you think I need to have a new one if the test becomes big. 2010-07-02T12:57:55 ^ ? 2010-07-02T12:58:26 Does it need CONFIGURE_MAXIMUM_BARRIERS? 2010-07-02T12:59:20 run dos2unix on the scn and drop off the line about the baudrate.. that is from the bsp 2010-07-02T12:59:51 aouch. missed these. added to the checklist 2010-07-02T12:59:59 shall i send the updated tests? 2010-07-02T13:00:14 no.. already fixed it. :) 2010-07-02T13:00:21 and it does not require the CONFIGURE_MAXIMUM_BARRIERS 2010-07-02T13:00:43 CONFIGURE_MAXIMUM_FIFOS does the trick now ;) 2010-07-02T13:01:03 CONFIGURE_MAXIMUM_TASKS 1 2010-07-02T13:01:07 it would be good to get your checklist on the wiki. There should be a test writing page with info on the templates, expectations on checking status, etc. 2010-07-02T13:01:27 Only this much of the ChangeLog is required .. just thenew entry 2010-07-02T13:01:29 *** bubaflub has joined #rtems 2010-07-02T13:01:52 My mistake, I didnt check the configure section of init.c 2010-07-02T13:01:53 is it January there? LOL your date was 2010-01-02 :D 2010-07-02T13:02:02 2010-07-02 Bharath Suri 2010-07-02T13:02:02 PR 1600/testing 2010-07-02T13:02:02 * spfifo04/init.c, spfifo04/spfifo04.doc, spfifo04/spfifo04.scn, 2010-07-02T13:02:02 spfifo04/Makefile.am: New files to improve coverage on fifo/pipe 2010-07-02T13:02:03 routines 2010-07-02T13:02:39 lol. i managed to do that too. 2010-07-02T13:03:00 haste makes these. i wanted to send these asap and.... 2010-07-02T13:03:09 the results are as above. :o 2010-07-02T13:08:55 that's ok.. small chunks are easy to review and for me to catch things in. I just committed it and have started a coverage run.. give it 20 minutes and they should be on the website 2010-07-02T13:13:28 just about the time for my dinner. 2010-07-02T13:13:33 * aniceberg is awk 2010-07-02T13:23:29 *** dr__house has quit IRC 2010-07-02T13:44:10 Bytes Analyzed : 119024 2010-07-02T13:44:10 Bytes Not Executed : 6960 2010-07-02T13:44:10 Percentage Executed : 94.15 2010-07-02T13:44:10 Percentage Not Executed : 5.848 2010-07-02T13:44:10 Uncovered ranges found : 289 2010-07-02T13:44:10 Total branches found : 2677 2010-07-02T13:44:12 Uncovered branches found : 592 2010-07-02T13:44:14 194 branches always taken 2010-07-02T13:44:16 398 branches never taken 2010-07-02T13:54:35 ah ha. closer to the cent. 2010-07-02T13:58:10 Yes.. and if the review on pr1532 ever completes, I will add a test for libmisc/untar. Right now, that isn't even in the set analyzed. 2010-07-02T13:58:34 Would a summary of uncovered bytes per file help? 2010-07-02T13:59:11 that should help me choose the next set... 2010-07-02T13:59:39 sure it would be helpful 2010-07-02T14:03:10 ok.. hold on.. 2010-07-02T14:32:45 new report now on the website with that in it...http://www.rtems.org/ftp/pub/rtems/people/joel/coverage/erc32-OsPD-20100702-1430/ 2010-07-02T14:41:04 had a look DrJoel... that would definitely help choose the next file to focus on... 2010-07-02T14:41:14 until it is like termios. :) 2010-07-02T14:41:32 i will cling on to fifo, for now... and then continue with assoc and the rest 2010-07-02T14:43:25 I highly recommend that. termios has 4 styles of device drivers it supports: polled, interrupt driven, task driven and "pppdisc" where the application can hook in at a low level to do its own protcol. I have provided test drivers for 3 of the 4. 2010-07-02T14:44:23 You could likely hit a lot of termios with what I have now but knocking the pppdisc style driver down should leave you in better shape. A lot of the tests are going to be changing termios modes.. 2010-07-02T14:47:56 http://www.gnu.org/s/libc/manual/html_node/Noncanonical-Input.html has the description of the 4 ways VMIN/VMAX interact. I am sure that at least 2 if not 3 of the four modes are untested and could be tested now. 2010-07-02T14:47:58 ah. ok... but are you writing the tests as and when you write the drivers? 2010-07-02T14:48:07 as could a number of character processing options 2010-07-02T14:48:19 no.. just hitting the driver creation and basic IO paths 2010-07-02T14:48:42 I added a little test code for some others to make sure the structure was there 2010-07-02T14:49:02 okay 2010-07-02T14:49:14 but do the assoc adn fifo first.. then let's evaluate 2010-07-02T14:49:28 sure DrJoel 2010-07-02T14:53:36 DrJoel, I might be away from internet this weekend, but I will continue to work on fifo. 2010-07-02T14:55:02 ok. holiday weekend here .. will be off Monday but likely on and off the net.. will keep up with email some though 2010-07-02T14:56:17 ok DrJoel. I am heading out now... some sleep and I am off for a game of soccer tomorrow 2010-07-02T14:56:24 nite DrJoel 2010-07-02T14:57:25 night 2010-07-02T14:57:28 good job 2010-07-02T14:57:41 thanks DrJoel 2010-07-02T14:57:44 bye for now. 2010-07-02T14:58:02 *** aniceberg has quit IRC 2010-07-02T15:03:34 *** madrazr has quit IRC 2010-07-02T15:39:21 *** methril_work has quit IRC 2010-07-02T15:54:06 *** methril_work has joined #rtems 2010-07-02T16:01:03 *** bubaflub has quit IRC 2010-07-02T16:17:57 *** madrazr has joined #rtems 2010-07-02T16:20:44 *** DrJoel has quit IRC 2010-07-02T16:30:36 *** Raj_ has quit IRC 2010-07-02T16:56:02 *** exception13 has left #rtems 2010-07-02T17:30:14 *** madrazr has quit IRC 2010-07-02T17:38:42 *** n3oo3n has quit IRC 2010-07-02T17:40:46 *** lekernel has quit IRC 2010-07-02T17:41:25 *** n3oo3n has joined #rtems 2010-07-02T17:41:25 *** lekernel has joined #rtems 2010-07-02T18:15:04 *** methril_work has quit IRC 2010-07-02T20:31:39 *** alinrus has quit IRC 2010-07-02T23:11:23 *** dr__house has joined #rtems 2010-07-03T02:33:04 *** madrazr has joined #rtems 2010-07-03T02:46:02 *** madrazr has quit IRC 2010-07-03T03:04:19 *** madrazr has joined #rtems 2010-07-03T03:10:05 *** madrazr has quit IRC 2010-07-03T04:38:26 *** alinrus has joined #rtems 2010-07-03T04:54:34 *** madrazr has joined #rtems 2010-07-03T05:43:38 *** tuxmaniac has joined #rtems 2010-07-03T06:04:07 *** tuxmaniac has quit IRC 2010-07-03T06:05:28 *** tuxmaniac has joined #rtems 2010-07-03T07:40:28 *** madrazr has quit IRC 2010-07-03T08:16:03 *** alinrus has quit IRC 2010-07-03T08:20:22 *** madrazr has joined #rtems 2010-07-03T08:28:56 *** alinrus has joined #rtems 2010-07-03T08:40:53 *** madrazr has quit IRC 2010-07-03T09:14:45 *** madrazr has joined #rtems 2010-07-03T09:22:44 *** madrazr has quit IRC 2010-07-03T10:13:13 *** DrJoel has joined #rtems 2010-07-03T10:13:13 *** ChanServ sets mode: +o DrJoel 2010-07-03T10:23:43 * Fallenou waves at DrJoel :) 2010-07-03T10:23:53 * DrJoel waves back.. 2010-07-03T10:24:06 * DrJoel was having a nice quiet Saturday morning until I noticed Bugzilla was down :( 2010-07-03T10:24:19 aouch =( 2010-07-03T10:30:09 ahhh ... the external network disk we use for backups is full and the automated backup was causing problems .... 2010-07-03T10:30:50 disk full of backups or logs is often a problem :) 2010-07-03T10:43:26 *** _Lucretia_ has quit IRC 2010-07-03T10:44:57 *** _Lucretia_ has joined #rtems 2010-07-03T12:17:01 * DrJoel notes that the gsoc progress report is published and that he has been on the computer too long for a saturday 2010-07-03T12:23:08 good :) 2010-07-03T12:23:20 i've published this on my blog http://sionneau.net/ 2010-07-03T12:23:33 i will link to the progress report 2010-07-03T12:26:27 not translated in french yet, but i guess you don't care :p 2010-07-03T12:32:11 DrJoel: you forgot the word "project" in your "Joel adds : " for the NetBSD Atapi/Sata Driver 2010-07-03T12:44:40 web site down 2010-07-03T12:45:19 you e-mail announce brought too much trafic :p 2010-07-03T13:15:50 *** _Lucretia__ has joined #rtems 2010-07-03T13:15:51 *** _Lucretia_ has quit IRC 2010-07-03T13:23:55 *** _Lucretia__ is now known as _Lucretia_ 2010-07-03T13:24:06 *** _Lucretia_ has joined #rtems 2010-07-03T13:27:59 Try now.. site is up (wasn't down for me) and fixed the text I think 2010-07-03T15:50:45 *** dr__house has quit IRC 2010-07-03T16:35:28 *** exception13 has joined #rtems 2010-07-03T16:55:15 *** tuxmaniac has quit IRC 2010-07-03T16:57:12 *** tuxmaniac has joined #rtems 2010-07-03T17:22:22 *** exception13 has left #rtems 2010-07-03T18:20:45 *** DrJoel has quit IRC 2010-07-03T18:33:50 *** DrJoel has joined #rtems 2010-07-03T18:33:50 *** ChanServ sets mode: +o DrJoel 2010-07-03T18:55:36 *** tuxmaniac has quit IRC 2010-07-03T19:06:33 DrJoel: is there a page in the wiki that documents the patch submition and merging process ? 2010-07-03T19:10:56 Fallenou: http://www.rtems.com/wiki/index.php/Contributing 2010-07-03T19:12:03 hey alinrus :) 2010-07-03T19:12:10 thx 2010-07-03T19:13:21 i am coming back from some friend's in paris, bus is long so i go on irc with the htc =) 2010-07-03T19:15:45 i'm learning for my final exam, so i'm pretty much stuck at home. on the bright side, i have a broadband connection :) 2010-07-03T19:18:30 hehe at home i have adsl @ 800 ko/s 2010-07-03T19:19:02 at university we had 13 mo/s :' 2010-07-03T19:19:39 good luck for your exams though 2010-07-03T19:19:46 has the academic year ended in france? 2010-07-03T19:20:28 yes it ended june the 25 th for me 2010-07-03T20:08:33 gn8 2010-07-03T20:54:15 *** mpanetta has quit IRC 2010-07-03T20:56:18 *** mpanetta has joined #rtems 2010-07-03T21:03:13 *** dr__house has joined #rtems 2010-07-03T21:10:48 *** alinrus has quit IRC 2010-07-03T22:18:53 *** madrazr has joined #rtems 2010-07-03T22:26:28 *** madrazr has quit IRC 2010-07-03T23:48:48 *** dr__house has quit IRC 2010-07-04T02:06:36 *** dr__house has joined #rtems 2010-07-04T02:52:19 *** dr__house has quit IRC 2010-07-04T03:39:28 *** methril__ has quit IRC 2010-07-04T03:51:36 *** methril__ has joined #rtems 2010-07-04T05:06:41 *** mwalle has joined #rtems 2010-07-04T05:25:55 *** dr__house has joined #rtems 2010-07-04T05:57:20 *** exception13 has joined #rtems 2010-07-04T06:05:45 *** mwalle has quit IRC 2010-07-04T06:06:56 *** mwalle has joined #rtems 2010-07-04T06:54:24 *** zwj has joined #rtems 2010-07-04T06:55:03 DrJoel:hi 2010-07-04T07:23:34 *** mpanetta has quit IRC 2010-07-04T07:25:14 *** alinrus has joined #rtems 2010-07-04T07:25:39 *** mpanetta has joined #rtems 2010-07-04T07:59:09 *** tuxmaniac has joined #rtems 2010-07-04T08:36:00 * DrJoel waves hi to zwj 2010-07-04T08:36:46 Fallenou: rtems.info is in my basement.. I have 6Mbps down an 1 Mbps down 2010-07-04T08:38:09 DrJoel: Hi, i have update the bugzlla 2010-07-04T08:38:18 attach a new patch 2010-07-04T08:38:26 ok.. will look at it once I wake up :D 2010-07-04T08:38:37 OK 2010-07-04T09:13:36 *** zwj has quit IRC 2010-07-04T09:13:50 *** zwj has joined #rtems 2010-07-04T09:16:18 *** zwj has quit IRC 2010-07-04T09:16:21 *** zwj has joined #rtems 2010-07-04T09:31:18 *** zhangwenjie has joined #rtems 2010-07-04T09:31:19 *** zwj has quit IRC 2010-07-04T09:38:16 *** zhangwenjie has quit IRC 2010-07-04T09:40:38 *** zwj has joined #rtems 2010-07-04T09:42:45 *** zwj has quit IRC 2010-07-04T09:44:16 *** zwj has joined #rtems 2010-07-04T09:59:34 *** madrazr has joined #rtems 2010-07-04T10:10:24 *** madrazr1 has joined #rtems 2010-07-04T10:10:24 *** madrazr has quit IRC 2010-07-04T10:23:12 *** tuxmaniac has quit IRC 2010-07-04T10:54:33 *** madrazr1 has quit IRC 2010-07-04T11:08:28 *** zwj has quit IRC 2010-07-04T11:27:29 *** kuzew has quit IRC 2010-07-04T11:28:27 *** kuzew has joined #rtems 2010-07-04T11:48:09 *** tuxmaniac has joined #rtems 2010-07-04T11:54:58 *** madrazr has joined #rtems 2010-07-04T12:27:22 *** lekernel has quit IRC 2010-07-04T12:34:29 *** lekernel has joined #rtems 2010-07-04T12:43:18 *** madrazr has quit IRC 2010-07-04T13:34:25 *** DrJoel has quit IRC 2010-07-04T13:48:30 *** DrJoel has joined #rtems 2010-07-04T13:48:32 *** ChanServ sets mode: +o DrJoel 2010-07-04T13:53:21 *** DrJoel has quit IRC 2010-07-04T14:06:17 *** DrJoel has joined #rtems 2010-07-04T14:06:17 *** ChanServ sets mode: +o DrJoel 2010-07-04T14:36:35 alinrus: can you please point me to the files that need code review? 2010-07-04T14:36:57 you have the entire code repo of rtems checke in code.google.com :) 2010-07-04T14:54:23 tuxmaniac: cpukit/poisx/src/aio_misc.c aio_write.c aio_read.c aio_error.c aio_return.c and cpukit/posix/include/aio_misc.h aio.h 2010-07-04T14:54:43 testsuites/samples/testrtemsaio 2010-07-04T15:06:00 alinrus: thanks 2010-07-04T15:51:09 *** exception13 has left #rtems 2010-07-04T16:26:30 *** tuxmaniac has quit IRC 2010-07-04T17:32:45 *** DrJoel has quit IRC 2010-07-04T20:14:58 *** DrJoel has joined #rtems 2010-07-04T20:14:58 *** ChanServ sets mode: +o DrJoel 2010-07-04T22:11:26 *** exception13 has joined #rtems 2010-07-04T23:38:39 *** alinrus has quit IRC 2010-07-04T23:57:45 *** exception13 has left #rtems