2013-04-15T00:43:03 *** monstr_ has joined #rtems 2013-04-15T02:04:44 *** sebhub has joined #rtems 2013-04-15T02:04:56 good morning 2013-04-15T02:05:24 kiwichris, you were interested to know what happened with the GDB io. I just tried using the file io functions in GDB with my RTEMS target (leon3 and grmon). It doesn't seem to be implemented 2013-04-15T02:05:49 I get a remote I/O error. I'm guessing it's GRMON's fault 2013-04-15T02:06:40 oh well, I can always just manually print out the text files in the console and PuTTY will log them, so it's not too much of a big deal for me 2013-04-15T02:08:09 Ah ok you are using GRMON. I have never seen or used it. 2013-04-15T02:09:30 but now that I think about it.. doesn't file io depend on support on the target OS as well? I mean it has to go through the filesystem otherwise the debugger wouldn't know where to get the data from 2013-04-15T02:10:19 correct, all I did was show you a protocol. 2013-04-15T02:10:44 I am not sure what is needed on the target but you have open etc there 2013-04-15T02:11:12 I see. Might be an interesting side-project for later 2013-04-15T02:11:55 or maybe it would be a good project for gsoc, assuming it's not implemented 2013-04-15T02:44:07 I suspect it depends on the gdb stub being used. 2013-04-15T02:44:21 can grmon be extended ? 2013-04-15T02:50:16 It can via loadable modules, but from what I can tell it's very low level (i.e. read and write to this memory location) 2013-04-15T02:50:36 and this only works in the full version, not the evaluation version 2013-04-15T02:51:12 Ah ok. 2013-04-15T02:51:44 I doubt the gdb stub can be extended this way at least 2013-04-15T03:07:10 *** mgahlot has joined #rtems 2013-04-15T03:12:40 *** arvind_khadri has joined #rtems 2013-04-15T03:24:08 *** mgahlot is now known as mgahlot|away 2013-04-15T03:32:55 *** pram has joined #rtems 2013-04-15T03:34:21 *** pram has left #rtems 2013-04-15T04:14:59 *** the9a3eedi has quit IRC (Ping timeout: 245 seconds) 2013-04-15T05:13:56 *** mgahlot|away has quit IRC (Remote host closed the connection) 2013-04-15T05:15:05 *** mgahlot has joined #rtems 2013-04-15T05:19:04 *** mgahlot is now known as mgahlot|away 2013-04-15T05:20:40 *** mgahlot|away has quit IRC (Remote host closed the connection) 2013-04-15T05:20:52 *** mgahlot has joined #rtems 2013-04-15T05:22:00 *** mgahlot|away has joined #rtems 2013-04-15T05:22:20 *** mgahlot has quit IRC (Remote host closed the connection) 2013-04-15T05:23:00 *** mgahlot|away has quit IRC (Remote host closed the connection) 2013-04-15T05:40:49 *** mgahlot has joined #rtems 2013-04-15T05:40:54 *** mgahlot is now known as mgahlot|away 2013-04-15T05:43:00 *** mgahlot|away has quit IRC (Remote host closed the connection) 2013-04-15T05:43:33 *** mgahlot has joined #rtems 2013-04-15T05:44:23 *** mgahlot has joined #rtems 2013-04-15T06:02:32 *** mgahlot is now known as mgahlot|away 2013-04-15T06:04:56 *** mgahlot has joined #rtems 2013-04-15T06:05:09 *** mgahlot is now known as mgahlot|away 2013-04-15T06:09:45 *** mgahlot|away has quit IRC (Remote host closed the connection) 2013-04-15T06:09:58 *** mgahlot has joined #rtems 2013-04-15T06:15:59 *** mgahlot has quit IRC (Remote host closed the connection) 2013-04-15T07:12:11 *** lcpfnvcy has quit IRC (Ping timeout: 260 seconds) 2013-04-15T07:17:27 *** smya has joined #rtems 2013-04-15T07:18:34 Hello there! 2013-04-15T07:19:42 I thought of select RTEMS for Gsoc. Could you please explain the procedure that I should follow? Do I need to fix a bug? 2013-04-15T07:19:53 thought of selecting* 2013-04-15T07:21:36 *** weiY has joined #rtems 2013-04-15T07:22:06 *** monstr_ has quit IRC (Ping timeout: 245 seconds) 2013-04-15T07:41:58 sebhub, hi 2013-04-15T07:49:11 *** smya has quit IRC () 2013-04-15T07:57:20 *** kktherock has joined #rtems 2013-04-15T08:01:29 *** lcpfnvcy has joined #rtems 2013-04-15T08:16:54 *** kktherock has quit IRC (Remote host closed the connection) 2013-04-15T08:17:18 *** antgreen has quit IRC (Ping timeout: 258 seconds) 2013-04-15T09:35:20 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2013-04-15T09:49:55 *** antgreen has joined #rtems 2013-04-15T09:50:40 *** sebhub has quit IRC (Remote host closed the connection) 2013-04-15T10:01:17 *** antgreen has quit IRC (Ping timeout: 245 seconds) 2013-04-15T10:07:15 *** monstr_ has joined #rtems 2013-04-15T10:07:43 *** weiY has quit IRC (Ping timeout: 256 seconds) 2013-04-15T10:17:04 *** antgreen has joined #rtems 2013-04-15T10:31:06 *** S_Somani has joined #rtems 2013-04-15T11:13:27 *** antgreen has quit IRC (Ping timeout: 260 seconds) 2013-04-15T11:25:36 *** antgreen has joined #rtems 2013-04-15T11:25:59 *** gedare has joined #rtems 2013-04-15T11:26:44 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-15T11:58:48 *** S_Somani has joined #rtems 2013-04-15T12:10:02 *** monstr_ has quit IRC (Ping timeout: 246 seconds) 2013-04-15T12:50:37 *** DrJoel has joined #rtems 2013-04-15T12:50:37 *** DrJoel has joined #rtems 2013-04-15T12:50:37 *** ChanServ sets mode: +o DrJoel 2013-04-15T13:00:40 DrJoel: you should put a link to the wiki rtems-soc page 2013-04-15T13:00:43 on the irc motd 2013-04-15T13:09:38 *** DrJoel changes topic to "Welcome to RTEMS (http://www.rtems.org). Please feel free to ask questions and, if you can, answer those from others. Google Summer of Code 2013 is ON and RTEMS is a participating organization! See http://wiki.rtems.org/wiki/index.php/RTEMSSummerOfCode for more information." 2013-04-15T13:09:42 How's that? 2013-04-15T13:11:30 good 2013-04-15T13:12:00 I added an asciidoc to rtems-testing/gcc. Mostly at this point to capture what it does. Cynthia was exploring 2013-04-15T13:12:45 ok. i have not looked at it before. 2013-04-15T13:13:44 No big deal.. just more writing to be out of date :) 2013-04-15T13:15:08 can you pm a sec? 2013-04-15T13:21:52 DrJoel: I have written my proposal and shared it with you.Could you please see it? https://docs.google.com/file/d/0B41ApxXt-m4jdlNBZm1qWGdTMkE/edit# 2013-04-15T13:22:15 S_Somani: I have requested permission too. 2013-04-15T13:23:00 gedare: shared 2013-04-15T13:25:41 Good thank you. 2013-04-15T13:31:41 S_Somani: it would be better for you to copy-paste your document text and enable commenting 2013-04-15T13:32:04 with the docx file, I am having trouble highlighting certain parts of your proposal in order to give detailed comments. 2013-04-15T13:32:30 ohh.. yes.. it is better as a native google doc.. in the end, you will have to cut and paste it to the gsoc web form anyway 2013-04-15T13:32:50 oh okay..just give me a minute 2013-04-15T13:35:16 please access- https://docs.google.com/document/d/1oFjUvydBm78fyIQzq2ol6M7tdkC_f6c21hzoHjXzuJc/edit?usp=sharing 2013-04-15T13:35:45 I can edit it and comment 2013-04-15T13:36:25 S_Somani: you can turn off editing if you like, commenting is good enough 2013-04-15T13:36:40 though it is not a big deal :) 2013-04-15T13:37:58 ok :) 2013-04-15T13:39:14 * DrJoel needs to start a list of the proposals needing review :) 2013-04-15T13:58:19 *** Hesham-- has joined #rtems 2013-04-15T14:00:57 Hi 2013-04-15T14:01:10 hey 2013-04-15T14:01:47 I tried to build rtems for qoriq bsp, but there were now output 2013-04-15T14:02:25 no output** 2013-04-15T14:03:16 No binaries or libraries have been compiled also 2013-04-15T14:03:32 qoirq if the "BSP Family". The specific BSP instances within that family are represented by the ".cfg" files in .../libbsp/qoriq/make/custom 2013-04-15T14:03:34 qoriq_core_0.cfg qoriq_core_1.cfg qoriq.inc qoriq_p1020rdb.cfg 2013-04-15T14:05:35 Ok, thanks DrJoel. 2013-04-15T14:05:46 *** antgreen has quit IRC (Ping timeout: 256 seconds) 2013-04-15T14:07:35 That's an area rtems can be hard to figure out.. we have generic code, cpu archecture specific code, cpu model specific code, peripherals found on multple boards, bsp families, and then specific bsps within a family. All of those (except board variant specific) are possible areas for reuse and the structure of rtems is designed to let you reuse where possible 2013-04-15T14:08:00 *** pkrao has joined #rtems 2013-04-15T14:09:40 *** pkrao has quit IRC (Client Quit) 2013-04-15T14:12:35 I see. 2013-04-15T14:13:19 Somewhat like object oriented 2013-04-15T14:18:55 *** antgreen has joined #rtems 2013-04-15T14:22:39 *** Hesham-- has quit IRC (Ping timeout: 276 seconds) 2013-04-15T14:29:23 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-15T14:31:19 *** S_Somani has joined #rtems 2013-04-15T14:37:27 *** antgreen has quit IRC (Ping timeout: 256 seconds) 2013-04-15T14:49:27 *** antgreen has joined #rtems 2013-04-15T15:31:53 *** antgreen has quit IRC (Ping timeout: 240 seconds) 2013-04-15T16:08:12 *** Hesham-- has joined #rtems 2013-04-15T16:14:44 *** S_Somani has quit IRC (Read error: Connection reset by peer) 2013-04-15T16:15:18 *** hiddenpearls has joined #rtems 2013-04-15T16:16:41 *** hiddenpearls has joined #rtems 2013-04-15T16:24:25 lots of comings and goings 2013-04-15T16:25:02 :) 2013-04-15T16:25:20 hey DrJoel did you see about Boston? crazy... 2013-04-15T16:25:49 gedare, agree that's bad 2013-04-15T16:25:50 i see that you did 2013-04-15T16:28:00 hiddenpearls: at least one master's thesis was done with a related gsoc project.. my student in 2011 2013-04-15T16:28:16 who ported a resource reservation framework and implement a constant bandwidth server scheduling algorithm 2013-04-15T16:28:29 Third explosion. :( 2013-04-15T16:28:39 you mean the one at the library? 2013-04-15T16:28:49 I have a friend who was in the marathon .. he was 1 mile from the site when they stopped the race 2013-04-15T16:28:55 DrJoel, oh I had heard about the 2 in news ? 2013-04-15T16:28:57 yeah.. JFK library? 2013-04-15T16:28:57 yikes. 2013-04-15T16:29:01 yeah 2013-04-15T16:29:13 its 5 miles from the finish line i saw 2013-04-15T16:29:13 nut jobs.. that explains it. 2013-04-15T16:29:36 tax day they usually go for the irs 2013-04-15T16:29:36 Given the timing, they were intended for the middle of the pack 2013-04-15T16:29:56 well it remains to be seen whether this is home grown or foriegn nut job. 2013-04-15T16:30:02 yeah. 2013-04-15T16:30:19 hiddenpearls: what is your thesis about? 2013-04-15T16:30:26 My early theory was a bunch of fat people who felt disenfranchised because they can't run a marathon. 2013-04-15T16:30:35 you're so bad 2013-04-15T16:30:43 * DrJoel knows that's in bad taste 2013-04-15T16:30:56 maybe better than my theory though 2013-04-15T16:30:59 but then I realized they would have spotted the scooters .. LOL\ 2013-04-15T16:31:10 * gedare sighs 2013-04-15T16:31:16 and yours is even tackier than mine? 2013-04-15T16:31:17 DHS did it. 2013-04-15T16:31:26 oh.. working on the conspiracy theory.. :) 2013-04-15T16:31:29 sequestration you know. 2013-04-15T16:31:41 yup 2013-04-15T16:31:42 that also explains north korea you know 2013-04-15T16:31:47 definitely 2013-04-15T16:31:55 i bet someone from huntsville is helping move that along. 2013-04-15T16:32:00 gedare, I need to select the topic yet :( It is Final Thesis to get MS degree 2013-04-15T16:32:09 can you pull in the vatican and Russian mob? 2013-04-15T16:32:18 hiddenpearls: i see.. does your advisor have some ideas for you? 2013-04-15T16:32:27 or do you need some help to pick a project that can turn into a thesis? 2013-04-15T16:32:34 and, when is your thesis due? 2013-04-15T16:32:50 Exactly, I need help to pick a project that can turn into thesis 2013-04-15T16:33:14 semester has been started a week ago 2013-04-15T16:33:23 and I have 1 year to complete it 2013-04-15T16:33:24 a suitable implementation project + performance study can turn into a thesis quite easily. 2013-04-15T16:33:48 what is your experience / strengths? 2013-04-15T16:34:04 or, what parts of RTEMS interest you most? 2013-04-15T16:34:04 I'm pretty good in Web Development 2013-04-15T16:34:16 heh.. not sure we have much that is thesis worthy in that department :) 2013-04-15T16:34:54 is working professionally from last 6 years .. running a Freelance Software Company 2013-04-15T16:35:13 Ya, I understand .. :( 2013-04-15T16:35:26 I have to dig in Operating System to do thesis 2013-04-15T16:35:48 are you pretty good with C? Python? C++? Ada? 2013-04-15T16:36:21 I was very good in C,C++ during Study periods … 2013-04-15T16:36:35 these days, PHP is my dream language 2013-04-15T16:36:46 hehe 2013-04-15T16:36:52 :) 2013-04-15T16:37:02 learning a new Language is not difficult .. My concepts are strong 2013-04-15T16:37:29 like I'm learning Objective C to do iPhone apps 2013-04-15T16:37:30 ok.. 2013-04-15T16:37:52 i will try to list some ideas... btw i'm doing this with a little one in my lap, so i may need to leave at any time :) 2013-04-15T16:38:29 compiling RTEMS with clang and using some code analysis tools like static analysis checkers available with llvm 2013-04-15T16:39:28 ok 2013-04-15T16:40:38 there are a few projects related to linkers... the run time linker and the capture engine... kiwichris is working on those 2013-04-15T16:41:02 rtl is basically "done", but maybe needs more architectures to be supported.. but probably would fit in scope of a thesis 2013-04-15T16:41:59 *** MegaAlex has joined #rtems 2013-04-15T16:42:07 Cool 2013-04-15T16:42:23 those both would involve host tools and some rtems coding i think. 2013-04-15T16:43:20 ok 2013-04-15T16:43:40 condition variables for classic API would be another good project.. i'm not sure if another student is going to propose it or not 2013-04-15T16:44:15 but, defining synchronization primitives in a real-time system is definitely a challenge and implementing and evaluating them would be good.. this would be kernel-level hacking 2013-04-15T16:44:47 interesting 2013-04-15T16:45:33 somethings are going above from my head :) 2013-04-15T16:45:36 an area with a lot of potential is the "rtems toolkits". we would like to create an infrastructure to support creating and maintaining "kits" that compose multiple external projects 2013-04-15T16:45:43 but I have to grasp them :) 2013-04-15T16:45:49 this infrastructure would be probably python-based 2013-04-15T16:46:12 i'm not sure what kind of evaluation you could do on the framework once it is done though 2013-04-15T16:47:14 hiddenpearls: you need to talk to your advisor. What qualifies for masters is their opinion 2013-04-15T16:47:44 yes. i think these are some project areas you might like to discuss though. 2013-04-15T16:47:45 some of these are just "projects" while others are going across newer ground. 2013-04-15T16:48:02 for the master's thesis it is not too important usually to break new ground 2013-04-15T16:48:05 The "rtems toolkits" is an important project though 2013-04-15T16:48:27 University advisor asks me to do some project in formal methods because he is doing research in that :) 2013-04-15T16:48:46 btw I can choose some other advisor though 2013-04-15T16:48:48 We haven' 2013-04-15T16:49:02 We haven't gotten formal enough for that to work. :) 2013-04-15T16:49:10 lol 2013-04-15T16:49:18 i think the clang/llvm project is closest in that area 2013-04-15T16:49:20 I'm not interested too ... 2013-04-15T16:49:40 because there are a lot of llvm-based tools that use formal methods approaches. but nothing quite so clearly formal methods 2013-04-15T16:49:40 I will discuss with my advisor and let u know .. 2013-04-15T16:50:27 One odd area which I hesitate to mention is decision and condition coverage analysis based on our current branch and instruction coverage information. 2013-04-15T16:51:09 I hesitate to mention because I don't have a reference to give you an answer on details. Someone in formal methods might. 2013-04-15T16:51:20 Ok np 2013-04-15T16:51:52 We have assembly instruction coverage informatino and (ideally) we could map that back to how the sub-expressions in an if were executed. 2013-04-15T16:52:26 MCDC type coverage says there is a bit map of path executed which are sufficient to demonstrate thorough testing (I hestitate to say correct code). 2013-04-15T16:53:07 This is the type of coverage required for FAA DO-178 flight certification. They have levels of criticality and the truth tables and name of coverage style change 2013-04-15T16:53:34 i'm pretty sure llvm can give you lots of coverage information from its IR form, by the way ;) 2013-04-15T16:54:33 is there any project for new device driver ? 2013-04-15T16:54:35 We still need to make assembly coverage back to that? 2013-04-15T16:54:45 good, thomas applied... peter d will apply but only to review proposals 2013-04-15T16:55:06 DrJoel: i don't think so, it depends on the mapping from IR to assembly though 2013-04-15T16:55:10 some people aren't clicking accept.. kiwichris is one of them 2013-04-15T16:55:17 hiddenpearls: what kind of new device driver? 2013-04-15T16:55:29 fwiw, neither did you 2013-04-15T16:55:43 for some reason you are not listed as an OA, but as a mentor who needs to take user action ;) 2013-04-15T16:56:23 gedare: don't know yet 2013-04-15T16:56:53 hiddenpearls: any reason you want to do a device driver project? :) 2013-04-15T16:57:03 I filed that with Melange.. that shouldn't have been allowed to be created and I can't access it :( 2013-04-15T16:57:07 hiddenpearls: a framework for device drivers might be a good project. 2013-04-15T16:57:13 want me to try withdrawing it? 2013-04-15T16:57:53 hiddenpearls: or an interface definition language for drivers :) 2013-04-15T16:58:06 hmmm.. i don't remember root password on this computer! cannot update! 2013-04-15T16:58:09 Device driver is a term that seems interesting to me .. and It will be good if I do some work on that field 2013-04-15T16:58:23 whew, i remembered to give mysefl sudo 2013-04-15T16:59:09 hiddenpearls: Honestly things of broad applicability are of more interest.. so a driver class framework is good if you can identify a class of devices. 2013-04-15T16:59:35 I don't know if anyone has been bold enough to ask about my "Summer of Simulators" :) 2013-04-15T17:03:49 Ok. 2013-04-15T17:04:00 It's 3 am here :) 2013-04-15T17:04:39 heading to bed now .. I will talk to my advisor and read more about device drivers etc 2013-04-15T17:04:50 talk u tomorrow 2013-04-15T17:07:48 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-15T17:23:57 b mjj bn 2013-04-15T17:24:19 having help at the keyboard from a little one gedare? 2013-04-15T17:24:24 yup 2013-04-15T17:24:31 :) 2013-04-15T17:25:01 in the office late tonight? 2013-04-15T17:25:43 yeah.. I didn't sleep well and didn't get here until about 9:30.. and tomorrow.. I have to do blood work for physical before i come in.. Plus taking Friday off completely 2013-04-15T17:26:09 ooo, fasting fun.. 2013-04-15T17:27:59 well.. only after midnight.. so does that count for much? 2013-04-15T17:28:14 depends on when your blood is drawn :) 2013-04-15T17:28:25 EARLY!!! As early as I can get to the lab. :) 2013-04-15T17:28:38 for the gui configuration project, do we want the student to define a new file/format for configuration parameters? 2013-04-15T17:28:48 the "parse conf.t to obtain parameters" seems suspect to me 2013-04-15T17:29:14 I don't think that's necessary if conf.t can suffice. Cynthia and the student independently parsed conf.t 2013-04-15T17:29:15 i think it would be better if we can define a general format to use, and then we can pull from that format into conf.t, and maybe even into confdefs.h or some other way to provide user configuration 2013-04-15T17:29:18 ok 2013-04-15T17:29:32 *** mgahlot has joined #rtems 2013-04-15T17:29:40 We can add whatever markup and detail. But if something parses it, it better be able to complain about issues it finds 2013-04-15T17:29:46 I don' 2013-04-15T17:30:05 *** mgahlot is now known as mgahlot|away 2013-04-15T17:30:11 I don't mind having an "origin" format but if we don't have to have it, then we save one manual transformation and creating a " 2013-04-15T17:30:25 ok 2013-04-15T17:30:25 an "origin" to conf.t translator. Plus maintenance of both 2013-04-15T17:30:27 time for me to go 2013-04-15T17:30:54 night 2013-04-15T17:31:00 Mark left before I asked 2013-04-15T17:31:07 you should probably email us both 2013-04-15T17:34:57 *** gedare has quit IRC (Ping timeout: 252 seconds) 2013-04-15T17:37:30 DrJoel, is this a mentoring click ? 2013-04-15T17:37:41 yes.. supposed to me 2013-04-15T17:37:51 login in at google-melange.com and see if you can accept it 2013-04-15T17:37:56 I did click and it said I was accepted 2013-04-15T17:39:10 try again... I did something 2013-04-15T17:39:46 Connection already exists 2013-04-15T17:39:55 Is the answer 2013-04-15T17:40:08 this connection stuff is fragile. :( 2013-04-15T17:40:13 Now what ? 2013-04-15T17:40:25 I did all this days and days ago 2013-04-15T17:40:29 I need to catch lennie or csmadhusen online and see if they think we are all ok 2013-04-15T17:40:41 When we where first accepted. 2013-04-15T17:40:43 I have a connection myself that I can't accept or delete 2013-04-15T17:40:50 My profile shows nothing 2013-04-15T17:41:04 Can you see organization connections? 2013-04-15T17:41:18 Is that a 'join' rather than a 'connection' ;) 2013-04-15T17:41:32 Where do I look ? 2013-04-15T17:42:24 My Dashboard 2013-04-15T17:42:29 Ah found something. On my dashboard it said my connections and then nothing. I thought that was my list but I need to connect on the link ? sigh 2013-04-15T17:42:47 I have 9 items on the dashboard 2013-04-15T17:43:06 Done. 2013-04-15T17:43:16 so you are ok from your view? 2013-04-15T17:43:18 I have 4 2013-04-15T17:43:51 now I have accepted. I needed to select the RTEMS project (the selection hoover thingy was fragile) and then I could accept. 2013-04-15T17:44:04 Long way to go to hit accept 2013-04-15T17:44:32 Do you see My Org Applications? Connectinos for my org? Mentor/Student Evaluations? 2013-04-15T17:45:16 Just My todos, important doc, my connections and my orgs 2013-04-15T17:45:22 Nothing else. 2013-04-15T17:45:32 then you aren't admin. :( 2013-04-15T17:46:07 Is that something you fix or I do ? 2013-04-15T17:46:48 I think it is a melange issue.. but I can't seem to find anyone in #gsoc or #melange .. 2013-04-15T17:47:05 I emailed the melange dev list last week with some issues and didn't see any response 2013-04-15T17:47:17 *** Hesham-- has quit IRC (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )) 2013-04-15T17:53:00 *** mgahlot|away has quit IRC (Ping timeout: 264 seconds) 2013-04-15T18:26:46 *** MegaAlex is now known as MegaAlex|away 2013-04-15T18:54:15 *** SYCrane has joined #rtems 2013-04-15T19:06:16 *** DrJoel has quit IRC (Quit: Some folks are wise, and some otherwise.) 2013-04-15T19:23:50 *** the9a3eedi has joined #rtems 2013-04-15T20:16:19 *** vivekkrishna93 has joined #rtems 2013-04-15T20:20:46 *** S_Somani has joined #rtems 2013-04-15T20:23:41 *** S_Somani has left #rtems 2013-04-15T20:46:23 *** pkrao63 has joined #rtems 2013-04-15T20:50:05 *** pkrao63 has quit IRC (Client Quit) 2013-04-15T20:50:39 *** pkrao63 has joined #rtems 2013-04-15T20:52:43 *** vivek has joined #rtems 2013-04-15T20:53:04 *** vivek_ has joined #rtems 2013-04-15T20:53:06 *** vivek is now known as Guest1529 2013-04-15T20:54:29 hi 2013-04-15T20:55:20 oh i see 2013-04-15T20:55:22 hi vivu 2013-04-15T20:55:37 ok ur all set with irc chat 2013-04-15T20:56:42 *** Guest1529 has quit IRC (Ping timeout: 245 seconds) 2013-04-15T21:00:03 *** pkrao63 has quit IRC (Ping timeout: 245 seconds) 2013-04-15T21:23:52 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-04-15T21:27:27 *** the9a3eedi has joined #rtems 2013-04-15T21:45:27 *** vivek_ has quit IRC (Ping timeout: 245 seconds) 2013-04-15T21:57:35 *** hiddenpearls has joined #rtems 2013-04-15T22:01:25 *** hiddenpearls has quit IRC (Client Quit) 2013-04-15T22:07:06 *** hiddenpearls has joined #rtems 2013-04-15T22:34:45 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-15T22:50:26 *** hiddenpearls has joined #rtems 2013-04-15T23:14:06 *** S_Somani has joined #rtems 2013-04-15T23:49:39 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-15T23:58:25 *** arvind_khadri has joined #rtems 2013-04-15T23:59:22 *** arvind_khadri has joined #rtems 2013-04-16T00:14:29 *** freenix has joined #rtems 2013-04-16T00:26:27 is there any way to get the capture engine to track ISR calls? this is probably technically impossible though, but I'd like to make sure 2013-04-16T00:44:26 None 2013-04-16T00:54:23 thanks 2013-04-16T01:02:17 *** monstr_ has joined #rtems 2013-04-16T01:09:16 *** kkthecoder has joined #rtems 2013-04-16T01:17:37 *** S_Somani1 has joined #rtems 2013-04-16T01:19:37 *** S_Somani has quit IRC (Ping timeout: 245 seconds) 2013-04-16T01:57:09 *** S_Somani1 has quit IRC (Ping timeout: 258 seconds) 2013-04-16T02:05:11 *** sebhub has joined #rtems 2013-04-16T02:06:23 good morning 2013-04-16T02:09:05 *** hiddenpearls has joined #rtems 2013-04-16T02:30:06 *** kkthecoder has quit IRC (Ping timeout: 252 seconds) 2013-04-16T03:07:01 *** SYCrane has quit IRC (Read error: Connection reset by peer) 2013-04-16T03:07:21 *** SYCrane has joined #rtems 2013-04-16T03:36:35 sebhub, are you about ? 2013-04-16T03:38:43 *** S_Somani has joined #rtems 2013-04-16T03:41:09 hi chris 2013-04-16T03:42:28 Looking at hypervisor virtual directories. I am concerned anything that attempts to go against the arch model will break not only itself but also RTEMS. 2013-04-16T03:43:11 Adding a special CPU called virtual is not real. My understanding is you need to use a i386 for a i386 based hypervisor 2013-04-16T03:43:32 Same for a leon, arm or ppc 2013-04-16T03:43:54 If they have common code then libcpu is the place for that 2013-04-16T03:59:12 *** S_Somani1 has joined #rtems 2013-04-16T04:01:41 *** S_Somani has quit IRC (Ping timeout: 246 seconds) 2013-04-16T04:08:21 *** S_Somani1 has quit IRC (Read error: Connection reset by peer) 2013-04-16T04:11:07 *** the9a3eedi has quit IRC (Ping timeout: 256 seconds) 2013-04-16T04:25:47 *** kkthecoder has joined #rtems 2013-04-16T04:39:48 *** S_Somani has joined #rtems 2013-04-16T04:52:57 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-16T04:55:14 *** cdcs has joined #rtems 2013-04-16T04:56:02 *** the9a3eedi has joined #rtems 2013-04-16T04:58:23 *** kkthecoder has quit IRC (Ping timeout: 255 seconds) 2013-04-16T05:00:23 *** hiddenpearls has joined #rtems 2013-04-16T05:04:07 *** hiddenpearls has quit IRC (Client Quit) 2013-04-16T06:05:01 *** hiddenpearls has joined #rtems 2013-04-16T06:17:27 *** mgahlot has joined #rtems 2013-04-16T06:17:50 *** mgahlot is now known as mgahlot|away 2013-04-16T06:32:53 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-16T07:33:17 *** sebhub has quit IRC (Remote host closed the connection) 2013-04-16T07:37:12 *** gedare has joined #rtems 2013-04-16T07:50:51 morning 2013-04-16T07:54:31 hi 2013-04-16T07:54:41 gsoc is in the air. ;-) 2013-04-16T07:54:52 yup ;) 2013-04-16T08:12:43 hi 2013-04-16T08:12:47 howdy 2013-04-16T08:12:58 what is the use case for rtems + paravirtulization ? 2013-04-16T08:13:20 to run RTEMS as a partition in a partitioned system such as an ARINC 653 platform 2013-04-16T08:17:24 freenix: can you send a proper patch for the 6410 to the rtems-devel mailing list? 2013-04-16T08:17:50 (without the modifications to hello world and ticker.. just the port so we can consider whether to include it?) 2013-04-16T08:28:54 *** mgahlot|away has quit IRC (Ping timeout: 264 seconds) 2013-04-16T08:29:20 *** mgahlot has joined #rtems 2013-04-16T08:31:36 *** mgahlot has quit IRC (Remote host closed the connection) 2013-04-16T08:33:13 *** antgreen has joined #rtems 2013-04-16T08:37:00 *** mgahlot has joined #rtems 2013-04-16T08:43:19 *** kkthecoder has joined #rtems 2013-04-16T08:43:33 *** mgahlot is now known as mgahlot|away 2013-04-16T08:46:22 *** mgahlot has joined #rtems 2013-04-16T08:53:15 *** hiddenpearls has joined #rtems 2013-04-16T09:01:36 Hi gedare 2013-04-16T09:01:42 hi freenix 2013-04-16T09:02:08 The code now just support timer and serial 2013-04-16T09:02:16 It is very simple now. 2013-04-16T09:03:05 most of the hardware configuration is done in uboot. 2013-04-16T09:03:28 everything starts out simple 2013-04-16T09:03:37 and if we're lucky we manage to keep it that way 2013-04-16T09:04:24 *** S_Somani1 has joined #rtems 2013-04-16T09:04:32 You may do not appreciate the coding style. I'll modify it and submit the patch 2013-04-16T09:04:33 *** S_Somani has quit IRC (Ping timeout: 240 seconds) 2013-04-16T09:05:58 Now I am make efforts to add nand support. the hello and ticker example is not modified. 2013-04-16T09:06:17 *** mgahlot has quit IRC (Remote host closed the connection) 2013-04-16T09:08:24 gsoc is all about code.. it is good for us to review your coding now and give you feedback 2013-04-16T09:08:44 this also gives you a better chance to get accepted.. if you are seen participating in RTEMS development 2013-04-16T09:09:12 incremental improvements to a BSP is fine.. especially if there are not other users yet. 2013-04-16T09:09:28 and this way, if someone else has the same SOC they can try out your code and give feedback or even submit their own improvements. 2013-04-16T09:09:41 anyway... did you read about the RTL? 2013-04-16T09:11:05 Thanks. Shoud I send the patch to the mailist? or where? 2013-04-16T09:11:18 yes, if you can send a git-format-patch 2013-04-16T09:11:37 also you can learn how to send a git patch directly with gmail on the RTEMS wiki. 2013-04-16T09:11:54 http://wiki.rtems.org/wiki/index.php/Git_Users#Sending_Email 2013-04-16T09:12:13 thanks, I'll read that and mail my patch 2013-04-16T09:12:27 it actually starts at http://wiki.rtems.org/wiki/index.php/Git_Users#Submitting_a_Patch 2013-04-16T09:17:19 Thanks for your advice, gedare. But now I have leave now.It's 22:00 now. 2013-04-16T09:17:25 ok 2013-04-16T09:17:28 *** cdcs has quit IRC (Quit: Page closed) 2013-04-16T09:17:54 *** freenix has quit IRC (Quit: Leaving) 2013-04-16T09:18:49 *** cdcs has joined #rtems 2013-04-16T09:27:08 *** cdcs has quit IRC (Ping timeout: 245 seconds) 2013-04-16T09:41:16 *** monstr_ has quit IRC (Ping timeout: 245 seconds) 2013-04-16T09:47:53 *** antgreen has quit IRC (Ping timeout: 240 seconds) 2013-04-16T09:55:06 *** SYCrane has quit IRC (Remote host closed the connection) 2013-04-16T10:00:33 *** S_Somani1 has quit IRC (Ping timeout: 240 seconds) 2013-04-16T10:03:22 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-16T10:07:56 *** DrJoel has joined #rtems 2013-04-16T10:07:56 *** DrJoel has joined #rtems 2013-04-16T10:07:56 *** ChanServ sets mode: +o DrJoel 2013-04-16T10:08:15 * DrJoel looks around 2013-04-16T10:10:24 *** weiY has joined #rtems 2013-04-16T10:23:57 *** antgreen has joined #rtems 2013-04-16T10:29:11 Hi gedare 2013-04-16T10:29:16 hey weiY 2013-04-16T10:29:19 Thank you for your comments 2013-04-16T10:29:24 welcome 2013-04-16T10:29:32 i have reply you 2013-04-16T10:30:40 How about the process of BSD port to RTEMS? 2013-04-16T10:30:57 because the lock debug tool is from FreeBSD 2013-04-16T10:31:17 Are you referring to the USB and TCP/IP stack effort? 2013-04-16T10:31:20 i want to know if there are some support about some freebsd services 2013-04-16T10:31:57 Drjoel, no. In my gsoc proposal i want to port the lock debug tool from freebsd to RTEMS 2013-04-16T10:31:59 what exists is either in the tree now or in rtems-libbsd 2013-04-16T10:32:20 Which is the latest or stable 2013-04-16T10:32:35 i don't know if it makes a lot of sense, because the lock debug tool is probably quite dependent on the lock implementation? 2013-04-16T10:33:00 There is only the rtems-libbsd. It is a work in process. Not stable yet at all. :) 2013-04-16T10:33:09 Hi Drjoel. If you have time your comments about my proposal is welcome. i have add you to my doc 2013-04-16T10:33:20 USB works.. tcp/ip for ipv4 will come up. Enable ipv6 and get a core dump 2013-04-16T10:33:37 OK.. I need to read (and write) more today 2013-04-16T10:34:41 gedare. the lock debug of FreeBSD is a relatively independent. YOU can define your own lock rules. So it is not spesic to any OS 2013-04-16T10:35:19 ok 2013-04-16T10:36:42 regarding the priority inversion, maybe it is a good idea to think about implementing priority check on lock acquire/release in order to double-check that locks are implementing the inheritance/ceiling protocol correctly... 2013-04-16T10:36:44 Drjoel. how about your option about the implementation of Lock debug tool used to debug deadlock or lock rescursion 2013-04-16T10:36:49 i'm not sure if it makes sense or not, though. 2013-04-16T10:38:46 *** hiddenpearls has joined #rtems 2013-04-16T10:39:08 gedare. if the priority inversion is implemented in lock debug tool, it will cross together lock-implementation. then it will not be independent and it will more complex 2013-04-16T10:41:11 *** hiddenpearls has quit IRC (Client Quit) 2013-04-16T10:41:36 I wouldn't mind seeing a tool like that but we need the "stack priority inheritance" algorithm to work and be the default. :) 2013-04-16T10:41:50 Now the lock debug tool has lock rules which user can add or modify and it also tracks the lock-primitive information. So if priority inversion detection rules is not complex and independent, it will also be ok 2013-04-16T10:41:59 *** Hesham-- has joined #rtems 2013-04-16T10:42:06 A broader project that addressed this from multiple views. 2013-04-16T10:42:29 I think you are proposing what effectively is an optional run-time check. 2013-04-16T10:42:54 I have long wanted a tool that ran inside gdb which analyzed a system to see if there was a deadlock. 2013-04-16T10:43:18 With the addition of python support in gdb, this is much more feasible than with the macro language it also had 2013-04-16T10:43:20 what is the mean of "stack priority inheritance" 2013-04-16T10:44:15 There are two algorithms in the score now. The default one which does not lower your priority until you release all resources. And another which steps your priority down as you release each mutex. 2013-04-16T10:44:16 gdb is another option. But you can alway debug all the code to detect the deadlock 2013-04-16T10:44:50 but if the lock debug tool is enabled. whenever the code is runing it will detect all the possible deadlock 2013-04-16T10:45:03 The "stepping down" algorithm was implemented by a student but never completely debugged or fully tested. 2013-04-16T10:45:22 so this tool can also detect the potential bug 2013-04-16T10:45:22 It is more theoretically correct but higher overhead. 2013-04-16T10:45:55 I understand on the tool .. but you may have a fielded production system and not be able to turn this code on. 2013-04-16T10:45:58 yeah, so in the prodect this option is not enabled default 2013-04-16T10:46:01 adding awareness of priority to the "witness" would be lock-independent. 2013-04-16T10:46:14 Both implementations are great options 2013-04-16T10:47:00 it might be simple to turn the "witness" implementation hooks into extensions for GDB to use also, in the future. 2013-04-16T10:47:27 witness will add some hook to lock struct so it also track the thread priority 2013-04-16T10:48:02 but adding awareness of priority should be evaluted 2013-04-16T10:48:24 I think a run-time option is great.. just trying to point out there there is never a single perfect solution in embedded systems as much as a salesman will tell you that about their RTOS. :) 2013-04-16T10:49:10 I don't know how much work the lock debug code is but fixing the stack algorithm would be a good test of it. :) 2013-04-16T10:49:53 I think the most useful of witness is where in the test stage . in this stage the overload is not problem and run all time to detect bugs. when in production the witness will be disable 2013-04-16T10:51:10 the lock debug code is relatively independent with OS or lock implementation. So port from freebsd is not diffcult 2013-04-16T10:51:52 *** Hesham-- has quit IRC (Ping timeout: 256 seconds) 2013-04-16T10:52:02 *** Hesham-- has joined #rtems 2013-04-16T10:52:10 Drjoel, could you give me a link to understand stack algorithm problem more and to see whether it is a good case for witness 2013-04-16T10:56:59 wei.a.yang@ gmail? 2013-04-16T10:57:01 hmmm 2013-04-16T10:57:07 the schedule this year is quite different for gsoc 2013-04-16T10:57:13 "begin coding" is not until june 17 2013-04-16T10:57:17 and "pencils down" is mid-september 2013-04-16T10:57:25 yes.. drags out until Oct for final data 2013-04-16T10:57:38 date .. but weekend for mentor summit is already on the calendar 2013-04-16T10:58:23 i had another idea for an open project... 2013-04-16T10:58:30 reviewing and implementing better idle tasks ;:) 2013-04-16T10:58:48 drjoel, yeah wei.a.yang@gmail.com 2013-04-16T10:58:55 Do you think there is a summer of work in that? 2013-04-16T10:59:16 no clue! 2013-04-16T10:59:23 probably someone needs to do the reviewing first to see 2013-04-16T10:59:38 it may be a small project instead 2013-04-16T11:00:27 another idea is to implement a boot-time memory checker 2013-04-16T11:00:35 it could do something like CRC over .text and .data 2013-04-16T11:01:16 .data can change though 2013-04-16T11:01:24 yup.. have to do the check early in boot 2013-04-16T11:01:30 but a background crc checker would be a nice addition to the toolkit 2013-04-16T11:01:34 or isolate the .data 2013-04-16T11:01:48 I would like to see the analog and discrete I/O support worked on. 2013-04-16T11:02:08 i feel like that is hard without having the hw to test it 2013-04-16T11:02:26 I have newer (but less complete) code on git.rtems.org in rtmes-multiio. And we have older but higher functinality code in a "hardware API" 2013-04-16T11:03:00 *** S_Somani has joined #rtems 2013-04-16T11:03:08 Just have a dummy driver. Bigger issue is what API do we want. No one ever gave feedback on the rtems-multiio code. But it was used on a couple of projects 2013-04-16T11:03:16 another idea is a background flash integrity checker 2013-04-16T11:03:38 it can run at low priority--i.e. idle--and is read-only on read-only memory so no locking is needed 2013-04-16T11:03:47 configuration table and a generic integrity checker? :) 2013-04-16T11:04:16 what do you mean by configuration table 2013-04-16T11:05:02 Just a table that specifies an array of memory blocks.. starting address, length, algorithm (maybe) and "check value" 2013-04-16T11:05:33 And maybe a frequency to check.. 2013-04-16T11:06:46 hehe. many run time checker is useful for RTEMS to see more information 2013-04-16T11:07:16 Hi,Drjoel. Is there any link on wiki about "stack priority inheritance"" 2013-04-16T11:08:02 ok, i have recevied your email 2013-04-16T11:08:39 i have to go to bed. see you tomorrow 2013-04-16T11:09:15 There may be an old project page.. but google isn't finding it 2013-04-16T11:09:22 The code is the truth :) 2013-04-16T11:12:11 ah, i see. yes a way to program the checker. maybe a hook also for what to do if a check fails 2013-04-16T11:13:05 running background integrity checks on general memory is not going to generalize 2013-04-16T11:14:09 who implemented the stack protocol? 2013-04-16T11:15:32 2007-12-21 Xi Yang 2013-04-16T11:15:32 * configure.ac, score/include/rtems/score/coremutex.h, 2013-04-16T11:15:32 score/include/rtems/score/thread.h, 2013-04-16T11:15:32 score/inline/rtems/score/coremutex.inl, 2013-04-16T11:15:32 score/src/coremutexsurrender.c, score/src/threadinitialize.c: Add 2013-04-16T11:15:32 support for proper stacking of priority inheritance on mutexes as 2013-04-16T11:15:32 well as enforce proper order of release. 2013-04-16T11:16:07 He is still around.. hmm.. let me ping him weiy and you for a status report 2013-04-16T11:16:35 yeah, i am here 2013-04-16T11:17:53 drjoel, it is code ? 2013-04-16T11:17:56 you should be in bed.. :) 2013-04-16T11:17:56 *** antgreen has quit IRC (Ping timeout: 256 seconds) 2013-04-16T11:18:19 weiy: Yes .. conditionally compiled algorithm in score. Plus some test code (sp38) 2013-04-16T11:18:23 sp38? 2013-04-16T11:18:50 hehe, in bed really 2013-04-16T11:19:09 i will look more deeply into those code 2013-04-16T11:19:29 I go to sleep and my wife stays up and writes .. she is working on a second book .. apparently it is quieter after I go to bed .. lol 2013-04-16T11:20:32 haha, it is the same to me. lol lol........... 2013-04-16T11:20:44 :) 2013-04-16T11:21:32 now it is really time to say byb. and i will contact Xi Yang to get more information and update the proposal. byb 2013-04-16T11:22:07 *** weiY has quit IRC (Quit: Linkinus - http://linkinus.com) 2013-04-16T11:33:10 *** Hesham-- has quit IRC (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )) 2013-04-16T11:34:37 *** prateekt has joined #rtems 2013-04-16T11:37:21 hi prateekt 2013-04-16T11:37:47 hi gedare... 2013-04-16T11:38:27 want to talk about scheduler project ideas? 2013-04-16T11:39:01 yes...can you please suggest me some of the ideas regarding scheduler?? 2013-04-16T11:40:23 yes. first though, what is your experience with C, Python, Bash scripting, and Makefiles? 2013-04-16T11:45:38 * gedare taps his microphone 2013-04-16T11:45:40 is this thing on? 2013-04-16T11:45:55 i have had good experience in C as i have done system programming quite alot...and regarding bash scripting i learned it in the past... 2013-04-16T11:46:03 ok 2013-04-16T11:46:38 just trying to gauge whether this would be a good project area for you, or what aspect of this might be good 2013-04-16T11:47:11 for example, the scheduler simulator is a useful project, which right now uses a mixture of bash scripting and Makefiles to stitch together the RTEMS scheduler with an external test driver 2013-04-16T11:47:17 i can see....what do you think now?? 2013-04-16T11:47:55 DrJoel developed the schedsim and knows better what might be done with it, but I think it might be better to wait until later to work on schedsim personally, because of pending changes to how we build rtems 2013-04-16T11:48:18 but, if you have decent experience with C and systems programming, then I think we can scope out a few scheduling related projects 2013-04-16T11:48:45 okay...i see.. 2013-04-16T11:49:32 what about the projects which are related to C and Systems Programming?? 2013-04-16T11:52:55 have you read some of the code and documentation for the current scheduler? 2013-04-16T11:53:12 *** prateekt_ has joined #rtems 2013-04-16T11:53:25 gedare.. not much bash in the scheduler simulator.. maybe none at all. 2013-04-16T11:53:39 oh, ok 2013-04-16T11:53:47 *** prateekt has quit IRC (Ping timeout: 245 seconds) 2013-04-16T11:53:53 prateekt_: have you read some of the code and documentation for the current scheduler? 2013-04-16T11:54:17 It has a hacked version of the RTEMS autoconf build system.. references an rtems source tree for some score, some Classic API, and some of the shell support. 2013-04-16T11:54:40 what is the work that needs to be done with it? 2013-04-16T11:54:44 It blends it all together to give you something you can configure like RTEMS using confdefs.h but is command driven 2013-04-16T11:54:58 and, would it be better to just wait until we switch to waf, and then convert schedsim to waf also? 2013-04-16T11:55:03 The build system sucks.. needs to be switched to waf 2013-04-16T11:55:17 yes...i read the documentation and your blog...regarding the current scheduler 2013-04-16T11:55:19 May be able to do it now. Chris would have to answer that 2013-04-16T11:55:43 But the autoconf build system is certainly out of sync with the current source tree 2013-04-16T11:55:43 ok prateekt_ good. some changes were made two summers ago when we added the EDF and CBS (constant bandwidth server) schedulers 2013-04-16T11:55:59 *** hiddenpearls has joined #rtems 2013-04-16T11:55:59 There are a few issues that I would like to investigate cleaning up and perhaps re-implementing. 2013-04-16T11:56:17 One is a bug that needs to be fixed, https://www.rtems.org/bugzilla/show_bug.cgi?id=2069 2013-04-16T11:56:36 Another is the number of hooks, and what they all do 2013-04-16T11:56:57 okay... 2013-04-16T11:56:59 The hooks for allocate, update, and free were all introduced to deal with how thread state changes affect scheduling 2013-04-16T11:57:13 * DrJoel would like to see the Block2N priority thread code refactored from the Thread Queue Priority support so it can be used in scheduling algorithms. And then a new scheduling algorithm like Simple Priority Scheduler added using it 2013-04-16T11:57:14 $ ls threadq*pri* 2013-04-16T11:57:14 threadqdequeuepriority.c threadqextractpriority.c 2013-04-16T11:57:14 threadqenqueuepriority.c threadqfirstpriority.c 2013-04-16T11:57:16 but these hooks are only used by the (deterministic) priority-based scheduler 2013-04-16T11:57:18 is the code to refactor 2013-04-16T11:57:52 another is what DrJoel just paste dumped ;) 2013-04-16T11:58:00 That is in the wiki though :) 2013-04-16T11:58:08 good thing too... 2013-04-16T11:58:27 yes i read about it in the wiki... 2013-04-16T11:58:41 I'd like to see if we can simplify the hooks, and maybe eliminate some of them 2013-04-16T11:59:11 another issue is that we added a "priority_compare" hook in the scheduler... 2013-04-16T11:59:48 really I think priority should be made more abstract by itself, with all priority operations in the score being handled by a pluggable Priority Manager or similar 2013-04-16T12:01:07 okay...how do you think i should begin with refactoring Block 2N thread code?? 2013-04-16T12:01:09 i thought i wrote some code like this, but now i can't find it :) 2013-04-16T12:01:53 probably to identify what would be the implementation of the scheduler hooks when using the block2n code 2013-04-16T12:02:38 I was thinking of a family of "thread set managers" .. there are at least 3 implementations in the score with different O() on time and space 2013-04-16T12:02:40 create a new scheduler based on the simple priority, and then copy-paste the block2n code.. once you get the block2n scheduler working, reduce the shared code 2013-04-16T12:03:29 ok i see... 2013-04-16T12:03:31 But they should have the same "operations" .. off the top of my head .. initialize .. enqueue.. dequeue.. extract.. get highest.. 2013-04-16T12:06:23 i was thinking what would be the advantages of a scheduler with block 2N code ??? 2013-04-16T12:06:48 simple priority scheduler i mean 2013-04-16T12:08:04 The default scheduler is the Deterministic Priority Scheduler. Its operations for block, unblock, etc are constant order - same number of instruction no matter how many tasks. 2013-04-16T12:08:27 Simple Priority is O(n) where n is the number of ready tasks for unblocking. 2013-04-16T12:09:20 the thread set abstraction might be a good way to reduce the number of scheduler hooks 2013-04-16T12:09:22 Block2N is O(bounded) for blocking.. but "bounded" by number of priorities in the system divided by a factor. It can take 16 iteratinos of a loop to insert a task at blocking 2013-04-16T12:10:05 I am starting to incorporate the concept of RTEMS managing thread sets into the class. It makes sense... there are sets of threads blocked waiting on resources or time. 2013-04-16T12:10:18 A ready set of threads 2013-04-16T12:10:22 And an executing set of threads 2013-04-16T12:11:05 The transitions between sets drive scheduling. in particular, the management of the ready set and the decision who to transition from ready to executing is the scheduler's job 2013-04-16T12:11:46 ok i get the picture now... 2013-04-16T12:12:01 I have a phone pic of a white board I drew.. need to get it into the computer 2013-04-16T12:12:54 The "thread set" used by the Deterministic Priority Scheduler should be one thread set handler, the simple priority another, and the block2n a third 2013-04-16T12:15:19 okay... 2013-04-16T12:16:09 can you send me the pic?? it would be interesting to have a look at it 2013-04-16T12:19:53 hold on.. 2013-04-16T12:20:34 it would certainly help my case.. 2013-04-16T12:21:34 my mail id is ptiwari2992@gmail.com 2013-04-16T12:23:18 i think this thread set abstraction might be helpful to cleaning up the scheduler hooks. 2013-04-16T12:24:43 yes i think so too... 2013-04-16T12:26:08 Can you give me some references i should not miss regarding the thread set abstraction?? 2013-04-16T12:26:36 i don't think we have much.. just the description the scheduler project page 2013-04-16T12:26:49 you'll need to flesh out some of this and work with us to figure it out 2013-04-16T12:27:01 In my head, I have a three column picture.. executing, ready, and blocked. With transition arrows between them which map to scheduler plugin points. 2013-04-16T12:27:03 but, the basic idea is to create a way to encapsulate a set of threads 2013-04-16T12:28:44 In object oriented terms, there is a virtual base class for "thread set" and at least 3 classes derived from it. All implementing the same API. Deterministic, Block2N and Simple all would provide priority based ordering of the set. 2013-04-16T12:29:33 okay... 2013-04-16T12:31:45 i should start working on it i guess.... 2013-04-16T12:33:24 you might want to start by diagramming the current scheduler and what constitute thread sets in the present implementation 2013-04-16T12:33:51 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-16T12:35:08 for the most part, the "executing" threads are just the head(s) of the ready thread set queue(s), and the ready threads are implemented within the scheduler framework (called the ready queue) 2013-04-16T12:35:17 and blocked threads are implemented by the threadq 2013-04-16T12:35:29 and have either priority or fifo discipline 2013-04-16T12:35:44 okay...i think doxygen would help getting the picture more clear... 2013-04-16T12:35:56 if i'm not mistaken, a fifo discipline threadq is equivalent to the simple scheduler readyq, right DrJoel ? 2013-04-16T12:39:24 no wait, i'm not right ;) 2013-04-16T12:39:34 the simple scheduler does sort by priority, of course.. 2013-04-16T12:39:41 we do not have a "fifo" scheduler 2013-04-16T12:40:35 rioght on fifi and priority scheduler .. same data structure.. different mechanics 2013-04-16T12:41:20 Thread set idea also helps uRTEMS or Tiny/RTEMS since you could do simple priority thread blocking. Save memory in code and data space. 2013-04-16T12:45:10 *** kkthecoder has quit IRC (Ping timeout: 256 seconds) 2013-04-16T12:45:53 by the way this doxygen thing is great.... 2013-04-16T12:46:00 *** hiddenpearls has joined #rtems 2013-04-16T12:47:14 DrJoel: watchdog chain might be another kind of thread set. 2013-04-16T12:47:14 Thanks.. a LOT of work has gone into it .. Doxygen didn't exist when RTEMS started. We started converting .h file comments to it a few years ago and this year Google Code-In high school students did a lot of grunt work on them. I did some work on the "grouping" of related functinoality .. if you see mistakes, patches are welcomed. This is a big area 2013-04-16T12:48:32 yes..i think its fascinating work.... 2013-04-16T12:49:02 gedare can you elaborate watchdog chain?? 2013-04-16T12:49:57 cpukit/score/src/watchdoginsert.c --- it implements a priority queue that is an insert-sorted linked list based on timeout values 2013-04-16T12:50:33 each "tick" checks if the head of the queue should be released 2013-04-16T12:50:52 * DrJoel thinks score watchdog is horribly named :( 2013-04-16T12:50:59 heh. 2013-04-16T12:51:04 people think it is a hardware device :( 2013-04-16T12:51:34 it's often known as a timer wheel 2013-04-16T12:51:57 okay...i get it... 2013-04-16T12:53:11 basically it is used by a task to queue events in the future and block until the event fires 2013-04-16T12:53:19 so, it is another way for threads to block 2013-04-16T12:53:25 I call it a delta chain .. wheel is when you put multiple ones in an array if I read the BSD docs correctly 2013-04-16T12:54:18 It is the source of almost all "timing" events in RTEMS. Timeouts on blocking, sleeping, Classic API timers, periods, etc. 2013-04-16T12:54:30 okay..tasks waiting for the event to happen... 2013-04-16T13:00:40 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-16T13:01:17 *** prateekt_ has quit IRC (Ping timeout: 245 seconds) 2013-04-16T13:17:08 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2013-04-16T13:20:24 *** kkthecoder has joined #rtems 2013-04-16T13:27:32 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-16T14:00:43 Hi 2013-04-16T14:02:21 Why do each semaphore has two task queus? FIFO and priority? And how does they prioritize between themselves? 2013-04-16T14:08:08 hello dhananjay 2013-04-16T14:08:26 gedare: Hi 2013-04-16T14:08:35 the task queue depends on what "discipline" the blocking threadq uses 2013-04-16T14:08:55 i think only one is active at a time, but i'm not so sure. 2013-04-16T14:09:00 DrJoel may know more. 2013-04-16T14:09:06 okay - but can it have both? 2013-04-16T14:09:42 i don't think so 2013-04-16T14:09:49 no.. Attributes.discipline determines which one is used 2013-04-16T14:10:03 *** hiddenpearls has joined #rtems 2013-04-16T14:10:04 tasks either wait at the semaphore in fifo or in priority order 2013-04-16T14:11:53 okay. I can find the discipline set to pririty here and still have two chains. 2013-04-16T14:12:16 here in the program I am trying to debug. 2013-04-16T14:14:04 what do you mean two chains? 2013-04-16T14:14:10 there should only be the Wait_queue 2013-04-16T14:15:08 oh i see 2013-04-16T14:15:31 With the priority-based discipline, the Wait_queue.Queues is actually an array of Chain_Control 2013-04-16T14:15:53 see... cpukit/score/include/rtems/score/tqdata.h 2013-04-16T14:16:08 but, I think there should actually be 4 chains 2013-04-16T14:16:26 It can only be fifo or priority discipline at a time. It is a create time attribute in the classic api. 2013-04-16T14:17:06 http://rtems.org/onlinedocs/doc-current/share/rtems/html/c_user/Semaphore-Manager-SEMAPHORE_005fCREATE-_002d-Create-a-semaphore.html#Semaphore-Manager-SEMAPHORE_005fCREATE-_002d-Create-a-semaphore is an example.. 2013-04-16T14:17:41 it is a union of one chain_control for FIFO and an array of 4 for priority discipline 2013-04-16T14:18:26 it should be replaced with a rbtree ;) 2013-04-16T14:19:07 You need to check the discipline 2013-04-16T14:19:29 serious question ... would that improve the performance? memory usage? 2013-04-16T14:19:52 Hmm.. thread set with rbtree as option #4 2013-04-16T14:20:58 not sure, but seems likely. or with the same set as the priority scheduler 2013-04-16T14:21:41 Interesting to consider in the thread set though. The O(memory) and O(time) are the keys to document for each one. Then you pick. 2013-04-16T14:21:59 ya 2013-04-16T14:22:03 dhananjay: did that help? Did my email response help on the "chain is empty" 2013-04-16T14:22:15 although O(time) for priority scheduler is dependent on some assumptions about priorities 2013-04-16T14:22:44 Yes.. .but 256 (or a power of 2 < 256) is a bounded design parameter :) 2013-04-16T14:23:39 later..getting some air 2013-04-16T14:23:42 Block2N has even weirder O() characteristics. It is order of the number of priority levels and their distribution blocking on a SINGLE PRIORITY BASED thread queue 2013-04-16T14:23:58 I understand.. take the daughter for a walk in the nice weather 2013-04-16T14:28:43 *** gedare has quit IRC (Ping timeout: 264 seconds) 2013-04-16T14:34:51 *** kiwichris has quit IRC (Quit: Leaving) 2013-04-16T14:43:22 *** hiddenpearls has quit IRC (Ping timeout: 245 seconds) 2013-04-16T14:43:25 *** hiddenpearls1 has joined #rtems 2013-04-16T14:46:49 *** hiddenpearls1 has quit IRC (Client Quit) 2013-04-16T14:54:53 Thanks DrJoel gedare. Fell asleep for a while, its 0120 here. ;) 2013-04-16T14:55:19 India? 2013-04-16T14:55:19 I will check 'em tomorrow and update you guys on result 2013-04-16T14:55:23 Yep 2013-04-16T14:55:57 no problem. irc isn't normally fast response.. faster than email. :) 2013-04-16T14:56:23 if you can layer the python up like the "classes" in the code, then you will shortly be reusing a lot. 2013-04-16T14:56:37 chain.. thread queue.. object handler.. 2013-04-16T14:56:52 plus posix thread == Classic API task.. both are supercore thread instances. 2013-04-16T14:57:11 posix mutex internally uses score mutex .. just as mutex mode of Classic API semapore 2013-04-16T14:57:29 Yes. I know I am duplicating some efforts,I was trying to prototype over already existing code bits. 2013-04-16T14:57:38 So when Classic API semaphore works, POSIX mutex and semaphore should be simple :) 2013-04-16T14:58:16 prototyping is good and replacing my macros with Python is good. 2013-04-16T14:58:41 Those macros are hard to write 2013-04-16T15:00:56 I think I will have to tweak the inheritance streams a bit over existing code to make it more general. 2013-04-16T15:02:02 Also, what does _Semaphore_Information really represent? Mutexes or semaphores? 2013-04-16T15:02:15 Are they same in RTEMS? 2013-04-16T15:02:17 I don't doubt that. I wondered if it would make sense to have one or more display commands for each structure/enum/bit field.. 2013-04-16T15:02:32 cpukit/rtems/include/rtems/rtems/sem.h? 2013-04-16T15:02:53 Classic API Semaphore can be either Mutex or Counting Semaphore based on an attribute specified at create time. 2013-04-16T15:04:06 Hmm.. and posix? 2013-04-16T15:04:19 posix mutex is a score mutex 2013-04-16T15:04:29 posix semaphore is a score semaphore 2013-04-16T15:04:39 with some fields added for "personality" 2013-04-16T15:05:54 is score == supercore? 2013-04-16T15:06:00 yes 2013-04-16T15:06:35 I suppose these fields areposix standard directives? 2013-04-16T15:07:40 score/supercore is RTEMS specific. The "classes" there are the building blocks for the APIs. They provide a SUPERset of the CORE capabilities required to implement the APIs supported 2013-04-16T15:09:23 Mm. Also I assume these wait_queues are circular? 2013-04-16T15:09:51 consdering the check function you mailed me\ 2013-04-16T15:10:39 no.. the chain_control just container an overlapping permanent head and tail node (e.g. immutable first) 2013-04-16T15:10:53 They are doubly linked chains/lists 2013-04-16T15:13:56 okay.I will lokk inot this tomorrow and will put a mail to the list. I should sleep now 2013-04-16T15:14:08 Thanks for the help 2013-04-16T15:16:07 sleep well.. night 2013-04-16T16:08:39 *** antgreen has joined #rtems 2013-04-16T16:50:49 antgreen: paying attention? 2013-04-16T16:54:27 oh well.. castch you another time 2013-04-16T16:54:30 *** DrJoel has quit IRC (Quit: When the chips are down, well, the buffalo is empty) 2013-04-16T18:17:51 *** Hesham-- has joined #rtems 2013-04-16T18:52:53 *** SYCrane has joined #rtems 2013-04-16T19:03:12 *** the9a3eedi has quit IRC (Ping timeout: 264 seconds) 2013-04-16T19:20:14 *** the9a3eedi has joined #rtems 2013-04-16T19:26:27 *** Hesham-- has quit IRC (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )) 2013-04-16T19:35:24 *** freenix has joined #rtems 2013-04-16T20:29:33 *** hiddenpearls has joined #rtems 2013-04-16T20:59:53 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-16T21:38:38 *** S_Somani has joined #rtems 2013-04-16T21:55:57 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-16T22:01:15 *** hiddenpearls has joined #rtems 2013-04-16T22:16:30 *** kkthecoder has quit IRC (Ping timeout: 272 seconds) 2013-04-16T22:36:27 *** hiddenpearls1 has joined #rtems 2013-04-16T22:38:42 *** hiddenpearls has quit IRC (Ping timeout: 256 seconds) 2013-04-16T22:43:39 *** kkthecoder has joined #rtems 2013-04-16T23:48:15 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2013-04-16T23:54:03 *** hiddenpearls has joined #rtems 2013-04-17T00:00:45 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T00:08:14 *** arvind_khadri has joined #rtems 2013-04-17T00:26:22 *** monstr_ has joined #rtems 2013-04-17T00:39:33 *** kkthecoder has quit IRC (Ping timeout: 240 seconds) 2013-04-17T00:59:14 *** kkthecoder has joined #rtems 2013-04-17T01:02:03 *** kkthecoder has quit IRC (Read error: Connection reset by peer) 2013-04-17T01:04:31 *** hiddenpearls has joined #rtems 2013-04-17T01:21:39 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T01:24:06 *** kkthecoder has joined #rtems 2013-04-17T02:03:55 *** sebhub has joined #rtems 2013-04-17T02:04:16 good morning 2013-04-17T02:15:33 *** kkthecoder has quit IRC (Ping timeout: 240 seconds) 2013-04-17T02:18:24 *** kkthecoder has joined #rtems 2013-04-17T02:50:11 *** kkthecoder has quit IRC (Ping timeout: 260 seconds) 2013-04-17T02:51:25 *** hiddenpearls has joined #rtems 2013-04-17T02:53:14 *** kkthecoder has joined #rtems 2013-04-17T03:00:46 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T04:16:17 *** the9a3eedi has quit IRC (Ping timeout: 246 seconds) 2013-04-17T04:21:38 *** hiddenpearls has joined #rtems 2013-04-17T04:35:54 *** hiddenpearls1 has joined #rtems 2013-04-17T04:37:29 *** hiddenpearls has quit IRC (Ping timeout: 245 seconds) 2013-04-17T04:47:03 *** the9a3eedi has joined #rtems 2013-04-17T05:09:05 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2013-04-17T05:22:03 *** kkthecoder has quit IRC (Quit: Leaving) 2013-04-17T06:01:38 *** hiddenpearls has joined #rtems 2013-04-17T06:16:55 *** Hesham has joined #rtems 2013-04-17T06:21:00 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T06:33:04 *** S_Somani has joined #rtems 2013-04-17T06:46:29 *** Hesham has quit IRC (Quit: Leaving.) 2013-04-17T07:36:07 *** gedare has joined #rtems 2013-04-17T07:55:34 good morning 2013-04-17T07:56:19 hi 2013-04-17T08:07:31 hi SYCrane 2013-04-17T08:07:58 hi 2013-04-17T08:08:30 i see you added yourself to the table of students on the wiki, but i don't think i have seen you discuss your project idea or post your hello world patch / screenshot unless I missed them 2013-04-17T08:12:32 sorry, i'm a little busy recently. i have sent a email to rtems-devel list with the hello world patch and screenshot. 2013-04-17T08:13:07 i do not know if this is fine 2013-04-17T08:15:56 *** hiddenpearls has joined #rtems 2013-04-17T08:20:14 i have discuss my project in this channle, should i discuss it using email list? 2013-04-17T08:20:14 gedare, sorry. i didn't check my email, i just checked email 2013-04-17T08:21:11 SYCrane: yes it would be preferable to give some idea of your project on the mailing list, where you can reach a wider audience and get early feedback. 2013-04-17T08:21:22 i have requested access to your draft proposal, also 2013-04-17T08:21:54 it's possible your email bounced, if you did not use the right email account or sign up to the mailing list, or if the message was too big. you can omit the screenshot and send it separately or post it somewhere and link to it 2013-04-17T08:24:45 thanks, i will send the patch and screenshot again, and also discuss my project on the mailing list. 2013-04-17T08:26:46 super 2013-04-17T08:27:13 please be aware that RTEMS cannot use code from Linux or other GPL products 2013-04-17T08:31:34 gedare: Hi, I talked to my Internal Supervisor and he said, University can provide me Supervisor from Software Engineering field 2013-04-17T08:35:01 *** gedare has quit IRC (Ping timeout: 245 seconds) 2013-04-17T08:36:56 *** DrJoel has joined #rtems 2013-04-17T08:36:56 *** DrJoel has joined #rtems 2013-04-17T08:36:56 *** ChanServ sets mode: +o DrJoel 2013-04-17T08:37:31 sebhub do you still have questions on the BSD code? lots of email just flew by and I am just getting here 2013-04-17T08:38:24 i sent some patches to the newlib list 2013-04-17T08:40:18 yes, i still have the y.tab.h problem 2013-04-17T08:45:17 *** kkthecoder has joined #rtems 2013-04-17T08:51:13 I recall there are two and they need to be generated. The Makefiles are supposed to do that. :( 2013-04-17T08:52:09 ok, i will have a look at this tomorrow, did you check in everything into the libbsd head? 2013-04-17T08:52:15 let me pull the source and look. 2013-04-17T08:52:34 I have switched machines since then. If I didn't, then it is on an external disk. I will check that also. 2013-04-17T08:54:09 i think we can use MSIZE 256 also for the current network stack 2013-04-17T08:54:33 That would simplify that issue. I just knew it was different. 2013-04-17T08:54:44 I think some .c files got added to the repository that are generated 2013-04-17T08:55:36 no.. take that back.. looking in lib/libipsec, I don't see policy_token.c so it is OK 2013-04-17T08:56:08 Is lib/libipsec/policy_token.c getting generated? 2013-04-17T08:56:52 find -name policy_token.c 2013-04-17T08:56:54 ./freebsd-userspace/lib/libipsec/policy_token.c 2013-04-17T08:58:27 [joel@rtbf64a libipsec]$ make policy_parse.c 2013-04-17T08:58:27 yacc -d -p __libipsecyy -o policy_parse.i policy_parse.y 2013-04-17T08:58:27 cat policy_parse.i | \ 2013-04-17T08:58:27 sed -e '/YY_BUF_SIZE/s/16384/1024/' >policy_parse.c 2013-04-17T08:58:27 rm -f policy_parse.i 2013-04-17T08:58:27 [joel@rtbf64a libipsec]$ ls 2013-04-17T08:58:28 ipsec_dump_policy.c ipsec_strerror.c pfkey.c policy_token.c 2013-04-17T08:58:28 ipsec_get_policylen.c ipsec_strerror.h pfkey_dump.c policy_token.l 2013-04-17T08:58:29 ipsec_set_policy.3 libpfkey.h policy_parse.c test-policy.c 2013-04-17T08:58:30 ipsec_strerror.3 Makefile policy_parse.y y.tab.h 2013-04-17T08:59:43 That rule must not be being tripped or an include path is wrong 2013-04-17T09:00:15 policy_token.l:61:19: fatal error: y.tab.h: No such file or directory 2013-04-17T09:00:30 this is the error, I don't have a y.tab.h 2013-04-17T09:01:32 If it isn't there, then the rule to generate isn't being tripped. :( 2013-04-17T09:02:59 Is policy_parse.c there? 2013-04-17T09:03:44 yes 2013-04-17T09:03:57 grep -r y.tab.h . 2013-04-17T09:03:58 ./freebsd-8.2/conf/files: clean "aicasm* y.tab.h" 2013-04-17T09:04:00 ./freebsd-8.2/conf/.svn/text-base/files.svn-base: clean "aicasm* y.tab.h" 2013-04-17T09:04:01 ./freebsd-userspace/lib/libipsec/policy_token.c:#include "y.tab.h" 2013-04-17T09:04:03 ./freebsd-userspace/lib/libipsec/policy_token.l:#include "y.tab.h" 2013-04-17T09:04:04 ./freebsd-userspace/Makefile:CLEAN_FILES += lib/libc/net/y.tab.h 2013-04-17T09:04:06 ./freebsd-userspace/Makefile:CLEAN_FILES += lib/libipsec/y.tab.h 2013-04-17T09:04:07 ./.gitignore:freebsd-userspace/lib/libipsec/y.tab.h 2013-04-17T09:04:09 ./.gitignore:freebsd-userspace/lib/libc/net/y.tab.h 2013-04-17T09:04:10 ./.gitignore:freebsd-userspace/lib/libipsec/y.tab.h 2013-04-17T09:04:12 I don't find a rule for y.tab.h 2013-04-17T09:04:27 is this a side effect from a command? 2013-04-17T09:08:59 It is a side effect of generating policy_parse. y.tab.h is yacc generated. 2013-04-17T09:27:13 *** S_Somani has quit IRC (Ping timeout: 240 seconds) 2013-04-17T09:27:16 *** freenix has quit IRC (Remote host closed the connection) 2013-04-17T09:29:07 *** S_Somani has joined #rtems 2013-04-17T09:33:15 *** kkthecoder has quit IRC (Read error: Connection reset by peer) 2013-04-17T09:33:15 DrJoel: Hi, I talked to my Internal Supervisor and he said, University can provide me Supervisor from Software Engineering field 2013-04-17T09:33:47 So, I'm interested in doing a framework for Device Drivers 2013-04-17T09:34:02 would that be sufficient ? 2013-04-17T09:35:17 If you can define a framework for something we are interested in. There are already patterns/frameworks in place for various classes of device drivers in RTEMS. I don't know of a particular need. You would have to identify the area. 2013-04-17T09:35:34 *** S_Somani1 has joined #rtems 2013-04-17T09:38:19 *** S_Somani has quit IRC (Ping timeout: 245 seconds) 2013-04-17T09:41:09 sebhub still there? 2013-04-17T09:43:10 yes 2013-04-17T09:46:46 Did any of that help on libbsd? 2013-04-17T09:47:13 Ok, I will look into it 2013-04-17T09:47:20 i will have a look at this tomorrow, since i have to leave soon 2013-04-17T09:47:26 *** gedare has joined #rtems 2013-04-17T09:50:17 *** SYCrane has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-04-17T09:53:53 *** kkthecoder has joined #rtems 2013-04-17T10:02:44 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T10:08:54 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-04-17T10:08:56 *** the9a3eedi-2 has joined #rtems 2013-04-17T10:36:29 *** sebhub has quit IRC (Remote host closed the connection) 2013-04-17T10:54:27 *** kkthecoder has quit IRC (Ping timeout: 256 seconds) 2013-04-17T10:55:57 *** zw_yao has joined #rtems 2013-04-17T10:59:39 Hi, I had a small doubt. Why are data types of each parameter present in config-v1.pdf but missing from conf.t? 2013-04-17T11:00:32 Mostly integer constant that you define.. either boolean.. some are method pointers 2013-04-17T11:01:43 *** zw_yao has quit IRC (Remote host closed the connection) 2013-04-17T11:01:56 DrJoel: these definitions are not present in conf.t...i just searched the document 2013-04-17T11:02:25 Start with cpukit/sapi/include/confdefs.h .. that is where these are interpreted. Then look at the includes that it has for the supporting structures 2013-04-17T11:03:50 oh..okay...but cant i directly parse the final generated transcript then...wouldn't it be easier? 2013-04-17T11:04:59 *** Hesham has joined #rtems 2013-04-17T11:05:04 Look at an RTEMS test or example. You define the defines, include and it generates code 2013-04-17T11:05:16 oh.. you mean parse ascii or pdf output from the document 2013-04-17T11:05:21 yup 2013-04-17T11:05:24 It would be better if we could stick to the original conf.t 2013-04-17T11:06:17 so i would have to look at confdefs.h for getting data types of each parameter? 2013-04-17T11:07:00 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2013-04-17T11:07:08 The data types in conf.t tell you what it is in English. You can see how it is used and get the precise rtems type in confdefs.h 2013-04-17T11:10:12 okay.. :) thanks. I was able to make a dynamically generated GUI. Now wished to add their default values hence the query. 2013-04-17T11:10:53 The default value is that you don't define it. :) 2013-04-17T11:11:19 that makes things easier for me 2013-04-17T11:13:58 You need the user to specify value.. integer.. or to define to enable the feature 2013-04-17T11:15:52 ok 2013-04-17T11:16:57 Look at the bottom of this file for a simple example of how it is used: http://git.rtems.org/examples-v2/tree/hello/hello_world_c/test.c 2013-04-17T11:20:26 I read the configuring RTEMS chapter, it looks familiar. 2013-04-17T11:25:00 That is a complete hello world. The examples-v2 should contain a number of simple (and complete) programs which give you an idea of how real applications use confdefs.h. If you move that to a self-contained "config.c" file, then you have the idea of what to generate 2013-04-17T11:27:36 Ok. I will through the process of configuring an application myself to get a better idea. 2013-04-17T11:34:34 *** monstr_ has quit IRC (Remote host closed the connection) 2013-04-17T11:35:26 *** Hesham has quit IRC (Ping timeout: 245 seconds) 2013-04-17T11:36:50 confdefs.h is designed so you (1) don't specify things you don't care about, (2) we can add new configuration points without impacting existing apps, (3) you either increase from nomimal or decrease from standards requirements, and (4) can specify "plugins" at certain points 2013-04-17T11:37:29 *** monstr has joined #rtems 2013-04-17T11:39:59 *** Hesham has joined #rtems 2013-04-17T11:40:29 Hi 2013-04-17T11:40:56 When I tried to find a BSP which is qoric model, going into libbsp/powerpc/qoriq/make/custom and found P1020RDB board mentioned there. But I think libbsp does not have such a bsp to build rtems for. 2013-04-17T11:42:39 DrJoel: ok :) Btw this is how the GUI looks like now - https://docs.google.com/file/d/0B41ApxXt-m4jaHA2QzJWSElHRTg/edit?usp=sharing 2013-04-17T11:43:09 configure with --enable-rtemsbsp=qoriq_p1020rdb 2013-04-17T11:45:00 Wow! You are moving very quickly. :) We need to figure out how to represent "unspecified - use default" 2013-04-17T11:45:17 Will it be easy to use tool tips or similar for help on each param? 2013-04-17T11:46:03 Ya I was thinking about tooltips too...I think it would be a good idea 2013-04-17T11:46:57 :) 2013-04-17T11:47:16 Most configuration parameters fall into a handful of types 2013-04-17T11:47:44 DrJoel: Yes I did configured rtems with this BSP and it seems like it did not found this BSP 2013-04-17T11:48:33 Yup ..i will create them dynamically based on their types... for ex- a checkbox will be created if the parameter is a boolean 2013-04-17T11:49:11 DrJoel: Now it works, a little typo error, thanks :) 2013-04-17T11:50:24 *** hiddenpearls has joined #rtems 2013-04-17T12:01:06 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T12:07:24 *** hiddenpearls has joined #rtems 2013-04-17T12:15:21 *** kkthecoder has joined #rtems 2013-04-17T12:38:23 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T13:10:14 *** hiddenpearls has joined #rtems 2013-04-17T13:28:53 *** Hesham has quit IRC (Ping timeout: 248 seconds) 2013-04-17T13:30:17 *** Hesham1 has joined #rtems 2013-04-17T13:32:51 *** S_Somani1 has quit IRC (Quit: Leaving.) 2013-04-17T13:44:34 *** DrJoel has quit IRC (Quit: Take my advice. I don't use it anyway) 2013-04-17T14:08:04 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T14:33:48 *** monstr has quit IRC (Ping timeout: 264 seconds) 2013-04-17T15:18:06 *** Hesham1 has quit IRC (Quit: Leaving.) 2013-04-17T15:18:20 *** Hesham has joined #rtems 2013-04-17T15:18:35 *** kkthecoder has quit IRC (Quit: Leaving) 2013-04-17T15:34:48 *** MegaAlex|away is now known as MegaAlex 2013-04-17T15:37:20 *** ares__ has joined #rtems 2013-04-17T15:38:49 *** ares__ has quit IRC (Client Quit) 2013-04-17T15:39:34 *** ares__ has joined #rtems 2013-04-17T15:41:43 *** AresHsu has joined #rtems 2013-04-17T16:10:33 any students have questions for me? i have a bit of time to kill. 2013-04-17T16:14:56 Yes, I got no feedbacks on my proposal, who else should I invite and what should I do else ? 2013-04-17T16:15:25 invite peter dufault and aanjhan ranganathan if you have their emails 2013-04-17T16:16:01 and maybe post the link in rtems-devel and invite others to review it 2013-04-17T16:16:27 Ok I will 2013-04-17T16:16:59 i'm taking a look now 2013-04-17T16:17:31 hmmmm 2013-04-17T16:17:49 oh i'm looking at your 2012 version :) 2013-04-17T16:17:56 i requested access 2013-04-17T16:18:28 hehe :) I have shared 2013 with you previously 2013-04-17T16:30:01 high-level feedback, I would like to see a clear list of goals for your project and expected outcome, and a preliminary design of what you would like to implement 2013-04-17T16:30:31 i think you have many of the pieces, but they are spread across your proposal. 2013-04-17T16:31:01 and, i think you should focus on the "static setup" use case first 2013-04-17T16:31:14 that is the most useful for an RTOS 2013-04-17T16:31:24 so it should be the most basic functionality available. 2013-04-17T16:31:54 you mean static setup at BSP init time ? 2013-04-17T16:32:13 es 2013-04-17T16:32:14 yes 2013-04-17T16:33:29 Ok, I will include "handling TLB misses" part to design goals as you suggested 2013-04-17T16:39:17 well, what i would say is that you do not need to focus on "high performance" or even mention the high/mid level layers, if your project design goal is first and foremost to support ease of porting for creating static memory maps 2013-04-17T16:39:45 then performance and services can come later 2013-04-17T16:39:58 we got too caught up last year in trying to define some useful high-level services 2013-04-17T16:40:04 when the underlying low-level code was just not ready 2013-04-17T16:41:08 i think it will be more useful for you to define a low/mid-level (e.g. BSP/CPU) framework for memory and cache management, and demonstrate the flexibility and portability of your framework using a couple of ARM and PPC bsps 2013-04-17T16:42:24 A new design other than the existing one ? 2013-04-17T16:42:56 it can be the existing one, if that one is suitable 2013-04-17T16:43:07 but remember... the existing one still has not been merged to rtems 2013-04-17T16:44:30 due to flags issue only ? I can work on it to be merged before the program starts 2013-04-17T16:44:45 Is that ok ? 2013-04-17T16:46:50 yep 2013-04-17T16:48:10 Good, I will send you a patch to review when I am done with it. 2013-04-17T16:49:34 super 2013-04-17T17:13:20 *** gedare has quit IRC (Ping timeout: 272 seconds) 2013-04-17T17:45:05 *** MegaAlex is now known as MegaAlex|away 2013-04-17T18:06:30 Hello, I have a question: I loaded centOS virtual machine in virtual box and ran: yum upgrade. But after that, when I powered it off and started it again, it was stuck in the starting process and never finished. Does anyone else have this problem? 2013-04-17T18:13:30 *** Hesham has left #rtems 2013-04-17T18:35:29 *** SYCrane has joined #rtems 2013-04-17T18:37:31 *** AresHsu has quit IRC (Quit: Leaving) 2013-04-17T19:21:09 *** the9a3eedi-2 has quit IRC (Ping timeout: 276 seconds) 2013-04-17T19:35:31 *** the9a3eedi has joined #rtems 2013-04-17T20:34:46 *** hiddenpearls has joined #rtems 2013-04-17T20:40:58 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T21:36:16 *** hiddenpearls has joined #rtems 2013-04-17T21:38:04 *** hiddenpearls has quit IRC (Client Quit) 2013-04-17T22:06:33 *** sevikkk has quit IRC (Ping timeout: 240 seconds) 2013-04-17T22:08:43 *** sevikkk has joined #rtems 2013-04-17T22:09:39 *** hiddenpearls has joined #rtems 2013-04-17T23:22:50 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-17T23:47:22 *** monstr has joined #rtems 2013-04-18T00:18:53 *** arvind_khadri has joined #rtems 2013-04-18T01:50:45 *** sebhub has joined #rtems 2013-04-18T01:53:32 good morning 2013-04-18T02:09:18 *** arvind_k has joined #rtems 2013-04-18T02:10:00 *** arvind_k has quit IRC (Max SendQ exceeded) 2013-04-18T02:10:20 *** arvind_k has joined #rtems 2013-04-18T02:11:40 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2013-04-18T02:41:54 *** MegaAlex|away is now known as MegaAlex 2013-04-18T03:05:59 *** freenix has joined #rtems 2013-04-18T03:19:22 *** monstr has quit IRC (Remote host closed the connection) 2013-04-18T03:42:40 *** monstr__ has joined #rtems 2013-04-18T04:00:43 *** hiddenpearls has joined #rtems 2013-04-18T04:28:31 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-18T05:06:41 *** hiddenpearls has joined #rtems 2013-04-18T05:13:20 *** cdcs has joined #rtems 2013-04-18T05:31:36 *** MegaAlex is now known as MegaAlex|away 2013-04-18T06:10:27 *** the9a3eedi has quit IRC (Ping timeout: 245 seconds) 2013-04-18T07:19:50 *** gedare has joined #rtems 2013-04-18T07:35:49 hello 2013-04-18T07:36:26 hi 2013-04-18T07:41:07 Hi gedare. 2013-04-18T07:41:31 freenix: looks like you are investigating the RTL? I think that is a good project to work on 2013-04-18T07:42:38 gedare: yes, I am trying to figure out how to do with this project. 2013-04-18T07:43:04 Busy preparing the proposal. Huh 2013-04-18T07:44:22 In the mailing list, you said the It may not make sense to include a new bsp variant for 6410. 2013-04-18T07:45:38 Hope you can give some suggestions about how to refactor it. 2013-04-18T08:01:18 freenix: i thought maybe it is better to have tiny6410 as a bsp variant of a generic 6410 bsp 2013-04-18T08:01:21 or something like that 2013-04-18T08:01:58 don't worry about it.. not an issue now 2013-04-18T08:02:36 ok.I'll try to make it as possible as I can. 2013-04-18T08:04:30 If want to be a generic bsp, I think 64xx should be all considered. But now, I do not have enough time and no hardware to do that 2013-04-18T08:07:56 If I have finished correcting the problem listed in the mailing list of the Tiny6410 patch, do I need to submit it again? 2013-04-18T08:15:41 freenix: yes send the updated patch 2013-04-18T08:16:02 freenix: probably sebhub will commit it for you once it is in good enough shape. 2013-04-18T08:16:22 freenix: i have seen the problem you described about failing to open a file. i'm thinking about it right now. 2013-04-18T08:17:04 although admittedly not with the rtl :) 2013-04-18T08:17:13 gedare: Thnaks for your help. 2013-04-18T08:18:03 I am not sure whether different tar version may incur that problem. 2013-04-18T08:18:25 what compiler toolchain are you using? 2013-04-18T08:18:41 sparc-rtmes4.11- 2013-04-18T08:18:51 if you untar fs-root.tar on your host, do you see the shell-init? 2013-04-18T08:18:57 yes 2013-04-18T08:19:14 the tar is compressed on my host pc. 2013-04-18T08:19:45 the tar compression is done when compiling rtl-target 2013-04-18T08:20:19 what is the conversion of fs-root.tar to fs-root-tarfile.o? 2013-04-18T08:20:45 sparc-rtmes4.11-* -I binary 2013-04-18T08:21:28 *** the9a3eedi has joined #rtems 2013-04-18T08:21:43 sparc-rtems4.11-objcopy -I binary -B sparc -O elf32-sparc fs-root.tar fs -root-tarfile.o 2013-04-18T08:22:07 sparc-rtems4.11-objcopy -I binary -B sparc -O elf32-sparc fs-root.tar fs-root-tarfile.o 2013-04-18T08:22:47 try to step through the Untar_FromMemory using sparc-rtems4.11-gdb 2013-04-18T08:23:10 ok. 2013-04-18T08:23:55 freenix: please send an updated patch to the list, the most important part is the license and copyright header, other things can be fixed later 2013-04-18T08:26:16 sebhub: ok. I am still do some modification to the code. I'll recheck the license and copyright. 2013-04-18T09:16:41 *** DrJoel has joined #rtems 2013-04-18T09:16:41 *** DrJoel has joined #rtems 2013-04-18T09:16:41 *** ChanServ sets mode: +o DrJoel 2013-04-18T09:19:45 sebhub: I am looking at the sparc64 and per config.gcc, the target has all the files. ' 2013-04-18T09:21:18 And the patch in rtems-crossrpms does not add it to config.gcc. 2013-04-18T09:21:21 Any ideas? 2013-04-18T09:22:41 sparc64 in gcc repo uses sparc/rtemself.h.. this is the only difference: 2013-04-18T09:22:54 +/* we are not svr4 */ 2013-04-18T09:22:54 +#undef CPP_SUBTARGET_SPEC 2013-04-18T09:22:54 +#define CPP_SUBTARGET_SPEC "" 2013-04-18T09:24:01 but the crossrpms patch does not do anything to config.gcc to use this file 2013-04-18T09:24:21 hm, strange 2013-04-18T09:24:35 i only noticed this file in the patch yesterday 2013-04-18T09:24:57 my goal is to get gcc 4.8 patch free 2013-04-18T09:25:14 I don't dispute that leavign __svr4__ is wrong but the patch doesn't appear to use the file. Do you see it using it? 2013-04-18T09:26:00 i didn't look at it closely 2013-04-18T09:26:25 I was looking at this because the cdefs.h is overwhelming :) 2013-04-18T09:27:00 i was able to build the libbsd tests 2013-04-18T09:27:13 but a lot of them fail with memory exceptions etc. 2013-04-18T09:27:19 great! just minor stuff I saw you committing 2013-04-18T09:27:23 *** the9a3eedi-2 has joined #rtems 2013-04-18T09:27:40 edit config.inc and disable ipv6. core dumps when ipv6 is enabled. 2013-04-18T09:27:40 i enabled ipv6 2013-04-18T09:27:52 ah, ok 2013-04-18T09:28:09 do you have a test for the ping or are there still issues with select()? 2013-04-18T09:28:44 There are issues with select .. but we tested ping using the command line .. hold on and I will cut and paste the command sequence 2013-04-18T09:28:56 *** the9a3eedi has quit IRC (Ping timeout: 256 seconds) 2013-04-18T09:29:18 in the netshell01 ? 2013-04-18T09:31:40 yes.. 2013-04-18T09:32:15 hold on.. let me see if Jennifer has the instructions handier than I do. They appear to be on an external drive for me. :( 2013-04-18T09:32:19 *** rtdin has joined #rtems 2013-04-18T09:34:42 Hello, I am interested in TCP/IP update project in RTEMS for GSOC'13 . I would like to what needs to be done in that ? 2013-04-18T09:35:34 sebhub: Here are Jennifer's notes on testing ipv4 from netshell. 2013-04-18T09:35:35 jmr3904-gdb */*exe 2013-04-18T09:35:35 ifconfig lo0 127.0.0.1 2013-04-18T09:35:35 route add default 127.0.0.1 2013-04-18T09:35:35 ping -c 1 127.0.0.1 2013-04-18T09:36:29 hi rtdin 2013-04-18T09:36:37 I have notes somewhere on configuring the ipv6 loopback and a real network interface. We also brought up at least two Intel ones to steady state. I don't recall trying to program an address though 2013-04-18T09:36:43 rtdin: we recommend avoiding that project right now 2013-04-18T09:36:58 gedare okay 2013-04-18T09:37:08 rtdin: the tcp/ip update project is transitioning toward production use, and until it settles there is not an easy way to scope student-sized projects 2013-04-18T09:37:19 Watch Monty Python's Holy Grail and study the bunny.. looks harmless enough ... but it can kill you :) 2013-04-18T09:38:01 what else are you interested in or have experience with doing? 2013-04-18T09:38:21 gdare thanks for the quick reply. I came to know that ther's a need for classic Condition Variabla APIs. but as i started looked into it .. i feel its very complex for me 2013-04-18T09:38:34 sebhub: I assume you saw my email.. you can play with netshell and ping but I would lean now to adding guts of the current loopback test as a shell command.. configure the network and then invoke it. It doesn't use select(). prove something simple works before select() 2013-04-18T09:38:43 i think another student might be looking at CV already too 2013-04-18T09:38:57 rtdin If that feels complex, you definitely want to avoid the network stack upgrade :) 2013-04-18T09:38:58 I think i only posted that in the mailing list 2013-04-18T09:39:37 sebhub we also need to decide how to configure the stack eventually... run a shell script? generate a shell script and run it? I currently lean to the first. 2013-04-18T09:40:13 rtdin how far along are you in school? 2013-04-18T09:40:15 gedare i also looked at the need for kqueue/taskqueue porting .. is it still open ? 2013-04-18T09:40:49 rtdin: yes that is an open project 2013-04-18T09:40:49 I have completed my bachelor's and have a 2 years experience. Now i have been accepted for my master's program for Fall'13 2013-04-18T09:40:54 looks good so far: 2013-04-18T09:41:02 RTEMS SHELL (Ver.1.0-FRC):/dev/console. Apr 17 2013. 'help' to list commands. 2013-04-18T09:41:04 [/] # ifconfig lo0 127.0.0.1 2013-04-18T09:41:05 ifconfig lo0 127.0.0.1 2013-04-18T09:41:07 [/] # route add default 127.0.0.1 2013-04-18T09:41:08 route add default 127.0.0.1 2013-04-18T09:41:09 add net default: gateway 127.0.0.1 2013-04-18T09:41:11 [/] # ping -c 1 127.0.0.1 2013-04-18T09:41:13 ping -c 1 127.0.0.1 2013-04-18T09:41:14 PING 127.0.0.1 (1.0.0.127): 56 data bytes 2013-04-18T09:41:16 ^C 2013-04-18T09:41:17 Program received signal SIGINT, Interrupt. 2013-04-18T09:41:19 0x880358f8 in _select (numfds=5, readfds=readfds@entry=0x881c4b8c, writefds=writefds@entry=0x0, exceptfds=exceptfds@entry=0x0, timeout=0x881c4b94) at lib/libc_r/uthread/uthread_select.c:69 2013-04-18T09:41:20 69 if ((readfds && FD_ISSET(i, readfds)) || 2013-04-18T09:42:08 That looks like where we were. 2013-04-18T09:42:24 rtdin: ok, i do not see any student listing CVs on the wiki.. btw have you gotten rtems built and posted the patch and screenshot? 2013-04-18T09:42:35 this is great, i want complete the newlib issues 2013-04-18T09:42:41 gedare in the condition variable classic API. i found that the trick part is making it PI aware .. yeah i have posted the patch and screenshot 2013-04-18T09:42:50 i got the wiki account also. i am yet to update 2013-04-18T09:43:03 You can pursue select which is layered with poll. Or you can port the loopback case for a simpler case. I worry select() is a bitch to debug. 2013-04-18T09:43:08 i have only a small budget to evaluate what is needed to satisfy our requirements 2013-04-18T09:43:27 rtdin: PI awareness is a question to us.. we want to know if it is possible. i think sebhub has some interest in the classic api implementation of condition variables 2013-04-18T09:43:46 i have only a small familiarity with the project idea. 2013-04-18T09:43:57 gedare yes. I was looking it at that point only since he mentioned to me . 2013-04-18T09:43:57 yes, condition variables for the classic api would be great 2013-04-18T09:44:05 sebhub ahhh.. ok.. well you duplicated our status and are pushing the newlib. If you want to add a simple Asciidoc and I will help you improve the instructions, it will make it easier for the next step. That's progress. 2013-04-18T09:44:52 senhub yes i was looking at that only. Doing it with PI aware seems a little complex to me. 2013-04-18T09:44:53 you did a great job already 2013-04-18T09:45:54 exists a priority inversion protection protocol for condition variables? 2013-04-18T09:46:42 sebhub wouldn't the mutex have that associated with it? Or do you mean a priority ordered blocking for CVs? 2013-04-18T09:47:00 i saw one paper proposed by darren hart for making a cnodition variable PI aware 2013-04-18T09:47:30 Drjoel sebhub yes it invovles the making the mutex PI aware 2013-04-18T09:47:32 Oh.. that is interesting and the kind of possible extension that is why the Classic API versions of objects in POSIX often more features. 2013-04-18T09:47:57 Our mutexes are PI aware.. I wonder what subtle think I am missing. :( 2013-04-18T09:48:06 *** arvind_k has quit IRC (Ping timeout: 264 seconds) 2013-04-18T09:48:17 I recall the blocking discipline (fifo, priority, etc) is loosely defined in posix. 2013-04-18T09:48:30 rtdin you do know that the POSIX std is online via opengroup.org, 2013-04-18T09:48:38 you release the mutex if you wait for cv signal 2013-04-18T09:49:02 yes.. the cv lets you "wait" at the acquire() stage 2013-04-18T09:49:07 *** SYCrane has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-04-18T09:49:20 DrJoel yes i am aware of it 2013-04-18T09:49:55 well posix is loose.. so we could definitely add discipline and options to this. 2013-04-18T09:49:56 rtdin: you can start with a basic CV implementation 2013-04-18T09:50:00 you have to deal with threads waiting for the cv signal 2013-04-18T09:50:01 and then add the PI awareness 2013-04-18T09:50:38 i think you can implement the basic CV primitive inside rtems supercore (score), and share it with classic and posix 2013-04-18T09:50:41 gedare: thinking out loud.. could the thread set concept help avoid duplicate PI implementations? 2013-04-18T09:50:45 yes, but adding posix like cv should be very simple, its only a new api wrapper for the score 2013-04-18T09:51:06 gedare if i implement the basic CV implementation what difference does it make from POSIX API? 2013-04-18T09:51:14 well we have to move the posix cv impl to the score first 2013-04-18T09:51:25 not much difference, it should be easy to do 2013-04-18T09:51:35 then you go on from there to improve the classic api version 2013-04-18T09:51:41 Just refactoring to enable future work. 2013-04-18T09:51:59 (1) refactor posix cv into score cv and posix cv.. pass posix cv tests 2013-04-18T09:52:05 gedare okay. Drjoel said we have PI aware mutex . 2013-04-18T09:52:10 DrJoel: maybe later it can deal with thread sets, right now we cannot introduce such a dependency 2013-04-18T09:52:16 (2) add classic api cv and basic tests.. 2013-04-18T09:52:32 (3) add capability and promote through Classic API 2013-04-18T09:52:46 yep that seems about right 2013-04-18T09:53:02 gedare DrJoel okay 2013-04-18T09:53:19 gedare: I am not even sure how to have a thread set help with PI.. just wondering it thread queue could know PI. 2013-04-18T09:53:44 oh.. dunno off hand 2013-04-18T09:53:50 rtdin Testing along way is important. you will want coverage reports and to be sure you didn't break anything. :) 2013-04-18T09:54:10 V said i should not try talking with anyone about intelligent things today. i was up for 2 hours around 3 am with A. 2013-04-18T09:54:26 Drjoel yeah okay. 2013-04-18T09:54:53 gedare ... maybe you will hallucinate and have deep insight.. don't hold back 2013-04-18T09:55:24 it may be that a thread set can temporarily hold threads in higher priority mode than their "real" priority 2013-04-18T09:55:32 that will give you de facto PI 2013-04-18T09:56:02 but... 2013-04-18T09:56:12 *** freenix has quit IRC (Remote host closed the connection) 2013-04-18T09:56:21 PI only matters for the thread inside the critical region 2013-04-18T09:56:24 i think? 2013-04-18T09:57:21 I have held to the concept that PI assumes a "resource" that a single thread has. If you view a CV as an extension of a mutex, then that concept still holds. But it breaks down horribly for message queues, counting semaphores 2013-04-18T09:57:24 so the CV should give the priority of the highest priority thread waiting 2013-04-18T09:57:44 *** kkthecoder has joined #rtems 2013-04-18T09:57:57 i think there are other algorithms that can handle those better, like PC 2013-04-18T09:58:00 yep. 2013-04-18T09:58:17 and general resource reservation schemes 2013-04-18T09:58:36 It may be possible to refactor the PI management from the coremutex and use it in both. That would need to be considered.. sure would make it easier to ensrue the same algorithm and behaviour 2013-04-18T09:59:38 https://www.rtems.org/bugzilla/show_bug.cgi?id=1467 2013-04-18T09:59:40 actually, it may be possible to refactor the core mutex to use the CV implementation, because CV generalizes mutexes and semaphores. 2013-04-18T10:01:19 good... i'll put some of this into a wiki page quickly 2013-04-18T10:01:24 then, i should get back to work :) 2013-04-18T10:01:57 gedare sebhub DrJoel thanks for this discussion 2013-04-18T10:02:35 rtdin: i'm starting a wiki page for the project, which you can then update as you learn things 2013-04-18T10:02:43 if you are going to propose the CV project, that is. 2013-04-18T10:02:54 gedare sure I am interested in that 2013-04-18T10:03:12 *** the9a3eedi-2 has quit IRC (Ping timeout: 276 seconds) 2013-04-18T10:03:20 gedare is github still broken for us? 2013-04-18T10:03:24 DrJoel: i've been slowly making changes to the main page of the wiki.. have a look at it to make sure you like it ;) 2013-04-18T10:03:42 ehm, verm__ did some manual updating, i haven't looked lately to see if the auto-updater still isn't working. 2013-04-18T10:05:47 no looks like it still isn't working right 2013-04-18T10:06:21 you are doing a good job. I can't tell what changed without looking at diff's :) 2013-04-18T10:06:52 I would (someday) like that architecture picture to be clickable so each box leads you to the right documentation 2013-04-18T10:09:26 http://wiki.rtems.org/wiki/index.php/Condition_Variables 2013-04-18T10:09:33 um 2013-04-18T10:09:39 i want to remove that arch picture to somewhere else 2013-04-18T10:09:43 it is too much for the front page 2013-04-18T10:10:28 ok.. I don't mind.. I am actually very flexible 2013-04-18T10:10:32 although turning it into a clickable diagram leading to other relevant areas is good 2013-04-18T10:10:47 but it is still a lot to process. i bet most people ignore it or are confused by it 2013-04-18T10:10:51 I think I added that to the small projects page 2013-04-18T10:11:03 k 2013-04-18T10:11:14 gedare thanks 2013-04-18T10:11:15 is http://wiki.rtems.org/wiki/index.php/FAQ useful? 2013-04-18T10:11:36 the questions seem not that common to me 2013-04-18T10:12:02 No.. definitely not so useful :) 2013-04-18T10:12:10 We need better questions and most of those are dead 2013-04-18T10:12:26 k 2013-04-18T10:13:49 *** rtdin has left #rtems 2013-04-18T10:16:48 DrJoel: the webring is DOA 2013-04-18T10:18:42 *** antgreen has quit IRC (Ping timeout: 272 seconds) 2013-04-18T10:19:41 *** S_Somani has joined #rtems 2013-04-18T10:22:22 *** rtdin has joined #rtems 2013-04-18T10:23:42 *** rtdin has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-04-18T10:33:00 *** sebhub has quit IRC (Remote host closed the connection) 2013-04-18T10:53:08 *** weiY has joined #rtems 2013-04-18T11:10:52 :9 on webring.. but webring itself is dead as far as I am concerned.. if you want to remove it.. feel free 2013-04-18T11:12:00 *** weiY has quit IRC (Quit: Leaving...) 2013-04-18T11:12:03 already did it 2013-04-18T11:12:12 now i really should get back to work... but i more or less overhauled the main page 2013-04-18T11:12:35 i moved the picture to a new page by itself that maybe someone will get around to filling... http://wiki.rtems.org/wiki/index.php/RTEMS_Software_Architecture 2013-04-18T11:32:50 *** monstr__ has quit IRC (Read error: Operation timed out) 2013-04-18T11:40:41 *** MegaAlex|away is now known as MegaAlex 2013-04-18T12:47:03 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-18T13:06:50 *** kkthecoder has quit IRC (Ping timeout: 256 seconds) 2013-04-18T13:11:06 *** cdcs has quit IRC (Quit: Page closed) 2013-04-18T13:38:34 *** antgreen has joined #rtems 2013-04-18T14:05:18 *** antgreen has quit IRC (Ping timeout: 258 seconds) 2013-04-18T14:18:23 *** antgreen has joined #rtems 2013-04-18T14:43:58 antgreen: paying attention today? :) 2013-04-18T14:44:11 kind of 2013-04-18T14:44:23 well, not really 2013-04-18T14:44:27 did I miss something? 2013-04-18T14:44:33 No.. just had a question 2013-04-18T14:44:36 shoot 2013-04-18T14:45:03 when I used the function section stuff on sparc sis BSP, I ended up with 80 byte executables. Obviously too small .. any suggestions on what went wrong? 2013-04-18T14:45:12 haha. 2013-04-18T14:45:19 the root set is probably empty 2013-04-18T14:45:25 hmmm 2013-04-18T14:45:46 the linker starts with a known set of "root set" symbols, 2013-04-18T14:45:58 and follows references to link everything together 2013-04-18T14:46:01 The root file should be start.o but I suspect the start.S confuses it.. 2013-04-18T14:46:11 if the root set is wrong for a target, then it won't get anything 2013-04-18T14:46:25 well, it's looking for specific symboles 2013-04-18T14:46:27 well, it's looking for specific symbols 2013-04-18T14:46:43 I'm in a meeting right now, 2013-04-18T14:46:50 but a distraction would be nice. 2013-04-18T14:47:04 let me see if I can find the root symbols in the linker 2013-04-18T14:47:06 source. 2013-04-18T14:49:08 The bsp_specs says start.o which includes "start" 2013-04-18T14:49:14 *** kkthecoder has joined #rtems 2013-04-18T14:50:52 rebuilding now to see what symbols actually make the cut 2013-04-18T14:57:21 bahh... I have to pay attention in this meeting. 2013-04-18T15:03:05 lol 2013-04-18T15:03:37 *** gedare has quit IRC (Ping timeout: 257 seconds) 2013-04-18T15:29:02 *** antgreen has quit IRC (Ping timeout: 272 seconds) 2013-04-18T16:32:51 *** kkthecoder has quit IRC (Quit: Leaving) 2013-04-18T16:34:49 *** kiwichris has joined #rtems 2013-04-18T16:55:15 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-18T17:04:51 *** remustata has joined #rtems 2013-04-18T17:14:53 *** remustata has left #rtems 2013-04-18T17:33:21 *** kiwichris has joined #rtems 2013-04-18T18:05:12 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-18T18:20:02 *** kiwichris has joined #rtems 2013-04-18T18:26:00 *** MegaAlex is now known as MegaAlex|away 2013-04-18T18:27:55 *** weiY has joined #rtems 2013-04-18T18:30:06 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-18T18:32:52 *** kiwichris has joined #rtems 2013-04-18T18:36:38 *** kiwichris has quit IRC (Client Quit) 2013-04-18T19:01:12 *** the9a3eedi has joined #rtems 2013-04-18T19:06:25 * DrJoel thinks it is time for him to start to head home 2013-04-18T19:10:53 *** weiY has quit IRC (Quit: Leaving...) 2013-04-18T19:11:14 *** DrJoel has quit IRC (Quit: Now if you will excuse me, I have a giant ball of oil to throw out my window) 2013-04-18T20:11:21 *** kiwichris has joined #rtems 2013-04-18T20:12:44 *** kiwichris has quit IRC (Client Quit) 2013-04-18T20:20:47 *** kiwichris has joined #rtems 2013-04-18T20:23:53 I found this http://rtemscentre.edisoft.pt/index.php?module=ContentExpress&file=index&func=display&ceid=16&meid=34 2013-04-18T20:24:06 A bunch of RTEMS-related tools that seem extremely useful. 2013-04-18T20:24:27 theyre GPL2 license but you can only access them with permission 2013-04-18T20:25:12 Shouldn't they at least be mentioned in the wiki? I think it would be helpful to users 2013-04-18T20:29:41 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-18T20:39:13 *** weiY has joined #rtems 2013-04-18T21:25:13 *** Robbie has joined #rtems 2013-04-18T21:34:47 *** kiwichris has joined #rtems 2013-04-18T21:43:05 *** weiY has quit IRC (Quit: Leaving...) 2013-04-18T21:46:38 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-18T21:50:48 *** antgreen has joined #rtems 2013-04-18T22:15:08 *** weiY has joined #rtems 2013-04-18T22:20:07 *** weiY has quit IRC (Read error: Connection timed out) 2013-04-18T22:27:49 *** Robbie has quit IRC (Quit: Page closed) 2013-04-18T22:41:44 *** Robbie has joined #rtems 2013-04-18T22:49:30 *** kiwichris has joined #rtems 2013-04-18T22:52:28 *** kiwichris has quit IRC (Client Quit) 2013-04-18T23:52:56 *** S_Somani has joined #rtems 2013-04-19T00:26:46 *** monstr__ has joined #rtems 2013-04-19T00:51:38 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-19T01:08:23 *** arvind_khadri has joined #rtems 2013-04-19T01:27:43 *** MegaAlex|away is now known as MegaAlex 2013-04-19T01:28:54 *** S_Somani has joined #rtems 2013-04-19T01:44:30 *** freenix has joined #rtems 2013-04-19T01:50:44 *** kiwichris has joined #rtems 2013-04-19T02:04:03 *** sebhub has joined #rtems 2013-04-19T02:04:08 good morning 2013-04-19T02:08:17 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-19T02:19:51 *** S_Somani1 has joined #rtems 2013-04-19T02:22:31 *** S_Somani has quit IRC (Ping timeout: 245 seconds) 2013-04-19T02:31:38 hi,sebhub 2013-04-19T02:32:17 i am very interested in Paravirtualization of RTEMS 2013-04-19T02:35:17 are there any references about the paravirtualization of RTEMS? 2013-04-19T02:37:09 sry, i have not looked into this area 2013-04-19T02:41:22 which mentor should i ask for this area? 2013-04-19T02:43:57 i have no idea, i was never interested in this area 2013-04-19T02:46:07 I've seen Gedare discuss it with another person on the mailing list. I'm not sure if he's the mentor, but he might know a thing or too 2013-04-19T02:46:32 thanks a lot 2013-04-19T02:46:56 there's a discussion here if you're interested http://www.rtems.org/pipermail/rtems-users/2013-April/011234.html 2013-04-19T02:56:32 *** S_Somani1 has quit IRC (Quit: Leaving.) 2013-04-19T03:07:23 sebhub: Hi, Thanks for reviewing the patch. 2013-04-19T03:08:43 the bsp_reset I modified is adding d-cache clean operation and mov pc to 0 2013-04-19T03:16:59 I have not added watchdog support now. Sorry, It may really need refactoring.I'll try to make it. Thanks 2013-04-19T03:22:17 we can do this also once the bsp is committed 2013-04-19T03:22:47 Now I am confusing about the common files. 2013-04-19T03:22:55 important is that all files have a clear copyright and license status, since this is hard to change once something is in the repository 2013-04-19T03:22:58 *** Robbie has quit IRC (Ping timeout: 245 seconds) 2013-04-19T03:23:24 *** S_Somani has joined #rtems 2013-04-19T03:24:23 Most of the files that I added is modified from gp32 lpc32xx smdk2410, because as a newbie I do not have enough expirence to make a new wheel from scratch. 2013-04-19T03:25:49 The filename and function name maybe same with other bsps. But inner register configruation and others have been changed, so i am confusing about how to use common files. 2013-04-19T03:26:32 Maybe the bsp_idle and bsp_reset maybe directed to gp32, but not new tiny6410. 2013-04-19T03:27:22 don't worry about the common files, just look if there is something left from the master file and add the appropriate copyright notice 2013-04-19T03:28:51 If I copied ,but modified from other bsps, should I include it's copyright? If so, I'll add it. 2013-04-19T03:30:29 *** kkthecoder has joined #rtems 2013-04-19T03:40:58 *** SYCrane has joined #rtems 2013-04-19T03:47:21 if there is something left from the original file, then yes 2013-04-19T03:59:26 sebhub:Thanks.I'll add it. 2013-04-19T03:59:42 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-19T04:08:03 *** kkthecoder has quit IRC (Ping timeout: 276 seconds) 2013-04-19T04:14:34 *** S_Somani has joined #rtems 2013-04-19T04:17:03 *** the9a3eedi has quit IRC (Ping timeout: 252 seconds) 2013-04-19T04:19:53 *** S_Somani1 has joined #rtems 2013-04-19T04:22:31 *** S_Somani has quit IRC (Ping timeout: 245 seconds) 2013-04-19T05:17:01 *** the9a3eedi has joined #rtems 2013-04-19T05:19:32 *** arvind_khadri has quit IRC (Ping timeout: 256 seconds) 2013-04-19T05:24:18 *** MegaAlex is now known as MegaAlex|away 2013-04-19T05:32:39 *** arvind_khadri has joined #rtems 2013-04-19T06:13:19 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2013-04-19T06:15:32 *** kkthecoder has joined #rtems 2013-04-19T06:25:07 *** kkthecoder has quit IRC (Max SendQ exceeded) 2013-04-19T06:28:39 *** kkthecoder has joined #rtems 2013-04-19T06:36:27 *** kkthecoder has quit IRC (Max SendQ exceeded) 2013-04-19T06:39:20 *** kkthecoder has joined #rtems 2013-04-19T06:48:09 *** kkthecoder has quit IRC (Max SendQ exceeded) 2013-04-19T06:50:38 *** S_Somani1 has quit IRC (Quit: Leaving.) 2013-04-19T06:51:26 *** kkthecoder has joined #rtems 2013-04-19T06:58:46 *** kkthecoder has quit IRC (Ping timeout: 272 seconds) 2013-04-19T07:42:18 *** S_Somani has joined #rtems 2013-04-19T08:07:59 *** Hesham has joined #rtems 2013-04-19T08:17:23 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-19T08:47:38 *** gedare has joined #rtems 2013-04-19T09:04:56 *** kkthecoder has joined #rtems 2013-04-19T09:11:12 *** SYCrane has quit IRC (Read error: Connection reset by peer) 2013-04-19T09:24:46 *** Hesham has quit IRC (Quit: Leaving.) 2013-04-19T09:33:37 *** S_Somani has joined #rtems 2013-04-19T09:35:54 *** weiY has joined #rtems 2013-04-19T09:49:37 *** freenix has quit IRC (Remote host closed the connection) 2013-04-19T09:49:40 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-19T09:49:55 *** S_Somani has joined #rtems 2013-04-19T09:51:23 *** weiY has quit IRC (Ping timeout: 260 seconds) 2013-04-19T09:54:20 *** S_Somani has quit IRC (Ping timeout: 256 seconds) 2013-04-19T10:07:51 *** weiY has joined #rtems 2013-04-19T10:16:26 Hi gedare. 2013-04-19T10:16:31 so quite, hehe 2013-04-19T10:31:34 *** weiY has quit IRC (Quit: Linkinus - http://linkinus.com) 2013-04-19T10:34:15 *** sebhub has quit IRC (Remote host closed the connection) 2013-04-19T10:45:24 *** S_Somani has joined #rtems 2013-04-19T10:53:36 *** Fallenou has quit IRC (Ping timeout: 252 seconds) 2013-04-19T10:55:08 *** Fallenou has joined #rtems 2013-04-19T11:04:55 *** MegaAlex|away is now known as MegaAlex 2013-04-19T11:13:35 *** mgahlot has joined #rtems 2013-04-19T11:29:26 hi 2013-04-19T12:10:23 *** budrys has joined #rtems 2013-04-19T12:10:24 *** mgahlot is now known as mgahlot|away 2013-04-19T12:10:44 hi all. 2013-04-19T12:12:32 I'm interested in porting RTEMS to OMAP3 (BeagleBoard) or OMAP4 (PandaBoard) as GSoC project 2013-04-19T12:12:41 is it suitable and desirable idea? 2013-04-19T12:12:48 or better think of something else? 2013-04-19T12:19:16 hi budrys 2013-04-19T12:19:46 last summer we had both projects, beagleboard and pandboard, we are waiting to merge the result 2013-04-19T12:19:57 budrys: so, better to think of something else :) 2013-04-19T12:20:07 hello dhananjay. 2013-04-19T12:21:10 gedare: hmm... so why there is active project idea on BeagleBoard website? 2013-04-19T12:21:37 oh 2013-04-19T12:21:52 sorry, I was reading previous year ideas page 2013-04-19T12:21:53 my fault 2013-04-19T12:21:54 :) 2013-04-19T12:22:31 ah it's ok 2013-04-19T12:23:29 budrys, did you already go through our GSoC getting started guide? 2013-04-19T12:24:41 gedare: yes. I've managed to run modify and run "hello world" application 2013-04-19T12:25:04 , 2013-04-19T12:28:00 gedare: I'm also interested in MMU support idea and I'm currently looking at work done at this area and informations about RTEMS memory model on MMU-enabled CPUs 2013-04-19T12:29:34 budrys, there is a student who will propose MMU project already, so that is better to avoid 2013-04-19T12:29:56 :) check the table on our SOC wiki page to see what other students are proposing 2013-04-19T12:30:00 we prefer to avoid overlap. 2013-04-19T12:31:28 gedare: so it's better not to cause you to choose between two students proposing the same idea? 2013-04-19T12:31:36 yes 2013-04-19T12:31:40 better for you too. 2013-04-19T12:57:16 budrys, please send your hello world patch and screenshot to the rtems-devel mailing list, and add yourself to the students table at http://wiki.rtems.org/wiki/index.php/RTEMSSummerOfCode 2013-04-19T12:57:44 and you can discuss project ideas on rtems-devel and get more details once you find a suitable project to propose 2013-04-19T13:33:32 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-19T14:12:19 *** monstr__ has quit IRC (Ping timeout: 258 seconds) 2013-04-19T14:23:10 *** xMYTHICx has joined #rtems 2013-04-19T15:41:54 *** gedare has quit IRC (Ping timeout: 272 seconds) 2013-04-19T16:00:13 *** kkthecoder has quit IRC (Quit: Leaving) 2013-04-19T16:33:23 *** kiwichris has joined #rtems 2013-04-19T16:54:38 *** mgahlot has joined #rtems 2013-04-19T16:58:12 *** mgahlot|away has quit IRC (Ping timeout: 256 seconds) 2013-04-19T18:51:49 *** mgahlot is now known as mgahlot|away 2013-04-19T19:08:38 *** MegaAlex is now known as MegaAlex|away 2013-04-19T20:38:05 *** mgahlot has joined #rtems 2013-04-19T20:40:13 *** mgahlot|away has quit IRC (Ping timeout: 240 seconds) 2013-04-19T20:43:32 *** mgahlot|away has joined #rtems 2013-04-19T20:45:49 *** mgahlot has quit IRC (Ping timeout: 245 seconds) 2013-04-19T22:15:12 *** dhananjay has quit IRC (*.net *.split) 2013-04-19T22:15:33 *** dhananjay has joined #rtems 2013-04-19T23:36:52 *** S_Somani has joined #rtems 2013-04-20T00:06:55 *** kkthecoder has joined #rtems 2013-04-20T00:17:53 *** rtemsLogger` has joined #rtems 2013-04-20T00:18:53 *** S_Somani has quit IRC (Ping timeout: 240 seconds) 2013-04-20T00:19:20 *** S_Somani has joined #rtems 2013-04-20T00:24:37 *** MegaAlexz has joined #rtems 2013-04-20T00:25:28 *** kkthecoder has quit IRC (*.net *.split) 2013-04-20T00:25:28 *** mgahlot|away has quit IRC (*.net *.split) 2013-04-20T00:25:29 *** MegaAlex|away has quit IRC (*.net *.split) 2013-04-20T00:25:29 *** verm__ has quit IRC (*.net *.split) 2013-04-20T00:25:29 *** rtemsLogger has quit IRC (*.net *.split) 2013-04-20T00:32:42 *** mgahlot is now known as mgahlot|away 2013-04-20T00:34:06 *** kkthecoder has joined #rtems 2013-04-20T00:41:25 *** S_Somani1 has joined #rtems 2013-04-20T00:43:32 *** S_Somani has quit IRC (Ping timeout: 252 seconds) 2013-04-20T00:46:39 *** freenix has joined #rtems 2013-04-20T00:53:22 *** kkthecoder has quit IRC (Read error: Connection reset by peer) 2013-04-20T00:54:15 *** S_Somani has joined #rtems 2013-04-20T00:55:52 *** S_Somani1 has quit IRC (Ping timeout: 252 seconds) 2013-04-20T01:13:25 *** S_Somani1 has joined #rtems 2013-04-20T01:13:32 *** S_Somani has quit IRC (Ping timeout: 260 seconds) 2013-04-20T01:44:03 *** verm__ has joined #rtems 2013-04-20T01:55:41 *** mgahlot|away has quit IRC (Read error: Connection reset by peer) 2013-04-20T01:58:22 *** mgahlot has joined #rtems 2013-04-20T02:02:44 *** S_Somani1 has quit IRC (Ping timeout: 252 seconds) 2013-04-20T02:35:57 kiwichris: Hi,I am confusing now, because I do not know how to make fine-grained tasks division for RTL host and target. 2013-04-20T02:36:17 I grap the meaning that elf object files should be arranged to a rap file in host side. Different sections should be merged and a relocation table may needs rebuilding. 2013-04-20T02:36:19 Do you mean which part does which bit ? 2013-04-20T02:37:06 I mean fined-grained tasks in RTL host and fined grained tasks in target. 2013-04-20T02:38:36 The RAP has a format. The host creates that format. 2013-04-20T02:38:49 The host needs to convert the ELF format to the RAP format 2013-04-20T02:39:35 sections in different object files should be merged ? 2013-04-20T02:40:07 The need for the RAP format rather than ELF format is the RAP format is compressed and this means you can only stream the data from the compressed to non-compressed side. 2013-04-20T02:40:34 The ELF format needs to be able to seek and you cannot seek a compressed file unless you uncompress it first. 2013-04-20T02:40:39 Does this make sense ? 2013-04-20T02:41:25 Yes. Thanks. 2013-04-20T02:42:10 The RAP format needs to be in an order that allows it to load 2013-04-20T02:42:42 To do this sections and merged into the only ones we need. 2013-04-20T02:43:12 In the ELF loader the different sections are loaded into .text, .data, .const, and .bss 2013-04-20T02:43:15 nothing more. 2013-04-20T02:43:35 All that is happening is this task is now happening on the host rather than the target. 2013-04-20T02:43:53 I have another question. Different object files have relocation tables. how to do with this on the host side? also merged into one? 2013-04-20T02:44:41 Thanks.But one object file may have many sections not just text.data.const and bss 2013-04-20T02:44:41 They are merged together as well 2013-04-20T02:45:12 In ELF yes and in the target ELF loader they all end up in the RAP listed ones 2013-04-20T02:45:50 Check rtl-elf.c for the details. 2013-04-20T02:46:14 ok. 2013-04-20T02:46:54 So the merging on the host is about iterating over the object files and figuring out the size of each of the RAP sections 2013-04-20T02:48:22 on the host side, if I want to implent arm support, should I mainly focus on the arch specific sections ? 2013-04-20T02:48:29 The complication is each section can have a specific alignment. A lot of the code in the RAP converter on the host is managing this. 2013-04-20T02:48:57 ARM should be covered because it is just ELF sections. The is nothing specific ?. I know at the moment 2013-04-20T02:49:20 The only ARM specific code should be the relocation record processing on the target 2013-04-20T02:49:26 As far as I :) 2013-04-20T02:49:56 We may find I am wrong and we need specific arch code but we need to try very hard to avoid this 2013-04-20T02:50:33 So you need look with readelf at the specific section in ARM that is failing and have a look. 2013-04-20T02:50:56 I am reading elf for arm mannual now. 2013-04-20T02:51:16 Just use readelf and see what it says about the section. 2013-04-20T02:51:35 Paste that specific bit to somewhere so I can have a look as well 2013-04-20T02:52:01 I'll post it on the mailing list. 2013-04-20T02:52:22 No need, just a pastebin thingy 2013-04-20T02:52:40 What is the github one, it has no ads 2013-04-20T02:52:49 https://gist.github.com/ 2013-04-20T02:52:52 ok. 2013-04-20T02:56:18 https://gist.github.com/anonymous/5425184 2013-04-20T03:09:57 Sorry was distracted 2013-04-20T03:10:21 Hmm 2013-04-20T03:12:49 Which section the error on ? 2013-04-20T03:13:25 ARM.attributes 2013-04-20T03:13:32 Thanks 2013-04-20T03:13:57 Error msg:error: libelf:elf_getdata: .ARM.attributes(xa.c.8.o): Invalid section descriptor 2013-04-20T03:14:17 What does the ARM ABI say about this section ? 2013-04-20T03:14:27 Also which doc are you reading for the ARM ABI ? 2013-04-20T03:16:05 The doc is here http://infocenter.arm.com/help/topic/com.arm.doc.ihi0044e/IHI0044E_aaelf.pdf 2013-04-20T03:16:26 thanks 2013-04-20T03:17:29 It is object file compatibility attributes. We can ignore it. 2013-04-20T03:18:04 oh. 2013-04-20T03:18:21 Our systems are simpler because the user should be in control of objects they use. Adding support for this at this point in time it not important. 2013-04-20T03:18:44 see page 17 of the doc 2013-04-20T03:18:56 Now how to ignore the section. 2013-04-20T03:20:51 I do not know what I should do with the host side. I have add this specific section related work in my proposal draft. That maybe removed. 2013-04-20T03:24:39 I am looking at the code now :) 2013-04-20T03:24:51 Lets just fix it and move onto more interesting things :) 2013-04-20T03:25:52 Thanks. I traced the code.It fails here. _libelf_xlate_shtype elf_data.c line 91 2013-04-20T03:26:24 Thanks. I will take a look. 2013-04-20T03:26:31 the sh_type is 0x70000003.But _libelf_xlate_shtype does not handle this. 2013-04-20T03:26:50 0x7000003 is the type of ARM.attributes 2013-04-20T03:27:34 Hmm ok. I wonder if upstream handles it 2013-04-20T03:29:33 I have to leave for a while.Back later. 2013-04-20T03:29:53 Thanks for all this work. 2013-04-20T03:30:35 No change. I think we need to handle this in the rtems-ld 2013-04-20T03:31:38 Hmm. only by do this work, I can learn how to handle with this project in coming gsoc. 2013-04-20T03:32:23 ok. I will also try to trace it. But now I have to leave. 2013-04-20T03:32:36 Once this is sorted the next part of the ARM relocation processing on the target. 2013-04-20T03:32:44 I will take a look soon 2013-04-20T03:32:50 Maybe tomorrow my time 2013-04-20T03:37:40 *** S_Somani has joined #rtems 2013-04-20T04:19:29 *** kkthecoder has joined #rtems 2013-04-20T05:27:14 *** sevikkk has quit IRC (Ping timeout: 256 seconds) 2013-04-20T05:32:05 *** sevikkk has joined #rtems 2013-04-20T06:13:15 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-20T07:01:16 *** mgahlot has quit IRC (Remote host closed the connection) 2013-04-20T07:08:52 *** S_Somani has joined #rtems 2013-04-20T07:24:25 *** weiY has joined #rtems 2013-04-20T08:14:44 *** zhangwenjie has joined #rtems 2013-04-20T08:14:45 *** weiY has quit IRC (Read error: Connection reset by peer) 2013-04-20T08:53:30 *** S_Somani has quit IRC (Ping timeout: 264 seconds) 2013-04-20T08:54:30 *** S_Somani has joined #rtems 2013-04-20T08:54:53 *** freenix has quit IRC (Remote host closed the connection) 2013-04-20T09:09:12 *** mgahlot has joined #rtems 2013-04-20T09:33:00 *** mgahlot is now known as mgahlot|away 2013-04-20T09:44:30 *** zhangwenjie has quit IRC (Read error: Connection reset by peer) 2013-04-20T09:44:56 *** weiY has joined #rtems 2013-04-20T10:03:18 *** kkthecoder has quit IRC (Ping timeout: 256 seconds) 2013-04-20T10:26:36 *** kkthecoder has joined #rtems 2013-04-20T10:57:38 Hi, anyone is here? 2013-04-20T10:57:59 how i update the toolchains for fedora 14 2013-04-20T10:58:32 it seems that there are no repository for fedora14 only 16-19 exsites 2013-04-20T11:04:52 *** weiY has quit IRC (Quit: Linkinus - http://linkinus.com) 2013-04-20T11:07:41 *** kkthecoder has quit IRC (Read error: Connection reset by peer) 2013-04-20T11:08:59 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-20T11:44:58 *** monstr__ has joined #rtems 2013-04-20T11:51:21 *** S_Somani has joined #rtems 2013-04-20T11:53:32 *** kkthecoder has joined #rtems 2013-04-20T11:58:29 *** S_Somani has left #rtems 2013-04-20T12:27:50 *** S_Somani has joined #rtems 2013-04-20T12:42:09 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-20T12:54:26 *** monstr__ has quit IRC (Ping timeout: 256 seconds) 2013-04-20T14:00:46 *** mgahlot has joined #rtems 2013-04-20T14:03:46 *** mgahlot|away has quit IRC (Ping timeout: 245 seconds) 2013-04-20T14:04:14 *** mgahlot|away has joined #rtems 2013-04-20T14:07:43 *** mgahlot|away_ has joined #rtems 2013-04-20T14:07:56 *** mgahlot has quit IRC (Ping timeout: 245 seconds) 2013-04-20T14:10:51 *** mgahlot|away has quit IRC (Ping timeout: 245 seconds) 2013-04-20T14:11:18 *** mgahlot has joined #rtems 2013-04-20T14:14:36 *** mgahlot|away_ has quit IRC (Ping timeout: 245 seconds) 2013-04-20T14:14:55 *** mgahlot|away has joined #rtems 2013-04-20T14:18:21 *** mgahlot has quit IRC (Ping timeout: 245 seconds) 2013-04-20T14:18:40 *** mgahlot has joined #rtems 2013-04-20T14:22:06 *** mgahlot|away has quit IRC (Ping timeout: 245 seconds) 2013-04-20T14:27:07 *** mgahlot is now known as mgahlot|away 2013-04-20T14:52:57 *** mgahlot has joined #rtems 2013-04-20T14:53:46 *** mgahlot|away has quit IRC (Ping timeout: 245 seconds) 2013-04-20T14:56:17 *** mgahlot|away has joined #rtems 2013-04-20T14:59:20 *** mgahlot has quit IRC (Ping timeout: 252 seconds) 2013-04-20T17:47:13 *** kkthecoder has quit IRC (Quit: Leaving) 2013-04-20T18:58:52 *** MegaAlexz is now known as MegaAlex|away 2013-04-20T18:59:04 *** SYCrane has joined #rtems 2013-04-20T20:19:49 *** mgahlot has joined #rtems 2013-04-20T20:22:27 *** mgahlot|away has quit IRC (Ping timeout: 260 seconds) 2013-04-20T22:28:56 *** S_Somani has joined #rtems 2013-04-20T22:48:09 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-20T23:30:37 *** weiY has joined #rtems 2013-04-20T23:53:44 *** S_Somani has joined #rtems 2013-04-21T00:01:05 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-04-21T00:01:43 *** the9a3eedi has joined #rtems 2013-04-21T00:28:58 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-21T00:40:21 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-21T00:43:48 *** kiwichris has joined #rtems 2013-04-21T01:34:14 *** weiY has quit IRC (Quit: Leaving...) 2013-04-21T02:06:40 *** weiY has joined #rtems 2013-04-21T02:18:58 *** mgahlot|away has joined #rtems 2013-04-21T02:22:06 *** mgahlot has quit IRC (Ping timeout: 245 seconds) 2013-04-21T02:48:43 *** weiY has quit IRC (Ping timeout: 246 seconds) 2013-04-21T03:06:21 *** weiY has joined #rtems 2013-04-21T03:10:57 *** weiY has quit IRC (Ping timeout: 258 seconds) 2013-04-21T03:33:30 *** MegaAlex|away is now known as MegaAlex 2013-04-21T03:50:33 *** hiddenpearls has joined #rtems 2013-04-21T04:17:16 *** mgahlot|away has quit IRC (Read error: Connection reset by peer) 2013-04-21T04:17:33 *** mgahlot|away has joined #rtems 2013-04-21T05:09:36 *** SYCrane has quit IRC (Ping timeout: 245 seconds) 2013-04-21T05:11:18 *** SYCrane has joined #rtems 2013-04-21T06:07:31 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-21T06:17:20 *** hiddenpearls has joined #rtems 2013-04-21T06:18:49 *** S_Somani has joined #rtems 2013-04-21T06:44:39 *** S_Somani1 has joined #rtems 2013-04-21T06:45:31 *** S_Somani has quit IRC (Ping timeout: 264 seconds) 2013-04-21T06:47:27 can we do RTEMS installation on MAC ? 2013-04-21T06:49:18 Yes. I am using a mac 2013-04-21T06:49:28 You can build on MacOS 2013-04-21T06:49:38 You cannot run RTEMS on a Mac 2013-04-21T07:16:49 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-21T07:17:23 *** mgahlot|away has quit IRC (Remote host closed the connection) 2013-04-21T07:54:56 *** kkthecoder has joined #rtems 2013-04-21T08:44:56 *** kkthecoder has quit IRC (Ping timeout: 260 seconds) 2013-04-21T08:58:29 *** kkthecoder has joined #rtems 2013-04-21T09:00:21 *** S_Somani1 has quit IRC (Quit: Leaving.) 2013-04-21T09:14:38 *** SYCrane has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-04-21T09:25:32 *** S_Somani has joined #rtems 2013-04-21T09:56:55 *** weiY has joined #rtems 2013-04-21T10:06:48 *** weiY has quit IRC (Ping timeout: 264 seconds) 2013-04-21T10:24:20 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-21T10:51:25 *** gedare has joined #rtems 2013-04-21T10:53:59 *** S_Somani has joined #rtems 2013-04-21T11:18:16 *** S_Somani has left #rtems 2013-04-21T11:19:03 *** S_Somani has joined #rtems 2013-04-21T11:33:31 *** mgahlot|away has joined #rtems 2013-04-21T11:35:59 *** gedare has quit IRC (Read error: Operation timed out) 2013-04-21T11:58:32 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-21T12:15:18 *** hiddenpearls has joined #rtems 2013-04-21T12:25:13 *** Shubham has joined #rtems 2013-04-21T12:38:37 *** Shubham has left #rtems 2013-04-21T12:38:46 *** Shubham has joined #rtems 2013-04-21T13:09:44 *** arvind_khadri has joined #rtems 2013-04-21T13:22:43 *** anomaly404 has joined #rtems 2013-04-21T13:24:42 *** anomaly404 has left #rtems 2013-04-21T13:40:01 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-21T13:58:25 *** Shubham has left #rtems 2013-04-21T14:19:02 *** mgahlot|away_ has joined #rtems 2013-04-21T14:22:30 *** mgahlot|away has quit IRC (Ping timeout: 272 seconds) 2013-04-21T14:23:58 *** mgahlot|away_ has quit IRC (Remote host closed the connection) 2013-04-21T14:24:29 *** mgahlot|away has joined #rtems 2013-04-21T14:25:51 *** kkthecoder has quit IRC (Ping timeout: 245 seconds) 2013-04-21T14:30:46 *** MegaAlex is now known as MegaAlex|away 2013-04-21T14:38:09 *** hiddenpearls has joined #rtems 2013-04-21T14:53:01 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-21T14:59:40 *** mgahlot|away is now known as mgahlot 2013-04-21T16:03:14 *** kkthecoder has joined #rtems 2013-04-21T17:09:30 *** mgahlot has quit IRC (Read error: Connection reset by peer) 2013-04-21T17:27:51 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-21T17:28:35 *** kiwichris has joined #rtems 2013-04-21T18:34:58 *** the9a3eedi has quit IRC (Ping timeout: 256 seconds) 2013-04-21T19:07:59 *** the9a3eedi has joined #rtems 2013-04-21T19:19:20 *** hiddenpearls has joined #rtems 2013-04-21T19:21:12 *** hiddenpearls has quit IRC (Client Quit) 2013-04-21T19:24:10 *** hiddenpearls has joined #rtems 2013-04-21T19:29:24 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-21T20:05:52 *** freenix has joined #rtems 2013-04-21T20:13:52 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-04-21T20:21:22 *** the9a3eedi has joined #rtems 2013-04-21T20:21:26 *** hiddenpearls has joined #rtems 2013-04-21T20:32:03 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-04-21T20:34:16 kiwichris: So, my mac is enough to do RTEMS stuff for learning/GSOC etc ?? 2013-04-21T20:34:48 I saw the post however have not taken the time to read the links. 2013-04-21T20:35:12 Sebastiain know more about the PPC so his comments are worth more than mine. 2013-04-21T20:35:34 *** the9a3eedi has joined #rtems 2013-04-21T20:35:48 I am not sure about the merits for GSoC of a specific project. They are all reviewed and judged by the mentors. 2013-04-21T20:55:25 ok 2013-04-21T21:30:42 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-21T22:50:23 *** hiddenpearls has joined #rtems 2013-04-21T23:09:12 *** arvind_khadri has quit IRC (Ping timeout: 264 seconds) 2013-04-21T23:48:51 *** hiddenpearls has quit IRC (Ping timeout: 260 seconds)