2013-04-29T00:00:24 *** monstr has joined #rtems 2013-04-29T00:19:44 *** hiddenpearls1 has joined #rtems 2013-04-29T00:21:17 *** hiddenpearls has quit IRC (Ping timeout: 256 seconds) 2013-04-29T00:25:40 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2013-04-29T00:45:10 *** hiddenpearls has joined #rtems 2013-04-29T01:19:01 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-29T01:19:33 *** hiddenpearls has joined #rtems 2013-04-29T01:47:06 *** mgahlot|away has joined #rtems 2013-04-29T01:57:10 *** mgahlot|away has quit IRC (Remote host closed the connection) 2013-04-29T02:03:19 *** sebhub has joined #rtems 2013-04-29T02:03:25 good morning 2013-04-29T02:20:45 *** weiY has joined #rtems 2013-04-29T02:32:22 kiwichris, hi 2013-04-29T02:35:34 freenix, hi 2013-04-29T02:36:05 I have a question about how to process local symbols 2013-04-29T02:36:38 The linker does not include local symbols in the final rap file. 2013-04-29T02:37:21 Now I am trying to make RTL work on mips jmr3904. 2013-04-29T02:38:15 Each object file contains an local symbol "$LC0" 2013-04-29T02:39:07 If not include this symbol, the loader will complain "unresolved symbol" 2013-04-29T02:39:27 *** S_Somani has joined #rtems 2013-04-29T02:39:49 What is the purpose of these symbols ? Do they have special attributes ? 2013-04-29T02:40:50 The local "$LC*" symols points to rodata section 2013-04-29T02:42:56 And the relocation table needs the symbols. for example, It likes: mov r0, $LC0; b printf; 2013-04-29T02:46:12 Oh it is to a string which is local. What do the symbols in the rodata section look like ? 2013-04-29T02:46:37 I wonder if a symbol label fiddle is needed 2013-04-29T02:47:31 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-29T02:48:06 I paste it here:https://gist.github.com/MrVan/5480244 2013-04-29T02:48:53 I have included these local symbols in the final rap file, But I am not sure whether this is the right way. 2013-04-29T02:54:41 The type and binding of the $LC* symbols are STT_NOTYPE and STB_LOCAL 2013-04-29T02:58:26 *** S_Somani1 has joined #rtems 2013-04-29T02:59:48 hmm, it works on mips now with only 3 reloc type supporting now. But the bad side is the rap file may include several $LC0 symbols. 2013-04-29T03:00:31 *** S_Somani has quit IRC (Ping timeout: 256 seconds) 2013-04-29T03:01:04 There maybe a need to add code to resolve these. 2013-04-29T03:01:23 *** weiY has quit IRC (Ping timeout: 255 seconds) 2013-04-29T03:01:57 Locals should not need to exist in the final RAP image 2013-04-29T03:02:08 But this can be something added later 2013-04-29T03:14:18 Thanks. I'll try to resolve this, but now it does not matter when loading. 2013-04-29T03:26:55 *** phipse has joined #rtems 2013-04-29T03:34:58 freenix, yeap 2013-04-29T03:35:14 yes? 2013-04-29T03:36:58 hi, kiwichris 2013-04-29T03:39:09 I have pasted mips related patch here:https://gist.github.com/MrVan . Thanks for your reviewing. Only a basic implementation now. 2013-04-29T03:39:38 *** S_Somani1 is now known as S_Somani 2013-04-29T03:39:43 it does not matter when loading .. it only saves space. 2013-04-29T03:40:00 I cannot review. I am off out for the evening 2013-04-29T03:40:10 It will happen later 2013-04-29T03:40:38 Thanks. If it does not matter, I'll just include them. 2013-04-29T03:42:14 Cool, cya 2013-04-29T03:42:32 hmm 2013-04-29T03:53:30 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-29T04:02:18 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-04-29T04:29:00 *** weiY has joined #rtems 2013-04-29T04:57:48 *** cdcs has joined #rtems 2013-04-29T05:03:39 *** arvind_khadri has joined #rtems 2013-04-29T05:06:35 *** freenix has quit IRC (Remote host closed the connection) 2013-04-29T05:09:49 *** S_Somani has joined #rtems 2013-04-29T05:18:04 *** hiddenpearls has joined #rtems 2013-04-29T06:33:54 *** hiddenpearls1 has joined #rtems 2013-04-29T06:36:23 *** hiddenpearls has quit IRC (Ping timeout: 252 seconds) 2013-04-29T06:38:56 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2013-04-29T06:46:03 *** hiddenpearls has joined #rtems 2013-04-29T07:02:56 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-29T07:36:49 *** hiddenpearls has joined #rtems 2013-04-29T07:40:34 *** antgreen has quit IRC (Ping timeout: 258 seconds) 2013-04-29T07:55:58 *** weiY has quit IRC () 2013-04-29T08:13:35 *** weiY has joined #rtems 2013-04-29T08:27:48 *** arvind_khadri has quit IRC (Ping timeout: 260 seconds) 2013-04-29T08:36:17 *** gedare has joined #rtems 2013-04-29T08:44:17 *** weiY has quit IRC (Quit: Leaving...) 2013-04-29T08:45:02 *** weiY has joined #rtems 2013-04-29T08:59:04 *** peerst has quit IRC (Read error: Connection reset by peer) 2013-04-29T08:59:39 *** peerst has joined #rtems 2013-04-29T09:11:57 *** hiddenpearls1 has joined #rtems 2013-04-29T09:14:03 *** hiddenpearls has quit IRC (Ping timeout: 256 seconds) 2013-04-29T09:29:13 *** DrJoel has joined #rtems 2013-04-29T09:29:14 *** DrJoel has joined #rtems 2013-04-29T09:29:14 *** ChanServ sets mode: +o DrJoel 2013-04-29T10:32:30 *** sebhub has quit IRC (Remote host closed the connection) 2013-04-29T10:54:07 *** weiY has quit IRC (Ping timeout: 252 seconds) 2013-04-29T11:06:02 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2013-04-29T11:15:27 quiet in here today 2013-04-29T11:25:28 *** cdcs has quit IRC (Ping timeout: 245 seconds) 2013-04-29T11:30:26 DrJoel: yea...I worked on the suggestions given by you... 2013-04-29T11:30:51 :) 2013-04-29T11:31:43 I seem to have been in a todo list/work break down mode the past few days. I was trying to put a Project together for the RTEMS Source Builder and realized that doing that with each GSOC project would help you all. 2013-04-29T11:32:21 * gedare bangs some pots and pans 2013-04-29T11:33:09 that sounds more like something the little one would do. 2013-04-29T11:33:24 i'm a child at heart. 2013-04-29T11:33:48 I needed some advice. In which direction should I move next...should I work for the MVC framework first or should I start building an application with curses? 2013-04-29T11:34:07 For you gedare.. someday you will understand .. lol ... http://www.gocomics.com/herman/2013/04/25 2013-04-29T11:34:25 i get it already. 2013-04-29T11:35:17 S_Somani: I don't know that which UI you do first matters. I would tend to work on the simplest one to implement first to get it out and get the most feedback. 2013-04-29T11:36:02 i saw you doing all that project planning and looking at simulators.. i had a thought of something else you could do if you're bored, but I forgot it. 2013-04-29T11:36:35 Back to our discussion on ensuring the Type strings in conf.t were basically an enumerated set. Do you know that set? I would really like to add this to conf.t and have your conversion program barf if it wasn't on the list. Then do the same with Ranges. 2013-04-29T11:36:51 hmm, a rapid prototype of both the GUI and TUI could be helpful just to have some idea.. 2013-04-29T11:36:59 but you may end up having to throw things away when you get to the MVC 2013-04-29T11:37:07 gedare: no thanks.. I also rebuilt a toilet this weekend. But the weather was rainy 2013-04-29T11:37:32 maybe you should have checked your roof?? 2013-04-29T11:37:43 gedare is right.. having a prototype of both will help you separate View from underlying logic. You could get to the end of either UI and unintentionally have it all intertwined 2013-04-29T11:38:10 and also a python CLI i think will be quite useful more and more 2013-04-29T11:38:47 It is pretty new.. should clean the gutters.. but cheaper to pay someone than fall from about 3.5 stories onto concrete. House is basement plus two stories and driveway in back slopes away from house.. rough fall 2013-04-29T11:38:50 i'm not sure yet if it might become overcomplicated 2013-04-29T11:39:17 DrJoel: My exams were nearing in so I wished do more work before the exams. I will try to make a small proto of curses too then. On conf.t- I havent really had a detailed study on it now. :( I thought I would do it in the soc. (assuming I got selected) 2013-04-29T11:39:18 i meant if the weather was rainy then maybe it wasn't your toilet that was leaking. :) 2013-04-29T11:39:39 S_Somani: Work on defining the set of tasks you will do and approximate order to do them 2013-04-29T11:39:41 S_Somani: you are doing the work. Whatever gives you the most confidence you have the right classes and separation 2013-04-29T11:39:54 we're suggesting the next thing you might want to do regardless of your time constraints 2013-04-29T11:40:12 but really, the "contract" to do work does not go into effect until Google says so... 2013-04-29T11:40:19 ok on conf.t.. I just think it is a good early task. You add a "transform" for each and we ensure the set is known, defined, and finite. I just want to make your work easier in that respect 2013-04-29T11:40:22 So you just do whatever amount of work you think you need to have a good proposal. 2013-04-29T11:40:28 +1 2013-04-29T11:43:42 I will see the conf.t also then and will try to identify which values are not coming in the set. 2013-04-29T11:44:30 But Gedare is right (1) school is your focus now and (2) planning the project is what is in this phase. 2013-04-29T11:47:03 Ok :) 2013-04-29T11:49:47 *** hiddenpearls has joined #rtems 2013-04-29T12:07:50 *** antgreen has joined #rtems 2013-04-29T12:26:24 antgreen: I played with function sections on sparc/sis and got close. Any hints on how to know what will be missed by ld? I had to add ENTRY() to linkcmds. I suspect the IRQ HW table stuff is special as well. 2013-04-29T12:29:22 "you may have problems with debugging if you specify both this option and -g. " 2013-04-29T12:29:49 in sparc the IRQ HW table goes in its own section already i think .... .trap_table 2013-04-29T12:30:08 maybe~ i forgot and am lazy to look 2013-04-29T12:30:32 This option has been around so long, I suspect that that warning is out of date. But we can learn. :) 2013-04-29T12:33:33 hmmm.. maybe on sparc64. I don't see any .section directives under libbsp/sparc 2013-04-29T12:33:44 building now to see what might be missing 2013-04-29T12:38:23 Hmm.. why would the C++ global ctors INIT call core dump 2013-04-29T12:38:26 127 if (doCons) /* && (volatile void *)_init) */ { 2013-04-29T12:38:27 (gdb) 2013-04-29T12:38:27 128 INIT_NAME (); 2013-04-29T12:38:27 (gdb) 2013-04-29T12:38:27 Breakpoint 1, _Internal_error_Occurred ( 2013-04-29T12:38:27 the_source=the_source@entry=RTEMS_FATAL_SOURCE_EXCEPTION, 2013-04-29T12:38:27 is_internal=is_internal@entry=false, the_error=the_error@entry=33670232) 2013-04-29T12:38:27 at ../../../../../../rtems/c/src/../../cpukit/score/src/interr.c:34 2013-04-29T12:39:36 Looks like we need KEEP, etc like sparc-rtems4.11/ldscripts/... 2013-04-29T12:40:36 yeah. they expect to be in a certain place 2013-04-29T12:40:49 and are not referenced directly because the language generates the references 2013-04-29T12:40:52 i think 2013-04-29T12:40:58 I think we are missing a lot of KEEP() directives. 2013-04-29T12:41:02 probably! 2013-04-29T12:41:23 We will only be able to use function sections on BSPs with the shared linker scripts because the effort to fix otherwise is too high\ 2013-04-29T12:41:29 More reason to push for shared 2013-04-29T12:42:10 yeah 2013-04-29T12:42:15 on sparc64 we did KEEP on init and fini 2013-04-29T12:42:28 but.. really we just did whatever we felt we had to do to get something working. 2013-04-29T12:45:44 I have sis working now with function-sections. I think I will do a build with and without to compare sizes 2013-04-29T12:46:38 DrJoel : sorry , I am sitting with a customer right now. It looks like you got it working, right? 2013-04-29T12:48:05 yeah! Pay attention to paying customers :) 2013-04-29T12:48:10 The world needs more of them 2013-04-29T13:42:38 DrJoel: thanks for your hints. I used caligraplan to build a tasklist, schedule and gantt. This draw out some issues, I haven't seen so far. I hope my proposal is on the way to become the detailed plan you want to read. 2013-04-29T13:43:28 hmmm.. caligraplan is a new one to me. Googleing it now 2013-04-29T13:44:16 caligra is the 'new' KDE office suite 2013-04-29T13:44:48 Looks like it has the right capabilities. Out of curiousity, does it handle MS-Project .plan files? 2013-04-29T13:45:19 I picked ProjectLibre because it had Linux, mac, and Windows versions 2013-04-29T13:46:06 Well the file is called .plan, but it doesn't look like a MS file type 2013-04-29T13:46:26 *** antgreen has quit IRC (Ping timeout: 268 seconds) 2013-04-29T13:46:56 I tried ProjectLibre, but it didn't like to show anything in that grey window. 2013-04-29T13:47:45 It may be the format. Not really important. It just helps you organize the tasks, break them down, order them. 2013-04-29T13:48:09 Programmers logically do that anyway. This just organizes it and makes it harder to miss something 2013-04-29T13:50:05 I haven't done it this way. But I haven't had any serious projects so far. So it's a good new experience, as I noticed. 2013-04-29T13:50:46 :) that's what gsoc is about 2013-04-29T13:51:06 *** monstr has quit IRC (Ping timeout: 264 seconds) 2013-04-29T13:53:28 You consider the wiki page when deciding on the proposal, right? 2013-04-29T13:54:02 Yes, we look at the proposal in melange, the google doc, the wiki page, and other resources you reference in your proposal. 2013-04-29T13:54:23 If you are updating it. But mostly it is the proposal and interaction with the student. We have to have confidence that you will complete the project in a satisfactory manner. 2013-04-29T13:57:42 Hopefully, the code will be used afterwards. Otherwise it was just a summer of code, not a great summer of code ;) 2013-04-29T13:58:25 :) We really like to get the code to the point where it can be merged. Sometimes the project is too large and there is work left to be done. 2013-04-29T13:58:39 The posix key student from last year is really being tenacious. 2013-04-29T13:59:28 *** antgreen has joined #rtems 2013-04-29T14:05:49 About last year, you mentioned the ending of last years ARINC API project. Is there something I should know about, besides the source code? 2013-04-29T14:06:53 I did analyse it a week ago and wrote about it in the wiki. 2013-04-29T14:07:07 The mentor is just getting back to work from a break. He will get in touch with you and help answer questions but he can't mentor this year. 2013-04-29T14:07:49 From what I remember, the final tarball is available on https://github.com/jolkaczad/rtems_hyper/tarball/v0.0.5 2013-04-29T14:07:49 Also, the student has some code repositories on https://github.com/jolkaczad (and especially https://github.com/jolkaczad/rtems/tree/rtems_hyper) 2013-04-29T14:07:49 For mentoring, I do not have the time right now and it will be difficult for me to mentor on a regular basis. I can answer to some sporadic e-mail like this one but do not prefer to take any duty about that. Also, people that are active on POK might be interested (I was thinking about the guys from TELECOM ParisTech oe OpenWide). A good idea might be to contact the developer mailing-list. 2013-04-29T14:07:49 Also, if the student has any question, he can ask. The main issue is that as my availabilities are not regular, he might some delays some times. 2013-04-29T14:08:10 That's what I got via email from him. Does that fill in any gaps? 2013-04-29T14:09:14 ALso there is a BSP/port which runs on Linux using qemu- which runs in user space on Linux. It might be useful from an abstraction viewpoint. 2013-04-29T14:11:09 I did get the sourcecode, but not from the branch. I used the rtems_hyper repository. But it doesn't look much different. 2013-04-29T14:12:10 I think he had hello working but couldn't figure out how to get a clock tick running 2013-04-29T14:14:08 I did run the hello yes. Can you hint me in the direction of that BSP/port? 2013-04-29T14:15:12 https://github.com/jolkaczad/rtems/tree/rtems_hyper/rtems_wspace/rtems4.11/c/src/lib/libbsp/i386 2013-04-29T14:15:36 pok is his BSP.. getting it working and running all tests was 1st goal. Then generalizing to other ports, VMs 2013-04-29T14:16:04 really? All tests? 2013-04-29T14:16:52 it takes nothing but ram, a console and a clock tick to run everything 2013-04-29T14:17:18 the timing tests need a high granularity time source to be useful but you can use the timer driver stub which returns 0 2013-04-29T14:17:46 or map to some Pok sstem call to get start/stop time 2013-04-29T14:19:48 I was working on the nsecs sample in L4RTEMS, recently. I provide a software clock, which uses the defined tick rate, but I haven't looked into that on pok. 2013-04-29T14:20:55 That's what needs to happen next on this code. 2013-04-29T14:21:39 And eventually abstract things so there is a thin layer in RTEMS which lets a single paravirtualized BSP support multiple hypervisors on an architecture 2013-04-29T14:21:43 grand goals.. I know :) 2013-04-29T14:21:49 but we hate duplication 2013-04-29T14:21:49 I disagree with his compile approach. It looks quite hacky. 2013-04-29T14:22:03 Pok is a bit strange to compile 2013-04-29T14:22:31 Second looks are always better and you probably have a stronger background with the L4 experience than last year's student 2013-04-29T14:24:05 Maybe, but I have stumbled across the issue, how to start the guest. In L4 I just pass the binary as argument and read the SP/IP to pass it to the vCPU, but that comfortable way isn't going to work on POK. 2013-04-29T14:24:07 "high granularity" should be "high resolution" or "fine granularity" ;) 2013-04-29T14:27:18 phipse: since that code did not get merged, feel free to rewrite it any way that makes sense. 2013-04-29T14:27:56 can we run rtems on xen i wonder 2013-04-29T14:28:30 xen wants the vm to implement its own page tables, or not? 2013-04-29T14:28:33 hmm.. 2013-04-29T14:28:43 xen might assume a lot about the guest 2013-04-29T14:29:34 I haven't worked with xen, so ... didn't I delete that passage? 2013-04-29T14:30:41 my understanding of Pok is that you generate a Pok system using their tools. It is mostly generated code. The RTEMS code for a partition would be "inserted" into that image befoer the final link. 2013-04-29T14:30:46 oh it was just a thought that occurred to me. 2013-04-29T14:30:57 You would build an RTEMS Pok "app" for each partition. 2013-04-29T14:31:30 Build replace "befoer" with "during" 2013-04-29T14:31:44 I wasn't think on reusing much of the code, as I want to introduce a clear layer between host and guest, so the syscall approach of last year isn't going to work out. 2013-04-29T14:31:50 The RTEMS app is "foriegn" code or whatever they call it 2013-04-29T14:32:24 I think the syscall interface is all you have to talk to the Pok hypervisor. You can layer on top of it and abstract it but ultimately you need it 2013-04-29T14:33:26 i haven't studied the sw architecture of last year's project, but if the idea is to put the syscall interface further from RTEMS core then that would be good 2013-04-29T14:33:32 Yes, but I want to have it implemented on the host side and provided as a library to RTEMS. 2013-04-29T14:33:48 that's good... 2013-04-29T14:33:52 like um.. 2013-04-29T14:34:03 host == hypervisor? 2013-04-29T14:34:09 yes 2013-04-29T14:34:28 ok.. probably doesn't change the ultimate layering.. just where the code resides 2013-04-29T14:34:39 sounds cleaner to me! 2013-04-29T14:35:26 So rtems links until the undefined references due to the missing code. But the function headers are available and defined by RTEMS. The host/hypervisor has to provide the implementation as a library 2013-04-29T14:36:10 the only issue is making sure the hypervisor owners want to provide the library but that's a small assumption with L4 and Pok. :) 2013-04-29T14:37:28 The slighly other way to do it, would be to write all the defined function implementations on the host, use - as last year's student - the .ralf file, generated by RTEMS if linking fails, and link that together with the POK code 2013-04-29T14:38:07 so a POK binary including the RTEMS stuff falls out at the end 2013-04-29T14:39:15 I get fuzzy when it comes to how complete the BSP link step with Pok must be. But that's just a detail. If it is "ld -r", then it may still need a real linking with linkcmds to be at a known address range to insert into the Pok link. 2013-04-29T14:39:23 I just don't know. It is a detail to work through 2013-04-29T14:40:15 Unfortunatly, I can't copy from last years script, as I don't want to 'cat' binaries. ;) 2013-04-29T14:43:29 Look into how Pok does things. It just has to work with your desired procedure and be automatable. :) 2013-04-29T14:45:51 Good point with the linkcmds. I put that on my wiki-brain-extension, thanks. 2013-04-29T14:48:14 I used the opportunity to present POK to the OS group at my university at the end of May. So there will be another deadline for the look-into-pok part :) 2013-04-29T14:49:00 AFAIK Pok is the best ARINC 653 libre solution. 2013-04-29T14:49:34 oh DrJoel why did you have to succumb to the "libre" crowd 2013-04-29T14:49:53 :) 2013-04-29T14:50:00 FLOSS? 2013-04-29T14:50:08 that's how you clean teeth properly. 2013-04-29T14:50:37 Because the other solutions I know of claim to be GPL until you want the source. 2013-04-29T14:51:31 wasn't the difference between BSD and GPL the libre vs open source thing? 2013-04-29T14:51:53 depends on who you ask. 2013-04-29T14:52:09 best answer 2013-04-29T14:52:51 i've been ruminating on it lately, cuz of license questions we get sometimes in rtems. maybe i'll blog about it. except i'm not sure iwant to get pulled in that debate :) 2013-04-29T14:53:16 very cut and dry for me, 100% bsd, no more gpl due to the issues with gplv3 2013-04-29T14:53:32 I am using it in the other solutions are closed. And unforuntately at least one I am aware of claimed GPL and promised code but never delivered AFAIK 2013-04-29T14:53:39 Wasn't googlefight invented to decide those things? ;) 2013-04-29T14:54:01 So in this case, I am on the burned side of the free. 2013-04-29T14:54:32 As far as I am concerned GPL/BSD are OK for host side code. BSD is ok for RTEMS target side code. GPL+run-time exception is OK for RTEMS target side code. 2013-04-29T14:54:45 the idealogy is.. GPL/FSF claims that free applies to the public, and BSD claims that free applies to the individual.. that's how I see it. 2013-04-29T14:54:58 Oh I am sorry I opened this can 2013-04-29T14:55:01 :) 2013-04-29T14:55:12 gedare: i suppose that's a good way to put it 2013-04-29T14:55:38 for me it's simple: free means no restrictions, period 2013-04-29T14:55:52 so bsd-style/public domain are the only ones that qualify 2013-04-29T14:56:38 yup.. my freedom stops where yours starts. 2013-04-29T14:57:10 yeah, in the end it doesn't matter -- do whatever makes you happy as long as you're not hurting anyone else 2013-04-29T14:57:13 we're all in this for fun 2013-04-29T14:58:08 I'm actually in both camps. 2013-04-29T14:58:16 fun? where do I have to sign? 2013-04-29T14:58:30 phipse: wherever you want? :) 2013-04-29T14:58:35 Because.. I also want to have the freedom to control my content. 2013-04-29T14:59:06 rather, i find that both POVs are valid and acceptable, if in opposition 2013-04-29T14:59:07 sounds like I don't want to ever mess around with licensing 2013-04-29T14:59:34 gedare: what do you mean control your content?.. 2013-04-29T14:59:51 the only reason i don't like the gpl is i can't count how many times i've seen perfectly useful gpl code/software thrown out due to license 2013-04-29T15:00:11 instead of taking $useful_library_x a company will rewrite something inferior from scratch 2013-04-29T15:00:18 this numbers *well* into the hundreds now 2013-04-29T15:00:29 bsd? no problem just use it 2013-04-29T15:00:41 Rather to have some belief that what I release cannot be exploited. 2013-04-29T15:00:45 the end result is better technology for everyone, no 15 implementations of something 2013-04-29T15:01:13 gedare: that's a more personal belif on that side, i feel that if i spend my time on something i want maximum usage 2013-04-29T15:01:16 well, with gpl you can buy it (assuming the copyright is held by one entity who is wiling to sell it). 2013-04-29T15:01:24 yeah. that's my point though 2013-04-29T15:01:33 there's freedom of different shades 2013-04-29T15:01:36 but now i have to go :) 2013-04-29T15:01:36 my time is worth more than money since i can only spend it once.. the end i don't care if someone takes my work and makes a million dollars from it, good for them 2013-04-29T15:01:42 gedare: yeah exactly 2013-04-29T15:01:50 that's why i always tell people to do what makes them happy. :) 2013-04-29T15:02:00 as long as they're writing code who cares it's all valuable in the end 2013-04-29T15:03:24 *** kkthecoder has joined #rtems 2013-04-29T15:04:44 *** gedare has quit IRC (Read error: Operation timed out) 2013-04-29T15:06:57 I guess, there are as many opinions as people out there. 2013-04-29T15:07:22 Thanks for the input. It's getting late. Bye. 2013-04-29T15:07:24 pretty much, i've released software under the beerware license 2013-04-29T15:07:34 and i've gotten beer from it, too i consider it the most useful license of all of them 2013-04-29T15:07:45 phipse: goodnight 2013-04-29T15:07:50 great! :) 2013-04-29T15:07:59 *** phipse has left #rtems 2013-04-29T15:09:08 *** hiddenpearls1 has joined #rtems 2013-04-29T15:12:09 *** hiddenpearls has quit IRC (Ping timeout: 268 seconds) 2013-04-29T15:15:27 *** kkthecoder has quit IRC (Ping timeout: 245 seconds) 2013-04-29T15:33:19 *** kkthecoder has joined #rtems 2013-04-29T15:54:29 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-29T16:01:33 *** antgreen has quit IRC (Quit: Leaving) 2013-04-29T16:43:45 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2013-04-29T17:23:53 DrJoel, you need newlib cvs head, add '--macros=snapshots/newlib-head.mc` 2013-04-29T17:24:12 There is a patch to fix the alignment on SIS in newlib 2013-04-29T17:24:43 and it isn't fixable in 2.0? 2013-04-29T17:25:15 Did you see my size results for sis with function sections? 2013-04-29T17:27:40 It is fixed in cvs so Sebastian, Alan and I are using newlib head. 2013-04-29T17:28:13 I understand but wonder why we don't have the patch in the newlib 2.0 set. 2013-04-29T17:28:37 We need to decide on how to manage newlib. We can snapshot to a tarball on a regular basic, snapshot when are happy and get newlib's cvs tagged or record a "snapshot" date 2013-04-29T17:29:31 newlib is a special animal with annual releases and no bug fix releases. We probably are just as well off with a dated version. I don't want to tag in their repo 2013-04-29T17:31:18 I sent you a patch for the gdb sim bsps to add to the test bsps 2013-04-29T17:31:24 starting a build now 2013-04-29T17:34:45 Great. What date ? :) 2013-04-29T17:36:15 for "stable", we will periodically just have to pick one that passes our tests and has all our patches in it. Given the reent stuff going on, I would guess that when that settles, we have a base version for a while. 2013-04-29T17:39:19 Yeap I agree. Do we make a tarball snapshot or just work from CVS ? 2013-04-29T17:41:55 If we know the CVS date string, I say leave it in CVS. Just a matter of putting a -D on it. :) 2013-04-29T17:42:48 Sure. 2013-04-29T17:43:27 Can you read an excel spreadsheet? I have one which compares the size change in text, data, and bss for all tests on sparc-sis with function section enabled 2013-04-29T17:43:51 Can you export as a PDF ? 2013-04-29T17:44:40 how about a google spreadsheet. It has columns up to M and 400+ rows :) 2013-04-29T17:48:18 https://docs.google.com/spreadsheet/ccc?key=0AngKwYy0qUGZdHFGNHhuRnd5dDcxdmhiN2JkdFRoRHc#gid=0 2013-04-29T17:48:23 world readable 2013-04-29T17:50:07 That is a saving 2013-04-29T17:50:21 Have you looked into the reason ? 2013-04-29T17:51:10 They do run ? :) 2013-04-29T17:51:21 Yep! That's where I found the alignmetn issue 2013-04-29T17:55:21 I haven't gone looking for the issues but the savings are computer worked. So fixing linker scripts is easier than hunting all of the links 2013-04-29T18:01:12 Sure. 2013-04-29T18:22:08 DrJoel, is it ~/.mailrc ? 2013-04-29T18:26:06 DrJoel, patch merged. Thanks. 2013-04-29T18:32:12 *** sreekonduri has joined #rtems 2013-04-29T18:33:27 *** sreekonduri has joined #rtems 2013-04-29T18:54:38 *** hiddenpearls has joined #rtems 2013-04-29T18:55:59 *** the9a3eedi has joined #rtems 2013-04-29T18:57:04 *** hiddenpearls has quit IRC (Client Quit) 2013-04-29T19:16:20 *** sreekonduri has quit IRC (Quit: Page closed) 2013-04-29T19:30:10 verm__, if I use your beer license do I get the beer or do you ? 2013-04-29T19:35:35 *** antgreen has joined #rtems 2013-04-29T20:02:36 me of course, PHK created it 2013-04-29T20:02:43 * "THE BEER-WARE LICENSE" (Revision 42): 2013-04-29T20:02:43 * wrote this file. As long as you retain this notice you 2013-04-29T20:02:43 * can do whatever you want with this stuff. If we meet some day, and you think 2013-04-29T20:02:44 * this stuff is worth it, you can buy me a beer in return Poul-Henning Kamp 2013-04-29T20:02:53 2013-04-29T20:02:55 best license ever 2013-04-29T20:03:05 i tried to buy him a beer once but it was the day before a talk so he declined, doh. 2013-04-29T20:09:48 *** kkthecoder has quit IRC (Ping timeout: 264 seconds) 2013-04-29T20:14:47 *** freenix has joined #rtems 2013-04-29T20:23:41 *** weiY has joined #rtems 2013-04-29T22:28:36 *** freenix has quit IRC (Remote host closed the connection) 2013-04-29T22:46:38 *** kkthecoder has joined #rtems 2013-04-29T23:52:17 *** hiddenpearls has joined #rtems 2013-04-30T00:03:29 *** monstr has joined #rtems 2013-04-30T01:09:31 *** arvind_khadri has joined #rtems 2013-04-30T01:11:09 *** freenix has joined #rtems 2013-04-30T01:15:13 *** antgreen_ has joined #rtems 2013-04-30T01:16:23 *** antgreen has quit IRC (Ping timeout: 245 seconds) 2013-04-30T02:02:08 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-30T02:04:49 *** sebhub has joined #rtems 2013-04-30T02:05:40 good morning 2013-04-30T02:09:11 sebhub, hi 2013-04-30T02:21:02 *** hiddenpearls has joined #rtems 2013-04-30T02:48:50 *** hiddenpearls1 has joined #rtems 2013-04-30T02:51:12 *** hiddenpearls has quit IRC (Ping timeout: 276 seconds) 2013-04-30T02:56:44 *** hiddenpearls1 has quit IRC (Ping timeout: 252 seconds) 2013-04-30T03:19:45 *** hiddenpearls has joined #rtems 2013-04-30T04:08:30 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-04-30T04:10:50 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-30T04:17:03 *** hiddenpearls has joined #rtems 2013-04-30T04:42:13 *** freenix has quit IRC (Remote host closed the connection) 2013-04-30T05:16:14 *** mgahlot|away has joined #rtems 2013-04-30T06:04:37 *** S_Somani has joined #rtems 2013-04-30T06:15:33 *** S_Somani has quit IRC (Ping timeout: 245 seconds) 2013-04-30T06:26:18 *** S_Somani1 has joined #rtems 2013-04-30T06:35:57 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-30T06:39:12 *** S_Somani1 has quit IRC (Ping timeout: 245 seconds) 2013-04-30T06:45:26 *** hiddenpearls has joined #rtems 2013-04-30T07:06:37 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-30T07:07:39 *** kiwichris has joined #rtems 2013-04-30T07:11:32 *** mgahlot|away has quit IRC (Remote host closed the connection) 2013-04-30T07:16:32 *** Hesham has joined #rtems 2013-04-30T07:23:46 *** gedare has joined #rtems 2013-04-30T07:26:46 morning 2013-04-30T07:28:49 morning 2013-04-30T07:44:53 *** Hesham has quit IRC (Ping timeout: 256 seconds) 2013-04-30T07:46:03 *** hiddenpearls1 has joined #rtems 2013-04-30T07:46:46 * DrJoel is in earlier than normal :( 2013-04-30T07:47:57 *** hiddenpearls has quit IRC (Ping timeout: 245 seconds) 2013-04-30T07:49:23 *** gedare has quit IRC (Ping timeout: 255 seconds) 2013-04-30T07:51:17 *** weiY has quit IRC (Ping timeout: 245 seconds) 2013-04-30T07:57:07 *** Hesham has joined #rtems 2013-04-30T08:27:03 *** Hesham has quit IRC (Quit: Leaving.) 2013-04-30T08:27:09 *** Hesham1 has joined #rtems 2013-04-30T08:38:50 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2013-04-30T08:52:06 *** hiddenpearls has joined #rtems 2013-04-30T08:54:14 *** antgreen_ has quit IRC (Ping timeout: 252 seconds) 2013-04-30T08:59:29 *** gedare has joined #rtems 2013-04-30T09:02:30 *** weiY has joined #rtems 2013-04-30T09:03:20 cramming in the hours? 2013-04-30T09:03:27 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-30T09:07:04 just woke up early for some odd reason 2013-04-30T09:07:21 ahh 2013-04-30T09:10:17 *** hiddenpearls has joined #rtems 2013-04-30T09:12:02 Jordan had an 8:00am exam which probably didn't help me sleep .. lol 2013-04-30T09:22:34 Hi, good moring 2013-04-30T09:23:23 hello weiY 2013-04-30T09:24:06 Hi drjoel. I am working on breaking my project plan more little 2013-04-30T09:24:20 good morning. 2013-04-30T09:24:49 The projectlibre is useful. 2013-04-30T09:24:54 good morning gedare 2013-04-30T09:25:38 i think i should commit your patch soon.. i forgot about it but no one complained 2013-04-30T09:25:54 now i have check the arm atomic implement in the freebsd , that i does not support smp mode 2013-04-30T09:26:27 does test the patch for all the architectures? 2013-04-30T09:26:42 i have not finished yet. maybe i'll try a couple more yet 2013-04-30T09:27:02 I did run powerpc and mips 2013-04-30T09:27:05 maybe someone did not pay attention on the patch, hehe 2013-04-30T09:27:15 probably. 2013-04-30T09:28:06 i'm at work and don't have a good setup for rtems here. 2013-04-30T09:28:09 so i can't test. 2013-04-30T09:28:15 if i have time i try to test it on other architectures, but now i am busy with to modify the proposal according to joel's adivce 2013-04-30T09:28:22 sure 2013-04-30T09:28:45 DrJoel, if you want you should prod Prateek and Rinesh 2013-04-30T09:29:00 first i will setup all other architectures toolchain, this is first thing need to do 2013-04-30T09:29:17 although i don't really care that much we have a good set of proposals to review already 2013-04-30T09:29:41 well, RPMs can help with that. or you can try the RTEMS source builder 2013-04-30T09:30:44 yes, i can try RPMS, i am familiar with this. later i can try the source builder 2013-04-30T09:31:01 all architectures? 2013-04-30T09:31:13 we should verify they build 2013-04-30T09:31:20 and run as many as possible on simulator 2013-04-30T09:31:33 i already did mips and powerpc.. i will do sparc64 also 2013-04-30T09:31:36 you did sparc already? 2013-04-30T09:31:50 yes, i use psim 2013-04-30T09:31:59 psim is powerpc 2013-04-30T09:32:03 and all testcases are passed 2013-04-30T09:32:30 i386 shouldn't change 2013-04-30T09:32:45 hmm, and powerpc either so i guess that is not as important 2013-04-30T09:32:51 because they have the original atomic implemented 2013-04-30T09:33:13 arm and sh should be testable with gdb 2013-04-30T09:33:18 i don't know about the rest 2013-04-30T09:33:21 no i forget i use sparc-debug 2013-04-30T09:33:31 oh ok that is good then 2013-04-30T09:34:14 ok, i try to test them build right for all architecture 2013-04-30T09:34:37 and if i have problem to simulator them i can call you 2013-04-30T09:34:54 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-30T09:35:02 ok. there is no guarantee they all build anyway. i don't know the last time all architectures were built correctly! 2013-04-30T09:35:07 soon we can add buildbot... 2013-04-30T09:35:21 have to release rtems-4.11 2013-04-30T09:35:26 * gedare pokes DrJoel 2013-04-30T09:35:35 hehe 2013-04-30T09:36:04 recently i have check the arm atomic implement in the freebsd , that i does not support smp mode 2013-04-30T09:36:10 DrJoel, do you think alphabetic characters would be acceptable for the "release candidate" versioning? 2013-04-30T09:36:21 it just use disable/enable to simulator 2013-04-30T09:36:24 like right now we could be on 4.11.0.a 2013-04-30T09:36:35 (or 4.11.a) 2013-04-30T09:36:52 and the 4.10 release could be on 4.10.3.a 2013-04-30T09:37:18 weiY, really? can you link me to source on web? 2013-04-30T09:37:33 *** hiddenpearls has joined #rtems 2013-04-30T09:37:38 ok wait 2013-04-30T09:38:10 do you mean freebsd does not support smp atomic, or that RTEMS does not? 2013-04-30T09:40:14 http://svnweb.freebsd.org/base/stable/9/sys/arm/include/atomic.h?revision=225736&view=markup 2013-04-30T09:40:34 "19 * 3. All advertising materials mentioning features or use of this software 2013-04-30T09:40:34 20 * must display the following acknowledgement: 2013-04-30T09:40:34 21 * This product includes software developed by Brini. 2013-04-30T09:40:34 22 * 4. The name of Brini may not be used to endorse or promote products 2013-04-30T09:40:34 23 * derived from this software without specific prior written permission." 2013-04-30T09:40:39 I don't know if we can use this code anyway. 2013-04-30T09:40:48 freebsd does not support smp atomic for all architectures 2013-04-30T09:41:08 for powerpc, i386, mips, sparc64. it does support 2013-04-30T09:41:14 ok 2013-04-30T09:41:18 but for arm it do not 2013-04-30T09:41:20 then you have some work to do! 2013-04-30T09:41:26 anyway... ARM SMP is not merged yet either 2013-04-30T09:42:11 * DrJoel is here and overloaded.. chat.. phone..email.. head exploding 2013-04-30T09:42:38 yeah, and i see netbsd support arm-V6 atomic ops 2013-04-30T09:42:46 link? 2013-04-30T09:43:02 and it seems that it supports SMP , but i am not sure. wait 2013-04-30T09:43:46 I don't know on version names gedare.. I used to put preN on them when I cut snapshots priority to releases 2013-04-30T09:43:57 Will it confuse any version parsing? 2013-04-30T09:47:37 gedare, http://cvsweb.netbsd.org/bsdweb.cgi/src/common/lib/libc/arch/arm/atomic/?only_with_tag=MAIN 2013-04-30T09:47:58 but the atomic API is not the same as freebsd 2013-04-30T09:48:31 the underlying instructions should be similar and it is a reference. 2013-04-30T09:49:07 FWIW rtems-testing/sim-scripts/Makefile shows which architectures have simulators in the gdb as built for RPMs or with RTEMS source builder 2013-04-30T09:49:15 And there are scripts there to help you 2013-04-30T09:50:08 ok, i will look into this Makefile 2013-04-30T09:50:40 *** mgahlot|away has joined #rtems 2013-04-30T09:51:37 It is just a list. :) The scripts are there. Most architectures have a simulator 2013-04-30T09:53:47 build or use the sim-scripts can i run all the architectures? 2013-04-30T09:54:29 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2013-04-30T09:54:44 most.. if you have tools w/gdb installed. 2013-04-30T09:55:44 Look for the comment ### GDB Simulator BSPs 2013-04-30T09:55:52 yeah i can try to install the toolchain for all the architectures 2013-04-30T09:55:52 and from there down you can see the architectures/bsps 2013-04-30T09:57:10 yeah i have see. most of architectures are supported 2013-04-30T09:58:48 *** mgahlot|away has quit IRC (Ping timeout: 268 seconds) 2013-04-30T09:59:08 They aren't all fabulously capable simulators from a device simulator viewpoint but they work for basic testing. 2013-04-30T09:59:18 sis jmr3904 and psim and the best of the lot. 2013-04-30T10:00:14 yeah, i kown your mean. i just test atomic UP mode ops. it should be ok 2013-04-30T10:00:52 *** mgahlot|away has joined #rtems 2013-04-30T10:02:49 joel, from the source code i see there is a member "extensions" in the TCB struct. whether i can use this member to store the lock list held by a thread? 2013-04-30T10:03:12 if ok i donot embedded extra member into the TCB 2013-04-30T10:05:56 If you write the code as an extension set. You would then have hooks for create, delete, start, restart, switch and exit 2013-04-30T10:06:40 Other code with this could be outside the extension. But some could/would be in the extension points. 2013-04-30T10:07:30 can those code be placed into score code? 2013-04-30T10:08:31 Good question. I suppose so. 2013-04-30T10:08:47 Is this something you enable at RTEMS compile time? 2013-04-30T10:09:16 yeah, i enable them through conditinal compile 2013-04-30T10:09:40 if disable it will not be compiled into binary 2013-04-30T10:10:13 and those code will insert into mutex RTEMS classic API and posix API 2013-04-30T10:10:18 There is no reason to make them extensions IMO. They will not be dynamically plugged in 2013-04-30T10:11:07 gedare's opinion matters here though.. so get 2. :) 2013-04-30T10:11:14 ok, but then it must allocte a new member in the TCB 2013-04-30T10:11:43 which gedare's opinion? 2013-04-30T10:11:47 and it is conditionally compiled in the structure. Not a big deal from my perspective. 2013-04-30T10:12:00 ok, i see 2013-04-30T10:12:02 on how to implement 2013-04-30T10:13:58 and how about the priority inheritance algorithm? i mean the thread set implementation 2013-04-30T10:14:27 *** S_Somani has joined #rtems 2013-04-30T10:15:32 what specifically about it? 2013-04-30T10:15:48 i mean if there are no overlapping work with thread set implementation i want to add the priority inheritance algorithm into proposal. even i am not sure i can completed it in this summer 2013-04-30T10:17:01 I don't think it will overlap enough to cause problems. This will be at the end of your effort and we can focus the other project on the area that impacts thread queue early enought o avoid conflict with you 2013-04-30T10:19:44 ok, i see. i can just focus on the algorithm itself. and if there are some standard API defined i can follow the API to manager the data struct 2013-04-30T10:20:41 yep. Worst case, we implement a standard api for what you need. 2013-04-30T10:22:05 hehe, ok. we can exchange idea as much as possible 2013-04-30T10:30:51 *** sebhub has quit IRC (Remote host closed the connection) 2013-04-30T10:44:01 avoid task extensions. 2013-04-30T10:44:20 well, for an optional interface it might not be so bad 2013-04-30T10:44:40 It is conditionally compiled in at RTEMS build time. So not a "confdefs.h" type extension 2013-04-30T10:44:51 yeah, i have talk with joel. i can embedded a new member into TCB 2013-04-30T10:45:03 *** mgahlot|away has quit IRC (Ping timeout: 268 seconds) 2013-04-30T10:45:42 gedare, have you see the netbsd atomic source code? 2013-04-30T10:45:48 no did you post link? 2013-04-30T10:45:55 yeah 2013-04-30T10:46:01 ok found it 2013-04-30T10:46:11 http://cvsweb.netbsd.org/bsdweb.cgi/src/common/lib/libc/arch/arm/atomic/?only_with_tag=MAIN 2013-04-30T10:47:21 looks good to start. especially if all the code is 2-clause bsd 2013-04-30T10:48:15 i want to encourage new code contributions to be the 2-clause BSD when possible 2013-04-30T10:48:51 where can i find the information about 2-clause BSD 2013-04-30T10:49:47 and freebsd and netbsd are all 2-clause BSD license? 2013-04-30T10:50:02 no they vary each file carries its own license information 2013-04-30T10:50:04 so you have to look 2013-04-30T10:50:14 http://en.wikipedia.org/wiki/BSD_licenses#2-clause_license_.28.22Simplified_BSD_License.22_or_.22FreeBSD_License.22.29 2013-04-30T10:50:21 3-clause is OK too 2013-04-30T10:50:23 4-clause is no good 2013-04-30T10:50:45 ok, i should learn more about the license. hehe 2013-04-30T10:51:22 maybe i'll write about them 2013-04-30T10:51:28 been on my mind lately 2013-04-30T10:51:44 really ever since we had to put in the latest application for gsoc. 2013-04-30T10:52:43 gedare: If you do this in the vein of "ok for RTEMS licenses and rules", there are some vendor licenses which are OK. Usually these are "as is, etc" 2013-04-30T10:52:44 Interesting... I typed California with my right hand shifted left one key.. Cakufirbua. sounds like a real word! 2013-04-30T10:52:57 yeah 2013-04-30T10:53:00 *** mgahlot|away has joined #rtems 2013-04-30T10:53:04 i know 2013-04-30T10:53:12 But there are some which are not. :( I am looking for an example now 2013-04-30T10:53:17 good 2013-04-30T10:55:46 arm toolchain has been installed. now try to build for ARM. 2013-04-30T10:58:18 gedare just forwarded you an example of a vendor license which is close to ok but manages to be unclear enough where I wouldn't merge it 2013-04-30T10:58:27 ok thanks 2013-04-30T10:58:42 i thought somewhere i read that UC Berkeley retroactively released the advertising clause? 2013-04-30T10:59:52 *** mgahlot|away has quit IRC (Read error: Connection reset by peer) 2013-04-30T11:00:35 They did. http://en.wikipedia.org/wiki/BSD_licenses . the original announcement appears to be down/gone 2013-04-30T11:00:57 i guess it is common knowledge though 2013-04-30T11:01:21 But code from other contributors sometimes continued to use the advertising clause. NetBSD has fixed most but can't fix those it can't find the owners or they won't give permission to drop it. So the kernel has a mix 2013-04-30T11:01:32 yes.. so do we 2013-04-30T11:01:35 and we should purge them 2013-04-30T11:02:00 We can't change it unless the upstream did or we get permission to drop it. We are in the same boat they are in. 2013-04-30T11:02:20 We will not get an IPV6 or USB stack w/o advertising clause. :( 2013-04-30T11:02:31 *** mgahlot|away has joined #rtems 2013-04-30T11:02:39 we should maintain those in a separate repo 2013-04-30T11:03:14 No TCP/IP stack or USB stack in rtems main repo? 2013-04-30T11:04:10 yup. separate checkout if you want to build 2013-04-30T11:05:19 we should strive to clean up licensing issues 2013-04-30T11:05:27 You won't be able to build many secondary languages without it. 2013-04-30T11:05:31 hmmm. 2013-04-30T11:05:36 those take two passes already? 2013-04-30T11:05:39 I agree. Not arguing. Just pointing out the complexities. 2013-04-30T11:05:50 Yep.. you are adding another step 2013-04-30T11:06:03 C/C++, then RTEMS, then TCP/IP, then secondary language 2013-04-30T11:06:37 I think that is ok. 2013-04-30T11:06:56 Also would have to figure out how that impacts integration of file descriptors 2013-04-30T11:07:03 but we aren't ready to merge that yet. 2013-04-30T11:07:38 If you want a list to make a list of candidates to move out, I am ok with that. Personally, I think once we get the RSB building RTEMS Add-ons, Mongoose could easily slip outside the tree. 2013-04-30T11:07:57 ok 2013-04-30T11:08:23 i'm thinking if i get an afternoon i will spend it looking at this license stuff, write a blog post and analyze rtems 2013-04-30T11:09:15 Looking forward, it makes sense if you want lwip as an alternative add-on 2013-04-30T11:09:50 Plus could let us keep old stack around in "cyogenic suspension" for a while :) 2013-04-30T11:10:02 yeah. 2013-04-30T11:20:22 i have to go to bed. see you tommorom 2013-04-30T11:23:20 okay 2013-04-30T11:32:30 *** monstr has quit IRC (Ping timeout: 276 seconds) 2013-04-30T11:38:41 *** mgahlot|off has joined #rtems 2013-04-30T11:46:11 *** sevikkk has quit IRC (Read error: No route to host) 2013-04-30T12:04:59 *** gedare has quit IRC (Ping timeout: 272 seconds) 2013-04-30T12:12:12 *** hiddenpearls has quit IRC (Ping timeout: 264 seconds) 2013-04-30T12:12:25 *** hiddenpearls has joined #rtems 2013-04-30T12:19:57 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-30T12:23:06 *** hiddenpearls has joined #rtems 2013-04-30T12:32:05 *** mgahlot|off has quit IRC (Read error: Connection reset by peer) 2013-04-30T12:33:07 *** mgahlot|away_ has joined #rtems 2013-04-30T12:36:40 *** mgahlot|away has quit IRC (Ping timeout: 268 seconds) 2013-04-30T12:37:36 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-30T12:41:36 *** mgahlot|away_ has quit IRC (Ping timeout: 268 seconds) 2013-04-30T13:18:45 *** dhananjay has joined #rtems 2013-04-30T13:20:59 *** dhananja1 has quit IRC (Ping timeout: 268 seconds) 2013-04-30T13:29:56 *** gedare has joined #rtems 2013-04-30T13:37:20 Hesham1: no the github is not up to date :( 2013-04-30T13:38:00 how far back was the question? 2013-04-30T13:38:25 That's why I cloned a vanilla repo from git.rtems.org/rtems and create patches for it 2013-04-30T13:38:32 from an email. i just thought i'd answer here :) 2013-04-30T13:38:40 ok thats fine 2013-04-30T13:39:01 you can also track git.rtems.org yourself on github, it is just a matter of setting up your remote properly 2013-04-30T13:40:36 I did not do it before, I will give it a try 2013-04-30T13:41:12 Will RTEMS keep track of GSoC projects on github this year ? 2013-04-30T13:41:12 yeah it is all right if not as long as i can find the patches in my mail box when it comes time to apply something :) 2013-04-30T13:41:24 i'd like to yes. 2013-04-30T13:41:44 if we can get the auto-updating fixed that would be ideal 2013-04-30T13:42:03 another option would be to have students point to git.rtems.org as their upstream... but then we can't do pull requests 2013-04-30T13:42:28 gedare do you have a github account? 2013-04-30T13:42:33 yup 2013-04-30T13:42:39 https://github.com/gedare 2013-04-30T13:43:13 i don't use it much although i might start 2013-04-30T13:43:27 I was going to add you to RTEMS on github if I can figure it out. 2013-04-30T13:43:32 Then maybe we can see what broke. :( 2013-04-30T13:43:40 i think i am in the group 2013-04-30T13:43:52 i think the problem is on the git.rtems.org server 2013-04-30T13:44:06 verm__'s script is broken or something 2013-04-30T13:44:20 The auto-updating would be good. I think when it reach to pull requests it will be easy to just create patches as usual 2013-04-30T13:44:38 we could just create an hourly cron job on git.rtems.org to push to github 2013-04-30T13:45:25 verm__ set up a script to push to github for every commit, but the script stops working for some unknown reasons 2013-04-30T13:45:49 given the rate of commits to rtems, an hourly cron job shouldn't be too bad 2013-04-30T13:46:13 I see no reference to rtems in the git hooks 2013-04-30T13:47:01 [remote "github"] 2013-04-30T13:47:02 url = git@github-update:RTEMS/rtems.git 2013-04-30T13:48:20 see /data/support/git-support/hooks 2013-04-30T13:48:45 specifically, post-receive-2 2013-04-30T13:48:53 it gets called from rtems.git/hooks/post-receive 2013-04-30T13:50:00 and i guess it is supposed to send a message to a daemon that will push to github 2013-04-30T13:50:36 i don't know why it doesn't just push to github directly, maybe something to do with blocking/synchrony 2013-04-30T13:51:05 Or the service isn't on after an update? \ 2013-04-30T13:51:12 that could be the problem 2013-04-30T13:51:30 verm__ said it worked on other tested distros, but redhat was causing some issues for him 2013-04-30T13:52:12 I have a doctor's appt and need to leave.. can you ping him to give us some hints on what to debug. pointers to documentation. We need to get this running again and no what we are looking for. 2013-04-30T13:52:23 he's working til midnight i think 2013-04-30T13:52:43 might not be yet though 2013-04-30T13:52:59 no prob.. this is just something that has gone on and we need to know how to fix it. Not asking him to fix it. He is busy. :) 2013-04-30T13:53:38 * DrJoel is out of here 2013-04-30T13:53:47 *** DrJoel has quit IRC (Quit: ASCII a stupid question, get a stupid ANSI!) 2013-04-30T14:03:32 *** sevikkk has joined #rtems 2013-04-30T14:03:59 gedare: I think that might help, https://github.com/andygrunwald/github-sync 2013-04-30T14:04:06 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-04-30T14:04:06 *** sevikkk has quit IRC (Read error: Connection reset by peer) 2013-04-30T14:04:32 *** sevikkk has joined #rtems 2013-04-30T14:05:26 thanks..i think we have the scripts they're just not configured properly 2013-04-30T14:05:52 *** kiwichris has joined #rtems 2013-04-30T14:30:06 *** arvind_khadri has joined #rtems 2013-04-30T14:52:18 *** hiddenpearls has quit IRC (Ping timeout: 264 seconds) 2013-04-30T14:55:17 *** weiY has quit IRC (Ping timeout: 248 seconds) 2013-04-30T14:55:39 *** weiY has joined #rtems 2013-04-30T14:55:44 *** Hesham1 has quit IRC (Ping timeout: 260 seconds) 2013-04-30T14:55:50 *** arvind_khadri has quit IRC (Ping timeout: 256 seconds) 2013-04-30T14:59:12 *** gedare has quit IRC (Ping timeout: 276 seconds) 2013-04-30T16:23:19 *** sevikkk has quit IRC (Quit: Leaving.) 2013-04-30T16:23:34 *** sevikkk has joined #rtems 2013-04-30T16:24:30 *** sevikkk has joined #rtems 2013-04-30T16:25:14 *** sevikkk has quit IRC (Client Quit) 2013-04-30T16:47:26 *** DrJoel has joined #rtems 2013-04-30T16:47:26 *** DrJoel has joined #rtems 2013-04-30T16:47:26 *** ChanServ sets mode: +o DrJoel 2013-04-30T16:47:31 kiwichris there? 2013-04-30T16:54:57 Yeah .. just have a battle with the Xilinx web site .. ahhhh 2013-04-30T16:55:05 lol 2013-04-30T16:55:21 wish I could laugh ? what a mess 2013-04-30T16:55:46 links that need java plugins and then are broken 2013-04-30T16:55:55 I was building i386 tools and the build ended when the rtems build failed and I ended up with no tools installed. What do I do exclude rtems when I just want tools to work on rtems? 2013-04-30T16:56:30 I have not added support to turn on and off rtems 2013-04-30T16:56:41 sorry 2013-04-30T16:57:13 The install happens at the end so broken bits and pieces are not installed 2013-04-30T16:57:24 A comment in the i386 bset file is best 2013-04-30T16:59:17 that is what i did. :) 2013-04-30T17:00:07 now you see how my gcc testing script got so many arguments .. lol 2013-04-30T17:01:23 I think the idea of a build set is valid, I am just not sure how I will manage this enable/disable part with endless conditional logic in the configurations 2013-04-30T17:01:36 with -> without 2013-04-30T17:02:17 Just fixing the email reports to include the build status at the top with the tail of the log 2013-04-30T17:02:25 when an error happens 2013-04-30T17:10:23 Should the email report's subject line indicate pass/fail ? 2013-04-30T17:11:58 yes.. but it gets grey when running make check 2013-04-30T17:13:34 Why ? 2013-04-30T17:14:24 *** gedare has joined #rtems 2013-04-30T17:15:34 how many test cases do you fail before it is a fail? 2013-04-30T17:15:55 The GCC reports for RTEMS would be much more useful with a BSP included in the description FWIW 2013-04-30T17:16:15 DrJoel: long day! 2013-04-30T17:16:20 go home! 2013-04-30T17:16:37 I am home! 2013-04-30T17:16:37 :) 2013-04-30T17:16:41 oh 2013-04-30T17:16:49 ok 2013-04-30T17:17:06 no bad news at doctor's visit. :) 2013-04-30T17:17:39 :) 2013-04-30T17:17:53 good 2013-04-30T17:18:08 heya kiwichris been awhile since we crossed paths. 2013-04-30T17:18:18 yeah, are are things with you ? 2013-04-30T17:18:37 pretty good. busy.. 2013-04-30T17:18:53 DrJoel, too hard to list the bsps, what if you need to run 4 to cover the arch ? 2013-04-30T17:18:56 which is good i suppose. 2013-04-30T17:19:03 The build set defines what happens 2013-04-30T17:19:10 Nice 2013-04-30T17:19:55 4 bsps to run tests on? 2013-04-30T17:20:00 DrJoel, how many lines of the log tail should I include ? 100 ? 2013-04-30T17:20:15 If they trigger different gcc flags. 2013-04-30T17:20:19 yes, eg arm 2013-04-30T17:21:07 On arm, I am guessing at least 4. That's one of my outstanding questions on gdbarmsim. How many variants does it need? Probably one for each multilib. So that's probably the answer. One for each multilib variant is a goal. 2013-04-30T17:21:47 Yeah and you may have VFP, VFP + Neon or no co-pro type support 2013-04-30T17:21:57 then A9 etc etc 2013-04-30T17:22:17 thumb thumb2 etc 2013-04-30T17:23:34 DrJoel: if you have time you should sketch some of the schedsim features and wishlist for that fault-tolerant scheduling student. 2013-04-30T17:24:30 moxie is failing on 4.11 with gcc 4.8. I am not sure why 2013-04-30T17:25:07 gedare on his proposal? or a wiki page? 2013-04-30T17:25:16 wherever you feel like 2013-04-30T17:25:25 wiki page is better for the future 2013-04-30T17:25:55 i'm trying to figure out the algorithm he proposes to implement 2013-04-30T17:25:59 make sure it is feasible 2013-04-30T17:26:40 Any wiki page in particular? 2013-04-30T17:26:47 sec.. 2013-04-30T17:27:09 http://wiki.rtems.org/wiki/index.php/SuperCore_Scheduler 2013-04-30T17:27:15 make a new subsection for Scheduler Simulator 2013-04-30T17:27:22 will do. 2013-04-30T17:27:27 or add to "Large related tasks" 2013-04-30T17:27:41 thanks. then just drop a note in the proposal 2013-04-30T17:27:55 i want to try to give all the students at least 2 days to "finalize" 2013-04-30T17:28:17 prateek answered back if you were cc'ed. The other student hasn't yet 2013-04-30T17:28:53 We apparently can let them edit after Friday is we want. But I think a big factor is proper mentoring. That may be the issue this year. 2013-04-30T17:29:14 agreed 2013-04-30T17:29:51 i was not cc'd. 2013-04-30T17:29:59 closed when ? 2013-04-30T17:30:02 but that's ok.. will see what he posts. 2013-04-30T17:35:28 DrJoel, sent you an example report with a fail. Ok ? 2013-04-30T17:38:34 I think that looks ok. I would take the date stamp off the subject line since it should be in the message body and likely matches a timestamp on the email (if we aremailing these). More useful would be the component that failed to build on the subject line. 2013-04-30T17:45:52 Build: 4.11/rtems-moxie x86_64-freebsd9.1 FAILED:moxie-rtems4.11-gcc-4.8.0-newlib-30-April-2013-1 2013-04-30T17:45:54 ? 2013-04-30T17:46:11 or ... 2013-04-30T17:46:28 Build: FAILED: 4.11/rtems-moxie x86_64-freebsd9.1 (moxie-rtems4.11-gcc-4.8.0-newlib-30-April-2013-1) 2013-04-30T17:46:42 and when ok ... 2013-04-30T17:46:54 Build: 4.11/rtems-moxie x86_64-freebsd9.1 2013-04-30T17:46:56 or ... 2013-04-30T17:47:01 Build: PASSED: 4.11/rtems-moxie x86_64-freebsd9.1 2013-04-30T17:49:35 I like the FAILED and PASSED verisons. My only question is on do you pass when we run make check and don't pass 100% of the tests. :) 2013-04-30T17:50:08 This is just a build report pass/fail right? 2013-04-30T17:52:10 *** Hesham has joined #rtems 2013-04-30T17:52:38 *** Hesham has left #rtems 2013-04-30T17:54:05 +1 FAILED PASSED 2013-04-30T18:04:56 http://wiki.rtems.org/wiki/index.php/SuperCore_Scheduler#Scheduler_Simulator 2013-04-30T18:05:28 I forgot the command set: 2013-04-30T18:05:31 $ ls 2013-04-30T18:05:31 commands.c main_executing.c main_semobtain.c main_tasksuspend.c 2013-04-30T18:05:31 getopt.c main_heir.c main_semrelease.c main_taskwakeafter.c 2013-04-30T18:05:31 include main_help.c main_taskcreate.c Makefile.am 2013-04-30T18:05:31 lookup_semaphore.c main_rtemsinit.c main_taskdelete.c schedsim_shell.h 2013-04-30T18:05:32 lookup_task.c main_semcreate.c main_taskmode.c shell_cmdset.c 2013-04-30T18:05:32 main_clocktick.c main_semdelete.c main_taskpriority.c shell_makeargs.c 2013-04-30T18:05:33 main_echo.c main_semflush.c main_taskresume.c 2013-04-30T18:06:04 Each task or sem command actually ends up doing a Classic API task or semaphore directive. This is setup so you could even play with priority inheritance or ceiling 2013-04-30T18:06:18 This really should be in waf 2013-04-30T18:07:56 *** gedare has quit IRC (Ping timeout: 245 seconds) 2013-04-30T18:16:01 yes 2013-04-30T18:16:43 it is only a top level configure and a handful of Makefile.am's ... 2013-04-30T18:19:59 hint.. hint 2013-04-30T18:20:43 DrJoel, we need to define what we say is a pass/fail for tests. This is needed for each BSP. 2013-04-30T18:21:05 It is not about 100% passing of the tests, it is about not regressing against a base line 2013-04-30T18:21:41 Re-the waf change over, is that the list or files to build etc ? 2013-04-30T18:24:11 Yep. This is the largest one. http://git.rtems.org/rtems-schedsim/tree/schedsim/rtems/Makefile.am 2013-04-30T18:25:47 There are only 7 Makefile.am's in the entire tree. The one I linked to there primarily lists the files in the RTEMS source tree needed plus a "sched_cpu" port 2013-04-30T18:26:45 And it needs RTEMS source ? 2013-04-30T18:28:17 Yes.. it uses the real source and wraps a scripting language around it. So you can test the scheduler's decisions natively. 2013-04-30T18:28:52 writing a scheduler is hard enough because you have to get the decision logic right. This lets you whip out scenarios in a few lines 2013-04-30T18:29:54 For SMP, you can do any number of cores in the script initialization. we tested the algorithm on 1, 2, and 4 core systems on the simulator with no HW. 2013-04-30T18:30:32 Hmm, waf likes all the source in the one tree, it is a design constrain. 2013-04-30T18:30:34 Hmm, waf likes all the source in the one tree, it is a design constraint. 2013-04-30T18:30:47 What if I copy into a local tree the needed source 2013-04-30T18:30:50 ? 2013-04-30T18:33:00 For building purposes, that would be OK. But they wouldhave to be in .gitignore and removed as part of clean up. 2013-04-30T18:33:21 Would a first step in waf do that? 2013-04-30T18:33:31 It can't be a permanent copy in the repo. 2013-04-30T18:51:09 It could be in the build dir and cleaned on clean 2013-04-30T18:51:53 That's fine. I am in the process of getting it building again. I think I am pretty close. 2013-04-30T18:51:55 I would have to take a look and figure it out 2013-04-30T18:52:17 I think it would help for me to get it building again... needed for gsoc, processor affinity work, etc 2013-04-30T19:01:27 scheduler simulator now runs again. :) 2013-04-30T19:10:03 Running the 4.11 build again on rtbf64a with the released files and I hope the pass/fail messages 2013-04-30T19:11:09 very cool! I will clean up the scheduler simulator some more and commit it tomorrow. I have two user documents .. one on writing a scheduler and one on using the scheduler simulator. I suppose the first could go in the regular docs but the other shouldn't and I hate to dupliate even more to build it. :( 2013-04-30T19:11:24 think on it. I need to quit for the night 2013-04-30T19:11:27 *** DrJoel has quit IRC (Quit: When the chips are down, well, the buffalo is empty) 2013-04-30T19:13:47 *** peerst has quit IRC (Ping timeout: 245 seconds) 2013-04-30T19:21:21 *** peerst has joined #rtems 2013-04-30T19:51:13 *** gedare has joined #rtems 2013-04-30T20:34:09 *** freenix has joined #rtems 2013-04-30T20:50:43 *** gedare has quit IRC (Read error: Operation timed out) 2013-04-30T21:37:46 *** S_Somani has joined #rtems 2013-04-30T21:56:52 *** S_Somani has quit IRC (Quit: Leaving.) 2013-04-30T21:56:54 *** hiddenpearls1 has joined #rtems 2013-04-30T22:04:29 *** hiddenpearls has joined #rtems 2013-04-30T22:06:24 *** hiddenpearls1 has quit IRC (Ping timeout: 256 seconds) 2013-04-30T22:17:29 *** mgahlot|away has joined #rtems 2013-04-30T22:18:10 *** mgahlot|away_ has joined #rtems 2013-04-30T22:18:48 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-04-30T22:20:12 *** hiddenpearls has joined #rtems 2013-04-30T22:30:31 *** mgahlot|away_ has quit IRC (Ping timeout: 268 seconds) 2013-04-30T22:30:31 *** mgahlot|away has quit IRC (Ping timeout: 268 seconds) 2013-04-30T23:02:31 *** hiddenpearls has quit IRC (Ping timeout: 245 seconds) 2013-05-01T00:04:34 *** monstr has joined #rtems 2013-05-01T00:12:40 *** hiddenpearls has joined #rtems 2013-05-01T00:16:40 *** hiddenpearls1 has joined #rtems 2013-05-01T00:17:06 *** hiddenpearls has quit IRC (Ping timeout: 245 seconds) 2013-05-01T00:21:20 *** hiddenpearls1 has quit IRC (Ping timeout: 260 seconds) 2013-05-01T00:27:19 *** hiddenpearls has joined #rtems 2013-05-01T00:32:03 *** hiddenpearls has quit IRC (Ping timeout: 260 seconds) 2013-05-01T00:44:20 *** arvind_khadri has joined #rtems 2013-05-01T00:51:02 *** sevikkk has joined #rtems 2013-05-01T01:09:41 *** hiddenpearls has joined #rtems 2013-05-01T01:42:24 *** sevikkk has quit IRC (Read error: No route to host) 2013-05-01T01:42:34 *** sevikkk has joined #rtems 2013-05-01T01:57:15 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-01T02:05:07 *** sevikkk has quit IRC (Quit: Leaving.) 2013-05-01T02:19:01 *** hiddenpearls has joined #rtems 2013-05-01T02:49:50 *** mgahlot|away has joined #rtems 2013-05-01T02:50:42 *** mgahlot|away_ has joined #rtems 2013-05-01T03:01:36 *** mgahlot|away_ has quit IRC (Remote host closed the connection) 2013-05-01T03:01:46 *** mgahlot|away is now known as mgahlot|off 2013-05-01T03:11:57 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-01T03:43:50 *** hiddenpearls has joined #rtems 2013-05-01T03:57:49 *** freenix has quit IRC (Remote host closed the connection) 2013-05-01T03:59:32 *** hiddenpearls has quit IRC (Ping timeout: 256 seconds) 2013-05-01T04:24:41 *** hiddenpearls has joined #rtems 2013-05-01T04:29:02 *** hiddenpearls has quit IRC (Ping timeout: 252 seconds) 2013-05-01T04:34:49 *** mgahlot|off has quit IRC (Remote host closed the connection) 2013-05-01T05:03:45 *** hiddenpearls has joined #rtems 2013-05-01T05:08:05 *** hiddenpearls has quit IRC (Ping timeout: 248 seconds) 2013-05-01T05:27:22 *** disdi has joined #rtems 2013-05-01T05:32:07 I want to to participate in GSOC 2013 with the project "BSP for the ARM Beagle Board simulated by QEMU " 2013-05-01T05:32:19 I have some questions 2013-05-01T05:32:45 can someone help 2013-05-01T06:03:30 *** hiddenpearls has joined #rtems 2013-05-01T06:05:10 disdi, ask the question, tell us what you have done, what do you want help with 2013-05-01T06:06:08 I am an electronic student from India 2013-05-01T06:06:30 My projects have been mainly around embedded systems 2013-05-01T06:06:51 I have worked on Embedded Linux too 2013-05-01T06:07:14 I have cross-compiled msp430, arm etc 2013-05-01T06:08:01 I have also customized the toolchains using buildroot , instead of just using pre-compiled toolchain 2013-05-01T06:08:33 *** hiddenpearls has quit IRC (Ping timeout: 276 seconds) 2013-05-01T06:08:51 I am intrested in "a BSP for the ARM Beagle Board on QEMU " 2013-05-01T06:09:07 need some guidance 2013-05-01T06:12:30 disdi, have you checked out what has to be done as gsoc student to apply? Have you finished any tests, if any and submitted the results 2013-05-01T06:12:58 the hello world test 2013-05-01T06:13:19 I shall be finishing it within one hour 2013-05-01T06:13:38 i just wanted to discuss the project topics in the meanwhile 2013-05-01T06:13:49 I am not a mentor, a mentor will come here in sometime, so hang around. 2013-05-01T06:14:30 on actually this group is active 2013-05-01T06:14:38 I am in IST time zone 2013-05-01T06:14:45 +5:30 2013-05-01T06:14:56 not group ,soory channel rather 2013-05-01T06:15:25 when actually this channel is active? 2013-05-01T06:34:57 *** disdi has quit IRC (Quit: Page closed) 2013-05-01T06:41:32 *** disdi has joined #rtems 2013-05-01T06:43:08 Dear Sir, I am interested in applying in GSOC 2013 with the project idea "BSP for ARM beagleboard" 2013-05-01T06:43:34 i have some questions 2013-05-01T06:44:43 can someone help 2013-05-01T07:26:10 I am interested in applying in GSOC 2013 with the project idea "BSP for ARM beagleboard" , i have some questions , can someone help 2013-05-01T07:29:53 *** hiddenpearls has joined #rtems 2013-05-01T07:32:42 *** gedare has joined #rtems 2013-05-01T07:56:06 I am interested in applying in GSOC 2013 with the project idea "BSP for ARM beagleboard" , i have some questions , can someone help 2013-05-01T07:57:00 disdi: that BSP has been completed, it just needs a little push from the developer to get it merged 2013-05-01T07:57:10 speaking of which..i should poke him with an email 2013-05-01T07:57:40 ? 2013-05-01T07:57:40 shall it be msp430 bsp 2013-05-01T07:57:52 sir which project should I take 2013-05-01T07:57:55 actually I want to work on BSP only 2013-05-01T07:58:09 well you can see his work at https://github.com/claas/rtems/tree/beagleboard 2013-05-01T07:58:22 You are interested in developing a BSP? 2013-05-01T07:58:41 yes sir with rtems support 2013-05-01T08:01:07 i think usually someone will develop a BSP because they have the board and want to use it with RTEMS 2013-05-01T08:01:15 is there a reason you want to do a BSP project? 2013-05-01T08:01:55 I have dona a lot of work on BSP 2013-05-01T08:02:16 I havent been a device driver or application developer 2013-05-01T08:02:37 so that is why I want to do it 2013-05-01T08:02:49 I dont see msp430 listed 2013-05-01T08:04:15 did you ask about that on the mailing list? 2013-05-01T08:04:20 there was a thread about it 2013-05-01T08:04:59 http://www.rtems.org/pipermail/rtems-devel/2013-April/003021.html 2013-05-01T08:05:52 ok.....do you have any suggesstions 2013-05-01T08:06:11 There are problems with the toolchain it looks like, and also concern about whether we can test the BSP. That makes it unlikely to be a good project. 2013-05-01T08:07:05 but if you can figure out those problems before we have to select projects then you might have a chance 2013-05-01T08:09:13 I will do that 2013-05-01T08:09:19 sir how to begin? 2013-05-01T08:10:16 Did you already go through our GSOC Getting Started? http://wiki.rtems.org/wiki/index.php/GSoC_Getting_Started 2013-05-01T08:10:54 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-01T08:11:22 yes sir 2013-05-01T08:12:35 send your proof to the rtems-devel mailing list, briefly describe your project, add yourself to the wiki page at http://wiki.rtems.org/wiki/index.php/RTEMSSummerOfCode#Students_Proposals and sign up officially at http://www.google-melange.com/gsoc/org/google/gsoc2013/rtems 2013-05-01T08:13:34 I am busy doing the hello world test only 2013-05-01T08:13:41 actually I use ubuntu 2013-05-01T08:13:46 so some problems 2013-05-01T08:13:51 but I will solve 2013-05-01T08:14:26 but I wanted to ask will I get support from rtems community if I move up with msp430 2013-05-01T08:14:41 I atleast need a mentor 2013-05-01T08:16:14 If your proposal is accepted you will be assigned a mentor. 2013-05-01T08:17:10 *** harrysg has joined #rtems 2013-05-01T08:17:38 ok sir 2013-05-01T08:18:07 *** harrysg has left #rtems 2013-05-01T08:19:47 thank you sir for the help 2013-05-01T08:19:56 *** disdi has quit IRC (Quit: Page closed) 2013-05-01T08:46:27 *** MegaAlex has joined #rtems 2013-05-01T08:47:16 Hi gedare, good evening 2013-05-01T08:48:23 I have build the arm, mips, bfin and sh successfully with the atomic patch 2013-05-01T08:48:57 but the gdb for those architecuture is not run ok, so i have not test the spatomic 2013-05-01T08:49:45 the compile should be ok for all the architectures 2013-05-01T08:51:20 hi weiY 2013-05-01T08:51:39 you need some special RPM for the mips 2013-05-01T08:52:04 but i ran that one already so it is ok 2013-05-01T08:52:23 "rtems-4.11-mipstx39-rtems4.11-gdb" 2013-05-01T08:52:42 and bsp jmr3904 2013-05-01T08:53:41 not all the bsp is run successfully on gdb 2013-05-01T08:54:27 i build bsp csb350 in the past hours and it did not run succssully on gdb 2013-05-01T08:55:26 what is the different between the mipstx39-rtems4.11-gdb and mips-rtems4.11-gdb 2013-05-01T08:55:39 dunno probably something to make the simulator part work 2013-05-01T08:55:52 there are problems with some of the arm...some work some do not 2013-05-01T08:56:31 for bfin and sh, there are also some work, some not 2013-05-01T08:57:34 ok. well i'm going to try to test one more, if it works I will push 2013-05-01T08:57:45 then we can just wait for someone to complain and fix it if there is a problem! 2013-05-01T08:58:10 for the generic atomic ops the implementation is the same, so if one architecture test is ok, then others are ok in theory 2013-05-01T08:58:19 yeah 2013-05-01T08:59:05 ok, i am build for lm32 and test whether the compile works 2013-05-01T09:00:40 and gedare, do you know how to build the schedule-sim source code 2013-05-01T09:01:05 i have done it before, but that is part of its problem... it gets bit rotted 2013-05-01T09:01:18 i use autoconf under the schedule-sim , it failed 2013-05-01T09:01:20 because it points at the source of RTEMS, but does not know anything about how RTEMS changes 2013-05-01T09:01:55 you can file a bug for Joel, maybe he will look at it and say what is wrong 2013-05-01T09:02:05 ok 2013-05-01T09:06:59 these tests are heavy! 2013-05-01T09:07:30 you should consider rewriting them as part of your GSOC 2013-05-01T09:08:38 *** hiddenpearls has joined #rtems 2013-05-01T09:11:58 yeah, i am considering rewriteing the atomic test 2013-05-01T09:12:10 it wastes so many times 2013-05-01T09:13:24 *** hiddenpearls has quit IRC (Ping timeout: 264 seconds) 2013-05-01T09:13:25 it also is not clear to me what it is testing 2013-05-01T09:15:06 at first those test is ported from the ConcurrencyKit . and it test the each atomic ops from its semantic 2013-05-01T09:15:27 like, atomic_fetch_add(a,b) == a+b 2013-05-01T09:15:51 ok. do any of them try to test the correctness during parallelism? 2013-05-01T09:16:09 well, i guess those should be written as smptests? 2013-05-01T09:16:16 but it does not guarantee its atomicity 2013-05-01T09:16:31 does not test its atomicity 2013-05-01T09:16:58 ok 2013-05-01T09:17:01 and the testcase of parallelism is not very easy to design 2013-05-01T09:17:07 i agree :) 2013-05-01T09:17:24 it might be better to do than the lock debug project though 2013-05-01T09:17:34 rather higher priority 2013-05-01T09:18:22 this can be a feature in the proposal, and i can research it those days. 2013-05-01T09:19:14 and see whether there are some exsit testcase for those scene 2013-05-01T09:19:48 ok. 2013-05-01T09:19:54 i ran a couple of the tests now on sparc64 also 2013-05-01T09:22:05 ok, great. thank you 2013-05-01T09:22:26 and the build time is a litter long 2013-05-01T09:23:01 when i build one architecture it use about half a hour in the virtualbox environment 2013-05-01T09:29:26 weiY: i have commited your changes with slight modification for whitespace fixes 2013-05-01T09:31:21 ok, thank you! i should pull the code. 2013-05-01T09:32:06 yeah 2013-05-01T09:34:34 gedare, do you know how to store a attachment in the network. and then i can put its link into my proposal 2013-05-01T09:41:04 you can upload a file to google drive 2013-05-01T09:41:16 and then share it with a link 2013-05-01T09:42:31 ok, i try. thank you 2013-05-01T09:44:18 *** hiddenpearls has joined #rtems 2013-05-01T09:48:35 *** hiddenpearls has quit IRC (Client Quit) 2013-05-01T10:02:24 *** disdi has joined #rtems 2013-05-01T10:11:41 *** disdi has left #rtems 2013-05-01T10:16:59 *** MegaAlex has quit IRC (Quit: Page closed) 2013-05-01T10:28:49 weiY: did you see the email about the patches? 2013-05-01T10:29:11 i think it will be necessary to remove the #ifdef__cplusplus...#endif blocks from the various score/cpu/..cpuatomic.h files 2013-05-01T10:40:32 yeah, i have seen the mail. if i i port the SMP mode atomic ops i do not know whether this is useful. now it seems that it can be removed 2013-05-01T10:41:56 if you port SMP you can do it in a separate include file and keep the UP mode? 2013-05-01T10:42:06 or does the SMP implement replace the UP? 2013-05-01T10:43:28 do not replace the UP. in my opinion i want to implement the SMP atomic in the same file cpuatomic.h 2013-05-01T10:44:05 which fill into the smp code into the #else ..... #endif 2013-05-01T10:44:27 because there is no generic smp atomic ops like UP mode 2013-05-01T10:44:59 do not need a separate file to implement it for each architecture 2013-05-01T10:46:39 now if there is a #error or #warn in the #else ... #endif it seems more better 2013-05-01T10:51:40 ok 2013-05-01T10:53:48 byb, i have to go 2013-05-01T10:53:52 ok 2013-05-01T10:53:59 *** weiY has quit IRC () 2013-05-01T10:58:58 *** haman has joined #rtems 2013-05-01T11:13:03 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2013-05-01T11:16:51 *** rtdin has joined #rtems 2013-05-01T11:23:04 *** haman has quit IRC (Quit: Page closed) 2013-05-01T11:23:11 hey rtdin .. have you finished your proposal? 2013-05-01T11:24:14 *** hiddenpearls has joined #rtems 2013-05-01T11:24:25 *** prasannatsm has joined #rtems 2013-05-01T11:24:30 I am working on that. I will complete it by today 2013-05-01T11:24:38 *** DrJoel has joined #rtems 2013-05-01T11:24:38 *** DrJoel has joined #rtems 2013-05-01T11:24:38 *** ChanServ sets mode: +o DrJoel 2013-05-01T11:24:55 hi DrJoel 2013-05-01T11:27:14 hey.. trying to be online and lay low.. trying to get something reviewed before noon here 2013-05-01T11:27:25 tick tock 2013-05-01T11:27:43 Scheduler simulator at least works again. So taht is off the plate for the gsoc students that could use ti. 2013-05-01T11:27:46 rtdin: good to hear 2013-05-01T11:27:52 good 2013-05-01T11:29:37 gedare :) 2013-05-01T11:42:24 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-01T11:52:06 *** prasannatsm has left #rtems 2013-05-01T11:54:47 *** prasannatsm has joined #rtems 2013-05-01T11:58:21 *** S_Somani has joined #rtems 2013-05-01T12:16:28 *** monstr has quit IRC (Ping timeout: 246 seconds) 2013-05-01T12:24:47 gedare did you notice the thread on lower application numbers this year? 2013-05-01T12:24:52 yup 2013-05-01T12:25:26 impressions? We appear to be slightly down but nothing drastic. We never got many spam or crap applications anyway though. 2013-05-01T12:25:34 i bet a good bit has to do with the 5 apps / student. 2013-05-01T12:25:51 And R is late in the alphabet. :) 2013-05-01T12:26:00 that will have culled many of the spam/weaker applications. 2013-05-01T12:26:03 heh, yeah 2013-05-01T12:26:09 I would love to see a graph of apps per letter 2013-05-01T12:26:12 anyway i expect we'll get another 2-3 2013-05-01T12:26:29 I think so too which puts us close to last year 2013-05-01T12:26:33 yep 2013-05-01T12:34:34 *** MegaAlex has joined #rtems 2013-05-01T12:36:25 Hello, me and Matt got our 'swag bags' from you guys, thanks :) Pictures comming soon:) 2013-05-01T12:36:27 i hate fruit flies 2013-05-01T12:36:47 cool..congrats 2013-05-01T12:37:08 Thank you:) 2013-05-01T12:40:57 How was Google? 2013-05-01T12:41:29 I will pass along to Jennifer that you liked the swag! She did all the work to put them together, order stuff, etc. 2013-05-01T12:48:42 Oh, we are still in the SF office for saying goodbye, group pictures etc. :) 2013-05-01T12:49:15 *** monstr has joined #rtems 2013-05-01T12:55:05 *** monstr has quit IRC (Ping timeout: 268 seconds) 2013-05-01T13:02:22 *** MegaAlex has quit IRC (Quit: AndroIRC - Android IRC Client ( http://www.androirc.com )) 2013-05-01T13:04:32 *** prasannatsm has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-05-01T13:08:44 http://www.kickstarter.com/projects/smogberrytrees/under-the-smogberry-trees-the-true-story-of-dr-dem 2013-05-01T13:24:39 *** rtdin has left #rtems 2013-05-01T13:26:03 just say no to kickstarters. 2013-05-01T13:26:11 ;) 2013-05-01T13:30:43 I must be a sucker .. lol .. the $20 for the reproduction of the 1975 t-shirt is very templting. 2013-05-01T13:31:09 You will laugh when the robotic bartender makes you an RTEMS :) 2013-05-01T13:31:44 i'm more of an ale or whiskey guy myself. 2013-05-01T13:34:30 I have reservations for the dedication of http://www.chattanoogawhiskey.com/. The law got changed and they can now really make it in Chattanooga. They are takign over a building that held an ice cream cone factory 2013-05-01T13:34:39 Not bad tasting stuff. :) 2013-05-01T13:37:36 nice 2013-05-01T13:40:06 I have a bottle of their 1816 Cask at home. Very easy to drink. 2013-05-01T13:42:09 looks like strong stuff. 2013-05-01T13:42:18 bet its good sippin 2013-05-01T13:43:32 It is indeed strong and slow sipping :) 2013-05-01T13:56:52 *** humam has joined #rtems 2013-05-01T14:50:32 *** DrJoel has quit IRC (Ping timeout: 252 seconds) 2013-05-01T14:54:42 *** gedare has quit IRC (Ping timeout: 264 seconds) 2013-05-01T15:13:23 *** humam has quit IRC (Ping timeout: 245 seconds) 2013-05-01T15:18:18 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-01T15:31:31 *** sevikkk has joined #rtems 2013-05-01T16:22:20 *** sevikkk has quit IRC (Quit: Leaving.) 2013-05-01T19:10:23 *** the9a3eedi has joined #rtems 2013-05-01T19:30:03 *** DrJoel has joined #rtems 2013-05-01T19:30:03 *** DrJoel has joined #rtems 2013-05-01T19:30:03 *** ChanServ sets mode: +o DrJoel 2013-05-01T19:37:30 *** freenix has joined #rtems 2013-05-01T19:47:12 *** DrJoel_ has joined #rtems 2013-05-01T19:47:22 kiwichris: awake and paying attention? 2013-05-01T19:47:40 *** DrJoel_ has quit IRC (Client Quit) 2013-05-01T19:47:52 Yeah here 2013-05-01T19:47:58 well .. I apparently was here beside myself 2013-05-01T19:48:07 Not any more 2013-05-01T19:48:20 I think I missed getting the 32-bit fix host fix in for ARM gcc into 4.8.0 2013-05-01T19:48:37 Can you confirm that? 2013-05-01T19:48:56 By building on a 32bit host ? 2013-05-01T19:49:11 Just checking the source if you have it handy 2013-05-01T19:49:34 I do not. I clean each build. It is easy to keep with --no-clean but I have not done this. 2013-05-01T19:50:00 I remember a one line patch to libgcc/config.host 2013-05-01T19:51:13 updating my gcc git checkout on rtbf64a 2013-05-01T19:51:46 FYI the google code in ceremony was today. Matt and Alex got their rtems t-shirt and stickers Jennifer sent to the google ospo 2013-05-01T19:51:56 Which one ? http://git.rtems.org/rtems-tools.git/tree/tools/4.11/gcc/arm ? 2013-05-01T19:52:30 http://git.rtems.org/rtems-tools.git/tree/tools/4.11/gcc/arm/gcc-4.7-rtems4.11-arm-eabi-libcpp-20130403.diff 2013-05-01T19:52:34 Nice. I have mine and being worn 2013-05-01T19:52:43 I think that patch is still needed 2013-05-01T19:53:09 Hmm ok. 2013-05-01T19:53:14 I need to check this. 2013-05-01T19:53:33 fyi the gsoc de-duplication irc is on my birthday and I would politely request to not be on irc that night :) 2013-05-01T19:54:06 just need to see which admin can sit in just in case.. we can always ask all students if they applied to multiple orgs but we need a rep there 2013-05-01T19:54:23 Ok. 2013-05-01T19:54:31 I think it is a bad time for me 2013-05-01T19:54:58 I suspect it is. It is after business hours in CA if I understood correctly. 2013-05-01T19:58:15 That is bad for Gedare 2013-05-01T19:58:53 Hmm.. well time will tell.. maybe Cynthia or Jennifer will be available 2013-05-01T19:59:19 FYI that patch is definitely needed for 4.8.0 and likely the head 2013-05-01T19:59:26 I will commit it tomorrow .. sigh 2013-05-01T19:59:36 Ok. 2013-05-01T19:59:49 Any timetable for 4.8.1 ? 2013-05-01T19:59:50 Did you see the email about the cia.vc email git hook? That needs to be killed/ 2013-05-01T20:00:02 Ah ok. I will do that. 2013-05-01T20:00:04 not that I know of.. but if I do this tomorrow it will be in 2013-05-01T20:00:18 verm__ should have github mirroring again.. we need to watch that. 2013-05-01T20:00:23 ok.. need to help with dinner 2013-05-01T20:00:25 cy 2013-05-01T20:00:29 cya 2013-05-01T20:00:30 *** DrJoel has quit IRC (Quit: OUCH!!!) 2013-05-01T20:01:26 *** the9a3eedi has quit IRC (Ping timeout: 272 seconds) 2013-05-01T20:06:53 *** the9a3eedi has joined #rtems 2013-05-01T22:00:35 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-05-01T22:00:50 *** the9a3eedi has joined #rtems 2013-05-01T22:12:33 *** hiddenpearls has joined #rtems 2013-05-01T22:34:04 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-01T22:35:53 *** hiddenpearls has joined #rtems 2013-05-01T22:59:57 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-01T23:12:58 *** hiddenpearls has joined #rtems 2013-05-01T23:35:15 *** hiddenpearls has quit IRC (Ping timeout: 252 seconds) 2013-05-01T23:37:45 *** hiddenpearls has joined #rtems 2013-05-01T23:47:19 *** the9a3eedi has quit IRC (Ping timeout: 264 seconds) 2013-05-01T23:49:50 *** monstr has joined #rtems 2013-05-01T23:51:45 *** the9a3eedi has joined #rtems 2013-05-01T23:52:20 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-02T00:11:33 *** sevikkk has joined #rtems 2013-05-02T00:15:56 *** S_Somani has joined #rtems 2013-05-02T00:16:32 *** S_Somani has left #rtems 2013-05-02T00:51:18 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-02T00:58:29 *** the9a3eedi-2 has joined #rtems 2013-05-02T00:58:29 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-05-02T01:18:01 *** kiwichris has joined #rtems 2013-05-02T01:37:33 *** arvind_khadri has joined #rtems 2013-05-02T01:48:31 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-02T02:05:17 *** kiwichris has joined #rtems 2013-05-02T02:13:42 *** hiddenpearls has joined #rtems 2013-05-02T02:14:49 *** hiddenpearls has quit IRC (Client Quit) 2013-05-02T02:14:51 *** S_Somani has joined #rtems 2013-05-02T02:16:05 *** hiddenpearls has joined #rtems 2013-05-02T02:20:55 *** hiddenpearls has quit IRC (Ping timeout: 264 seconds) 2013-05-02T02:24:31 *** S_Somani has quit IRC (Ping timeout: 264 seconds) 2013-05-02T02:26:09 *** rtdin has joined #rtems 2013-05-02T02:48:39 *** S_Somani has joined #rtems 2013-05-02T03:46:17 *** disdi has joined #rtems 2013-05-02T03:48:01 *** hiddenpearls has joined #rtems 2013-05-02T03:54:33 *** S_Somani has left #rtems 2013-05-02T04:11:30 *** the9a3eedi-2 has quit IRC (Read error: Connection reset by peer) 2013-05-02T04:49:04 *** disdi has quit IRC (Quit: Page closed) 2013-05-02T04:52:35 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-02T04:58:41 *** hiddenpearls has joined #rtems 2013-05-02T05:10:24 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-02T05:19:23 *** hiddenpearls has joined #rtems 2013-05-02T05:26:31 *** antgreen has joined #rtems 2013-05-02T06:19:45 *** rtdin_ has joined #rtems 2013-05-02T06:20:44 *** rtdin has quit IRC (Ping timeout: 255 seconds) 2013-05-02T06:20:55 *** rtdin_ is now known as rtdin 2013-05-02T06:21:28 *** antgreen has quit IRC (Ping timeout: 272 seconds) 2013-05-02T06:28:33 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-02T06:30:01 *** kiwichris has joined #rtems 2013-05-02T06:33:31 *** antgreen has joined #rtems 2013-05-02T06:52:32 *** disdi has joined #rtems 2013-05-02T06:56:06 *** disdi has quit IRC (Client Quit) 2013-05-02T07:58:55 *** hiddenpearls has quit IRC (Quit: Leaving.) 2013-05-02T08:15:53 *** weiY has joined #rtems 2013-05-02T08:42:36 *** sevikkk has quit IRC (Quit: Leaving.) 2013-05-02T08:46:31 *** rtdin_ has joined #rtems 2013-05-02T08:48:33 *** rtdin has quit IRC (Ping timeout: 240 seconds) 2013-05-02T08:48:38 *** rtdin_ is now known as rtdin 2013-05-02T09:20:36 *** freenix has quit IRC (Quit: Leaving) 2013-05-02T09:47:36 *** arvind_khadri has quit IRC (Ping timeout: 264 seconds) 2013-05-02T09:56:10 *** monstr has quit IRC (Ping timeout: 268 seconds) 2013-05-02T10:08:44 *** rtdin has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-05-02T10:45:10 *** gedare has joined #rtems 2013-05-02T10:46:03 *** S_Somani has joined #rtems 2013-05-02T10:46:11 *** S_Somani has left #rtems 2013-05-02T11:39:48 *** phipse has joined #rtems 2013-05-02T11:40:08 hi 2013-05-02T11:43:47 *** arvind_khadri has joined #rtems 2013-05-02T12:05:57 *** S_Somani has joined #rtems 2013-05-02T13:52:24 *** S_Somani has quit IRC (Ping timeout: 264 seconds) 2013-05-02T13:52:26 *** DrJoel has joined #rtems 2013-05-02T13:52:27 *** DrJoel has joined #rtems 2013-05-02T13:52:27 *** ChanServ sets mode: +o DrJoel 2013-05-02T13:53:46 *** S_Somani has joined #rtems 2013-05-02T13:58:52 hi phipse 2013-05-02T13:59:10 looks like i'm only on a 2 hour buffered delay 2013-05-02T13:59:12 ;) 2013-05-02T14:02:05 gedare did you see the new SMP project yet. Just a yes or no.. no public commentary obviously 2013-05-02T14:11:20 DrJoel: You around? 2013-05-02T14:12:10 apaprently a bit more than gedare.. 2013-05-02T14:12:20 i did not see it 2013-05-02T14:12:25 i'm sending paperwork for a refinance 2013-05-02T14:12:34 gedare ok.. 2013-05-02T14:12:39 i'll check projects hopefully still today, but I may not get around to it before the deadline. 2013-05-02T14:12:41 good time for refinance. 2013-05-02T14:12:49 yup..its going to be almost free 2013-05-02T14:12:57 we'll recoup the cost in about 4 months 2013-05-02T14:13:44 We moved 12 years ago next week and started with a 30 since we hadn't sold the old house. A few years later cut ours back to 15 years and really didn't change the payments. 2013-05-02T14:13:59 dhananjay ask away 2013-05-02T14:20:53 I was looking at the consistancy checks in sim-scripts 2013-05-02T14:21:16 ok and .. what's the question :) 2013-05-02T14:21:42 I could only find one command rtems_check_state. Is that it? 2013-05-02T14:22:01 Do we have any other documetnation about the good practices in RTEMS? 2013-05-02T14:22:20 so that we can pull them into a checks module? 2013-05-02T14:22:27 rtems_check_state is the key consistency checker 2013-05-02T14:22:54 Off the top of my head, the command to dump the watchdog chain is another one that is not object related 2013-05-02T14:22:56 let me look 2013-05-02T14:24:10 ok 2013-05-02T14:29:11 looking at the commands at the top.. I see : 2013-05-02T14:30:24 rtems_task_bracktrack, rtems_malloc_walk, rtems_internal_ticks_chain, rtems_internal_seconds_chain, rtems_workspace_walk, rtens_tod, and rtems_check_state 2013-05-02T14:30:44 the other commands implemented all are focused on objects w/Ids and are sharing underlying helpers 2013-05-02T14:31:10 but rtems_check_state is the only true consistency checker 2013-05-02T14:31:50 Hmm. 2013-05-02T14:32:33 Where can I find $isrnl > $tddl is bad? 2013-05-02T14:32:35 melange is at a crawl right now. going to just read what i got in the email 2013-05-02T14:32:44 but won't give any feedback at melange til later 2013-05-02T14:33:21 the documntation about whats going on. 2013-05-02T14:33:27 is there a leon4 bsp? 2013-05-02T14:34:44 this proposal reads like a research whitepaper, will have to see what the student comes up with from looking at the git head.. 2013-05-02T14:34:53 leon4 can use leon3 .. no impacting hw changes to bsp per Daniel. Also no free simulator so testing will be an issue 2013-05-02T14:35:09 ok 2013-05-02T14:35:26 scheduler shouldn't be hw specific, so it doesn't matter much 2013-05-02T14:35:44 unless you just mean for RTEMS to have a leon4 BSP the lack of simulator is an isuse. 2013-05-02T14:35:50 well he does have to identify a path to test it (and show it isn't the same as what is there now) 2013-05-02T14:36:04 ah. 2013-05-02T14:36:10 dhananjay: $isrnl == ISR nest level 2013-05-02T14:36:19 dhananjay: $tddl is thread dispatch disable level 2013-05-02T14:39:40 does eclipse use gdb as the debugger? 2013-05-02T14:39:43 I understood that.am ot able to undestand the connection. 2013-05-02T14:41:26 dhananjay: have you filed FSF paperwork to be able to submit patches to GDB? 2013-05-02T14:41:37 (Is the python extensions part of core GDB?) 2013-05-02T14:45:05 gedare: No. I am little cfused about it. 2013-05-02T14:45:24 about the paparwork,.can you give me a link? 2013-05-02T14:45:59 eclipse uses gdb through the "MI" machine interface. Under the hood it si gdb 2013-05-02T14:46:00 No we can keep it seperate if we wish to 2013-05-02T14:46:53 dhananjay: keeping it separate is better.. otherwise we have to sync our changes with gdb.. 2013-05-02T14:46:58 dhananjay: you should email the GDB mailing list and ask for FSF copyright assignment forms. Someone from GDB maintainers will send you a text file to fill in which will have info on where to send it.. (assign@gnu.org, and CC the GDB maintainer) 2013-05-02T14:47:11 H 2013-05-02T14:47:15 Hmm. 2013-05-02T14:47:32 Okaywill do. 2013-05-02T14:47:41 you can read about contributing to gdb: http://sourceware.org/cgi-bin/cvsweb.cgi/~checkout~/src/gdb/CONTRIBUTE?content-type=text/x-cvsweb-markup&cvsroot=src 2013-05-02T14:48:00 :) 2013-05-02T14:48:05 isr nest level is now in _Per_CPU_Information which is technically an array with one instance per core 2013-05-02T14:48:31 some items which used to be explicit variables are not in that structure and instance on a per cpu basis. That is NOT reflected in the macros. 2013-05-02T14:48:57 two issues to figure our are (1) any architecture specific differences which have to be address. (e.g. stack backtraces) 2013-05-02T14:49:16 and (2) smp impact .. one per cpu structure versus N 2013-05-02T14:49:43 (3) optional apis.. what is POSIX is disabled? Does that impact your code? 2013-05-02T14:50:19 DrJoel: did you see the patch about the warning fix? did you have any opinion? 2013-05-02T14:50:34 Yes.. I need to lose weight.. 2013-05-02T14:50:57 try weight watchers 2013-05-02T14:51:07 oh.. on the code.. I imagine it is a comparision with zero or useless comparison on unsigned variable 2013-05-02T14:51:22 exactly.. do we care to fix it though? 2013-05-02T14:51:46 adding a couple of CPP commands to avoid a comparison the compiler should just eliminate..i dunno. 2013-05-02T14:52:09 I would fix it but not that way 2013-05-02T14:52:12 heh 2013-05-02T14:52:20 afraid you'd say that! 2013-05-02T14:52:23 I would break it into something like: 2013-05-02T14:52:28 #if 2013-05-02T14:52:49 if (vector <= (rtems_vector_number) BSP_INTERRUPT_VECTOR_MAX) return false; 2013-05-02T14:52:51 #endif 2013-05-02T14:52:57 return vector <= (rtems_vector_number) BSP_INTERRUPT_VECTOR_MAX; 2013-05-02T14:53:19 And document why that first one is conditional. I think there are other places in the code base that do that. 2013-05-02T14:53:42 That first if should have been _MIN.. bad cut and paste 2013-05-02T14:53:49 yeah i understand 2013-05-02T14:53:56 i have to go though, maybe i'll revisit it 2013-05-02T14:54:01 but we need the warning to be logged with the comment and commit. 2013-05-02T14:54:10 agreed.. 2013-05-02T14:54:12 do you know his BSP 2013-05-02T14:54:17 DrJoel: I will look into it. the (2) is a new challenge that I havent anticipated till now. 2013-05-02T14:54:43 there is something distasteful about superfluous patches to fix warnings :) 2013-05-02T14:54:50 no i don't. 2013-05-02T14:55:08 but now i'm going to go enjoy the day 2013-05-02T14:55:20 I would say just be aware of the Per CPU structure and wrap references to it. May want to investigate later changing the CPU focus. I dont' 2013-05-02T14:55:26 don't know how gdb addresses that 2013-05-02T14:58:20 Mm. I cant find anything on google. 2013-05-02T14:58:48 Nothing on a quick look, I will have to dig deeper. 2013-05-02T15:00:10 *** gedare has quit IRC (Ping timeout: 272 seconds) 2013-05-02T15:01:26 I know with grsim, you would have to issue a command to change the focus. its command line focuses on 1 core at time 2013-05-02T15:08:05 Havent used it - is there any tutorial? 2013-05-02T15:10:01 I am way too sleepy. Should sleep. 2013-05-02T15:10:09 It is an expensive closed source simulator for the LEON3. Their docs are online at gaisler.com 2013-05-02T15:10:13 Night. You are doing well 2013-05-02T15:10:22 SOrry. Will mail you. 2013-05-02T15:10:40 no problem.. irc is faster when we both are here and aware 2013-05-02T15:10:43 awake 2013-05-02T15:17:19 *** Hesham has joined #rtems 2013-05-02T15:57:13 DrJoel: Can I get your opionion on the scale of my paravirtualization project? You mentioned it is challenging and I feel a little bit uncomfortable with the deliverables. 2013-05-02T16:03:21 *** Hesham has quit IRC (Ping timeout: 256 seconds) 2013-05-02T16:03:22 *** Hesham1 has joined #rtems 2013-05-02T16:09:31 *** Hesham1 has quit IRC (Quit: Leaving.) 2013-05-02T16:12:29 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-02T16:14:03 *** kiwichris has joined #rtems 2013-05-02T16:34:55 phipse still there? 2013-05-02T16:36:20 jep 2013-05-02T16:36:42 sorry.. was down the hall 2013-05-02T16:38:09 I just mean that it involves a number of areas. Pok, RTEMS, integration 2013-05-02T16:38:38 When I look at my list of deliverables it feels like the balance between midterm and final isn't very good. 2013-05-02T16:38:42 I saw your mail to Pok this morning. If you can address that one issue and keep the current hello world running, you have a big step up 2013-05-02T16:39:06 did you try a project planning sw package? 2013-05-02T16:39:18 yes 2013-05-02T16:39:58 but that startup the RTEMS guest issue is kind of frightening ;) 2013-05-02T16:40:53 I can't estimate the time consumption by this 2013-05-02T16:40:55 I agree. :) That was always an issue for me. 2013-05-02T16:41:25 My approach is to ask for outside help early and work on other things while waiting for a response 2013-05-02T16:41:45 Which I think you are doing also 2013-05-02T16:41:52 Well, the wiki page is getting crowded ;) 2013-05-02T16:42:18 Have you run the Pok examples? 2013-05-02T16:43:27 yeah, the partition-threads one and I wrote a hello world 2013-05-02T16:43:35 just one partition to get a feeling for aadl. 2013-05-02T16:44:06 By now i read the OSADL11 paper by Julien, which explains a lot 2013-05-02T16:44:57 DrJoel, hi 2013-05-02T16:45:06 RSB 4.11 gcc .... 2013-05-02T16:45:07 Source0: git://gcc.gnu.org/git/gcc.git?branch=master?pull?reset?branch=gcc-4_8-branch 2013-05-02T16:45:09 ? 2013-05-02T16:46:59 Julien is at the CMU SW Eng Institute in Pittsburg now. He can't mentor this year but will try to answer questions 2013-05-02T16:47:33 Do they have an example with foriegn code? 2013-05-02T16:48:17 kiwichris: that should have the patch.. gcc 4.8.0 did not 2013-05-02T16:48:41 I will commit this until 4.8.1 comes out. Ok ? 2013-05-02T16:48:56 fyi http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48231 has follow up from Jeff Law.. he says the h8 patch posted there is incorrect 2013-05-02T16:49:04 sounds like a plan for gcc 2013-05-02T16:49:33 I haven't come across one, but I didn't look into most of the examples yet. 2013-05-02T16:49:44 IMO We should consider dropping the h8 multilibs for the extremely small targets.. some only have 64k address space 2013-05-02T16:50:34 phipse can you identify how they link their code? What are the steps/commands? It has been a long time since I looked :( 2013-05-02T16:51:08 I agree. 2013-05-02T16:51:35 Why this still exists is beyond me 2013-05-02T16:51:54 Did you get the build result emails ? 2013-05-02T16:52:18 http://www.rtems.org/pipermail/rtems-tooltestresults/2013-May/date.html 2013-05-02T16:52:34 Look ok ? 2013-05-02T16:52:36 I am not subscribed to that but it is archiving 2013-05-02T16:52:55 How do you post without being subscribed ? 2013-05-02T16:53:02 I had to subscribe 2013-05-02T16:54:11 I am on the exceptions list.. :) We need to fgure out how to let anyone post and get through without an admin. Perhaps a filter matching the subjects 2013-05-02T16:54:23 On the m32c, why did it continue to build RTEMS when gcc failed? 2013-05-02T16:55:29 Because --regression turns on --keep-goiing --always-clean --no-install 2013-05-02T16:56:22 ok.. bit if binutils fails.. gcc/newlib+rtems is pointless.. if gcc/newlib fails, rtems is pointless 2013-05-02T16:57:21 Sure I know this but the tool has been asked to keep going. It is an outer build set calling the inner build sets. 2013-05-02T16:57:41 I make no different in the processing of the build set files. 2013-05-02T16:58:02 ok. Without marking some "inner sets" as failure for outer set, there is no way to know. 2013-05-02T16:58:04 ie if arm fails build avr etc 2013-05-02T16:58:28 Yeah and I have avoided it knowing we have this issue 2013-05-02T16:59:09 :) Yep.. I tried hard to account for this in rtems-testing/gcc. Get as much info as possible. But stop when it was pointless and proceed to the next step 2013-05-02T16:59:55 Sure. Being generic some if benefits and side effects 2013-05-02T17:00:03 some _> comes 2013-05-02T17:00:27 yeah.. you would have to add a "prereqs" concept I suppose 2013-05-02T17:00:51 And this tool is not a package manager :) 2013-05-02T17:01:19 :) 2013-05-02T17:02:19 http://git.rtems.org/chrisj/rtems-source-builder.git/commit/?id=cd3d3ae4b2ca2ce62a6543e27372a67ec3927623 2013-05-02T17:02:37 The change to switch to a git branch. Nice and simple. 2013-05-02T17:03:47 I didn't dive into the linker script, but from the looks of it, every partition is a binary on its own. So there might be a way of inserting our binary. 2013-05-02T17:04:45 but that's more speculation than reliable info. I'll look into it. 2013-05-02T17:08:15 It may help to diagram their linking flow. And how you get the addresses per partition. I suspect that we will end up generating Pok partition "templates" for a few cases. One partition and two partition setup being obvious. Figure out what address/RAM parameters to give the RTEMS image when linking it. Then insert it into the pre-generated template 2013-05-02T17:11:53 The RTEMS binary will contain the POK definitions? 2013-05-02T17:13:28 For a first cut, they could. Incrementally start with what he left. Fix the binary cat to something more correct. Then we can change the "merge" step. 2013-05-02T17:14:32 IThen we can figure out how to produce a nearly completely linked RTEMS image which is bound to the library from Pok and to addresses/length to be a Pok multi-partition program. 2013-05-02T17:15:06 A first cut of that process would be the same hello binary compiled by RTEMS, "massaged" into two partition images", and then running in two partitions. 2013-05-02T17:23:09 So the RTEMS linkcmds for the BSP has to care for the POK partition structure. 2013-05-02T17:24:26 Thanks, I'll look into it tomorrow. Right now the fever is messing with my brain. 2013-05-02T17:24:29 Good night 2013-05-02T17:24:35 For a first step to remove the cat. Yes. Assume a "Pok one partition template" and link the BSP to fit that memory layout. 2013-05-02T17:24:43 Get to feeling better 2013-05-02T17:24:56 I was just suggesting an incremental way to nibble at it. 2013-05-02T17:25:21 *** phipse has left #rtems 2013-05-02T18:01:21 *** DrJoel has quit IRC (Quit: If at first you don't succeed, skydiving is not for you) 2013-05-02T18:16:41 *** weiY has quit IRC (Ping timeout: 255 seconds) 2013-05-02T18:25:25 *** new_lido_ has joined #rtems 2013-05-02T19:07:31 *** Hesham has joined #rtems 2013-05-02T19:14:51 *** Hesham has left #rtems 2013-05-02T19:16:44 *** the9a3eedi has joined #rtems 2013-05-02T19:28:03 *** arvind_khadri has quit IRC (Ping timeout: 260 seconds) 2013-05-02T19:28:31 *** new_lido_ has quit IRC (Ping timeout: 260 seconds) 2013-05-02T19:58:16 *** arvind_khadri has joined #rtems 2013-05-02T20:01:02 antgreen, hi 2013-05-02T20:05:53 *** freenix has joined #rtems 2013-05-02T20:09:45 *** the9a3eedi-2 has joined #rtems 2013-05-02T20:10:22 *** the9a3eedi has quit IRC (Ping timeout: 268 seconds) 2013-05-02T20:44:05 *** the9a3eedi-3 has joined #rtems 2013-05-02T20:47:36 *** the9a3eedi-2 has quit IRC (Ping timeout: 264 seconds) 2013-05-02T20:49:54 hi kiwichris 2013-05-02T20:50:16 trying to fix my stupid uart 2013-05-02T20:50:33 hehe 2013-05-02T20:50:43 Looking at the current moxie rtems tools ... 2013-05-02T20:50:44 http://www.rtems.org/pipermail/rtems-tooltestresults/2013-May/000638.html 2013-05-02T20:51:17 rohoh 2013-05-02T20:51:24 I did not pull in the moxie patch I had. Is this going to be committed into gcc ? 2013-05-02T20:51:41 which patch was that again? 2013-05-02T20:52:13 Ah I need to find it. It is not in the rtems-tools git repo. 2013-05-02T20:52:45 I build RTEMS tools from upstream sources, no patches 2013-05-02T20:52:46 Oh it is failing building libgcc. I wonder why 2013-05-02T20:53:17 The gdb patch is failing on the latest gdb. Maybe I was confused. 2013-05-02T20:53:51 I am building with newlib taken from cvs 2013-05-02T20:54:38 Looks like I need to remove the gdb patch as it is merged. 2013-05-02T20:54:40 Nice 2013-05-02T20:56:00 Hmm I see the RSB does not append the build report when a fail happens. That is silly because the record of why is not present. 2013-05-02T20:58:42 if you have a log showing how libgcc failed to build, please send it my way 2013-05-02T20:59:23 I will generate a build locally and see what is happening. 2013-05-02T21:00:54 oh.. and I just figured out my uart problem, after almost of week of head scratching. time to take break! 2013-05-02T21:01:22 great and well done 2013-05-02T21:23:40 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-02T21:24:06 *** S_Somani has joined #rtems 2013-05-02T21:28:00 *** the9a3eedi-3 has quit IRC (Ping timeout: 260 seconds) 2013-05-02T21:29:21 *** S_Somani has quit IRC (Ping timeout: 276 seconds) 2013-05-02T22:18:23 *** rtdin has joined #rtems 2013-05-02T22:34:13 *** arvind_khadri has quit IRC (Ping timeout: 240 seconds) 2013-05-02T23:03:42 *** rtdin has quit IRC (Ping timeout: 264 seconds) 2013-05-02T23:51:14 *** the9a3eedi has joined #rtems 2013-05-03T00:09:43 *** arvind_khadri has joined #rtems 2013-05-03T00:16:55 *** monstr has joined #rtems 2013-05-03T00:45:59 *** S_Somani has joined #rtems 2013-05-03T00:47:27 *** rtdin has joined #rtems 2013-05-03T01:03:17 *** monstr has quit IRC (Ping timeout: 248 seconds) 2013-05-03T01:04:22 *** monstr_ has joined #rtems 2013-05-03T01:06:43 Say I had a system running RTEMS in the field, and I wanted to upload a new task, or a new version of a task, and have it get loaded and run (without rebooting). Is this currently possible? 2013-05-03T01:08:05 I found this, http://www.rtems.org/wiki/index.php/Dynamic_Object_File_Loading , but I was wondering if there are other methods. 2013-05-03T01:08:23 because I don't think this allows for modules to be unloaded or updated 2013-05-03T01:11:14 dlclose unloads 2013-05-03T01:11:44 You need to make sure you stop the task and release any resources it has been calling dlclose 2013-05-03T01:11:55 been -> before 2013-05-03T01:12:42 It is rare in embedded systems like this to change functionality mid-function and make it do something new. 2013-05-03T01:26:13 I see. 2013-05-03T01:26:52 Well it doesn't have to do something entirely new.. but things like bugfixes while the system is running would be nice 2013-05-03T01:27:23 but I guess dlclose does the job 2013-05-03T01:29:32 stopping the task before unloading/reloading is fine, as long as there's a way to do it safely without rebooting the whole thing 2013-05-03T01:31:28 Task, group of tasks. You could do this but is it worth it ? Hard to test you are full independent. It is like unloading and reloading the menu handling part of your web browser. 2013-05-03T01:33:47 yeah it really depends on the system overall.. it would have to be designed that way 2013-05-03T01:38:57 Correct. When you at overall system requirements sometimes there are other ways to solve this problem. 2013-05-03T01:45:35 *** sebhub has joined #rtems 2013-05-03T01:48:58 good morning 2013-05-03T02:15:51 morning 2013-05-03T02:49:16 *** S_Somani has quit IRC (Read error: Connection reset by peer) 2013-05-03T03:31:03 *** the9a3eedi-2 has joined #rtems 2013-05-03T03:34:20 *** the9a3eedi has quit IRC (Ping timeout: 260 seconds) 2013-05-03T05:23:35 *** the9a3eedi-2 has quit IRC (Read error: Connection reset by peer) 2013-05-03T06:39:54 *** the9a3eedi has joined #rtems 2013-05-03T06:46:19 *** phipse has joined #rtems 2013-05-03T06:46:29 hi 2013-05-03T06:46:30 *** S_Somani has joined #rtems 2013-05-03T07:33:45 *** freenix has quit IRC (Remote host closed the connection) 2013-05-03T07:34:39 *** freenix has joined #rtems 2013-05-03T07:37:37 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-03T07:38:28 *** kiwichris has joined #rtems 2013-05-03T08:13:40 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-05-03T08:35:13 *** prateekt has joined #rtems 2013-05-03T08:49:41 *** rtdin has left #rtems 2013-05-03T08:51:10 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-03T09:25:57 *** gedare has joined #rtems 2013-05-03T09:27:42 *** kkthecoder has joined #rtems 2013-05-03T09:30:48 *** DrJoel has joined #rtems 2013-05-03T09:30:49 *** DrJoel has joined #rtems 2013-05-03T09:30:49 *** ChanServ sets mode: +o DrJoel 2013-05-03T09:35:23 DrJoel, good response we should inquire for the student too 2013-05-03T09:35:35 *** weiY has joined #rtems 2013-05-03T09:35:57 i'm skimming the gsoc rules 2013-05-03T09:41:47 I thought he asked Carol himself 2013-05-03T09:41:55 DrJoel, matthew's project will make an awesome demo... 2013-05-03T09:42:10 what's the project? 2013-05-03T09:42:19 did he? I did not see any mention of where you found the requirements. 2013-05-03T09:42:22 see rtems-users ml 2013-05-03T09:42:26 "Cheap avionics board" 2013-05-03T09:44:49 the smp project now severely overlaps with the supercore scheduler project 2013-05-03T09:46:07 brb 2013-05-03T09:47:15 https://groups.google.com/forum/?hl=en&fromgroups=#!searchin/google-summer-of-code-discuss/thangavel%7Csort:relevance 2013-05-03T09:48:06 ahh.. so the problem is not being legally able to do work while in school for aug/sep 2013-05-03T09:48:27 what is the date for "pencils down"? 2013-05-03T09:49:26 i requested the SMP scheduling student to come to IRC if possible. 2013-05-03T09:49:44 i'll try to stay lurking here though i have a couple meetings so i might be away from time to time. 2013-05-03T10:01:30 DrJoel, should we ask students if they are receiving money from another source that might have a claim to the work they do, e.g. a fellowship? 2013-05-03T10:01:32 afk 2013-05-03T10:03:42 *** arvind_khadri has quit IRC (Ping timeout: 264 seconds) 2013-05-03T10:04:01 gedare sounds like it on the conflicts thread. 2013-05-03T10:04:16 phipse: looks like you are making progress :) 2013-05-03T10:04:35 gedare: the newer student? who started with 4.9? 2013-05-03T10:05:09 *** freenix has quit IRC (Remote host closed the connection) 2013-05-03T10:07:46 DrJoel: yes, that session yesterday evening got me thinking and this morning I just tried if it would work. 2013-05-03T10:16:41 it looked like it was close. Did the original procedure from last year work for you? 2013-05-03T10:19:31 I followed his description and got it working, but I didn't come around to modify his code. 2013-05-03T10:20:08 that's the baseline :) and where you have to start from. 2013-05-03T10:20:54 Figure out the linking issue. A clock tick mechanism in Pok. And work through your abstraction library and you have nailed it I think. 2013-05-03T10:21:25 Pok and more advanced services like TCP/IP and non-volatile storage is likely a bigger fish to fry 2013-05-03T10:22:25 His deep integration of POK in RTEMS worries me a little 2013-05-03T10:22:46 I try to find a way around this. 2013-05-03T10:48:29 I thought it wasn't much beyond the syscalls. Are those not well isolated? 2013-05-03T10:48:55 It they are contained enough, then you can use one base BSP with variants which change the system calls to match other hypervisors. 2013-05-03T10:50:18 it's ok to throw out the past gsoc's code with good reason... just make sure you get at least as far as he did rather quickly ;) 2013-05-03T10:50:50 i'm going to go through each proposal and ask a "form question"... 2013-05-03T10:51:02 *** S_Somani has joined #rtems 2013-05-03T10:51:14 what is the question? 2013-05-03T10:51:21 working on a draft, will post here 2013-05-03T10:54:18 "Hi, Please state whether you may have any conflicts that may not allow you to contribute code you develop during GSoC. Possible reasons would include participating in a summer internship, or a scholarship/fellowship at your school that restricts your ability to claim copyright on your work." 2013-05-03T10:54:20 something like that? 2013-05-03T10:54:22 brb..lunch 2013-05-03T10:54:47 DrJoel: gedare:Hi, Any last minute suggestions? 2013-05-03T10:55:17 *** sebhub has quit IRC (Remote host closed the connection) 2013-05-03T11:00:50 *** sreeharshakondur has joined #rtems 2013-05-03T11:05:28 hey gedare, 2013-05-03T11:08:50 S_Somani: I read your last comment. Just to double check .. the proposal itself reflects the current state of your work and prototyping. 2013-05-03T11:10:17 Umm I have started the work on the TUI proto...but since not much work has been done on that front so didnt add it in the proposal 2013-05-03T11:11:21 ok.. I think you are in better shape as a student than we are as mentors reviewing things. lol 2013-05-03T11:11:49 :p 2013-05-03T11:12:53 I try to read and comment on all. I haven't made it completely through this year yet. Asking me questinos on IRC is more effective apparently. 2013-05-03T11:13:20 I don't think I have commented officially on phipse's proposal on Melange but we have discussed things a lot 2013-05-03T11:13:41 I would recommend making sure students (ALL) are comfortable with their mid-term and final goals. 2013-05-03T11:15:59 *** kkthecoder has quit IRC (Ping timeout: 256 seconds) 2013-05-03T11:16:41 Joel, does RTEMS currently have per core scheduling and a possibility for migration in place? 2013-05-03T11:21:16 Simple SMP Scheduler treats all cores as a set to which any thread can be assigned. Since it is one address space for all of the cores in SMP, there is no migration except to assign it to a core. The application has a set of threads which are ready to execute and (ignoring the possibility of no-preempt) the Simple SMP Scheduler assigns the N highest priority to the N cores. There is more to the scheduler than that. When evaluating the thread to cor 2013-05-03T11:21:58 The code is in score/src/schedulersimplesmpschedule.c 2013-05-03T11:22:20 You can see it in action using the rtems-schedsim and the scripts 2013-05-03T11:22:53 thank you, i'll look at it. 2013-05-03T11:23:22 hi sreeharshakondur 2013-05-03T11:23:28 hey 2013-05-03T11:23:46 i think it would be good for you not to worry too much about the Thread Set refactoring 2013-05-03T11:23:57 but instead to focus on the SMP scheduling 2013-05-03T11:24:24 fwiw the code is fairly well documented and intentionally simple minded in implementation. Easy to read before clever. The second SMP scheduling algorithm can be clever 2013-05-03T11:24:26 and the path to testing your scheduler, which should include schedsim, and i think pc386/qemu also supports running RTEMS with SMP 2013-05-03T11:24:51 i would like to see a "real" real-time multiprocessor scheduling algorithm. 2013-05-03T11:25:15 I think Global EDF is plausible. it depends on a global EDF queue with a lock, and per-core queues 2013-05-03T11:25:22 * gedare is eating lunch 2013-05-03T11:25:23 yes .. I tested smp with qemu and pc585-sse or similar. qemu must be told to be a core2duo to get the instructinos you want 2013-05-03T11:26:02 if your project is accepted, DrJoel will give you some guidance since he is most familiar with the SMP stuff. 2013-05-03T11:26:35 gedare: If doing Global EDF, would it make sense to add scenarios which exercise regular EDF to the scheduler simulator? Then work on global EDF and scenarios for it? 2013-05-03T11:26:59 and you can consider the Thread Set refactoring as a possibility if you get the global edf working in good time. 2013-05-03T11:27:16 DrJoel, do you have RM scenarios for schedsim? 2013-05-03T11:27:35 i.e. has it been tested with periodic tasks? 2013-05-03T11:27:58 *** sreeharshakondur has quit IRC (Ping timeout: 245 seconds) 2013-05-03T11:28:18 No. That would be a nice addition. It only has scenarios for deterministic priority and simple smp. 2013-05-03T11:28:55 I **recall** that Simple SMP on a single core would run the DPS scenarios with the same output. 2013-05-03T11:29:27 Adding enough to do the same for Simple Priority (uniprocessor) would be good as well. 2013-05-03T11:30:09 should only be a matter of getting periods in the command language. 2013-05-03T11:30:21 the student "quit" not sure why 2013-05-03T11:30:30 proof read for me please: 2013-05-03T11:30:31 Hi 2013-05-03T11:30:31 Please state whether you may have any conflicts that may not allow you to contribute code you develop during GSoC. Possible reasons would include participating in a summer internship, or a scholarship/fellowship at your school that restricts your ability to claim copyright on your work. Although working or being funded by your school can be OK with GSoC, we need to know about your other commitments, and also need to be certain that the code you write can b 2013-05-03T11:30:31 e contributed by you to RTEMS without a claim by your employer or University. If you are in doubt, you may work with us to try to figure out the situation. 2013-05-03T11:30:33 Thanks, 2013-05-03T11:30:35 *** sreeharshakondur has joined #rtems 2013-05-03T11:30:35 Gedare 2013-05-03T11:30:59 DPS? 2013-05-03T11:31:05 welcome back sreeharshakondur 2013-05-03T11:31:16 Sorry i lost the connection, thanks 2013-05-03T11:31:21 what was the last thing you read? 2013-05-03T11:31:39 code being easy to real 2013-05-03T11:31:49 read** and examine SMP scheduling 2013-05-03T11:31:58 DPS = Deterministic Priority Scheduler .. the default one 2013-05-03T11:32:20 okay 2013-05-03T11:32:32 [11:28] <@DrJoel> No. That would be a nice addition. It only has scenarios for deterministic priority and simple smp. 2013-05-03T11:32:32 [11:28] <@DrJoel> I **recall** that Simple SMP on a single core would run the DPS scenarios with the same output. 2013-05-03T11:32:32 [11:29] <@DrJoel> Adding enough to do the same for Simple Priority (uniprocessor) would be good as well. 2013-05-03T11:32:32 [11:30] <@DrJoel> should only be a matter of getting periods in the command language. 2013-05-03T11:32:36 ah. ok. DrJoel what do you think of my draft question? 2013-05-03T11:33:08 i think it would be good for you not to worry too much about the Thread Set refactoring 2013-05-03T11:33:08 but instead to focus on the SMP scheduling 2013-05-03T11:33:08 fwiw the code is fairly well documented and intentionally simple minded in implementation. Easy to read before clever. The second SMP scheduling algorithm can be clever 2013-05-03T11:33:08 and the path to testing your scheduler, which should include schedsim, and i think pc386/qemu also supports running RTEMS with SMP 2013-05-03T11:33:08 i would like to see a "real" real-time multiprocessor scheduling algorithm. 2013-05-03T11:33:09 I think that's a good question. I am glad you included employment 2013-05-03T11:33:10 I think Global EDF is plausible. it depends on a global EDF queue with a lock, and per-core queues 2013-05-03T11:33:15 Ok 2013-05-03T11:33:27 I'm just going to put it on every student's proposal. 2013-05-03T11:33:38 We should remember to include it in our requirements next year. 2013-05-03T11:33:42 Can global edf and edf share code? 2013-05-03T11:33:53 You can add it to the proposal template 2013-05-03T11:34:01 that did not get updated for this year ;) 2013-05-03T11:34:07 + \lazy 2013-05-03T11:35:01 I had a few questions in mind to add.. but they were already there when I checked :) 2013-05-03T11:35:11 I did update it to fix the times and phases to discuss 2013-05-03T11:35:21 oh ok..some of the students got the old copy somehow. 2013-05-03T11:35:34 did the deadline pass for submitting? 2013-05-03T11:35:37 so currently the simple smp scheduler only mimics the DPS? am i right? 2013-05-03T11:35:50 Probably started before I updated it 2013-05-03T11:35:58 It is 3pm EST I thin gedare 2013-05-03T11:36:24 ok. sreeharshakondur I don't think it mimics the simple scheduler..just a bunch of priority-sorted linked lists, but DrJoel knows better 2013-05-03T11:36:26 sreeharshakondur: yes. Extends priority scheduling to multiple cores. 2013-05-03T11:36:36 One priority sorted linked list 2013-05-03T11:36:39 hm. what data structure it uses? 2013-05-03T11:36:51 ok, so it is like the simples cheduler 2013-05-03T11:37:01 Same as Simple Priority - a sorted chain 2013-05-03T11:37:01 basically a "global priority" scheduler 2013-05-03T11:38:05 yes.. and very simple one.. goal was functional SMP -- not optimal SMP. algorithm is correct not efficient 2013-05-03T11:38:39 then guess more functionality can be added to the smp scheduler in terms of the scheduling policies? 2013-05-03T11:38:45 gedare: also takes into account no-preempt, time on core (i.e. FIFO per priority) 2013-05-03T11:39:13 sreeharshakondur, you should implement an efficient SMP scheduler. a single global priority queue is bad for scalability/performance 2013-05-03T11:39:30 sreeharshakondur: a better data structure and algorithm to get the same behavior is desirable 2013-05-03T11:39:42 okay, 2013-05-03T11:39:55 i will try to enhance it 2013-05-03T11:40:08 one of the best places to get info is probably http://www.litmus-rt.org/ 2013-05-03T11:40:08 The Simple SMP has O(# tasks) to unblock (insert on ready chain) 2013-05-03T11:40:15 they implement a lot of smp schedulers in linux 2013-05-03T11:40:18 O(constant) to unblock (remove from chain) 2013-05-03T11:40:38 just note you cannot use their code because of licensing issues 2013-05-03T11:40:44 O(n-cores) or O(n-cores^2) to schedule 2013-05-03T11:41:23 schedule == assign threads to cores when a change in the ready thread set occurs 2013-05-03T11:45:02 DrJoel, if you stick around that Friday night, you can come to attend my graduation if you like. 2013-05-03T11:45:32 and a reminder that i will be needing to leave early'ish on thursday and friday 2013-05-03T11:46:45 *** Hesham has joined #rtems 2013-05-03T11:47:04 I know.. I don't think it is a problem for you to leave. 2013-05-03T11:47:23 I still have to make travel plans.. I will let you know 2013-05-03T11:47:49 k 2013-05-03T11:50:23 howdy Hesham feel good about your proposal? 2013-05-03T11:50:54 Not sure :) 2013-05-03T11:51:39 I am writing an e-mail regarding commitment now, or should I reply through in melange ? 2013-05-03T11:51:46 either is fine 2013-05-03T11:52:18 we just need to know that the code you write is ok for us to use, and want to know about other commitments from a project management point of view 2013-05-03T11:52:42 having other jobs is not a blocker, but not being able to commit code would be a problem 2013-05-03T11:53:20 have you had a chance yet to order a raspberry pi? 2013-05-03T11:53:31 better through melange .. more private and recorded 2013-05-03T11:53:37 only mentors will see it here 2013-05-03T11:53:39 there 2013-05-03T11:53:43 less private, but recorded ;) 2013-05-03T11:54:05 yes.. irc and email are very public and recorded :) 2013-05-03T11:54:26 well, direct personal email is relatively more private than melange. 2013-05-03T11:54:30 any ways 2013-05-03T11:54:31 I have sent you an e-mail, do not know whether it would be a problem to commit code or not. I can ask about that 2013-05-03T11:54:36 Ok 2013-05-03T11:54:39 please do 2013-05-03T11:55:24 Ok I will copy and submit the message to melange 2013-05-03T11:56:50 About Raspberry, I would go to an electronic store in Cairo and hope to get one next week 2013-05-03T11:57:00 I think It will be Model A 2013-05-03T11:57:04 cool 2013-05-03T11:57:27 i don't know which one the BSP was developed for 2013-05-03T11:57:50 oh, did you see the emails recently about CP15 initialization problems? 2013-05-03T11:58:03 Nope 2013-05-03T11:58:09 a couple of platforms, including raspberry pi, were having some bugs due to a misconfigured bit set in CP15 to control unaligned memory access 2013-05-03T11:58:22 we should take this into consideration for low-level arm MM support 2013-05-03T11:58:46 i'll forward the relevant emails to you 2013-05-03T11:59:18 that's Ok, I will check these e-mails 2013-05-03T11:59:28 ok 2013-05-03T11:59:41 it was on rtems-users, search for Raspberry 2013-05-03T12:00:24 and "unaligned problem when porting rtems for arm" 2013-05-03T12:00:27 ok 2013-05-03T12:00:52 I got it 2013-05-03T12:02:11 DrJoel, give me permission to edit the proposal template ;) 2013-05-03T12:02:41 sorry.. on phone to amazon 2013-05-03T12:02:53 Michele's book went down in price 10% .. amazon put it on sale. :) 2013-05-03T12:04:25 give me the link.. having trouble finding it 2013-05-03T12:05:04 https://docs.google.com/document/d/1d_-wtMV_BH4WzkMoysTqFS8BtZhh3i7uCzoAWLqVSe4/edit?pli=1 2013-05-03T12:05:37 done 2013-05-03T12:07:58 added a COI section 2013-05-03T12:10:18 :) 2013-05-03T12:10:39 We dealt with one in the past. Student got a job at Microsoft and needed to start about 2 weeks before deadline. 2013-05-03T12:10:59 gedare: I think Alan worked on Model A, anyway I asked him about the Model he used on his blog. 2013-05-03T12:11:30 super 2013-05-03T12:11:49 yeah. as long as we are aware of COI that is the important part so we can deal with it 2013-05-03T12:12:00 I did some minor updates on my proposal and I think it's ready to go. :) 2013-05-03T12:12:25 phipse, great! 2013-05-03T12:12:55 depending on whether there are fundamental gaps or questions, we may be in contact 2013-05-03T12:13:06 otherwise it is quite usual not to hear anything until Google announces the accepted list. 2013-05-03T12:13:12 I hope I can deliver that picture I have in my head :) 2013-05-03T12:13:30 but you should feel free to continue to make improvements and chip away at your project.. 2013-05-03T12:13:40 just because GSOC doesn't "start" doesn't mean you cannot start. 2013-05-03T12:13:46 the usual unpleasant silence in the application process to anything ;) 2013-05-03T12:13:53 yep 2013-05-03T12:14:03 I expect we will be asking questions on multiple proposals. phipse as you get updates or figures, just post them to Melange and/or the mailing list. 2013-05-03T12:14:16 or your project wiki 2013-05-03T12:14:24 which will be the "living document" of your project. 2013-05-03T12:14:31 I will choose the wiki ;) 2013-05-03T12:14:58 just poke the ML or melange so we know something happened on the wiki :) 2013-05-03T12:15:03 And probably a journal in the repo ;) 2013-05-03T12:15:36 about the repo, it should be just a simple code.google.com repo? Or are there special gsoc repos? 2013-05-03T12:15:44 blogging isn't bad either.. encouraged 2013-05-03T12:16:00 phipse, actually i don't think you need to be on code.google.com 2013-05-03T12:16:09 github code.google.com, whereever.. no special rules except at end to submit proof and get paid 2013-05-03T12:16:13 in fact, if/when we get our github updater working again, it would be better on github 2013-05-03T12:16:22 i think the proof is submitted through melange 2013-05-03T12:16:53 i find it MUCH easier to find student projects if they fork github/RTEMS/rtems 2013-05-03T12:16:55 okay, then I will just fork your github repo and open a branch for my changes 2013-05-03T12:17:00 you are right.. that changed.. can be just a tarball of patches submitted... anything capturing the work 2013-05-03T12:17:01 danke 2013-05-03T12:17:09 the github isn't updating properly right now. :( 2013-05-03T12:17:19 hopefully it will be by the time coding starts! 2013-05-03T12:17:41 DrJoel, got an RTEMS build script anywhere? I don't have mine on this computer :) 2013-05-03T12:18:22 i should really have it posted somewhere, hm 2013-05-03T12:20:08 ugh i have to build a toolchain too. maybe time to try RSB again 2013-05-03T12:21:51 thanks 2013-05-03T12:21:52 see attached for lazy scripts 2013-05-03T12:22:16 about the same as my lazy scripts ;0 2013-05-03T12:22:42 Do you have a "./j" and "./j1" 2013-05-03T12:22:57 hah no 2013-05-03T12:32:08 *** sreeharshakondur has quit IRC (Ping timeout: 245 seconds) 2013-05-03T12:34:03 *** S_Somani1 has joined #rtems 2013-05-03T12:35:55 *** S_Somani has quit IRC (Ping timeout: 264 seconds) 2013-05-03T12:36:33 *** S_Somani1 is now known as S_Somani 2013-05-03T13:25:26 *** peerst has quit IRC (Remote host closed the connection) 2013-05-03T13:28:37 * DrJoel thinks it is about 30 minutes to student app deadline 2013-05-03T13:29:21 oh that's tonight? sigh it's been a crazy week 2013-05-03T13:55:41 *** prateekt has joined #rtems 2013-05-03T14:02:51 ah, finally, deadline over, one thing less to worry about :D 2013-05-03T14:06:32 dramatic chipmunk face... OR IS THERE? :) 2013-05-03T14:08:16 drama? excitement? you can't tell it apart. 2013-05-03T14:08:26 :) 2013-05-03T14:08:57 http://upload.wikimedia.org/wikipedia/en/b/b6/Dramatic_Chipmunk.png 2013-05-03T14:10:39 ;) 2013-05-03T14:12:28 error: shell cmd failed: /bin/sh -ex /home/gedare/work/rtems/rtems-source-builder/rtems/build/arm-rtems4.11-gdb-7.6-1-7.6-1/doit 2013-05-03T14:12:28 Build FAILED 2013-05-03T14:12:30 *** prateekt has quit IRC (Quit: Page closed) 2013-05-03T14:12:30 onoes. 2013-05-03T14:12:47 configure: error: no termcap library found 2013-05-03T14:12:47 make[1]: *** [configure-gdb] Error 1 2013-05-03T14:17:56 what OS? 2013-05-03T14:18:22 http://rtems.org/ftp/pub/rtems/people/chrisj/source-builder/source-builder.html covers most cases 2013-05-03T14:18:41 yum install autoconf automake binutils gcc gcc-c++ gdb make patch \ 2013-05-03T14:18:41 bison flex xz unzip ncurses-devel texinfo zlib-devel python-devel git 2013-05-03T14:21:13 ubuntu. 2013-05-03T14:21:27 the apt get is just below that in the manual :) 2013-05-03T14:21:46 aha ncurses-dev 2013-05-03T14:23:03 i want to try out the new realview bsp 2013-05-03T14:23:13 it may prove useful to me 2013-05-03T14:23:28 looks interesting :) 2013-05-03T14:23:30 but, I won't have time to finish building today 2013-05-03T14:23:51 i want to run it on http://www.m5sim.org/Main_Page 2013-05-03T14:23:55 RSB needs help for building simulators. I am trying to get a handle on skyeye status in the background 2013-05-03T14:24:09 it can boot Android Linux configured for the realview pbx a9 2013-05-03T14:24:17 so hopefully I can get an rtems to boot 2013-05-03T14:24:23 Should be able to 2013-05-03T14:24:31 And that niagara looks interesting .. hint 2013-05-03T14:24:39 oh ahhh 2013-05-03T14:24:47 * gedare hides 2013-05-03T14:25:02 actually, i used it to run niagara bsp from master yesterday 2013-05-03T14:25:16 i have instructions somewhere on my blog, and maybe some day i can bother to write a sim-script 2013-05-03T14:25:27 i'm not sure ... there is a lot of setup for that simulator 2013-05-03T14:25:39 but, I can set it up on a test system if you need some day 2013-05-03T14:25:45 e.g. for continuous integration, hint hint. 2013-05-03T14:26:02 *** arvind_khadri has joined #rtems 2013-05-03T14:26:17 getting a sim-script helps with the continuous integration :) 2013-05-03T14:26:27 well.. about time for me to call it a day 2013-05-03T14:26:46 tickets to Addams Family musical tomorrow night... touring show :) 2013-05-03T14:26:57 and we apparently have a lot of student proposals to read 2013-05-03T14:28:09 hmmm... only about 2/3 of last year.. no mad rush at the end 2013-05-03T14:29:23 I know a good way, to lessen the work. :) 2013-05-03T14:29:39 :) 2013-05-03T14:29:47 have funsies 2013-05-03T14:30:34 funsies? 2013-05-03T14:31:11 yep. 2013-05-03T14:31:17 what's that? 2013-05-03T14:31:21 the noun form of fun. (slang) 2013-05-03T14:31:21 ;) 2013-05-03T14:31:38 well, i suppose "fun" is an idea, which is a noun, of sorts 2013-05-03T14:31:54 ah I get the drift, thanks 2013-05-03T14:32:01 np..american slang 2013-05-03T14:32:17 usually of the form.... I did for funsies. 2013-05-03T14:32:18 ;) 2013-05-03T14:32:38 well, that's one good way, but I have an easier one :P 2013-05-03T14:33:18 choose me, one less to read :D 2013-05-03T14:35:40 or, discard you, one less to read? 2013-05-03T14:36:03 works both ways it seems! 2013-05-03T14:36:29 all right, i'm off too.. enjoy the weekend! 2013-05-03T14:36:31 damn, you're right 2013-05-03T14:36:32 :D 2013-05-03T14:36:39 thanks, you too 2013-05-03T14:36:46 *** gedare has quit IRC (Quit: Leaving) 2013-05-03T14:43:48 *** Hesham has left #rtems 2013-05-03T14:53:23 *** kkthecoder has joined #rtems 2013-05-03T15:00:06 *** monstr_ has quit IRC (Read error: Operation timed out) 2013-05-03T15:10:50 *** DrJoel has quit IRC (Quit: When the chips are down, well, the buffalo is empty) 2013-05-03T15:24:07 *** phipse has left #rtems 2013-05-03T15:35:25 *** peerst has joined #rtems 2013-05-03T16:43:58 *** gedare has joined #rtems 2013-05-03T17:00:05 *** S_Somani has left #rtems 2013-05-03T17:30:12 *** gedare has quit IRC (Ping timeout: 264 seconds) 2013-05-03T17:59:44 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-03T18:00:26 *** kiwichris has joined #rtems 2013-05-03T20:10:52 *** freenix has joined #rtems 2013-05-03T20:44:30 *** SYCrane has joined #rtems 2013-05-03T21:27:14 *** kkthecoder has quit IRC (Quit: Leaving) 2013-05-03T21:27:33 *** kkthecoder has joined #rtems 2013-05-03T22:11:50 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-03T22:15:49 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2013-05-03T23:26:44 *** rtdin has joined #rtems 2013-05-03T23:33:23 *** S_Somani has joined #rtems 2013-05-03T23:33:46 *** S_Somani has left #rtems 2013-05-04T00:10:24 *** kkthecoder has quit IRC (Ping timeout: 264 seconds) 2013-05-04T00:53:23 *** kiwichris has joined #rtems 2013-05-04T03:29:35 *** S_Somani has joined #rtems 2013-05-04T03:36:08 *** rtdin has quit IRC (Ping timeout: 268 seconds) 2013-05-04T03:36:38 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-04T03:58:51 *** S_Somani has joined #rtems 2013-05-04T04:38:24 *** S_Somani1 has joined #rtems 2013-05-04T04:38:25 *** S_Somani has quit IRC (Ping timeout: 268 seconds) 2013-05-04T05:49:43 *** weiY has quit IRC (Ping timeout: 264 seconds) 2013-05-04T07:42:12 *** S_Somani1 has quit IRC (Quit: Leaving.) 2013-05-04T07:53:03 *** freenix has quit IRC (Quit: Leaving) 2013-05-04T07:55:58 *** arvind_khadri has joined #rtems 2013-05-04T07:57:58 *** S_Somani has joined #rtems 2013-05-04T08:01:39 *** S_Somani has left #rtems 2013-05-04T08:34:03 *** DrJoel has joined #rtems 2013-05-04T08:34:03 *** DrJoel has joined #rtems 2013-05-04T08:34:03 *** ChanServ sets mode: +o DrJoel 2013-05-04T08:34:23 verm__: Thanks.. github looks like it mirrored at least once.. :) 2013-05-04T08:34:48 The next time a commit happens, I will try to keep an eye on it for an update 2013-05-04T08:43:24 *** arvind_khadri has quit IRC (Ping timeout: 264 seconds) 2013-05-04T08:45:48 *** arvind_khadri has joined #rtems 2013-05-04T08:46:39 *** arvind_khadri has joined #rtems 2013-05-04T08:47:56 *** arvind_khadri has joined #rtems 2013-05-04T08:48:57 *** arvind_khadri has joined #rtems 2013-05-04T08:49:23 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2013-05-04T08:52:05 *** arvind_khadri has joined #rtems 2013-05-04T09:04:23 *** arvind_khadri has quit IRC (Ping timeout: 256 seconds) 2013-05-04T09:19:53 *** SYCrane has quit IRC (Read error: Connection reset by peer) 2013-05-04T09:52:17 *** weiY has joined #rtems 2013-05-04T10:20:01 *** gedare has joined #rtems 2013-05-04T10:20:51 *** gedare has joined #rtems 2013-05-04T10:21:04 DrJoel, hi 2013-05-04T10:23:43 hey .. 2013-05-04T10:23:50 got time got private chat? 2013-05-04T10:24:14 shouldn't take too long 2013-05-04T10:24:28 yeah i'm lurking and watching A play 2013-05-04T10:25:03 :) Today is Jess' birthday and Michele is making a carrot cake for her.. May the Fourth be with you! 2013-05-04T10:28:37 *** weiY has quit IRC (Ping timeout: 248 seconds) 2013-05-04T10:56:08 gedare: I woke up thinking about time passage and RTEMS in a virtualized environment.. especially the difference between in an ARINC 653 partition and a pure virtualized environment 2013-05-04T10:56:27 ok 2013-05-04T10:56:35 You have at least processor usage, wall time, time-slicing, and relative time (delays) 2013-05-04T10:57:02 In ARINC 653, the partitions are scheduled and there is a "true" wall time across the system 2013-05-04T10:57:16 In a regular hypervisor, I don't know that you would have that. 2013-05-04T10:57:24 probably not 2013-05-04T10:57:37 maybe one with real-time capabilities 2013-05-04T10:57:51 Your cpu usage and timeslicing accounting shouldn't increment in ARINC 653 while your partition isn't executing. 2013-05-04T10:58:08 But delays probably should still be moving along. 2013-05-04T10:58:35 I recall issues on line about VMWare and Virtualbox having jumpy time.. this may be the issue 2013-05-04T10:59:13 delays are usually like "every 100 milliseconds touch a sensor" and that is hypervisor time -- not virtualized time 2013-05-04T10:59:49 Are there more types of time passage? 2013-05-04T11:03:46 it would be good to try to classify the different times in such a system 2013-05-04T11:04:01 i wonder if there are theoretical results about these systems. i should take a look some time 2013-05-04T11:04:22 *** DrJoel has quit IRC (Quit: Life without danger is a waste of oxygen) 2013-05-04T11:04:45 *** gedare has quit IRC (Quit: Ex-Chat) 2013-05-04T11:40:09 *** S_Somani has joined #rtems 2013-05-04T12:05:04 *** DrJoel has joined #rtems 2013-05-04T12:05:05 *** DrJoel has joined #rtems 2013-05-04T12:05:05 *** ChanServ sets mode: +o DrJoel 2013-05-04T12:06:10 Welcome to RTEMS (http://www.rtems.org). Please feel free to ask questions and, ifyoucan, 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 *** MENTORS WANTED!!! PROPOSALS UNDER REVIEW!!! *** 2013-05-04T12:06:22 *** DrJoel has quit IRC (Client Quit) 2013-05-04T12:30:17 *** rtdin has joined #rtems 2013-05-04T13:28:52 *** rtdin has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-05-04T14:47:39 *** S_Somani has left #rtems 2013-05-04T15:28:14 *** ppisa has joined #rtems 2013-05-04T16:19:20 *** DrJoel has joined #rtems 2013-05-04T16:19:21 *** DrJoel has joined #rtems 2013-05-04T16:19:21 *** ChanServ sets mode: +o DrJoel 2013-05-04T16:19:39 ppisa sorry.. my daughter's was here for birthday cake.. 24 2013-05-04T16:29:17 no problem, I excuse for silence. 2013-05-04T16:29:48 I am sorry for not being available quicker. I know it is late for you on a Saturday night. :) 2013-05-04T16:30:16 Thanks for the clarifications on the CANBus project. I just want to feel comfortable about the entire thing long term 2013-05-04T16:34:28 As for CAN API, I am not sure what is the best option. From user perspective, SocketCAN has nice API. But it is wuite complex and generally GPL only. There existed variant for Xenomai, which is not so much depandant on Linux, uses some small/separate network buffers implementation. 2013-05-04T16:35:12 I will admit to being largely ignorant in this domain. :( 2013-05-04T16:35:49 I just want something that is attractive to someone with RTEMS using CAN.. 2013-05-04T16:37:03 As for LinCAN, it is tested on Linux, in original RT-Linux environment and system-less environment. It is based on the concept of FIFOs. I think, that they are quite well implemented and portable. On the other hand, there is an option to use RTEMS API message queues. 2013-05-04T16:38:13 If it runs on RTEMS now, what would the GSOC project do? 2013-05-04T16:39:02 The CAN has two distinct use cases and each map well to other HW chips and SW concept. One is the build of the mostly controlled/configured device with limited number of sent and received message IDs. 2013-05-04T16:40:02 The other use is monitor, configuration system or coordinator, top level controller. This needs to provide some infrastructure for FIFOs and messages filtering in SW. 2013-05-04T16:40:45 So line 1553 bus in some respects. 2013-05-04T16:40:47 As for LinCAN, it is not adapted for RTEMS yet. It was used only in system-less/IRQ driven environment. 2013-05-04T16:40:49 like 1553 2013-05-04T16:40:59 ahhh... ok so that's the project. 2013-05-04T16:44:01 As for RTEMS, there are already some CAN drivers included in the tree with different API. They are mostly blocking, without FIFOs as I have looked at them in the past. But for some applications it is quite good solution or even may it be better if only simple "slave" device (device with static configuration or at least messages ID assigment plan) functionality is required. 2013-05-04T16:44:59 There are many higher level layers for CAN -- CANopen, DeviceNet, OSEK/AUTOSAR based ID mapping for automotive etc. 2013-05-04T16:45:45 And it sounds like these higher level layers can be put on top of LinCAN. Right? 2013-05-04T16:48:01 As for higher layers, for static IDs mapping CANFestival provides CANopen implementation and more tools and is quite complete. It is GNU LGPL licensed. I have seen some of authors and we can speak with them about RTEMS compatibility. 2013-05-04T16:48:22 ok.. that's a nice long term approach 2013-05-04T16:48:57 if we can have a complete solution that would be awesome 2013-05-04T16:48:57 LinCAN has been tested with CANfestival which includes minimal/basic LinCAN support in the mainline. 2013-05-04T16:50:25 That gives us a technical path. The licensing is (as always) our concern. 2013-05-04T16:51:18 Our (OrtCAN) upper layer code for CANopen is quite configurable, allows to build device dictionary structure at runtime, but it is not so complete. It is compatible with LinCAN and we have some working code to provide access to CAN and CANopen over Ethernet. 2013-05-04T16:51:40 fyi the long term plan is for Chris' RTEMS Source Builder to move beyond building the tools and rtems, to include building add-on packages. RSB already can fetch tar balls, cvs, svn, git and patch as needed. THen build. So add-on layers are hopefully becoming easier to deal with. 2013-05-04T16:52:25 you have satisfied my concerns about the CAN code itself. Now we worry about the student and the viability of his proposal. 2013-05-04T16:53:23 I would like to replace the current graphics toolkit with a set of entries in the RSB build system when the time comes. Chris has built OpenSSL for RTEMS. We could add NTP, and these packages. Fetch, patch and build 2013-05-04T16:54:22 As for OrtCAN higher layers, they have been developed at CTU and I have proposed and kept licensing with RTEMS in mind. So license approved by our group head and all authors is GPL with linking exception, see 2013-05-04T16:54:22 http://ortcan.git.sourceforge.net/git/gitweb.cgi?p=ortcan/ortcan-vca;a=blob;f=libvca/vca_pdo.c 2013-05-04T16:55:45 Marvelous and very explicitly clear! Completely compatible with RTEMS licensing and obviously so. 2013-05-04T16:56:03 Are you up late to watch the Kentucky Derby? :) 2013-05-04T16:58:50 As for LinCAN, licensing is more complex. It has some traces from the past. I didnot introduce yet another API, so I started from original LDDK code. But it has been rewritten completely except some lines in the obscure old cards support. But some traces are to GPL origin about 15 years old. But from side and all later ubdates the license is 2013-05-04T16:58:51 http://ortcan.git.sourceforge.net/git/gitweb.cgi?p=ortcan/lincan;a=blob;f=lincan/src/can_queue.c 2013-05-04T16:59:21 No I do not watch derby, I usually end with my work late in the night. 2013-05-04T16:59:34 I think this is my first time. My wife wanted to watch it. :) 2013-05-04T16:59:59 can_queue.c has the same license. 2013-05-04T17:00:09 Knowing you guys the target side requirements are pretty light 2013-05-04T17:00:47 If you want to watch now then I can be available tommorou after 19:00 or 20:00 CET till 2:00 CET. 2013-05-04T17:01:33 nah.. I am ok.. it is all pomp and a 60 second race. :) 2013-05-04T17:01:49 OK 2013-05-04T17:01:54 do you think there is a summer of work? 2013-05-04T17:03:58 As for the CAN in Qemu, I have not found time to dig if somewan already has it for some chip. We have some experience with PCI data acquisition card implementation (ADC, DAC, IN, OUT) 2013-05-04T17:03:58 http://rtime.felk.cvut.cz/hw/index.php/Humusoft_MF6xx 2013-05-04T17:03:58 http://lwn.net/images/conf/rtlws-2011/proc/Pisa.pdf 2013-05-04T17:07:13 The tricky decission is which chip to try to emulate. I would expect that old SJA1000 would be most usable. It is single object FIFO style chip, maps weel to SocketCAN or other driver API on the host side. But it has quite complicated registers set. Complicates emulation and use with 8-bit only registers and more unnecessary modes for compatability with even older chips. 2013-05-04T17:07:53 s/maps weel/map well/ 2013-05-04T17:08:21 :) i trust you to make the right decisions. 2013-05-04T17:08:43 How did you drive input into the multi IO card? 2013-05-04T17:09:06 I do not trust myself so much. As one knows more he/she knows that there is even more to not know. 2013-05-04T17:09:51 I started building a multi I/O library which is in the RTEMS git. Would like to have some simulated target to test against and let others play with it. 2013-05-04T17:10:20 Well knowing that you do not know keeps you from walking stupidly into an area you need to learn about. :) 2013-05-04T17:10:23 The card has on the host side of QEMU TCP/IP port where written/output values are sent and which accepts changes from external application (is simple GUI tool). 2013-05-04T17:12:42 Sounds useful. I need to write up a small project around that. As foundation to work more on rtems-multiio. It includes application API and even commands in shell to get status, tinker output, drive patterns, etc 2013-05-04T17:13:13 All is in the GIT http://rtime.felk.cvut.cz/gitweb/mf6xx.git , unfortuantelly QEMu evolves very quickly so there is required update for actual version again. 2013-05-04T17:13:49 Yes. Qemu is hard to get in sync with and get code into because of that. They also are Xen focused and the true simulators lose sometimes 2013-05-04T17:14:49 I think Skyeye is more focused on embedded but it needs some help. They rearchitected and broke some stuff. I have been trying to get a handle on what broke when and help them track down fixies. 2013-05-04T17:14:51 As for the size of CAN project, I have feeling that doing both -- drivers and QEMU in single GSoC is quite huge amount of work. 2013-05-04T17:15:04 ok. then we will be happy if he gets it done. ;) 2013-05-04T17:16:41 OK, I try to think about it little more. As for LinCAN, if chosen, it needs some cleanup to make it better for porting. I would be happy if we can restructure it such way, that RTEMS support is included in the main repo. 2013-05-04T17:17:42 that is very desirable. We do not want a fork or patch. ;) 2013-05-04T17:19:02 OK, that is probably all for CAN now. Should we switch to RTEMS graphics? 2013-05-04T17:22:28 sure. :) 2013-05-04T17:22:39 What do we need to do to merge it. :) 2013-05-04T17:24:27 The first are users, people willing to test newer version. Nor older nor newer is perfect. 2013-05-04T17:25:41 I don't know how many testers you will get. If myself (and possibly Cynthia if she is willing) can duplicate the results, then we should be happy 2013-05-04T17:26:20 I expect that from longer perspective on i386 linead mode configured through VESA BIOS (aghr) is the must. It requires to extend RTEMS global descriptors table for i386. 2013-05-04T17:27:47 There is really no problem extending the global descriptors. There is only 1 BSP and it is not memory constrained. 2013-05-04T17:27:55 Yes at least you and someone else would be great. Sebastian Huber have stated in some e-mail that he expects that he need RTEMS graphics for some project. 2013-05-04T17:29:23 Yes descriptors extension is not problem. It could be static still with some small sane limit. But some finction to find empty slot and setup it according to some structure/parameters should be implemented. 2013-05-04T17:29:48 s/finction/function/ 2013-05-04T17:30:09 Email the users list with a plan and request for testers. Hopefully it will be more than who we think is possible 2013-05-04T17:32:41 In the fact, I have tried to initiate testing with some of my previsous e-mails. But for PC target there is the problem with keyboard translation I have analyzed and documented which is lettle unfortunate for testers. Only subset of keys is working. 2013-05-04T17:35:03 My E-mail from 2013-03-20 11:48 subject "Re: Back to RTEMS Graphics Toolkit and i386 MWin keyboard problems" 2013-05-04T17:36:45 By the way, please, contact GMANE archive users to add rtems-devel list to the database 2013-05-04T17:36:45 http://dir.gmane.org/search.php?match=rtems 2013-05-04T17:36:45 I consider GMANE archive as the best and RTEMS has only rtems-user registered. 2013-05-04T17:36:54 Well.. we can only test and try to fix. Is it a parsing problem? 2013-05-04T17:37:09 And I don't know who registered rtems-users :) 2013-05-04T17:38:20 I do not know too. It can be requested by everyone but I think it is better when it is from list owner. 2013-05-04T17:39:20 But GMANE is nice and even Linux Weekly News and many Linux developers use its therads pointers to reference in other articles. 2013-05-04T17:40:30 Done for rtems-devel. Should we add rtems-vc and rtems-toolstesting? 2013-05-04T17:42:03 As for the keyboard, there has been redone PC console stuff in RTEMS some time ago and MicroWindows and UID was not updated. On the other hand the required translation is PC cruft specific and if we define PC style mapping as the must then it complicates other targets. 2013-05-04T17:42:14 Thanks for registering. 2013-05-04T17:43:12 Jennifer and I redid the console. As we test the update, she should be able to look into fixing it 2013-05-04T17:44:52 In the fact it is not bug in the console. It is incompatibility between RTEMS and MicroWindows side. 2013-05-04T17:45:53 ahh.. well I will try to remember to put it on Jennifer's list to follow up with you. 2013-05-04T17:46:05 But we might has well move along with merging it and then debug it 2013-05-04T17:46:48 May it be, I should prepare and fill PR, if it is better for coordination. 2013-05-04T17:48:04 please do. Assign it to Jennifer and put the email link in it. 2013-05-04T17:48:54 Jennifer and I reworked the console as part of supporting Com3/4 for a user who actually had amost 40 serial ports. Multiple 8 UART PCI boards and some z8530 boards. 2013-05-04T17:49:01 All with libchip table and probes :) 2013-05-04T17:49:11 dynamically installed 2013-05-04T17:49:54 well I better start heading off. We have tickets to a touring Broadway production of Addams Family the musical in a few hours :) 2013-05-04T17:49:58 All pending changes from Alex's work are targetted to the separate/RTEMS graphhic toolkit git. What we need to decide if actual code should be put to the branch and Alex's code be a master or vice versa. Or if the patches should be yet more cleaned up. 2013-05-04T17:50:33 Our convention says that the "desired" code is the master. But we need to review his changes versus upstream on everything. 2013-05-04T17:51:03 We should baseline import the original update tarball. Then patch it. 2013-05-04T17:51:28 This is short term to me until we can add building it to the RTEMS Source Builder. Fetch from upstream, minimal patch, then build 2013-05-04T17:52:05 OK, I try to check that it match the practice. Have a nice evening and I hope to find more time windows to sort the things with you. 2013-05-04T17:52:05 I should talk to Chris to see if biting that off now would be easier. And then "merge" this via the RTEMS Source Builder 2013-05-04T17:52:15 night.. thanks for the cat 2013-05-04T17:52:21 chat != cat lol 2013-05-04T17:52:25 *** DrJoel has quit IRC (Quit: Light travels faster then sound, which is why some people appear bright, until you hear them speak) 2013-05-04T17:52:48 *** ppisa has left #rtems ("Kopete 0.12.7 : http://kopete.kde.org") 2013-05-04T18:09:02 *** ppisa has joined #rtems 2013-05-04T18:09:14 *** ppisa has left #rtems ("Kopete 0.12.7 : http://kopete.kde.org") 2013-05-04T18:38:29 *** gedare has joined #rtems 2013-05-04T18:43:12 *** gedare has quit IRC (Client Quit) 2013-05-04T22:04:43 *** weiY has joined #rtems 2013-05-04T22:53:25 *** freenix has joined #rtems 2013-05-04T23:23:45 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-04T23:24:30 *** kiwichris has joined #rtems 2013-05-05T00:03:56 *** freenix has quit IRC (Quit: Leaving) 2013-05-05T00:11:07 *** rtdin has joined #rtems 2013-05-05T01:16:25 *** S_Somani has joined #rtems 2013-05-05T01:16:50 *** S_Somani has left #rtems 2013-05-05T02:22:20 *** monstr_ has joined #rtems 2013-05-05T02:27:01 *** monstr_ has quit IRC (Ping timeout: 248 seconds) 2013-05-05T02:55:07 *** weiY has quit IRC (Read error: Connection reset by peer) 2013-05-05T02:57:01 *** S_Somani has joined #rtems 2013-05-05T03:37:52 *** rtdin has quit IRC (Ping timeout: 260 seconds) 2013-05-05T03:39:00 *** rtdin has joined #rtems 2013-05-05T03:43:55 *** rtdin has quit IRC (Ping timeout: 260 seconds) 2013-05-05T03:49:10 *** mgahlot|away has joined #rtems 2013-05-05T03:56:59 *** mgahlot|away is now known as mgahlot|off 2013-05-05T05:15:07 *** Fallenou_ has joined #rtems 2013-05-05T05:17:29 *** Fallenou has quit IRC (Remote host closed the connection) 2013-05-05T05:34:23 *** mgahlot|off has quit IRC (*.net *.split) 2013-05-05T05:34:24 *** rokka_ has quit IRC (*.net *.split) 2013-05-05T05:34:25 *** antgreen has quit IRC (*.net *.split) 2013-05-05T05:34:28 *** peerst has quit IRC (*.net *.split) 2013-05-05T05:34:31 *** noselasd has quit IRC (*.net *.split) 2013-05-05T05:34:32 *** budrys has quit IRC (*.net *.split) 2013-05-05T05:34:34 *** ChanServ has quit IRC (*.net *.split) 2013-05-05T05:34:35 *** dhananjay has quit IRC (*.net *.split) 2013-05-05T05:34:36 *** verm__ has quit IRC (*.net *.split) 2013-05-05T05:41:22 *** mgahlot|off has joined #rtems 2013-05-05T05:41:22 *** peerst has joined #rtems 2013-05-05T05:41:22 *** antgreen has joined #rtems 2013-05-05T05:41:22 *** dhananjay has joined #rtems 2013-05-05T05:41:22 *** budrys has joined #rtems 2013-05-05T05:41:22 *** rokka_ has joined #rtems 2013-05-05T05:41:22 *** verm__ has joined #rtems 2013-05-05T05:41:22 *** noselasd has joined #rtems 2013-05-05T05:41:22 *** ChanServ has joined #rtems 2013-05-05T05:41:22 *** barjavel.freenode.net sets mode: +o ChanServ 2013-05-05T05:47:00 *** freenix has joined #rtems 2013-05-05T06:58:31 *** rtdin has joined #rtems 2013-05-05T07:43:20 *** rtdin has quit IRC (Ping timeout: 260 seconds) 2013-05-05T08:30:59 *** rtdin has joined #rtems 2013-05-05T08:37:52 kiwichris,hi 2013-05-05T08:39:31 I encounter a problem when processing arm thumb interworking. 2013-05-05T08:44:16 I write a testcase. rap ld ./y.rap. first rtld arm code jump to testcase thumb code, and this is ok. But when testcase call in-rtld functions, system halts. 2013-05-05T09:21:21 *** S_Somani has quit IRC (Ping timeout: 256 seconds) 2013-05-05T09:29:26 *** S_Somani has joined #rtems 2013-05-05T09:42:24 *** Hesham has joined #rtems 2013-05-05T09:47:48 *** phipse has joined #rtems 2013-05-05T09:50:50 *** mgahlot|away has joined #rtems 2013-05-05T09:53:27 *** mgahlot|off has quit IRC (Ping timeout: 268 seconds) 2013-05-05T10:01:07 *** freenix has quit IRC (Quit: Leaving) 2013-05-05T10:08:07 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-05T10:08:29 *** S_Somani has joined #rtems 2013-05-05T10:12:29 *** S_Somani1 has joined #rtems 2013-05-05T10:12:51 *** S_Somani has quit IRC (Ping timeout: 256 seconds) 2013-05-05T10:23:17 *** S_Somani1 has quit IRC (Ping timeout: 255 seconds) 2013-05-05T10:24:47 *** S_Somani has joined #rtems 2013-05-05T10:33:55 *** rtdin has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-05-05T10:43:07 *** Hesham has quit IRC (Ping timeout: 264 seconds) 2013-05-05T10:43:39 *** Hesham has joined #rtems 2013-05-05T12:11:17 *** mgahlot|away is now known as mgahlot|off 2013-05-05T13:37:35 *** phipse has quit IRC (Quit: leaving) 2013-05-05T13:53:24 *** mgahlot|away has joined #rtems 2013-05-05T13:54:02 *** mgahlot|off has quit IRC (Ping timeout: 252 seconds) 2013-05-05T14:44:07 *** ppisa has joined #rtems 2013-05-05T15:01:49 *** S_Somani has left #rtems 2013-05-05T15:22:38 *** S_Somani has joined #rtems 2013-05-05T15:22:45 *** S_Somani has left #rtems 2013-05-05T16:03:37 *** mgahlot|off has joined #rtems 2013-05-05T16:06:11 *** mgahlot|away has quit IRC (Ping timeout: 255 seconds) 2013-05-05T17:23:14 *** lcpfnvcy has quit IRC (Ping timeout: 256 seconds) 2013-05-05T17:28:23 *** lcpfnvcy has joined #rtems 2013-05-05T17:35:08 *** Hesham has quit IRC (Ping timeout: 251 seconds) 2013-05-05T17:35:08 *** ppisa has left #rtems ("Kopete 0.12.7 : http://kopete.kde.org") 2013-05-05T19:11:33 *** Hesham has joined #rtems 2013-05-05T19:12:37 *** Hesham1 has joined #rtems 2013-05-05T19:16:38 *** Hesham has quit IRC (Ping timeout: 272 seconds) 2013-05-05T19:27:12 *** the9a3eedi has joined #rtems 2013-05-05T19:40:31 *** Hesham1 has quit IRC (Ping timeout: 268 seconds) 2013-05-05T19:57:10 *** mgahlot|off has quit IRC (Ping timeout: 268 seconds) 2013-05-05T19:58:25 *** freenix has joined #rtems 2013-05-05T20:51:58 *** Hesham has joined #rtems 2013-05-05T20:52:04 *** Hesham has left #rtems 2013-05-05T21:10:45 *** the9a3eedi has quit IRC (Ping timeout: 256 seconds) 2013-05-05T23:56:35 *** the9a3eedi has joined #rtems