2014-05-26T01:19:25 *** zoso has joined #rtems 2014-05-26T01:21:36 *** edwardk has quit IRC (Quit: Computer has gone to sleep.) 2014-05-26T01:23:23 *** sebhub has joined #rtems 2014-05-26T01:23:56 good morning 2014-05-26T01:34:21 sebhub, hi 2014-05-26T01:36:09 The gdbarmsim issue is a dud stack pointer. I suspect the gdbarmsim linkcmd which is the standard ones for ARM is not working with the old start.S in that BSP. 2014-05-26T01:37:40 it shouldn't have its own start.S 2014-05-26T01:41:42 Yeap I agree. 2014-05-26T01:41:52 I will move it over and it should clean it up 2014-05-26T01:44:40 Shezzz I hate this build system, adding and changing things like new source or a header may or may not work so a bootstrap and rebuild ... grr 2014-05-26T01:47:21 yes, its a sometimes it works system 2014-05-26T01:48:09 ... more times it does not .. at least the parallel bootstrap in the RSB speeds that part up 2014-05-26T01:49:06 i normally do a "make preinstall" after configure.ac or Makefile.am changes, and this works most of the time 2014-05-26T01:49:46 Ah yes running the preinstall directly would help 2014-05-26T01:53:06 Is there a default bsp_interrupt_dispatch ? 2014-05-26T01:53:44 no 2014-05-26T01:54:03 Hmm ok. Given there are no interrupts supported I will provide a stub 2014-05-26T01:54:08 every useful bsp needs an interrupt controller 2014-05-26T01:54:19 yes, this is a bit special 2014-05-26T01:56:11 *** edwardk has joined #rtems 2014-05-26T01:57:53 By the way the SWI on the zynq in tm27 does not work 2014-05-26T01:59:00 I hacked tm27 to print.... 2014-05-26T01:59:02 ][ ICDISPR0=00000000 ICDABR0=00000000 ICCHPIR=000003ff ICCICR=00000001 2014-05-26T01:59:02 ][ ICDISPR0=00002000 ICDABR0=00000000 ICCHPIR=0000000d ICCICR=00000001 2014-05-26T01:59:20 The first line is before and the second in the loop waiting. 2014-05-26T01:59:41 I cannot see what is not being enabled. It all looks ok 2014-05-26T02:01:39 cya 2014-05-26T02:01:41 *** kiwichris has quit IRC () 2014-05-26T02:04:25 *** edwardk_ has joined #rtems 2014-05-26T02:06:10 *** AoLaD has joined #rtems 2014-05-26T02:06:17 *** edwardk has quit IRC (Ping timeout: 264 seconds) 2014-05-26T02:06:56 *** AoLaD has left #rtems 2014-05-26T03:05:01 *** kiwichris has joined #rtems 2014-05-26T04:30:29 *** edwardk_ has quit IRC (Quit: Computer has gone to sleep.) 2014-05-26T05:20:10 *** edwardk has joined #rtems 2014-05-26T05:32:47 *** edwardk has quit IRC (Quit: Computer has gone to sleep.) 2014-05-26T06:15:58 *** edwardk has joined #rtems 2014-05-26T06:23:53 *** edwardk has quit IRC (Ping timeout: 258 seconds) 2014-05-26T06:32:24 sebhub, changed gdbarmsim over the the standard startup and that fixed a couple of tests. spfatal26 is not failing. I suspect the simulator does not raise the exception. I think this test should be remove from the bsp 2014-05-26T06:32:50 sorry spfatal26 is not ending and timing out 2014-05-26T06:36:19 *** edwardk has joined #rtems 2014-05-26T06:41:16 *** edwardk has quit IRC (Ping timeout: 276 seconds) 2014-05-26T06:51:13 *** zoso has quit IRC (Read error: Connection reset by peer) 2014-05-26T06:51:45 *** zoso has joined #rtems 2014-05-26T07:51:58 yes, the simulator seems to lack general support for exceptions 2014-05-26T08:00:16 Ok for me to commit the patch for no-isr ? 2014-05-26T08:01:02 *** zoso has quit IRC (Ping timeout: 252 seconds) 2014-05-26T08:10:07 yes, this spfatal26 is a problem for some simulators 2014-05-26T08:23:39 sebhub, the tm27 problem is there is no Lower_tm27_intr call so the priority is never lowered. 2014-05-26T08:26:10 *** edwardk has joined #rtems 2014-05-26T08:30:37 *** edwardk has quit IRC (Ping timeout: 252 seconds) 2014-05-26T08:41:46 *** atgreen has joined #rtems 2014-05-26T09:36:47 *** sebhub has quit IRC (Remote host closed the connection) 2014-05-26T09:52:37 *** edwardk has joined #rtems 2014-05-26T10:04:27 *** edwardk has quit IRC (Quit: Computer has gone to sleep.) 2014-05-26T10:07:11 *** edwardk has joined #rtems 2014-05-26T10:12:15 *** edwardk has joined #rtems 2014-05-26T11:12:02 *** edwardk has quit IRC (Quit: Computer has gone to sleep.) 2014-05-26T12:24:24 *** edwardk has joined #rtems 2014-05-26T12:48:22 *** monstr has quit IRC (Ping timeout: 258 seconds) 2014-05-26T14:57:50 *** edwardk has quit IRC (Quit: Computer has gone to sleep.) 2014-05-26T14:59:17 *** edwardk has joined #rtems 2014-05-26T16:37:23 *** edwardk has quit IRC (Quit: Computer has gone to sleep.) 2014-05-26T16:46:30 *** mattttttt has joined #rtems 2014-05-26T17:50:57 *** atgreen` has joined #rtems 2014-05-26T17:52:43 *** atgreen has quit IRC (Ping timeout: 252 seconds) 2014-05-26T17:53:18 *** mattttttt has quit IRC (Ping timeout: 240 seconds) 2014-05-26T18:00:18 *** kiwichris has quit IRC () 2014-05-26T18:43:30 *** kiwichris has joined #rtems 2014-05-26T18:44:47 *** atgreen` has quit IRC (Ping timeout: 252 seconds) 2014-05-26T18:46:31 *** edwardk has joined #rtems 2014-05-26T19:07:58 *** edwardk has quit IRC (Quit: Computer has gone to sleep.) 2014-05-26T19:27:17 *** gigetoo has quit IRC (Remote host closed the connection) 2014-05-26T19:28:24 *** gigetoo has joined #rtems 2014-05-26T23:42:56 *** monstr has joined #rtems 2014-05-27T00:01:24 *** monstr has quit IRC (Remote host closed the connection) 2014-05-27T00:04:12 *** monstr has joined #rtems 2014-05-27T00:41:51 *** sebhub has joined #rtems 2014-05-27T00:45:04 good morning 2014-05-27T01:24:29 *** kiwichris has quit IRC () 2014-05-27T01:52:12 *** zoso has joined #rtems 2014-05-27T01:53:14 *** zoso has joined #rtems 2014-05-27T02:02:05 *** kiwichris has joined #rtems 2014-05-27T02:29:58 *** kiwichris has quit IRC () 2014-05-27T02:58:32 *** zoso has quit IRC (Read error: Connection reset by peer) 2014-05-27T02:58:53 *** zoso has joined #rtems 2014-05-27T03:57:14 *** kiwichris has joined #rtems 2014-05-27T03:58:02 *** edwardk has joined #rtems 2014-05-27T05:39:48 sebhub, still looking at the SGI. I see in the GIC int dispatch the EOI is written on return. With tm27 I need to nest so the ack should be before dispatch is called. 2014-05-27T05:42:31 sebhub, hmmm this seems to fix the test 2014-05-27T05:51:13 *** atgreen has joined #rtems 2014-05-27T06:48:49 *** atgreen has quit IRC (Ping timeout: 276 seconds) 2014-05-27T06:51:21 *** AoLaD has joined #rtems 2014-05-27T06:57:58 *** cdcs has joined #rtems 2014-05-27T07:00:15 *** asuolen has joined #rtems 2014-05-27T07:04:46 Hesham, here ? 2014-05-27T07:06:18 *** Hesham has joined #rtems 2014-05-27T07:06:29 Hesham, hi 2014-05-27T07:06:36 Hi kiwichris 2014-05-27T07:06:49 Happy to start a little before the others ? 2014-05-27T07:07:00 Sure 2014-05-27T07:07:36 Thanks for the RSB patch. I have looked over it and it seems fine. I need to test it before pushing because of the change in the gdb config. 2014-05-27T07:08:06 Take your time, and let me know if there is any other issue to fix 2014-05-27T07:08:11 I have also reviewed the newlib patch and it also looks much better. 2014-05-27T07:08:40 There is a file called or1k-support.h that should be removed ? 2014-05-27T07:08:52 It has been a crazy day here and I hope tomorrow is a little slower so I can sort this out. 2014-05-27T07:08:59 Where ? 2014-05-27T07:09:15 One second ... 2014-05-27T07:10:03 No I have already discarded it in the latest newlib patch 2014-05-27T07:10:10 It has a GPL licence 2014-05-27T07:10:13 Great. 2014-05-27T07:10:15 Excellent. 2014-05-27T07:10:38 One things I wanted to ask about the newlib patch. Did you look in the nosys directory ? 2014-05-27T07:11:46 No, I do not know what to do with it, I just redirected system calls from libgloss to rtems' in the configuration files as I remember 2014-05-27T07:13:25 I suspect a lot of the code in the libgloss part of the patch could be removed and the nosys source referenced. 2014-05-27T07:14:04 This is the purpose of the nosys directory, to stub out those functions for libgloss. For RTEMS there is not need. 2014-05-27T07:14:24 What are you working on now ? 2014-05-27T07:15:22 Sure, I can do that, does this mean to delete libgloss/or1k, or keep the minimum files to build newlib ? 2014-05-27T07:16:42 I am in the middle of my final exams these days (end by June 8), so I am afraid I have no much time for the project currently 2014-05-27T07:17:04 I am not sure. I would take it in steps. Find the call in nosys, add to the makefile and delete the code from the or1k dir. When you reach the end you will see how things are placed. 2014-05-27T07:17:25 That is fine and I full understand. 2014-05-27T07:17:39 Thanks :) 2014-05-27T07:18:20 I will try to work on newlib and nosys when I have some free time 2014-05-27T07:18:35 Did you see the offlist mail regarding or1ksim patches ? 2014-05-27T07:18:51 Hang on. 2014-05-27T07:19:39 I have the github download issue to look at. Where was the email sent ? 2014-05-27T07:19:58 Oh off list. No I do not remember the email. 2014-05-27T07:20:48 Yes that's it, I have sent it to your e-mail with subject " Problem with downloading a github repo from RSB" 2014-05-27T07:21:08 Yes I have this flagged to look at. 2014-05-27T07:21:25 That's it. 2014-05-27T07:21:27 I have a few work related things taking priority 2014-05-27T07:22:03 I think this is all. Please let Gedare know we have talked. Thanks 2014-05-27T07:22:16 I believe it's a simple issue, I will be working on other tasks meanwhile. 2014-05-27T07:22:19 Sure 2014-05-27T07:28:02 hi Hesham 2014-05-27T07:28:06 and everyone :) 2014-05-27T07:29:16 it seems i missed the meeting ;) 2014-05-27T07:41:16 *** JustJanek has joined #rtems 2014-05-27T07:42:59 hi beng-nl how are you? 2014-05-27T07:43:10 The meeting has not been started yet 2014-05-27T07:47:39 oh yeah no i mean your project 2014-05-27T07:49:42 Yeah, we did not talk much about it :) 2014-05-27T07:51:24 Can I ask about something about newlib? 2014-05-27T07:55:55 *** Sched has joined #rtems 2014-05-27T07:57:52 *** phipse has joined #rtems 2014-05-27T07:57:55 hi 2014-05-27T07:58:10 *** gedare has joined #rtems 2014-05-27T07:58:19 hi everyone 2014-05-27T07:59:22 hi 2014-05-27T07:59:31 hi 2014-05-27T07:59:31 here is the link for today's meeting doc: https://docs.google.com/document/d/1QyTZPPsxHR7KOLZfGW799Jex0nsIxqOZ7vjxO2_NNQ4/edit?usp=sharing 2014-05-27T07:59:46 looks like all the students are present today 2014-05-27T08:00:04 hi Hesham, AoLaD, asuolen, Sched, JustJanek 2014-05-27T08:00:17 Hi,gedare 2014-05-27T08:00:20 hmm no access to edit 2014-05-27T08:00:21 Hi gedare 2014-05-27T08:00:21 Hi everyone 2014-05-27T08:00:30 oops 2014-05-27T08:00:37 hey 2014-05-27T08:00:48 ok edit should be doable now 2014-05-27T08:01:03 you might need to refresh i don't know. 2014-05-27T08:01:15 yeah, refresh is required 2014-05-27T08:01:56 students, the basic format of these meetings is pretty much the same.. update about what you've done since last meeting, what you said you were going to do but didn't / trouble you encountered, and what you will work on until next time 2014-05-27T08:02:03 Hesham: would you like to go first? 2014-05-27T08:02:08 Sure 2014-05-27T08:02:21 hi kiwichris 2014-05-27T08:02:28 First I and kiwichris talked about the project an hour ago 2014-05-27T08:02:36 Here... 2014-05-27T08:02:40 ok great 2014-05-27T08:03:19 Second, regarding my status, as I said last week, I could not do much due to my exams, but I have submitted patches to RSB in the mailing list 2014-05-27T08:03:35 and fixed it according to Chris feedback. 2014-05-27T08:03:41 yes i saw, that is good 2014-05-27T08:04:01 are you mirroring RSB and updating on github also? 2014-05-27T08:05:00 gedare: do you have last weeks meeting notes handy? i lost the url 2014-05-27T08:05:17 I have cloned RSB, and always generate patch against remote, about github, I did not work a lot, but yes, all the code I have done with RTEMS is pushed to github 2014-05-27T08:06:10 Here's the link from last week's docs: https://docs.google.com/document/d/1313SYe2afNsjxKRRgQTSYA8CletOtknSQXykuZ3zZFo/edit?usp=sharing 2014-05-27T08:06:11 I will be working on deleting some libgloss code from my newlib patch and use nosys as Chris suggested 2014-05-27T08:06:47 *** DrJoel has joined #rtems 2014-05-27T08:06:47 *** DrJoel has joined #rtems 2014-05-27T08:06:47 *** ChanServ sets mode: +o DrJoel 2014-05-27T08:06:59 *** jenniferA has joined #rtems 2014-05-27T08:07:28 *** gedare_ has joined #rtems 2014-05-27T08:07:52 JustJanek: thanks 2014-05-27T08:08:03 welcome back gedare_, should I sent what I wrote again? 2014-05-27T08:08:08 send* 2014-05-27T08:08:12 no i read the log 2014-05-27T08:08:43 Good, any comments? 2014-05-27T08:08:46 ok Hesham that sounds good 2014-05-27T08:08:48 umm 2014-05-27T08:08:55 did you get any resolution about that setjmp issue? 2014-05-27T08:08:58 or is it still waiting 2014-05-27T08:09:12 Most likely I will write it again from scratch 2014-05-27T08:09:18 ok i think that is good 2014-05-27T08:09:19 I got no response from the author 2014-05-27T08:09:44 ok great. 2014-05-27T08:09:48 any volunteer to go next? 2014-05-27T08:09:52 *** gedare has quit IRC (Ping timeout: 245 seconds) 2014-05-27T08:09:58 *** gedare_ is now known as gedare 2014-05-27T08:10:02 i wonder if the author is on irc 2014-05-27T08:10:09 this is for a file in newlib, right? 2014-05-27T08:10:55 Sched: are you here? 2014-05-27T08:11:23 verm__: the author is not on IRC of course, he is Damjan founder of opencores and openrisc 2014-05-27T08:11:26 Yeah 2014-05-27T08:11:38 ok Sched would you like to go now? 2014-05-27T08:11:43 hi phipse 2014-05-27T08:11:44 verm__: I poked around Google and linkedin for the author of setjmp and didn't find him. It is not that much assembly. Just save all and restore all. 2014-05-27T08:11:47 hi cdcs 2014-05-27T08:11:51 I have implement a hypercall system, which is imitate from syscall. I want to use this system to separate apart the normal function and hypervisor function in POK kernel. I have finish the most jobs but still needs testing job to make it work rightly. 2014-05-27T08:12:10 Sched: is your code available on github? 2014-05-27T08:12:12 Hi all 2014-05-27T08:12:19 Did you read the discussion about including paravirtualization headers in RTEMS? 2014-05-27T08:12:30 That was my next question. :) 2014-05-27T08:12:51 Yes, and my mentor have review it and point out some mistake 2014-05-27T08:13:10 Sched: Also, do you have some design notes about your implementation? A blog post or document somewhere to describe the ideas? 2014-05-27T08:13:38 for the next question, For now, No 2014-05-27T08:14:03 No you have not read the rtems-devel thread about paravirtualization? 2014-05-27T08:14:36 Of course, as the the convention, I will post a blog not later then Friday about what I have do last week. 2014-05-27T08:14:59 Sched: Where are you on the virtual CPU for POK question? 2014-05-27T08:15:14 Did you read up on virtual CPUs in Xen or some other hypervisor? 2014-05-27T08:15:37 DrJoel: ah, ok, what is the file? 2014-05-27T08:15:50 Hesham: i see, hmm 2014-05-27T08:16:37 verm__: setjmp.S 2014-05-27T08:16:59 Sched: You should definitely study the mailing list thread, it may be is important for your work. 2014-05-27T08:17:04 Yes, I have read some of them, but not very proficient, I consider the hypercall system is the basic of paravirtualization, so I want to finish it first, then implement the virtual CPU according to needing by hypercall. 2014-05-27T08:17:21 Sched: It is important that you describe what you will do before you do it 2014-05-27T08:17:32 Sched: Otherwise you may waste a lot of time on implementing something that is not correct 2014-05-27T08:17:45 Yes, that's my mistake. 2014-05-27T08:18:01 Sched: It is best to outline the requirements for your code and to determine what the interface functions will be. 2014-05-27T08:19:01 OK.. for next week I want you to focus on describing your design and reading about how to do it 2014-05-27T08:19:09 Yes, I will notice it next week. 2014-05-27T08:19:29 Sched: I strongly recommend, that you have a look at virtual CPUs. Design/Concept and the implementation including kernel and user-land side. 2014-05-27T08:19:31 And of course whatever else phipse says. :) 2014-05-27T08:19:47 OK, That's what I learn last week. 2014-05-27T08:20:12 phipse, I will learning it, before I do the jobs for next week 2014-05-27T08:20:26 because it's necessary for my next action. 2014-05-27T08:20:45 and please write about it before you implement it 2014-05-27T08:21:10 Sorry for my mistake last week. 2014-05-27T08:21:52 I thought it might be easy so I just want to do it directly. 2014-05-27T08:21:56 OK. JustJanek Are you here? 2014-05-27T08:22:04 Sched: Nothing is just easy :) 2014-05-27T08:22:10 Yeah 2014-05-27T08:22:16 OK JustJanek Can you go now? 2014-05-27T08:22:22 Yep, sure 2014-05-27T08:22:31 OK, fyi cdcs is your mentor Claudio 2014-05-27T08:22:39 and phipse is philipp 2014-05-27T08:23:21 now what i've done for the past week is try to setup the environment with rtems running virtualized on POK on a dedicated system (qemu kvm problems were caused because i was running within a virtual ubuntu machine ) 2014-05-27T08:24:01 so now I can run the "hello world", though it crashes eventually but I can clearly see the hello world text 2014-05-27T08:24:50 with regard to the crash philipp may be able to help 2014-05-27T08:25:08 how is your school allocation during the next weeks? 2014-05-27T08:25:10 I also tried to simply run an arinc653 call by including one of the arinc653 header files and trying to use one of the functions but that couldn't compile which isnt completely surprising 2014-05-27T08:25:15 Does your crash referr to "An Error occured: 0!" ? 2014-05-27T08:25:41 I mean, do you think it crashed because of that line in the output? 2014-05-27T08:26:04 no, qemu crashes 2014-05-27T08:26:18 hmmmm 2014-05-27T08:26:23 I dont remember seeing a string saying "An Error occured" 2014-05-27T08:26:53 Post to the mailing list with the details on that. Include the qemu invocation and version information. 2014-05-27T08:27:03 even more hmmm ... are you able to make a screenshot of the output ? 2014-05-27T08:27:33 There is also a debug/trace option to qemu which you can turn on. Maybe that can capture the instruction sequence right before the crash 2014-05-27T08:27:39 currently I cannot because I'm at school but when I'm back home I definitely can 2014-05-27T08:27:58 Better for devel than this IRC anyway 2014-05-27T08:28:27 OK JustJanek. Have you seen the rtems-devel mailing list thread about paravirtualization? 2014-05-27T08:28:46 If not, please do read it thoroughly 2014-05-27T08:28:47 not yet, no 2014-05-27T08:29:00 is it the thread about including paravirtualization to rtems? 2014-05-27T08:29:04 yeah 2014-05-27T08:29:05 yes 2014-05-27T08:29:16 it is important for both you and Sched 2014-05-27T08:29:28 OK , 2014-05-27T08:29:33 I got it 2014-05-27T08:29:45 okay, i will check it out when I'm back home 2014-05-27T08:29:47 JustJanek: what will you be working on until next week? 2014-05-27T08:30:16 phipse I want to discuss it with you for more detail. after meeting :-) 2014-05-27T08:30:39 I want to look in more detail to calling arinc653 calls 2014-05-27T08:30:40 btw, DrJoel, do you agree with our "conclusions" in that thread? 2014-05-27T08:30:56 last time i hadnt looked into properly configuring the aadl file 2014-05-27T08:30:58 for arinc653 2014-05-27T08:31:09 which i assume is why it didnt compile 2014-05-27T08:31:28 cdcs Need to read the entire thread again. And reply publicly. I must admit to enjoying the holiday weekend here 2014-05-27T08:31:34 ok JustJanek sounds good, thanks. 2014-05-27T08:31:41 DrJoel =) 2014-05-27T08:31:44 AoLaD: are you here? 2014-05-27T08:31:59 Hello 2014-05-27T08:32:17 hi AoLaD since you missed last week, can you introduce yourself with name and title of project 2014-05-27T08:32:19 FWIW Chris is right concerned about making sure the entire process of getting an RTEMS exe into a Pok "container" and testing is well documented and easily reproducible. His focus/concern is on automated testing 2014-05-27T08:32:23 and then say what you've done and what you will be doing? 2014-05-27T08:32:35 AoLaD: DrJoel is your co-mentor, along with ppisa who is not here today 2014-05-27T08:32:44 but whom i guess you probably see in person some times. :) 2014-05-27T08:32:53 JustJanek, Sched, after the meeting we can discuss a little more about your projects if you want 2014-05-27T08:32:54 sure :) 2014-05-27T08:33:11 Of course :-)\ 2014-05-27T08:33:53 well, i'm a slow writer, so intruduction will take some time. I've got preapered small status report 2014-05-27T08:34:30 For an introduction I'd like to thank you all for accepting me. 2014-05-27T08:35:37 OK AoLaD in the future you can also prepare your report and copy-paste into here. 2014-05-27T08:36:12 Last wednesday i met with Mr. Pisa. We looked into some ways to compile RTEMS for our MCU. It looked promising but we’ll see when test board arrive. Mr. Pisa has already negotiated development board from TI (Texas Instruments). It should arrive anytime soon. In meantime I should have dig into serial peripheral for future debugging support. Next task was to construct header file with registers 2014-05-27T08:36:12 definitions from TI’s reference manual. I’m afraid to say, that I’ve get sick and I haven’t been able to do that. I’ve originally dedicated 2 day in a week for this project and the rest for exams and study (until I pass all exams, say if i should reschedule). Sadly, because of illness I did not meet plans for last week. 2014-05-27T08:36:12 For this week I’d like to make that header file. I hope that I won't have to stay in bed for all weekend. 2014-05-27T08:36:20 yeah, like that.. ;) 2014-05-27T08:36:59 when will your exams be done? 2014-05-27T08:37:44 AoLaD: have you read the BSP developers guide? 2014-05-27T08:37:47 Hesham: from which architecture? 2014-05-27T08:37:57 I have 6 exams this semestr. I need time only for two of them. 2014-05-27T08:37:59 AoLaD: http://rtems.org/onlinedocs/doc-current/share/rtems/html/bsp_howto/index.html 2014-05-27T08:38:05 thank you 2014-05-27T08:38:18 it also has pdf available as a link 2014-05-27T08:38:22 http://rtems.org/onlinedocs/doc-current/share/rtems/html/ 2014-05-27T08:38:23 from there 2014-05-27T08:38:57 AoLaD.. have you answered the first question of what gcc flags to pass to specify the CPU model? 2014-05-27T08:39:16 j 2014-05-27T08:40:03 I also got a exam tomorrow, I did not review it until now, this year the GSoC is start early then usual :-( 2014-05-27T08:40:18 DrJoel: which architecture is setjmp.s needed from? 2014-05-27T08:40:25 weve looked for that. After time we discovered it should be quiete simple... 2014-05-27T08:40:48 But I supouse, that I need board to test that. 2014-05-27T08:41:11 verm__: or1k .. 2014-05-27T08:41:40 AoLaD: Does qemu support somethign similar? 2014-05-27T08:41:56 DrJoel: which one is it here?: https://sourceware.org/cgi-bin/cvsweb.cgi/src/newlib/libc/machine/?cvsroot=src 2014-05-27T08:42:03 AoLaD: But you need the board.. soon. :) 2014-05-27T08:42:12 verm__: not upstream 2014-05-27T08:42:56 Ive looked in quemu and i dont think cortex R4 is supported 2014-05-27T08:43:15 ok AoLaD read that BSP guide, look at the ARM bsp's also which are a bit specialized, and investigate whether there are any simulator options 2014-05-27T08:43:16 it's not upstream? hmm 2014-05-27T08:43:26 just looking for the history for setjmp.S 2014-05-27T08:43:33 verm__ version on net has murky history.. and current (unmerged) source has been contaiminated by being declared GPLv3. Such a small file, it isn't worth the trouble of dealing with hunting. A couple of hours to write and test tops. 2014-05-27T08:43:38 AoLaD: And rest to get better and study for your exams. :) 2014-05-27T08:43:41 i'm usually pretty good at chasing people down 2014-05-27T08:44:16 AoLaD get better but you must have a way to test and debug. ASAP .. and Pavel and I will have to obtain the same setup 2014-05-27T08:44:30 thank you. I've had one of the big ones today...meybe I'me half way done :) 2014-05-27T08:44:45 DrJoel: I sent an email awhile ago about purchasing a board for testing, but you should ask Pavel about what are good options 2014-05-27T08:45:08 we should also purchase a raspberry pi for the lab and perhaps an FPGA for the or1k testing. 2014-05-27T08:45:17 Thank you AoLaD. Next up, asuolen 2014-05-27T08:45:53 hello all! 2014-05-27T08:46:02 This past week I have been thinking about the GPIO API: what it should do and look. The discussion is currently going on the dev mailling list and some draft code is already in github. 2014-05-27T08:46:23 AoLaD: Keep me honest on getting hardware .. dig the info up and start a thread on devel with it. 2014-05-27T08:47:01 DrJoel: yeah nevermind then better we just do it ourselves and not worry about it 2014-05-27T08:47:07 especially since the gplv3 is involved. 2014-05-27T08:47:09 verm__: or1k (openrisc) 2014-05-27T08:47:18 I also got some hardware this week that will be needed to have a degub environment with the raspberry, and have been discussing that configuration with alan 2014-05-27T08:47:18 Hesham: yeah got it, thanks :) 2014-05-27T08:47:26 asuolen: I mentioned early on (and shold have piped up) that there is existing code for multiio (GPIO, DAC, ADC) at http://git.rtems.org/multiio/ 2014-05-27T08:47:59 asuolen: it may be worth merging the multiio into rtems/cpukit, if the code is any good there. 2014-05-27T08:48:00 It has drivers for two PC104 based multiio boards along with a set of shell commands which assist in testing. 2014-05-27T08:48:28 DrJoel: will have a look at that then 2014-05-27T08:48:33 but definitely keep working to define what is a good and useful GPIO interface for users across rtems 2014-05-27T08:48:34 I am not convinced the API is perfect. It does not include some nice features we had on some in-house code. But it has been used on a robotic project and was sufficient there. 2014-05-27T08:48:48 (I looked at it once and recall thinking it was a bit incomplete.) 2014-05-27T08:49:15 It is more mature than starting from scratch. Eveolving to improve API and support other boards would be a big plus to maturing it. 2014-05-27T08:49:41 yes. and having the pc104 boards to compare with is good too. DrJoel do you still have those boards for testing? 2014-05-27T08:50:01 gedare I am sure I have one of those. 2014-05-27T08:50:15 I would have to think about why I did the other one. :) 2014-05-27T08:50:59 ok good. So asuolen for next week are you planning to continue with the GPIO discussion and coding, and setting up your debug infrastructure? 2014-05-27T08:51:31 AoLaD: I forgot to mention, please make sure to post your code as you start your new BSP to your github repo. 2014-05-27T08:52:18 asuolen: I did all my testing with a multimeter to test DAC/discrete out, push button (with proper resistance) to test discrete input, and when ADC was tested looped a DAC back and checked to see if read matched written 2014-05-27T08:52:57 So you don't need much test equipment. One of the commands in multiio writes a stairstep voltage pattern on a DAC and you can watch it on a cheap multimeter 2014-05-27T08:53:17 yes I intend to continue the API discussion and will look into multiio. I also have interest in debug support for the Pi as soon as possible as my university project can also benefit from it 2014-05-27T08:54:29 A concept worth considering once the basics work is a server which polls inputs at a configured rate and triggers user "change handlers". This is a common application need. Inputs often don't cause interrupts 2014-05-27T08:54:32 OK asuolen sounds good. 2014-05-27T08:54:43 Thanks everyone that concludes the meeting.. any post-meeting meetings can now commence! 2014-05-27T08:55:46 such a server is a higher-level service that should be kept in mind as a user of the interfaces, but not part of the interfaces themselves i suspect. 2014-05-27T08:56:29 Also update your section of the meeting documen 2014-05-27T08:56:35 https://docs.google.com/document/d/1QyTZPPsxHR7KOLZfGW799Jex0nsIxqOZ7vjxO2_NNQ4/edit?usp=sharing 2014-05-27T08:58:24 sorry, had a phone call back now 2014-05-27T08:59:10 DrJoel: a polling function on the API? 2014-05-27T08:59:32 *** zoso has quit IRC (Ping timeout: 265 seconds) 2014-05-27T08:59:55 *** zoso has joined #rtems 2014-05-27T09:01:51 Hi 2014-05-27T09:02:21 JustJanek cdsk 2014-05-27T09:02:35 should we have a discussion 2014-05-27T09:03:31 Sched, i have a side task for you 2014-05-27T09:03:58 identify the registers that are protected in the x86 architecture 2014-05-27T09:04:40 asuolen: What we had was a standard API for applications to call. But some pins/fields/ADCs were read on demand. Others were polled by the server thread and the app/API read the value from memory. 2014-05-27T09:04:40 Yes? 2014-05-27T09:04:44 when i say protected, i mean those registers that are only acessible in lower rings 2014-05-27T09:04:54 https://docs.google.com/document/d/10ehcM1f2eKNwcNgv5stphGtsVAnYc_K7KLhxmjvE1k8/edit 2014-05-27T09:05:09 the task is to "identify the registers that are protected in the x86 architecture" 2014-05-27T09:05:15 I thought you can have a look about my proposal 2014-05-27T09:05:19 We did this because we had ~128 discrete I/Os and ~16-24 ADCs/DACs. Bus bandwidth hitting the individually was killing the system. 2014-05-27T09:05:28 this will help you to understand the virtualCPU concept 2014-05-27T09:05:34 in the future 2014-05-27T09:05:38 Plus the framework we had could also read multi-bit wide subfields. 2014-05-27T09:05:40 my mentor have ask me to describe the x86 hole in the proposal 2014-05-27T09:06:14 Sched: you descibed senstive/protected instructions not registers 2014-05-27T09:06:54 Yes, but they are interconnect 2014-05-27T09:06:59 OK 2014-05-27T09:07:10 so my to-do is : 2014-05-27T09:07:13 ok, that should anyway listed somewhere in the wiki 2014-05-27T09:07:34 Sched: Your todo goes here too: https://docs.google.com/document/d/1QyTZPPsxHR7KOLZfGW799Jex0nsIxqOZ7vjxO2_NNQ4/edit 2014-05-27T09:07:42 1, rectify the jobs about hypercall system 2014-05-27T09:07:57 Yep I know 2014-05-27T09:08:00 k 2014-05-27T09:08:15 as part of your project, Sched, we will need to define which registers we will include in the virtualCPU 2014-05-27T09:08:16 I just want to tell you to discuss if I miss any thing 2014-05-27T09:08:28 that you will implement in POK 2014-05-27T09:08:49 RTEMS will then access these "virtual registers" through hypercalls 2014-05-27T09:08:50 2, find the sensitive register ? is this right? 2014-05-27T09:09:01 Yep I know. 2014-05-27T09:09:09 ok :) 2014-05-27T09:09:20 we need to define which ones we need to include in the virutal CPU 2014-05-27T09:09:50 No 2. should be read the mail list first 2014-05-27T09:10:01 yes, the email is important 2014-05-27T09:10:07 then 3. is find the sensitive register 2014-05-27T09:10:21 4. implement some necessary hypercall 2014-05-27T09:11:47 That's the to-do for now. 2014-05-27T09:12:26