2013-03-25T00:44:45 *** the9a3eedi-2 has joined #rtems 2013-03-25T00:46:21 *** the9a3eedi has quit IRC (Ping timeout: 245 seconds) 2013-03-25T01:50:27 *** arvind_khadri has joined #rtems 2013-03-25T02:04:04 *** sebhub has joined #rtems 2013-03-25T02:06:10 good morning 2013-03-25T02:07:54 sebhub, hi 2013-03-25T02:25:47 the source builder works fine on opensuse 2013-03-25T02:29:05 *** freenix has joined #rtems 2013-03-25T02:40:00 *** freenix has quit IRC (Quit: Leaving) 2013-03-25T02:58:26 thanks. 2013-03-25T02:58:42 Any thing I need to add to the doco in the hosts section ? 2013-03-25T02:59:48 i had to install the python-devel package for gdb, that was all 2013-03-25T03:00:50 and there is a warning for chown 2013-03-25T03:00:54 Command line ? What is the exact version of Opensuse ? 2013-03-25T03:01:06 Oh, what ? 2013-03-25T03:01:41 the standard python package works, i don't know the version 2013-03-25T03:01:54 I mean opensuse 2013-03-25T03:02:29 12.3 2013-03-25T03:02:35 Thanks 2013-03-25T03:02:47 warning .. absolute path found .. chown 2013-03-25T03:03:14 Hmm, option in the man page to supressed it ? 2013-03-25T03:03:57 i think the source builder produces this warning 2013-03-25T03:04:28 really ? I will take a look. 2013-03-25T03:08:08 *** dhananjay has quit IRC (Ping timeout: 260 seconds) 2013-03-25T03:08:55 *** dhananjay has joined #rtems 2013-03-25T03:10:10 sebhub, teaching the source builder to canadian cross on cygwin to 64bit MinGW 2013-03-25T03:11:13 this is a really nice feature 2013-03-25T03:13:21 kiwichris: I was looking into the GDB project, I've looked at the GDB python APIs. They seem nice, but not much documentation. But was able to follow some tutorials using my host system. 2013-03-25T03:14:05 It is nice and yes you need to dig into the python code or look at existing pretty printers 2013-03-25T03:14:49 dhananjay, are you appling for GSoC (if we get accepted) ? 2013-03-25T03:14:50 But when I try to debug programs from existing pretty printes? Are you referring to what you've already written or the GDB tree? 2013-03-25T03:15:01 In the GDB tree 2013-03-25T03:15:49 I am planning to :) 2013-03-25T03:16:15 Great, please make sure you have done the hello sample change and filled in the wiki page 2013-03-25T03:16:16 But I have to look at my university schedules first 2013-03-25T03:16:21 Ah ok 2013-03-25T03:16:49 I would get this one anyway, if it works out great, if not it is no big deal 2013-03-25T03:17:06 .. this done anyway .. 2013-03-25T03:19:01 *** dhananjay_ has joined #rtems 2013-03-25T03:19:31 kiwichris: I might have some examinations during soc. 2013-03-25T03:19:58 ah ok, your study is always the most important thing 2013-03-25T03:19:59 *** dhananjay has quit IRC (Read error: Connection reset by peer) 2013-03-25T03:19:59 *** dhananjay_ is now known as dhananjay 2013-03-25T03:21:21 hmm... But I guess I can request the university to make it fast (Individual tests) and complete them taking a break from GSoC 2013-03-25T03:21:55 I hope thats okay. It will be ~2 weeks 2013-03-25T03:22:54 The way this normal works is you and your mentor set down the required tasks and if you need to take a study break you need to make up the time. 2013-03-25T03:23:09 Okay 2013-03-25T03:23:38 It is always best to be opne and up front about this. In the past there have been students who have not said this and they just drop off for a while. 2013-03-25T03:23:49 or explain this after they are accepted. 2013-03-25T03:24:11 So I should patch the hello world and sent it to mailing list 2013-03-25T03:24:14 This makes the mentoring role difficult 2013-03-25T03:24:31 Yes with the screen shot of it running in a simulator 2013-03-25T03:24:40 Yes. I plan to mention my complete schedule in the application. 2013-03-25T03:24:48 Excellent 2013-03-25T03:26:21 I was trying to play around the gdb a bit (Spark, built using the source-builder) and the exapmles-v2 repo 2013-03-25T03:26:48 Ah ok. I would love to have the examples converted to waf 2013-03-25T03:27:36 I see -g passed the gcc while building them, but cant see the any more debug info (when I do list it just prints a link to Start.S) 2013-03-25T03:27:57 is it the normal behaviour? 2013-03-25T03:28:17 Do you have any articles using gdb with rtems? 2013-03-25T03:28:37 In the examples ? 2013-03-25T03:29:01 No. in general 2013-03-25T03:29:19 You need the RTEMS examples to have -g 2013-03-25T03:29:54 What about 'l Init' or 'l main' does that show source code ? 2013-03-25T03:30:56 l Init works 2013-03-25T03:30:56 :) 2013-03-25T03:31:05 Great 2013-03-25T03:34:07 okay. But what do I have to use to print the entire program? 2013-03-25T03:34:07 kiwichris: Why waf? 2013-03-25T03:35:40 *** the9a3eedi-2 has quit IRC (Ping timeout: 245 seconds) 2013-03-25T03:35:42 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2013-03-25T03:40:26 *** sebhub has quit IRC (*.net *.split) 2013-03-25T03:41:44 *** rokka_ has quit IRC (Ping timeout: 264 seconds) 2013-03-25T03:42:48 *** rokka has joined #rtems 2013-03-25T03:43:13 I have some troubles with chat. See you guys later. :-( 2013-03-25T03:43:16 I have been using it on a number of projects and it is fast and simple and stable 2013-03-25T03:43:17 *** dhananjay has left #rtems 2013-03-25T03:43:39 * kiwichris is off to hit tennis balls 2013-03-25T03:54:29 *** arvind_k has joined #rtems 2013-03-25T06:17:59 *** freenix has joined #rtems 2013-03-25T06:55:35 *** freenix has quit IRC (Quit: Leaving) 2013-03-25T07:49:22 *** antgreen has quit IRC (Quit: Leaving) 2013-03-25T08:49:57 *** antgreen has joined #rtems 2013-03-25T09:02:06 *** antgreen has quit IRC (Quit: Leaving) 2013-03-25T09:04:51 *** weiY has joined #rtems 2013-03-25T10:14:51 *** gedare has joined #rtems 2013-03-25T10:35:23 gedare, good moring 2013-03-25T10:35:57 hey weiY 2013-03-25T10:36:14 Recently i have some to do some work on atomic support 2013-03-25T10:36:19 time 2013-03-25T10:36:45 cool... i'm not up to speed with what the status is 2013-03-25T10:37:30 ah .. there was one problem I wanted fixed 2013-03-25T10:37:34 there is not some big progress because of lack of time. 2013-03-25T10:38:01 first i want add the general atomic function for the UP 2013-03-25T10:38:19 right now we resorted to some compiler autotools magic to avoid the problem when cpuatomic does not exist 2013-03-25T10:38:20 yeah 2013-03-25T10:38:29 we need some sort of generic functional fallback 2013-03-25T10:38:50 that would be really good to try to implement 2013-03-25T10:38:57 yeah, i know. Now there is a macro to avoid compiler compilation 2013-03-25T10:39:19 it would be good to get rid of it :) 2013-03-25T10:39:40 *** antgreen has joined #rtems 2013-03-25T10:39:42 i think we agreed it is ok to wrap some ISR_Disable / ISR_Enable to implement the generic version? 2013-03-25T10:39:49 so i want to know what architectures are UP which will not have RTEMS_SMP macro to define 2013-03-25T10:40:16 right now smp is only available in sparc, x86, and maybe arm 2013-03-25T10:40:23 yeah, first the general ISR_Disable/ISR_Enable function must be implementted 2013-03-25T10:40:24 the arm support is not merged to the tree yet 2013-03-25T10:41:03 ok, sparc, x86, arm ,powerpc are/ will be SMP support? 2013-03-25T10:41:45 yes 2013-03-25T10:41:46 Then i will try implement general ISR related function for other architecutres? 2013-03-25T10:41:52 ok 2013-03-25T10:41:54 no... 2013-03-25T10:41:57 well 2013-03-25T10:42:04 we need it available for any architecture i guess 2013-03-25T10:42:10 until the smp is implemented... 2013-03-25T10:42:24 and even for some variants of the architectures that we implement smp on, they might not support the atomics 2013-03-25T10:42:33 like..older sparc boards don't have the casa i think 2013-03-25T10:42:46 so they would still need to fall back to the generic version 2013-03-25T10:42:53 anyway, the generic version should just be available 2013-03-25T10:43:13 and it will be linked whenever the arch does not support the atomic ops? 2013-03-25T10:43:17 yeah, but in the SMP support architecture we may be using two implementations (SMP and UP ) 2013-03-25T10:44:42 i mean in the SMP support architectures they maybe have its own atomic instructions so they do not need general implementation 2013-03-25T10:45:06 correct 2013-03-25T10:45:21 and even some non-SMP architecture, or without SMP enabled, might have the atomic operations eventually 2013-03-25T10:45:22 even SMP support is not implementation but we can use its own atomic instructions to implementation 2013-03-25T10:46:27 yeah, if they have its own atomic instructions we just use its own atomic ops not general implementation 2013-03-25T10:48:36 gedare, recently GSOC will be starting. Any new project will in open projects? 2013-03-25T10:49:14 maybe i can also partispart in this year GSOC, haha 2013-03-25T10:49:17 if you are interested, you might be able to propose more atomic support 2013-03-25T10:50:21 there might be some more projects listed... everyone's been too busy to think many new ones up :) 2013-03-25T10:50:31 yeah, i will be propose more atomic support. but i am afraid whether this proposal will be enough 2013-03-25T10:51:49 haha, i think this year GSOC. RTEMS will accept more students to do more developments 2013-03-25T10:52:42 Because of you guy so good mentors, haha 2013-03-25T10:53:23 Thanks :) 2013-03-25T10:54:31 did you get an SMP version of rtems to ever run? 2013-03-25T10:55:20 yeah, last year i have test atomic on x86 SMP version 2013-03-25T10:55:30 but i do not try the latest version 2013-03-25T10:55:47 is there any news on SMP support? 2013-03-25T10:55:51 i have not run smp myself 2013-03-25T10:55:57 i think it has slowed somewhat... 2013-03-25T10:56:04 there was a big push because of some work on the sparc i think 2013-03-25T10:56:09 more work in smp is good 2013-03-25T10:56:10 yeah, it should be more faster 2013-03-25T10:56:38 better smp scheduler, better infrastructure for ipi's and other services... 2013-03-25T10:56:50 you mean there will be new big progress on smp support about sparc? 2013-03-25T10:57:16 yeah, maybe i can proposal more work on smp support 2013-03-25T10:57:42 but i donot know a good direction about RTEMS smp improvement 2013-03-25T10:58:33 But i think first better smp scheduler should more needed for smp support 2013-03-25T10:59:31 and delete lots of big lock in kernel like the road of linux turn to SMP 2013-03-25T11:00:43 not sure if anyone is working on new SMP work right now 2013-03-25T11:01:13 and more Synchronization primitives i think 2013-03-25T11:01:30 there was some patch set for thread affinity that we still did not merge... 2013-03-25T11:01:35 and some bkl type code too 2013-03-25T11:01:42 let me look for it one sec 2013-03-25T11:02:11 https://www.rtems.org/bugzilla/show_bug.cgi?id=1806 2013-03-25T11:02:38 https://www.rtems.org/bugzilla/show_bug.cgi?id=1808 2013-03-25T11:03:02 i don't know why we did not merge these, maybe because the affinity was not done quite right 2013-03-25T11:03:27 this one too I guess is related: https://www.rtems.org/bugzilla/show_bug.cgi?id=1803 2013-03-25T11:04:04 were those patches tested? 2013-03-25T11:05:17 i will do some more look about those link, but now i have to go to bed. see you tomorrom 2013-03-25T11:07:13 i don't think so, except by the contributor 2013-03-25T11:07:15 i just read them 2013-03-25T11:07:21 take care. 2013-03-25T11:07:46 it would be good to see progress on the general fallback solution for the atomic ops :) 2013-03-25T11:10:01 *** antgreen has quit IRC (Ping timeout: 245 seconds) 2013-03-25T11:12:30 *** weiY has quit IRC (Quit: Linkinus - http://linkinus.com) 2013-03-25T11:45:58 *** arvind_k has quit IRC (Ping timeout: 245 seconds) 2013-03-25T11:46:11 *** MegaAlex|away is now known as MegaAlex 2013-03-25T12:45:36 *** antgreen has joined #rtems 2013-03-25T13:17:33 *** Hesham has joined #rtems 2013-03-25T13:25:18 *** Hesham has quit IRC (Ping timeout: 264 seconds) 2013-03-25T14:10:58 *** wilx has joined #rtems 2013-03-25T14:11:05 Hi. 2013-03-25T14:11:40 I am developing a logging library for C++ and I would like to test a cross compilation for RTEMS but I know little to nothing about it. 2013-03-25T14:11:50 Where should I start? 2013-03-25T14:12:06 This is Ubuntu and I can run VirtualBox images, if it helps. 2013-03-25T14:13:15 *** arvind_k has joined #rtems 2013-03-25T14:58:55 *** arvind_k has quit IRC (Read error: Operation timed out) 2013-03-25T15:46:54 *** antgreen has quit IRC (Ping timeout: 264 seconds) 2013-03-25T17:43:25 *** Hesham has joined #rtems 2013-03-25T17:46:36 hi gedare 2013-03-25T18:30:38 hey Hesham 2013-03-25T18:31:09 hi wilx 2013-03-25T18:31:12 seems like you are too busy ;) did you have a chance to see my e-mail ? 2013-03-25T18:32:16 Hesham: no.. i don't see it where did you send it? 2013-03-25T18:32:31 wilx: http://wiki.rtems.org/wiki/index.php/GSoC_Getting_Started is the best place to start... 2013-03-25T18:32:48 it is oriented toward gsoc students but is the best getting started guide we have in general 2013-03-25T18:32:54 you can ask on the rtems-users mailing list for some help too 2013-03-25T18:33:09 gwu.edu 2013-03-25T18:33:13 hmmm 2013-03-25T18:33:18 it hasn't come through yet. odd 2013-03-25T18:33:23 how long ago you sent it? 2013-03-25T18:33:36 weird, that was 10 days ago 2013-03-25T18:33:40 ohhhh 2013-03-25T18:33:44 yes it is buried in my inbox i see it 2013-03-25T18:33:58 i had to shelf it for a bit.. hadn't gotten back to it yet 2013-03-25T18:34:10 you talked to the other student? 2013-03-25T18:34:13 It's ok 2013-03-25T18:34:16 yes 2013-03-25T18:34:25 good.. i think there may be room for two to split up work on this 2013-03-25T18:35:07 he contacted me in private and I suggested he could work on high-level API 2013-03-25T18:35:10 is there a problem with passing the pointer from high to low? 2013-03-25T18:35:22 i mean, can the low level just get a pointer, attribute, and do its thing? 2013-03-25T18:35:42 he sent me some notes and a google doc, preliminary proposal so i cannot share that with you, but... 2013-03-25T18:35:47 i think i sent him an email i can share with you 2013-03-25T18:36:32 Pointer for what ? 2013-03-25T18:36:54 whatever memory region is being dealt with 2013-03-25T18:36:57 I preferred to keep the data structure private to low-level API 2013-03-25T18:38:05 hmm 2013-03-25T18:38:25 I mean the data structure that holds memory entries 2013-03-25T18:38:29 when i looked at arm, it looked possible to recreate the page table entry of interest just from the virtual memory address 2013-03-25T18:39:15 i think the lowest level just needs to deal with virtual memory address ranges and attributes 2013-03-25T18:39:43 the higher level can provide those to it in whatever way is most useful for the services we want to implement 2013-03-25T18:40:15 maybe.. i don't really know for sure 2013-03-25T18:40:31 it would be good to solicit input from rtems-devel mailing list. i think there was some good input from the last round of patches i sent too 2013-03-25T18:41:12 from feedbacks ? 2013-03-25T18:41:30 don't worry about that yet i guess.. just when you start to think you have an idea of what you want to design/implement 2013-03-25T18:41:43 to me it seems the lowest level just needs to know the bare minimum about how to deal with the mm hardware 2013-03-25T18:42:14 Currently I am writing a proposal but it's not complete 2013-03-25T18:42:18 i have to run... but i'll try to find some time to send you an email 2013-03-25T18:42:35 basically i think focus on minimizing the size of the low-level 2013-03-25T18:42:48 so bare minimum of data structures, take advantage of whatever the hw will provide you easily 2013-03-25T18:42:49 I see 2013-03-25T18:43:03 on arm it looked like there is an array of page table entries that gets set up 2013-03-25T18:43:10 yes 2013-03-25T18:43:12 you can index the table by some bits of the vma 2013-03-25T18:43:19 so all you need to get the pte is the vma 2013-03-25T18:43:33 so that tells you what the low-level iface for arm should be like 2013-03-25T18:43:41 for ppc you can do something similar, wit hsome kind of hashing i think it uses 2013-03-25T18:43:48 at kleast for the ppc we're using 2013-03-25T18:43:50 i gotta go.. 2013-03-25T18:43:53 hope that helps 2013-03-25T18:44:03 yes sure, thanks :) 2013-03-25T18:46:02 *** antgreen has joined #rtems 2013-03-25T18:48:02 *** gedare has quit IRC (Read error: Operation timed out) 2013-03-25T19:35:34 *** MegaAlex is now known as MegaAlex|away 2013-03-25T20:00:03 *** the9a3eedi has joined #rtems 2013-03-25T20:00:42 *** Hesham has left #rtems 2013-03-25T20:18:11 Does the RTEMS source have any API support for CANbus devices? 2013-03-25T21:20:00 I found this.. http://www.rtems.org/wiki/index.php/CanFestival .. which makes it seem like RTEMS doesn't have any out-of-the-box support.. so I guess I'll have to get a CAN stack externally from somewhere. I guess CANFestival is a good place to start 2013-03-25T21:20:14 unless anyone has other suggestions? 2013-03-25T21:20:49 I think this is what typically happens. Until someone working woth canbus adds support this situation will not change 2013-03-25T21:22:36 I see.. thanks.. Maybe I should contact the guy who posted that info to see if he has any more details 2013-03-25T21:23:38 Please do. If you can get a suitable solution we would welcome it in RTEMS. 2013-03-25T21:31:55 I'll see how it goes. Alternatively though I can just work directly with the CAN registers on the LEON3 hardware.. Judging from the look of things over here I might have to go with the latter 2013-03-25T21:32:11 Thanks for the info 2013-03-25T21:32:14 Ok 2013-03-25T21:51:58 *** antgreen has quit IRC (Ping timeout: 256 seconds) 2013-03-25T22:15:35 *** antgreen has joined #rtems 2013-03-25T23:56:09 *** antgreen has quit IRC (Ping timeout: 240 seconds) 2013-03-26T00:00:34 *** antgreen has joined #rtems 2013-03-26T00:08:02 *** arvind_k has joined #rtems 2013-03-26T00:50:00 *** arvind_k has quit IRC (Ping timeout: 264 seconds) 2013-03-26T01:52:49 *** arvind_k has joined #rtems 2013-03-26T01:58:47 *** monstr_ has joined #rtems 2013-03-26T02:02:17 Alright. 2013-03-26T02:02:59 The CentOS VM works for me but I cannot update it. If I update it stops booting. It stops after libvirtd service start and does not proceed with the startup any further. 2013-03-26T02:03:34 wilx, what host are you running on ? 2013-03-26T02:03:53 Another thing, my configure script does not detect POSIX threads for sparc-rtems4.11. How do I convince it that there are Pthreads available on RTEMS? 2013-03-26T02:04:11 kiwichris: VirtualBox on Ubuntu (Xubuntu actually) 64 bit. 2013-03-26T02:04:20 OK 2013-03-26T02:04:35 Is this for GSoC ? 2013-03-26T02:04:41 No, unrelated. 2013-03-26T02:04:45 Ok, 2013-03-26T02:04:46 *** the9a3eedi-2 has joined #rtems 2013-03-26T02:04:54 I am testing log4cplus cross compilation to different environments. 2013-03-26T02:05:27 so you are running log4cplus's configure script ? 2013-03-26T02:05:32 Yes. 2013-03-26T02:05:52 Have you built RTEMS ? 2013-03-26T02:06:04 I have run the ./bootstrap script. 2013-03-26T02:06:43 You need to then build with a suitable --prefix and install. For example I use --prefix=$HOME/development/rtems/build/4.11 2013-03-26T02:07:09 *** the9a3eedi has quit IRC (Ping timeout: 240 seconds) 2013-03-26T02:07:17 You then need to let the configure script know this path. Typically this is via suitable CFLAGS on the configure command line 2013-03-26T02:07:20 Really? I thought that the CentOS VM that was available was like pre-installed with stuff. 2013-03-26T02:07:47 ie CFLAGS='-I /prefix path' 2013-03-26T02:08:07 I am not 100% sure what is on the VM. It may have a built RTEMS image. 2013-03-26T02:08:17 The log4cplus configure script detects stuff OK but it then fails to find a way to provide pthread_create() etc. 2013-03-26T02:09:12 No many configure scripts are good at configuring for cross-compiling. For example they may try and execute a test which fails for RTEMS. 2013-03-26T02:09:29 This is a link test only. 2013-03-26T02:09:42 How would you compile RTEMS app that is using Pthreads? 2013-03-26T02:10:01 Great. Next you need to let configure know the paths to the RTEMS libraries that contain this function 2013-03-26T02:10:32 Almost the same way, but I would need to add the specific target flags for the board support package you are linking for 2013-03-26T02:10:41 ie -Lpath -lrtems etc 2013-03-26T02:11:26 RTEMS supports many different runtime targets, while Linux on x86_64 is just the single set up therefore gcc knows these things by default. 2013-03-26T02:14:32 How does Xubuntu differ from Ubuntu ? 2013-03-26T02:15:15 kiwichris: The X11 shell that is preinstalled is Xfce instead of Gnome. Nothing else. 2013-03-26T02:15:37 Ok, well then for Xubuntu tools you could try ?. 2013-03-26T02:15:49 http://www.rtems.org/ftp/pub/rtems/people/chrisj/source-builder/source-builder.html 2013-03-26T02:16:25 This should build the tools directly for your host. 2013-03-26T02:16:43 If you use 4.10 it will also build and install RTEMS. 2013-03-26T02:55:46 *** wilx has quit IRC (Quit: Lost terminal) 2013-03-26T03:02:24 *** sebhub has joined #rtems 2013-03-26T03:03:27 good morning 2013-03-26T03:38:23 * kiwichris sees wilx is the maintainer for log3cplus 2013-03-26T03:39:46 * kiwichris looked over http://log4cplus.sourceforge.net/ and it looks interesting 2013-03-26T03:58:32 *** rtemsLogger has joined #rtems 2013-03-26T04:06:24 *** rtemsLogger` has quit IRC (*.net *.split) 2013-03-26T04:09:09 *** the9a3eedi-2 has quit IRC (Ping timeout: 245 seconds) 2013-03-26T04:32:57 *** MegaAlex|away is now known as MegaAlex 2013-03-26T05:17:36 no, never had this kind of error 2013-03-26T05:22:54 *** the9a3eedi has joined #rtems 2013-03-26T05:41:51 *** MegaAlex is now known as MegaAlex|away 2013-03-26T06:01:26 sebhub, ok, Alan is getting it on a clean VM install on Ubuntu but his other VM which is older does not show it 2013-03-26T06:01:38 cya 2013-03-26T06:28:29 *** arvind_k is now known as arvind_khadri 2013-03-26T09:03:55 *** DrJoel has joined #rtems 2013-03-26T09:03:55 *** DrJoel has joined #rtems 2013-03-26T09:03:55 *** ChanServ sets mode: +o DrJoel 2013-03-26T09:05:11 sebhub.. am I caught up on your gcc patches? 2013-03-26T09:15:16 *** dhananjay has joined #rtems 2013-03-26T09:15:59 Hi. 2013-03-26T09:16:38 kiwichris the other day, my connection was disrupted and the conversation was not reaching me. 2013-03-26T09:24:30 kiwichris is in Sydney Australia so should be sound asleep. 2013-03-26T09:24:44 I can take a stab at your questions 2013-03-26T09:28:12 DrJoel: Thanks. we were chatting up a bit about the GDB project he proposed 2013-03-26T09:28:34 then my connection was lost. Anyway - I will sent him a mail. 2013-03-26T09:28:35 Which one? :) 2013-03-26T09:29:07 Provide rtems specific extension to gdb, using python 2013-03-26T09:29:25 Ahh.. that is a good project. And needed :) 2013-03-26T09:30:57 I am interested. :) But the docs are little rough. I am reading up the pretty printers in gdb tree now 2013-03-26T09:34:04 Internal type docs are almost always painful. Ask questions on the rtems and/or gdb list. The gdb list is probably better :) 2013-03-26T09:34:13 I know I on on that list. Chris I think is also. 2013-03-26T09:36:06 *** dhananjay has quit IRC (Ping timeout: 272 seconds) 2013-03-26T09:44:39 hello joel 2013-03-26T09:44:54 this -pthread option for gcc helps to avoid some test suite failures 2013-03-26T09:45:35 Ok.. that is in the queue.. what about your v2 testsuite patch to sort. is that ok now? 2013-03-26T09:45:55 yes, sorry to not send you the updated version 2013-03-26T09:46:16 no problem. Just trying to keep track of what I need to do as we push forward. 2013-03-26T09:46:25 Do you want this on the 4.8 branch as well as the head? 2013-03-26T09:46:44 yes, 4.8 and head 2013-03-26T09:46:46 OK 2013-03-26T09:47:07 FYI did you know Michele now has a book on Amazon. :) 2013-03-26T09:47:27 Should pop out on the Nook this week. I am working on the conversion for the Apple Store 2013-03-26T09:48:33 nice 2013-03-26T09:50:43 I have learned a lot through the process 2013-03-26T09:51:04 http://www.amazon.de/Rose-Her-Own-Michele-Pendelton/dp/1482569167 2013-03-26T09:51:17 Notice the .de :) 2013-03-26T09:51:39 Are there any other patches from you or issues I need to deal with? 2013-03-26T09:52:26 not for gcc 2013-03-26T09:52:40 only the newlib thread cancellation patch 2013-03-26T09:52:58 and that has to settle a bit more, right? 2013-03-26T09:53:33 we should evaluate what we do with the locks 2013-03-26T09:54:48 agreed 2013-03-26T10:10:21 *** the9a3eedi has quit IRC (Ping timeout: 252 seconds) 2013-03-26T10:14:48 http://gcc.gnu.org/ml/gcc-patches/2013-03/msg00970.html 2013-03-26T10:15:12 this bug is fixed since last october and nobody cares to commit it 2013-03-26T10:15:16 Thanks for the URL.. It is on my list after figuring out how to deal with the speeding ticket I got last week. :( 2013-03-26T10:15:56 http://sourceware.org/ml/newlib/2013/msg00139.html 2013-03-26T10:16:01 this is the newlib patch 2013-03-26T10:16:21 Is the newlib patch ok by Corrinna to commit? 2013-03-26T10:17:09 http://sourceware.org/ml/newlib/2013/msg00144.html 2013-03-26T10:18:10 ok.. in the queue as well. I hope to get to this today. 2013-03-26T10:19:52 *** weiY has joined #rtems 2013-03-26T10:32:29 arm gcc tests look pretty good: 2013-03-26T10:32:31 http://gcc.gnu.org/ml/gcc-testresults/2013-03/msg02743.html 2013-03-26T10:33:01 cu 2013-03-26T10:33:12 *** sebhub has quit IRC (Remote host closed the connection) 2013-03-26T10:36:59 *** weiY has quit IRC (Quit: Linkinus - http://linkinus.com) 2013-03-26T10:40:32 *** dhananjay has joined #rtems 2013-03-26T11:09:43 *** gedare has joined #rtems 2013-03-26T11:09:57 hey DrJoel 2013-03-26T11:10:11 hey gedare.. got marshmallows for the fire? 2013-03-26T11:10:23 haven't sat down on my log yet 2013-03-26T11:10:37 just got done with inbox, moving to rtems labels :) 2013-03-26T11:10:47 I just committed a couple of sebastian patches. newlib cancellation and gcc -pthread option 2013-03-26T11:11:18 ahh 2013-03-26T11:11:52 i'm not on the gcc patches list, 2013-03-26T11:11:59 pinging on a gcc PowerPC PR.. 2013-03-26T11:12:03 i just joined newlib 3 weeks ago 2013-03-26T11:12:04 pushing.. one of those was newlib 2013-03-26T11:12:22 That and a speeding ticket are on my todo list today 2013-03-26T11:12:23 :( 2013-03-26T11:14:35 how much over what? 2013-03-26T11:14:50 rather how fast over what limit ":) 2013-03-26T11:15:17 do BSPs need to do something particular to propagate exceptions to signals... I don't know that I ever saw such a thing 2013-03-26T11:16:01 when you compile with lpthread, doesn't it make main like a pthread? 2013-03-26T11:16:02 We don't talk about it much but if you get a FP exception, it can be turned into a SIGFPE, 2013-03-26T11:16:11 mm, i see 2013-03-26T11:16:16 Depends on the initialization task table. 2013-03-26T11:16:31 The book I cited specifically says "hardware originated signals are directed to the executing thread" 2013-03-26T11:16:44 but I can't find that in the web version of POSIX. 2013-03-26T11:17:13 I know the Ada run-time catches SIGFPE and turns them into language exceptions. I suspect C++, Go, and Java do the same 2013-03-26T11:17:41 it's probably a good thing to support 2013-03-26T11:18:04 that kind of thing would be hard to test 2013-03-26T11:18:39 i might expect sigfpe returns to my executing task, but maybe under some situation in the field, due to multithreading not present during testing, an FP exception might happen in one task and get directed to another I suppose 2013-03-26T11:19:08 what's the point of sending these tool release versions to the devel ml? 2013-03-26T11:20:15 you see sebhubs commit today? is this TLS something gcc supports or is it standard any clue? 2013-03-26T11:20:37 GCC supports TLS 2013-03-26T11:21:21 i see 2013-03-26T11:21:27 SIGFPE and other hardware exceptions are something to discuss. Mostly they are handled in a non-uniform manner. It probably should be an application configure time option to fault or generate a signal 2013-03-26T11:22:01 yes, and some more framework in BSP / CPU parts would be useful 2013-03-26T11:22:08 i had no idea this is even an issue until now 2013-03-26T11:22:11 The mix of task/thread APIs means the POSIX behavior may or may not be appropriate for Classic API tasks. I think Sebastian has tripped across long standing behavior that Classic API tasks ignore signals. 2013-03-26T11:22:26 which, considering i added a whole new architecture, is somewhat disturbing ;) 2013-03-26T11:22:39 It is a good thing to discuss after so many years and come up with a plan 2013-03-26T11:22:57 i put it in a list of things to discuss later :) 2013-03-26T11:24:30 i think our open projects are getting a little lean, for the runtime oriented 2013-03-26T11:26:36 cindy has a lot of...energy 2013-03-26T11:27:06 i'm hopeful she can be a co-mentor this year 2013-03-26T11:27:21 i suggested she might be eligible as a student, but her work plate is full 2013-03-26T11:29:45 one or two more good pushes and i think the posix keys will be done 2013-03-26T11:29:55 i need to double check the posix spec to make sure the approach i suggested is valid 2013-03-26T11:31:19 This would be a good GSOC project. Some targets already have the exception handlers. Most of the gdb simulators can handle this. Having a uniform behavior, etc would be a fair amount of work. And cross target 2013-03-26T11:31:35 i was thinking that, but was not sure if it is "ready" to suggest or not 2013-03-26T11:31:45 if you don't care i'll add it as an idea if i find some spare minutes. 2013-03-26T11:32:02 I just emailed her. I want to voice chat with her soon and find out what role she would like to take. She was a high school math teacher before getting into CS. SHe has a unique background. 2013-03-26T11:32:13 neat 2013-03-26T11:33:19 Well.. there are multiple parts: (1) per architecture code to catch the exceptions in BSPs, (2) regular fatal exception handler, (3) signal generator, and (4) configuratino magic to select between them at an applicaiton level 2013-03-26T11:33:49 I think a lot of code is in place. SPARC can generate signals for gnat. 2013-03-26T11:34:01 1 should exist. sebastian put in something like 2 but i'm not sure how deeply he integreated it 2013-03-26T11:34:07 3 and 4 do not afaik exist 2013-03-26T11:34:29 (3) exists for some architectures in some form 2013-03-26T11:34:33 *nod* 2013-03-26T11:34:44 but.. does it need to be arch-dependent? 2013-03-26T11:34:58 seems like that 3 and 4 can be platform-independent to me 2013-03-26T11:35:03 as long as 1 is done right 2013-03-26T11:35:07 see sparc/shared/gnatcommon.c 2013-03-26T11:35:09 k 2013-03-26T11:36:13 Sebastian did a lot of the framework. I don't know that there is a clean handler in every architecture. And he didn't map them to signals. 2013-03-26T11:36:18 this could easily be made independent. 2013-03-26T11:36:34 I think it would challenge a student even if it is light in places 2013-03-26T11:37:07 the only arch-specific stuff is vector #... how do PIC provide the exception source? 2013-03-26T11:37:17 ok 2013-03-26T11:37:31 there is not enough detail yet for a student to make a good proposal 2013-03-26T11:38:09 Yeah. we need to work on that :) 2013-03-26T11:49:36 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2013-03-26T12:17:35 gedare.. yet another git question 2013-03-26T12:17:53 I have I microblaze branch in my personal repo. 2013-03-26T12:18:19 I don't seem to be able to do a simple git push anymore even though I have rebased it 2013-03-26T12:18:38 what is the error 2013-03-26T12:18:39 $ git status 2013-03-26T12:18:39 # On branch microblaze 2013-03-26T12:18:39 # Your branch and 'origin/microblaze' have diverged, 2013-03-26T12:18:39 # and have 11 and 5 different commit(s) each, respectively. 2013-03-26T12:18:39 # 2013-03-26T12:18:39 nothing to commit (working directory clean) 2013-03-26T12:18:39 [joel@localhost rtems]$ git push --dry-run origin microblaze 2013-03-26T12:18:40 To ssh://joel@git.rtems.org/home/joel/git/rtems.git 2013-03-26T12:18:41 ! [rejected] microblaze -> microblaze (non-fast-forward) 2013-03-26T12:18:42 error: failed to push some refs to 'ssh://joel@git.rtems.org/home/joel/git/rtems.git' 2013-03-26T12:18:42 To prevent you from losing history, non-fast-forward updates were rejected 2013-03-26T12:18:42 Merge the remote changes before pushing again. See the 'Note about 2013-03-26T12:18:43 fast-forwards' section of 'git push --help' for details. 2013-03-26T12:18:59 $ git pull --rebase upstream master 2013-03-26T12:18:59 remote: Counting objects: 19, done. 2013-03-26T12:18:59 remote: Compressing objects: 100% (10/10), done. 2013-03-26T12:18:59 remote: Total 10 (delta 8), reused 0 (delta 0) 2013-03-26T12:18:59 Unpacking objects: 100% (10/10), done. 2013-03-26T12:19:00 From ssh://git.rtems.org/data/git/rtems 2013-03-26T12:19:00 * branch master -> FETCH_HEAD 2013-03-26T12:19:00 First, rewinding head to replay your work on top of it... 2013-03-26T12:19:01 Applying: Microblaze initial import of "stub" port 2013-03-26T12:19:02 Applying: microblaze/cpu.h: Add comments 2013-03-26T12:19:02 Applying: xdr_float.c: Add Microblaze 2013-03-26T12:19:03 Applying: Microblaze: Add _CPU_Exception_frame_print 2013-03-26T12:19:03 Applying: score/microblaze: Rename files to be unique in tree 2013-03-26T12:19:15 So it is rebased OK I think.. just can't push and get it in sync with origin 2013-03-26T12:20:12 um 2013-03-26T12:20:20 hmm 2013-03-26T12:21:05 git's great when it works :) 2013-03-26T12:21:20 so when you push now it still rejects non -ff? 2013-03-26T12:22:21 git push --dry-run origin microblaze 2013-03-26T12:22:24 Yep 2013-03-26T12:23:14 git status? 2013-03-26T12:23:36 git pull --rebase upstream microbranch 2013-03-26T12:23:42 you rebased the wrong branch 2013-03-26T12:23:50 $ git checkout microblaze 2013-03-26T12:23:50 Switched to branch 'microblaze' 2013-03-26T12:23:50 Your branch and 'origin/microblaze' have diverged, 2013-03-26T12:23:50 and have 12 and 5 different commit(s) each, respectively. 2013-03-26T12:25:11 grrr... ok 2013-03-26T12:25:14 now all is ok 2013-03-26T12:25:25 this is hard to keep my head around 2013-03-26T12:26:34 thank you 2013-03-26T12:27:06 yep 2013-03-26T12:27:21 git is easy..when you keep your branches straight 2013-03-26T12:27:22 :) 2013-03-26T12:28:29 *** MegaAlex|away is now known as MegaAlex 2013-03-26T12:51:14 *** monstr_ has quit IRC (Ping timeout: 245 seconds) 2013-03-26T12:57:50 *** gedare has quit IRC (Remote host closed the connection) 2013-03-26T13:01:56 *** dhananjay has quit IRC (Ping timeout: 272 seconds) 2013-03-26T13:50:47 *** dhananjay has joined #rtems 2013-03-26T14:38:27 *** dhananjay has quit IRC (Remote host closed the connection) 2013-03-26T14:38:42 hi 2013-03-26T14:39:58 *** arvind_khadri has joined #rtems 2013-03-26T14:51:59 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2013-03-26T15:25:14 *** DrJoel has quit IRC (Ping timeout: 246 seconds) 2013-03-26T15:49:46 *** gedare has joined #rtems 2013-03-26T16:33:17 *** MegaAlex is now known as MegaAlex|away 2013-03-26T17:05:58 *** DrJoel has joined #rtems 2013-03-26T17:05:58 *** DrJoel has joined #rtems 2013-03-26T17:05:58 *** ChanServ sets mode: +o DrJoel 2013-03-26T18:28:57 Does any understand in a shell "if [ -z ${foo+xxx} ];" does ? What does the expansion '${foo+xxx}' mean ? 2013-03-26T18:30:35 Where's that from? I can't make anything in the shell come out using HOME and PATH or xxx 2013-03-26T18:31:01 FYI my fee for the ticket and to go to driving school was $184 .. that's a lot of beer 2013-03-26T18:32:25 kiwichris what? the ${a+b}? 2013-03-26T18:32:29 b is used if a is not set 2013-03-26T18:33:51 Look at a stackoverflow post about a portable way to handle an empty or not present variable in a shell 2013-03-26T18:33:58 Look -> looking 2013-03-26T18:34:14 $ echo ${XXX+PATH} 2013-03-26T18:34:14 bash-4.1$ echo ${PATH+XXX} 2013-03-26T18:34:14 XXX 2013-03-26T18:34:32 That seems to show it in action 2013-03-26T18:34:33 Hmm 2013-03-26T18:34:47 kiwichris -z "${a}" should work fine 2013-03-26T18:34:59 Yeah I agree it should. 2013-03-26T18:35:01 sometimes people do the hack 2013-03-26T18:35:11 test "a" = "${var}"a 2013-03-26T18:35:16 or 2013-03-26T18:35:17 test "a" = "${var}a" 2013-03-26T18:35:33 one or 2 '=' ? 2013-03-26T18:35:36 because some verions of test can't handle empty strings, this causes the evaulation to be a=a if it's empty 2013-03-26T18:35:57 Hmm one, ouch that I keep forgetting and use 2, ie == 2013-03-26T18:36:06 use -eq 2013-03-26T18:36:07 verm__ that's what I learned years ago "X${x}" = "X" 2013-03-26T18:36:21 DrJoel: yeah, lots of weird hacks :( 2013-03-26T18:36:39 it's not required on any modern (newer than 10 years) unix though 2013-03-26T18:36:48 will do 2013-03-26T18:36:58 so it's broken on FreeBSD .. as he runs.. lol 2013-03-26T18:37:09 No just centos 2013-03-26T18:37:21 honestly I learned UNIX from Western Electric manuals so a lot of old tricks are still ingrained 2013-03-26T18:37:34 We welcome all operating systems in RTEMS these days :) 2013-03-26T18:38:07 I wonder if the old Yggdrasil distro CDs are around 2013-03-26T18:38:25 I have one with a 0.99 linux kernel 2013-03-26T18:38:32 ~1994 if I recall correctly. I actually bought a SCSI 3x or 4x CD-ROM for it 2013-03-26T18:38:41 I think older than that 2013-03-26T18:38:53 Did python exist then ? 2013-03-26T18:39:00 this is supposed to be the first CD distro.. 2013-03-26T18:39:07 I think it may be older than python 2013-03-26T18:40:22 fixed the vm issue for the log4cplus guy. :) 2013-03-26T18:40:24 bbl 2013-03-26T18:44:44 Found it, Yggdrasil Summer 1994 2013-03-26T18:45:03 The orig cd, but do not have the wrapper any more :( 2013-03-26T18:45:17 Need 8M of RAM 2013-03-26T18:54:56 Does not boot in a VM. :( 2013-03-26T18:58:14 probably need something like bochs 2013-03-26T18:58:24 or an old version of qemu 2013-03-26T18:58:51 Or I need to extract a floppy disk image and boot from that !! 2013-03-26T18:58:59 Or boot from freedos 2013-03-26T19:13:43 *** the9a3eedi has joined #rtems 2013-03-26T19:17:06 DrJoel: i like monday in case you did not see my email. 2013-03-26T19:17:33 I think seb has a problem with that day 2013-03-26T19:17:49 But he also has a problem with the time so it may not matter 2013-03-26T19:19:33 yeah 2013-03-26T19:19:36 gedare, phyiscal journals vs logical journals for RFS on flash ? physical better ? 2013-03-26T19:19:44 I tend to think so. 2013-03-26T19:19:44 we can try to alternate some times but there are not many that work well 2013-03-26T19:20:03 i don't think i know the differense 2013-03-26T19:20:12 logical = in ram? 2013-03-26T19:20:13 Yeah with daylight saving kicking in the times have become difficult 2013-03-26T19:20:18 physical = in flash? 2013-03-26T19:20:44 physical is the block differences, logical is the metadata level 2013-03-26T19:20:49 ahhh 2013-03-26T19:20:54 physical is seen as safer 2013-03-26T19:20:57 so physical is a lot larger...? 2013-03-26T19:21:19 logical would embed semantic knowledge i guess? 2013-03-26T19:21:26 Maybe, if the log is a diff it might not be, ie offset, len, data 2013-03-26T19:21:47 i'm not much of an fs person... i have read some about journaling fs's but mostly to learn about their data structures. 2013-03-26T19:21:50 physical could sit in the bdbuf layer and support any file system 2013-03-26T19:22:06 yeah less duplication of effort 2013-03-26T19:22:07 *** the9a3eedi-2 has joined #rtems 2013-03-26T19:22:13 what is the downside to physical then? 2013-03-26T19:22:18 *** the9a3eedi-2 has quit IRC (Client Quit) 2013-03-26T19:22:42 Just take a section of storage somewhere and record the sequence, confirm it is written then update the fs's media 2013-03-26T19:22:50 Considered slower 2013-03-26T19:23:05 ah 2013-03-26T19:23:09 logical is more complex in the way you capture the changes and then replay 2013-03-26T19:23:19 file system specific 2013-03-26T19:23:23 right 2013-03-26T19:23:54 i guess a physical journal could be implemented just as a subtree of bdbuf blocks? 2013-03-26T19:23:55 I think physical is a good first pass to implement and then if needed a logical can be implement and used. 2013-03-26T19:23:58 yes 2013-03-26T19:24:12 i agree with this based on what you told me... but i may make a mental note to try reading up on it myself 2013-03-26T19:24:19 since what you told me might bias me toward your conclusion too :) 2013-03-26T19:24:38 Yes and you optionally add a device interface to bypass any special driver code, ie flash driver 2013-03-26T19:25:04 Yeah I have corrupted you !! ;) 2013-03-26T19:25:51 *** the9a3eedi has quit IRC (Ping timeout: 252 seconds) 2013-03-26T19:35:39 kiwichris: bdbuf uses an avl tree? 2013-03-26T19:39:51 would it make sense to replace with an rbtree? 2013-03-26T19:40:05 i know.. it ain't broke, don't fix it ;0 2013-03-26T19:41:01 does it make sense to bypass the flash driver and write directly to the flash for the journal? 2013-03-26T19:41:24 that seems like it would need some extra care to me... maybe 2013-03-26T19:41:33 again though, i'm not an fs or disk person 2013-03-26T19:43:50 i think you are right about phys vs log journals... we should provide phys and if someone wants to implement/use logical caveat emptor 2013-03-26T19:44:26 And get that implemented. 2013-03-26T19:44:48 write it up as an open project? 2013-03-26T19:45:41 The bypass means the user and driver need to be set up and it puts the journal where they want, ie battey backed up RAM, special section of flash eyc 2013-03-26T19:45:58 Someone is interested in this. 2013-03-26T19:46:05 oh, ok 2013-03-26T19:46:09 even better :) 2013-03-26T19:46:28 i thought you just had a brainstorm for some reason wanted to see journaling available 2013-03-26T19:46:44 I want a good solution but also a pratical and not large change. 2013-03-26T19:47:14 Yeah AVL tree. We could change this or make it configurable but it is working so has not changed. 2013-03-26T19:47:42 *** the9a3eedi has joined #rtems 2013-03-26T19:47:54 by the way... it might be efficient to implement a journal using a functional data structure... one that keeps around old copies, like how RCU locking is done 2013-03-26T19:47:55 i have no idea, haven't read up on them ;) 2013-03-26T19:48:12 yeah exactly..if it ain't broke. 2013-03-26T19:48:14 links on RCU ? 2013-03-26T19:48:24 http://en.wikipedia.org/wiki/Read-copy-update 2013-03-26T19:50:57 I think journal is a sequence so you do not overwrite in the journal, that happens in the file system copy 2013-03-26T19:52:53 *** gedare has quit IRC (Ping timeout: 248 seconds) 2013-03-26T19:56:08 *** antgreen has quit IRC (Ping timeout: 272 seconds) 2013-03-26T19:56:27 Hi. I've been looking at generated RTEMS executable size. When viewing the actual .exe file size, it's from 2-6MB (which is fairly large). When using the sparc-rtems-size command, the executable size is between 100-500 kilobytes. what's taking up the rest of the space? 2013-03-26T19:56:58 The ELF data such as symbols etc 2013-03-26T19:57:29 Try 'file rtems-exe' and see if the exe has been stripped. 2013-03-26T19:58:12 it hasn't been stripped 2013-03-26T19:58:30 I wasn't expecting ELF data to be this big though 2013-03-26T19:58:49 It needs to know a lot to track the source cleanly 2013-03-26T19:59:08 None of this is needed on the target, and is only of use on the host. 2013-03-26T19:59:45 When put on the target it is normally stripped and if your target loads elf just run the strip the command on it. 2013-03-26T19:59:45 I see.. but say I wanted to flash the rom with the file.. how may I strip the .exe from the "extra fat"? 2013-03-26T20:00:01 oh ok 2013-03-26T20:00:02 arm-rtems4.11-strip 2013-03-26T20:00:19 or what ever arch you use 2013-03-26T20:01:04 Got it. thanks! 2013-03-26T20:10:30 *** antgreen has joined #rtems 2013-03-26T20:38:45 *** antgreen has quit IRC (Ping timeout: 248 seconds) 2013-03-26T20:46:12 *** antgreen has joined #rtems 2013-03-26T20:56:09 *** DrJoel has quit IRC (Quit: REALITY.SYS Corrupted: Re-boot universe? (Y/N/Q)) 2013-03-26T20:56:16 *** antgreen has quit IRC (Ping timeout: 245 seconds) 2013-03-26T20:59:29 *** antgreen has joined #rtems 2013-03-27T01:33:48 *** monstr_ has joined #rtems 2013-03-27T01:44:36 *** arvind_khadri has joined #rtems 2013-03-27T04:05:33 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-03-27T05:52:23 *** freenix has joined #rtems 2013-03-27T07:56:16 *** antgreen has quit IRC (Ping timeout: 245 seconds) 2013-03-27T09:08:07 *** freenix has quit IRC (Quit: Leaving) 2013-03-27T09:14:51 *** weiY has joined #rtems 2013-03-27T09:49:31 *** gedare has joined #rtems 2013-03-27T10:46:59 *** DrJoel has joined #rtems 2013-03-27T10:46:59 *** DrJoel has joined #rtems 2013-03-27T10:46:59 *** ChanServ sets mode: +o DrJoel 2013-03-27T10:47:05 morning DrJoel 2013-03-27T10:47:13 hey.. 2013-03-27T10:47:29 Do you think we finished the GSOC application properly? 2013-03-27T10:47:32 *** weiY has quit IRC (Quit: Linkinus - http://linkinus.com) 2013-03-27T10:47:36 I was feeling the need to double check. :) 2013-03-27T10:47:46 pfft, i'm in "fire and forget" mode these days 2013-03-27T10:48:06 maybe when i can get back to more than 5 hours of sleep a night 2013-03-27T10:48:19 It's been a long 2 weeks :) 2013-03-27T10:48:21 lol.. I'm just in the forget mode :( 2013-03-27T10:48:52 Yes, well, you are, after all, over the Federally-protected age for agism 2013-03-27T10:49:10 I had to do some mandated harassment training 2013-03-27T10:49:16 mandatory? 2013-03-27T10:49:26 Nothing I did to earn it, just rqeuirements to stay employed 2013-03-27T10:49:31 Which I found kind of harassing. 2013-03-27T10:49:58 Damn! Now I feel old. And discriminated against. Do you have any money my ambulance chasing lawyer should target? LOL 2013-03-27T10:50:34 Years ago, some of the government folks I knew had to take a class on phone usage and manners. And another group went to "credit card school" 2013-03-27T10:51:02 hahah 2013-03-27T10:51:07 I had to take a human trafficing webinar as required training for one customer to keep access to their facilities. 2013-03-27T10:51:17 interesting 2013-03-27T10:51:42 trafficking is disturbing. 2013-03-27T10:51:48 enough to make me want to low-jack my dd 2013-03-27T10:51:51 ;-) 2013-03-27T10:51:52 Honestly the travel safety web class I have taken has good general advice. 2013-03-27T10:52:03 i'm sure 2013-03-27T10:52:05 just put a tracker in the diaper 2013-03-27T10:52:12 lol 2013-03-27T10:52:18 the paci 2013-03-27T10:52:25 that won't get lost!!! 2013-03-27T11:52:04 *** antgreen has joined #rtems 2013-03-27T11:54:26 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2013-03-27T12:15:33 *** antgreen has quit IRC (Remote host closed the connection) 2013-03-27T12:19:36 *** monstr_ has quit IRC (Ping timeout: 245 seconds) 2013-03-27T12:53:58 looks like i'm outvoted on the mondays anyway 2013-03-27T12:54:19 It doesn't matter much to me.. At some point, every day of the week sucks. :) 2013-03-27T12:54:41 Next week I will be in Dayton Ohio and wouldn't be able to participate 2013-03-27T12:55:19 that's ok we don't need weekly or even bi-weekly 2013-03-27T12:55:31 my mom goes there for work sometimes. 2013-03-27T12:55:38 I was hoping every 2-3 weeks 2013-03-27T13:00:20 that is sensible to me 2013-03-27T13:00:37 I just hope we can push at 4.11.. it needs cutting soon 2013-03-27T13:00:42 but the tools are the big issue at this point 2013-03-27T13:03:06 as soon as the code mergers are done, we can cut the release. i think all we really need to do is create a new branch at a particular point 2013-03-27T13:03:15 we could use tags... but i think that would be better for dot releases 2013-03-27T13:03:50 once the code is merged we can do testing and minor fixes commit to both head and the .11 branch 2013-03-27T13:03:58 Agreed.. even though it sounds like cvs, it is right. :) 2013-03-27T13:04:01 and continue with active development on head 2013-03-27T13:04:07 Right 2013-03-27T13:04:30 it will match better with what has been done in the past, adn with how the .9 and .10 are handled right now 2013-03-27T13:04:40 Yep. 2013-03-27T13:04:50 I need to look into cutting releases from git 2013-03-27T13:05:04 i'd like to get the posix keys done just to have it dealt with 2013-03-27T13:05:09 but i don't want to block the release for it 2013-03-27T13:05:18 I don't think it will end up doing that. 2013-03-27T13:05:43 i think creating a new branch is the right way, but haven't looked closely 2013-03-27T13:05:58 i think linux uses tags, but i don't know how they deal with dot releases then 2013-03-27T13:06:08 probably they branch the maintained releasses 2013-03-27T13:07:36 https://www.kernel.org/pub//software/scm/git/docs/git-archive.html looks to be the starting point 2013-03-27T13:08:39 For us, likely a starting point.. update versions in tree, tag, make a tar file, untar it, bootstrap it.. re-tar it.. I also untar and make the docs 2013-03-27T13:17:55 *** Hesham has joined #rtems 2013-03-27T13:23:57 yeah 2013-03-27T13:23:59 that makes sense 2013-03-27T13:24:48 i suggest sending this in a quick email to rtems-devel 2013-03-27T13:25:11 Let me play a bit. I think I can actually hack this together pretty quickly once I get a chance. :) 2013-03-27T13:25:16 k 2013-03-27T13:25:20 still.. send an email 2013-03-27T13:25:41 Well only Amar had offered to help and he doesn't track the list. :) 2013-03-27T13:25:49 When I get some more details fleshed out, I will 2013-03-27T13:25:49 for the record. 2013-03-27T13:26:00 and, it hasn't been discussed much on devel ml yet 2013-03-27T13:26:03 might be others interested 2013-03-27T13:26:16 or even send to rtems-users 2013-03-27T13:26:18 yeah.. let me try some steps by hand and make sure I am not offbase. :) 2013-03-27T13:26:21 yep 2013-03-27T13:26:22 rtems-users is more than appropriate 2013-03-27T13:26:34 yeah. would be good to get a little feedback if anyone wants to 2013-03-27T13:27:37 FYI I updated the VM we gave to the high school student who interned here a couple of weeks ago. Nothing special. Now it is 600 Mb larger. I need to check again what I did. I think I did some magic to re-layout the disk blocks. 2013-03-27T13:28:07 MAKE IT SMALLER 2013-03-27T13:28:29 I was removing packages.. it actually had gimp in it. That's bad 2013-03-27T13:29:21 I thought of putting a script on it they should run when they first log in -- it could do the clones, tool builds, etc. Would take about an hour to run on my machine 2013-03-27T13:29:32 But wouldn't save that much disk space 2013-03-27T13:32:18 Hi gedare & DrJoel, I have a question... 2013-03-27T13:32:24 what's up? 2013-03-27T13:32:28 Is it ok to include cache operations ( Cache manager) at libmm ? 2013-03-27T13:32:41 you mean to use them? 2013-03-27T13:33:01 it should be fine, in fact i think it would be necessary 2013-03-27T13:33:12 I mean moving the existing Cache code to the new libmm 2013-03-27T13:33:13 and... if the cache op you need is not implemented, you should feel free to provide it :) 2013-03-27T13:33:17 ahhhhh 2013-03-27T13:33:26 i'm not sure. 2013-03-27T13:33:36 Yes that's exactly a part of my proposal :) 2013-03-27T13:33:51 they're related, but not so tightly that it makes sense to make them the same thing 2013-03-27T13:33:57 at least, not to me. 2013-03-27T13:34:16 i think it is better to keep them modular, unless you have some good reason or motivation 2013-03-27T13:34:50 DrJoel, you should start with a barebones VM, like a stripped down CentOS or Ubuntu Server 2013-03-27T13:35:09 I have this idea because I think MMU/Caches relate to memory management 2013-03-27T13:35:18 ah, i see. 2013-03-27T13:35:24 it might be sensible 2013-03-27T13:35:52 but we have to avoid introducing much overhead for CPUs that do not need (or want) the MMU support 2013-03-27T13:36:43 We could have libmm as namespace and keep cache/MMU as separate modules inside 2013-03-27T13:37:25 Yes by mean of configuration header 2013-03-27T13:37:30 Ok 2013-03-27T13:37:42 I think it could make sense. 2013-03-27T13:37:48 Especially because cache becomes more complex on SMP 2013-03-27T13:37:56 and is more complicated when there is an MMU 2013-03-27T13:38:19 you can also think if there is some better way to design the libmm. But! 2013-03-27T13:38:30 note that the cache manager functionality is necessary inside score 2013-03-27T13:38:38 so libmm probably needs to have some visibility there. 2013-03-27T13:39:08 I think that is the case 2013-03-27T13:39:08 I see 2013-03-27T13:39:20 you'll have to read a little more on that to be sure 2013-03-27T13:39:30 but... fleshing out cache manager support is useful too 2013-03-27T13:39:36 not enough for a project by itself 2013-03-27T13:39:42 but put together with MMU support it is a good idea 2013-03-27T13:39:51 you have convinced me anyway 2013-03-27T13:39:57 :) 2013-03-27T13:39:57 just need to convince everyone else now 2013-03-27T13:40:39 * gedare pokes DrJoel 2013-03-27T13:40:44 Read more about what ? Caches/MMU or what ? 2013-03-27T13:40:50 both! 2013-03-27T13:40:59 but really, how the cache manager is integrated right now 2013-03-27T13:41:10 if yo'ure going to include refactoring it 2013-03-27T13:41:17 you have to be sure you know how it works and is used :) 2013-03-27T13:41:30 and be able to test ti 2013-03-27T13:41:31 it 2013-03-27T13:42:13 I will read more about how Caches/MMU and even SMP relate to each other... 2013-03-27T13:43:17 Anyway I have a prototype and a draft of my proposal, is it ok to share it with you ? or wait to make it clearer ? 2013-03-27T13:46:40 it's ok to share i will read it when i have some time 2013-03-27T13:47:10 Please share it. 2013-03-27T13:47:35 At SoC page ? 2013-03-27T13:47:37 The only issue with the cache code being integrated with the MMU code is that there may be CPUs in a family with the cache instructions but not the MMU. 2013-03-27T13:47:49 Hesham: Google Doc at this point. 2013-03-27T13:47:59 just share with me and DrJoel 2013-03-27T13:48:01 Keep your proposal from being completely open 2013-03-27T13:48:25 ok 2013-03-27T13:48:26 I can see there being cases where the cache/mmu are hard to pull apart though. 2013-03-27T13:48:34 is there a way to make a group for gdoc? be nice to have an RTEMS Mentors group to share docs 2013-03-27T13:48:53 ok.. i have to go 2013-03-27T13:49:04 I have no idea on that gedare .. I couldn't even figure out how to add a set of people as a contact in gmail. I know I have done this before. :( 2013-03-27T13:54:43 *** gedare has quit IRC (Ping timeout: 245 seconds) 2013-03-27T13:56:35 *** antgreen has joined #rtems 2013-03-27T14:30:24 *** DrJoel has quit IRC (Quit: ASCII a stupid question, get a stupid ANSI!) 2013-03-27T16:04:13 *** dhananjay has joined #rtems 2013-03-27T16:04:25 *** Hesham has left #rtems 2013-03-27T16:04:50 Hi 2013-03-27T16:36:04 *** antgreen has quit IRC (Remote host closed the connection) 2013-03-27T16:38:44 *** dhananjay has quit IRC (Ping timeout: 272 seconds) 2013-03-27T18:11:17 *** antgreen has joined #rtems 2013-03-27T21:26:38 *** the9a3eedi has joined #rtems 2013-03-27T22:05:15 *** dhananjay has joined #rtems 2013-03-27T22:06:48 *** the9a3eedi has quit IRC (Ping timeout: 272 seconds) 2013-03-27T22:35:18 *** dhananjay has quit IRC (Ping timeout: 272 seconds) 2013-03-27T22:47:46 *** dhananjay has joined #rtems 2013-03-27T23:17:40 *** the9a3eedi has joined #rtems 2013-03-27T23:42:53 Hi guys 2013-03-27T23:56:31 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-03-28T00:13:40 *** kiwichris has joined #rtems 2013-03-28T00:33:19 *** the9a3eedi has quit IRC (Ping timeout: 245 seconds) 2013-03-28T00:47:59 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-03-28T01:12:12 *** kiwichris has joined #rtems 2013-03-28T01:33:36 *** monstr_ has joined #rtems 2013-03-28T01:58:50 *** dhananjay has quit IRC (Ping timeout: 258 seconds) 2013-03-28T02:08:17 *** arvind_khadri has joined #rtems 2013-03-28T02:15:25 *** dhananjay has joined #rtems 2013-03-28T02:27:08 *** the9a3eedi has joined #rtems 2013-03-28T02:42:42 *** sebhub has joined #rtems 2013-03-28T02:43:05 good morning 2013-03-28T02:47:19 *** weiY has joined #rtems 2013-03-28T03:44:22 *** dhananjay has quit IRC (Ping timeout: 272 seconds) 2013-03-28T04:12:19 *** the9a3eedi has quit IRC (Ping timeout: 246 seconds) 2013-03-28T04:18:41 *** dhananjay has joined #rtems 2013-03-28T05:34:14 *** dhananjay has quit IRC (Ping timeout: 246 seconds) 2013-03-28T06:11:12 *** weiY has quit IRC (Quit: Leaving...) 2013-03-28T06:23:44 *** verm__ has quit IRC (Ping timeout: 255 seconds) 2013-03-28T06:23:44 *** lcpfnvcy has quit IRC (Ping timeout: 255 seconds) 2013-03-28T06:23:46 *** lcpfnvcy_ has joined #rtems 2013-03-28T06:28:30 *** sevikkk has quit IRC (*.net *.split) 2013-03-28T06:28:31 *** peerst has quit IRC (*.net *.split) 2013-03-28T06:31:35 *** sevikkk has joined #rtems 2013-03-28T06:31:35 *** peerst has joined #rtems 2013-03-28T06:44:02 *** the9a3eedi has joined #rtems 2013-03-28T07:56:14 *** gedare has joined #rtems 2013-03-28T08:18:18 *** verm__ has joined #rtems 2013-03-28T08:39:20 *** verm__ has quit IRC (Read error: Connection reset by peer) 2013-03-28T08:41:30 *** verm__ has joined #rtems 2013-03-28T08:47:53 *** weiY has joined #rtems 2013-03-28T08:50:05 *** weiY has quit IRC (Read error: Connection reset by peer) 2013-03-28T08:50:12 *** weiY has joined #rtems 2013-03-28T09:06:21 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2013-03-28T09:09:51 *** arvind_khadri has joined #rtems 2013-03-28T09:24:05 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2013-03-28T09:48:51 Hi 2013-03-28T09:52:49 *** sebhub has quit IRC (Remote host closed the connection) 2013-03-28T10:29:39 Hi, gedare. 2013-03-28T10:29:48 how's it going weiY? 2013-03-28T10:30:21 i had another thought for a project idea for you... 2013-03-28T10:30:27 to implement some kind of SMP locks 2013-03-28T10:30:37 I am looking up every architecture rtems supports 2013-03-28T10:30:38 these would probably be using the atomic ops 2013-03-28T10:30:42 so that would be pretty good 2013-03-28T10:30:51 for what? 2013-03-28T10:31:13 yeah, i also have similiar idea last year 2013-03-28T10:31:29 but i have no enough time to implement last year 2013-03-28T10:31:53 some sync primitives based on atomic ops 2013-03-28T10:32:55 i look up the architecture supported by rtems for general atomic ops which do not have atomic instructions 2013-03-28T10:33:58 i want to summary which architecutures do not have atomic instructions 2013-03-28T10:34:39 ah. that might be good to put in the wiki somewhere too 2013-03-28T10:35:17 the atomic ops might be implemented on CPU families with an architecture though, or some variants of an architecture 2013-03-28T10:35:23 So it is not just the architecture level 2013-03-28T10:35:48 for example, the sparc v8 does not specify atomic ops, but the leon family implements the casa instructions specified in sparc v9. 2013-03-28T10:36:06 although i don't know if leon uses 32-bit or 64-bit casa. i suspect only 32 bit 2013-03-28T10:37:32 yeah, this will be more difficult if one architecture has different atomic instruction support 2013-03-28T10:38:43 but powerpc and x86 just implement the same among all the cpu families 2013-03-28T10:38:51 yes i think most will be the same 2013-03-28T10:39:01 I just know off-hand that the sparc32 is a special case for RTEMS 2013-03-28T10:39:09 because of what leon3 does 2013-03-28T10:39:18 *** the9a3eedi has quit IRC (Ping timeout: 252 seconds) 2013-03-28T10:39:30 however... i don't know if we even care about other sparc32 CPUs anymore, other than the Leon3 (and maybe Leon4 some day) 2013-03-28T10:39:57 anyway, i think a table like what you are talking about will be quite useful 2013-03-28T10:40:07 i can reference the implementation of freebsd and linux to see whether they support different cpu families 2013-03-28T10:40:08 where are you looking for your info? 2013-03-28T10:40:48 yes, i think that would be good.. and you can add as a column in your table at least whether one of the BSD implements the atomic ops, since that gives a place to borrow code 2013-03-28T10:41:10 http://wiki.rtems.org/wiki/index.php/SupportedCPUs 2013-03-28T10:41:40 i just find this link , but i do not know whether this is outdate? 2013-03-28T10:43:26 so first step i must know how many CPU familieres are supporting now 2013-03-28T10:45:15 and then summary them to a table about their atomic instruction information and implementation reference 2013-03-28T10:49:01 gedare, I have to go to bed, see you later, take care. 2013-03-28T10:49:05 bye 2013-03-28T10:49:51 i think this table is messy 2013-03-28T10:50:08 ok, i have to check the code . haha 2013-03-28T10:50:11 it's reasonably up to date though 2013-03-28T10:50:20 yeah, really 2013-03-28T11:03:53 *** weiY has quit IRC (Quit: Linkinus - http://linkinus.com) 2013-03-28T12:13:52 *** DrJoel has joined #rtems 2013-03-28T12:13:52 *** DrJoel has joined #rtems 2013-03-28T12:13:52 *** ChanServ sets mode: +o DrJoel 2013-03-28T12:45:49 *** monstr_ has quit IRC (Read error: Operation timed out) 2013-03-28T13:24:58 hi DrJoel 2013-03-28T13:25:06 hey.. 2013-03-28T13:25:10 pm please :) 2013-03-28T13:26:07 hit me up on gchat.. moving computers for a telecon 2013-03-28T13:26:12 I hope all is OK. 2013-03-28T13:26:23 k, it is 2013-03-28T13:30:25 :) 2013-03-28T13:30:34 rebooting telecon computer.. sigh.. webex is so fragile 2013-03-28T15:44:25 kiwichris awake? 2013-03-28T15:45:00 yeah, and heading out soon to play tennis on this good friday 2013-03-28T15:45:14 it is, isn't it 2013-03-28T15:45:25 there is an option --no-clean but the default appears to leave the build directories 2013-03-28T15:45:26 well, not yet here, but soon. 2013-03-28T15:45:27 how may I help ? 2013-03-28T15:46:00 The dirs are left if the build fails and there an almost empty dir I need to chase down 2013-03-28T15:46:02 I am just cycling over all the targets on CentOS 6.4 32-bit and want it to clean the build directory between targets. Otherwise the VM runs out 2013-03-28T15:46:11 It is empty just a little is left 2013-03-28T15:46:47 Cleaning the dirs on fail makes debugging harder :) 2013-03-28T15:47:08 * gedare meets his deadline 2013-03-28T15:47:23 just in time to enjoy the rest of the evening 2013-03-28T15:47:25 hold on 2013-03-28T15:48:29 Hmm.. I just don't have a lot of disk space :( 2013-03-28T15:48:38 $ ls build/ 2013-03-28T15:48:39 arm-rtems4.11-binutils-2.23.1-1-2.23.1-1 avr-rtems4.11-gcc-4.7.2-newlib-2.0.0-1-4.7.2-1 tmp 2013-03-28T15:48:39 arm-rtems4.11-gcc-4.7.2-newlib-2.0.0-1-4.7.2-1 expat-2.1.0-i686-linux-gnu-1-2.1.0-1 2013-03-28T15:48:39 avr-rtems4.11-binutils-2.23.1-1-2.23.1-1 lm32-rtems4.11-binutils-2.23.1-1-2.23.1-1 2013-03-28T15:48:39 no clean is set to 0 in defaults.py ? "'no-clean' : '0'," 2013-03-28T15:48:54 They are fails 2013-03-28T15:49:00 *** sevikkk1 has joined #rtems 2013-03-28T15:49:12 You need to clean up after a fail. Not much I can do here. 2013-03-28T15:49:14 probably a couple of legit failures and others when disk space got low 2013-03-28T15:49:38 Maybe for this sweep, I should just have the wrapper script remove build in between targets 2013-03-28T15:50:08 Are you using keep going ? 2013-03-28T15:50:20 Yes.. just wanting to see pass/fail 2013-03-28T15:50:42 Ah ok then, this is a valid use case and I think --always-clean is needed 2013-03-28T15:51:15 so --keep-going --always-clean would be the way. I will look at adding this option when I can 2013-03-28T15:51:38 need to go cya 2013-03-28T15:51:39 *** sevikkk has quit IRC (Read error: Operation timed out) 2013-03-28T15:51:48 no problem or hurry... we need to define what "acceptance criteria" we are shooting for 2013-03-28T15:54:18 *** gedare has quit IRC (Quit: Leaving) 2013-03-28T16:00:09 wonder how this will turn out for the source builder 2013-03-28T16:00:21 Linux gcc1-power7.osuosl.org 3.7.2-204.fc18.ppc64 #1 SMP Fri Jan 18 10:44:51 MST 2013 ppc64 ppc64 ppc64 GNU/Linux 2013-03-28T17:02:23 64 cores.. building rtems tools just fine so far 2013-03-28T17:07:00 [joel@gcc1-power7 b-arm7tdmi]$ gdbarmsim -i `find . -name ticker.exe` arm-rtems4.11-run is /home/joel/rtems/tools/bin/arm-rtems4.11-run 2013-03-28T17:07:00 *** CLOCK TICK TEST *** 2013-03-28T17:07:00 TA1 - rtems_clock_get_tod - 09:00:00 12/31/1988 2013-03-28T17:07:00 TA2 - rtems_clock_get_tod - 09:00:00 12/31/1988 2013-03-28T17:07:00 TA3 - rtems_clock_get_tod - 09:00:00 12/31/1988 2013-03-28T17:07:00 TA1 - rtems_clock_get_tod - 09:00:05 12/31/1988 2013-03-28T17:07:01 TA2 - rtems_clock_get_tod - 09:00:10 12/31/1988 2013-03-28T17:07:01 TA1 - rtems_clock_get_tod - 09:00:10 12/31/1988 2013-03-28T17:07:02 TA3 - rtems_clock_get_tod - 09:00:15 12/31/1988 2013-03-28T17:07:02 TA1 - rtems_clock_get_tod - 09:00:15 12/31/1988 2013-03-28T17:07:03 TA2 - rtems_clock_get_tod - 09:00:20 12/31/1988 2013-03-28T17:07:03 TA1 - rtems_clock_get_tod - 09:00:20 12/31/1988 2013-03-28T17:07:04 TA1 - rtems_clock_get_tod - 09:00:25 12/31/1988 2013-03-28T17:07:05 TA3 - rtems_clock_get_tod - 09:00:30 12/31/1988 2013-03-28T17:07:05 TA1 - rtems_clock_get_tod - 09:00:30 12/31/1988 2013-03-28T17:07:05 TA2 - rtems_clock_get_tod - 09:00:30 12/31/1988 2013-03-28T17:07:05 *** END OF CLOCK TICK TEST *** 2013-03-28T17:07:06 [joel@gcc1-power7 b-arm7tdmi]$ uname -a 2013-03-28T17:07:06 Linux gcc1-power7.osuosl.org 3.7.2-204.fc18.ppc64 #1 SMP Fri Jan 18 10:44:51 MST 2013 ppc64 ppc64 ppc64 GNU/Linux 2013-03-28T17:07:26 I'm all alone.. all alone.. no one here beside me.. 2013-03-28T17:20:14 Clearly not only is there no one who knows the musical Spamalot, I am truly all alone 2013-03-28T17:25:51 sparc/sis successful on the same host 2013-03-28T17:39:12 *** DrJoel has quit IRC (Quit: If you can't laugh at yourself, make fun of other people.) 2013-03-28T19:10:53 *** the9a3eedi has joined #rtems 2013-03-28T19:52:29 * kiwichris likes Joel's results on that machine ? wonder how long it took to build 2013-03-28T20:11:44 What sort of encoding does RTEMS use for the console? I'm having some trouble reading the output from the UART. Text is all garbled up. I tried UTF-8 and Latin-1 2013-03-28T20:12:36 No encoding. This happens when the baud rate is wrong. This can be the wrong rate or the settings do not match the clock rate 2013-03-28T20:13:31 I see... interesting.. didn't think the baud has something to do with it 2013-03-28T20:15:33 actually, I've been using grmon with leon3, and there's an option to have it set the console output to the same serial line used for debugging.. and that works but is horrendously slow, which I thought was because it was sharing the same line as the debug line 2013-03-28T20:17:33 I do not know anything about gmin 2013-03-28T20:17:36 gmon 2013-03-28T20:18:47 well.. the baud rate still turned out to be wrong.. it was actually 38400 and not 115200. thanks! 2013-03-28T20:19:05 working now 2013-03-28T20:19:08 no problems 2013-03-28T22:36:35 *** dhananjay has joined #rtems 2013-03-28T22:41:15 *** dhananjay has quit IRC (Ping timeout: 260 seconds) 2013-03-28T22:44:39 *** dhananjay has joined #rtems 2013-03-28T22:48:53 *** dhananjay has quit IRC (Ping timeout: 248 seconds) 2013-03-28T22:52:44 *** dhananjay has joined #rtems 2013-03-28T23:05:36 *** dhananjay has quit IRC (Ping timeout: 264 seconds) 2013-03-28T23:23:46 *** Fallenou has quit IRC (Ping timeout: 256 seconds) 2013-03-28T23:23:53 *** Fallenou has joined #rtems 2013-03-29T00:50:41 *** arvind_khadri has joined #rtems 2013-03-29T00:52:00 *** arvind_khadri has joined #rtems 2013-03-29T00:52:38 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2013-03-29T00:53:10 *** arvind_khadri has joined #rtems 2013-03-29T00:54:29 *** arvind_khadri has joined #rtems 2013-03-29T00:56:00 *** arvind_khadri has joined #rtems 2013-03-29T01:06:48 *** arvind_khadri has quit IRC (Ping timeout: 264 seconds) 2013-03-29T01:11:50 *** monstr_ has joined #rtems 2013-03-29T03:44:26 *** dhananjay has joined #rtems 2013-03-29T04:07:41 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-03-29T04:55:24 *** monstr_ has quit IRC (Ping timeout: 264 seconds) 2013-03-29T05:06:29 *** dhananjay has quit IRC (Ping timeout: 255 seconds) 2013-03-29T06:31:34 *** MegaAlex|away is now known as MegaAlex 2013-03-29T08:09:54 *** freenix has joined #rtems 2013-03-29T08:37:44 *** dhananjay has joined #rtems 2013-03-29T09:56:10 *** dhananjay has quit IRC (Ping timeout: 258 seconds) 2013-03-29T10:07:42 *** DrJoel has joined #rtems 2013-03-29T10:07:42 *** DrJoel has joined #rtems 2013-03-29T10:07:42 *** ChanServ sets mode: +o DrJoel 2013-03-29T10:58:19 *** dhananjay has joined #rtems 2013-03-29T11:16:48 *** S_Somani has joined #rtems 2013-03-29T11:18:58 Hi, I downloaded the RTEMS vmware image from https://sourceforge.net/projects/rtems-vms/files/. But CentOS seems to hang on boot on opening my first time (without installing virtual box guest OS additions). 2013-03-29T11:27:56 sorry for being asleep at the keyboard 2013-03-29T11:28:12 does it hang at libvrtd? 2013-03-29T11:28:17 libvirtd? 2013-03-29T11:28:25 When you do an F1 while booting to see detail? 2013-03-29T11:29:07 nothing seems to work except ctrl+alt+del 2013-03-29T11:29:11 during that time 2013-03-29T11:29:29 Press F1 as the progress bar starts .. it provides more detail 2013-03-29T11:30:52 after libvirtd daemon 2013-03-29T11:30:59 it hangs at virt-who 2013-03-29T11:31:08 ok that's close enough 2013-03-29T11:31:16 still a virtualbox guest extension problem 2013-03-29T11:31:48 I started writing instructions on this. Let's see if they work and can be improved. :) 2013-03-29T11:31:49 http://wiki.rtems.org/wiki/index.php/Virtual_Machines_for_RTEMS_Development#System_Hangs_at_libvirtd_During_Boot 2013-03-29T11:31:51 umm ok...but i havent installed virtualbox guest extensions yet 2013-03-29T11:32:01 They were installed before you got the VM 2013-03-29T11:32:25 Maybe I shouldn't leave them installed when I upload the VMs. 2013-03-29T11:32:30 What host are you on? 2013-03-29T11:32:43 windows 2013-03-29T11:32:52 ok .. that's the same as me 2013-03-29T11:33:07 visit that wiki page and let's walk you through fixing it as the wiki fails :) 2013-03-29T11:34:11 so i should reboot and go to the command line 2013-03-29T11:34:35 basically .. at init 3 2013-03-29T11:35:00 and help fix the instructions in the wiki. I wrote them from memory and I know they suck 2013-03-29T11:37:09 umm so to go to init 3 i have to press ctrl alt f2? 2013-03-29T11:38:03 reset the VM.. I think you press "e" to edit the boot line for the kernel. add a "3" 2013-03-29T11:38:19 then work your way out of the grub edit system and press "b" to boot to init 3 2013-03-29T11:39:34 *** S_Somani1 has joined #rtems 2013-03-29T11:40:12 down arrow, "e" to edit that entry, down arrow to the kernel line, "e" to edit that line, " 3" to add init level 3, return, and "b" to boot to init 3 2013-03-29T11:40:54 just did it. That should be very close but pay attention to make sure the screen hints that what I said is correct. :) 2013-03-29T11:42:21 *** S_Somani has quit IRC (Ping timeout: 258 seconds) 2013-03-29T11:42:58 sorry my net got disconnected due to some problem 2013-03-29T11:43:06 no problem 2013-03-29T11:43:19 I didn't hit return on the line until you were back 2013-03-29T11:44:44 i entered the kernel thingy 2013-03-29T11:45:07 i now have to write 3 ahead of "rd_NO_DM"? 2013-03-29T11:46:51 *** gedare has joined #rtems 2013-03-29T11:49:24 hello 2013-03-29T11:51:06 S_Somani1 at the end of the line 2013-03-29T11:51:09 and press b 2013-03-29T11:51:12 hey gedare 2013-03-29T11:51:30 this is supposed to be a day off but I am on irc as I listen to a telecon on my call at home. sigh 2013-03-29T11:51:31 hey DrJoel having a good friday? 2013-03-29T11:51:36 ahhh 2013-03-29T11:52:03 today is a work day for me, but i had a deadline yesterday, so it is more of a thinking day 2013-03-29T11:53:47 S_Somani1: any luck 2013-03-29T11:54:26 gedare we are walking him through the virtualbox extension problem where it is incompatible after a kernel upgrade. http://wiki.rtems.org/wiki/index.php/Virtual_Machines_for_RTEMS_Development#System_Hangs_at_libvirtd_During_Boot 2013-03-29T11:54:32 And trying to improve the docs 2013-03-29T11:54:35 ya worked 2013-03-29T11:54:46 i do not have experience with the problem. 2013-03-29T11:55:12 :D 2013-03-29T11:55:19 you should just not install guest extensions in your image 2013-03-29T11:55:26 S_Somani1: can you make sure the wiki works for the next poor soul who trips over this. :) 2013-03-29T11:55:40 Yes.. that's a starting point but if YOU install them and do a kernel upgrade, then it is still broken 2013-03-29T11:55:46 ok 2013-03-29T11:55:51 fiar enough 2013-03-29T11:55:52 Okay :) 2013-03-29T11:56:04 I am going to build a new VM from scratch and won't ship with them installed. 2013-03-29T11:56:06 brb 2013-03-29T11:56:56 DrJoel, make your vm SMALLER 2013-03-29T11:56:58 ;-) 2013-03-29T12:14:49 DrJoel, i had an idea for a tinyRTEMS project. it requires some investigation though. Idea: Replace doubly-linked chains with singly-linked lists in cases the back-pointers are not needed 2013-03-29T12:14:54 gedare: trying 2013-03-29T12:15:22 i'm not sure if it can be applied anywhere... since removal from singly linked list is icky if not from the front 2013-03-29T12:15:26 but watchdog chain might be one case 2013-03-29T12:15:35 Do you have any cases where they aren't needed? The scheduling/blocking assumes you can extract in the middle easily 2013-03-29T12:15:55 Cancel when your blocking is satisfied is an arbitrary extract 2013-03-29T12:16:30 i haven't looked at it, just had the thought the other day. 2013-03-29T12:16:36 while looking at bsd queues 2013-03-29T12:17:06 I would like to see -ffunction-sections (is that the option) in use on all targets 2013-03-29T12:17:26 Plus a BSD-style initialization so the init routines of all managers aren't linked in 2013-03-29T12:17:49 And an attempt to use avr-libc for avr-rtems. :) 2013-03-29T12:18:10 i'm not familiar with ffunction-sections or what you mean by it 2013-03-29T12:18:19 it is vaguely familiar though 2013-03-29T12:19:10 is the init issue supposed to be solved by the sequenced initializatoin project? 2013-03-29T12:19:24 i think cindy took a brief look at the avr-libc... i hav enot though 2013-03-29T12:19:39 but, i do have an arduino uno 2013-03-29T12:19:44 which is a 16-bit avr iirc 2013-03-29T12:21:40 yes on sequenced init project.. Chris prototyped it and that was close for Classic API 2013-03-29T12:22:12 avr-libc would just be smaller and better optimized and more appropriate for target. Plus a good easy first step to try something other than newlib on a small target 2013-03-29T12:22:23 you can "bold it" oln the open projects if you are very interested 2013-03-29T12:23:02 :) Need to do some work on that page when I get done helping Jordan 2013-03-29T12:25:12 *** MegaAlex is now known as MegaAlex|away 2013-03-29T12:38:22 ordered my regalia..got raked over the coals for it 2013-03-29T12:38:33 it was like the cost of a wedding dress! 2013-03-29T12:38:55 lol.. yes it was expensive.. how much these days? 2013-03-29T12:39:26 at least you can't get divorced from a PhD. :) 2013-03-29T12:39:34 Wow! Now I sound like a cynic 2013-03-29T12:41:05 haha indeed... just over 700 2013-03-29T12:41:46 I recall it being about $400-$500 in 1999. I have used my Masters robe as a wizards costume since them so it was money well spent. 2013-03-29T12:42:39 nice 2013-03-29T12:42:44 lol 2013-03-29T12:42:46 brb 2013-03-29T12:42:56 they threw the hood in for me.. they give me the hood at the hooding ceremony 2013-03-29T12:43:06 master's grads have to purchase their hoods though! 2013-03-29T12:43:16 at least that's what i understood 2013-03-29T13:03:13 i think the newest student that is considering an mmu project may be reading leaked vxworks code 2013-03-29T13:03:20 have to keep an eye on that :/ 2013-03-29T13:03:50 S_Somani1, I have confirmed your wiki account 2013-03-29T13:05:02 gedare: Thanks. I'll edit the wiki now. :) 2013-03-29T13:08:38 ouch.. we can't let any of that code in. 2013-03-29T13:08:51 Did you slap their wrist? 2013-03-29T13:21:57 gedare I know you are swamped.. but do you want to be part of brad testing of the RSB? 2013-03-29T13:40:34 what testing of what? 2013-03-29T13:40:48 rsb=rtems source builder, ok got that... 2013-03-29T13:40:53 what is brad testing? 2013-03-29T13:41:22 broad 2013-03-29T13:41:31 using GCC Compile Farm for variety 2013-03-29T13:42:21 ah.. i can only say maybe 2013-03-29T13:43:08 if you even want to think about it.. you may want to go ahead and get an account on the farm 2013-03-29T13:45:30 are we happy with the gsoc app.. only 15 minutes left 2013-03-29T13:45:31 there a form for that? 2013-03-29T13:47:53 it was fine the last time i saw it 2013-03-29T13:53:33 I just checked again.. popped into #gsoc to watch as the bell tolls.. usually entertaining 2013-03-29T13:57:52 heh. 2013-03-29T13:57:58 i avoid that place 2013-03-29T14:01:57 deadline has passed.. so far no whining.. just a few people expecting someone to whine. :) 2013-03-29T14:09:09 *** dhananjay has quit IRC (Ping timeout: 258 seconds) 2013-03-29T14:34:27 *** gedare has quit IRC (Ping timeout: 258 seconds) 2013-03-29T14:55:18 *** gedare has joined #rtems 2013-03-29T15:29:38 What is the source directory for gcc in RTEMS vm? I'm getting this error - checking for MPFR... no configure: error: libmpfr not found or uses a different ABI. 2013-03-29T15:30:00 Hmmm... 2013-03-29T15:30:11 what are you building? And how? 2013-03-29T15:31:15 I was building mpc-0.8.1 and then i got this error. So i thought it was probably because of the old files. 2013-03-29T15:31:26 so did an upgrade and tried again 2013-03-29T15:31:32 Why are you building that? 2013-03-29T15:31:58 The RTEMS cross tools should be in /opt/rtems-4.11 and be up to date. The CentOS VMs to date have used the RPMs for tools. 2013-03-29T15:32:35 because it was written in the wiki that i need to build it the first time I launch 2013-03-29T15:33:33 ok :) 2013-03-29T15:34:04 ouch! touche.. where in the wiki.. hopefully that doesn't happen on the VM pages 2013-03-29T15:34:05 :( 2013-03-29T15:34:16 Do you have tools in /opt/rtems4.11/bin 2013-03-29T15:35:01 yes i got the location of gcc 2013-03-29T15:35:28 I think this is what is happening with me http://stackoverflow.com/questions/9297933/cannot-configure-gcc-mpfr-not-found 2013-03-29T15:36:43 Oh.. that has nothing to do with an RTEMS toolset... it can happen when you build them but you don't need to 2013-03-29T15:37:04 *** gedare has quit IRC (Quit: Leaving) 2013-03-29T15:37:47 umm..didnt quite get you 2013-03-29T15:38:46 You don't have to build a gcc for rtems using this vm. It is already there. 2013-03-29T15:40:03 ok. But how to get mpfr required for mpc? 2013-03-29T15:40:16 type sparc-rtems4.11-gcc 2013-03-29T15:40:30 what error started you wanting it? 2013-03-29T15:41:07 cd ../b-mpc ../mpc-0.8.1/configure && make && make install 2013-03-29T15:41:25 ^ sorry 2013-03-29T15:41:36 But before that.. why did you want mpc? 2013-03-29T15:42:11 the VM is supposed to have everything you need to build RTEMS for the sparc BSP sis and run tests on the simulator in gdb without building any tools from source 2013-03-29T15:42:37 Somewhere you took a wrong turn.. and I think it was early. What were you trying to do? 2013-03-29T15:43:01 I wanted it because it was written in the wiki. 2013-03-29T15:43:31 It was written the first time you run the virtual environment 2013-03-29T15:43:52 log in as root and do an update and build gmp,mpc 2013-03-29T15:43:55 So you tried a yum upgrade and it failed? 2013-03-29T15:44:02 I don't have the VM with me this weekend. 2013-03-29T15:44:31 My upgrade was successfull 2013-03-29T15:44:34 You don't need to build gmp and mpc. If it failed during the upgrade, we are misisng a yum repository. But I haven't seen it failed. 2013-03-29T15:44:35 but http://wiki.rtems.org/wiki/index.php/Virtual_Machines_for_RTEMS_Development 2013-03-29T15:44:43 oh okay 2013-03-29T15:45:11 ok.. now cut and paste the original error :) 2013-03-29T15:45:28 I have built all the VMs so shoould be able to help you fix it. 2013-03-29T15:45:45 checking for MPFR... no configure: error: libmpfr not found or uses a different ABI. 2013-03-29T15:46:14 crap.. that's not necessary.. I wonder who added that. Damn it! 2013-03-29T15:46:34 Even if it is necessary it isn't the right way to do it. FYI I didn't write that :) 2013-03-29T15:46:48 OK.. so you completed the "yum upgrade" successfully 2013-03-29T15:46:49 haha 2013-03-29T15:46:52 DrJoel: ps -I'm new to open source you are the most helpful person i have met in the open source community so far. 2013-03-29T15:48:02 It is Easter Weekend so light duty and it is also spring break for a lot of us. I know my sons at university are off this week 2013-03-29T15:48:28 But I am the project lead (if that's the term) and one of the original developers .. > 20 years working on rtems in some way. 2013-03-29T15:48:38 OK.. let's go back to the root problem and then we can fix the wiki 2013-03-29T15:48:51 Did the yum upgrade complete OK 2013-03-29T15:48:56 yup 2013-03-29T15:49:01 can you do this? 2013-03-29T15:49:34 what? fix the wiki? 2013-03-29T15:49:53 [joel@localhost rtems-4.11-work]$ sparc-rtems4.11-gcc -o m m.c 2013-03-29T15:49:53 [joel@localhost rtems-4.11-work]$ cat m.c 2013-03-29T15:49:53 int main() {return 0;} 2013-03-29T15:50:04 bad cut and paste. 2013-03-29T15:50:17 Let's see if the sparc-rtems4.11-gcc compiler works that is in /opt/rtems4.11-bin 2013-03-29T15:50:19 echo $PATH 2013-03-29T15:50:26 type sparc-rtems4.11-gcc 2013-03-29T15:53:52 still there? 2013-03-29T15:54:20 ya 2013-03-29T15:54:37 u want me to copy paste from 2:19:54 right? 2013-03-29T15:54:46 Enter the command: "echo $PATH" 2013-03-29T15:55:03 Nah.. let's take it slower .. the cut and paste messed things up.. some lines didn't paste 2013-03-29T15:55:20 oh ok 2013-03-29T15:55:35 You want /opt/rtems-4.11/bin at the front of your PATH 2013-03-29T15:55:49 kk 2013-03-29T15:56:19 ya i have a path 2013-03-29T15:56:26 should i copy paste it 2013-03-29T15:56:27 ? 2013-03-29T15:56:32 yes.. just to be sure 2013-03-29T15:57:17 sorry (a newbie question) there is no cursor inside...how do i copy paste? :p 2013-03-29T15:58:34 Highlight and either to copy to paste.. or just use the Edit/Copy on the menu bar 2013-03-29T15:58:40 Assuming you are using xchat in the VM 2013-03-29T16:00:49 /usr/lib/qt-3.3/bin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin 2013-03-29T16:01:21 There should be a file in the directory rtems-4.11-work which will help you here. Is "setenv" in that directory 2013-03-29T16:01:27 It is short for "set environment" 2013-03-29T16:01:41 Enter the command ". ~/rtems-4.11-work/setenv" 2013-03-29T16:01:48 Then look at your PATH again 2013-03-29T16:06:13 actually i dont have a directory named rtems-4.11-work 2013-03-29T16:06:22 is sparc rtems4.11 2013-03-29T16:06:25 the same? 2013-03-29T16:06:27 pwd 2013-03-29T16:06:55 ls ${HOME} 2013-03-29T16:07:21 If you logged in as rtems, it is in your ${HOME} unless you removed it 2013-03-29T16:07:56 i was logged in as root 2013-03-29T16:08:13 my bad 2013-03-29T16:10:36 the rtems account should be where you live 99% of the time 2013-03-29T16:11:56 kk :) 2013-03-29T16:15:23 It is hanged again at libvirtd. though i'm pretty sure vmware guest was uninstalled cleanly 2013-03-29T16:15:58 sigh 2013-03-29T16:16:15 well you know the trick with adding the 3. :-D 2013-03-29T16:16:32 We can manually reinstall the extensions at that level. I had to do that once 2013-03-29T16:17:05 rechecking 2013-03-29T16:18:22 http://askubuntu.com/questions/22743/how-do-i-install-guest-additions-in-virtualbox is for Ubuntu but the basics are the same if you don't get to the GUI 2013-03-29T16:18:36 if there isn't a /dev/cdrom, it is /dev/sr0 for the virtual cd rom 2013-03-29T16:18:38 brb 2013-03-29T16:18:45 kk 2013-03-29T16:27:09 keep chatting.. I am helping son with something and will walk by periodically 2013-03-29T16:28:26 it shows that you must specify the filesystem type in mount 2013-03-29T16:41:55 make sure on Virtualbox that "Devices" > "CD/DVD Devices" has the extensions iso checked 2013-03-29T16:42:07 Then (as root) mount /dev/sr0 /mnt 2013-03-29T16:42:38 That should work. 2013-03-29T16:42:40 # mount /dev/cdrom /mnt 2013-03-29T16:42:40 mount: block device /dev/sr0 is write-protected, mounting read-only 2013-03-29T16:42:51 From my virtualbox session 2013-03-29T16:51:50 when i go to edit->preferences->devices->cd/dvd (there is no such option for unchecking an iso) 2013-03-29T16:52:19 *checking 2013-03-29T16:54:33 it is checked in mine 2013-03-29T16:54:50 VBoxGuestAdditions.iso 2013-03-29T16:56:52 DrJoel, what broke on python 2.5.x ? 2013-03-29T16:58:41 it threw an exception.. hold on and I will cut and paste 2013-03-29T16:59:00 interestingly this machine failed with the ARM and I think it is SPARC64. So it is not host size 2013-03-29T17:00:58 * kiwichris almost has MinGW tools built under cygwin with the RSB 2013-03-29T17:01:05 C:\opt\rtems\src\rsb\rtems>.\build\tmp\source-build-chris\4.10\rtems-arm\cygdriv 2013-03-29T17:01:06 e\c\Users\chris\development\rtems\4.10\bin\arm-rtems4.10-gcc --version 2013-03-29T17:01:06 arm-rtems4.10-gcc (GCC) 4.4.7 20120313 (RTEMS 4.10-RSB(1e1ea1e1a292fddb58e79cbda 2013-03-29T17:01:06 6feaaffcb9d23b2-modified)-1,gcc-4.4.7/newlib-1.18.0) 2013-03-29T17:01:06 Copyright (C) 2010 Free Software Foundation, Inc. 2013-03-29T17:01:06 This is free software; see the source for copying conditions. There is NO 2013-03-29T17:01:08 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. 2013-03-29T17:03:15 Hmm.. I have a build in process.. I will have to work to recreate. let me get back to you on that 2013-03-29T17:03:16 brb 2013-03-29T17:03:17 Joel, I suspect the python version has other issues with old hosts. Currently the RSB GDB build script assumes Python is present to link with GDB. 2013-03-29T17:03:32 I don't doubt that. I will recreate it. 2013-03-29T17:04:06 If there is no suitable Python to link with is the gdb built without Python support ? 2013-03-29T17:04:35 In the case of MinGW built on Cygwin do I need to built Python ? I was hoping to avoid doing this. 2013-03-29T17:15:13 I am guessing that the only reasonable solution. Either that or not build at all. Will gdb autonatically disable Python support or fail 2013-03-29T17:15:20 S_Somani1.. any luck? 2013-03-29T17:16:15 Yup mounted and installed... 2013-03-29T17:16:53 I thought it would end my woes, but on rebooting it showed that vmware guest additons not present... :/ 2013-03-29T17:19:29 do you mean not installed? or not able to mount 2013-03-29T17:19:44 Chris.. what from the host environment could negatively impact the arm builds? 2013-03-29T17:21:14 I dont know what went wrong. It mounted correctly and then I ran ./VBoxLinuxAdditions-x86.run 2013-03-29T17:22:20 It now shows Virtualbox guest PCI device not found 2013-03-29T17:31:10 wow.. that's a new one. :( 2013-03-29T17:31:17 Do you still have the .ova file you downloaded? 2013-03-29T17:31:26 Can you just mentally reset and reimport it? 2013-03-29T17:35:35 I have it. But then I would have to do the upgrade again? :( 2013-03-29T17:48:23 reimported a .ova file , uninstalled guest additions and was able to log into rtems :) :) 2013-03-29T18:14:23 you can uninstall the guest extensions and reboot before it upgrading. 2013-03-29T18:14:27 So a bit better now 2013-03-29T18:16:05 yup 2013-03-29T18:17:39 I wondered if something went wrong. 2013-03-29T18:18:08 I actually gave up updating the VM I was releasing this week. And will do a fresh install. Will delay it a week or so but hopefully smaller image 2013-03-29T19:39:13 DrJoel, no idea. 2013-03-29T19:39:43 I will disable python on Canadian cross (Cxc) builds until I sort something out 2013-03-29T20:00:42 sounds reasonable. I am trying to reproduce the earlier Puthon exception 2013-03-29T20:00:48 is anything cached? 2013-03-29T20:03:49 kiwichris around? 2013-03-29T20:21:28 yeah 2013-03-29T20:21:37 been play tennis again :) 2013-03-29T20:21:41 sending python error and another anomaly. 2013-03-29T20:21:53 I just in in from the hot tub and am about to leave the computer alone for a while 2013-03-29T20:22:52 FWIW if an environment check fails, then it would be nice to note that you have a set builder option to address it. I built xz, m4, and Python by hand. Each time someone does that, it is a chance for another error. 2013-03-29T20:23:14 Just send the Python 2.5.2 log message 2013-03-29T20:23:18 and I am out of here :) 2013-03-29T20:23:19 Thanks 2013-03-29T20:23:52 We need to figure out where milestone 0 is. And make a plan for going forward. We need to decide whether we are depth or breadth first. 2013-03-29T20:23:57 I can provide configurations to build tools. One exists for m4. 2013-03-29T20:24:34 Agreed. I am chasing my tail at the moment adding functionality and managing needed features. 2013-03-29T20:24:35 I am prone to do a little breadth but mostly depth first. Better to have a few covered well and have good framework. Addressing qemu itself might be ugly. 2013-03-29T20:24:48 To build ? 2013-03-29T20:25:04 Pusn the gdb simulators and BSPs as far as we can. 2013-03-29T20:25:27 To identify working versions for each BSP, patches, coverage preferred version, etc. I bet you will end up with more than one version 2013-03-29T20:25:57 I think this is the best path. It would make sure we have all the need parts and they work. Adding more sims will be simpler. 2013-03-29T20:26:07 The gdb simulators are fairly orthogonal, already built and would let us figure out how to report results, run tests, add add-ons, etc 2013-03-29T20:26:21 build so much with test results that folks want to fill in holes. :) 2013-03-29T20:26:22 cya 2013-03-29T20:26:26 *** DrJoel has quit IRC (Quit: Some folks are wise, and some otherwise.) 2013-03-29T21:10:20 *** DrJoel has joined #rtems 2013-03-29T21:10:20 *** DrJoel has joined #rtems 2013-03-29T21:10:20 *** ChanServ sets mode: +o DrJoel 2013-03-29T21:10:30 kiwichris.. found this looking at the failure log 2013-03-29T21:10:32 os.path.relpath(path[, start]) 2013-03-29T21:10:32 Return a relative filepath to path either from the current directory or from an optional start point. 2013-03-29T21:10:32 start defaults to os.curdir. 2013-03-29T21:10:32 Availability: Windows, Unix. 2013-03-29T21:10:32 New in version 2.6. 2013-03-29T21:10:59 Apparently os.path.relpath ties you to Python 2.6 or newer. Don't know what else might be broken after that 2013-03-29T21:11:16 Yeap. I need to add a relpath to the path module and provide it on version that do not provide it 2013-03-29T21:12:21 I am building Python 2.7.3 and continuing on for this machine 2013-03-29T21:12:30 Sure 2013-03-29T21:12:48 Deban 5.0 4-core amd64 2013-03-29T21:12:52 gcc12 2013-03-29T21:13:38 Did you see my email that this machine doesn't have a new enough autoconf to build automake .. the dependencies 2013-03-29T21:13:46 Ok. I hope the subprocess module is the same in 2.5.x. Back dating that code to something older would be tricky 2013-03-29T21:13:57 Yes I see that. Steps. 2013-03-29T21:14:17 The way to manage this is to make a build set that steps you forard 2013-03-29T21:14:20 forward 2013-03-29T21:14:42 A config per autoconf build then just step forward until you get a built version. 2013-03-29T21:15:11 autoconf is horrible because you need to have it installed in the real prefix to use it. 2013-03-29T21:15:20 back later cya 2013-03-29T21:17:51 *** DrJoel has quit IRC (Quit: If you can't laugh at yourself, make fun of other people.) 2013-03-29T21:47:05 *** kiwichris_ has joined #rtems 2013-03-29T21:55:58 *** kiwichris has quit IRC (*.net *.split) 2013-03-29T22:12:09 *** S_Somani1 has left #rtems 2013-03-29T22:15:35 *** dhananjay has joined #rtems 2013-03-30T00:59:18 *** dhananjay has quit IRC (Ping timeout: 245 seconds) 2013-03-30T01:07:21 *** dhananjay has joined #rtems 2013-03-30T02:46:39 *** dhananjay has quit IRC (Ping timeout: 245 seconds) 2013-03-30T03:06:39 *** dhananjay has joined #rtems 2013-03-30T03:12:03 *** dhananjay has quit IRC (Remote host closed the connection) 2013-03-30T03:19:42 *** S_Somani has joined #rtems 2013-03-30T04:39:13 *** S_Somani has quit IRC (Read error: Connection reset by peer) 2013-03-30T04:39:58 *** S_Somani has joined #rtems 2013-03-30T08:40:53 *** S_Somani has left #rtems 2013-03-30T11:01:14 *** DrJoel has joined #rtems 2013-03-30T11:01:14 *** DrJoel has joined #rtems 2013-03-30T11:01:14 *** ChanServ sets mode: +o DrJoel 2013-03-30T11:01:42 *** DrJoel changes topic to "Welcome to RTEMS (http://www.rtems.org). Please feel free to ask questions and, if you can, answer those from others. Google Summer of Code 2013 has been announced. RTEMS has applied." 2013-03-30T11:23:19 *** DrJoel has quit IRC (Quit: On the other hand, you have different fingers.) 2013-03-30T11:39:38 *** dhananjay has joined #rtems 2013-03-30T11:54:58 *** antgreen has quit IRC (Ping timeout: 272 seconds) 2013-03-30T11:56:57 *** S_Somani has joined #rtems 2013-03-30T12:35:43 *** antgreen has joined #rtems 2013-03-30T12:37:07 kiwichris: sometimes I install autoconf under ~/ and it works fine. (re: autoconf is horrible because you need to have it installed in the real prefix to use it.) 2013-03-30T14:09:14 *** dhananjay has quit IRC (Remote host closed the connection) 2013-03-30T14:17:34 *** S_Somani has left #rtems 2013-03-30T15:02:21 *** S_Somani has joined #rtems 2013-03-30T15:02:26 *** S_Somani has left #rtems 2013-03-31T08:08:40 *** S_Somani has joined #rtems 2013-03-31T09:35:44 *** monstr_ has joined #rtems 2013-03-31T11:42:46 *** antgreen has quit IRC (Ping timeout: 256 seconds) 2013-03-31T11:55:41 *** antgreen has joined #rtems 2013-03-31T13:03:36 *** S_Somani has left #rtems 2013-03-31T13:19:56 *** antgreen has quit IRC (Ping timeout: 258 seconds) 2013-03-31T13:31:53 *** antgreen has joined #rtems 2013-03-31T14:45:53 *** monstr_ has quit IRC (Ping timeout: 240 seconds) 2013-03-31T17:12:50 *** MegaAlex|away has quit IRC (Quit: ZNC - http://znc.in) 2013-03-31T19:29:06 *** the9a3eedi has joined #rtems 2013-03-31T20:02:26 *** the9a3eedi has quit IRC (Ping timeout: 255 seconds) 2013-03-31T20:29:19 *** the9a3eedi has joined #rtems 2013-03-31T21:39:54 *** Fallenou_ has joined #rtems 2013-03-31T21:49:32 *** Fallenou has quit IRC (*.net *.split)