2013-05-06T00:02:27 *** the9a3eedi has quit IRC (Ping timeout: 256 seconds) 2013-05-06T00:03:29 *** monstr_ has joined #rtems 2013-05-06T00:25:43 *** freenix has quit IRC (Read error: Connection reset by peer) 2013-05-06T00:30:26 *** freenix has joined #rtems 2013-05-06T01:05:43 *** rtdin has joined #rtems 2013-05-06T01:32:23 *** the9a3eedi-2 has joined #rtems 2013-05-06T02:05:41 *** sebhub has joined #rtems 2013-05-06T02:06:37 good morning 2013-05-06T02:13:55 does usleep go into a busy loop? i.e. does usleep cause the running task to switched to blocked? or does it stay running? 2013-05-06T02:21:21 i suppose usleep uses nanosleep and this one blocks and waits for some system ticks 2013-05-06T02:29:32 what if we're sleeping for less than a tick? 2013-05-06T02:30:00 i think you always sleep f 2013-05-06T02:30:04 or one tick 2013-05-06T02:35:00 sebhub, looks like 4.8-branch is broken on mingw32 2013-05-06T02:35:57 did you try a native build or with a linux to mingw cross compiler? 2013-05-06T02:36:57 native 2013-05-06T02:38:27 I am not what the issue is 2013-05-06T02:38:59 In libiberty in fibhead.c .. 2013-05-06T02:39:16 ../../gcc-4.8.0/libiberty/fibheap.c:220:30: error: 'LONG_MIN' undeclared (first use in this function) 2013-05-06T02:39:39 It seems this should be ok. Have also just upgraded my MinGW set up 2013-05-06T02:40:23 I will have to look into this tomorrow as I need to head out 2013-05-06T02:40:40 ok 2013-05-06T02:41:32 I have had to refactor the mingw native support code in the RSB to make sure the paths stay under 200 and something chars cause that is the max on Windows 2013-05-06T02:41:45 Cygwin is too unstable to build the compiler 2013-05-06T02:42:35 yes, i gave up native builds on mingw 2013-05-06T02:42:47 cygwin worked the last time i tried it 2013-05-06T02:43:13 Using Unix to build MinGW is nice however it avoids a range of test and lets things bot rot 2013-05-06T02:43:18 need to go cya 2013-05-06T03:00:51 *** phipse has joined #rtems 2013-05-06T03:07:26 *** lcpfnvcy has quit IRC (Read error: Connection reset by peer) 2013-05-06T03:07:36 *** lcpfnvcy has joined #rtems 2013-05-06T03:53:50 hmm.. in the git sources, the file cpukit\score\cpu\sparc\rtems\score\sparc.h uses uint32_t, but does not include stdint.h, which gave me an error when compiling my application. Is this a bug? 2013-05-06T03:57:12 this is a bug, but you should not use this file directly 2013-05-06T03:57:18 use 2013-05-06T03:58:27 actually, I included leon.h, which included that file 2013-05-06T04:01:34 ok, this is a bug, can you send a patch? 2013-05-06T04:02:04 sure, but I don't know the proper procedure to sending a patch. just a diff? 2013-05-06T04:02:10 and where do I send it? 2013-05-06T04:02:25 do you use the git version of rtems? 2013-05-06T04:02:59 I'm currently using 4.10.2 which is where I encountered the bug, but then I confirmed that the file in the git version doesn't have the include line either 2013-05-06T04:03:16 so yeah I guess 2013-05-06T04:03:52 you can use "git format-patch" and "git send-email", the list is rtems-devel@rtems.org 2013-05-06T04:04:07 sry, i have to leave now 2013-05-06T04:04:16 thanks for the info, will do now 2013-05-06T04:38:55 *** freenix has quit IRC (Ping timeout: 264 seconds) 2013-05-06T04:56:22 hmm.. msysgit (git port to windows) crashes when I use git send-email. I hope it's ok if I send the patch by copy pasting the text to outlook 2013-05-06T04:57:06 i would add the patch as an attachment 2013-05-06T04:58:00 ok, guess I'll do that 2013-05-06T05:00:05 all done 2013-05-06T05:00:44 now I can contribute back to the project :D thanks sebhub 2013-05-06T05:02:32 thanks, there is only a minor problem, the includes should be outside the __cplusplus stuff 2013-05-06T05:11:35 I see 2013-05-06T05:11:46 can you please send a new patch, then i can check it in 2013-05-06T05:11:54 ok 2013-05-06T05:27:30 resubmitted. Hope it's good now. 2013-05-06T05:27:43 I can't believe I messed up a silly include line :( 2013-05-06T05:34:27 *** freenix has joined #rtems 2013-05-06T05:54:05 *** arvind_khadri has joined #rtems 2013-05-06T05:54:37 *** arvind_khadri has joined #rtems 2013-05-06T05:59:18 *** noselasd has quit IRC (Ping timeout: 245 seconds) 2013-05-06T06:15:31 *** the9a3eedi-2 has quit IRC (Ping timeout: 264 seconds) 2013-05-06T06:28:26 *** rtdin_ has joined #rtems 2013-05-06T06:29:00 *** rtdin has quit IRC (Ping timeout: 264 seconds) 2013-05-06T06:29:05 *** rtdin_ is now known as rtdin 2013-05-06T06:55:32 *** peerst has quit IRC (Read error: Connection reset by peer) 2013-05-06T06:55:52 *** peerst has joined #rtems 2013-05-06T07:01:30 *** gedare has joined #rtems 2013-05-06T07:16:21 *** antgreen has quit IRC (Quit: Leaving) 2013-05-06T07:41:45 *** the9a3eedi-2 has joined #rtems 2013-05-06T07:41:52 *** weiY has joined #rtems 2013-05-06T07:45:11 *** freenix has quit IRC (Quit: Leaving) 2013-05-06T07:51:09 *** Hesham has joined #rtems 2013-05-06T07:52:43 Hi gedare, What's is the value for --enable-rtemsbsp to build for new realview-pbx-a9 BSP ? 2013-05-06T07:57:28 realview_pbx_a9_qemu 2013-05-06T07:59:40 thanks sebhub 2013-05-06T08:02:25 sebhub: Still can not find the BSP. My configuration line is ... 2013-05-06T08:02:39 did you run the bootstrap script? 2013-05-06T08:02:42 ../../rtems/configure --target=arm-rtems4.11 --enable-rtemsbsp=realview_pbx_a9_qemu --enable-tests --disable-posix --disable-itron --disable-networking 2013-05-06T08:03:00 looks good so far 2013-05-06T08:03:26 no I forgot I will run the bootstrap and try again 2013-05-06T08:15:57 hi good keep me informed :) 2013-05-06T08:20:23 It's compiling now... 2013-05-06T08:23:52 good 2013-05-06T08:24:03 and then to figure out how to boot it on qemu! :) 2013-05-06T08:24:15 the sim-script should "just work" ideally 2013-05-06T08:24:23 you know? 2013-05-06T08:26:39 I found the command to run on QEMU in README file at the bsp dir ... 2013-05-06T08:26:42 qemu-system-arm -S -s -no-reboot -net none -nographic -M realview-pbx-a9 -m 256M -kernel ticker.exe 2013-05-06T08:27:31 Yes sure I know about sim-script 2013-05-06T08:28:26 The problem now after building and trying to get ticker.exe run on Qemu is that "nothing happens" :( 2013-05-06T08:28:36 oss: Could not initialize DAC 2013-05-06T08:28:37 oss: Failed to open `/dev/dsp' 2013-05-06T08:28:37 oss: Reason: No such file or directory 2013-05-06T08:28:37 oss: Could not initialize DAC 2013-05-06T08:28:37 oss: Failed to open `/dev/dsp' 2013-05-06T08:28:37 oss: Reason: No such file or directory 2013-05-06T08:28:37 audio: Failed to create voice `lm4549.out' 2013-05-06T08:29:34 use the sim-script, or look at the qemu documentation for the options "-S -s" 2013-05-06T08:31:14 sebhub: Ok I will. BTW, I have an earlier version of qemu-system-arm, would that be a problem ? 2013-05-06T08:32:10 it can always be a problem, but you should be able to tell if the realview pbx a9 is supported 2013-05-06T08:32:27 i don't know, you have to try it out 2013-05-06T08:33:05 I am updating qemu now 2013-05-06T08:35:52 github is back, good news 2013-05-06T08:36:25 Does somebody happen to know, why the saved interrupt level in bootcard.c is never restored? 2013-05-06T08:38:00 it should make sure that interrupts are disabled during system start 2013-05-06T08:38:17 its superfluous from my point of view 2013-05-06T08:40:06 in L4RTEMS I have the problem, that the IRQs stay disabled, because this level never gets restored 2013-05-06T08:41:26 during the first context switch interrupt should initialized with the value stored in the thread context 2013-05-06T08:41:52 the interrupt state at boot_card is irrelevant for the threads 2013-05-06T08:44:16 is this the call to _Modes_Set_interrupt_level in modes.inl? 2013-05-06T08:46:37 _Thread_Handler() -> _ISR_Set_level() 2013-05-06T08:51:59 thanks 2013-05-06T09:04:27 okay, so the thread starts with IRQs disabled, but they never get enabled again. As enable just restores the old state, I never get to an enabled state. 2013-05-06T09:05:56 I just can't figure out, if I messed something up in RTEMS or if the call to the l4vcpu is going wrong. 2013-05-06T09:16:29 phipse: the interrupt level is set by each thread 2013-05-06T09:16:44 see: cpukit/score/src/threadinitialize.c:181 or so 2013-05-06T09:16:58 :q 2013-05-06T09:17:22 and it is processed when the thread is 'dispatched' 2013-05-06T09:17:27 in score/src/threaddispatchdisablelevel.c 2013-05-06T09:17:47 *** the9a3eedi-2 has quit IRC (Ping timeout: 256 seconds) 2013-05-06T09:19:02 the "idle" task usually is the first created, and it will have interrupt enabled. your application threads may be specifying to start with interrupts disabled? 2013-05-06T09:20:21 *** antgreen has joined #rtems 2013-05-06T09:26:13 *** weiY has quit IRC (Ping timeout: 246 seconds) 2013-05-06T09:26:15 I start the nsecs sample on L4 in qemu 2013-05-06T09:30:48 but there is not interrupts? 2013-05-06T09:31:10 probably an issue with the implementation of the isr_set_level or whatever function 2013-05-06T09:32:26 *** weiY has joined #rtems 2013-05-06T09:34:23 Hesham: how busy are you these days? 2013-05-06T09:35:20 rtdin: sorry about your status :( if you still like to participate in RTEMS we are glad to help 2013-05-06T09:36:18 I will begin my practical exams next week and right after it my final exams starting at 25/5 2013-05-06T09:36:18 thats okay. But yeah I am interested :) 2013-05-06T09:38:14 *** freenix has joined #rtems 2013-05-06T09:38:24 ok Hesham then you should be studying? :) I do have a task for you to try if you have some spare time though... 2013-05-06T09:38:45 See if you can get the realview pbx a9 BSP to boot on http://www.m5sim.org/Main_Page 2013-05-06T09:39:30 I have used the simulator quite a lot myself and can help you to figure out how to build it, the first step would be to build and boot linux 2013-05-06T09:40:02 if you're interested. this is not something we need for GSoC, just a little bit of "fun" 2013-05-06T09:40:03 Yes I should by studying but I do not :D I will work on that task now :) 2013-05-06T09:40:33 ok.. start out here: http://www.m5sim.org/Documentation 2013-05-06T09:40:54 prod me if you get stuck. what we want is the "full system" simulator 2013-05-06T09:41:23 actually, if yo ugo http://www.m5sim.org/Tutorials 2013-05-06T09:41:35 the HiPEAC Tutorial is quite good material 2013-05-06T09:42:41 I will read both, would it be good for RTEMS to get some BSPs running into that simulator ? 2013-05-06T09:42:44 rtdin: if you like you can post your development project idea on rtems-devel to solicit feedback. we can try to get you some mentorship and guidance if you are interested in implementing the CVs for RTEMS. 2013-05-06T09:42:56 Hesham: Yes. I have already done it with sparc64/niagara 2013-05-06T09:43:20 but, I am keenly interested in the arm/realview 2013-05-06T09:43:33 and, we might also investigate whether we can get any MIPS bsp to boot on it 2013-05-06T09:43:36 gedare: okay i will post my project proposal soon 2013-05-06T09:44:06 i don't know if we can boot any of our i386 bsp's, since their x86 is focused on the 64-bit implementation it might not have everything for backward compatibility. 2013-05-06T09:44:10 rtdin: cool. 2013-05-06T09:44:25 :) 2013-05-06T09:44:33 Great ! I will see what I can I do and come back to you again. 2013-05-06T09:44:46 hi,gedare. can rtems run on gem5? I am also interested in it. :) 2013-05-06T09:45:11 freenix: it depends on what you want to do with it, but maybe you would like to work together with Hesham to see if the realview bsp can boot on it? 2013-05-06T09:45:29 freenix: I have booted the niagara BSP on the SPARC Full System simulator mode. 2013-05-06T09:45:42 freenix: are you familiar with gem5? 2013-05-06T09:46:08 in the past, out lab mainly use it to learn cpu arch 2013-05-06T09:46:31 http://gedare-csphd.blogspot.com/2013/02/rtems-on-gem5-sparc.html 2013-05-06T09:46:43 Not familar, But I used it. 2013-05-06T09:46:46 i did this about 3 years ago, but verified it works last week 2013-05-06T09:46:54 ok. i use it in my research. 2013-05-06T09:47:03 you can see more on some of my recent blog posts there. 2013-05-06T09:47:23 It sounds great. 2013-05-06T09:48:46 But gem5 is slow in simulating full system 2013-05-06T09:49:12 If run android, it may take hours to boot.:) 2013-05-06T09:49:47 maybe 2013-05-06T09:50:33 a question not about gem5. Does the rtems arm gdb simulator support timer? 2013-05-06T09:52:02 not with an accurate time 2013-05-06T09:52:09 it fast-forwards time 2013-05-06T09:52:51 when I use gdb to run rtl, it always halt when excuting shell. 2013-05-06T09:53:17 So I just doubt , but not sure whether timer is supported. 2013-05-06T09:54:58 *** sebhub has quit IRC (Remote host closed the connection) 2013-05-06T09:56:57 *** sebhub has joined #rtems 2013-05-06T09:57:24 timer should be fine. I don't know why the shell is halting 2013-05-06T09:58:03 I think the "capture" testsuites/sample/capture uses the shell, does that work? 2013-05-06T09:59:46 I did not try it. I will try this tommorw, for I have no development envrioment now. 2013-05-06T10:01:39 ok 2013-05-06T10:04:18 *** DrJoel has joined #rtems 2013-05-06T10:04:18 *** ChanServ sets mode: +o DrJoel 2013-05-06T10:04:18 gedare: I read your blog just now. 2013-05-06T10:04:40 *** S_Somani has joined #rtems 2013-05-06T10:04:48 But gem5 only provides a cpu model, but not soc model. 2013-05-06T10:04:52 hopefully not the whole thing! :) 2013-05-06T10:05:03 not exactly... 2013-05-06T10:05:11 haha, just the rtems sparc page 2013-05-06T10:05:19 it does model devices and busses 2013-05-06T10:05:30 it does not model a specific soc usually i think 2013-05-06T10:05:49 but at least the ARM Android port relies on the Realview PBX A9 configuration of linux 2013-05-06T10:06:04 so I suspect we should be able to use the new realview bsp there also... i hope :) 2013-05-06T10:06:47 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2013-05-06T10:06:48 * DrJoel waves 2013-05-06T10:07:00 yeah, I missed this, I have run a full android on gem5. 2013-05-06T10:07:17 hey DrJoel how's it going 2013-05-06T10:07:17 so interesting. 2013-05-06T10:07:44 exactly. 2013-05-06T10:07:49 DrJoel: gedare: Hi, I made TUI prototype. I commented the link on the proposal. Did you see it? 2013-05-06T10:07:59 haven't had time to investigate 2013-05-06T10:08:15 oh okay.. 2013-05-06T10:08:44 me either. I still have a couple I am not sure I have commented on. Even if I takled to the student. :( 2013-05-06T10:09:12 S_Somani: there has been some discussion regarding your project and direction. if we accept the project, we may need to provide a slightly different scope of work. it will still be GUI/TUI/CLI stuff, but might rely on some pre-existing, non-published code that a developer has been working on 2013-05-06T10:09:32 *** rtdin has quit IRC (Ping timeout: 260 seconds) 2013-05-06T10:09:32 so, just a heads-up not to get too invested in any particular techniques yet :) 2013-05-06T10:10:10 it is good to make a couple of in-roads though. and hopefully, if we do choose a slightly different direction, we can rework that asap. 2013-05-06T10:10:43 oh..okay...I just thought in order to make my proposal strong it would be better if I started working asap 2013-05-06T10:10:55 I think the work to generating the input to the gui is a big deal. 2013-05-06T10:11:14 i know this is all a little vague, but I just don't want you to do a lot of work now that we don't need later. 2013-05-06T10:11:14 *** freenix has quit IRC (Quit: Leaving) 2013-05-06T10:11:16 It is stronger. 2013-05-06T10:11:31 yep, and your experience will be quite helpful then too 2013-05-06T10:11:38 Ok :) 2013-05-06T10:11:45 And it may turn out that the approach of generating a file from conf.t pays off double. 2013-05-06T10:11:47 * gedare slaps verm__ on the back of the head 2013-05-06T10:12:08 My understanding is that he was focused on the BSP specific options and RTEMS build time. 2013-05-06T10:12:25 Not the confdefs.h parameters and application build/link time 2013-05-06T10:12:28 i think it will be useful at any rate, yes, though it may become more of a secondary sanity check. 2013-05-06T10:12:54 Better to have a single plan that covers both and avoids unnecessary duplication. I think that's the concern. 2013-05-06T10:13:25 ok...can you please send me the code...so that I start getting familiar with it. 2013-05-06T10:13:31 is that a fair description gedare? 2013-05-06T10:13:36 he was supposed to give an overview last week, but that got lost. both he and kiwichris seem to agree. S_Somani I will work on getting that to you. 2013-05-06T10:13:49 ok :) 2013-05-06T10:14:18 DrJoel: I'm not sure. I talked with verm__ and he said it also includes application configuration options, but there was some incompleteness there. but still, the options/parameters can be generated by the config db or something like that 2013-05-06T10:14:39 this gets back to an issue i mentioned awhile back 2013-05-06T10:14:57 which is whether parameters should come from conf.t or from a programmatic interface (I had mentioned XML) 2013-05-06T10:15:16 The progammatic interface can be generated from conf.t 2013-05-06T10:15:38 yeah, but my understanding is it already exists in something called "rtems config" :) 2013-05-06T10:15:45 I just don't see the value of defining a forma 2013-05-06T10:15:49 I'll ask verm__ to get the code and some basic documentation up on ftp site or something. 2013-05-06T10:15:58 If it isn't published, it doesn't exit 2013-05-06T10:16:00 exist 2013-05-06T10:17:41 about to reboot but.. should leon.h include rtems/score/sparc.h? I thought we didn't allow inclusion of files at that level 2013-05-06T10:18:02 i don't know. i complained about that in my first email. Ralf deleted my complaint in his way. 2013-05-06T10:18:29 the problem is to avoid having S_Somani work on something that becomes obsolete within the next RTEMS release cycle. 2013-05-06T10:18:32 I asked again and then Sebastian committed the patch before anyone had a chance 2013-05-06T10:18:53 sry, but this is a bug in sparc.h 2013-05-06T10:18:55 yeah i know. well, including stdint.h is fine 2013-05-06T10:19:13 the leon.h is a different issue 2013-05-06T10:19:18 yes, exactly. 2013-05-06T10:19:25 the bug is real, the symptom is not :) 2013-05-06T10:19:49 but rtems/score/sparc.h should never be directly included 2013-05-06T10:20:15 rebooting.. hold on 2013-05-06T10:20:20 *** DrJoel has quit IRC (Quit: Leaving) 2013-05-06T10:26:11 *** DrJoel has joined #rtems 2013-05-06T10:26:12 *** DrJoel has joined #rtems 2013-05-06T10:26:12 *** ChanServ sets mode: +o DrJoel 2013-05-06T10:34:51 I reboot and it gets quiet :) 2013-05-06T10:35:01 yup. 2013-05-06T10:36:57 *** sebhub has quit IRC (Remote host closed the connection) 2013-05-06T10:46:23 what do I have to do, to get qemu output of RTEMS to my console? the -serial stdio switch doesn't change a thing. 2013-05-06T10:47:24 phipse pc386 must be built with USE_CONSOLE_AS_COM1=1 on configure line to switch the default console device 2013-05-06T10:48:39 ha, I won't find that in the qemu man page ;) 2013-05-06T10:50:09 Umm.. a doubt just popped into my head. I'm fine on working with a pre-existing code, but don't you think I should mention that in my proposal too (but now since the deadline is closed how to do that.. or its not a big deal) ? 2013-05-06T10:52:12 S_Somani: we can permit modification of the proposal, but as long as you are in contact with us it is ok 2013-05-06T10:53:24 ok :) I'll get familiar with the code and prepare for my exams in the acceptance waiting period then. 2013-05-06T10:55:05 *** weiY has quit IRC (Ping timeout: 276 seconds) 2013-05-06T10:57:29 *** Hesham has quit IRC (Quit: Leaving.) 2013-05-06T10:57:50 sounds good 2013-05-06T11:01:33 *** rtdin has joined #rtems 2013-05-06T11:23:40 DrJoel: I configured RTEMS with the USE_CONSOLE_AS_COM1=1 option and built it, but I still don't get the output. 2013-05-06T11:23:44 gedare: Once we pick projects, would it make sense to have the students put their proposal (minus personal info) on the wiki page for their project and update it for the community there 2013-05-06T11:25:03 phipse: http://wiki.rtems.org/wiki/index.php/Qemu#Using_the_rtems-testing_Module 2013-05-06T11:25:16 it's a little bit old, but the ideas should be the same 2013-05-06T11:25:18 Hmmm.. what was your complete configure command for rtems 2013-05-06T11:25:24 you might have to hook up a vnc viewer 2013-05-06T11:25:42 i forget exactly.. 2013-05-06T11:25:58 DrJoel: yes it probably would make sense 2013-05-06T11:26:34 what are the boot arguments on your grub command line? I forgot the BSP now can use those to select the console 2013-05-06T11:26:52 ../rtems/configure --target=i386-rtems4.11 --enable-rtemsbsp=pc386 --enable-maintainer-mode --enable-rtems-debug --disable-posix --disable-networking --enable-tests=samples --disable-cxx -prefix=/home/phipse/gsoc/b-i386-install target_alias=i386-rtems4.11 USE_CONSOLE_AS_COM1=1 2013-05-06T11:27:02 qemu-kvm -m 256 -serial stdio -boot a -fda rtems-boot.img -hda hda.img -hdb fat:b-i386/i386-rtems4.11/c/pc386/testsuites/samples/nsecs/ 2013-05-06T11:27:06 That looks ok. 2013-05-06T11:27:28 can you look at the file grubmenu in the rtems-boot.img? 2013-05-06T11:27:42 usually using mtools and mcopy to get it back into linux 2013-05-06T11:27:52 or me bold and mount it in loopback :) 2013-05-06T11:29:39 * gedare always mounts 2013-05-06T11:30:45 can you give him the commadn line? 2013-05-06T11:31:28 gedare: how close is the posix key code? :) 2013-05-06T11:31:52 no because i lost my qemu helping scripts awhile back and never regenerated them 2013-05-06T11:32:32 Ash is rewriting the code to use the "new" approach that I suggested. I don't think it will take too long 2013-05-06T11:32:39 Then I can do some quick testing and commit 2013-05-06T11:32:44 mount -o loop /path/to/file.img /some/mount/point 2013-05-06T11:32:52 great! He has been really dedicated! 2013-05-06T11:32:57 yesh 2013-05-06T11:33:11 * gedare takes a knife to that h 2013-05-06T11:48:32 *** S_Somani has left #rtems 2013-05-06T11:53:00 ha, phone interrupts ... which file should I look for in the image? grub.cfg? 2013-05-06T11:54:12 I think its grubmenu but its been a while 2013-05-06T11:54:14 http://whnt.com/2013/05/05/your-pictures-cold-core-low-causes-hail-in-the-tennessee-valley/ 2013-05-06T11:54:37 about 30 miles from here.. this weekend .. enough hail to have "snow" pictures 2013-05-06T11:55:17 jnice 2013-05-06T11:55:33 or should i say nICE 2013-05-06T12:00:07 the only file I can access is grub.cfg, the rest is binary. 2013-05-06T12:00:16 brb .. phone 2013-05-06T12:00:43 and this just defines where the rtems-grub.cfg should be 2013-05-06T12:01:54 *** monstr_ has quit IRC (Ping timeout: 264 seconds) 2013-05-06T12:08:37 *** antgreen has quit IRC (Ping timeout: 246 seconds) 2013-05-06T12:09:21 *** phipse is now known as phipse_away 2013-05-06T12:26:13 *** arvind_khadri has joined #rtems 2013-05-06T12:38:06 *** antgreen has joined #rtems 2013-05-06T12:48:11 *** rtdin has quit IRC (Quit: ChatZilla 0.9.90 [Firefox 20.0.1/20130409194949]) 2013-05-06T12:49:07 *** antgreen has quit IRC (Ping timeout: 268 seconds) 2013-05-06T12:56:04 *** phipse_away is now known as phipse 2013-05-06T13:08:58 *** monstr_ has joined #rtems 2013-05-06T13:51:15 *** monstr_ has quit IRC (Ping timeout: 258 seconds) 2013-05-06T14:04:10 phipse: did you find what you need? 2013-05-06T14:11:19 Today is the anniversary of getting my B.S. Michele located this gem: https://fbcdn-sphotos-g-a.akamaihd.net/hphotos-ak-frc1/470098_10150764877931589_1439455490_o.jpg 2013-05-06T14:14:56 i had a hard time noticing you without a beer in your hand 2013-05-06T14:15:12 gedare: sadly not, but after 10h coding I can't concentrate anymore. As it's not critical to my progress I will come back to it, if I need it. 2013-05-06T14:15:53 verm__: thanks folks.. he'll be here all week. :) 2013-05-06T14:19:28 bye 2013-05-06T14:19:36 *** phipse has left #rtems 2013-05-06T14:20:05 *** gedare_ has joined #rtems 2013-05-06T14:20:24 amazing the updater worked 2013-05-06T14:21:19 *** gedare_ has quit IRC (Client Quit) 2013-05-06T14:22:21 *** gedare2 has joined #rtems 2013-05-06T14:22:31 yes at long last! ;) 2013-05-06T14:23:14 Just in time for gsoc.. thanks! 2013-05-06T14:24:00 well, technically it was fixed last year but those github changes broke it, sigh 2013-05-06T14:26:16 verm__, is it a problem to post the code for rtems config base? 2013-05-06T14:27:00 i (and the student, and other mentors) need to wrap our heads around it before making any further decisions 2013-05-06T14:27:01 i can tear out the portion that would be used in the gsoc project which is what i was planning on doing 2013-05-06T14:27:15 https://github.com/MentorEmbedded/Debug-Agent-Library/wiki looks interesting 2013-05-06T14:27:32 "tear out" as in remove some code you already wrote? Or as in provide a clean interface? 2013-05-06T14:27:43 provide a clean interface that has nothing to do with my current work 2013-05-06T14:28:10 DrJoel, interesting for what? 2013-05-06T14:28:11 gedare2: well the choice is he'll do his gsoc project using conf.t or whatever 2013-05-06T14:28:24 and then when i publish my work conf.t gets deleted along with all other hand-maintained / generated files 2013-05-06T14:28:27 and that leaves his project where? 2013-05-06T14:28:48 you assume your work will be the accepted one ;) 2013-05-06T14:28:51 to me the choice is pretty clear: just use the code i've already written to avoid that 2013-05-06T14:29:15 yes. and we need to see it still. right now, to me, it is vaporware. i have no means to understand it 2013-05-06T14:29:19 you know? 2013-05-06T14:29:28 i'm not sure if you're being sarcastic or not 2013-05-06T14:29:37 it's vaporware because you weren't at the mentors summit 2013-05-06T14:29:44 you keep ignoring that sebastian, chris and joel saw it 2013-05-06T14:29:59 and you keep ignoring that it's running right now i've tried to get you interested in it in the past 2013-05-06T14:30:01 you were always busy 2013-05-06T14:30:30 yeah. even so, none of them (Except chris, briefly) have given any input about how it will work with application configuration. That's all I want to be certain about. 2013-05-06T14:30:34 i'm sorry but please do not push this all on me, you're more than welcome to refuse the work but i'm making it clear now what the situation is 2013-05-06T14:31:27 it's not my intent to be pushy, and i'm sorry if you feel pressure 2013-05-06T14:31:43 i have to admit i've found your approach to this rather abrasive -- rather than ask me how my work will help you've asked me to present it so you can decide if it's useful or not 2013-05-06T14:32:04 i'm thinking about the student's interest at this point in time. I agree 100% that we need to incorporate your work, but we also need the student to have the solid footing right away, do you agree? 2013-05-06T14:32:30 i've put nearly three years of effort into this and you're happy to decide -- seemingly on a whim -- if it's even worth anything 2013-05-06T14:32:37 agreed and i'm thinking about that, too which is why i brought it up 2013-05-06T14:33:03 and i'm happy to provide everything everyone asks but i've had two seperate people dismiss it without even seeing it i find that very frusterating 2013-05-06T14:33:31 good. that's my main concern. i have no doubt that your work will be useful. 2013-05-06T14:33:46 i also have no doubt the student will be able to work with the framework 2013-05-06T14:34:04 since he has shown competence with python and scripting, which i assume are going to be the main skills needed. 2013-05-06T14:34:50 I have seen nothing but the basics of the waf build system. I have seen nothing related to the parameters in confdefs.h. You may have mentioned it and discussed it with Chris but I have never seen that or even realized it was in the scope of what you were addressing. 2013-05-06T14:35:14 My understanding is that you were addressing the BSP specific configuration options in their configure.ac files 2013-05-06T14:35:27 Those are not documented except in the code. 2013-05-06T14:35:47 that's because most of my presentation was on the the build system 2013-05-06T14:35:58 there are two parts to it 90% of what i talked about was the build 2013-05-06T14:36:23 i wouldn't have even brough this up now if it wasn't for the gsoc project -- the portion i'm talking about is all the backend glue 2013-05-06T14:36:51 that part was designed to be hidden from developers, guess it served its purpose in the wrong way.. 2013-05-06T14:36:56 We want to share that. One thing we hate is two sets of code with similar capabilities 2013-05-06T14:37:13 that's the problem, there can't be because my project is single definition for everything 2013-05-06T14:37:35 unit tests, documentation, configurations are all generated from one centrally stored repository of build time, run time and test information 2013-05-06T14:37:51 sure, you can have two but you'll be running both not one or the other 2013-05-06T14:38:57 while this is earlier than i was expecting since the code is already there there's no reason it can't be used for the gui project 2013-05-06T14:39:53 it sounds promising, like i said, i just need to see and understand how the application configures the resources it needs--as you point out, that is the part that is relevant to the student's proposed project 2013-05-06T14:39:55 it means conf.t will become a python file 2013-05-06T14:40:09 documentation and automated tests will be generated from that 2013-05-06T14:40:38 see, that never before is clear to me, how the configuration options are specified. I think one way to phrase the conversion of effort from the students POV is that the parsing of conf.t will turn into a way to generate that python file 2013-05-06T14:40:57 and in the future, we can maintain the python file manually I would imagine, if that is what will generate the documentation for end user 2013-05-06T14:41:05 that's the wrong way around 2013-05-06T14:41:19 the python options are classes 2013-05-06T14:41:48 with programic restrictions on values, (int, string, list of ints, strings.. min+max) 2013-05-06T14:42:19 this means for a gui you can have limited spinners, dropdown with static lists it would be fairly easy the original gui i wrote was using wxpython (i've been using wx for quite a while in c++) 2013-05-06T14:42:29 well, as long as there is a match somewhere between where configuration is implemented (confdefs.h still?), configured (python-based rtems config), and documented. 2013-05-06T14:43:03 good. those are all basically the same kinds of constraints that they were putting into conf.t, but manually. i was interested in the programmatic approach and am glad to find it basically "in hand" 2013-05-06T14:43:20 yeah it will be confdefs.h for now but in the future that will be autogenerated too 2013-05-06T14:43:28 that's fine when we get there. 2013-05-06T14:43:55 not needed for now -- i'll send an email out in a few hours or so 2013-05-06T14:44:32 ok. one last question for you, does the output of configuration produce a .h file for compiling with the application? Or does it embed the configuration options some other way into the compilation of the application? 2013-05-06T14:45:08 heh actually, it doesn't even address that.. waf does that part 2013-05-06T14:45:18 i.e. the specific setting of all the configuration options produces a _______ file 2013-05-06T14:45:22 converting from internal options to .h 2013-05-06T14:45:30 i guess what i'm digging for is what the output of the gui tool will be 2013-05-06T14:45:39 or will it just call more python code 2013-05-06T14:45:48 .ini file 2013-05-06T14:45:59 ok. that is what i was urging the student to use also already 2013-05-06T14:45:59 that's how the current build-time configuration is kept 2013-05-06T14:46:11 i already have code written to support that 2013-05-06T14:46:18 going from my internal config (class-based) format to a .ini file 2013-05-06T14:46:23 because i had in my head an idea about using a python-based configuration tool *cough cough* 2013-05-06T14:46:24 and loading that .ini file over static defaults 2013-05-06T14:46:31 so he doesn't have to do any of that 2013-05-06T14:46:43 all he has to do is go from conf.t -> python classes (1 per option) 2013-05-06T14:46:46 and the rest is 100% complete 2013-05-06T14:47:02 writing out .h files is trivial using waflib and avoiding using waf fully 2013-05-06T14:47:03 ok. feel free to help him to scope an appropriate project then 2013-05-06T14:47:35 i need to take out the files from waf so he can do this, shouldn't be too bad 2013-05-06T14:47:48 i was in the process of splitting it out anyway so i could write unit tests 2013-05-06T14:48:27 i have api documentation now as well using Sphinx 2013-05-06T14:48:46 (which i did during the holidays) 2013-05-06T14:49:07 ok great. feel free to give the student some small tasks to try out too, to make sure he is going to be capable 2013-05-06T14:49:20 we've been trying to do that this year during this period of reviewing proposals. 2013-05-06T14:49:52 well you guys said he is that's good enough for me 2013-05-06T14:49:55 also, will it be plausible for legacy applications to be ported forward to the new system, or will that be hard? 2013-05-06T14:50:04 esp if they already know wx 2013-05-06T14:50:20 yeah i think he used that to mock up a gui already, and did an ncurses tui recently 2013-05-06T14:50:24 that's the only part i'd be worried, dealing with wx can be painful or gui programming that uses sizers for layout 2013-05-06T14:50:37 *** gedare2 has quit IRC (Quit: Ex-Chat) 2013-05-06T14:50:46 legacy applications? can you elaborate? 2013-05-06T14:51:17 the simplest case is to take the hello sample and convert its configuration approach (i.e. the system.h file) to the new way 2013-05-06T14:51:30 actually, hello might not use system.h, it probably embeds configuration options directly in its C source code 2013-05-06T14:52:29 i imagine at some point we will be breaking the ability to compile existing applications. hopefully that will not be until after the confdefs.h rewrite. 2013-05-06T14:53:29 all right. these kinds of issues can wait. 2013-05-06T14:53:41 if he ends up using this no, there will be no backporting 2013-05-06T14:53:52 *** antgreen has joined #rtems 2013-05-06T14:53:57 however, if it's important since he'll be writing it as a stand alone tool it would be possible 2013-05-06T14:54:14 what i mean is, as long as it remains stand alone that would be possible 2013-05-06T14:54:23 we can consider some methods to ease user integration 2013-05-06T14:54:33 the moment it's integreated with everything else (stops using a fork of rtems-config) then it wouldn't be 2013-05-06T14:54:44 ok, makes sense. 2013-05-06T14:55:06 the reason is the build system at that point would be waf which doesn't support legacy anything 2013-05-06T14:55:21 ah, i had a question related to that too... 2013-05-06T14:55:42 can an application compile against a waf-built RTEMS using an application's Makefile? 2013-05-06T14:55:58 well that's not true, you can use legacy if you want by using rtems-config (the commandline binary) similar to how many open source programs work.. rtems-config --bsp=sparc/sis --cflags|--libs etc 2013-05-06T14:56:04 yes 2013-05-06T14:56:06 ok 2013-05-06T14:56:10 CFLAGS=`rtems-config --bsp=sparc/sis --cflags` 2013-05-06T14:56:40 this is the normal, expected "application time" compilation for users I think... most of them will compile RTEMS once and install the headers/libraries to compile the application against. 2013-05-06T14:57:06 yeah, rtems-config is a stand alone binary all the information is compiled into it when it's installed 2013-05-06T14:57:28 okay. it is an important distinction to continue carrying forward. 2013-05-06T14:57:37 that works now 2013-05-06T14:57:49 it has for over a year had you been interested in trying it out :) 2013-05-06T14:58:04 heh. i tried it once! 2013-05-06T14:58:27 it's gone through a 100% rewrite since then 2013-05-06T14:58:40 yeah, so just as well that i wasn't trying to track it with my own work ;) 2013-05-06T14:58:42 i mean that in the most literal sense i started over from scratch 2013-05-06T14:58:59 well the config file format never changed :) (config.ini) 2013-05-06T15:03:01 ok, thanks for the clarifications. 2013-05-06T15:03:45 it will be good for the student to get his hands on the code and rewrite his timeline. hopefully you guys can come up with some good tasks to carry out assuming the GUI work is as easy as you've said 2013-05-06T15:07:51 well, it also means they'll be able to focus a lot more on the GUI portion 2013-05-06T15:08:37 there are other things that can be worked on, too. i had a gui frontend for building with a pretty scroll bar :) 2013-05-06T15:13:51 providing a path for integration with eclipse might be interesting. 2013-05-06T15:17:28 already have that working 2013-05-06T15:17:31 and msvc 2013-05-06T15:17:46 oh sorry, not eclipse -- i have compiled rtems under msvc and used it as the ide 2013-05-06T15:17:53 waf has a headless mode suitable for this 2013-05-06T15:20:44 ok 2013-05-06T15:21:03 but also integrating configuration in the IDE? 2013-05-06T15:21:30 i don't know if that is possible in msvc, but probably eclipse has suitable flexibility for it. I don't know how desirable it would be though, or what the maintenance effort might be 2013-05-06T15:21:36 that would be possible with eclipse, msvc... good question 2013-05-06T15:21:47 oh, i know exclipse can do it for sure 2013-05-06T15:22:03 maint wouldn't be hard because it would be a shim in front of rtemsconfig 2013-05-06T15:22:22 the only maintained portion would the connector to the eclipse gui which tends to be extremely stable 2013-05-06T15:22:45 plus we're dealing with typed options so all you have to do is write handlers for those types the code base would be extremely small 2013-05-06T15:23:07 it was specifically designed to have gui(s) put in front of it 2013-05-06T15:24:44 i think msvc only supports custom config guis for msbuild, in the case where it's not using the ms C compiler/msbuild it's not possible 2013-05-06T15:25:04 ok. i'd suggest guiding the student toward that as a possible project if time permits then. assuming we accept that project, of course. *nods* 2013-05-06T15:25:18 or even if not *nods sagely* 2013-05-06T15:25:21 the clipse gui? 2013-05-06T15:25:25 eclipse 2013-05-06T15:25:27 yeah 2013-05-06T15:25:41 as something to do if everything else goes fast enough 2013-05-06T15:25:45 OK, does the version of eclipse matter? 2013-05-06T15:26:10 no clue he should ask around for what is being used in RTEMS.. there is some plugin for RTEMS so I'd start there 2013-05-06T15:26:21 if that plugin works on the most recent eclipse, do that then 2013-05-06T15:26:27 i think i looked at that one a couple of years ago 2013-05-06T15:26:36 that's what i would use the most recent version of eclipse and start over 2013-05-06T15:26:53 sure, he can update the RTEMS plugin then too if it makes sense. 2013-05-06T15:27:02 i haven't looked at it so i don't know enough 2013-05-06T15:27:12 but i know qutie a few users are using it 2013-05-06T15:27:16 i think that would be a waste of time since that plugin would be obsoleted by waf 2013-05-06T15:27:30 which works w/out even needing a plugin :) 2013-05-06T15:30:46 anyway, i have to go for a while we'll chat via email 2013-05-06T15:34:07 ah, ok 2013-05-06T15:34:15 *** antgreen has quit IRC (Quit: Leaving) 2013-05-06T15:37:28 hey DrJoel 2013-05-06T15:56:43 *** gedare has quit IRC (Quit: Leaving) 2013-05-06T16:53:38 *** DrJoel has quit IRC (Quit: There's nothing dirtier then a giant ball of oil) 2013-05-06T17:01:48 *** arvind_k has joined #rtems 2013-05-06T17:04:43 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2013-05-06T17:56:26 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-06T19:05:02 *** the9a3eedi-2 has joined #rtems 2013-05-06T20:02:45 *** the9a3eedi-3 has joined #rtems 2013-05-06T20:05:35 *** the9a3eedi-2 has quit IRC (Ping timeout: 256 seconds) 2013-05-06T20:17:29 *** antgreen has joined #rtems 2013-05-06T20:22:46 *** freenix has joined #rtems 2013-05-06T20:30:51 *** kiwichris has joined #rtems 2013-05-06T21:48:02 *** the9a3eedi-3 has quit IRC (Ping timeout: 252 seconds) 2013-05-06T21:48:22 *** the9a3eedi has joined #rtems 2013-05-06T23:40:07 *** monstr_ has joined #rtems 2013-05-07T00:08:48 *** arvind_k has quit IRC (Ping timeout: 272 seconds) 2013-05-07T01:41:57 *** arvind_khadri has joined #rtems 2013-05-07T01:42:30 *** sebhub has joined #rtems 2013-05-07T01:42:35 good morning 2013-05-07T01:42:36 *** arvind_khadri has joined #rtems 2013-05-07T02:52:57 morning 2013-05-07T04:09:36 *** arvind_khadri has quit IRC (Quit: Leaving) 2013-05-07T04:26:44 *** arvind_khadri has joined #rtems 2013-05-07T06:26:36 *** the9a3eedi has quit IRC (Ping timeout: 256 seconds) 2013-05-07T06:43:40 *** the9a3eedi has joined #rtems 2013-05-07T08:00:26 *** gedare has joined #rtems 2013-05-07T08:01:24 good day 2013-05-07T08:24:04 *** antgreen has quit IRC (Ping timeout: 272 seconds) 2013-05-07T08:25:03 *** weiY has joined #rtems 2013-05-07T08:27:47 gedare, hi 2013-05-07T08:28:41 hey weiY how is it going? 2013-05-07T08:29:37 I am doing the smp atomic ops port. now is mips 2013-05-07T08:29:51 and i found a problem about the 64bit atomic ops 2013-05-07T08:29:51 how about arm? 2013-05-07T08:29:57 yes? 2013-05-07T08:30:34 arm smp atomic is not supported by freebsd, i must learn the netbsd source code or other references 2013-05-07T08:30:44 ok. 2013-05-07T08:31:04 if the architecture is not running the 64bit mode the 64bit ops should not be supported ? 2013-05-07T08:31:14 probably 2013-05-07T08:31:32 unless the processor has implemented the op somehow 2013-05-07T08:31:38 which i think would be rare 2013-05-07T08:31:47 and if the implementation of 64bit atomic ops is not support how to do its API 2013-05-07T08:32:00 cannot 2013-05-07T08:32:23 warn or error some message when compile? 2013-05-07T08:32:43 document it and move on probably. until RTEMS supports the 64-bit mode there is no concern 2013-05-07T08:33:12 ok, i see. now all the 64bit API is not useful 2013-05-07T08:33:29 probably those processors will not be used in SMP, at least not for awhile 2013-05-07T08:34:08 yep.. for now 32-bit is the main target 2013-05-07T08:34:19 only the experimental sparc64 port runs in 64-bit mode 2013-05-07T08:34:21 and in the freebsd there is a gcc buitin macro to indicate whether is compiled for 64bit or not. and in rtems whether there is a similiar macro 2013-05-07T08:34:27 nope 2013-05-07T08:35:17 ok, i can leave it blank and fill it later 2013-05-07T08:36:22 sure. now the only targets that really matter are the ones that will be SMP-capable 2013-05-07T08:36:38 for the rest we probably will not ever execute atomic ops 2013-05-07T08:36:44 except in testcase 2013-05-07T08:37:26 yeah, arm, mips ,sparc, sparc64 is my foucs 2013-05-07T08:38:23 ok 2013-05-07T09:35:52 *** DrJoel has joined #rtems 2013-05-07T09:35:52 *** DrJoel has joined #rtems 2013-05-07T09:35:52 *** ChanServ sets mode: +o DrJoel 2013-05-07T09:36:31 * DrJoel waves good morning 2013-05-07T09:37:33 good morning 2013-05-07T09:39:23 Hmm.. I thought I updated the topic. :( 2013-05-07T09:39:41 Welcome to RTEMS (http://www.rtems.org). Please feel free to ask questions and, ifyoucan, answer those from others. Google Summer of Code 2013 is ON and RTEMS is a participating organization!See http://wiki.rtems.org/wiki/index.php/RTEMSSummerOfCode for more information. *** MENTORS WANTED *** 2013-05-07T09:39:53 Welcome to RTEMS (http://www.rtems.org). Please feel free to ask questions and, ifyoucan, answer those from others. Google Summer of Code 2013 is ON and RTEMS is a participating organization! See http://wiki.rtems.org/wiki/index.php/RTEMSSummerOfCode for more information. *** MENTORS WANTED *** 2013-05-07T09:41:14 is that a new word, "ifyoucan"? 2013-05-07T09:43:02 Welcome to RTEMS (http://www.rtems.org). Please feel free to ask questions and, if you can, answer those from others. Google Summer of Code 2013 is ON and RTEMS is a participating organization! See http://wiki.rtems.org/wiki/index.php/RTEMSSummerOfCode for more information. *** MENTORS WANTED *** 2013-05-07T09:43:44 The cut and paste on this client seems to not put a space where a line break was. So two cut and pastes later, there are three words run together. :( 2013-05-07T09:44:39 ah. 2013-05-07T09:48:31 *** antgreen has joined #rtems 2013-05-07T09:52:53 *** freenix has quit IRC (Quit: Leaving) 2013-05-07T10:11:04 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-05-07T10:36:52 *** noselasd has joined #rtems 2013-05-07T10:56:26 *** sebhub has quit IRC (Remote host closed the connection) 2013-05-07T11:01:54 *** monstr_ has quit IRC (Ping timeout: 264 seconds) 2013-05-07T11:04:12 *** DrJoel has quit IRC (Quit: Clap on! , Clap off! Clap@#&$NO CARRIER) 2013-05-07T11:17:05 *** noselasd has quit IRC (Read error: Connection reset by peer) 2013-05-07T11:21:02 *** noselasd has joined #rtems 2013-05-07T11:50:00 *** DrJoel has joined #rtems 2013-05-07T11:50:01 *** DrJoel has joined #rtems 2013-05-07T11:50:01 *** ChanServ sets mode: +o DrJoel 2013-05-07T11:50:25 * DrJoel is back 2013-05-07T12:07:30 ping gedare.. there? 2013-05-07T12:36:09 DrJoel, kinda 2013-05-07T12:36:22 slogging through some old libstdc++ code that I need to push upstream 2013-05-07T12:36:32 wanted to ask a private question 2013-05-07T12:36:36 ok 2013-05-07T12:39:55 *** S_Somani has joined #rtems 2013-05-07T12:47:02 S_Somani How are you? 2013-05-07T12:48:16 fine :) 2013-05-07T12:52:36 Any questions on your project? 2013-05-07T12:54:14 Umm.. what are the changes that might occur now that I have to link with some pre existing code? 2013-05-07T12:54:33 I was hoping you would ask that. 2013-05-07T12:56:00 so..your ideas? 2013-05-07T12:56:03 verm__ (Amar) has a configuration library in Python which is technically independent of RTEMS but he has used it on a configuration tool for the BSP specific options in the various BSP configure.ac files. These may be specified at RTEMS configuration time to tweak the build of a BSP. 2013-05-07T12:57:02 The library should be a reusable class that you can just leverage, enhance, test. It would end up as (hopefully) shared infrastructure for both types of configurations. 2013-05-07T12:57:12 One at RTEMS build. One at applciation configure/build. 2013-05-07T12:58:25 If you look far enough down the line, there could be an RTEMS "mega UI" which could let you build tools, rtems w/ BSP specific configure options, add-on libraries, and generate the aplciation configuration parameters 2013-05-07T12:59:13 i think the technical details for his tool will be similar to what you have been looking at... python-based, .ini stores parameters, feeds easily into something like wxWidget 2013-05-07T12:59:46 he is going to send an email soon to explain some more details, and plans to get the code out by the end of this month so you can study it before gsoc starts if your project goes forward. 2013-05-07T12:59:47 Yes. I suspect that if all goes well, we may find you reading .ini files for the BSP configure options by the end of the summer. 2013-05-07T13:00:01 code w/documentation which is what was slowing him 2013-05-07T13:01:52 okay. Could you please elaborate on the tasks that I would have to do to make his libraries inclusive in this project? 2013-05-07T13:04:57 gedare: any thoughts? 2013-05-07T13:05:48 my thought is that you should have "classes" to do this type of operation. His are probably farther along although you may have functionality already that he doesn't and it may be missing capabilities you will need. 2013-05-07T13:06:18 i think it won't be very clear until we have documentation and code to compare 2013-05-07T13:06:27 At this point, my impression is that it effectively replacing your new class with his and growing that one as needed 2013-05-07T13:06:48 but that's a broad swipe at it 2013-05-07T13:07:37 ok. I think I should first try and understand his code and then come up with any doubts or ideas that I may have. 2013-05-07T13:09:50 It isn't as bad as it sounds. Technically we are looking to a framework for a family of build and configuration support tools. Anything which is generally useful will be leveraged across multiple tools 2013-05-07T13:14:59 ok i submitted my code to gcc..because if i didn't who knows when it would get done. 2013-05-07T13:16:41 headed home. 2013-05-07T13:20:53 *** gedare has quit IRC (Ping timeout: 245 seconds) 2013-05-07T14:02:58 *** arvind_khadri has quit IRC (Ping timeout: 276 seconds) 2013-05-07T14:10:26 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-07T14:38:52 *** tabit_bkco has joined #rtems 2013-05-07T14:40:19 *** tabit_bkco has quit IRC (Client Quit) 2013-05-07T14:40:49 *** arvind_khadri has joined #rtems 2013-05-07T14:48:20 *** gedare has joined #rtems 2013-05-07T15:07:05 *** gedare has quit IRC (Read error: Operation timed out) 2013-05-07T15:48:19 *** gedare has joined #rtems 2013-05-07T15:55:59 http://gedare-csphd.blogspot.com/2013/05/software-licenses-with-rtems.html 2013-05-07T15:56:04 hope i did it justice 2013-05-07T16:03:13 oh my nothing like light reading. 2013-05-07T16:23:09 *** antgreen has quit IRC (Ping timeout: 258 seconds) 2013-05-07T16:30:57 *** DrJoel has quit IRC (Quit: If at first you don't succeed, skydiving is not for you) 2013-05-07T17:19:59 *** gedare has quit IRC (Read error: Operation timed out) 2013-05-07T17:47:19 *** gedare has joined #rtems 2013-05-07T18:07:36 *** gedare has quit IRC (Read error: Operation timed out) 2013-05-07T18:50:40 *** antgreen has joined #rtems 2013-05-07T20:32:21 *** antgreen has quit IRC (Ping timeout: 248 seconds) 2013-05-07T20:45:04 *** antgreen has joined #rtems 2013-05-07T20:55:26 *** weiY has quit IRC (Ping timeout: 276 seconds) 2013-05-07T21:26:13 *** gedare has joined #rtems 2013-05-07T21:32:58 *** gedare has quit IRC (Ping timeout: 245 seconds) 2013-05-07T21:35:50 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2013-05-07T21:51:02 *** arvind_khadri has joined #rtems 2013-05-07T22:55:21 *** freenix has joined #rtems 2013-05-07T23:12:24 kiwichris,hi 2013-05-07T23:37:22 moorning 2013-05-07T23:55:26 *** S_Somani has joined #rtems 2013-05-08T00:00:01 *** monstr_ has joined #rtems 2013-05-08T00:12:06 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-08T01:14:37 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2013-05-08T01:48:27 *** S_Somani has joined #rtems 2013-05-08T02:02:33 *** sebhub has joined #rtems 2013-05-08T02:04:07 good morning 2013-05-08T02:07:29 sebhub, hi 2013-05-08T02:07:55 I have moved the RSB's newlib date to 8-May-2013 to pick up the latest atexit changes 2013-05-08T02:11:46 ok, but the the relevant patch is not yet checked in 2013-05-08T02:12:55 Hmm ok. I thought Jeff had checked it in 2013-05-08T02:13:06 Which bit is missing ? 2013-05-08T02:13:26 the first, but not the second patch 2013-05-08T02:13:32 the first one is only cosmetic 2013-05-08T02:13:54 Ah ok. I see 2013-05-08T02:14:10 the Add and use _REENT_GLOBAL_ATEXIT ? 2013-05-08T02:15:29 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-08T02:17:42 yes 2013-05-08T02:19:47 Ok. I missed this with the threading 2013-05-08T02:32:39 *** arvind_khadri has joined #rtems 2013-05-08T03:32:35 *** S_Somani has joined #rtems 2013-05-08T04:05:20 *** kiwichris_ has joined #rtems 2013-05-08T04:13:03 *** rtemsLogger has joined #rtems 2013-05-08T04:14:38 *** rtemsLogger` has quit IRC (Ping timeout: 256 seconds) 2013-05-08T04:14:58 *** kiwichris has quit IRC (*.net *.split) 2013-05-08T05:13:05 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-08T05:38:14 *** monstr_ has quit IRC (Read error: Operation timed out) 2013-05-08T05:40:56 *** S_Somani has joined #rtems 2013-05-08T06:18:57 *** monstr_ has joined #rtems 2013-05-08T06:21:36 *** S_Somani has left #rtems 2013-05-08T07:18:45 *** antgreen has quit IRC (Ping timeout: 248 seconds) 2013-05-08T07:38:44 *** gedare has joined #rtems 2013-05-08T08:12:03 *** DrJoel has joined #rtems 2013-05-08T08:12:03 *** DrJoel has joined #rtems 2013-05-08T08:12:03 *** ChanServ sets mode: +o DrJoel 2013-05-08T08:14:36 sebhub: Do you want the multilib patch committed to 4.6 and newer? 2013-05-08T08:14:44 * DrJoel is waiting for an ARM maintainer to comment. 2013-05-08T08:15:23 i think 4.6 is closed 2013-05-08T08:15:51 it is 4.7 and newer 2013-05-08T08:16:14 since this is a pure rtems file, i don't think it is the domain of the arm maintainer 2013-05-08T08:16:53 the qemu support for the realview pbx a9 is really nice 2013-05-08T08:16:56 I was hoping for a multilib sanity/suggestion review from them. 2013-05-08T08:17:24 Does it support all of its peripherals? 2013-05-08T08:17:46 i am only interested in clock and console 2013-05-08T08:17:55 and smp 2013-05-08T08:18:14 it has a working mmu emulation 2013-05-08T08:18:22 including support for the exceptions 2013-05-08T08:18:40 Does qemu still simulate multiple cores by alternating cores at the instruction translation block level? 2013-05-08T08:18:42 you can also use vfp + neon instructions 2013-05-08T08:18:53 Having more instructions and mmu should help us test more 2013-05-08T08:19:03 i thinks so 2013-05-08T08:19:58 As the testing infrastructure increases and becomes more automated, I want to review all multilibs and see if we can test each multilib variant with one BSP. At least have a master table and know which ones we can cover, cover automated, etc 2013-05-08T08:20:27 on arm the realview bsp can test all multilibs 2013-05-08T08:21:11 awesome! 2013-05-08T08:22:41 i told hesham to investigate it, he has the realview pbx a9 qemu running also 2013-05-08T08:23:07 i think a hacked psim can also simulate all powerpc multilibs 2013-05-08T08:23:48 that might be a good project 2013-05-08T08:23:52 I think so also on psim. Would likely be faster but unable to do coverage. We shuold be able to do coverage using realview BSP and qemu 2013-05-08T08:24:26 sebhub, DrJoel: we are also investigating if the realview bsp can work on http://www.m5sim.org/Main_Page 2013-05-08T08:24:35 ok, i never used the ppc qemu bsp 2013-05-08T08:24:41 that simulator can boot android in single and multicore 2013-05-08T08:24:47 i don't know if they support altivec and e500 spe 2013-05-08T08:24:58 but it is slow..meant for computer architecture research ;) 2013-05-08T08:26:08 i will post any info if we get somewhere, and hopefully some scripts for building the simulator and running rtems on it. 2013-05-08T08:27:49 I am hoping the RSB will soon be able to build simulators 2013-05-08T08:28:01 I would also like to get it building add-on libraries .. 2013-05-08T08:29:24 *** weiY has joined #rtems 2013-05-08T08:29:27 maybe first it should build RTEMS? :) 2013-05-08T08:34:16 gedare it can :) 2013-05-08T08:34:49 so now i think we have two 64-bit targets in rtems? 2013-05-08T08:35:01 well, two "families" of them 2013-05-08T08:35:26 or does the a9 still 32 bit 2013-05-08T08:35:36 i thought it was 64 when i read its spec, but maybe i was hasty :) 2013-05-08T08:40:27 nvm. 2013-05-08T08:41:00 hi, I am using powerpc psim to run rtems, but it tells 'Unknow PVR value of 0x0, Please add it to '. How to resolve this? thanks. 2013-05-08T08:41:55 It seems '_read_PPC_PVR()' returns 0. 2013-05-08T08:42:05 use the rtems-testing.git:/sim-script/psim script? 2013-05-08T08:42:27 No, I just use powerpc-rtems4.11-gdb hello.exe 2013-05-08T08:42:32 you should use the script 2013-05-08T08:42:36 psim requires a device tree 2013-05-08T08:42:43 which the script i think includes 2013-05-08T08:42:49 oh, thanks. I will have a try. 2013-05-08T08:51:19 the qemu should support e500 spe. last year i saw a patch to support e500 spe in qemu. but i donot know whether this patch is merged into mainline 2013-05-08T08:51:32 psim requires a very complicated device tree. :) 2013-05-08T09:01:39 hi, I just want to use a simulator to help me test rtl code. I get stuck in how to use psim. 2013-05-08T09:02:10 './bin/psim psim.in hello.exe' . Is this the right way to use it? 2013-05-08T09:02:19 sebhub: the context validation explanation still does not say "why" we want to do this validation. also, do we have a "class" of tests that are expected not to terminate under correct operation? 2013-05-08T09:03:33 oh.. i see the test uses a timeout to detect the successful condition 2013-05-08T09:03:58 It outputs the message to a log file. 2013-05-08T09:04:00 freenix "psim -i hello.exe" 2013-05-08T09:04:05 psim -? for help 2013-05-08T09:04:34 The sim-scripts are designed to be easiest to invoke for automated batch runs. "psim *.exe" can run 100s with timeouts 2013-05-08T09:04:37 just "psim hello.exe" will cat the results to "log/hello" 2013-05-08T09:04:45 Thanks, 2013-05-08T09:04:52 but that won't help with the RTL if he is using the shell commands for it. 2013-05-08T09:04:57 It works. 2013-05-08T09:05:15 ok right. then you need the -i(nteractive) flag. 2013-05-08T09:10:55 *** Fallenou_ is now known as Fallenou 2013-05-08T09:19:06 *** zw_yao has joined #rtems 2013-05-08T09:19:18 since the timer interrupt is the only portable interrupt source on rtems we have to do the new spcontext01 test this way 2013-05-08T09:21:24 what about the default validate? loop or fatal error? 2013-05-08T09:22:00 I guess loop. When is this called? I am missing context and usage. 2013-05-08T09:22:22 its in the no_cpu header, is only for sptests/spcontext01 2013-05-08T09:23:20 currently we have no explicit test for context switching and interrupt handling 2013-05-08T09:23:24 That doesn't answer the question. I read the description. I don't know why I would call this, when and why I would call something that has two behaviors -- lock up or fatal error 2013-05-08T09:25:46 +/** 2013-05-08T09:25:48 + * @ingroup CPUContext 2013-05-08T09:25:49 + * 2013-05-08T09:25:51 + * @brief Initializes and validates the CPU context with values derived from 2013-05-08T09:25:52 + * the pattern parameter. 2013-05-08T09:25:54 + * 2013-05-08T09:25:55 + * This function will not return if the CPU context remains consistent. In 2013-05-08T09:25:57 + * case this function returns the CPU port is broken. 2013-05-08T09:25:59 + * 2013-05-08T09:26:00 + * This function is used in test sptests/spcontext01. 2013-05-08T09:26:01 + * 2013-05-08T09:26:03 + * @param[in] pattern Pattern used to generate distinct register values. 2013-05-08T09:26:04 + */ 2013-05-08T09:26:06 +void _CPU_Context_validate( uintptr_t pattern ); 2013-05-08T09:26:26 So these are only to support a test? No operational use? 2013-05-08T09:26:34 yes 2013-05-08T09:27:41 Add ONLY to the "is used sentence. The name may need to indicate test fixture somehow. Not sure. 2013-05-08T09:28:07 ok, i add a only 2013-05-08T09:30:44 what is the relevant *.t file in doc for this? 2013-05-08T09:34:55 https://www.haiku-os.org/news/2007-05-08/haiku_getting_a_freebsd_network_driver_compatibility_layer 2013-05-08T09:35:29 haiku also develop a compatibility layer for network porting from freebsd 2013-05-08T09:37:19 *** freenix has quit IRC (Remote host closed the connection) 2013-05-08T09:40:54 sebhub any chance of pushing on the v850 multilib patch 2013-05-08T09:41:20 i write a reminder for thomas 2013-05-08T09:41:29 porting/taskcontext.t 2013-05-08T09:42:14 Thanks. That will leave the wchar patch I don't understand at all the purpose of, the sparc64 patch I do understand, and the h8300 one which the h8 maintainers think is invalid 2013-05-08T09:42:57 got no answer regarding the wchar patch 2013-05-08T09:43:25 me either. The answer on the mailing list was something like "it fixes stuff that's broken" 2013-05-08T09:43:27 this patch is around since 2009 2013-05-08T09:43:53 :( Need to figure out why it exists :( 2013-05-08T09:43:53 yes, but without mentioning what is broken 2013-05-08T09:44:10 And without that, I don't push to apply it 2013-05-08T09:44:12 or was broken 2013-05-08T09:44:31 you are right. Even if broken in 2009, It may have been fixed since then 2013-05-08T09:45:12 Oddest thing someone has passed along to me in a while: http://www.legacy.com/obituaries/nytimes/obituary.aspx?n=antonia-larroux&pid=164596250#fbLoggedOut 2013-05-08T09:45:22 I am not sure if it is real, a joke or what. 2013-05-08T09:49:57 joel, what do you use for cpu bit maps, i propose to use this api and add it to newlib: http://man7.org/linux/man-pages/man3/CPU_SET.3.html 2013-05-08T09:50:37 i want to use this for a interrupt handler api extension 2013-05-08T09:51:09 e.g. for: i want to install this interrupt handler and the interrupt should fire on cpu 1, 2, and 17 2013-05-08T09:59:03 *** zw_yao has left #rtems 2013-05-08T10:01:58 I was planning on adding something derived from that. It needs to be available whether POSIX is enabled or not. 2013-05-08T10:02:31 If you add it to newlib, I will just use it in the score for both thread affinity maps and cpu online/available maps 2013-05-08T10:02:50 will you limit the number of cores to 32 so the structure is only 4 bytes long :) 2013-05-08T10:04:20 is should be defined by the newlib port, but for rtems i would start with 32 2013-05-08T10:04:35 the api is quite flexible 2013-05-08T10:04:48 used also on bsd in a variant 2013-05-08T10:05:09 i don't think we should reinvent the wheel for rtems 2013-05-08T10:06:24 I am fine with that. The BSD API is closer to what I think the POSIX group would accept as an API. 2013-05-08T10:07:25 If everyone is OK using that at the core, I am fine with it. 2013-05-08T10:07:29 pthread_attr_setaffinity_np(3), 2013-05-08T10:07:29 pthread_setaffinity_np(3) 2013-05-08T10:07:51 are NP and in pthread_np.h. What about those? We don't have a pthread_np.h. 2013-05-08T10:07:53 yes, they use also the cpu_set api 2013-05-08T10:08:14 I don't have a problem with the API, just the location and non-portable part. We have never crossed that line 2013-05-08T10:08:39 well, posix is to lame and linux is a defacto standard 2013-05-08T10:08:53 We just need to make a decision on where to prototype them 2013-05-08T10:09:08 use the linux or bsd location 2013-05-08T10:09:29 there is no posix standard for this yet 2013-05-08T10:09:41 BSD is pthread_np.h which is included from pthread.h. Do we follow that and put it in newlib? 2013-05-08T10:09:43 #define _GNU_SOURCE /* See feature_test_macros(7) */ 2013-05-08T10:09:45 #include 2013-05-08T10:10:03 sounds good 2013-05-08T10:10:14 the pthread_np.h included from pthread.h 2013-05-08T10:10:49 ok.. and wrap the include in defined(__rtems__) and provide our pthread_np.h in the rtems directory? 2013-05-08T10:11:10 FWIW I expect blowback from normal sources since this is beyond standards. 2013-05-08T10:11:55 *** arvind_khadri has quit IRC (Ping timeout: 264 seconds) 2013-05-08T10:12:42 I don't mind pushing this now in anticipation of implementing it soon. real soon. 2013-05-08T10:12:45 i will start with the most brutal approach 2013-05-08T10:15:21 If you are addressing interrupt routing, anything is better than nothing. 2013-05-08T10:15:32 Keep Daniel Hellstom in the loop. He wants to do this on LEON3/4 2013-05-08T10:16:24 sry, have to leave now 2013-05-08T10:19:20 *** sebhub has quit IRC (Remote host closed the connection) 2013-05-08T10:23:51 the api looks clean. the header filename confuses me. what does the "np" mean? 2013-05-08T10:24:33 ah, non-portable 2013-05-08T10:24:41 These functions are nonstandard GNU extensions; hence the suffix "_np" (nonportable) in the names. 2013-05-08T10:25:21 we could offer some linux-y api's in RTEMS... 2013-05-08T10:25:31 * gedare mutters and grumbles in his head 2013-05-08T10:25:38 cpukit/linux! 2013-05-08T10:40:11 *** antgreen has joined #rtems 2013-05-08T10:56:02 gedare: I believe the _np in the name is a POSIX requirement/recommendation. If approved, it can have it removed. Clearly denotes extensions 2013-05-08T10:56:23 And what would those APIs be? :) 2013-05-08T10:57:05 1) moving targets 2013-05-08T10:57:19 2) goto 1 2013-05-08T10:57:24 lol.. 2013-05-08T10:57:40 more seriously... sched_*, rbtree_* come to mind from my POV. 2013-05-08T10:57:43 I am happy to see sebhub propose adding cpuset. 2013-05-08T10:58:18 Linux sched_ beyond POSIX? 2013-05-08T10:58:31 Linux has rbtree? or do you mean making it available at rtems api level 2013-05-08T10:58:54 http://lxr.missinglinkelectronics.com/#linux/include/linux/rbtree.h 2013-05-08T10:59:28 it's somewhat less usable or featureful as the RTEMS version 2013-05-08T10:59:44 freebsd has a header-only one, but that is hard to understand..all CPP macros 2013-05-08T11:03:51 I would be prone to promote yours to user land. 2013-05-08T11:04:05 it already is 2013-05-08T11:04:48 http://git.rtems.org/rtems/tree/cpukit/sapi/inline/rtems/rbtree.inl 2013-05-08T11:05:13 and http://git.rtems.org/rtems/tree/cpukit/sapi/include/rtems/rbtree.h 2013-05-08T11:05:26 ok.. blanking out and I suppose it is in the manual also :) 2013-05-08T11:05:31 yup. 2013-05-08T11:05:43 so much code.. so little time 2013-05-08T11:06:14 it's ok. probably a "linux" api is not very useful anyway since linux is a moving target... 2013-05-08T11:06:26 and the implementation would have to be hand-written due to licensing problems. 2013-05-08T11:07:05 on a more practical note.. how should the cpuset methods behave when SMP is disabled? 2013-05-08T11:07:35 I think you can build them but setting affinity to a core doesn't make sense if there is only one. 2013-05-08T11:08:39 i'd just leave it the same...the scheduler will ignore it anyway 2013-05-08T11:08:55 only affinity-aware smp schedulers will bother to check the cpu set 2013-05-08T11:09:32 EINVAL (pthread_setaffinity_np()) cpuset specified a CPU that was out- 2013-05-08T11:09:32 side the set supported by the kernel. (The kernel configuration 2013-05-08T11:09:32 option CONFIG_NR_CPUS defines the range of the set supported by 2013-05-08T11:09:32 the kernel data type used to represent CPU sets.) 2013-05-08T11:09:33 the cpu set methods will probably need to call scheduler hooks, which may need to be invented for the purpose 2013-05-08T11:09:35 From Linux 2013-05-08T11:10:04 sure. and setting the affinity for the one core of a up is essentially a nop. 2013-05-08T11:10:29 If the score thread includes a cpuset per thread when smp is enabled, then the scheduler should be ok. I don't see a special callout at this point .. but that is subject to change. :) 2013-05-08T11:11:11 when the cpuset changes for an executing thread on any core, the scheduler may need to be invoked. 2013-05-08T11:19:38 *** antgreen has quit IRC (Ping timeout: 245 seconds) 2013-05-08T11:20:40 *** S_Somani has joined #rtems 2013-05-08T12:36:29 oof. looks like it's time for me to learn some perl 2013-05-08T12:36:49 no.. not that.. run.. 2013-05-08T12:36:57 * DrJoel is freezing.. stepping outside 2013-05-08T12:37:29 but baby, it's cold outside 2013-05-08T12:37:50 Eartha Kitt version? 2013-05-08T12:38:01 whichever one sounds the creepiest 2013-05-08T12:38:38 i should just recreate this little perl program in python, and then do the modifications I want on it 2013-05-08T12:38:39 ;) 2013-05-08T12:39:25 69 lines of perl 2013-05-08T12:50:26 Probably not much more python 2013-05-08T12:51:30 *** dividebyzero has joined #rtems 2013-05-08T12:53:24 *** dividebyzero has left #rtems 2013-05-08T12:57:53 actually i think i figured out kind of how to do what i want by wrapping the perl script in some shell code and just beating on it with a stick. 2013-05-08T13:00:15 yup.. maybe i'll blog about it. give you some more light reading :) 2013-05-08T13:00:49 *** S_Somani1 has joined #rtems 2013-05-08T13:01:17 *** S_Somani has quit IRC (Ping timeout: 246 seconds) 2013-05-08T13:06:17 one thing you didn't mention in your license post was that it's trivial for a project to move from any gpl version to a later one -- unless RTEMS has removed that, as well 2013-05-08T13:06:29 a gplv2 poject can become gplv3 at any time 2013-05-08T13:06:37 (v3 is a disaster) 2013-05-08T13:13:36 ah, yeah. though the v2 project can be forked 2013-05-08T13:13:57 but, the same is true for any permissive licensed project 2013-05-08T13:14:10 in the general sense. 2013-05-08T13:14:52 RTEMS could be migrated to a gplv3 and make use of the section 7 for special-purpose exceptions. I do not know the other ramifications 2013-05-08T13:17:45 ah. RTEMS cannot accept gplv3 for some technical issues related to the relinking requirements 2013-05-08T13:19:35 that is to say... RTEMS Project whole-heartedly supports the "tivoization" of RTEMS ;) 2013-05-08T13:21:15 gpl3 will mean that RTEMS can't be used by me or any of my customers 2013-05-08T13:21:25 right. 2013-05-08T13:21:38 nor I suspect anyone else that seriously uses RTEMS 2013-05-08T13:21:58 because of the patent thing and the drm clauses 2013-05-08T13:22:13 to say it in sort 2013-05-08T13:22:25 we can't use Uboot because of these 2013-05-08T13:22:26 * DrJoel steps on soapbox... RTEMS will never include code with a license that imposes any restrictions or burdens on end users 2013-05-08T13:22:38 The BSD code with advertising is only marginally acceptable. 2013-05-08T13:22:58 RTEMS license is "ok" as-is.. it would be better to get out of the GPL camp because of the gplv3 issues, but I think users can easily understand gplv2 sticks around still because of Linux. 2013-05-08T13:23:01 And that's because it is embedded in the FreeBSD and NetBSD code base still. They want it removed and can't either. 2013-05-08T13:23:04 because rms somehow duped the Uboot maintainers into adopting gpl3 instead of gpl2 2013-05-08T13:23:31 btw, we should collect those advertising clauses and all the copyright notices and other disclaimers and put them somewhere for users to copy from if their legal dept tells them to do so 2013-05-08T13:23:49 DrJoel: good to hear (the soapbox line) 2013-05-08T13:24:20 i'd like to put the advertising "obligation" code in a separate repo if technically possible, but that is another issue 2013-05-08T13:25:40 I agree in principle.. when to execute is another matter. After the add-on building in RSB is there would be an attractive time to do it. 2013-05-08T13:26:13 RTEMS woudl have long ago been dead if we had applied a strict GPL license. I truly believe that. 2013-05-08T13:26:30 Let me add.. for target resident code. On the host, I am more accepting 2013-05-08T13:26:50 +1 2013-05-08T13:30:15 peerst: No real details yet but there will be a class in Munich the last of June. Will you be around? 2013-05-08T13:30:40 I'm in sweden from June 11 to Jun 27th 2013-05-08T13:32:10 DrJoel: so the *last* of June I will be here 2013-05-08T13:32:16 ok. I will probably go home on the Saturday or Sunday following Thursday 27th 2013-05-08T13:32:54 * DrJoel has to be careful on travel .. I have lost 9 pounds and don't want to screw up eating heavy on the road. :) 2013-05-08T13:32:58 27th last day of class? 2013-05-08T13:33:09 28th Friday should be. 2013-05-08T13:33:09 haha Schweinsbraten 2013-05-08T13:33:40 knoedel (did I spell that right) and schnitzel.. and .. my arch enemy for weight gain.. BEER ... :( 2013-05-08T13:33:42 Schweinsbraten mit Knödel will tempt you 2013-05-08T13:34:06 yes.. Bavaria is a land of good eating :) 2013-05-08T13:34:07 yeah Knoedel is correct if you don't have ö 2013-05-08T13:42:01 thanks verm__ i have added that RTEMS uses v2 and will never use v3, and gave some short reasoning. 2013-05-08T13:54:47 DrJoel (or anyone) know if I can change ${PREFIX} of something that is already compiled, without rerunning configure and make? or..if i rerun configure, will make install work with the new prefix? 2013-05-08T13:54:54 * gedare starts googling in the meantime 2013-05-08T13:55:36 *** S_Somani1 has quit IRC (Quit: Leaving.) 2013-05-08T13:57:00 maybe i can trick it with chroot. 2013-05-08T13:57:32 sometimes you can just move the binaries. Depends on the program 2013-05-08T13:57:38 gcc ;) 2013-05-08T13:58:18 It may have changed but I think you want to recompile. You can move it and symlink old to new directory if it is a disk space thing 2013-05-08T13:58:38 it is a not wanting to recompile gcc again thing. that's all 2013-05-08T13:58:55 i agree probably best just to do that 2013-05-08T14:03:13 ok i kicked off my script 2013-05-08T14:03:29 actually, looks like just rerunning configure might be working. 2013-05-08T14:15:13 * DrJoel is heading to a meeting see you all 2013-05-08T14:15:18 *** DrJoel has quit IRC (Quit: I used to think I was indecisive, but now I'm not too sure.) 2013-05-08T14:17:19 *** monstr_ has quit IRC (Ping timeout: 264 seconds) 2013-05-08T14:46:32 gedare: sure -- all of my clients have banned gpl at the door now because of the gplv3 2013-05-08T14:47:15 it's not the same as using a permissive license, for instance going from bsd to gpl and continuing development is difficult, gpl2->3 is trivial or any other future incarnation/version 2013-05-08T14:47:49 it's not a matter of us saying we won't use gplv3 the problem is we could change our minds on a whim and do it 2013-05-08T14:56:10 i don't quite understand the difference, but ok 2013-05-08T14:56:24 i gotta go but might read back later. 2013-05-08T15:10:09 *** antgreen has joined #rtems 2013-05-08T18:17:33 antgreen, is there a patch for bintutils I am missing ? 2013-05-08T18:18:24 kiwichris_, which version of binutils? development head? 2013-05-08T18:19:08 http://www.rtems.org/pipermail/rtems-tooltestresults/2013-May/000681.html 2013-05-08T18:19:40 The config.log for libgcc said something about an option GCC was using that as did not understand 2013-05-08T18:20:13 I am yet to look at the gdb breakage 2013-05-08T18:20:29 I added bi-endian support to the toolchain some time ago 2013-05-08T18:20:36 let me look at your link 2013-05-08T18:20:41 Ok 2013-05-08T18:20:48 Hmm that would make sense 2013-05-08T18:22:50 Do you have this patch? 2013-05-08T18:22:51 http://sourceware.org/ml/binutils/2012-09/msg00161.html 2013-05-08T18:28:18 The only patch I have is http://git.rtems.org/rtems-tools.git/plain/tools/4.11/binutils/moxie/binutils-2.22-rtems4.11-moxie-20130214.diff 2013-05-08T18:28:41 Is your patch in 2.23.2 ? 2013-05-08T18:28:53 I don't know :-) 2013-05-08T18:28:57 let me download that version 2013-05-08T18:29:03 Thanks 2013-05-08T18:29:09 I committed it in september it seems 2013-05-08T18:29:32 Ok. I do not follow the binutils release process 2013-05-08T18:30:17 It would be good if the patch I just posted the link of was merged. It would make the moxie patch clena 2013-05-08T18:33:00 your patch is in the development head 2013-05-08T18:33:41 oh.. hang on 2013-05-08T18:34:23 http://sourceware.org/ml/binutils/2013-02/msg00302.html 2013-05-08T18:35:39 hang on. bathtime duty for my kiddies. 2013-05-08T18:43:35 *** gedare has quit IRC (Ping timeout: 246 seconds) 2013-05-08T18:44:11 *** weiY has quit IRC (Ping timeout: 256 seconds) 2013-05-08T19:28:55 *** the9a3eedi has joined #rtems 2013-05-08T20:01:10 *** the9a3eedi-2 has joined #rtems 2013-05-08T20:01:30 *** the9a3eedi has quit IRC (Ping timeout: 276 seconds) 2013-05-08T21:00:33 *** the9a3eedi-2 has quit IRC (Ping timeout: 252 seconds) 2013-05-08T21:22:36 *** the9a3eedi has joined #rtems 2013-05-08T21:38:34 *** freenix has joined #rtems 2013-05-08T23:14:50 *** the9a3eedi has quit IRC (Ping timeout: 256 seconds) 2013-05-08T23:31:01 *** the9a3eedi has joined #rtems 2013-05-08T23:54:32 *** monstr_ has joined #rtems 2013-05-09T00:52:44 *** arvind_khadri has joined #rtems 2013-05-09T01:06:09 *** kiwichris_ has quit IRC (Quit: This computer has gone to sleep) 2013-05-09T01:24:17 *** kiwichris has joined #rtems 2013-05-09T01:30:08 *** S_Somani has joined #rtems 2013-05-09T01:37:33 *** arvind_khadri has quit IRC (Quit: Leaving) 2013-05-09T01:39:35 *** arvind_khadri has joined #rtems 2013-05-09T01:41:24 *** arvind_khadri has joined #rtems 2013-05-09T01:42:21 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-09T01:42:31 *** arvind_khadri has joined #rtems 2013-05-09T01:49:37 *** S_Somani has left #rtems 2013-05-09T01:52:44 *** kiwichris has quit IRC (Quit: This computer has gone to sleep) 2013-05-09T02:09:07 *** kiwichris has joined #rtems 2013-05-09T02:56:54 hi,kiwichris 2013-05-09T02:57:18 ppc works now. haha 2013-05-09T03:57:26 *** the9a3eedi has quit IRC (Ping timeout: 252 seconds) 2013-05-09T05:00:45 *** the9a3eedi has joined #rtems 2013-05-09T05:48:51 *** Hesham has joined #rtems 2013-05-09T06:36:26 *** Hesham has left #rtems 2013-05-09T07:15:27 *** gedare has joined #rtems 2013-05-09T08:21:20 *** antgreen has quit IRC (Ping timeout: 255 seconds) 2013-05-09T08:25:49 morning 2013-05-09T08:26:24 freenix: are you going to have anything left to do by the time gsoc would start?? 2013-05-09T08:26:29 :) 2013-05-09T08:44:27 *** weiY has joined #rtems 2013-05-09T08:59:49 hi gedare 2013-05-09T08:59:53 howdy. 2013-05-09T09:00:39 I have done something about rtl 2013-05-09T09:01:14 i saw. i was making a joke because you have made a lot of progress already with it. 2013-05-09T09:01:19 it is good. 2013-05-09T09:01:41 if your project is accepted you will have a good start i think 2013-05-09T09:02:15 oh. haha. my poor english may catch me. 2013-05-09T09:02:43 I think the hard point is the second part of the project. 2013-05-09T09:03:32 good...that's how a good project should be! 2013-05-09T09:03:47 And I do not have the hardware or simulator to test the code, Although I am trying to use the simulators that listed in the wiki page. 2013-05-09T09:04:07 good. did you see the realview bsp that runs on qemu? 2013-05-09T09:04:12 it is new since a few days ago 2013-05-09T09:04:19 Yes. I have made it run. 2013-05-09T09:05:22 it helps me to test the relocation types and instructions which are not supported on the 6410 board. 2013-05-09T09:13:16 ok, good 2013-05-09T09:13:31 what's next on your list? 2013-05-09T09:14:27 arm mips ppc has been partly finished now. i386,sparc and m68k were implemented by chris. 2013-05-09T09:15:00 the arm and thumb switch needs some work 2013-05-09T09:15:19 I do not have a good idea about how to process the switch. 2013-05-09T09:16:32 I'll do some test to the implemented archs. And begin nios2. 2013-05-09T09:23:18 what do you mean, "switch"? 2013-05-09T09:23:34 from arm mode to thumb mode 2013-05-09T09:24:22 a function in arm mode invokes a function in thumb mode. 2013-05-09T09:24:28 ah.. like this? http://www.rtems.org/pipermail/rtems-devel/2013-May/003096.html 2013-05-09T09:24:58 but you don't know which one it is? 2013-05-09T09:25:34 I can get the relocation types. 2013-05-09T09:25:42 for example 2013-05-09T09:25:53 function func1() is in arm mode 2013-05-09T09:26:02 function func2() is in thumb mode 2013-05-09T09:26:13 if func1 invokes func2. 2013-05-09T09:26:28 ahhh. 2013-05-09T09:26:36 there will be an relocation entry for the invokation. 2013-05-09T09:26:39 tricky tricky 2013-05-09T09:26:50 the type indicates it is a call 2013-05-09T09:27:14 And now I should judge it is a call to arm or to thumb func 2013-05-09T09:27:47 i'm not familiar enough.. do you have the disassembly? 2013-05-09T09:27:54 wait 2013-05-09T09:29:03 7c: ebfffffe bl 0 2013-05-09T09:29:20 7c: R_ARM_CALL rtems_thumb 2013-05-09T09:29:43 This is the reloc entry 2013-05-09T09:30:59 some instructions can be modified from arm insn encoding to thumb insn encoding. 2013-05-09T09:31:07 but some can not 2013-05-09T09:32:00 Thus I will have to write a stub code to handle this. each such reloc entry may need a stub. 2013-05-09T09:32:35 will there be much overhead involved? 2013-05-09T09:33:11 can you do a pass over all of the functions to determine whether they are in thumb or arm mode, somehow? 2013-05-09T09:34:21 To figure out whether the func is in arm or thumb mode is an easy task 2013-05-09T09:34:58 no overhead for this. 2013-05-09T09:35:21 the lsb bit: 1 means thumb and 0 means arm 2013-05-09T09:35:45 This is done when compiling c files to object files. 2013-05-09T09:35:50 ok.. i don't understand the problem yet then 2013-05-09T09:36:15 so, function foo, and function bar... 2013-05-09T09:36:17 foo calls bar 2013-05-09T09:36:21 I am trying to see how the arm linux gcc resolve this. 2013-05-09T09:36:30 you know that foo is arm and bar is thumb function 2013-05-09T09:36:38 oh 2013-05-09T09:36:55 so, what is the problem then? 2013-05-09T09:38:41 arm to thumb, it just modifes the instruction encoding. But thumb to arm, it uses stub code. 2013-05-09T09:39:58 each stub code consumes 8 bytes. 2013-05-09T09:40:27 each such switch mode reloc types needs such a stub code. 2013-05-09T09:43:07 i see. so every time a thumb function calls an arm function, you need to insert some extra stub in the relocation? 2013-05-09T09:43:16 yes. 2013-05-09T09:43:40 I am confusing how to manage all the stub code 2013-05-09T09:44:26 how similar are the stubs? 2013-05-09T09:44:49 bx pc; nop; b armfunc 2013-05-09T09:45:11 bx pc and nop each consumes 2 bytes. 2013-05-09T09:45:19 b armfunc consumes 4 bytes 2013-05-09T09:45:39 b armfunc is what I should care about 2013-05-09T09:47:41 it would probably be a good idea to write up the problem in an email 2013-05-09T09:49:09 ok. i'll have to leave to sleep now,haha. Wish you have a good day. 2013-05-09T09:49:40 if you send an email maybe it will be answered before you wake up ;) 2013-05-09T09:50:30 :) 2013-05-09T09:50:37 *** DrJoel has joined #rtems 2013-05-09T09:50:37 *** DrJoel has joined #rtems 2013-05-09T09:50:37 *** ChanServ sets mode: +o DrJoel 2013-05-09T09:53:53 good morning DrJoel have I thoroughly spent your a.m. brain cells? 2013-05-09T09:54:55 Pretty much. I was all set to do something else when I started reading email. :) 2013-05-09T09:55:37 that's the way of it 2013-05-09T09:56:04 We need to discuss with sebhub the entire set of cpuset and affinity calls. Just at an API level. 2013-05-09T09:56:17 And we will need to include them somewhere in the manual. 2013-05-09T09:56:26 yeah makes sense. 2013-05-09T09:57:57 know if it works to specify CFLAGS on the command line when invoking make for rtems? 2013-05-09T09:58:01 * gedare about to try anyway 2013-05-09T09:58:32 THat's in the archives somewhere .. :( 2013-05-09T09:59:04 too hard to search. need gmane or a google search widget ;) 2013-05-09T09:59:27 whether it will build mips successfully when enable "--enable-smp" 2013-05-09T09:59:38 crap.. I started emailing gmane about importing everything this weekend.. Forgot to follow up Monday 2013-05-09T09:59:48 weiY: I don't think anyone has tried MIPS with SMP 2013-05-09T10:00:08 weiY there is autoconf logic to disable SMP silently unless on sparc or 386 2013-05-09T10:00:08 now i can not build successfully, and i want to test the atomic smp 2013-05-09T10:01:17 when i enable "--enable-smp" when build jmr3904 bsp it can not build capture 2013-05-09T10:02:13 mips-rtems4.11-gcc -B../../../../../jmr3904/lib/ -specs bsp_specs -qrtems -march=r3900 -Wa,-xgot -G0 -O2 -g -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -march=r3900 -Wa,-xgot -G0 -o capture.exe init.o test1.o 2013-05-09T10:02:14 (.text+0x8): undefined reference to `__stack_chk_guard' 2013-05-09T10:02:14 (.text+0xc): undefined reference to `__stack_chk_guard' 2013-05-09T10:02:14 (.text+0x68): undefined reference to `__stack_chk_guard' 2013-05-09T10:02:15 (.text+0x88): undefined reference to `__stack_chk_fail' 2013-05-09T10:02:16 (.text+0x9c): undefined reference to `__stack_chk_guard' 2013-05-09T10:02:46 I do not know those symbols. :( 2013-05-09T10:03:28 ../rtems/configure --target=mips-rtems4.11 --enable-rtemsbsp=jmr3904 --enable-tests --disable-posix --enable-smp 2013-05-09T10:03:38 this is my configure command 2013-05-09T10:03:48 It appears to be part of gcc's stack checking support. 2013-05-09T10:04:03 *** freenix has quit IRC (Remote host closed the connection) 2013-05-09T10:04:34 seems not related to smp? 2013-05-09T10:05:07 SMP was silently disabled. Check cpuopts.h 2013-05-09T10:05:30 how did you get your mips tools? 2013-05-09T10:06:25 fwiw DrJoel I have a generic freelist implementation sitting around somewhere. i'll see if it might be usable here. 2013-05-09T10:06:32 AC_ARG_ENABLE(smp, 2013-05-09T10:06:32 [AS_HELP_STRING([--enable-smp],[enable smp interface])], 2013-05-09T10:06:32 [case "${enableval}" in 2013-05-09T10:06:32 yes) case "${RTEMS_CPU}" in 2013-05-09T10:06:32 sparc|i386) RTEMS_HAS_SMP=yes ;; 2013-05-09T10:06:33 *) RTEMS_HAS_SMP=no ;; 2013-05-09T10:06:33 esac 2013-05-09T10:06:34 ;; 2013-05-09T10:07:03 Aww.. SMP is silently disabled. you need to edit all enable-smp.m4 to enable it on the mips. 2013-05-09T10:07:09 But that won't fix stack guard. 2013-05-09T10:07:10 i add the apt source list to ubuntu. and "apt-get install mip-rtems4.11*" 2013-05-09T10:07:34 Any idea where the repository is? 2013-05-09T10:07:48 mips-rtems4.11-gcc --version 2013-05-09T10:08:21 mips-rtems4.11-gcc (GCC) 4.7.2 2013-05-09T10:08:21 Copyright (C) 2012 Free Software Foundation, Inc. 2013-05-09T10:08:21 This is free software; see the source for copying conditions. There is NO 2013-05-09T10:08:21 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. 2013-05-09T10:08:42 http://rtems.org/ftp/pub/rtems/linux/4.11/ubuntu/ precise rtems4.11 2013-05-09T10:09:04 this is my apt source list content 2013-05-09T10:09:07 ok.. so Ralf built them. 2013-05-09T10:09:09 $ mips-rtems4.11-gcc --version 2013-05-09T10:09:09 mips-rtems4.11-gcc (GCC) 4.8.0 20130322 (RTEMS 4.11-RSB(df56f7ee093089081640925a9d845c10f27e0cd6-modified)-1,gcc-4.8.0/newlib-30-April-2013) 2013-05-09T10:09:28 I am building myself using the RTEMS Source Builder. Rebuilding my tools now to see if I can duplicate this 2013-05-09T10:09:32 Do any tests link? 2013-05-09T10:09:53 hello world is ok 2013-05-09T10:10:35 and before i build i also exec the "apt-get update" 2013-05-09T10:10:37 I don't see any extra options in capture/Makefile.am. 2013-05-09T10:10:55 gedare: toolset bug? 2013-05-09T10:11:01 mips-rtems4.11-gcc -B../../../../../jmr3904/lib/ -specs bsp_specs -qrtems -march=r3900 -Wa,-xgot -G0 -O2 -g -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -march=r3900 -Wa,-xgot -G0 -o capture.exe init.o test1.o 2013-05-09T10:11:06 this is link command 2013-05-09T10:11:21 weiY "make -k" .. do a lot fail? 2013-05-09T10:11:44 i did not think any apt repos were available 2013-05-09T10:12:04 i just exec make command. do i need "make -k" again? 2013-05-09T10:12:06 but yes, it sounds like the toolchain was compiled wrong 2013-05-09T10:12:07 I am surprised also 2013-05-09T10:12:40 that's what I think.. building one here using the RSB. If that works, you should build your own tools. Otherwise, I may need to report a gcc bug 2013-05-09T10:14:35 ok, i can try to build using RSB. and or should i try download the toolchain (deb) and install manually? 2013-05-09T10:15:33 RSB probably is more reliable albeit a bit more to learn how to use 2013-05-09T10:16:08 Way easier than building by hand. 2013-05-09T10:16:32 *** noselasd has quit IRC (Ping timeout: 255 seconds) 2013-05-09T10:16:38 But hold on.. let me finish building and see if it happens with gcc 4.8.0++ on CentOS. Then we know it is nothing unique to you 2013-05-09T10:17:37 ok, and got RSB from http://git.rtems.org? 2013-05-09T10:18:25 yep.. look under chris 2013-05-09T10:18:27 make -k. lots of fail with the same error like "stack check" 2013-05-09T10:18:48 http://rtems.org/ftp/pub/rtems/people/chrisj/source-builder/source-builder.html 2013-05-09T10:18:57 Docs 2013-05-09T10:19:32 It could fail the same way though. :( 2013-05-09T10:27:51 gedare: the msp430 student didn't apply to gsoc, did they? I don't recall seeing it 2013-05-09T10:28:41 they did not 2013-05-09T10:28:44 i'm just relaying the info 2013-05-09T10:33:49 *** noselasd has joined #rtems 2013-05-09T10:34:58 *** antgreen has joined #rtems 2013-05-09T10:44:06 *** arvind_khadri has quit IRC (Ping timeout: 276 seconds) 2013-05-09T11:17:30 *** S_Somani has joined #rtems 2013-05-09T11:19:42 *** antgreen has quit IRC (Remote host closed the connection) 2013-05-09T11:19:52 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-09T11:22:36 *** antgreen has joined #rtems 2013-05-09T11:25:00 *** weiY has quit IRC (Ping timeout: 268 seconds) 2013-05-09T11:37:14 *** weiY has joined #rtems 2013-05-09T11:43:51 *** weiY has quit IRC (Ping timeout: 252 seconds) 2013-05-09T11:47:46 *** weiY has joined #rtems 2013-05-09T11:56:56 *** S_Somani has joined #rtems 2013-05-09T12:04:44 DrJoel: Here is some light reading for lunch; http://gedare-csphd.blogspot.com/2013/05/whos-calling-me-visualizing-function.html 2013-05-09T12:21:06 *** monstr_ has quit IRC (Ping timeout: 256 seconds) 2013-05-09T12:56:18 *** MegaAlex|away has joined #rtems 2013-05-09T12:59:35 *** MegaAlex|away is now known as MegaAlex 2013-05-09T13:51:00 *** S_Somani has left #rtems 2013-05-09T13:54:44 That's interesting! Would that be useful for analysis purposes? I assume it generated dot files which we could parse for bad patterns if we had any known. 2013-05-09T13:55:14 FYI I could not duplciate that mips problem. I think it is either 4.7.x (I use 4.8++) or a packaging issue with Ralf's debian stuff 2013-05-09T13:57:03 The graph for each scheduler would be interesting. And separating the qres library from regular RTEMS (if possible). 2013-05-09T13:57:14 The various schedulers should look similar. 2013-05-09T14:08:12 *** weiY has quit IRC (Ping timeout: 256 seconds) 2013-05-09T14:12:02 *** weiY has joined #rtems 2013-05-09T14:23:01 you could infer some general ideas about connectedness/complexity I suppose 2013-05-09T14:23:22 I developed the approach to create a visual to convince someone that a piece of code is terribly complex. :) 2013-05-09T14:23:34 but thought the scheduler is a nice way to demo it for the public. 2013-05-09T14:26:56 Is this the plugin someone was discussing on the mailing list? 2013-05-09T14:50:13 no 2013-05-09T15:20:51 my mips jmr3904 built correctly. Looks like wiey was suffering from a packaging introduced bug 2013-05-09T15:22:25 *** gedare has quit IRC (Ping timeout: 268 seconds) 2013-05-09T16:39:30 *** antgreen has quit IRC (Ping timeout: 268 seconds) 2013-05-09T16:48:59 *** gedare has joined #rtems 2013-05-09T16:56:15 *** gedare has quit IRC (Ping timeout: 252 seconds) 2013-05-09T17:32:51 *** DrJoel has quit IRC (Quit: If at first you don't succeed, skydiving is not for you) 2013-05-09T19:03:17 *** the9a3eedi has quit IRC (Ping timeout: 252 seconds) 2013-05-09T19:07:59 *** antgreen has joined #rtems 2013-05-09T19:41:09 *** MegaAlex is now known as MegaAlex|away 2013-05-09T20:03:33 i download the RSB one night, it still do not complete, so slow internet speed 2013-05-09T20:03:42 i have to go, byb 2013-05-09T20:08:05 *** weiY has quit IRC (Ping timeout: 252 seconds) 2013-05-09T20:41:22 *** freenix has joined #rtems 2013-05-09T23:56:15 *** monstr_ has joined #rtems 2013-05-10T00:12:57 *** monstr_ has quit IRC (Remote host closed the connection) 2013-05-10T00:19:07 *** monstr has joined #rtems 2013-05-10T02:06:25 *** sebhub has joined #rtems 2013-05-10T02:06:52 good morning 2013-05-10T03:05:33 sebhub, hi 2013-05-10T03:06:16 sebhub, I am doing a 'bootstrap -p' and see an extra \n at the end of the file on FreeBSD. Do you see this on Linux ? 2013-05-10T03:06:45 yes, this is a strange problem i have for a couple of days now 2013-05-10T03:06:52 didn't investigate it further 2013-05-10T03:06:57 ok. 2013-05-10T03:07:16 the build system is a don't touch area currently 2013-05-10T03:07:21 The other issue is the generated files depend on the order of ls 2013-05-10T03:07:39 yes, this leads to conflicts on different hosts 2013-05-10T03:07:56 Well my 'sb-bootstrap' is nice performance win but this is a pain. 2013-05-10T03:08:09 did you get the qemu running? 2013-05-10T03:08:12 Yes 2013-05-10T03:08:29 the freebsd 'qemu-devel' port works 2013-05-10T03:08:55 just 'portinstall emulation/qemu-devel' and it is built from source and works 2013-05-10T03:09:06 nice 2013-05-10T03:09:17 Yeah.Ticker works. 2013-05-10T03:09:40 I have not tried the zync machine in qemu yet 2013-05-10T03:09:40 you can run all tests with the rtems-testing framework 2013-05-10T03:09:46 Nice. 2013-05-10T03:10:06 i want to work on the smp startup today 2013-05-10T03:10:55 nice 2013-05-10T03:11:13 try sb-bootstrap from the rsb from the top of the RTEMS source tree 2013-05-10T03:13:57 just push a change to match the output better 2013-05-10T03:16:29 Looks like the commit from WieY (9b605b28) is the reason 2013-05-10T03:38:36 *** rtemsLogger` has joined #rtems 2013-05-10T03:38:48 *** kiwichris_ has joined #rtems 2013-05-10T03:44:49 verm__, about ? 2013-05-10T03:46:19 *** kiwichris has quit IRC (*.net *.split) 2013-05-10T03:46:19 *** rtemsLogger has quit IRC (*.net *.split) 2013-05-10T03:46:56 sebhub, http://git.rtems.org/rtems/ is that the last commit ? 2013-05-10T03:47:33 no 2013-05-10T03:47:47 What is this ? 2013-05-10T03:47:58 hm, this is up to date: http://git.rtems.org/rtems/log/ 2013-05-10T03:48:04 If you click on the [?] you get the what the log shows in git 2013-05-10T04:00:25 *** rtemsLogger has joined #rtems 2013-05-10T04:01:18 *** kiwichris_ has joined #rtems 2013-05-10T04:01:18 *** monstr has joined #rtems 2013-05-10T04:01:18 *** freenix has joined #rtems 2013-05-10T04:01:18 *** antgreen has joined #rtems 2013-05-10T04:01:18 *** MegaAlex|away has joined #rtems 2013-05-10T04:01:18 *** noselasd has joined #rtems 2013-05-10T04:01:18 *** peerst has joined #rtems 2013-05-10T04:01:18 *** lcpfnvcy has joined #rtems 2013-05-10T04:01:18 *** ChanServ has joined #rtems 2013-05-10T04:01:18 *** verm__ has joined #rtems 2013-05-10T04:01:18 *** rokka_ has joined #rtems 2013-05-10T04:01:18 *** budrys has joined #rtems 2013-05-10T04:01:18 *** dhananjay has joined #rtems 2013-05-10T04:01:18 *** Fallenou has joined #rtems 2013-05-10T04:01:18 *** sendak.freenode.net sets mode: +o ChanServ 2013-05-10T04:01:37 *** sebhub has quit IRC (Read error: Operation timed out) 2013-05-10T04:01:55 *** rtemsLogger` has quit IRC (Ping timeout: 246 seconds) 2013-05-10T04:14:45 *** Fallenou_ has joined #rtems 2013-05-10T04:16:12 *** kiwichris__ has joined #rtems 2013-05-10T04:21:47 *** Fallenou has quit IRC (Ping timeout: 264 seconds) 2013-05-10T04:21:48 *** kiwichris_ has quit IRC (Ping timeout: 264 seconds) 2013-05-10T04:44:58 *** MegaAlex|away is now known as MegaAlex 2013-05-10T05:23:00 *** sebhub has joined #rtems 2013-05-10T05:32:50 *** MegaAlex is now known as MegaAlex|away 2013-05-10T06:02:10 *** MegaAlex|away has quit IRC (Ping timeout: 256 seconds) 2013-05-10T06:15:05 the sb-bootstrap is not faster than the normal bootstrap script (i use --jobs 24) 2013-05-10T07:00:35 *** the9a3eedi has joined #rtems 2013-05-10T07:22:47 *** S_Somani has joined #rtems 2013-05-10T07:37:36 *** S_Somani has left #rtems 2013-05-10T07:52:54 *** gedare has joined #rtems 2013-05-10T08:02:04 *** weiY has joined #rtems 2013-05-10T08:03:02 *** MegaAlex|away has joined #rtems 2013-05-10T08:10:07 *** MegaAlex|away has quit IRC (Ping timeout: 256 seconds) 2013-05-10T08:13:29 hi, chris 2013-05-10T08:13:56 i saw there is a problem which seems about me "Looks like the commit from WieY (9b605b28) is the reason" 2013-05-10T08:14:06 what is the problem about? 2013-05-10T08:14:38 problem is solved, there was a problematic newline at the end of some files 2013-05-10T08:15:15 is it caused by my patch? 2013-05-10T08:15:37 it seem so 2013-05-10T08:16:33 and how to solve it? i just want to understand this problem 2013-05-10T08:17:01 i have no idea how this could happen, the preinstall.am are generated by a script 2013-05-10T08:17:22 what version of autoconf do you have weiY ? 2013-05-10T08:17:43 I noticed that discrepancy also when I looked at your patch. I thought I fixed it, or maybe I introduced it :) 2013-05-10T08:18:01 hmm, actually... the script calls other things.. 2013-05-10T08:18:05 one of ralf's little tools 2013-05-10T08:18:22 autoconf (GNU Autoconf) 2.69 2013-05-10T08:18:23 Copyright (C) 2012 Free Software Foundation, Inc. 2013-05-10T08:18:23 License GPLv3+/Autoconf: GNU GPL version 3 or later 2013-05-10T08:18:23 , 2013-05-10T08:18:23 This is free software: you are free to change and redistribute it. 2013-05-10T08:18:23 There is NO WARRANTY, to the extent permitted by law. 2013-05-10T08:18:23 Written by David J. MacKenzie and Akim Demaille. 2013-05-10T08:18:59 sry, preinstall.am is not from autoconf 2013-05-10T08:19:24 ok, and could you show me the differece about this issue 2013-05-10T08:19:45 http://git.rtems.org/rtems/commit/?id=3806e01a77e8929f222f9224665cea5c8ec3b969 2013-05-10T08:21:12 for some reason, your patch did not have a blank line at the end of each preinstall.am file 2013-05-10T08:22:01 it does not make much sense, but is hard to debug 2013-05-10T08:22:11 ok i have seen. it lost a blank line. and in pre-patch i just modify the preinstall.am manually. so i want to know whether this is right way 2013-05-10T08:22:18 *** antgreen has quit IRC (Ping timeout: 264 seconds) 2013-05-10T08:22:52 and if preinstall.am lost a blank line what problem will casue 2013-05-10T08:26:13 ok, i know "see an extra \n at the end of the file on FreeBSD". strange 2013-05-10T08:26:29 *** the9a3eedi has quit IRC (Ping timeout: 248 seconds) 2013-05-10T08:29:37 yeah , i have check other preinstall.am and there is really a blank line in the last line of preinstall.am file. so i think i forgot insert a blank line in the last line of preinstall.am file 2013-05-10T08:34:54 hi, are there any simulators for bfin and nios2 that can run rtems now? 2013-05-10T08:35:48 weiY: the script should do it automatically for you. 2013-05-10T08:35:57 freenix: i think the bfin-gdb can 2013-05-10T08:36:19 bfin-rtems4.11-gdb? 2013-05-10T08:36:20 i'm not sure about nios2 though. i think most of that development was done with the board. 2013-05-10T08:36:21 yeah 2013-05-10T08:36:36 i thought it has a gdb simulator... i could be wrong though 2013-05-10T08:37:07 any arguments should be passed to gdb? I have tried gdb and skyeye for bfin, but failed. 2013-05-10T08:37:27 hmm 2013-05-10T08:38:13 I follow this page http://docs.blackfin.uclinux.org/doku.php?id=toolchain:sim 2013-05-10T08:38:28 bfin-elf-run --env operating --model bf537 ./hello.exe 2013-05-10T08:38:29 skyeye has some known (and unknown) problems 2013-05-10T08:38:58 I find that some instructions can not be decoded on skyeye. 2013-05-10T08:39:29 i think you might want to try rtems-testing.git: /sim-scripts/bf537Stamp-gdb 2013-05-10T08:39:53 ok. I will take a try now. 2013-05-10T08:42:16 gedare: you mean "bootstrap" script to generate preinstall.am? 2013-05-10T08:43:52 yes, actually it is ampolish3 2013-05-10T08:44:02 and what is input file to generate preinstall.am 2013-05-10T08:44:08 Makefile.am? 2013-05-10T08:44:11 yep 2013-05-10T08:44:30 the ampolish3 script is a custom script. what platform are you using? 2013-05-10T08:45:06 ok, i will do it next time. but if the preinstall.am is generate automatically why it is include into git? 2013-05-10T08:45:33 ubuntu now. and i use fedora in the past days 2013-05-10T08:45:35 gedare: bf537Stamp-gdb hello.exe failed 2013-05-10T08:45:48 what error? 2013-05-10T08:45:54 gedare: _cec_raise: double fault at 0xffb00000 ! :( 2013-05-10T08:46:03 because i use fedora14 but it seems not support now by RTMES 2013-05-10T08:46:03 ok i will try 2013-05-10T08:46:24 weiY, yes the fedora moves too fast.. you have to keep upgrading or you get left behind 2013-05-10T08:46:29 ubuntu should work 2013-05-10T08:46:36 gedare: It seems .l1lcode section is not loaded. 2013-05-10T08:46:44 weiY: because it is not regenerated very often, only when header file changes 2013-05-10T08:46:45 yeah, so now i have switch to ubuntu 2013-05-10T08:47:34 ok, and if the preinstall.am is not included into git what problem will cause? 2013-05-10T08:47:57 freenix: what BSP did you build 2013-05-10T08:49:27 weiY: everyone has to just run bootstrap -p before bootstrap 2013-05-10T08:49:38 usually bootstrap does not regen the preinstall.am 2013-05-10T08:49:54 preinstall.am files were kept in CVS repo, so we keep them in Git repo also 2013-05-10T08:51:03 yeah, i understand. this is not a problem to new user which clone the git repo. but it will casue the exist user to run bootstrap -p every update time 2013-05-10T08:53:00 *** freenix has quit IRC (Remote host closed the connection) 2013-05-10T08:55:24 *** freenix has joined #rtems 2013-05-10T08:55:55 gedare: my computer crashed just now, sorry. I used eZKit533 2013-05-10T08:57:53 freenix: I just used the bf537Stamp bsp and ran it with the sim-script 2013-05-10T08:58:02 there are some complaints by gdb, but it does run 2013-05-10T08:58:49 my configure paraments:../rtems/configure --prefix=/opt/rtems-4.11 --target=bfin-rtems4.11 --enable-rtemsbsp="eZKit533" --enable-tests=samples --enable-cxx --enable-maintainer-mode --enable-rtems-debug 2013-05-10T08:59:03 use --enable-rtemsbsp=bf537Stamp 2013-05-10T08:59:12 um 2013-05-10T08:59:13 hang on 2013-05-10T08:59:15 ok 2013-05-10T08:59:18 there is also an ezkit533-gdb 2013-05-10T08:59:21 yo ucan try to use that one 2013-05-10T08:59:25 in sim-scripts 2013-05-10T08:59:38 Thanks. 2013-05-10T09:21:54 hi gedare, I tried bf537stamp just now. timer is not supported? I tried ticker.exe, it stops at TA3. 2013-05-10T09:24:11 dunno. did you try ezkit533-gdb? 2013-05-10T09:26:30 something is broken there. the time is advancing, but the timer did not fire. 2013-05-10T09:26:42 I tried ezkit533-gdb. _cec_raise: double fault at 0xffb00000 ! 2013-05-10T09:28:06 ok dunno what to say about that one. 2013-05-10T09:31:35 i think for the bf537Stamp, if we can set the BSPOPTS for "BFIN_ON_SKYEYE" that will force the clock driver to use SIM_FAST_IDLE 2013-05-10T09:31:44 but, i don't really know how to set BSPOPTS :P 2013-05-10T09:31:59 on skyeye? 2013-05-10T09:32:27 no no it is an option 2013-05-10T09:32:33 that controls the clock for bfin 2013-05-10T09:32:43 oh 2013-05-10T09:32:43 there should be a better way though... 2013-05-10T09:33:04 I am not familar with bfin dsp. 2013-05-10T09:33:36 yeah i understand 2013-05-10T09:35:50 It means the bfin gdb should be recompiled with that BFIN_ON_SKYEYE? 2013-05-10T09:39:53 maybe. i'm trying a workaround now to see if it helps 2013-05-10T09:40:00 if not..you can ask on rtems-devel if anyone knows 2013-05-10T09:40:07 Thanks. 2013-05-10T09:40:10 *** antgreen has joined #rtems 2013-05-10T09:46:36 i'm not sure freenix you shall have to ask 2013-05-10T09:46:52 ok 2013-05-10T09:47:17 on user list or devel list? 2013-05-10T09:47:48 for this, probably user is good 2013-05-10T09:48:00 ok 2013-05-10T09:49:54 *** freenix has quit IRC (Quit: Leaving) 2013-05-10T10:09:14 *** DrJoel has joined #rtems 2013-05-10T10:09:14 *** DrJoel has joined #rtems 2013-05-10T10:09:14 *** ChanServ sets mode: +o DrJoel 2013-05-10T10:09:36 sebhub: arm and v850 patches should be committed to 4.7, 4.8 and head. Please confirm for me. 2013-05-10T10:14:15 thanks joel 2013-05-10T10:14:36 no problem. We just need to work through the rest 2013-05-10T10:14:55 Just let me know quickly if they are ok. With gcc 4.8.1 soon, we don't want to screw that up 2013-05-10T10:15:19 i have to leave for the train in one minute 2013-05-10T10:15:45 ok.. have a nice weekend! 2013-05-10T10:15:53 cu and have a nice weekend 2013-05-10T10:16:09 *** sebhub has quit IRC (Remote host closed the connection) 2013-05-10T10:29:30 *** monstr has quit IRC (Ping timeout: 252 seconds) 2013-05-10T10:47:22 *** phipse has joined #rtems 2013-05-10T10:47:25 hi 2013-05-10T10:47:33 hey.. nice email 2013-05-10T10:47:47 gedare: awake? 2013-05-10T10:48:04 gedare: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57237 2013-05-10T10:48:14 DrJoel: what do you think of the approach? 2013-05-10T10:48:31 DrJoel: yes trying to wrap things up since next week is going to be crazy 2013-05-10T10:48:49 hahah 2013-05-10T10:49:37 That patch has been in the RPMs for 10 months!!! 2013-05-10T10:49:49 doesn't matter no one reads the gcc bugzilla anymore ;) 2013-05-10T10:49:59 i finally subscribed to gcc-patches.. 2013-05-10T10:50:02 I don't want to be an A^% but it is either good enough to commit to gcc or should be thrown away 2013-05-10T10:50:07 i have a filter for "rtems" in place 2013-05-10T10:50:36 what email? was it offlist 2013-05-10T10:51:15 No.. read the bug comments.. would have gone to gcc-bugs 2013-05-10T10:53:35 yes i saw that... i was wondering what you meant by "nice email." 2013-05-10T10:54:13 sorry.. I broke out in hives/rash and got flush and hot about an hour ago. Still not 100%. :( 2013-05-10T10:54:40 the nice email was phipse's to Pok. Well written description of his process and challenges 2013-05-10T10:54:59 "well written" << thanks :) 2013-05-10T10:58:26 Did you figure out how to debug inside a partition with symbols ? 2013-05-10T10:59:36 heading to lunch.. bbl 2013-05-10T11:00:13 I will just take the original elf file and not the one in partitions.bin 2013-05-10T11:04:09 what'd you eat! 2013-05-10T11:10:08 ok. i'm not on pok's ml 2013-05-10T11:12:37 gedare: the main part of the email was this: http://wiki.rtems.org/wiki/index.php/GSOC_2013_-_Paravirtualization_of_RTEMS#Build_process 2013-05-10T11:19:43 I want to implement storage handling in L4RTEMS. can you hint me at an starting point to read up about the device handling? 2013-05-10T11:31:54 Do I just tell RTEMS this memory chunk is a block device and IMFS should be used? 2013-05-10T12:36:42 i don't know i have not done too much with I/O in rtems 2013-05-10T12:45:41 i think it should work to use imfs though 2013-05-10T12:45:56 that's mainly what i have used for file i/o in rtems 2013-05-10T12:47:05 *** kkthecoder has joined #rtems 2013-05-10T13:09:18 gedare grilled chicken on a sald.. plenty of hot sauce and blue cheese dressing. Low carb 2013-05-10T13:09:27 i meant before you got flushed 2013-05-10T13:09:38 i should eat better..i was for about 2 weeks. ;) 2013-05-10T13:09:47 IMFS is pure RAM -- mallocs file storage 2013-05-10T13:10:11 i'm out for the night. bye 2013-05-10T13:10:15 You would be looking at say RTEMS File System and the RAM block device driver 2013-05-10T13:10:20 night.. see you Tuesday 2013-05-10T13:12:54 *** gedare has quit IRC (Read error: Operation timed out) 2013-05-10T13:13:35 ok, I will have a look at it. 2013-05-10T13:14:47 I'm out too, bye. 2013-05-10T13:15:38 *** phipse has left #rtems 2013-05-10T13:28:18 *** MegaAlex|away has joined #rtems 2013-05-10T13:28:21 *** MegaAlex|away is now known as MegaAlex 2013-05-10T15:14:58 *** kkthecoder has quit IRC (Ping timeout: 256 seconds) 2013-05-10T15:57:47 *** DrJoel has quit IRC (Quit: I used to think I was indecisive, but now I'm not too sure.) 2013-05-10T16:28:38 *** antgreen has quit IRC (Ping timeout: 252 seconds) 2013-05-10T19:30:35 *** the9a3eedi has joined #rtems 2013-05-10T20:02:20 Hi chris, are you around 2013-05-10T20:02:48 how to rebuild RSB with reuse the source already downloaded. 2013-05-10T20:03:22 i encounter a error and i want to rebuild RSB and do not want to re-download the source file 2013-05-10T20:13:19 *** MegaAlex has quit IRC (*.net *.split) 2013-05-10T20:13:20 *** noselasd has quit IRC (*.net *.split) 2013-05-10T20:13:46 *** noselasd has joined #rtems 2013-05-10T20:18:04 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-05-10T20:21:40 *** freenix has joined #rtems 2013-05-10T20:24:33 *** MegaAlex has joined #rtems 2013-05-10T20:33:19 *** rtemsLogger` has joined #rtems 2013-05-10T20:37:20 *** rtemsLogger has joined #rtems 2013-05-10T20:37:36 *** noselasd has quit IRC (Ping timeout: 245 seconds) 2013-05-10T20:39:28 *** rokka has joined #rtems 2013-05-10T20:40:24 *** Fallenou has joined #rtems 2013-05-10T20:41:04 *** Fallenou_ has quit IRC (Ping timeout: 268 seconds) 2013-05-10T20:41:09 *** rokka_ has quit IRC (Ping timeout: 268 seconds) 2013-05-10T20:44:12 *** rtemsLogger` has quit IRC (*.net *.split) 2013-05-10T20:55:26 *** rtemsLogger` has joined #rtems 2013-05-10T20:56:09 *** kiwichris_ has joined #rtems 2013-05-10T20:56:47 *** rtemsLogger has quit IRC (*.net *.split) 2013-05-10T21:03:03 *** kiwichris__ has quit IRC (*.net *.split) 2013-05-10T21:09:35 *** budrys has quit IRC (Ping timeout: 245 seconds) 2013-05-10T21:22:12 *** budrys has joined #rtems 2013-05-10T21:25:16 *** antgreen has joined #rtems 2013-05-10T21:27:05 *** kiwichris__ has joined #rtems 2013-05-10T21:29:52 *** kiwichris_ has quit IRC (Ping timeout: 268 seconds) 2013-05-10T21:53:18 *** MegaAlex|away has quit IRC (*.net *.split) 2013-05-10T22:00:23 *** MegaAlex|away has joined #rtems 2013-05-11T05:11:44 *** S_Somani has joined #rtems 2013-05-11T06:55:13 *** MegaAlex|away is now known as MegaAlex 2013-05-11T07:27:49 *** freenix has quit IRC (Remote host closed the connection) 2013-05-11T08:24:21 *** weiY has quit IRC (Ping timeout: 248 seconds) 2013-05-11T08:25:16 *** weiY has joined #rtems 2013-05-11T08:29:07 *** S_Somani has left #rtems 2013-05-11T08:30:58 *** antgreen has quit IRC (Ping timeout: 245 seconds) 2013-05-11T08:45:45 *** antgreen has joined #rtems 2013-05-11T09:42:47 *** antgreen has quit IRC (Ping timeout: 272 seconds) 2013-05-11T09:46:47 *** budrys__ has joined #rtems 2013-05-11T09:47:18 *** budrys has quit IRC (Read error: Connection reset by peer) 2013-05-11T10:22:12 *** antgreen has joined #rtems 2013-05-11T10:28:54 *** antgreen has quit IRC (Ping timeout: 264 seconds) 2013-05-11T10:31:07 *** antgreen has joined #rtems 2013-05-11T10:33:14 *** verka has joined #rtems 2013-05-11T11:18:24 *** budrys__ has quit IRC (Ping timeout: 245 seconds) 2013-05-11T11:18:28 *** budrys has joined #rtems 2013-05-11T11:47:25 *** S_Somani has joined #rtems 2013-05-11T12:41:17 *** weiY has quit IRC (Ping timeout: 256 seconds) 2013-05-11T13:08:08 *** dhananja1 has joined #rtems 2013-05-11T13:14:22 *** ChanServ has quit IRC (*.net *.split) 2013-05-11T13:14:22 *** verka has quit IRC (*.net *.split) 2013-05-11T13:14:22 *** dhananjay has quit IRC (*.net *.split) 2013-05-11T13:14:23 *** verm__ has quit IRC (*.net *.split) 2013-05-11T13:31:18 *** ChanServ has joined #rtems 2013-05-11T13:31:18 *** rajaniemi.freenode.net sets mode: +o ChanServ 2013-05-11T13:32:30 *** antgreen has quit IRC (Ping timeout: 264 seconds) 2013-05-11T13:44:15 *** ChanServ has quit IRC (*.net *.split) 2013-05-11T13:47:52 *** antgreen has joined #rtems 2013-05-11T13:49:43 *** ChanServ has joined #rtems 2013-05-11T13:49:43 *** rajaniemi.freenode.net sets mode: +o ChanServ 2013-05-11T14:14:19 *** ChanServ has quit IRC (*.net *.split) 2013-05-11T14:30:06 *** antgreen has quit IRC (Ping timeout: 264 seconds) 2013-05-11T14:42:41 *** antgreen has joined #rtems 2013-05-11T15:58:25 *** S_Somani has quit IRC (Quit: Leaving.) 2013-05-11T16:27:12 *** ChanServ has joined #rtems 2013-05-11T16:27:12 *** rajaniemi.freenode.net sets mode: +o ChanServ 2013-05-11T16:27:18 *** ChanServ has quit IRC (shutting down) 2013-05-11T17:01:34 *** budrys__ has joined #rtems 2013-05-11T17:01:46 *** kiwichris_ has joined #rtems 2013-05-11T17:07:58 *** ChanServ has joined #rtems 2013-05-11T17:07:58 *** rajaniemi.freenode.net sets mode: +o ChanServ 2013-05-11T17:10:02 *** budrys has quit IRC (*.net *.split) 2013-05-11T17:10:02 *** kiwichris__ has quit IRC (*.net *.split) 2013-05-11T17:22:46 *** verm__ has joined #rtems 2013-05-11T17:49:38 *** MegaAlex has quit IRC (Read error: Operation timed out) 2013-05-11T17:56:23 *** MegaAlex has joined #rtems 2013-05-11T18:00:42 *** MegaAlex has quit IRC (Read error: Operation timed out) 2013-05-11T18:07:22 *** MegaAlex has joined #rtems 2013-05-11T18:27:32 *** ChanServ has quit IRC (*.net *.split) 2013-05-11T18:28:43 *** rtemsLogger has joined #rtems 2013-05-11T18:30:02 *** rokka has joined #rtems 2013-05-11T18:31:24 *** Fallenou has joined #rtems 2013-05-11T18:31:24 *** noselasd_ has joined #rtems 2013-05-11T20:04:22 *** rokka has quit IRC (Ping timeout: 245 seconds) 2013-05-11T20:04:57 *** rokka has joined #rtems 2013-05-11T20:09:25 *** MegaAlex is now known as MegaAlex|away 2013-05-11T21:46:59 *** weiY has joined #rtems 2013-05-11T21:59:37 *** weiY has quit IRC (Ping timeout: 246 seconds) 2013-05-11T22:09:15 *** ChanServ has quit IRC (*.net *.split) 2013-05-11T22:40:05 *** ChanServ has joined #rtems 2013-05-11T22:40:05 *** gibson.freenode.net sets mode: +o ChanServ 2013-05-11T23:27:24 *** weiY has joined #rtems 2013-05-12T01:55:58 *** S_Somani has joined #rtems 2013-05-12T01:59:48 *** S_Somani1 has joined #rtems 2013-05-12T02:01:50 *** S_Somani has quit IRC (Ping timeout: 256 seconds) 2013-05-12T02:29:22 *** S_Somani1 has left #rtems 2013-05-12T02:39:57 *** monstr has joined #rtems 2013-05-12T05:33:27 *** freenix has joined #rtems 2013-05-12T06:56:43 *** MegaAlex|away is now known as MegaAlex 2013-05-12T07:39:56 *** S_Somani has joined #rtems 2013-05-12T08:23:47 *** gedare has joined #rtems 2013-05-12T08:55:37 *** S_Somani has left #rtems 2013-05-12T08:58:53 *** freenix has quit IRC (Remote host closed the connection) 2013-05-12T10:17:26 *** Fallenou_ has joined #rtems 2013-05-12T10:18:02 *** noselasd has joined #rtems 2013-05-12T10:24:51 *** Fallenou has quit IRC (*.net *.split) 2013-05-12T10:24:51 *** noselasd_ has quit IRC (*.net *.split) 2013-05-12T10:32:12 *** MegaAlex has quit IRC (Ping timeout: 256 seconds) 2013-05-12T10:36:35 *** MegaAlex has joined #rtems 2013-05-12T10:56:21 *** weiY has quit IRC (Ping timeout: 256 seconds) 2013-05-12T11:19:47 *** S_Somani has joined #rtems 2013-05-12T11:50:01 *** Fallenou_ is now known as Fallenou 2013-05-12T12:59:11 *** gedare has quit IRC (Ping timeout: 245 seconds) 2013-05-12T13:32:35 *** S_Somani has left #rtems 2013-05-12T18:21:48 *** gedare has joined #rtems 2013-05-12T19:06:02 *** the9a3eedi has joined #rtems 2013-05-12T20:23:25 *** freenix has joined #rtems 2013-05-12T21:02:13 *** the9a3eedi has quit IRC (Ping timeout: 248 seconds) 2013-05-12T21:04:34 *** the9a3eedi has joined #rtems 2013-05-12T21:05:53 Hi. I'm having a bit of trouble figuring out how to use the RTEMS linker. I get the error "error: undefines: symbol not found: rtems" ... kinda stumped. 2013-05-12T21:06:40 I've specified the compiler, executable image, and the object file I want to turn into a library. Aren't the rtems symbols supposed to be extracted from the executable image? 2013-05-12T21:12:24 as something to load via the runtime loader ? 2013-05-12T21:15:23 *** MegaAlex is now known as MegaAlex|away 2013-05-12T21:20:12 yeah 2013-05-12T21:21:32 What is the rtems-ld command ? 2013-05-12T21:21:48 what arch 2013-05-12T21:25:39 my command is rtems-ld -b ..\FileTransferTest.ex 2013-05-12T21:25:40 e -o hellomodule.rap --cc C:\opt\rtems-4.10\bin\sparc-rtems4.10-gcc.exe hellomodule.o -v -v -v 2013-05-12T21:26:56 Can you please paste the errors to a pastebin type web site ? 2013-05-12T21:42:33 *** gedare has quit IRC (*.net *.split) 2013-05-12T21:47:53 *** gedare has joined #rtems 2013-05-12T21:51:21 *** gedare has quit IRC (*.net *.split) 2013-05-12T21:52:42 *** gedare has joined #rtems 2013-05-12T22:03:57 *** the9a3eedi has quit IRC (Ping timeout: 252 seconds) 2013-05-12T23:33:26 *** the9a3eedi has joined #rtems 2013-05-12T23:49:49 *** monstr__ has joined #rtems