2015-03-02T00:03:10 *** monstr has joined #rtems 2015-03-02T00:21:32 *** kiwichris has quit IRC () 2015-03-02T00:28:48 *** zoso has joined #rtems 2015-03-02T00:44:24 *** beng-nl_ has joined #rtems 2015-03-02T00:45:35 *** beng-nl has quit IRC (Read error: Connection reset by peer) 2015-03-02T02:40:31 *** Hesham has joined #rtems 2015-03-02T02:45:19 *** lynx` has quit IRC (Remote host closed the connection) 2015-03-02T02:50:33 [trac] #2270/enhancement (accepted) updated by sebastian.huber (From the SPARC Architecture Manual Version 8 we have: "PSR_enable_floating-point ...) http://devel.rtems.org/ticket/2270 2015-03-02T03:01:39 *** Davidbrcz has joined #rtems 2015-03-02T03:16:04 *** beng-nl_ is now known as beng-nl 2015-03-02T03:22:21 *** zoso has quit IRC (Ping timeout: 252 seconds) 2015-03-02T03:22:48 [trac] #2270/enhancement (accepted) updated by daniel (I guess STFSR instruction is in a way what PPC call context synchronizing. I don't ...) http://devel.rtems.org/ticket/2270 2015-03-02T03:26:13 [trac] #2270/enhancement (accepted) updated by daniel (The LEON4-N2X and GR740 puts part of the FPU into power-down when PSR.EF is cleared.) http://devel.rtems.org/ticket/2270 2015-03-02T03:57:32 *** zoso has joined #rtems 2015-03-02T05:24:35 [git] 4ccb79b by Sebastian Huber: [score: Enable SMP barriers for all configurations] http://devel.rtems.org/changeset/4ccb79b/rtems 2015-03-02T06:13:44 [trac] #2270/enhancement (accepted) updated by sebastian.huber (Are the FPU instructions really that slow? I would disable the PSR.EF in the ...) http://devel.rtems.org/ticket/2270 2015-03-02T06:13:52 *** Davidbrcz has quit IRC (Ping timeout: 240 seconds) 2015-03-02T06:31:02 *** itazzz has joined #rtems 2015-03-02T06:47:29 *** itazzz is now known as ita 2015-03-02T08:24:27 *** zoso has quit IRC (Ping timeout: 245 seconds) 2015-03-02T08:28:10 *** Hesham has quit IRC (Ping timeout: 255 seconds) 2015-03-02T08:41:11 *** Hesham has joined #rtems 2015-03-02T08:46:16 *** zoso has joined #rtems 2015-03-02T08:58:46 *** Hesham has quit IRC (Ping timeout: 265 seconds) 2015-03-02T09:07:48 *** Hesham has joined #rtems 2015-03-02T09:11:13 *** zoso has quit IRC (Ping timeout: 264 seconds) 2015-03-02T09:15:39 *** dr__house|away is now known as dr__house 2015-03-02T09:16:15 *** DrJoel has joined #rtems 2015-03-02T09:16:15 *** DrJoel has joined #rtems 2015-03-02T09:16:15 *** ChanServ sets mode: +o DrJoel 2015-03-02T09:17:08 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2015-03-02T09:43:35 *** Hesham has joined #rtems 2015-03-02T09:54:25 *** antgreen has quit IRC (Ping timeout: 264 seconds) 2015-03-02T10:08:57 *** Hesham has quit IRC (Ping timeout: 244 seconds) 2015-03-02T10:21:36 *** Hesham has joined #rtems 2015-03-02T10:30:25 *** Hesham has quit IRC (Ping timeout: 264 seconds) 2015-03-02T10:41:04 *** Hesham has joined #rtems 2015-03-02T11:49:10 *** monstr has quit IRC (Remote host closed the connection) 2015-03-02T11:52:50 [trac] #2278/defect (v:4.11) created by joel.sherrill (GCC FORTRAN Test Output Failures) http://devel.rtems.org/ticket/2278 2015-03-02T12:02:34 *** Davidbrcz has joined #rtems 2015-03-02T12:04:24 quiet in here today. 2015-03-02T12:05:07 We were accepted into the ESA Summer of Code in Space. Waiting to hear about GSOC in a couple of hours before I update the topic. 2015-03-02T12:13:13 there is a ESA SOC ? 2015-03-02T12:13:29 didn't know that 2015-03-02T12:28:02 You have to be a university student in a member country. The exact list of countries and rules is on their site. 2015-03-02T12:33:40 *** ita is now known as ita|zzz 2015-03-02T13:01:22 Congratulation on being accepted to GSoC and ESA SoC 2015-03-02T13:02:11 *** gedare has joined #rtems 2015-03-02T13:02:11 *** ChanServ sets mode: +o gedare 2015-03-02T13:02:29 hey DrJoel back home now? 2015-03-02T13:07:18 Yep. First day at work and very little of my voice is back yet 2015-03-02T13:08:33 *** DrJoel changes topic to "Welcome to RTEMS (http://www.rtems.org). We have been ACCEPTED as an organization in both the 2015 Google Summer of Code and ESA Summer of Code in Space student programs. Our Open Projects page is https://devel.rtems.org/wiki/Developer/OpenProjects. High priority projects are bolded and space related projects are italicized. Please ask questions as needed on IRC or mailing lists." 2015-03-02T13:08:45 I guess that counts as an announcement. :) 2015-03-02T13:10:41 gedare.. do we have anything we want to change in our GSoC profile now that we are accepted? 2015-03-02T13:10:57 heh. 2015-03-02T13:11:02 And the application template is now open. What did we put in it last year? 2015-03-02T13:11:07 ooof. 2015-03-02T13:11:32 I don't know anything I would change in the previously entered information but the last tab is new. Please take a look. 2015-03-02T13:11:40 looking. 2015-03-02T13:14:01 what am i looking for? 2015-03-02T13:14:32 "Preferences" tab? 2015-03-02T13:15:41 *** kiran4399 has joined #rtems 2015-03-02T13:16:09 Yes. That one is new. 2015-03-02T13:16:13 Is Ben Gras active ? 2015-03-02T13:16:19 ping beng-nl 2015-03-02T13:16:47 He boxes and I think he works out a lot :) But that may not be what you are asking LOL 2015-03-02T13:16:58 i was thinking the same thing Joel. 2015-03-02T13:17:16 hmm my gdocs are a mess. looking for past gsoc stuff. 2015-03-02T13:19:12 Glad the girls did well on their doctor visit. :) 2015-03-02T13:20:04 heh thanks. 2015-03-02T13:20:16 i can't find any docs related to this preferences tab 2015-03-02T13:20:25 i can't even find our application doc though 2015-03-02T13:20:37 oh i just found that. 2015-03-02T13:21:07 *** DrJoel_ has joined #rtems 2015-03-02T13:21:12 Weird. Not sure what just happened. 2015-03-02T13:21:13 https://docs.google.com/document/d/1ECsmPB0ngIrVNXHlJFz0ArwN6GfkDURvpKKVHIl_0vo/edit 2015-03-02T13:22:46 ok we should copy this and update the dates i suppose. 2015-03-02T13:23:56 Do we put it in the template or just a URL to a Gdoc to start from? 2015-03-02T13:24:19 URL 2015-03-02T13:24:27 put... 2015-03-02T13:24:41 a link in the page at https://devel.rtems.org/wiki/GSoC/2015 2015-03-02T13:24:46 and put a URL to that page 2015-03-02T13:25:12 ok.. hold on 2015-03-02T13:25:26 i've got last years preferences tab open now 2015-03-02T13:25:30 i can copy-paste from there. 2015-03-02T13:25:35 we should probably capture it somewhere too. 2015-03-02T13:26:10 verm__ shared an RTEMS folder some time ago. kiwichris had one too. 2015-03-02T13:26:30 verm__'s had a GSOC subdirectory. hmm 2015-03-02T13:26:34 OK 2015-03-02T13:26:53 I just copied this to a GSOC2015 folder and renamed it 2015-03-02T13:27:13 https://docs.google.com/document/d/1F5XCodvX8AYNqWX5ssu7dfjkmFT__83uf8ABKbB_Pkg/edit 2015-03-02T13:27:34 I marked same permissions as 2014 doc so should be ok for you. And students to copy 2015-03-02T13:31:49 [trac] GSoC/2015 edited by joel.sherrill (Add link to proposal template) http://devel.rtems.org/wiki/GSoC/2015 2015-03-02T13:34:20 ok i also stashed a copy in the folder verm__ had shared previously. it has chris, sebastian, you, me, and him sharing in it. 2015-03-02T13:35:29 great. Just sent the announcement 2015-03-02T13:35:37 Ok 2015-03-02T13:37:11 did you fix up the dates yet? 2015-03-02T13:37:58 Where? 2015-03-02T13:38:24 Ahhh.. project deliverables.. no 2015-03-02T13:39:58 ok i'll do it quick 2015-03-02T13:40:13 and then i should try to work... debugging something tricky at the moment. 2015-03-02T13:40:20 i updated the preferences tab 2015-03-02T13:40:30 mentors can start to join now? 2015-03-02T13:41:48 I just did it. double check me. 2015-03-02T13:41:52 I think so. 2015-03-02T13:42:07 Why don't you post asking for mentors. Everything shouldn't come from me. 2015-03-02T13:42:15 beng-nl .. time to register 2015-03-02T13:44:20 ahh i was wondering why all the dates looked right! 2015-03-02T13:44:26 sure i'll make that post in a bit. 2015-03-02T13:44:52 Post GSOC - What will you do? 2015-03-02T13:44:55 is that new? i like it. 2015-03-02T13:45:33 Was there last year 2015-03-02T13:46:12 did we ever resolve on a CoC? 2015-03-02T13:46:33 dates look good i fixed one only. 2015-03-02T13:46:52 Code of Conduct? 2015-03-02T13:47:06 No one except he who shall no be named commented 2015-03-02T13:47:17 I say post it to the wiki 2015-03-02T13:54:55 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-02T13:55:11 Where would it be put on the top page? 2015-03-02T13:56:16 [trac] WikiStart edited by joel.sherrill (Add Code of Conduct link) http://devel.rtems.org/wiki/WikiStart 2015-03-02T13:58:40 [trac] CodeOfConduct created http://devel.rtems.org/wiki/CodeOfConduct 2015-03-02T13:59:09 Try that gedare. See where I added it to the top page and how the layout moved over from gdocs 2015-03-02T14:01:10 [trac] GSoC edited by gedare (chop out student proposals section, it goes into year-specific pages now.) http://devel.rtems.org/wiki/GSoC 2015-03-02T14:02:05 looking now. btw, i don't think a link to the RTEMS Users Manuals is necessarily a good thing to put in the RTEMS SOC Page... 2015-03-02T14:02:15 We don't really expect students to read them, certainly not all of them. 2015-03-02T14:02:31 OK.. clobber it.. if they can't find them, then we don't need them 2015-03-02T14:03:09 [trac] GSoC edited by gedare (Remove link to User Manuals) http://devel.rtems.org/wiki/GSoC 2015-03-02T14:04:49 [trac] WikiStart edited by gedare (Move code of conduct link into list of similar links.) http://devel.rtems.org/wiki/WikiStart 2015-03-02T14:05:01 i moved the coc link i dno't think it needs to be quite so prevalent 2015-03-02T14:05:37 THat's fine. I just needed it somewhere to put a link to create the page 2015-03-02T14:05:38 the page itself looks nice. 2015-03-02T14:06:23 ok i'm done on this for today 2015-03-02T14:08:19 Thanks. I just wanted to make a quick pass. We had done a lot of work before this. 2015-03-02T14:08:37 Thanks for the reminder on the CoC. It was a discussion that was shutdown using normal tactics 2015-03-02T14:09:45 syeah 2015-03-02T14:09:56 ugh, 89 unread patches. ;0 2015-03-02T14:10:06 although really just one beast left for me to review. the drvmgr 2015-03-02T14:10:24 i hope to work through it this week. we'll see. 2015-03-02T14:14:26 Many of These have been pending since before Daniel's wife had their child 2015-03-02T14:14:51 We all didn't get to the PCI and drvmgr before that so the backlog got worse 2015-03-02T14:16:14 yeah i know 2015-03-02T14:19:25 It was unfortunate but well beyond our control. I think we eliminated 40 patches before he took leave 2015-03-02T14:21:03 any clue if there are any blockers for a 4.11 release? 2015-03-02T14:21:13 at this point i suspect it wouldn't happen this summer.. 2015-03-02T14:22:03 no.. just the tickets and I am trying not to add any 2015-03-02T14:22:29 I am adding tickets here and there but setting the milestone after 4.11 2015-03-02T14:23:02 is there any reason we can't just move all the tickets to 4.11.1 wholesale? 2015-03-02T14:23:16 at this rate there will always be 'one more important' patch 2015-03-02T14:24:23 I don't know. Need to see what's left for code. The server issues are definitely 4.11 blockers 2015-03-02T14:24:23 it would be good if the build errors are fixed. 2015-03-02T14:24:37 But this is a devel@ discussion 2015-03-02T14:25:20 can we discuss it there? if we can get it down to 2 tickets and focus on those that would be good 2015-03-02T14:25:27 I was going to leave a build of all BSPs going overnight. 2015-03-02T14:25:40 Doing it by email would be the right thing to do. 2015-03-02T14:26:10 at the very least i'd like to move the default milestone for new ticket sto 4.11.1 anyone have any objections to that? 2015-03-02T14:26:11 We should try to branch before GSoC gets moving. 2015-03-02T14:30:39 that would be good 2015-03-02T14:30:44 i was thinking of adding some waf gsoc projects 2015-03-02T14:30:54 Go ahead. 2015-03-02T14:31:00 i will have phabricator up this week 2015-03-02T14:32:09 Awesome! 2015-03-02T14:32:13 verm__: yes bump the default milestone 2015-03-02T14:32:18 i had to write support for buildbot 2015-03-02T14:32:21 gedare: great 2015-03-02T14:32:34 i've done what i can to close tickets that had patches/code on them 2015-03-02T14:32:42 so when someone submits a review buildbot will build it automatically 2015-03-02T14:32:51 looking at them, none are really deal-breakers, but we should get the build problems fixed 2015-03-02T14:33:05 we shouldn't cut a release with 3 archs not building 2015-03-02T14:33:14 is it worth it to just move them all to 4.11.1 and then move ones that need to be kept back? 2015-03-02T14:33:19 well, according to open tickets, 5-6 archs aren't building 2015-03-02T14:33:22 and focus on getting it building? 2015-03-02T14:33:25 no 2015-03-02T14:33:40 i'd rather bump to 4.11.1 than to back-bump 2015-03-02T14:33:50 but i can maybe do some of that now 2015-03-02T14:34:00 p.s. is our SSL cert fixed now? 2015-03-02T14:34:24 you can check that yourself but i sent 3 emails to the list about it 2015-03-02T14:34:32 Which architectures have issues? Is this all libdl related? 2015-03-02T14:34:41 i haven't noticed any problems recently. i did see your email about getting a new cert up 2015-03-02T14:34:42 it's been fixed for over a week now 2015-03-02T14:34:52 ok i'll close that ticket then. 2015-03-02T14:34:55 well getting the new cert up required rebuilding everything 2015-03-02T14:35:05 i'll close it 2015-03-02T14:35:09 forgot there was a ticket 2015-03-02T14:35:09 I was behind a draconian firewall last week and didn't notice any issues except that the students had internet access and I didn't. 2015-03-02T14:35:13 ok 2015-03-02T14:35:28 oh 2015-03-02T14:35:34 [trac] #2195/infra (closed) updated by amar (This was fixed last week. We received a free certificate from Globalsign (thanks!)) http://devel.rtems.org/ticket/2195 2015-03-02T14:35:44 verm__: github still isn't updating automagically, right? 2015-03-02T14:36:04 [trac] #2206/infra (closed) updated by amar (Updating automatically via Buildbot (not public). It may miss an update here or ...) http://devel.rtems.org/ticket/2206 2015-03-02T14:36:04 we should at least manually bump it now, and again around gsoc coding start time 2015-03-02T14:36:32 weird wonder why it says rtems was only updated 26 days ago 2015-03-02T14:38:12 and DrJoel_ you should do a check whether any warnings need to be fixed 2015-03-02T14:40:00 Yes. But we are so far down from before that unless it is in a core place, I am prone to be forgiving. 2015-03-02T14:40:01 i only recently built a sparc toolchain... i haven't been building/using rtems in awhile. 2015-03-02T14:40:15 sure. 2015-03-02T14:40:22 maybe we should aim for 0-warnings in dot releases. :0 2015-03-02T14:40:51 That reminds me that one of the printf() warning sources is in gcc's newlib-stdint.h and he whose name shall not be mentioned shut down the discussion on that one and no one else piped up 2015-03-02T14:41:06 phabricator will fix that thankfully, it won't let commits be added until it's error (and in the future, warning) free 2015-03-02T14:42:09 :) Chris adding libdl broke some targets. I think we may have to disable it for some architectures or make you force it on to work on it. I don't see it getting fixed on all 2015-03-02T14:42:14 soon enough I mean 2015-03-02T14:42:45 [trac] #1247/defect (assigned) updated by gedare (bump milestone) http://devel.rtems.org/ticket/1247 2015-03-02T14:42:45 [trac] #2002/defect (assigned) updated by gedare (bump milestone) http://devel.rtems.org/ticket/2002 2015-03-02T14:42:45 [trac] #2189/defect (new) updated by gedare (bump milestone) http://devel.rtems.org/ticket/2189 2015-03-02T14:42:45 [trac] #1523/defect (assigned) updated by gedare (bump milestone) http://devel.rtems.org/ticket/1523 2015-03-02T14:42:45 [trac] #1747/defect (accepted) updated by gedare (bump milestone) http://devel.rtems.org/ticket/1747 2015-03-02T14:42:46 [trac] #2147/defect (new) updated by gedare (bump milestone) http://devel.rtems.org/ticket/2147 2015-03-02T14:42:46 [trac] #2158/defect (reopened) updated by gedare (bump milestone) http://devel.rtems.org/ticket/2158 2015-03-02T14:42:47 [trac] #2159/defect (reopened) updated by gedare (bump milestone) http://devel.rtems.org/ticket/2159 2015-03-02T14:42:47 [trac] #2173/defect (reopened) updated by gedare (bump milestone) http://devel.rtems.org/ticket/2173 2015-03-02T14:44:42 [trac] #2277/defect (closed) updated by gedare (dupe of #2195) http://devel.rtems.org/ticket/2277 2015-03-02T14:45:19 [trac] #2236/defect (accepted) updated by amar (I'll fix this.) http://devel.rtems.org/ticket/2236 2015-03-02T14:45:45 gedare: https://devel.rtems.org/ticket/2277 2015-03-02T14:45:48 this is not a duplicate 2015-03-02T14:45:57 oh.. oops 2015-03-02T14:46:03 it is a separate issue with ssl? 2015-03-02T14:46:12 it's a problem on Chris' local machine 2015-03-02T14:46:17 oh. 2015-03-02T14:46:18 [trac] #2277/defect (reopened) updated by amar (Not a duplicate.) http://devel.rtems.org/ticket/2277 2015-03-02T14:46:33 [trac] #2277/defect (closed) updated by amar (This is a local error on your machine. Please check your certificate CA install.) http://devel.rtems.org/ticket/2277 2015-03-02T14:47:26 verm__: close 2225 2015-03-02T14:48:04 *** Hesham has quit IRC (Ping timeout: 250 seconds) 2015-03-02T14:48:07 [trac] #2213/defect (new) updated by gedare (empty comment) http://devel.rtems.org/ticket/2213 2015-03-02T14:48:24 [trac] #2225/defect (closed) updated by amar (Since nobody has objected I will close this.) http://devel.rtems.org/ticket/2225 2015-03-02T14:50:26 [trac] #2228/defect (new) updated by gedare (bump milestone) http://devel.rtems.org/ticket/2228 2015-03-02T14:50:26 [trac] #2263/defect (new) updated by gedare (bump milestone) http://devel.rtems.org/ticket/2263 2015-03-02T14:50:31 grrrr 2015-03-02T14:50:33 [trac] #2183/defect (new) updated by joel.sherrill (I don't see this workflow item. Do I only see it if I own a ticket?) http://devel.rtems.org/ticket/2183 2015-03-02T14:50:59 [trac] #2228/defect (new) updated by gedare (adjust from 5.0 to 4.11.1) http://devel.rtems.org/ticket/2228 2015-03-02T14:50:59 [trac] #2263/defect (new) updated by gedare (adjust from 5.0 to 4.11.1) http://devel.rtems.org/ticket/2263 2015-03-02T14:56:12 DrJoel_: did you mean to comment on #2225 instead of #2183? 2015-03-02T14:57:15 [trac] #2225/defect (closed) updated by joel.sherrill (I tried to change the assigned from sebastian to unassigned on 2183 and don't see ...) http://devel.rtems.org/ticket/2225 2015-03-02T14:57:31 Yes. I was going backwards and forwards reading the ticket 2015-03-02T14:57:33 ahhh i see. 2015-03-02T14:57:52 ok well i'm bootstrap'ping RTEMS for the first time in a long time. 2015-03-02T14:57:59 :P 2015-03-02T14:58:12 do you know off-hand a good way to make RSB build with a custom newlib? 2015-03-02T14:58:23 i need to test some patches i wrote 2015-03-02T15:00:05 ahh i can add the patches directly there. 2015-03-02T15:02:44 *** kiwichris has joined #rtems 2015-03-02T15:02:57 And there's the man to blame 2015-03-02T15:03:09 heh. i'm trying to reverse engineer it. 2015-03-02T15:03:13 maybe i should read the docs 2015-03-02T15:06:42 Docs.. docs.. we don't need no stinking docs 2015-03-02T15:20:31 urgh, i think i put the patch commands in the wrong config file. *sigh* 2015-03-02T15:31:24 kiwichris: ping.. how can I tell my RSB to apply a local patch to newlib? I tried %patch add newlib /home/gedare/work/rtems/patches/newlib/0001-rtems-make-MAXNAMLEN-match-with-NAME_MAX.patch inrtems-gcc-4.9.2-newlib-2.2.0-1-1.cfg but the filepath got rejected. 2015-03-02T15:31:48 gedare, hi 2015-03-02T15:31:56 ooo 2015-03-02T15:32:01 put it in my patches directory i see 2015-03-02T15:32:09 and there will be magic to find it there 2015-03-02T15:32:15 The "approved hack" (cause I have no other way) is to add it to the patch fir 2015-03-02T15:32:27 gotcha. i finally resorted to reading the doco ;0 2015-03-02T15:32:36 hehe you found the hack 2015-03-02T15:33:25 I wonder if 'file:' is supported ... 2015-03-02T15:33:45 ita|zzz, hi 2015-03-02T15:34:10 %patch add newlib %{rtems_newlib_patches}/0001-rtems-make-MAXNAMLEN-match-with-NAME_MAX.patch 2015-03-02T15:34:12 that should work? 2015-03-02T15:34:27 ifi put the patch in patches? 2015-03-02T15:34:28 *tries* 2015-03-02T15:36:18 [trac] #2253/defect (new) updated by joel.sherrill (memsize is referenced in start.S for all mips BSPs as a .global but is never ...) http://devel.rtems.org/ticket/2253 2015-03-02T15:38:18 gedare, is the patch in the latest snapshot from Jeff ? 2015-03-02T15:38:37 I am happy to bump the snapshot version. 2015-03-02T15:39:10 Also there is support to pull patches from some web based git interfaces. 2015-03-02T15:40:00 And yes the local patch cache is always checked first then the hash checked if defined 2015-03-02T15:40:47 DrJoel_, looks like gdb-7.9 builds for MSYS2 windows for the ARM at least. 2015-03-02T15:43:19 That's great! 2015-03-02T15:43:34 I think I have the mips dl tests build issue resolved. :) 2015-03-02T15:44:04 A symbol was marked .global in start.S files but never referenced or defined 2015-03-02T15:46:27 kiwichris: no it is a local patch i am testing 2015-03-02T15:46:55 ugh. i need to apply it against newlib.git instead. 2015-03-02T15:46:55 gedare, ah ok, then the hack is the only way. 2015-03-02T15:47:00 ok gotcha. 2015-03-02T15:47:09 now i have to hunt out the right config to use the git head 2015-03-02T15:47:13 gedare, there are configs for newlib bit 2015-03-02T15:47:15 git 2015-03-02T15:47:21 yeah i've seen 'em 2015-03-02T15:47:26 just have to figure out how to use them 2015-03-02T15:47:48 gedare, the URL reads like a list of git commands 2015-03-02T15:48:11 The reset etc is to bring the repo back to the original so you can apply patches 2015-03-02T15:49:04 DrJoel_, tripped over a minor waf vs MSYS2 issue with rtems-tools. 2015-03-02T15:49:06 ok good. i updated the .cfg file for newlib.git now i just need to point to it somewhere i guess. 2015-03-02T15:49:31 The base URL should be in one of the configs 2015-03-02T15:49:55 got it.. 2015-03-02T15:50:47 in the bset file i change the newlib version to -git-1. 2015-03-02T15:51:00 and then cross my fingers. :0 2015-03-02T15:51:34 good enough to hack my way through. probably need a better flow if i end up doing this daily though. 2015-03-02T15:51:49 thanks for the hints 2015-03-02T15:51:59 and for the RSB tool ;) 2015-03-02T15:52:25 :-) 2015-03-02T15:52:32 the RSB tool rocks 2015-03-02T15:53:39 +1 2015-03-02T15:53:40 beng-nl: someone was looking for you earlier. but they disappeared. 2015-03-02T15:53:55 beng-nl you can try to register as a mentor now. 2015-03-02T15:54:07 FYI Network Time made it this year also 2015-03-02T15:54:16 did they miss last year? 2015-03-02T15:54:20 yes 2015-03-02T15:54:24 i see 2015-03-02T15:54:49 kiwichris: it worked for me. 2015-03-02T15:55:03 Transition from NTP to NT Foundation with multiple projects resulted in NTF not having a good open projects page. That was what I was told. 2015-03-02T15:55:15 i see. 2015-03-02T15:55:16 That is also why I think it is imperative to polish the Open Projects page each year 2015-03-02T15:55:32 gedare: hmm ok 2015-03-02T15:55:32 i try to come back to it a few times a year myself. 2015-03-02T15:55:47 *** DrJoel_ has quit IRC (Read error: Connection reset by peer) 2015-03-02T15:55:50 *** DrJoel has quit IRC (Read error: Connection reset by peer) 2015-03-02T15:55:51 gedare: i have been emailing a bit with some people, hopefully that overlaps with this person 2015-03-02T15:56:06 *** DrJoel has joined #rtems 2015-03-02T15:56:06 *** DrJoel has joined #rtems 2015-03-02T15:56:06 *** ChanServ sets mode: +o DrJoel 2015-03-02T15:56:16 Now I am me again. :) 2015-03-02T15:56:43 *** Hesham has joined #rtems 2015-03-02T15:59:05 Hesham.. you may meet the rules for ESA SOCIS.. not sure.. check rules for countries/students eligibility 2015-03-02T15:59:40 i don't think UK participaets? 2015-03-02T16:00:25 I get confused thought about when the UK does and does not join in with the rest of Europe. 2015-03-02T16:00:37 ok gotta go, bye 2015-03-02T16:00:40 I am not sure. The language is: "You hereby represent, warrant, and agree that: (a) you are enrolled in or accepted into an accredited institution including but not necessarily limited to a college, university, masters program, PhD program and/or undergraduate program; (b) the institution you are enrolled in is based in in one of the ESA funding states that are allowed to participate to the ESA basic Technology Research Programme; (c) you are eligib 2015-03-02T16:00:41 *** gedare has quit IRC (Quit: Leaving) 2015-03-02T16:00:52 I don't know what that does to UK or Canada 2015-03-02T16:01:43 DrJoel: What is the ESA funding states? 2015-03-02T16:02:02 UK is not part of EU I think 2015-03-02T16:06:35 UK is not but ESA membership is not all EU and not strictly EU. 2015-03-02T16:06:36 http://www.esa.int/About_Us/Welcome_to_ESA/New_Member_States 2015-03-02T16:06:55 UK is a member state of ESA as best I can tell. 2015-03-02T16:07:05 Yes I found out that now 2015-03-02T16:07:24 man. makes me want to enroll into a masters/phd program :-) 2015-03-02T16:07:42 That's good actually because I found another GSoC project that I was going do it part of my research anyway ;) 2015-03-02T16:07:59 ESA SOCIS != GSOC 2015-03-02T16:08:30 are socis and gsoc running at the same time this year? 2015-03-02T16:08:37 Yeah 2015-03-02T16:08:47 But now I can join both I think 2015-03-02T16:08:53 Announed within hours of both. 2015-03-02T16:08:57 you can apply for both but not do both 2015-03-02T16:09:01 Bold of you. :) 2015-03-02T16:09:14 the time commitment would be far too much to do both 2015-03-02T16:09:16 You are a candidate for both but these are supposed to be considered full time jobs 2015-03-02T16:09:40 yeah 2015-03-02T16:09:42 I asked about ESA because it is sometimes hard to get applicants. World is candidate for GSOC. 2015-03-02T16:09:44 I was going to work in parallel, they are both part of my research proposal anyway. 2015-03-02T16:10:06 Aha I see 2015-03-02T16:10:16 the problem with doing both is our commitment to both GSoC or SOCIS 2015-03-02T16:10:24 DrJoel, MSYS2 has a couple of issues, one is the temp file name which limits the RSB is a single core and using the mingw64/python2 implementation interacts with waf in an interesting way. 2015-03-02T16:10:29 GSoC payment is what I would consider great summer intern funding here. 2015-03-02T16:10:44 we wouldn't want google or the esa to think we're not taking it seriously by allowing someone to do both, they need to be commited to one or the other 2015-03-02T16:11:20 I wasn't going to say we wouldn't accept the same person to RTEMS for both programs. It is too much. 2015-03-02T16:11:38 It isn't my decision alone for RTEMS but it is extremely unwise and I wouldn't personally approve of it. 2015-03-02T16:12:05 neither 2015-03-02T16:12:24 Both proposals are not for RTEMS 2015-03-02T16:12:28 having said that applying for both and accepting one or the other if you're accepted to both is OK :) 2015-03-02T16:13:02 One is rump kernels, and the other is for lowRISC (I am designing RISC-V ISA hardware currently) 2015-03-02T16:13:10 Does anyone know if the RPi v2 has JTAG on a connector ? The circuit shows it wired to a connector but I could not find a suitable mechanical pic showing the connectors. 2015-03-02T16:14:22 verm__: I don't think it makes difference for me. The usual student is doing school work + GSoC project. I am fully devoted to research (I don't attend lectures or have any other commitments that can waste my time). 2015-03-02T16:15:05 So, if I was accepted on one project, I would still work on the other even it wasn't part of GSoC/ESA 2015-03-02T16:15:26 Because that's what I proposed part of my MSc anyway 2015-03-02T16:18:21 Hesham: i understand, what i'm saying is i don't think it's appropriate to do both -- if we had a student who applied for GSoC and got accepted for SOCIS on another project i'd ask them to choose 2015-03-02T16:18:31 since GSoC (and SOCIS) are considered 40 hour a week minimum commitments 2015-03-02T16:18:51 to do both would be 80 hours a week minimum for the duration of both that's too much 2015-03-02T16:21:10 I see 2015-03-02T16:21:37 I will have to discuss it then with my supervisor 2015-03-02T16:22:12 *** jhollister has joined #rtems 2015-03-02T16:22:57 Hesham: to be cliear it's OK to apply for both just not accept both to work on at the same time. :) 2015-03-02T16:23:46 i think a lot of people were asking the ESA to move it to the summer.. originally it was around the holidays to allow for what you're suggesting but applications were low 2015-03-02T16:24:38 Yeah, that what I will discuss with my supervisor, which project to pick up if both were accepted. 2015-03-02T16:34:58 BTW, I tried to run ticker.exe for Xilinx Zynq BSP recently on QEMU and it didn't output anything. However, realview with the same QEMU args executes normally. 2015-03-02T16:40:54 *** krzysiekm13 has joined #rtems 2015-03-02T16:41:22 hi 2015-03-02T16:47:19 well I am now out of here. Maybe kiwichris can hang and answer question 2015-03-02T16:47:24 *** DrJoel has quit IRC (Quit: A day without sunshine is like .... night) 2015-03-02T16:50:45 *** krzysiekm13 has quit IRC (Quit: Leaving) 2015-03-02T17:12:08 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2015-03-02T17:35:22 *** Davidbrcz has quit IRC (Ping timeout: 255 seconds) 2015-03-02T17:41:58 *** kiwichris has quit IRC (Read error: Connection reset by peer) 2015-03-02T17:42:15 *** kiwichris has joined #rtems 2015-03-02T17:43:49 kiwichris: pong 2015-03-02T17:44:06 *** ita|zzz is now known as ita 2015-03-02T17:44:53 ita, hi 2015-03-02T17:45:21 ita, just testing a full tools build on MSYS2 with python.org python2 2015-03-02T17:50:03 kiwichris: yes 2015-03-02T17:53:39 ita, I will take a look at what is needed in waf if anything. 2015-03-02T17:54:11 ita, the MSYS2 mingw64/python2 is native and dependent on msys (cygwin.dll). 2015-03-02T17:54:53 It looks for an env var called MSYSTEM and changes what it does. I am not sure I can work around this removing that from env it runs in. 2015-03-02T17:59:43 with a .bat file? 2015-03-02T19:09:17 *** kiran4399_ has joined #rtems 2015-03-02T19:09:45 beng-nl... There ?? 2015-03-02T19:25:20 *** jhollister has quit IRC (Remote host closed the connection) 2015-03-02T19:45:58 *** kiran4399_ has quit IRC (Ping timeout: 246 seconds) 2015-03-02T20:00:13 aw. 2015-03-02T20:47:31 ita, no it is using the RTEMS Source Builder which is a Python program 2015-03-02T21:55:01 *** kiran4399 has joined #rtems 2015-03-02T21:55:21 is beng-nl there ? 2015-03-02T21:58:30 I want to talk about a project for gsoc'15 2015-03-02T22:04:55 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-02T23:51:40 *** kiran4399 has joined #rtems 2015-03-02T23:51:47 *** monstr has joined #rtems 2015-03-02T23:52:04 Hello beng-nl !! 2015-03-03T00:03:13 *** SH__ has quit IRC (Ping timeout: 246 seconds) 2015-03-03T00:03:25 *** kiran4399 has quit IRC (Quit: Page closed) 2015-03-03T00:20:15 *** dr__house is now known as dr__house|away 2015-03-03T00:50:25 [trac] #2270/enhancement (accepted) updated by daniel (Haven't calculated what the worst case is, perhaps it is not that much. You can ...) http://devel.rtems.org/ticket/2270 2015-03-03T01:21:24 *** kiwichris has quit IRC () 2015-03-03T01:22:15 *** kiran4399 has joined #rtems 2015-03-03T01:22:43 *** kiran4399 has quit IRC (Client Quit) 2015-03-03T02:16:27 *** Hesham has joined #rtems 2015-03-03T02:30:51 *** kiwichris has joined #rtems 2015-03-03T02:32:08 *** chan1 has joined #rtems 2015-03-03T02:32:46 hello, I see my pdf document for 'file system design' doesn't have figures in many places. 2015-03-03T02:33:07 Does anybody have the 'file system design' document with figures? 2015-03-03T02:34:15 chan1: https://wiki.samba.org/index.php/Writing_a_Samba_VFS_Module ? 2015-03-03T02:34:36 wrong channel, ignore me 2015-03-03T03:37:49 *** Hesham has quit IRC (Ping timeout: 264 seconds) 2015-03-03T06:20:19 http://freehackers.org/~tnagy/0001-Run-configure-automatically-when-config.cfg-is-modif.patch 2015-03-03T07:41:53 *** antgreen has joined #rtems 2015-03-03T07:48:43 *** kiran4399 has joined #rtems 2015-03-03T07:49:10 beng-nl... there? 2015-03-03T07:55:40 *** gedare has joined #rtems 2015-03-03T07:55:40 *** ChanServ sets mode: +o gedare 2015-03-03T07:57:18 Is anyboody mentoring for Beagleboard BSP projects ? 2015-03-03T07:59:56 kiran4399: i am here 2015-03-03T08:00:03 kiran4399: and i am available as beagle bsp mentor 2015-03-03T08:00:38 So I worked on Beaglebone black for around 2 years.. 2015-03-03T08:00:49 Can you get me started ?? 2015-03-03T08:01:54 probably, with what actually :-) 2015-03-03T08:01:57 rtems & beagle? 2015-03-03T08:02:32 with beagle.. 2015-03-03T08:03:07 okay, i wrote a detailed blog post about that as it happens 2015-03-03T08:03:11 I am new to rtems.. but I can catch up soon becuase I have a good understanding of real-time os... :-) 2015-03-03T08:03:51 Can you please paste the link .. 2015-03-03T08:03:57 how to build the toolchain & other dependencies 2015-03-03T08:04:01 the beagle code is mainlined fortunately 2015-03-03T08:04:07 not all of the other stuff 2015-03-03T08:04:14 kiran4399: http://www.shrike-systems.com/beagleboard-xm-beaglebone-black-and-everything-else-rtems-on-the-beagles.html 2015-03-03T08:06:04 some rtems + beagle bsp items that would be good for gsoc from my POV are listed @ https://devel.rtems.org/wiki/Developer/Projects/Open/ImproveBeagleBSP 2015-03-03T08:08:53 what exactly should we do in this project ? 2015-03-03T08:09:11 good question! - that is really mostly up to you 2015-03-03T08:09:32 [trac] Testing/Tester edited by gedare (Fix links.) http://devel.rtems.org/wiki/Testing/Tester 2015-03-03T08:09:37 a heavy factor for GSOC is what would make for a successful project, i.e. what motivates you and what your expertise is 2015-03-03T08:09:57 so it is largel up to you to come with a proposal 2015-03-03T08:10:05 the ideas on the wiki page are suggestions from the rtems project 2015-03-03T08:10:11 well, from me ;-) 2015-03-03T08:10:36 but if you have another suggestion that you think would be good for rtmes & you are motivated to do, it could well be a better idea 2015-03-03T08:11:38 [trac] Developer/Projects/Open/TesterImprovements edited by gedare (Point to the wiki page instead of external.) http://devel.rtems.org/wiki/Developer/Projects/Open/TesterImprovements 2015-03-03T08:14:58 I really want to do some thing on the beaglebone.. So can you please tell me more specifically what to do ?? 2015-03-03T08:15:38 did you take a look at the wiki page? 2015-03-03T08:15:56 yes.. 2015-03-03T08:16:06 do you want me to choose :) ? 2015-03-03T08:16:11 (surely not :)) 2015-03-03T08:16:41 is there anything there that interests you? 2015-03-03T08:17:06 Does RTEMS has support for all the peripherals ? 2015-03-03T08:17:12 in Beaglebone ? 2015-03-03T08:18:04 no, there are quite a few missing ones. 2015-03-03T08:18:31 with the wiki page i meant https://devel.rtems.org/wiki/Developer/Projects/Open/ImproveBeagleBSP which lists a few 2015-03-03T08:19:13 there are so many peripherals that i may not even know about some interesting ones 2015-03-03T08:19:20 or may have forgotten them 2015-03-03T08:22:51 I would like to go for the missing pheripheral support for beaglebone black !!! 2015-03-03T08:24:00 kiran4399: it would be best for you to start with our GSOC Quick Start, which you can find through a link at https://devel.rtems.org/wiki/GSoC 2015-03-03T08:24:26 kiran4399: the quick start will orient you to building RTEMS, and then you can also follow beng-nl blog post about building it for the beagle 2015-03-03T08:26:05 kiran4399: If you want to go after peripheral support, you should consider how to "organize" the effort. there are too many to try for in a single summer. 2015-03-03T08:28:18 how many do you want me to work on... I mean do have any high-prioritized peripheral support in your mind ? 2015-03-03T08:31:27 I don't. Maybe beng-nl does. I would guess you could go for 2 or maybe 3 of the items on the list. It depends on how much code exists you can re-use. 2015-03-03T08:31:55 I believe someone has networking operational with lwIP on a beagle 2015-03-03T08:32:05 And there was i2c support in the initial port 2015-03-03T08:32:22 so those might be "low hanging fruit" that you can easily pick up, maybe. 2015-03-03T08:34:15 You mean somebody has done some pheripheral support for beagle in rtems before ? 2015-03-03T08:35:04 There may be some code that has not been merged. And you may also be able to re-use code from another source, if the license is acceptable. (Non-copyleft) 2015-03-03T08:38:31 is any pheripheral support for beagle merged in the main repo ? 2015-03-03T08:41:55 you can lookhttps://git.rtems.org/rtems/tree/c/src/lib/libbsp/arm/beagle 2015-03-03T08:42:23 OK.. 2015-03-03T08:43:49 *** kiran4399 has left #rtems 2015-03-03T09:18:16 *** DrJoel has joined #rtems 2015-03-03T09:18:30 *** DrJoel has joined #rtems 2015-03-03T09:18:30 *** ChanServ sets mode: +o DrJoel 2015-03-03T09:19:06 [git] e4154b8 by Joel Sherrill: [mips BSPs: Remove .global for unprovided symbol closes 2253] http://devel.rtems.org/changeset/e4154b8/rtems 2015-03-03T09:20:00 verm__ grrrr "closes 2253" in a comment didn't close it 2015-03-03T09:20:05 [trac] #2253/defect (closed) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2253 2015-03-03T09:29:39 [trac] #2279/defect (v:4.11) created by joel.sherrill (mips-rtems dl01 failure on MIPS Endian Issue) http://devel.rtems.org/ticket/2279 2015-03-03T09:44:45 *** Hesham has joined #rtems 2015-03-03T09:45:30 good morning DrJoel 2015-03-03T10:12:51 [trac] 0001-psxtests-psxclock01-use-clock-and-CLOCKS_PER_SEC.patch (Test case.) attached to #2182 by gedare http://devel.rtems.org/ticket/2182 2015-03-03T10:35:58 [trac] #1394/defect (assigned) updated by gedare (I don't have a test case for this.) http://devel.rtems.org/ticket/1394 2015-03-03T11:03:47 DrJoel: there is a bug somewhere in the script i wrote 2015-03-03T11:03:59 i'll fix it but it's not a high priority since it mostly works 2015-03-03T11:06:16 OK.. Just watched the ticket. It just didn't close like I thought. 2015-03-03T11:06:31 Unfortunately, I only fixed 1/2 the MIPS BSPs. There is still a bug for Chris 2015-03-03T11:06:35 That was the new one. 2015-03-03T11:06:45 i have a better script i want to implement that will be more stable 2015-03-03T11:06:54 Hi Gedare. Did you notice only 137 organizations this year in GSoC? 2015-03-03T11:07:10 the problem is our git repo + trac live on different machines when pushing so there is no easy fix 2015-03-03T11:07:22 DrJoel: yeah that is such a huge drop 2015-03-03T11:07:45 i wonder if there were fewer applications or stricter requirements 2015-03-03T11:10:14 Carol said just down because last year was special as the tenth. It is a lot to manage 2015-03-03T11:10:43 'just down'? as in fewer applications or fewer accepted projects? 2015-03-03T11:11:35 fewer accepted. No stats given 2015-03-03T11:11:56 Did you see the email from Cyrille? 2015-03-03T11:11:58 they've said in previous years how many projects applied 2015-03-03T11:12:05 i did what project are they from? 2015-03-03T11:14:11 Multiple. He discussed a testing framework with me that sounded like it supported what was called API robustness testing years ago. 2015-03-03T11:14:41 Figured we can evalute the value and ideas. No harm in entertaining something. We may learn and could get something nice. 2015-03-03T11:14:49 He is a 5 year GSoC veteran per a blog 2015-03-03T11:17:21 416 org applications for GSoC this year 2015-03-03T11:17:32 http://google-opensource.blogspot.co.uk/2015/03/mentoring-organizations-for-google.html 2015-03-03T11:18:48 DrJoel: ah 2015-03-03T11:21:24 no DrJoel i did not notice. how many last year? 2015-03-03T11:22:12 i can't imagine fewer projects applied, but who knows. 2015-03-03T11:22:25 175 or so in last year 2015-03-03T11:22:32 seems they just accepted fewer this year 2015-03-03T11:22:36 ah, ok 2015-03-03T11:22:40 probably money considerations 2015-03-03T11:22:51 yes. there is some buzz about google's $ recently. 2015-03-03T11:22:57 i guess ad revenue is down. 2015-03-03T11:23:01 it's been that way for the last few years 2015-03-03T11:23:08 they've been clamping down quite a bit 2015-03-03T11:23:17 yeah. tinstaafl. 2015-03-03T11:23:33 i don't know that it's revenue being down more like more important projects going on 2015-03-03T11:24:25 i'm happy to bring on fresh mentors though. 2015-03-03T11:24:35 yeah 2015-03-03T11:24:44 more testing projects is not a bad thing 2015-03-03T11:24:47 heh. 2015-03-03T11:25:09 anyone working on it could use the waf build 2015-03-03T11:25:55 i don't think there is any reason a student couldn't use it now if they wanted then just merge their changes to 4.11 2015-03-03T11:26:30 yeah we'll see what happens. if the build errors can be fixed, i'd be happy to push for 4.11 release 2015-03-03T11:26:42 DrJoel is making some progress there. 2015-03-03T11:26:51 same, i talked with chris he said he won't have time to look at his tickets 2015-03-03T11:27:03 but i don't think he'd object if we move them wholesale i'll ask tonight 2015-03-03T11:27:22 would be really nice to just condense it down to tickets that must be done (build/tool errors) 2015-03-03T11:27:47 hey DrJoel are the build problems in the primary rtems build, or in the tests? 2015-03-03T11:27:55 I think we have to decide if libdl and associated tools need to be further restricted by targets. It was put in late and he doesn't have time to deal with cross target issues 2015-03-03T11:28:03 Yes. 2015-03-03T11:28:09 heh. 2015-03-03T11:28:12 moxie doesn't have a defined toolset 2015-03-03T11:28:19 that's the worst case 2015-03-03T11:28:23 well it also only has 1 user... 2015-03-03T11:28:40 Most others are libdl in tests by architecture 2015-03-03T11:29:05 I need to do a full build sweep. I was focusing on testing the mips fix I had last night 2015-03-03T11:29:35 omg success i finally resolved a 2-week debugging session. while eating pizza and chatting here. ;) 2015-03-03T11:29:42 ok 2015-03-03T11:30:07 nice! 2015-03-03T11:30:56 DrJoel: are there any tickets that you want to keep in 4.11 that you've opened? 2015-03-03T11:31:22 you opened a lot for chris, he doesn't have any time right now to look at them but if you're ok with any being fixed in 4.11.1 i don't think he'll care if they're moved 2015-03-03T11:33:23 I still have these stupid _POSIX_PATH_MAX / SSIZE_MAX declaration error when trying to compile "un-hacked" version of RTEMS using epiphany-rtems4.11-*. DrJoel any ideas why? 2015-03-03T11:34:05 problem in tools / newlib? 2015-03-03T11:34:16 GCC 2015-03-03T11:34:34 is it missing decls, or multiple decls that aren't the same? 2015-03-03T11:35:26 Messing declaration, DrJoel told me before it may be something missing part the process of adding RTEMS to the the Epiphany toolchain 2015-03-03T11:35:34 Missing* 2015-03-03T11:35:54 DrJoel: I think we could bump the tickets that are dl01 build failures. or add dl01 to the "ignore test" list for the target. 2015-03-03T11:38:08 [trac] #1526/defect (closed) updated by gedare (We can open a new ticket if these warnings are still present, but with the warning ...) http://devel.rtems.org/ticket/1526 2015-03-03T11:38:27 [trac] #1526/defect (closed) updated by gedare (Fix accidental bump to milestone...) http://devel.rtems.org/ticket/1526 2015-03-03T11:38:27 I wonder if we should add a "dltest" master list like the simulator clock tick. That would mark the ones which have PRs associated with them better than the ones which should work by architecture but don't have enough memory to link the tests 2015-03-03T11:39:19 hmm 2015-03-03T11:39:50 These are mostly architectural issues that Chris hasn't gotten to yet. 2015-03-03T11:40:01 I have filed PRs on them all 2015-03-03T11:40:20 ok. i don't see one for sparc-sis though 2015-03-03T11:40:37 It built for sparc last I tried. 2015-03-03T11:40:40 hmm 2015-03-03T11:40:51 not for me but maybe i'm missing something. 2015-03-03T11:41:27 DrJoel: if you commit the patch i sent to newlib, i can commit the rtems-side too and close #2182. 2015-03-03T11:41:38 Was it approved by Corrina? 2015-03-03T11:41:39 I should probably add the patch to RSB in the interim... 2015-03-03T11:41:45 ahem, good question! 2015-03-03T11:41:56 she may be waiting for another RTEMS person to chime in 2015-03-03T11:42:03 it's solely ifdef__rtems__ controlled. 2015-03-03T11:42:34 CLOCKS_PER_SEC is the number of clock ticks per second AFAIK 2015-03-03T11:42:51 read the PR. 2015-03-03T11:42:52 It can't be a hard-coded number because it has to be derived from the application configuration 2015-03-03T11:45:15 The POSIX page is contradictory. THe description says "best approximation" and the note in return value says it is hard-coded. That means clock() must return something in microseconds? 2015-03-03T11:45:25 Does our clock do that? Is my interpreation correct? 2015-03-03T11:45:50 yeah. i change __times() to return microseconds instead of ticks now 2015-03-03T11:46:35 [trac] #2275/defect (closed) updated by amar (Closing ticket. Please re-open if there are any issues.) http://devel.rtems.org/ticket/2275 2015-03-03T11:47:40 [trac] #2272/defect (closed) updated by amar (Please re-open if this is still an issue.) http://devel.rtems.org/ticket/2272 2015-03-03T11:48:31 [trac] #2233/defect (reopened) updated by amar (This is a bug in the script I will be changing it to use a different method. ...) http://devel.rtems.org/ticket/2233 2015-03-03T11:50:02 And that resolves clock()? 2015-03-03T11:51:43 yeah near as I can figure. 2015-03-03T11:52:08 OK. Asked to commit 2015-03-03T11:53:46 DrJoel: https://devel.rtems.org/ticket/2256 2015-03-03T11:53:50 Hesham asked a qeustion there 2015-03-03T11:54:56 [trac] #2256/defect (accepted) updated by joel.sherrill (At this point, yes. That seems to be the best solution. I don't think they should ...) http://devel.rtems.org/ticket/2256 2015-03-03T11:55:53 If it is easy for you, apply the 2256 change gedare 2015-03-03T11:55:53 Hesham: can you take a look at https://devel.rtems.org/ticket/2257 2015-03-03T11:56:09 i don't know where it goes 2015-03-03T11:56:11 [trac] #2250/task (new) updated by amar (Move to milestone:4.11.1) http://devel.rtems.org/ticket/2250 2015-03-03T11:58:13 [trac] #2256/defect (accepted) updated by joel.sherrill (Hesham.. attach a patch to the file in librpc and anyone can apply it.) http://devel.rtems.org/ticket/2256 2015-03-03T11:58:52 DrJoel: https://devel.rtems.org/ticket/2246 can you commit that patch? 2015-03-03T11:59:01 [trac] #2257/defect (new) updated by joel.sherrill (Post a patch to newlib and I will try to be sure to see it and commit it. The ...) http://devel.rtems.org/ticket/2257 2015-03-03T11:59:43 gedare: IIRC the last time we contacted Christian about this issue, he said he wasn't interested anymore in doing any or1k-related stuff for some reason. 2015-03-03T12:00:36 So I'll submit a straight forward patch for now. 2015-03-03T12:00:47 [trac] #2199/defect (new) updated by amar (empty comment) http://devel.rtems.org/ticket/2199 2015-03-03T12:01:07 [trac] #2236/defect (closed) updated by amar (Close in lieu of ticket #2199) http://devel.rtems.org/ticket/2236 2015-03-03T12:01:54 ok thanks Hesham. it happens, and we may just need to "limp along" the or1k until someone cares to sponsor/use it in production. 2015-03-03T12:02:54 [git] d5a7951 by Gedare Bloom: [Fix broken URLs to FTP site] http://devel.rtems.org/changeset/d5a7951/rtems-source-builder 2015-03-03T12:03:07 *** idleguy has joined #rtems 2015-03-03T12:03:25 gedare: FYI closes #.. works in non-rtems.git repos, too 2015-03-03T12:03:34 There's a new great effort done by Stefan (and still going on), I think because OpenRISC is part of his project http://www.optimsoc.org/ 2015-03-03T12:03:52 [trac] #2246/defect (closed) updated by gedare (Thanks, I fixed all similar links in RSB config files also.) http://devel.rtems.org/ticket/2246 2015-03-03T12:04:13 verm__: The github repo is not synchronized with git server? 2015-03-03T12:04:28 not at the moment i'm working on buildbot 2015-03-03T12:04:34 i can update it now 2015-03-03T12:04:49 i'll just crontab it for now 2015-03-03T12:04:54 ah, OK, never mind, I just wanna report that in case you didn't know 2015-03-03T12:05:09 hey DrJoel is there an official way for us to "deprecate" API features e.g. notepads (#2265) 2015-03-03T12:05:25 Gotta go bye. 2015-03-03T12:05:41 verm__: yeah i didn't even think to use it til after i pushed. :0 2015-03-03T12:05:45 bye Hesham 2015-03-03T12:07:45 set it to update every hour 2015-03-03T12:11:21 *** Hesham has quit IRC (Ping timeout: 250 seconds) 2015-03-03T12:22:28 [trac] #2264/defect (closed) updated by Gedare Bloom (In [changeset:"001f588d99ecdf502e727a66465df4d0ff756c8e/examples-v2"]: {{{ ...) http://devel.rtems.org/ticket/2264 2015-03-03T12:22:29 [git] 001f588 by Gedare Bloom: [hello: remove spurious code. Close #2264] http://devel.rtems.org/changeset/001f588/examples-v2 2015-03-03T12:37:59 [git] e08fe1b by Joel Sherrill: [dtc: Update to dtc 1.4.1 and use it on moxie The site referenced in the 1.2.x configuration is no longer accessible. The new configuration gives both an updated version and a more permanent location.] http://devel.rtems.org/changeset/e08fe1b/rtems-source-builder 2015-03-03T12:38:11 [trac] #2209/defect (closed) updated by gedare (empty comment) http://devel.rtems.org/ticket/2209 2015-03-03T12:42:07 *** Hesham has joined #rtems 2015-03-03T12:51:00 Hesham: did you build/test your patch? 2015-03-03T12:52:12 Nope, I need the configure options for generating this kind of mno-delay build. 2015-03-03T12:52:26 oh. 2015-03-03T12:52:33 well i pushed it anyways ;) 2015-03-03T12:52:49 [trac] #2256/defect (closed) updated by Hesham ALMatary (In [changeset:"396557412a6bab02116f1ce9c1020df68c8ce36d/rtems"]: {{{ ...) http://devel.rtems.org/ticket/2256 2015-03-03T12:52:50 [git] 3965574 by Hesham ALMatary: [librpc: Include or1knd part of the recongnized CPUs at xdr_float.c Close #2256] http://devel.rtems.org/changeset/3965574/rtems 2015-03-03T12:53:27 It won't be an issue, because this or1knd is already defined part of the toolchain patches, everywhere. 2015-03-03T12:53:47 Just tell me what options should I use there to build now 2015-03-03T12:53:58 I mean the configure line 2015-03-03T13:04:36 *** monstr has quit IRC (Remote host closed the connection) 2015-03-03T13:09:32 *** tester123 has joined #rtems 2015-03-03T13:11:05 *** dr__house|away is now known as dr__house 2015-03-03T13:18:46 *** sgulati___ has joined #rtems 2015-03-03T13:24:36 *** tester123 has quit IRC (Quit: Page closed) 2015-03-03T13:28:31 I had to build with --enable-multilib to trigger that. It is not a commonly used build configurations but cycles over all the multilib variants. :) 2015-03-03T13:28:48 gedare.. how is the 4.11 ticket count down going? 2015-03-03T13:36:50 https://devel.rtems.org/query?status=new&status=reopened&milestone=4.11&group=status&order=priority 2015-03-03T13:36:52 20 tickets. 2015-03-03T13:37:14 https://devel.rtems.org/query?owner=chrisj&status=accepted&status=assigned&status=new&status=reopened&milestone=4.11&col=id&col=summary&col=status&col=type&col=priority&col=milestone&col=component&order=priority 2015-03-03T13:37:23 is 10 assigned to chris. 2015-03-03T13:37:27 hmm 2015-03-03T13:37:39 i cleaned up what i could that was in rsb 2015-03-03T13:37:48 verm__ asked me to check what could be deferred. I see: 2015-03-03T13:37:52 most of those have to do with dl01/dl02 problems. 2015-03-03T13:37:56 1394 scandir() which you were working on 2015-03-03T13:38:07 yeah i have a patch attached there4 2015-03-03T13:38:08 2205 - tools don't build on mingw 2015-03-03T13:38:11 but i don't have any way to test it. 2015-03-03T13:39:07 I can delay all dl0* ones IF we have a way to avoid that stopping the build which makes it clear that the libdl COULD work on those architectures. Please enable this option and go for it. 2015-03-03T13:39:32 Or disable the libdl on those architectures. A simple decision would probably address 5 of those. 2015-03-03T13:39:40 maybe 6 2015-03-03T13:41:12 I see figuring that out as a big release hurdle. Clears a lot of tickets. :) 2015-03-03T13:43:06 brb 2015-03-03T13:43:06 hmm 2015-03-03T13:43:30 *** idleguy has left #rtems 2015-03-03T13:43:51 [trac] #2235/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2235 2015-03-03T13:44:06 That was for warnings building libdl on mips. Easy to push forward. 2015-03-03T13:44:12 nice 2015-03-03T13:44:27 Now 6 of 9 are a group solution. :) 2015-03-03T13:44:49 the build needs to be modified to work around them? 2015-03-03T13:45:15 I think so but chris needs to suggest something rational.. disable the architectures for libdl support for now? Is he ok for that? 2015-03-03T13:46:00 2205 I think Chris may be willing to close 2015-03-03T13:46:21 Moxie we can ping anthony again and push out IMO 2015-03-03T13:46:26 that's 2254 2015-03-03T13:51:19 *** krzysiekm13 has joined #rtems 2015-03-03T13:54:56 2005 is fixed? 2015-03-03T13:55:59 [trac] #2254/defect (new) updated by amar (Anthony? Can you update this ticket with any info you have please?) http://devel.rtems.org/ticket/2254 2015-03-03T13:56:26 *pokes antgreen* 2015-03-03T13:58:25 gedare: is 2182 required for 4.11? 2015-03-03T13:58:43 no but it is going to be fixed today or tomorrow. 2015-03-03T13:58:49 ah, ok 2015-03-03T13:58:56 the newlib patch has been approved. DrJoel can commit it now I think. 2015-03-03T13:59:05 and i'll commit the rtems ones too 2015-03-03T14:00:23 * antgreen wakes up 2015-03-03T14:00:50 antgreen: hey! any comments for this ticket? https://devel.rtems.org/ticket/2254 2015-03-03T14:02:01 hmm... I always use upstream HEAD. I think we still need a new binutils release. 2015-03-03T14:02:14 DrJoel: Building with --enable-multilib (after this fix) actually raised other errors, one of them relates to the old linkersymbols.h include 2015-03-03T14:02:21 is the build machinery running fedora/centos? because I have yum repos for working tools. 2015-03-03T14:02:41 The other is "mprotect.c:31:22: fatal error: sys/mman.h: No such file or directory" 2015-03-03T14:03:05 http://moxielogic.org/blog/pages/toolchain.html 2015-03-03T14:03:27 it's mostly tested on freebsd+centos at the moment i think.. 2015-03-03T14:03:55 can you just pull my rpms then? 2015-03-03T14:04:02 those work for sure 2015-03-03T14:04:08 that won't work for osx, windows or freebsd 2015-03-03T14:04:37 bah. 2015-03-03T14:04:46 ok, let me ask binutils guys 2015-03-03T14:04:53 we may be able to work around it by generating a patch for binutils with the relevant changes or see if they'll make a release 2015-03-03T14:05:01 cool that would be great 2015-03-03T14:06:43 i think we can bump the moxie binutils version in RSB too 2015-03-03T14:06:55 to point to a git ref 2015-03-03T14:07:11 if we have a "good one" 2015-03-03T14:13:39 *** krzysiekm13 has quit IRC (Remote host closed the connection) 2015-03-03T14:15:38 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2015-03-03T14:26:20 Yes. We can point the Moxie tools to a specific binutils git hash. Just need the hash. 2015-03-03T14:26:27 Well.. Chris can do it quickly. :) 2015-03-03T14:28:01 gedare can you push hesham's latest patch to the bsp? I am in the middle of a build sweep and don't want to screw anything up 2015-03-03T14:28:56 sure. 2015-03-03T14:29:25 Rocket testing today.. interesting noises :) 2015-03-03T14:29:51 i think you can commit my patch to newlib now. 2015-03-03T14:29:53 heh 2015-03-03T14:30:32 man.. i'd be so pissed if they were doing that while I was doing my taxes... ;) 2015-03-03T14:31:31 does that patch fix a ticket? or was it just an outstanding problem 2015-03-03T14:31:47 [git] 262a0d7 by Hesham ALMatary: [or1k/or1k-exception-default.c: Eliminate dependency on BSP provided .h file] http://devel.rtems.org/changeset/262a0d7/rtems 2015-03-03T14:32:09 oh 2015-03-03T14:47:21 DrJoel: yea/nay on the naming of drvmgr? 2015-03-03T14:48:00 was there a thread? :( 2015-03-03T14:48:16 I am working on you a scandir test case. 2015-03-03T14:48:23 great! 2015-03-03T14:48:29 no not a thread 2015-03-03T14:48:37 i'm reviwing the code now and just thinking about it. 2015-03-03T14:48:41 What suggestions do you have? 2015-03-03T14:48:49 well, we usually avoid abbreviations 2015-03-03T14:49:11 crap .. xdr_float.c doesn't compile for sparc now 2015-03-03T14:49:21 ruh roh 2015-03-03T14:49:57 no || at the end of the new line :( 2015-03-03T14:54:00 [git] 396ce49 by Joel Sherrill: [xdr_float.c: Fix typo] http://devel.rtems.org/changeset/396ce49/rtems 2015-03-03T14:54:17 ahh 2015-03-03T15:13:08 *** Davidbrcz has joined #rtems 2015-03-03T15:17:09 [trac] 0001-Add-simple-test-for-scandir-on-all-file-systems-test.patch (fstest suite scandir() family of tests) attached to #1394 by joel.sherrill http://devel.rtems.org/ticket/1394 2015-03-03T15:17:42 Gedare.. that adds a test case for imfs, mounted imfs, dosfs, rfs, and jffs2 for your scandir() case. 2015-03-03T15:18:02 Unfortunately, it seems to work for me. Double check the code in fsscandir01. 2015-03-03T15:18:28 *** sgulati___ has quit IRC (Ping timeout: 246 seconds) 2015-03-03T15:19:03 DrJoel: it may be due to the HAVE_SCANDIR macro 2015-03-03T15:19:21 i don't know. 2015-03-03T15:20:01 OK.. well see how it behaves for you. I don't have a toolset with your patch and it works. 2015-03-03T15:20:08 ok 2015-03-03T15:21:09 If the tests looks reasonable and are not duplicative, I think they should be added. 2015-03-03T15:21:41 It may be an old and crusty PR. :) 2015-03-03T15:22:08 maybe, but as sebastian said, the mismatch in the #define exists 2015-03-03T15:22:20 and it still does today 2015-03-03T15:22:29 OK.. well there is test code now 2015-03-03T15:22:30 i think there is an edge case somehow that gets hit 2015-03-03T15:22:40 maybe it can be made to hit it 2015-03-03T15:25:20 *** kiwichris has quit IRC (Ping timeout: 256 seconds) 2015-03-03T15:25:22 ahh 2015-03-03T15:25:45 the DIRSIZ macro was modified in 2013 2015-03-03T15:25:57 i think it got rid of the problem 2015-03-03T15:26:45 Do we have test code in fstests comparable to what I added? 2015-03-03T15:27:02 i doubt it. sebastian referenced only psxreaddir 2015-03-03T15:28:02 ok so it looks like the mismatch isn't really a problem anymore. but we should consider fixing it anyways 2015-03-03T15:28:11 I would love to see fstest code all in fstests but history had a lot of the system call interface tested with the IMFS. So there is duplication. psxtests tend to do 100% of the IMFS but other filesystems rely on the fstests. 2015-03-03T15:29:49 gedare@gedare-pc:~/work/rtems/rtems/testsuites$ ack-grep scandir -l 2015-03-03T15:29:49 psxtests/psxreaddir/test.c 2015-03-03T15:29:49 psxtests/psxreaddir/psxreaddir.scn 2015-03-03T15:29:49 fstests/configure.ac 2015-03-03T15:29:49 fstests/jffs2_fsscandir01/Makefile.am 2015-03-03T15:29:50 fstests/jffs2_fsscandir01/Makefile.in 2015-03-03T15:29:52 fstests/fsscandir01/init.c 2015-03-03T15:29:54 fstests/Makefile.am 2015-03-03T15:29:58 fstests/Makefile.in 2015-03-03T15:30:00 fstests/mimfs_fsscandir01/Makefile.am 2015-03-03T15:30:02 fstests/mimfs_fsscandir01/Makefile.in 2015-03-03T15:30:04 fstests/mrfs_fsscandir01/Makefile.am 2015-03-03T15:30:06 fstests/mrfs_fsscandir01/Makefile.in 2015-03-03T15:30:08 fstests/configure 2015-03-03T15:30:10 fstests/mdosfs_fsscandir01/Makefile.am 2015-03-03T15:30:12 fstests/mdosfs_fsscandir01/Makefile.in 2015-03-03T15:30:14 fstests/imfs_fsscandir01/Makefile.am 2015-03-03T15:30:16 fstests/imfs_fsscandir01/Makefile.in 2015-03-03T15:30:18 so these are probably adding some value. 2015-03-03T15:31:24 Yep. At least basic cases and a starting point to work from. 2015-03-03T15:31:36 As you feel like the issue is resolved, just push them as part of your patch set 2015-03-03T15:34:19 cannot open dl.tar for reading 2015-03-03T15:35:46 Which test has that issue? 2015-03-03T15:43:09 make[6]: Entering directory `/home/gedare/work/rtems/builds/b-sis-2/sparc-rtems4.11/c/sis/testsuites/libtests/dl01' 2015-03-03T15:43:57 *** BEGIN OF TEST FSSCANDIR ROOT IMFS *** 2015-03-03T15:43:57 Initializing filesystem ROOT IMFS 2015-03-03T15:43:57 scandir: Success 2015-03-03T15:43:57 Shutting down filesystem ROOT IMFS 2015-03-03T15:43:57 *** END OF TEST FSSCANDIR ROOT IMFS *** 2015-03-03T15:44:06 do you have tar? pax? dl01 needs that 2015-03-03T15:44:55 hmm.. is that a runtime or build time error? 2015-03-03T15:45:17 *** BEGIN OF TEST libdl (RTL) 1 *** 2015-03-03T15:45:17 load: /dl-o1.o 2015-03-03T15:45:17 handle: 0x2046f40 loaded 2015-03-03T15:45:17 Loaded module: argc:2 [../../../../../../../rtems/c/src/../../testsuites/libtests/dl01/dl-o1.c] 2015-03-03T15:45:17 0: Line 1 2015-03-03T15:45:17 1: Line 2 2015-03-03T15:45:18 Loaded module: argc:3 [../../../../../../../rtems/c/src/../../testsuites/libtests/dl01/dl-o1.c] 2015-03-03T15:45:18 0: Call 2, line 1 2015-03-03T15:45:19 1: Call 2, line 2 2015-03-03T15:45:19 [trac] #1394/defect (assigned) updated by gedare (Further examination in newlib shows that this bug should no longer manifest itself. ...) http://devel.rtems.org/ticket/1394 2015-03-03T15:45:19 2: Call 2, line 3 2015-03-03T15:45:20 handle: 0x2046f40 closed 2015-03-03T15:45:20 *** END OF TEST libdl (RTL) 1 *** 2015-03-03T15:45:36 no pax. 2015-03-03T15:46:46 build error? 2015-03-03T15:47:17 I have pax on my host.. not sure if it is needed though 2015-03-03T15:47:57 looks to be used in build process for three tests 2015-03-03T15:59:47 ok. well i gotta go done for the day 2015-03-03T16:03:06 bye 2015-03-03T16:04:23 hopefully chris comes around soon he's usually online by now 2015-03-03T16:04:42 *** gedare has quit IRC (Ping timeout: 252 seconds) 2015-03-03T16:29:01 [trac] GSoC/Mentors edited by cartho (empty comment) http://devel.rtems.org/wiki/GSoC/Mentors 2015-03-03T16:36:27 *** kiwichris has joined #rtems 2015-03-03T16:38:36 Hi 2015-03-03T16:38:51 hey 2015-03-03T16:38:59 Built ARM tools on Windows natively. 2015-03-03T16:39:02 read back on the open 4.11 tickets that are you 2015-03-03T16:39:28 I saw that. Have a few local issues I am working and the tools in the back ground 2015-03-03T16:39:41 We need gdb-7.9 for Windows but it does not build for SPARC 2015-03-03T16:39:48 We need Jiri's updated patches 2015-03-03T16:40:12 Windows needs 7.9 2015-03-03T16:40:24 Most come down to what to do about (1) architectures which can't link any BSP with dl0? and (2) targets which most BSPs can link dl0? but some can't due to a bug NOT a memory issue 2015-03-03T16:40:36 One is for windows tools. 2015-03-03T16:40:41 One gedare is running down. 2015-03-03T16:40:44 One is Moxie 2015-03-03T16:49:49 [trac] GSoC edited by joel.sherrill (empty comment) http://devel.rtems.org/wiki/GSoC 2015-03-03T16:50:10 [trac] GSoC edited by joel.sherrill (empty comment) http://devel.rtems.org/wiki/GSoC 2015-03-03T16:50:44 [trac] GSoC edited by joel.sherrill (empty comment) http://devel.rtems.org/wiki/GSoC 2015-03-03T16:52:11 *** Davidbrcz has quit IRC (Ping timeout: 256 seconds) 2015-03-03T16:52:21 [trac] GSoC edited by joel.sherrill (empty comment) http://devel.rtems.org/wiki/GSoC 2015-03-03T16:54:03 [trac] SOCIS/2015 created http://devel.rtems.org/wiki/SOCIS/2015 2015-03-03T16:54:19 [trac] SOCIS/2015 edited by joel.sherrill (empty comment) http://devel.rtems.org/wiki/SOCIS/2015 2015-03-03T16:55:12 [trac] GSoC/2015 edited by joel.sherrill (Correct tense and time) http://devel.rtems.org/wiki/GSoC/2015 2015-03-03T17:07:29 *** antgreen has quit IRC (Remote host closed the connection) 2015-03-03T17:30:18 DrJoel: can we consolodate all the tickets into 1, add a workaround and bump it? 2015-03-03T17:31:07 No. They are distinct bugs and will have to be fixed individually. If we can get a workaround to disable them that Chris approves, then we can just bump them. Consolidating them loses information 2015-03-03T17:32:09 ok that's fair 2015-03-03T17:32:51 he has to fix a v850 issue, a mips issue, etc. 2015-03-03T17:35:04 *** kiwichris has quit IRC (Read error: Connection reset by peer) 2015-03-03T17:35:10 *** kiwichri_ has joined #rtems 2015-03-03T17:37:10 Are you here a minute? 2015-03-03T17:38:27 https://devel.rtems.org/query?owner=chrisj&status=accepted&status=assigned&status=new&status=reopened&milestone=4.11&col=id&col=summary&col=status&col=type&col=priority&col=milestone&col=component&order=id 2015-03-03T18:31:39 good night 2015-03-03T18:31:42 *** DrJoel has quit IRC (Quit: Pull the pin and count to what?) 2015-03-03T18:49:30 *** gedare has joined #rtems 2015-03-03T18:49:30 *** ChanServ sets mode: +o gedare 2015-03-03T18:53:52 *** gedare has quit IRC (Ping timeout: 256 seconds) 2015-03-03T19:28:39 *** antgreen has joined #rtems 2015-03-03T20:34:36 verm__, about ? 2015-03-03T20:37:57 here 2015-03-03T20:39:47 It would be really nice if the RSB can be run by buildbot. I have a slave machine here that can be used and I can also provide a Windows slave. 2015-03-03T20:44:11 does the rsb support steps yet or is it still all-or-nothing? 2015-03-03T20:58:52 It can build a single arch. 2015-03-03T20:59:36 There is a command to build configs but the build sets are important as they define the build 2015-03-03T21:01:35 It would be nice to have the build sets being built as that is a step up from what we currently have. 2015-03-03T21:01:58 They can be built with --no-install to avoid installing. 2015-03-03T21:05:48 ok 2015-03-03T21:05:52 well, i can do that after phabricator 2015-03-03T21:05:59 and 4.11 2015-03-03T21:07:44 *** antgreen` has joined #rtems 2015-03-03T21:07:54 *** antgreen has quit IRC (Remote host closed the connection) 2015-03-03T21:16:27 Many thanks. Let me know what I need on my side for the slave. 2015-03-03T23:18:23 *** dr__house is now known as dr__house|away 2015-03-04T00:27:35 *** monstr has joined #rtems 2015-03-04T00:41:12 *** sebhub has joined #rtems 2015-03-04T00:41:50 good morning 2015-03-04T00:53:45 [trac] #2280/defect (v:4.11) created by sebastian.huber (Object extend is broken on SMP) http://devel.rtems.org/ticket/2280 2015-03-04T01:02:50 *** kiwichri_ has quit IRC () 2015-03-04T01:15:59 *** zoso has joined #rtems 2015-03-04T02:01:02 *** kiran4399 has joined #rtems 2015-03-04T02:01:35 beng-nl there ?? 2015-03-04T02:08:57 *** kiwichris has joined #rtems 2015-03-04T02:42:43 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-04T02:57:30 *** Hesham has joined #rtems 2015-03-04T03:02:19 *** Hesham has quit IRC (Client Quit) 2015-03-04T03:16:11 *** Hesham has joined #rtems 2015-03-04T04:10:14 *** idleguy has joined #rtems 2015-03-04T04:14:25 *** idleguy has quit IRC (Ping timeout: 246 seconds) 2015-03-04T04:29:25 *** idleguy has joined #rtems 2015-03-04T04:29:32 hi 2015-03-04T04:29:48 where can i find the rtems implementation of rename() function? 2015-03-04T04:30:44 i could find _rename_r . is this the one 2015-03-04T04:45:46 *** idleguy has quit IRC (Quit: Page closed) 2015-03-04T04:59:12 [trac] #2281/defect (v:4.11) created by sebastian.huber (Test objectivs are not checked in smptests/smpcapture01) http://devel.rtems.org/ticket/2281 2015-03-04T05:04:34 yes, the _rename_r 2015-03-04T05:04:46 for rename, see newlib sources 2015-03-04T05:05:31 [trac] #2273/enhancement (accepted) updated by Sebastian Huber (In [changeset:"e50297e36caf2c989a03825cc8589cd730c03389/rtems"]: {{{ ...) http://devel.rtems.org/ticket/2273 2015-03-04T05:05:32 [git] 0324410 by Sebastian Huber: [smptests: Do not use Giant directly This ensures that thread dispatching is disabled. A Giant acquire/release pair must take place in the same processor.] http://devel.rtems.org/changeset/0324410/rtems 2015-03-04T05:05:32 [git] c5436ff by Sebastian Huber: [smptests/smpcapture01: Avoid livelock condition] http://devel.rtems.org/changeset/c5436ff/rtems 2015-03-04T05:05:32 [git] 2b7fe35 by Sebastian Huber: [score: Add parameter to Giant acquire/release Update #2273.] http://devel.rtems.org/changeset/2b7fe35/rtems 2015-03-04T05:05:32 [git] b99be6e by Sebastian Huber: [score: Inline _Thread_Disable_dispatch() for SMP] http://devel.rtems.org/changeset/b99be6e/rtems 2015-03-04T05:05:32 [git] d2ffb7d by Sebastian Huber: [score: Add and use _Thread_Enable_dispatch_body() Update #2273.] http://devel.rtems.org/changeset/d2ffb7d/rtems 2015-03-04T05:05:33 [git] e50297e by Sebastian Huber: [score: ISR lock C/C++ compatiblity issue Empty structures are implementation-defined in C. GCC gives them a size of zero. In C++ empty structures have a non-zero size. Add ISR_LOCK_DEFINE() to define ISR locks for structures used by C and C++. Update #2273.] http://devel.rtems.org/changeset/e50297e/rtems 2015-03-04T05:11:39 [trac] #2273/enhancement (accepted) updated by sebastian.huber ([2b7fe35688fbeb3450ca79b7c5907811433e4fa9/rtems] ...) http://devel.rtems.org/ticket/2273 2015-03-04T06:10:42 *** kiran4399 has joined #rtems 2015-03-04T06:12:27 beng-nl: I would like to do pheripheral support for PRU, ADC and GPIO in Beagle support library as a project... 2015-03-04T06:12:45 for GSOC'13 2015-03-04T06:12:47 oops sorry for GSOC'15 2015-03-04T06:48:43 kiran4399: i see 2015-03-04T06:49:31 So how would you like me to get started ?? 2015-03-04T06:50:08 *** antgreen` has quit IRC (Remote host closed the connection) 2015-03-04T06:54:14 have you built & run rtems with the beagle bsp? 2015-03-04T06:54:25 or, let me ask differently, what have you done so far to get started? 2015-03-04T06:56:29 I saw the source code for rtems and also checkout out the bsp code.. 2015-03-04T06:56:46 okay 2015-03-04T06:56:53 which of the hardware do you have? 2015-03-04T06:57:20 Beaglebone black.. 2015-03-04T06:57:28 okay; with console? 2015-03-04T06:57:33 i mean, serial :) 2015-03-04T06:58:01 you mean to say the type of interface with the computer ? 2015-03-04T06:58:23 yeah, i mean do you have the cable that you connect to the serial pins on the BBB 2015-03-04T06:58:39 so you can see the serial output of (in this case) rtems on your computer 2015-03-04T06:59:31 *** Davidbrcz has joined #rtems 2015-03-04T06:59:53 You mean usb to serial.. No :-( 2015-03-04T07:00:26 yeah, specifically usb to ttl serial that's on the bbb 2015-03-04T07:00:53 okay, you will need that to do any work 2015-03-04T07:01:36 or were you planning something else for rtems console i/o? 2015-03-04T07:01:36 Or wait !! My friend has it.. I'll borrow that from him. 2015-03-04T07:01:56 (maybe i'm missing something of your intent) 2015-03-04T07:04:08 I don't think there is any good way... Do you have any thing in mind ? 2015-03-04T07:04:54 no, just the cable :) 2015-03-04T07:05:04 so, borrow that asap 2015-03-04T07:05:23 Can we make it with an old usb cable ? 2015-03-04T07:05:31 then the first step is to run rtems on the beagleboard xm emulator 2015-03-04T07:05:54 kiran4399: not sure what you mean by old usb cable..? 2015-03-04T07:06:41 I mean can we cut a usb cable and attach 4 female pins on the other side ? 2015-03-04T07:07:06 nooo 2015-03-04T07:07:21 it's *serial* to usb so there is some logic needed 2015-03-04T07:07:24 this thing: https://www.tigal.com/product/3466 2015-03-04T07:09:51 OK !! I got it.. I'll get it asap .. 2015-03-04T07:10:05 ok cool 2015-03-04T07:10:05 After that what would you want me to do? 2015-03-04T07:10:16 build & run rtems on the BBB 2015-03-04T07:10:30 before that, you could build & run rtems on the beagleboard xm emulator (linaro qemu) 2015-03-04T07:10:42 because that is all-software 2015-03-04T07:11:28 So instead of the hardware can we do it on the emulator ? 2015-03-04T07:11:32 not instead 2015-03-04T07:11:35 but first 2015-03-04T07:11:37 to get started 2015-03-04T07:11:46 so you are not blocked on the cable 2015-03-04T07:11:50 you need the cable asap 2015-03-04T07:12:37 OK 2015-03-04T07:17:48 so step one, build & run on the bbxm emulator 2015-03-04T07:17:57 in parallel, get the cable 2015-03-04T07:18:05 OK.. 2015-03-04T07:18:14 then.. 2015-03-04T07:18:17 step two, do something for rtems that you can see on the bbxm emulator 2015-03-04T07:18:39 debugging a failing test would be great 2015-03-04T07:18:55 the caveat there is that it might be the fault of linaro qemu :) 2015-03-04T07:19:14 what is that ? 2015-03-04T07:19:17 instructions to build & run all of this, including running the tests, is also on the blog post i linked to btw 2015-03-04T07:19:25 kiran4399: what? a failing test? 2015-03-04T07:21:02 yes. 2015-03-04T07:21:09 aha 2015-03-04T07:21:18 rtems has a test suite 2015-03-04T07:21:29 instructions to build & run it are in the post 2015-03-04T07:21:35 not all of them pass on the bbxm emulator 2015-03-04T07:21:53 if you can debug the cause of one of those failures, that would be great 2015-03-04T07:22:13 with the warning, that the failure might be due to imperfect emulation by qemu 2015-03-04T07:29:01 oh! 2015-03-04T07:30:29 which kernel is rtems using ? 2015-03-04T07:49:42 *** gedare has joined #rtems 2015-03-04T07:49:42 *** ChanServ sets mode: +o gedare 2015-03-04T08:06:28 *** Davidbrcz has quit IRC (Ping timeout: 252 seconds) 2015-03-04T08:07:05 kiran4399: 2015-03-04T08:14:25 kiran4399: which kernel? 2015-03-04T08:15:16 kiran4399: the rtems application is linked with its own system code and so runs autonomously without any further kernel, if i understand you correctly 2015-03-04T08:15:39 *** acenos has joined #rtems 2015-03-04T08:17:10 hello 2015-03-04T08:18:52 hello acenos 2015-03-04T08:19:04 hello, beng-nl 2015-03-04T08:19:11 hi acenos 2015-03-04T08:19:25 I am interested in the GSoC. Anybody online I can consult about the topic? 2015-03-04T08:19:31 Sure! 2015-03-04T08:19:33 hi gedara 2015-03-04T08:19:52 if you haven't found it yet, https://devel.rtems.org/wiki/GSoC 2015-03-04T08:19:56 that is the main portal 2015-03-04T08:20:02 for gsoc-related work 2015-03-04T08:20:08 eyup, I come from there 2015-03-04T08:20:10 there's a lot of info there of course... 2015-03-04T08:20:30 do you have some questions now? 2015-03-04T08:21:00 Yes, I am specially concern about the difficulty of the proposed ideas 2015-03-04T08:22:38 What is your background? 2015-03-04T08:22:47 my experience with RTEM is that I have used it in my "work". This would be my first time in the GSoC or working in a open source project. So I am a bit intimidated : ) 2015-03-04T08:22:52 I am a grad student 2015-03-04T08:23:12 OK well don't worry too much. We want you to aim for a challenging-but-doable project. 2015-03-04T08:23:40 what kinds of projects most interest you? 2015-03-04T08:23:47 I have work in a project of the ESA to warrant the timing composability in COTS multicores. There I used RTEMS 2015-03-04T08:24:46 (You may also be interested in applying through SOCIS.) 2015-03-04T08:26:01 Well, I was interested into improve my coding and knowledge about OS 2015-03-04T08:26:35 (Yeah, I notice the SOCIS program thanks to your website : ) ) 2015-03-04T08:27:14 Then I guess you may want to focus on the "RTEMS Run-Time Oriented" projects, if you want to dig in to the OS / kernel code. 2015-03-04T08:27:45 are you a grad student in CS? 2015-03-04T08:27:49 aye 2015-03-04T08:28:05 are you familiar with C programming? 2015-03-04T08:28:15 what is the background level expected? 2015-03-04T08:28:18 yes 2015-03-04T08:28:30 then I think you will be able to scope a project appropriate to yourself 2015-03-04T08:29:01 we are happy to include anyone with sufficient interest. it is best to have some C experience and CS knowledge, anything else is a bonus 2015-03-04T08:30:34 glad to hear that 2015-03-04T08:31:25 run-time tracing is a good project for SOCIS too 2015-03-04T08:31:36 on-target trace capabilities are important for space-borne devices 2015-03-04T08:32:02 hmm yeah 2015-03-04T08:32:24 although I think the mentors there may have some concrete ideas about what they want to see done. 2015-03-04T08:32:28 I have worked around that but using PMC 2015-03-04T08:32:49 PMC? 2015-03-04T08:32:59 Program Monitoring Counters 2015-03-04T08:33:23 ahhh 2015-03-04T08:34:31 i believe they want to integrate our tracing with lttng 2015-03-04T08:34:55 there may be someone working on this already 2015-03-04T08:36:18 hmm curious 2015-03-04T08:38:44 so it would be to code a trace generator? 2015-03-04T08:38:59 we already have a trace generator 2015-03-04T08:39:11 It would be to continue to work on it 2015-03-04T08:39:57 You should contact devel@rtems.org to ask questions about what remains to do on it. or contact Chris Johns directly. 2015-03-04T08:40:20 noted thanks 2015-03-04T08:41:16 I was reading what you have in the web about this project, to compare about what I know: until now I have work with traces a very low level : D 2015-03-04T08:42:20 yeah. well we want something that works across platforms, sometimes that is trickier 2015-03-04T08:42:21 *** antgreen has joined #rtems 2015-03-04T08:42:48 yeah, that was something that was going around my mind: what architecture are you aiming to 2015-03-04T08:42:52 ? 2015-03-04T08:43:25 as many as possible :) for SOCIS they care more about the space-ready ones of course, so sparc/leon is the primary target 2015-03-04T08:43:54 oh! that is good and bad 2015-03-04T08:43:58 but RTEMS also has lots of users of the other targets especially powerpc and arm 2015-03-04T08:44:22 nowadays you only need to care about leon3 and 4 2015-03-04T08:44:27 I have been working with the Leon3 and the NGMP, that is good. The bad is that I have already some bad blood with them XD 2015-03-04T08:44:34 haha 2015-03-04T08:44:51 we have very mature leon support 2015-03-04T08:44:58 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-04T08:45:11 what uni are you at? 2015-03-04T08:45:27 UPC (universitat politecnica de Catalunya) 2015-03-04T08:45:35 It is in Barcelona, Spain 2015-03-04T08:48:20 ok 2015-03-04T08:48:52 I suppose that it is not very famous XD 2015-03-04T08:49:06 fame is overrated 2015-03-04T08:49:39 *** zoso has quit IRC (Quit: Leaving) 2015-03-04T08:49:50 oh, talking about colleges, I wanted to ask about how much time is expected to invest in the project 2015-03-04T08:50:20 I have been told that this has been made with USA colleges in mind, so it is expected to have 1 to 2 free months free for it 2015-03-04T08:50:21 we can be flexible to your uni schedule to some extent. we know the EU schools are still in session while gsoc starts 2015-03-04T08:50:49 we expect you to be full-time on it when you are out of school though. 2015-03-04T08:51:03 and you must be making some progress until the midterm regardless of your school schedule 2015-03-04T08:51:09 hmm here the Phd works a bit different 2015-03-04T08:51:18 oh right you're a grad student 2015-03-04T08:51:46 it is like be employed to work 40 hours per week, 28 vacation days per year 2015-03-04T08:51:59 well, you are responsible for setting your schedule and milestones for gsoc, and to make sure you meet them. 2015-03-04T08:52:32 well, sleep is also overrated 2015-03-04T08:52:38 heh 2015-03-04T08:53:25 it is better if you can pick a project that overlaps with your research area 2015-03-04T08:54:09 if your uni is paying you, then you may want to consult with your supervisor/advisor about whether it is a good idea to do SOC. 2015-03-04T08:54:10 agree, that is one reason to pick rtems 2015-03-04T08:54:35 Advisor: do whatever you want as long as you dont miss any deadline 2015-03-04T08:54:39 I did GSoC as a grad student with RTEMS on a project that overlaps with mine 2015-03-04T08:54:43 yeah, that's good. 2015-03-04T08:54:44 we are really into the Real Time thing 2015-03-04T08:54:47 XD 2015-03-04T08:55:11 Then I suggest you should not worry too much, but set your own schedule. You can probably expect to succeed even at 20 hrs/wk averaged over the whole summer 2015-03-04T08:55:29 we basically have the same answer as your advisor... "do whatever you want as long as you dont miss any deadline" 2015-03-04T08:55:47 your proposal should contain the detailed schedule and milestones 2015-03-04T08:55:53 so we'll judge you by that 2015-03-04T08:56:06 i have to go soon, any other questions? 2015-03-04T08:57:07 so I have to make a proposal about one of the proposed projects, right? 2015-03-04T08:59:07 May I propose the milestones or may I contact any mentor and negotiate them? 2015-03-04T08:59:47 [trac] Developer/Projects/Open/ImproveBeagleBSP edited by beng (full GSOC getting started instructions, better hierarchy, fix TOC) http://devel.rtems.org/wiki/Developer/Projects/Open/ImproveBeagleBSP 2015-03-04T08:59:48 negotiate. you can propose your own project too, but it is only going to succeed if you can get a mentor for it. 2015-03-04T09:00:20 whereas most of the open projects there should be a mentor available 2015-03-04T09:00:46 *** kiran4399 has joined #rtems 2015-03-04T09:01:06 yeah I see the one in the trace tools is Chris Jhons 2015-03-04T09:01:28 Any idea about the "CPU statics" one? 2015-03-04T09:01:40 beng-nl.. there ?? 2015-03-04T09:01:50 beng-nl: 2015-03-04T09:02:08 kiran4399: i am 2015-03-04T09:03:19 So right now after the basic tasks are done ... should I focus more on writing proposal or to contributing ? 2015-03-04T09:04:02 the basic tasks i was talking about were really just pre-requisites to get started; to show that you're to to speed 2015-03-04T09:04:10 kiran4399: i added them to the wiki page 2015-03-04T09:04:19 kiran4399: what basic tasks are you thinking of? 2015-03-04T09:04:52 Whatever you told :-D 2015-03-04T09:06:06 kiran4399: the way i see it, the proposal is a plan of what you're going to do along with some evidence that you're motivated & able to do it 2015-03-04T09:06:29 kiran4399: part of that is getting up to speed (see steps) & making a change (my suggestion is, debug a failing test) - that would make for a great proposal 2015-03-04T09:06:39 Oh.. 2015-03-04T09:06:46 no good? 2015-03-04T09:06:53 Yes sure !!! 2015-03-04T09:08:23 together with the project idea that motivates you of course :) 2015-03-04T09:08:31 and explain why it does, what your background is etc 2015-03-04T09:10:41 kiran4399: of course there are people on this channel (incl me) you can talk to at length to get comfortable / more details 2015-03-04T09:11:51 about BSP ? 2015-03-04T09:11:59 yeah, and rtems, and gsoc 2015-03-04T09:13:02 OK.. 2015-03-04T09:21:57 *** sebhub has quit IRC (Quit: Konversation terminated!) 2015-03-04T09:26:13 *** Davidbrcz has joined #rtems 2015-03-04T09:34:04 *** starprince has joined #rtems 2015-03-04T09:36:31 Where should RTEMS get the definition of SSIZE_MAX? On my Linux machine it's defined at "/usr/include/bits/posix1_lim.h" and for RTEMS it's defined at utf8proc.h. I got undeclared error of this SSIZE_MAX at libio_.h which just #include . Does this mean that limit.h is not included part of "the newly ported epiphany-rtems4.11-gcc" ? 2015-03-04T09:38:01 acenos, I wrote the cpu statistics page. it is a good project to scope out. the space community also is interested in it, to do in situ performance monitoring. 2015-03-04T09:39:12 Hesham, those defines usually come in through either newlib or gcc itself, but should be in gcc 2015-03-04T09:39:38 grep for SSIZE_MAX in newlib.git/newlib and see what you find 2015-03-04T09:40:11 So, the problem may be with newlib, that makes sense because I almost didn't touch their newlib code. 2015-03-04T09:40:23 I'll have a look, thanks gedare 2015-03-04T09:40:32 yup 2015-03-04T09:43:57 *** DrJoel has joined #rtems 2015-03-04T09:43:57 *** DrJoel has joined #rtems 2015-03-04T09:43:57 *** ChanServ sets mode: +o DrJoel 2015-03-04T09:44:25 * DrJoel waves hi 2015-03-04T09:46:46 gedare, so you would be the mentor ergo the one to negotiate the milestone with, right? 2015-03-04T09:46:50 hi DrJoel 2015-03-04T09:46:55 *** starprince has quit IRC (Ping timeout: 246 seconds) 2015-03-04T09:47:19 acenos, hard to say for sure, but i generally try to get involved in all the projects a little bit. 2015-03-04T09:47:53 the on-target monitoring/tracing/statistics projects tend to be for either DrJoel or kiwichris 2015-03-04T09:48:40 Yes. We may also be able to get a co-mentor from the LTTng group. If the proposal makes sense for them to be involved. 2015-03-04T09:48:46 I had dinner with them last week. 2015-03-04T09:59:10 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-04T10:05:42 *** starprince has joined #rtems 2015-03-04T10:07:07 hi DrJoel 2015-03-04T10:07:16 That's weird. in gcc/config/epiphany/epiphany.h There is no notion about newlib at all! 2015-03-04T10:07:54 Even in their rtems.h, it just #undef LINK_GCC_C_SEQUENCE_SPEC 2015-03-04T10:08:17 Unlike or1k, which include "#undef DRIVER_SELF_SPECS" 2015-03-04T10:12:46 *** kiran4399 has joined #rtems 2015-03-04T10:16:31 hey beng-nl 2015-03-04T10:16:44 Hesham UK can apply to ESA SOCIS as partner nation. 2015-03-04T10:17:06 We are recommending ESA eligible students to apply to both and let us sort out slots when the dust settle. 2015-03-04T10:17:19 config.gcc I think is what knows about newlib 2015-03-04T10:17:24 That's what I intend to do anyway 2015-03-04T10:17:56 for or1k, all newlib details are included in config/or1k/or1k.h 2015-03-04T10:18:23 The corresponding epiphany.h file doesn't have anything about newlib 2015-03-04T10:20:02 *** dr__house|away is now known as dr__house 2015-03-04T10:22:40 A maybe silly question for gedare and/or DrJoel about redacting the proposal for the CPU Statics or Trace project. I have not very clear what to propose so: Should I play a bit with what is already done and then propose what comes to my mind? 2015-03-04T10:23:59 acenos, yeah. and engage the community about what might be interesting. 2015-03-04T10:26:37 thanks, then I will come back to you when I get something : ) 2015-03-04T10:38:38 great. and also do go through the "getting started for gsoc": 2015-03-04T10:39:46 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-04T10:43:19 *** antgreen has quit IRC (Remote host closed the connection) 2015-03-04T10:54:01 DrJoel: on config.gcc 2015-03-04T10:54:01 epiphany-*-rtems* ) 2015-03-04T10:54:01 tm_file="${tm_file} epiphany/rtems.h rtems.h" 2015-03-04T10:54:09 *** kiran4399 has joined #rtems 2015-03-04T10:54:16 tm_file should contain t-rtems? 2015-03-04T11:03:26 let me look at source 2015-03-04T11:03:59 tmake_file 2015-03-04T11:04:11 tm_file="dbxelf.h elfos.h v850/v850.h" 2015-03-04T11:04:11 tm_file="${tm_file} rtems.h v850/rtems.h newlib-stdint.h" 2015-03-04T11:04:11 tmake_file="${tmake_file} v850/t-v850" 2015-03-04T11:04:11 tmake_file="${tmake_file} v850/t-rtems" 2015-03-04T11:04:15 from v850 2015-03-04T11:06:24 Yeah, the patch you provided me with doesn't contain t-rtems there 2015-03-04T11:06:56 What's is t-rtems doing? 2015-03-04T11:07:23 Just says we always have limits.h so not a huge impact. Wrong but not fatal 2015-03-04T11:09:36 That may relate to the error I get when compiling RTEMS -> undeclared SSIZE_MAX and _POSIX_NAME_MAX? 2015-03-04T11:25:54 *** jhollister has joined #rtems 2015-03-04T11:32:53 *** monstr has quit IRC (Remote host closed the connection) 2015-03-04T11:41:05 Maybe. not sure the impact 2015-03-04T12:00:40 DrJoel: were you able to talk with chris about the build issues yesterday? 2015-03-04T12:01:10 i think we should just go ahead and work around it if not and make a note of it for the release notes 2015-03-04T12:02:12 I asked on irc about them but he never replied. I was online until about 6:45CST which should have been about 3 hours overlap with him. 2015-03-04T12:02:27 0 2015-03-04T12:02:31 er, ignore 2015-03-04T12:02:39 I will look at a workaround. I built all BSPs overnight and the next step it to look at the failures 2015-03-04T12:03:12 he verified that he got the email if he had any objections he'd have spoken up earlier if we do a workaround and move the tickets i'll talk with him tonight about it 2015-03-04T12:03:41 I have a couple of ideas on how to workaround this. Deciding how painful each if it is a subset of architecture or total architecture failure 2015-03-04T12:05:10 that's a tough call if the fix will fix all the bsps i'd just mark the entire arch as failed.. if not, individually 2015-03-04T12:10:54 that`s how i am leaning but have to see current results 2015-03-04T12:11:09 i think 20+ powerpc bsps fail but most work. 2015-03-04T12:11:20 50% MIPS BSPs fail 2015-03-04T12:11:29 all bfin and v850 fail 2015-03-04T12:15:31 *** chan1 has quit IRC () 2015-03-04T12:20:37 ah, well it may be worth it to keep some of the powerpc bsps but i don't think so 2015-03-04T12:20:53 especially if the eventual fix will happen at the same time 2015-03-04T12:21:58 *** idleguy has joined #rtems 2015-03-04T12:34:37 *** Davidbrcz has quit IRC (Ping timeout: 264 seconds) 2015-03-04T12:38:44 starting the wading. 2015-03-04T12:38:55 39 BSPs with some build issue 2015-03-04T12:40:17 whoa 2015-03-04T12:40:29 should we have a commit freeze? 2015-03-04T12:42:02 Seriously committing libdl introduced most of these. 2015-03-04T12:42:19 A few are BSPs that were on the cusp of some tests not fitting into memory and now they don't 2015-03-04T12:43:30 i don't see a point in you fixing the builds while we layer more changes onto it 2015-03-04T12:43:36 we can branch 4.11.1 and have any future changes go there 2015-03-04T12:43:55 that or just branch 4.11 and work off of that 2015-03-04T12:44:07 we can merge any change we make to it to fix the build into master 2015-03-04T12:44:26 not arguing .. 2015-03-04T12:44:48 I amm just cleaning up now. Won't take long. This will let us put off at least 6 tickets 2015-03-04T12:45:18 ok, i'd just hate for you to go through all the effort and yet another chang lands that breaks something else 2015-03-04T12:45:51 Most of these seriously are libdl. The first one unfortunately wasn't .. typical. Small ARMs 2015-03-04T12:50:29 [trac] #2251/defect (closed) updated by joel.sherrill (There is another PR for the same issue on dl01 (2247). I will close this and update ...) http://devel.rtems.org/ticket/2251 2015-03-04T12:51:07 [trac] #2247/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2247 2015-03-04T12:51:18 That was an easy way to kill one. :) 2015-03-04T12:52:13 [trac] #2210/defect (closed) updated by joel.sherrill (This is described better in 2258. Closing this as duplicate.) http://devel.rtems.org/ticket/2210 2015-03-04T12:52:28 nice 2015-03-04T12:53:22 It will take build time to address 2252 and 2260 but they will get an update and then a defer in the next couple of hours. 2015-03-04T13:27:28 i'll try to commit my CLOCKS_PER_SEC fix tonight or tomorrow... 2015-03-04T13:28:21 that will close a PR. and you can add your fscandir test if you want DrJoel and then either close the PR or move it to 4.10.3 if we want to do a bug-fix patch for newlib in RSB for that toolchain. 2015-03-04T13:28:44 looks like i'll probably have a snow day tomorrow so i may not get to work much. 2015-03-04T13:28:54 I need to capture screens for those tests but I can commit them after that. 2015-03-04T13:29:03 ok 2015-03-04T13:29:06 We are supposed to be getting ice here also. Looks like it will hit overnight. 2015-03-04T13:29:20 do you want to apply the fix as an RSB patch for the 4.10 toolchain? 2015-03-04T13:29:32 as in, is that something we should do? 2015-03-04T13:29:34 I am disabling v850 and bfin for libdl with a commit to cpukit/configure.ac with a comment pointing to the ticket 2015-03-04T13:29:41 ok. 2015-03-04T13:29:42 Ask on devel. I am not sure. 2015-03-04T13:29:46 ok. 2015-03-04T13:30:46 moxie tools probably just need an RSB patch to point toward newer binutils. 2015-03-04T13:31:20 I think so. Anthony said yesterday he would provide a hash. Could try just pointing at today's head. :) 2015-03-04T13:31:28 Try that and see if it builds. :) 2015-03-04T13:32:37 *** Davidbrcz has joined #rtems 2015-03-04T13:33:43 hi Davidbrcz 2015-03-04T13:34:30 heh yeah 2015-03-04T13:35:02 he said that's all he used, the latest 2015-03-04T13:36:37 heading home. cya 2015-03-04T13:36:43 *** gedare has quit IRC (Remote host closed the connection) 2015-03-04T13:42:50 *** krzysiekm13 has joined #rtems 2015-03-04T13:45:42 *** Hesham has quit IRC (Ping timeout: 245 seconds) 2015-03-04T13:51:08 DrJoel: nice about canada 2015-03-04T13:51:23 DrJoel: btw canada is a contributing country just an 'associate member' (the only one?) 2015-03-04T13:55:19 ESA is worse than the Vatican or Roman Empire for beurocracy. Lots to wade through. :) 2015-03-04T13:55:34 So Canadians appear to be eligible. 2015-03-04T13:56:52 we should probably note that on the wiki and give a list of member states 2015-03-04T13:59:10 If you feel comfortable doing that, feel free. Too close to legal advice for my taste. :) 2015-03-04T14:00:30 [trac] #2258/defect (new) updated by chrisj (I suspect there are flags on gcc command line which effect the asm code being ...) http://devel.rtems.org/ticket/2258 2015-03-04T14:05:14 [trac] SOCIS created http://devel.rtems.org/wiki/SOCIS 2015-03-04T14:07:57 [trac] WikiStart edited by amar (Cleanup GSoC and add SOCIS entry.) http://devel.rtems.org/wiki/WikiStart 2015-03-04T14:08:11 there 2015-03-04T14:11:52 :0 2015-03-04T14:11:57 gedare.. around? 2015-03-04T14:15:48 *** Hesham has joined #rtems 2015-03-04T14:37:15 verm__ One of the keywords is "updates" right? 2015-03-04T14:37:28 And "closes" is the one we really want. :) 2015-03-04T14:39:14 [trac] #2182/defect (closed) updated by Gedare Bloom (In [changeset:"56c719260668c35e87acda32bb12bdf81f08d9e0/rtems"]: {{{ ...) http://devel.rtems.org/ticket/2182 2015-03-04T14:39:14 [git] 76f3765 by Gedare Bloom: [libcsupport: scale times() call to microseconds] http://devel.rtems.org/changeset/76f3765/rtems 2015-03-04T14:39:14 [git] 56c7192 by Gedare Bloom: [psxtests/psxclock01: use clock() and CLOCKS_PER_SEC Closes #2182] http://devel.rtems.org/changeset/56c7192/rtems 2015-03-04T14:45:34 DrJoel: yep 2015-03-04T14:45:43 Thanks 2015-03-04T14:45:46 there are several different ones 2015-03-04T14:45:53 update, updates, ref, refs etc 2015-03-04T14:46:19 the reason it's not triggering every time is due to the workaround i had to do for it to work since it's split across several machines 2015-03-04T14:46:39 [git] 57a914a by Joel Sherrill: [Add simple test for scandir() on all file systems tested updates 1394] http://devel.rtems.org/changeset/57a914a/rtems 2015-03-04T14:47:33 That didn't update the PR. :( 2015-03-04T14:47:47 [trac] #1394/defect (assigned) updated by joel.sherrill (Added new set of tests fsscandir01 in fstests for code shown in this PR.) http://devel.rtems.org/ticket/1394 2015-03-04T14:48:03 i think that one is a long standing bug 2015-03-04T14:48:11 you can't have the ticket be the last item 2015-03-04T14:48:17 updates 1394. would have worked 2015-03-04T14:48:46 It was the last line in the commit message. Where should it go? 2015-03-04T14:49:02 it doesn't matter where it goes you just need an extra character at the end 2015-03-04T14:49:14 or put it before the end either way 2015-03-04T14:49:49 XXX closes 2015-03-04T14:49:54 no 2015-03-04T14:49:57 closes XXX. 2015-03-04T14:50:07 'updates 1394.' vs 'updates 1394' 2015-03-04T14:50:15 doesn't matter what the character is 2015-03-04T14:50:23 ok. we will see if this next round works. :) 2015-03-04T14:51:44 Gedare is addressing the scandir fix. But he said there was no test code so I added the test case provided for every file system. He will have to close it when he is happy. 2015-03-04T14:53:48 *** gedare has joined #rtems 2015-03-04T14:53:48 *** ChanServ sets mode: +o gedare 2015-03-04T14:53:49 [git] 461a930 by Joel Sherrill: [capture01: Update screen file] http://devel.rtems.org/changeset/461a930/rtems 2015-03-04T15:11:13 verm__: that commit did not update a referenced ticket. did DrJoel miss something? 2015-03-04T15:11:27 which one? 461a930 ? 2015-03-04T15:16:40 yea 2015-03-04T15:16:58 does the ticket ref need a # before the number to get picked up? 2015-03-04T15:17:45 he said it needs a period after the ticket number 2015-03-04T15:17:50 updates xxx. 2015-03-04T15:18:59 [trac] #1394/defect (assigned) updated by gedare (Move to 4.10.3 in case we want to provide a fix there. This problem is resolved in ...) http://devel.rtems.org/ticket/1394 2015-03-04T15:19:17 Once my build is finished, I can close one ticket and defer two others. 2015-03-04T15:19:54 ok 2015-03-04T15:20:04 The powerpc dl one (2258) will take a bit more build time to work through so I will commit what I have before I start on it. 2015-03-04T15:20:06 DrJoel: what's going on here: https://devel.rtems.org/ticket/2245 2015-03-04T15:20:16 was that your anonymous comment? 2015-03-04T15:21:07 yes. I need to dig through my branches on my local machine VM and deal with that. I consider that a 4.11 issue 2015-03-04T15:21:21 0k 2015-03-04T15:21:29 Hesham: do you have eyes on #2257? 2015-03-04T15:21:36 https://devel.rtems.org/ticket/2257 2015-03-04T15:22:09 gedare: Nope, first time to see it. 2015-03-04T15:22:20 DrJoel: feel free to mark notepads and task variables as obsolete/deprecated 2015-03-04T15:22:31 DrJoel: do we have a "standard" way to do that? 2015-03-04T15:22:40 Hesham: ah, ok. think you can handle it? 2015-03-04T15:23:35 if i had more time i'd love to take a stab at implementing #2183, it seems like it would be some fun hacking.. 2015-03-04T15:23:42 Notepads can be. Probably should discuss task variables. Not SMP safe but not as bad as notepads 2015-03-04T15:24:09 File a ticket for each of those 2015-03-04T15:24:12 kiwichris wants to axe task variables. 2015-03-04T15:24:14 there is one 2015-03-04T15:24:23 https://devel.rtems.org/ticket/2265 2015-03-04T15:24:26 2183 is just a matter of a little assembly code for locking 2015-03-04T15:24:36 Same algorithm as sparc, arm, and powerpc. 2015-03-04T15:25:01 Then ask on devel for how we do that. There is a precedence with rtems_clock_get() and some other IRQ? APIs. 2015-03-04T15:25:16 I think you can force a linker warning if they are referenced but I don't remember the magic 2015-03-04T15:25:37 gedare: I'll have a look and try to generate this error again, DrJoel does it occur with --enable-multilib also? 2015-03-04T15:26:30 no idea. I am already building two architectures. I hope to start a build soon for that. 2015-03-04T15:26:55 Did the setjmp change get committed to newlib? That was a related issue and the RSB needs a bump 2015-03-04T15:26:57 https://devel.rtems.org/query?status=accepted&status=assigned&status=new&status=reopened&milestone=4.11&col=id&col=summary&col=status&col=type&col=priority&col=milestone&col=component&order=priority 2015-03-04T15:26:59 we're getting close! 2015-03-04T15:27:07 *** starprince has quit IRC (Ping timeout: 246 seconds) 2015-03-04T15:27:21 DrJoel: I only just now pointed the ticket out to Hesham. 2015-03-04T15:27:49 gedare: Ah, I think this relate to or1knd definition also 2015-03-04T15:29:11 It should be added here https://github.com/heshamelmatary/or1k-rtems/blob/master/patches/newlib-cvs17072014.diff#L61 2015-03-04T15:30:29 I will pull the latest newlib head and see if Stefan has fixed it or not. 2015-03-04T15:31:04 i see it 2015-03-04T15:31:07 on newlib head 2015-03-04T15:31:13 #if defined(__or1k__) || defined(__or1knd__) 2015-03-04T15:31:14 #define _JBLEN 31 /* 32 GPRs - r0 */ 2015-03-04T15:31:14 #define _JBTYPE unsigned long 2015-03-04T15:31:14 #endif 2015-03-04T15:31:37 so maybe it works now, eh? 2015-03-04T15:31:48 Then it is a matter of bumping to latest newlib monthly tarball. Did it manage to make that? 2015-03-04T15:32:09 Yeah, that's what my proposed solution 2015-03-04T15:32:28 probably i think those went in a few weeks ago 2015-03-04T15:32:48 9502c33b (Jeff Johnston 2014-12-12 2015-03-04T15:32:53 been there since december. 2015-03-04T15:33:11 == a few weeks in my world.... 2015-03-04T15:33:40 ok.. then I need to build a new toolset 2015-03-04T15:34:13 Hesham.. try it and let us know. I am trying to implement workarounds for all the libdl test issues so they aren't blocking us branching 2015-03-04T15:34:16 we'll want to bump the newlib version when we release anyway, to pick up recent changes 2015-03-04T15:34:25 probably the next monthly snapshot 2015-03-04T15:34:47 Yep 2015-03-04T15:36:11 i'm happy with a release if we can get the libdl workarounds done. 2015-03-04T15:38:59 Ok, it's solved in the latest newlib svn head 2015-03-04T15:39:17 It's a matter of getting recent version of newlib from RSB 2015-03-04T15:39:32 Have we worked through most of the tickets? 2015-03-04T15:39:45 I still have some server infrastructure ones I want to make sure are dealt with 2015-03-04T15:39:48 irc logs 2015-03-04T15:40:54 %define newlib_version ebbb290f6e15e71b140a911c0de530453b9f8443 2015-03-04T15:41:05 This is at config/tools/rtems-gcc-4.8.3-newlib-git-1.cfg 2015-03-04T15:41:36 How should I get a newer version of this newlib git repo? 2015-03-04T15:43:13 [trac] #2257/defect (closed) updated by gedare (This has been fixed in Newlib since 12/12/2014.) http://devel.rtems.org/ticket/2257 2015-03-04T15:44:04 Hesham: change it to b75c57950318ed657807f87a4c694e7b8d68fcae 2015-03-04T15:44:14 and submit a patch 2015-03-04T15:44:38 gedare: Ok, but how did you get it? ;) 2015-03-04T15:44:47 erm, is that the right config though? 2015-03-04T15:45:35 ahh i see 2015-03-04T15:45:36 It's the latest, I'll update it, build the toolchain, built RTEMS and see if things are working as expected 2015-03-04T15:45:50 i have newlib.git checked out on my computer 2015-03-04T15:46:14 Ah, Ok, I only have svn version 2015-03-04T15:46:42 you could also consider switching or1k to use rtems-gcc-4.9-newlib-git-1.cfg 2015-03-04T15:46:53 anyways. 2015-03-04T15:47:19 most of the bsets are using rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-04T15:47:37 we should aim to get as many as possible on the same toolchain versions 2015-03-04T15:48:20 DrJoel: should i bump the broken smp thread migration to 4.11.1? 2015-03-04T15:48:39 *** krzysiekm13 has quit IRC (Quit: Leaving) 2015-03-04T15:48:40 mmm, this will take sometime, because I'll have to re-generate a whole patch for or1k to gcc-4.9* 2015-03-04T15:50:13 oh 2015-03-04T15:50:19 it's not in the upstream? 2015-03-04T15:50:22 don't worry about that then. 2015-03-04T15:50:28 unless chris asks for it :0 2015-03-04T15:53:15 No 2015-03-04T15:53:27 ok 2015-03-04T15:53:36 just bump the newlib version then. 2015-03-04T15:54:02 RSB is building the or1k toolchain now 2015-03-04T15:54:27 it looks like two other toolchains use that config, m32c and h8300 2015-03-04T15:54:36 might want to build those too, to be sure they don't break now 2015-03-04T15:55:59 Ok, I'll but not sure I can make it today 2015-03-04T15:56:22 sure. 2015-03-04T16:00:12 [trac] #2219/defect (new) updated by gedare (You could do something like what is done for libdl in cpukit/configure.ac. Or you ...) http://devel.rtems.org/ticket/2219 2015-03-04T16:00:40 i gotta go myself. cya 2015-03-04T16:02:00 See you 2015-03-04T16:05:44 *** gedare has quit IRC (Ping timeout: 265 seconds) 2015-03-04T16:31:50 Seems like there is a recent commit the broke or1k 2015-03-04T16:31:53 cpukit/posix/src/mprotect.c:31:22: fatal error: sys/mman.h: No such file or directory 2015-03-04T16:32:44 bootstrap -p? in cpukit? 2015-03-04T16:35:16 I am cloning and bootstraping a new repo, maybe that's the problem because I was building from and old repo after doing git pull 2015-03-04T16:36:33 Maybe that will help 2015-03-04T16:37:17 *** Davidbrcz has quit IRC (Ping timeout: 246 seconds) 2015-03-04T16:56:34 [trac] #2279/defect (assigned) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2279 2015-03-04T17:00:59 This commit broke or1k build also 4e3d9a4d6c76fba8e31138d503f736405dafc213 2015-03-04T17:01:07 score: Make available 2015-03-04T17:01:23 Make a ticket for it. We should get that fixed. 2015-03-04T17:01:24 atomic.h:20:35: fatal error: rtems/score/cpuatomic.h: No such file or directory 2015-03-04T17:01:28 Can you tell what broke it? 2015-03-04T17:01:54 I have that error also 2015-03-04T17:02:30 Yes 2015-03-04T17:02:39 or1k port does not have a cpuatomic.h file 2015-03-04T17:03:07 Every score/cpu/* has to include include_rtems_score_HEADERS += rtems/score/cpuatomic.h 2015-03-04T17:03:12 or1k doesn't 2015-03-04T17:03:33 The one for mips is the default version. Just copy it over and add the line. Post a patch and I will merge it immediately 2015-03-04T17:03:44 Ok 2015-03-04T17:04:30 I would do it but the machine with or1k tools is swamped. I might get it right or not 2015-03-04T17:05:41 *** jhollister has quit IRC (Ping timeout: 252 seconds) 2015-03-04T17:06:04 Don't worry I have the proper env to build, doing it now 2015-03-04T17:06:53 I am closing or deferring tickets. Up to four or five so far but the testing is slow. Won't push until I have addressed all 2015-03-04T17:08:12 I believe there are other "broken" architectures due to this commit. 2015-03-04T17:09:28 Great, now or1k builds find. Have to try with --enable-multilib now 2015-03-04T17:10:17 Should I create a ticker for this issue? 2015-03-04T17:17:05 ticket* 2015-03-04T17:18:22 *** kiwichris has quit IRC (Ping timeout: 240 seconds) 2015-03-04T17:24:33 [git] 3358d62 by Hesham ALMatary: [score/cpu/or1k: Add cpuatomic.h to fix broken build. 4e3d9a4d6c76fba8e31138d503f736405dafc213 broke or1k, cpuatomic.h has to be added to all architectures.] http://devel.rtems.org/changeset/3358d62/rtems 2015-03-04T17:24:55 Bam 2015-03-04T17:24:58 Not now. :) 2015-03-04T17:25:04 top 2015-03-04T17:27:04 *** beowulf_ has joined #rtems 2015-03-04T17:28:46 *** kiwichris has joined #rtems 2015-03-04T17:29:15 About time you showed 2015-03-04T17:29:25 kiwichris ... 2015-03-04T17:29:35 Hi beowulf_ 2015-03-04T17:31:10 Hey 2015-03-04T17:31:24 Task variables should be killed off. 2015-03-04T17:32:56 Ok. Just a ticket and procedural now. How do we document deprecation and put them in the "enable-obsolete" category 2015-03-04T17:33:28 *** beowulf_ has quit IRC (Ping timeout: 246 seconds) 2015-03-04T17:34:17 FYI I am working on work arounds for all the libdl tickets. So you can ignore them a bit longer 2015-03-04T17:36:40 DrJoel: Perfect. I have a full or1k build with --enable-multilib 2015-03-04T17:36:51 Great! One more issue down 2015-03-04T17:37:12 I'll test other architectures gedare told me about tomorrow and submit the RSB patch 2015-03-04T17:37:56 If you can bump binutils for the moxie to the git head and build a moxie toolset and moxiesim that would kill a ticket and we would be greatful 2015-03-04T17:38:36 Sure, I'll do it tomorrow 2015-03-04T17:38:54 Have to go sleep now it's almost 12AM ;) 2015-03-04T17:39:15 See you 2015-03-04T17:39:55 grateful* ;) 2015-03-04T17:41:24 DrJoel, thanks for the libdl work. I should be getting some more time soon 2015-03-04T17:42:32 You're welcome. Don't worry about it until I say something. I intend to fix or defer all of them with workarounds 2015-03-04T17:43:05 This ticket is yours though: https://devel.rtems.org/ticket/2205 2015-03-04T17:44:24 *** Hesham has quit IRC (Ping timeout: 245 seconds) 2015-03-04T17:46:48 Understood 2015-03-04T17:47:05 Just testing my Windows patch on FreeBSD to make sure I have not broken anything 2015-03-04T17:47:34 for the RSB 2015-03-04T17:48:25 After that it is just the broken archs on Windows and being able to more than one core (MSYS2 or binutils temp file bug) 2015-03-04T17:50:59 Are all targets movign up on gdb? 2015-03-04T17:53:14 They have to for Windows to build 2015-03-04T17:53:34 Which means SPARC is broken until we can get a suitable patch mix from Jiri 2015-03-04T17:56:14 OK. That will settle out 2015-03-04T17:56:24 I thought he mentioned posting some and getting them merged to git 2015-03-04T17:58:03 Heading to grocery store. Ice storm hits in the morning and this is normal "we need stuff" shopping :( 2015-03-04T17:58:08 bye 2015-03-04T18:00:25 *** DrJoel has quit IRC (Quit: Friends help you move. Real friends help you move bodies.) 2015-03-04T18:01:08 [trac] #2277/defect (closed) updated by amar (For anyone curious this is an issue on FreeBSD see here: ...) http://devel.rtems.org/ticket/2277 2015-03-04T18:10:24 [trac] Developer/Release created http://devel.rtems.org/wiki/Developer/Release 2015-03-04T18:26:08 The RSB removeall code to delete a tree is now broken on Unix ... who would have thought deleting a tree of files was so hard. 2015-03-04T20:16:05 *** gedare has joined #rtems 2015-03-04T20:16:05 *** ChanServ sets mode: +o gedare 2015-03-04T20:21:45 *** gedare has quit IRC (Ping timeout: 256 seconds) 2015-03-05T00:00:17 *** monstr has joined #rtems 2015-03-05T00:38:33 *** dr__house is now known as dr__house|away 2015-03-05T00:52:01 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-05T00:53:07 *** sebhub has joined #rtems 2015-03-05T00:55:32 *** zoso has joined #rtems 2015-03-05T01:11:14 good morning 2015-03-05T01:14:46 *** sgulati_ has joined #rtems 2015-03-05T01:19:28 *** sgulati_ has quit IRC (Client Quit) 2015-03-05T01:29:23 *** kiwichris has quit IRC () 2015-03-05T01:36:49 [git] 8ac3549 by Sebastian Huber: [score: Delete unused CPU_UNROLL_ENQUEUE_PRIORITY] http://devel.rtems.org/changeset/8ac3549/rtems 2015-03-05T01:42:59 *** zoso has quit IRC (Ping timeout: 244 seconds) 2015-03-05T01:46:03 [git] 18d0a49 by Sebastian Huber: [Move test support from sapi to libmisc/testsupport] http://devel.rtems.org/changeset/18d0a49/rtems 2015-03-05T01:46:03 [git] 51acbdc by Alexander Krutwig: [score: Documentation] http://devel.rtems.org/changeset/51acbdc/rtems 2015-03-05T01:48:54 *** sgulati has joined #rtems 2015-03-05T02:00:04 *** zoso has joined #rtems 2015-03-05T02:01:03 *** zoso has joined #rtems 2015-03-05T02:02:18 *** zoso has joined #rtems 2015-03-05T02:03:38 *** zoso has joined #rtems 2015-03-05T02:04:42 *** zoso has joined #rtems 2015-03-05T02:05:16 *** zoso has quit IRC (Max SendQ exceeded) 2015-03-05T02:05:45 *** zoso has joined #rtems 2015-03-05T02:06:41 *** zoso has joined #rtems 2015-03-05T02:07:08 *** zoso has quit IRC (Max SendQ exceeded) 2015-03-05T02:08:00 *** zoso has joined #rtems 2015-03-05T02:09:30 *** zoso has joined #rtems 2015-03-05T02:10:25 *** zoso has joined #rtems 2015-03-05T02:11:39 *** zoso has joined #rtems 2015-03-05T02:12:34 *** zoso has joined #rtems 2015-03-05T02:13:08 *** zoso has quit IRC (Max SendQ exceeded) 2015-03-05T02:13:45 *** zoso has joined #rtems 2015-03-05T02:14:34 [git] 7f577d3 by Alexander Krutwig: [tests: Refactor parallel test execution] http://devel.rtems.org/changeset/7f577d3/rtems 2015-03-05T02:14:37 *** sgulati has quit IRC (Ping timeout: 246 seconds) 2015-03-05T02:14:42 *** zoso has joined #rtems 2015-03-05T02:15:07 *** zoso has quit IRC (Max SendQ exceeded) 2015-03-05T02:15:51 *** zoso has joined #rtems 2015-03-05T02:16:44 *** zoso has joined #rtems 2015-03-05T02:17:10 *** zoso has quit IRC (Max SendQ exceeded) 2015-03-05T02:17:41 *** zoso has joined #rtems 2015-03-05T02:18:46 *** zoso has joined #rtems 2015-03-05T02:19:10 *** zoso has quit IRC (Max SendQ exceeded) 2015-03-05T02:19:57 *** zoso has joined #rtems 2015-03-05T02:21:08 *** zoso has joined #rtems 2015-03-05T02:22:03 *** zoso has joined #rtems 2015-03-05T02:23:10 *** zoso has joined #rtems 2015-03-05T02:24:14 *** zoso has joined #rtems 2015-03-05T02:34:30 *** Hesham has joined #rtems 2015-03-05T03:41:55 *** kiran4399 has joined #rtems 2015-03-05T03:55:22 *** kiwichris has joined #rtems 2015-03-05T03:56:49 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-05T04:30:08 *** Davidbrcz has joined #rtems 2015-03-05T04:36:00 [trac] #2219/defect (closed) updated by Sebastian Huber (In [changeset:"77510c72dc2c49c2ade13caada77efb18a1adeb8/rtems"]: {{{ ...) http://devel.rtems.org/ticket/2219 2015-03-05T04:36:01 [git] 5b5d2fd by Sebastian Huber: [Filesystem: Delete unused null_op_fsmount_me()] http://devel.rtems.org/changeset/5b5d2fd/rtems 2015-03-05T04:36:02 [git] b8aa1c1 by Sebastian Huber: [score: C/C++ compatibility macros for ISR locks Update #2273.] http://devel.rtems.org/changeset/b8aa1c1/rtems 2015-03-05T04:36:02 [git] 5d42857 by Sebastian Huber: [libblock: Fix warning] http://devel.rtems.org/changeset/5d42857/rtems 2015-03-05T04:36:02 [git] 81c73ac by Sebastian Huber: [IMFS: Fix warning] http://devel.rtems.org/changeset/81c73ac/rtems 2015-03-05T04:36:02 [git] 29d36fa by Sebastian Huber: [IMFS: Include missing header file] http://devel.rtems.org/changeset/29d36fa/rtems 2015-03-05T04:36:02 [git] 2c88438 by Sebastian Huber: [IMFS: Fix warning] http://devel.rtems.org/changeset/2c88438/rtems 2015-03-05T04:36:02 [git] b4fa5a5 by Sebastian Huber: [bsps: Fix warning] http://devel.rtems.org/changeset/b4fa5a5/rtems 2015-03-05T04:36:03 [git] 5155069 by Sebastian Huber: [libchip/serial: Fix warning] http://devel.rtems.org/changeset/5155069/rtems 2015-03-05T04:36:03 [git] d006128 by Sebastian Huber: [tests: Fix warnings] http://devel.rtems.org/changeset/d006128/rtems 2015-03-05T04:36:04 [git] 6fd5b4e by Sebastian Huber: [tests: Fix warning] http://devel.rtems.org/changeset/6fd5b4e/rtems 2015-03-05T04:36:04 [git] c735cd5 by Sebastian Huber: [dosfs: Fix warnings] http://devel.rtems.org/changeset/c735cd5/rtems 2015-03-05T04:36:05 [git] 16b2de3 by Sebastian Huber: [capture: Fix warnings] http://devel.rtems.org/changeset/16b2de3/rtems 2015-03-05T04:36:05 [git] 77510c7 by Sebastian Huber: [libmd: Disable SHA modules for m32c Close #2219.] http://devel.rtems.org/changeset/77510c7/rtems 2015-03-05T04:39:08 *** andrei_ has joined #rtems 2015-03-05T04:39:19 Hello! 2015-03-05T04:39:49 *** zoso has quit IRC (Quit: Leaving) 2015-03-05T04:46:36 I'm looking over your GSoC ideas (https://devel.rtems.org/wiki/Developer/OpenProjects) and I'm having a difficulty in choosing one. 2015-03-05T04:48:05 I'll properly introduce myself on the mailinglist, but I'm trying to figure out if there s anything I can tackle. 2015-03-05T04:53:05 I've been an Linux kernel developer at Freescale and I'm currently developing real time voip apps(C++). On short, I believe I have a decent understanding of OS and real time concepts. What I'm looking for is a project that I could deliver. 2015-03-05T04:53:25 I've been a GSoC participant twice before(at BRL-CAD), I will also detaliate this 2015-03-05T04:53:48 and I feel it s common to "propose" a lot more than it can be done, then end up in an in-between state 2015-03-05T04:59:50 *** Davidbrcz has quit IRC (Ping timeout: 244 seconds) 2015-03-05T05:04:04 At a quick glance , https://devel.rtems.org/wiki/Developer/Projects/Open/Condition_Variables this is the idea I feel like working on the most. Mainly because I'm also passionate about concurency, parallelism and such. Any thoughts, questions or suggestions? Thank you! 2015-03-05T05:13:31 Hi andrei_, I would say that you ask on the mailing list directly as currently IRC is not logged, and potential mentors for related-projects may not be here (or busy). 2015-03-05T05:16:51 Hesham, thanks ! I will do that, then. 2015-03-05T05:23:56 *** SprinkleBerryMuf has joined #rtems 2015-03-05T05:58:35 Is RTEMS the only OS org accepted this year in GSoC? I can't find minix, fedora, Linux, HelenOS, NetBSD,... 2015-03-05T06:07:02 Hesham: there is FreeBSD 2015-03-05T06:07:50 Yeah I noticed. Still great for RTEMS, I think. 2015-03-05T06:28:39 *** SprinkleBerryMuf has quit IRC (Quit: leaving) 2015-03-05T07:02:13 *** Davidbrcz has joined #rtems 2015-03-05T08:08:17 *** gedare has joined #rtems 2015-03-05T08:08:17 *** ChanServ sets mode: +o gedare 2015-03-05T08:08:54 that's interesting 2015-03-05T08:09:16 i wonder if we'll see an uptick in applications. 2015-03-05T08:10:14 fedora is there. 2015-03-05T08:10:23 *** kiran4399 has joined #rtems 2015-03-05T08:10:49 and RIOT 2015-03-05T08:11:17 gedare: how many students did you have last year? 2015-03-05T08:12:14 andrei_: we scaled back last year.. i think we had 5 GSoC and 1 SOCIS student 2015-03-05T08:12:50 i think we had 10 one year recently. but it is hard on the mentors we get spread thin 2015-03-05T08:14:02 Weird I searched for Fedora before but didn't find it. 2015-03-05T08:14:23 the linux kernel is missing 2015-03-05T08:14:30 gedare: BTW, I tried to build the BSPs you told me about yesterday, but both fail. 2015-03-05T08:14:51 I don't think it's related to the newlib update 2015-03-05T08:15:11 *** andrei___ has joined #rtems 2015-03-05T08:15:25 Have you tried to build BSPs for these architectures recently? 2015-03-05T08:15:42 ok thanks Hesham i have not 2015-03-05T08:16:01 can you submit the RSB patch to bump newlib for or1k? 2015-03-05T08:16:49 Yeah, it's already there, but what about the h8300 and m32c failures? 2015-03-05T08:17:16 For h8300, the error is " ibmisc/shell/main_ping.c:1148:1: error: insn does not satisfy its constraints " 2015-03-05T08:17:31 seems like an internal GCC compiler error, not sure 2015-03-05T08:18:37 *** andrei_ has quit IRC (Ping timeout: 246 seconds) 2015-03-05T08:20:14 andrei___: are you planning to participate as a gsoc student again this year? 2015-03-05T08:20:41 Hesham: that sounds like an asm problem 2015-03-05T08:20:46 *** andrei___ is now known as andrei_ 2015-03-05T08:20:54 *** andrei_ has joined #rtems 2015-03-05T08:20:54 *** andrei_ has joined #rtems 2015-03-05T08:21:11 :q 2015-03-05T08:21:13 gedare, I hope so 2015-03-05T08:21:19 oops. wrong window. 2015-03-05T08:21:37 do you think you will do more with raspberry pi or something else? 2015-03-05T08:22:55 hrmm 2015-03-05T08:23:02 or am I mistaking you for someone else? 2015-03-05T08:23:08 you are 2015-03-05T08:23:12 I was just about to write that 2015-03-05T08:23:16 I think there's a confusion around. I haven't worked with raspberry pi so far. 2015-03-05T08:23:27 I've been a gsoc student at BRL-CAD, in 2012 and 2014 2015-03-05T08:23:44 ohh ok! 2015-03-05T08:23:55 yes I was thinking of Andre. :) 2015-03-05T08:24:31 yeah, it's a really common name 2015-03-05T08:24:46 ok, well, welcome! 2015-03-05T08:24:51 thank you ! 2015-03-05T08:24:53 is BRL-CAD in gsoc this year? 2015-03-05T08:24:56 it is 2015-03-05T08:25:08 but it's their last year, so I thought it would be better to leave them a chance to take fresh talent 2015-03-05T08:25:11 are you just looking for something new then? ahh. 2015-03-05T08:25:24 I've also been a Google Code in mentor at them for the past 3 years 2015-03-05T08:25:24 what kind of projects are you interested in? 2015-03-05T08:25:29 so I' 2015-03-05T08:25:36 so I'm "quite bond" with them. 2015-03-05T08:25:57 ok great. i met Sean at the GCI meet last spring. 2015-03-05T08:26:08 ah, awesome ! Small world, isn't it? 2015-03-05T08:26:18 yeah. 2015-03-05T08:27:48 Regarding the project 2015-03-05T08:28:32 I'm not 100% sure yet, that's the main reason I haven't sent an introduction email so far 2015-03-05T08:28:57 ok 2015-03-05T08:29:09 I'm more experienced in low level 2015-03-05T08:29:15 so that would be the general area 2015-03-05T08:29:21 then you probably want the runtime-oriented projects? 2015-03-05T08:30:04 yes 2015-03-05T08:32:41 I find synchronization interesting 2015-03-05T08:33:24 that's why I picked Condition Variables 2015-03-05T08:34:40 great. that is a good project i think, but you should ask sebhub if it is good to undertake for gsoc 2015-03-05T08:35:02 depending on whether he is here, or you might want to email devel@rtems.org. 2015-03-05T08:35:18 Why I'm reluctant is because, in my limited experience so far, multithreaded/concurency can also cause obscure, tough to handle bugs 2015-03-05T08:35:30 that's the fun part! 2015-03-05T08:35:41 due to it's nondeterministic nature 2015-03-05T08:36:22 At least in my case, with GSoC there's always the tendency to propose so much 2015-03-05T08:36:55 (pressure to get accepted, will to do something as great as you possibly could) 2015-03-05T08:37:11 and end up in a state that it's OK to pass a GSoC, but not realistically usable for the org. 2015-03-05T08:37:11 [trac] #2257/defect (closed) updated by Hesham ALMatary (In [changeset:"ee85d9cfb62d2cb914f54a24a74a8091c4751490/rtems-source-builder"]: {{{ ...) http://devel.rtems.org/ticket/2257 2015-03-05T08:37:11 [git] ee85d9c by Hesham ALMatary: [RSB: Fetch recent version of newlib that solves or1knd build failure. Updates #2257.] http://devel.rtems.org/changeset/ee85d9c/rtems-source-builder 2015-03-05T08:37:26 thanks Hesham i modified the patch a little so you'll need to re-pull your master. 2015-03-05T08:37:45 I really want to avoid that, I feel that I got some experience (with GSoC) and this is something that shouldn' 2015-03-05T08:37:48 t happen. 2015-03-05T08:37:58 we prefer you to propose the right amount to match your capabilities 2015-03-05T08:38:06 we have rejected students in the past who propose too much 2015-03-05T08:38:16 because we don't think they understood the scope of their work 2015-03-05T08:38:44 gedare: Ah, so "Updates #" can have an effect even if the ticket was closed 2015-03-05T08:38:46 we do everything we can to get the code into usable state 2015-03-05T08:38:50 yeah 2015-03-05T08:39:43 I was wondering how this rtems-bot recognizes me from the committed patch 2015-03-05T08:40:33 the author field of the patch 2015-03-05T08:40:53 in git the author and committer are separated 2015-03-05T08:40:56 So it compares my name there, and the member of this channels 2015-03-05T08:41:02 no 2015-03-05T08:41:18 your IRC client is highlighting your name whenever it sees it 2015-03-05T08:41:33 Ah, that's what I mean 2015-03-05T08:41:38 Got it, thank you. 2015-03-05T08:41:41 the bot just copies the author field, which includes your name. right. 2015-03-05T08:42:43 DrJoel asked me to update RSB moxie binutils, I think it's already fixed now? 2015-03-05T08:43:20 hmm 2015-03-05T08:43:47 tools/rtems-binutils-2.24-1 2015-03-05T08:44:13 i think we want rtems-binutils-git-1 2015-03-05T08:44:33 and to bump the binutils_version 2015-03-05T08:44:48 i don't know where kiwichris has been tracking these versions though 2015-03-05T08:55:44 gedare: Ok, I'll do it now. 2015-03-05T09:03:04 I think this is the first time anybody builds moxie toolchain from RSB 2015-03-05T09:03:39 Because its referring to the wrong config file of dtc, and abort building from the early begining 2015-03-05T09:07:59 oh right that... 2015-03-05T09:08:04 was supposed to be fixed 2015-03-05T09:08:28 see #2209 and #2202 2015-03-05T09:11:01 Ok, I fixed it and it's building now 2015-03-05T09:11:53 *** monstr has quit IRC (Remote host closed the connection) 2015-03-05T09:12:48 thanks much appreciated 2015-03-05T09:18:59 gedare: mmmmm, I reverted back to the old version of dtc, because the recent one fire a compilation error. Is that Ok? 2015-03-05T09:19:31 I have no idea! See if the tools and rtems build with it 2015-03-05T09:19:53 btw how did it go with epiphany-newlib? 2015-03-05T09:20:36 Yeah RTEMs builds with the older version just fine. 2015-03-05T09:21:21 No progress regarding epiphany-newlib, I believe what's missing it an include to limits.h 2015-03-05T09:21:53 ok 2015-03-05T09:22:02 go ahead and submit patch for the older version 2015-03-05T09:22:18 I discovered that t-rtems has to be added part of config.gcc file, but when I tried to compile GCC again, it generates pre-processing error when processing this t-rtems file 2015-03-05T09:22:23 Ok 2015-03-05T09:22:28 and cc green AT moxielogic com 2015-03-05T09:22:58 he is sort of the moxie person. 2015-03-05T09:23:09 i see. 2015-03-05T09:26:04 Is beng-nl there ? 2015-03-05T09:26:05 Ok, how can I cc from the commit message? 2015-03-05T09:26:11 beng-nl: 2015-03-05T09:26:22 when you do git-send-email it should prompt you for anyone else to cc 2015-03-05T09:26:57 Ah ok 2015-03-05T09:28:58 kiran4399: 2015-03-05T09:31:58 So I asked for the cable ... My friend says he does not have it.. What should I do ?? :-( 2015-03-05T09:34:44 Order it? 2015-03-05T09:36:12 Hesham: that message for me ? :-) 2015-03-05T09:36:50 *** DrJoel has joined #rtems 2015-03-05T09:36:51 *** DrJoel has joined #rtems 2015-03-05T09:36:51 *** ChanServ sets mode: +o DrJoel 2015-03-05T09:37:56 well it was 70F yesterday and 27F now. Just enough ice and snow to mess things up here 2015-03-05T09:37:57 Yeah, I had to order tty to USB adapter when I worked on Raspberry Pi project. It didn't take more than three days for me, not sure about your cable. 2015-03-05T09:38:52 [RSB] git: clone: git://sourceware.org/git/binutils-gdb.git -> sources/git/binutils-gdb.git 2015-03-05T09:38:52 is taking forever :/ 2015-03-05T09:39:17 sourcware.org is an overloaded machine 2015-03-05T09:40:40 So why are we switching moxie binutils to it? 2015-03-05T09:41:13 Some instruction formats changed in assembly and the only way to get a binutils that assembles gcc generated code is with a snapshot 2015-03-05T09:42:34 I see 2015-03-05T09:42:55 Hesham: It will take atleast 5-6 days for me.. what do you think ? 2015-03-05T09:43:35 DrJoel: This may be a reason for the assembler error I got when tried to build h8300 2015-03-05T09:44:18 kiran4399: I'm sure beng-nl would give you a better advice here. 2015-03-05T09:44:59 you absolutely need it to do anything on the BBB kiran4399 2015-03-05T09:45:54 so my suggestion is to order it if you're serious about the project, and use qemu in the meantime 2015-03-05T09:45:59 If this is ordering hardware, you will need to do a hello world project. beng-nl can he use qemu for bb to do that? 2015-03-05T09:46:16 Obviously yes. 2015-03-05T09:46:41 yep 2015-03-05T09:47:05 kiran4399: for the proposal stage you can use qemu, but you must order the cable to have any hope to be successful with the hardware. 2015-03-05T09:50:26 gedare: Yes sure.. 2015-03-05T09:54:38 DrJoel: we had rain all night and now snow is layering over the top of the freezing asphalt. it will be pretty nasty out there this afternoon. 2015-03-05T09:55:31 Ours started 5-6am and the driveways and road I see are not drivable. No cars have passed in the hour I have sat here 2015-03-05T09:57:40 *** andrei_ has quit IRC (Ping timeout: 246 seconds) 2015-03-05T10:03:01 *** sgulati has joined #rtems 2015-03-05T10:03:15 gedare: Which ticket should I update regarding the dtc failure? 2202? 2015-03-05T10:05:14 probably. 2015-03-05T10:05:39 well DrJoel you are in a state of emergency right 2015-03-05T10:05:52 seems like a good day to finish closing the tickets and cut the release. ;) 2015-03-05T10:06:15 yes. makes it possible to give cities/counties extra leeway and school day doesn't have to be made up. 2015-03-05T10:06:24 Hesham: also 2254 2015-03-05T10:06:33 if you get it working close that one. 2015-03-05T10:06:40 Gedare .. commit Hesham's as they show up 2015-03-05T10:06:45 i have been 2015-03-05T10:06:47 FYI newlib is moving to git 2015-03-05T10:06:57 DrJoel: So, I have built the moxie toolchain with binutils git repo, I still get an error when building RTEMS for moxie.... 2015-03-05T10:06:59 start.S:21: Error: unknown opcode sub $r2,$r0 2015-03-05T10:08:16 mmmmmm 2015-03-05T10:08:36 So, I guess binutils_version has to be updated? 2015-03-05T10:08:53 at rtems-binutils-git-1.cfg file 2015-03-05T10:10:29 DrJoel: nice 2015-03-05T10:11:10 where did you hear about that? 2015-03-05T10:11:34 also, i started a release checklist if anyone has anything to add: https://devel.rtems.org/wiki/Developer/Release 2015-03-05T10:12:40 I think I have a workaround for all the libdl related issues plus an h300 build failure. It should never have attempted to build tcpip. It is too small a CPU to even hold it and 16 bit so beyond design constraints for BSD 2015-03-05T10:13:03 I am about to start a complete build of all BSPs. 2015-03-05T10:13:44 Is there a ticket for lm32 tools? We have a gcc ICE building libdl. 2015-03-05T10:13:55 There should be a gcc PR for this. 2015-03-05T10:14:06 We need a PR to track that. When we have it, I will add a workaround 2015-03-05T10:15:35 Is the or1k tools patch committed? I am building new nios2 tools now before the sweep so might as well do that as well 2015-03-05T10:15:36 Ok, I am building moxie toolchain with a recent binutils-gdb git version and will wee if that would fix the assembler error I get when building RTEMS for moxie. 2015-03-05T10:16:24 The one for newlib git is not yet committed 2015-03-05T10:17:13 I was waiting to build other gcc4.8.3 dependent architectures 2015-03-05T10:17:42 h8300 fails with this 16 bit size array 2015-03-05T10:17:49 verm__: corinna annoucned it on newlib ml. 2015-03-05T10:17:52 I have a fix pending for that 2015-03-05T10:17:58 It will happen next week though 2015-03-05T10:18:15 corinna is apparently rolling it herself. i guess she doesn't know about the existing mirror. 2015-03-05T10:18:25 Ok, so I can submit the newlib patch that fixed or1knd build? 2015-03-05T10:19:10 RSB reference to a recent version of newlib I mean 2015-03-05T10:19:25 yes that would be fine. wasn't that already done? 2015-03-05T10:20:05 mmmm, yeah, sorry I didn't sleep well :/ 2015-03-05T10:20:19 So now all or1k patches are commited 2015-03-05T10:20:45 And RSB fetches a newlib with the fix? snashot if possible 2015-03-05T10:22:38 right now it is getting a particular git version. 2015-03-05T10:22:53 we can ask for a snapshot before they take cvs offline 2015-03-05T10:23:09 and the patch didn't make the last snapshot? 2015-03-05T10:23:10 Like with current binutils-gdb-git 2015-03-05T10:23:10 but i think what will happen is we will use the next monthly snapshot in the 4.11 tools 2015-03-05T10:23:26 umm, actually yes the patch should be in the last snapshot 2015-03-05T10:23:29 if not, then let's use git for a month. No big deal. Only impacts a few people. We can switch to snapshot in a month 2015-03-05T10:23:44 but we have other fixes in newlib that we committed, like the CLOCKS_PER_SEC 2015-03-05T10:23:53 and marked as fixed in 4.11 2015-03-05T10:23:54 OK. 2015-03-05T10:24:01 Do I need to build all new tools? 2015-03-05T10:24:04 so all tools should aim to use newlib if possible. 2015-03-05T10:24:10 um 2015-03-05T10:24:22 lets get the tickets closed 2015-03-05T10:24:23 Commit the or1k rsb change and I will start a tool build 2015-03-05T10:24:27 and worry about the tool versions later 2015-03-05T10:24:33 i already did commit it. 2015-03-05T10:24:37 *** sgulati has quit IRC (Ping timeout: 246 seconds) 2015-03-05T10:25:02 *** sgulati has joined #rtems 2015-03-05T10:25:47 ok. I will build or1k tools when nios finishes and then start a BSP sweep 2015-03-05T10:28:50 I have a two-hour lab now, I am leaving moxie toolchain building with newer binutils-gdb-git version. 2015-03-05T10:31:00 *** DrJoel_ has joined #rtems 2015-03-05T10:31:41 This is really me in another room. Was on the old laptop I use for remote access. Now on my work laptop to deal with other things 2015-03-05T10:31:47 I will switch back and forth 2015-03-05T10:31:56 ok thanks Hesham you can submit patch when you get back please 2015-03-05T10:32:03 ok 2015-03-05T10:47:59 DrJoel_/DrJoel: did you find any branches with relevant code for https://devel.rtems.org/ticket/2245 2015-03-05T10:48:35 no one responded to my inquiry about deprecating an interface. 2015-03-05T10:49:41 yes.. after I deal with the current set of issues. 2015-03-05T10:49:56 I am going to add psxhdrs tests for all the methods in socket.h and ensure they are correct 2015-03-05T11:01:00 *** dr__house|away is now known as dr__house 2015-03-05T11:16:51 *** ragunath has joined #rtems 2015-03-05T11:23:29 *** ragunath has left #rtems 2015-03-05T11:29:02 *** Davidbrcz has quit IRC (Ping timeout: 245 seconds) 2015-03-05T11:29:34 *** ragunath has joined #rtems 2015-03-05T11:48:25 sigh.. got about a dozen BSPs in to the build before I noticed that I needed to rebootstrap. :( 2015-03-05T11:56:16 *** Davidbrcz has joined #rtems 2015-03-05T12:02:30 *** kiran4399 has left #rtems 2015-03-05T12:03:46 DrJoel: where are you building? 2015-03-05T12:03:57 gedare: ah, nice 2015-03-05T12:13:06 i wonder if they need any help switching 2015-03-05T12:13:18 converting from old cvs repos to git can be a real PITA. 2015-03-05T12:15:34 *** sgulati has quit IRC (Ping timeout: 246 seconds) 2015-03-05T12:22:02 *** Davidbrcz has quit IRC (Ping timeout: 246 seconds) 2015-03-05T12:25:49 on rtbf64a.. everything is setup there. Need to get used to using sync more 2015-03-05T12:26:16 verm__: Email Corrrina and ask. I am sure she won't turn anyone down if she needs it 2015-03-05T12:26:29 FYI gdb is starting to use Sphinx for some docs 2015-03-05T12:27:58 crap.. now I have run into gedare's change for CLOCKS_PER_SEC. I am going to have to build all new tools. 2015-03-05T12:28:09 gedare: is RSB up to date to grab that? 2015-03-05T12:38:49 *** andrei_ has joined #rtems 2015-03-05T12:45:07 DrJoel: no 2015-03-05T12:46:30 DrJoel: i'm not sure what procedures are in place for defining the "authoritative set" of tool versions. afaik it is what RSB uses, but i know kiwichris wanted to have a well-defined way of specifying it. 2015-03-05T12:46:53 bumping versions in cfg files feels a bit like "Calvinball" 2015-03-05T13:05:46 rtems-arm.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:47 rtems-avr.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:47 rtems-bfin.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:47 rtems-h8300.bset:tools/rtems-gcc-4.8.3-newlib-git-1 2015-03-05T13:05:47 rtems-i386.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:47 rtems-lm32.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:49 rtems-m32c.bset:tools/rtems-gcc-4.8.3-newlib-git-1 2015-03-05T13:05:51 rtems-m32r.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:53 rtems-m68k.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:55 rtems-microblaze.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:57 rtems-mips.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:05:59 rtems-moxie.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:06:01 rtems-nios2.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:06:03 rtems-or1k.bset:tools/rtems-gcc-4.8.3-newlib-git-1 2015-03-05T13:06:05 rtems-powerpc.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:06:07 rtems-sh.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:06:09 rtems-sparc64.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:06:13 rtems-sparc.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:06:15 rtems-v850.bset:tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 2015-03-05T13:06:17 DrJoel: DrJoel_: looks like most of the tools are using a snapshot of newlib 2015-03-05T13:06:53 Hesham updated the 4.8.3-newlib-git-1 to point to the recent commit, but he said there were some build problems with m32c and h8300... 2015-03-05T13:07:17 so you won't be picking up the recent commits to newlib... any since the first snapshot I guess that was taken in January? 2015-03-05T13:08:09 gedare: The error with h8300 does not relate to newlib 2015-03-05T13:08:10 probalby what this means is that the psxclock01 test won't build. 2015-03-05T13:09:01 Not sure about m32c one 2015-03-05T13:09:13 Its assembler thing as you said 2015-03-05T13:09:26 ok 2015-03-05T13:09:39 did your other build work? 2015-03-05T13:09:40 Ok, now moxie tools does not build 2015-03-05T13:09:44 unknown opcode mul.l $r12,$r6 2015-03-05T13:09:50 moxxie 2015-03-05T13:09:56 hmm. 2015-03-05T13:10:12 that is with binutils master? 2015-03-05T13:10:14 This is with the a recent binutils-gdb git version 2015-03-05T13:10:58 Yeah from master 2015-03-05T13:11:16 %source set binutils git://sourceware.org/git/binutils-gdb.git?fetch?reset=hard?branch=master?checkout=%{binutils_version} 2015-03-05T13:11:42 It was building before I change the version to a recent one. 2015-03-05T13:11:48 what is binutils_version set to? 2015-03-05T13:11:52 i see 2015-03-05T13:12:01 beb8418f4799b50ce414b7a63ac7a2a363dc8a05 2015-03-05T13:12:18 I did it today, 3 hours ago maybe 2015-03-05T13:15:05 hmm there is no mul.l in opcodes / moxie-opc.c 2015-03-05T13:15:51 is the mul.l in some asm in rtems? 2015-03-05T13:16:26 No, I got this error from the RSB/moxie build. 2015-03-05T13:16:54 The problem with RTEMS was at start.S I think it was add instruction 2015-03-05T13:16:54 ahh. 2015-03-05T13:17:28 m32c is a compiler error I think 2015-03-05T13:17:38 DrJoel Yes 2015-03-05T13:17:50 So basically I fixed build errors for ~60 BSPs and now to test them I have to ignore failures in all 180+? Shit 2015-03-05T13:18:15 m32c is a fragile gcc port 2015-03-05T13:18:16 Hesham: need a newer version of gcc 2015-03-05T13:18:43 I guess the only practical solution is to remove psxclock01 from my build locally and file a critical blocking ticket on the RSB 2015-03-05T13:19:16 Hesham: it needs to pick up 2014-12-24 Anthony Green 2015-03-05T13:19:27 gedare: tools/rtems-gcc-4.9.2-newlib-2.2.0-1-1 is not enough? 2015-03-05T13:20:13 no it looks like there were a few modifications, the last one was picked up in 2015-01-10 green Fix mul.x flag for moxie 2015-03-05T13:20:24 pretty recent changes. 2015-03-05T13:20:54 Aha 2015-03-05T13:20:58 the problem is the moxie instruction set changed. pretty uncommon thing to happen 2015-03-05T13:20:59 nios2 barfed at psxclock01. 2015-03-05T13:21:04 Trying sparc/sis 2015-03-05T13:21:10 they all will 2015-03-05T13:22:20 rtems-gcc-4.9-newlib-git-1 ? 2015-03-05T13:22:38 Hesham: for you? or for DrJoel ? 2015-03-05T13:22:39 if that's what is in the RSB, I will continue building tools 2015-03-05T13:22:59 joel, none of the tools (expect or1k) will work. 2015-03-05T13:23:05 on psxclock01 2015-03-05T13:23:25 gedare: I was suggesting to refer to rtems-gcc-4.9-newlib-git-1 to get a recent gcc version for moxie 2015-03-05T13:23:32 OK.. Understood.. then I repeat 2015-03-05T13:23:33 Is "rtems-gcc-4.9-newlib-git-1" the right cfg file ? 2015-03-05T13:23:39 i can fix it by bumping them all to rtems-gcc-4.9-newlib-git-1.cfg 2015-03-05T13:23:40 I guess the only practical solution is to remove psxclock01 from my build locally and file a critical blocking ticket on the RSB 2015-03-05T13:23:51 That would be appreciated. It is a fatal issue now 2015-03-05T13:24:10 I need to take a shower so push whatever you want and I will build tools when I get out 2015-03-05T13:24:26 Hesham: yes, with the right setting for the gcc_version 2015-03-05T13:24:38 DrJoel: ok i will get it 'fixed' 2015-03-05T13:26:30 Thanks .. trust me.. it isnt making it worse 2015-03-05T13:26:55 gedare: Ok to set gcc_version to the latest commit? 2015-03-05T13:27:18 FYI for BSPs on architectures where some BSPs work with libdl, I added testsuites/testdata/dltests-broken-on-this-bsp.tcfg 2015-03-05T13:27:22 Is that obvious enough? :) 2015-03-05T13:28:31 Hesham: I think so, especially I don't think any other tools are using that file. 2015-03-05T13:31:14 Ok, set to a today commit c0ec50454e47e3d0f7d01be8fe2b557df9d48ca9 2015-03-05T13:31:17 emailed 2015-03-05T13:31:49 ok if the build works submit the patch 2015-03-05T13:32:00 i'm going to update all the rest of the bsets to point to newlib-git-1. 2015-03-05T13:32:31 ok.. should I wait or take a shower? :) 2015-03-05T13:33:15 no 2015-03-05T13:33:18 DrJoel: lots of projects are switching to Sphinx 2015-03-05T13:33:23 i'm building sparc/sis toolchain to test my change 2015-03-05T13:33:28 i'm going to be emailing out a migration proposal today some time 2015-03-05T13:33:42 GCI tasks imo. 2015-03-05T13:33:42 privatly to work out the kinks then publically 2015-03-05T13:34:43 gedare: no I should wait for the commit? Or take a shower? 2015-03-05T13:34:45 gedare: rtems-gcc-4.9-newlib-git-1.cfg does not set gcc_version, should I add it? 2015-03-05T13:37:48 DrJoel: shower! 2015-03-05T13:38:33 Hesham: yes or set it in moxie's bset 2015-03-05T13:39:20 hmm 2015-03-05T13:39:40 Hesham: i think that file is not quite right. 2015-03-05T13:40:15 i don't see any configs that get gcc from git.. 2015-03-05T13:40:51 Hesham: no you shoultn't 2015-03-05T13:41:13 I was going to modify this file with the correct settings 2015-03-05T13:41:14 Hesham: that file is getting called from files like rtems-gcc-4.9.2-newlib-git-1.cfg 2015-03-05T13:41:30 Oh 2015-03-05T13:41:41 i don't see any configs that will get a gcc git version 2015-03-05T13:41:48 it may be undesirable 2015-03-05T13:41:54 we do prefer to have a release.. 2015-03-05T13:42:29 looks li9ke we'll have to wait on gcc-5 2015-03-05T13:42:43 hmmm 2015-03-05T13:42:46 what a mess 2015-03-05T13:42:56 So, suspend the work on moxie for now? 2015-03-05T13:43:01 yeah... 2015-03-05T13:43:16 don't waste more time on it now 2015-03-05T13:43:23 Ok 2015-03-05T13:43:44 Any other tasks I can help with? 2015-03-05T13:45:41 [git] 7f8c066 by Gedare Bloom: [Switch to newlib-git-1 and bump version for all 4.9.2 tools] http://devel.rtems.org/changeset/7f8c066/rtems-source-builder 2015-03-05T13:45:57 DrJoel: pull and try to build all the tools now. 2015-03-05T13:46:07 expect moxie, h800 and m32c to fail i guess. 2015-03-05T13:46:42 Hesham: how well do you know x86 asm? :0 2015-03-05T13:47:06 Not to much 2015-03-05T13:47:36 The last time I had an eye on it was 4-5 years ago 2015-03-05T13:48:02 Gotta catch the bus, coming back withing half an hour 2015-03-05T13:48:03 Bye 2015-03-05T13:52:19 *** Hesham has quit IRC (Ping timeout: 245 seconds) 2015-03-05T13:52:33 *** Davidbrcz has joined #rtems 2015-03-05T14:14:01 hmm. there might be enough work in this PCI and drvmgr merging to justify a GSOC yet.. 2015-03-05T14:14:57 *** Hesham has joined #rtems 2015-03-05T14:15:53 Back.. 2015-03-05T14:19:08 Just started tool build ... hopefully can start BSPs in a while. 2015-03-05T14:19:26 Cleared all build and source directories so there will be some download time. :) 2015-03-05T14:21:30 25F here and still icy even where sun clears things the fastest. Still haven't seen a car pass in front of my house 2015-03-05T14:27:59 gedare, see the sb-report command in the RSB 2015-03-05T14:28:23 It can create a few types 2015-03-05T14:29:26 kiwichris as soon as I get new tools built, I will build all the BSPs again. I think I have fixed or worked around all the libdl failures 2015-03-05T14:29:51 I need to file a ticket on the lm32. It gets an ICE compiling libdl. We should at least track the GCC ticket. 2015-03-05T14:30:31 DrJoel_, ok 2015-03-05T14:30:46 gedare, what version did you bump in the RSB ? 2015-03-05T14:30:57 Just slow with lots of building. 2015-03-05T14:31:18 We need to review the server tickets. IRC logs need to be available as GSoC spins up. 2015-03-05T14:32:00 I do not want git for newlib any more. I get too many issues because the repo is not a good source 2015-03-05T14:32:33 DrJoel, what is the specific change we need ? 2015-03-05T14:33:01 Gedare's clocks_per_sec change. He can point you at it 2015-03-05T14:33:04 There might be another 2015-03-05T14:33:38 Can it exist as a patch ? Can we wait until the next snapshot ? 2015-03-05T14:34:15 *** Davidbrcz has quit IRC (Ping timeout: 264 seconds) 2015-03-05T14:34:21 I pushed Jeff for the snapshots because we all started to get failures with newlib git. 2015-03-05T14:34:26 You can't build any BSP and tests without this patch. Yes. It could be a patch 2015-03-05T14:34:38 Why ? 2015-03-05T14:34:45 What changed in the tests ? 2015-03-05T14:35:05 Gedare can explain it better. He ran it down 2015-03-05T14:35:55 But code with calls to clock() dont' compile. 2015-03-05T14:35:59 I appreciate the work being done to close out this issue but pushing hard this way make us now depend on the state of sourceware.org 2015-03-05T14:36:30 Yep. gedare .. can you turn that newlib into a snapshot plus patch? 2015-03-05T14:36:33 It compiles for me a few weeks ago. 2015-03-05T14:36:46 So something must have changed 2015-03-05T14:36:48 It did for me two days ago. :) 2015-03-05T14:37:07 gedare dropped off. Ask on devel@ for an explanation. 2015-03-05T14:37:13 And as we know sourceware.org is over loaded 2015-03-05T14:37:43 He is still here, so I assume he will drop by. It is not urgent. 2015-03-05T14:37:57 CLOCKS_PER_SEC was implemented using sysconf() and not returning the right value anyway per POSIX. You shouldn't have to include the .h for sysconf() to use clock(). I am guessing he fixed the code to drop the .h it shouldn't have needed. 2015-03-05T14:38:31 I am building tools against this. If it gets fixed in the next day, it likely won't impact anyone or just a few. 2015-03-05T14:40:33 Great. I suppose in hindsight upstream the newlib patch first then applying the RTEMS would have avoided this. Switching to git is a hammer and I was wanting to move the tools to a stable state. 2015-03-05T14:41:28 We all do. Stable tool definition is a requirement for 4.11. 2015-03-05T14:41:43 Jeff makes the snapshots around the middle of the month so we do not have long to wait. 2015-03-05T14:41:45 Do we have a ticket for the gdb situation? If not, can you file one. I think it is a blocker 2015-03-05T14:41:54 Except for the conversion to git next week. :) 2015-03-05T14:42:01 Huh ? 2015-03-05T14:42:11 See mailing list. Announced this morning. 2015-03-05T14:42:48 haha, great subject line 2015-03-05T14:46:30 it isn't evil as much as overdue.. :) 2015-03-05T14:47:50 hi kiwichris 2015-03-05T14:47:55 reading back 2015-03-05T14:48:31 kiwichris: yes i had that thought, but DrJoel_ really wanted to get a working toolchains and the bugfix for CLOCKS_PER_SEC broke them. 2015-03-05T14:48:42 kiwichris: yes i can provide an RSB patch 2015-03-05T14:48:52 rather, a newlib patch for RSB to apply against the previous snapshot 2015-03-05T14:50:39 ok, so what happened is that I pushed a newlib fix upstream, it was accepted, then I naively committed the RTEMS fix + test case 2015-03-05T14:50:43 I am not sure which way to move now we have the changes in. Are your newlib changes isolated and can be just address the patch ? 2015-03-05T14:50:51 however, the test case (rightly) is broken without the updated newlib. 2015-03-05T14:51:00 yes just one patch. easy 2015-03-05T14:51:14 i can just revert what I pushed 2015-03-05T14:51:18 I appreciate this. It is not about right or wrong, just about the management. 2015-03-05T14:51:26 and then add a patch to the snapshot 2015-03-05T14:51:31 yup 2015-03-05T14:51:58 The complicated is going to be the git conversion next week for newlib. If we are on a snapshot we have no issues. 2015-03-05T14:52:01 i expected this to be an interim solution anyway, to get us through the hump of closing out tickets and feeling good about releasing 4.11 2015-03-05T14:52:08 the tools i expect for 4.11 should use the next snapshot from Jeff. 2015-03-05T14:52:14 Understood 2015-03-05T14:52:24 or the start of the git. ;) 2015-03-05T14:52:40 yeah. the new git may differ from the sourceware one 2015-03-05T14:52:45 We need to plan around the snapshot being delayed due to Jeff needing to figure out a git procedure 2015-03-05T14:52:45 in terms of sha 2015-03-05T14:52:56 we should ask for a snapshot off the end of the cvs perhaps 2015-03-05T14:52:57 Yes 2015-03-05T14:53:12 Nice idea 2015-03-05T14:53:24 Want me to pipe up on the ml? or you do it privately? 2015-03-05T14:53:28 Please. 2015-03-05T14:53:31 okie 2015-03-05T14:54:03 so with that in mind, do you also want me to revert to the snapshot and provide a newlib patch in RSB? 2015-03-05T14:54:21 or expect to just switch to the next snapshot, which should be about equivalent to using the git mirror? 2015-03-05T14:55:00 DrJoel_: have you started building your tools with the newlib-git?" 2015-03-05T14:56:41 Yes. But you can change git. It won't impact me. Once I get the BSPs built, I will resolve the tickets. Then build tools and repeat BSP build 2015-03-05T14:57:13 ok waiting to hear what kiwichris will prefer for me to do. 2015-03-05T14:57:23 Sorry 2015-03-05T14:57:28 i have to go make a phone call for about 15 min. then i'll have about 30min I can work before I have to be done. 2015-03-05T14:57:31 I say wait. 2015-03-05T14:57:33 okay 2015-03-05T14:57:47 No need for more churn and extra work 2015-03-05T14:57:51 works for me. 2015-03-05T14:58:10 The issue is just about the performance of git over time from soureware.org 2015-03-05T14:58:25 So kill my tool build and wait for gedare to push this? 2015-03-05T14:58:35 t isn't a big deal. I am probably only on the third target 2015-03-05T14:58:47 DrJoel_: No 2015-03-05T14:58:53 DrJoel_: keep building your tools. i'm not going to change RSB 2015-03-05T14:59:09 DrJoel_: we're going to wait for the next snapshot and then update RSB to that for the official 4.11 tools 2015-03-05T14:59:10 I say leave the changes as they are and wait for a newlib snapshot 2015-03-05T14:59:26 *most likely official* :) 2015-03-05T14:59:30 brb. 2015-03-05T15:00:22 off to the office; back soon 2015-03-05T15:00:31 *** kiwichris has quit IRC () 2015-03-05T15:01:49 *** ragunath has quit IRC (Ping timeout: 246 seconds) 2015-03-05T15:13:20 *** ita has quit IRC (Remote host closed the connection) 2015-03-05T15:18:51 *** kiwichris has joined #rtems 2015-03-05T15:20:19 *** ita has joined #rtems 2015-03-05T15:30:33 hi kiwichris 2015-03-05T15:30:41 hi 2015-03-05T15:30:43 kiwichris: Hesham spent some time trying to get moxie tools to build, 2015-03-05T15:30:59 and it looks like we need to pick-up some recent patches from gcc (from this december) 2015-03-05T15:31:03 Excellent. What is needed ? 2015-03-05T15:31:25 any suggestions how to go about it? should we cherry-pick the patches and see if we can apply them to 4.9.2? 2015-03-05T15:31:30 Can we cherry pick from git ? 2015-03-05T15:31:53 If you can we can look at a way to get them out of git. 2015-03-05T15:31:55 probably. they seem to be isolated to moxie 2015-03-05T15:31:57 okay 2015-03-05T15:32:00 Thanks 2015-03-05T15:32:07 Upstream is always best 2015-03-05T15:32:11 yeah. 2015-03-05T15:32:30 antgreen changed the instruction set of moxie, and added/removed some opcodes, which horribly broke a few things. 2015-03-05T15:32:42 so we just need to find the right set of changes to get a working toolchain. 2015-03-05T15:33:42 *** ita has quit IRC (Quit: Lost terminal) 2015-03-05T15:39:45 well i don't have time today but i'll try to get back to this. i'll make a note on the ticket. Hesham if you have some time and want to give it a try, you can probably search for green's commits in gcc.git and try to pull them out as patches for RSB to use 2015-03-05T15:40:20 gedare: I was about to suggest this, I'll give it a try tomorrow. 2015-03-05T15:40:40 Where should I put the patch to be referenced from RSB? 2015-03-05T15:41:09 [trac] #2254/defect (new) updated by gedare (Hesham dug into this a little. It looks like we need to get some recent commits made ...) http://devel.rtems.org/ticket/2254 2015-03-05T15:41:59 Hesham: you can start by putting it in the rtems/patches directory, that will be searched first to resolve %patch directives 2015-03-05T15:42:34 and for the final solution i think the upstream patches are supposed to be added somewhere in rtems-tools, kiwichris ? 2015-03-05T15:42:46 Ok 2015-03-05T15:43:32 Hesham, if they are upstream and available you can reference them. The RSB supports patchworks, gitweb and may be should add git cherrypick. 2015-03-05T15:44:35 Otherwise for release we need them in RTEMS somewhere and rtems-tools/tools/4.11 is the place 2015-03-05T15:45:12 i think gcc has gitweb. 2015-03-05T15:45:13 kiwichris: Any special syntax for doing this trick? Or just a reference to the upstream patches? 2015-03-05T15:45:31 grep for patch in rtems/config 2015-03-05T15:45:37 Hmm... I haven't had a new bfin toolset buildand install since Aug.. hmm.. need to check into that when the build finishes 2015-03-05T15:45:46 Hesham, it depends on where they are and what they are held in. 2015-03-05T15:45:47 https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;a=commitdiff;h=d09fd72c630c4886367f1977cdb366aa82950e32? 2015-03-05T15:45:59 That is supported. 2015-03-05T15:46:01 https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;a=commitdiff;h=d09fd72c630c4886367f1977cdb366aa82950e32 loke this? 2015-03-05T15:46:05 like* 2015-03-05T15:46:06 Let me check, hang obn 2015-03-05T15:47:07 https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git&a=search&h=d09fd72c630c4886367f1977cdb366aa82950e32&st=commit&s=+Anthony+Green 2015-03-05T15:47:32 https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git&a=search&h=HEAD&st=commit&s=moxie 2015-03-05T15:47:40 i think you just want to pick up a few of the recent commits 2015-03-05T15:47:48 you might also consider asking anthony for guidance. 2015-03-05T15:48:31 Ok, I will do it tomorrow and come back to you if I was stuck 2015-03-05T15:48:52 yeah.. if they don't apply cleanly to 4.9.2 then it may be hard. 2015-03-05T15:48:56 bye 2015-03-05T15:49:29 Ok, Good night everyone. Bye 2015-03-05T15:49:35 nite! 2015-03-05T15:51:19 good nigh 2015-03-05T15:53:10 Hesham, still looking 2015-03-05T15:53:46 *** gedare has quit IRC (Ping timeout: 272 seconds) 2015-03-05T15:53:56 *** Hesham has quit IRC (Ping timeout: 244 seconds) 2015-03-05T15:59:30 https://git.rtems.org/rtems-source-builder/tree/source-builder/sb/download.py#n139 2015-03-05T15:59:46 This is the doco in the code for gitweb.cgi patches 2015-03-05T16:04:06 *** andrei_ has quit IRC (Remote host closed the connection) 2015-03-05T16:35:36 The bfin for gcc 4.9.2 build dies with an ICE: 2015-03-05T16:35:37 ./../../gcc-4.9.2/libgcc/fp-bit.c -fvisibility=hidden -DHIDE_EXPORTS 2015-03-05T16:35:37 0x9bdf11 reorder_loops 2015-03-05T16:35:37 ../../gcc-4.9.2/gcc/hw-doloop.c:500 2015-03-05T16:35:37 0x9bdf11 reorg_loops(bool, hw_doloop_hooks*) 2015-03-05T16:35:38 ../../gcc-4.9.2/gcc/hw-doloop.c:641 2015-03-05T16:35:40 0x93dce3 bfin_reorg_loops 2015-03-05T16:35:42 ../../gcc-4.9.2/gcc/config/bfin/bfin.c:3888 2015-03-05T16:35:44 0x93dce3 bfin_reorg 2015-03-05T16:35:46 ../../gcc-4.9.2/gcc/config/bfin/bfin.c:4630 2015-03-05T16:35:48 0x70e9f9 rest_of_handle_machine_reorg 2015-03-05T16:35:50 ../../gcc-4.9.2/gcc/reorg.c:3936 2015-03-05T16:35:52 0x70e9f9 execute 2015-03-05T16:35:54 ../../gcc-4.9.2/gcc/reorg.c:3966 2015-03-05T16:35:56 Please submit a full bug report, 2015-03-05T16:35:58 with preprocessed source if appropriate. 2015-03-05T16:36:14 bfin-rtems4.11-gcc (GCC) 5.0.0 20150120 (experimental) 2015-03-05T16:36:18 built 2015-03-05T16:36:55 When things settle after this build of BSPs, I will see if the 4.9 branch builds for it 2015-03-05T16:40:14 starting a build of gcc 4.9 branch now .. seem to be plenty of cycles 2015-03-05T16:40:59 kiwichris: Is it possible to switch a gcc target to a snapshot? bfin may require a random head to give us a toolset 2015-03-05T16:41:14 Yes 2015-03-05T16:41:23 Just set the version. 2015-03-05T16:41:45 OK. Once my build finishes and I have a guess, I will ask again for deails. 2015-03-05T16:42:07 First, make a gcc git config file then set the hash in the arch bset file. 2015-03-05T16:42:30 the gcc git config file lives in rtems/config/tools/ 2015-03-05T16:42:40 git from sourceware? gcc.gnu.org is the same machine I think 2015-03-05T16:42:58 I would gcc.gnu.org/ 2015-03-05T16:43:09 If there are issues with that then gcc should get it fixed. 2015-03-05T16:43:11 I mean from a lload view point 2015-03-05T16:43:17 I understand 2015-03-05T16:43:20 It is on their list 2015-03-05T16:43:32 We should always focus on the upstream which means gcc.gnu.org 2015-03-05T16:43:41 Simpler to report. 2015-03-05T17:13:10 [trac] #2282/defect (v:4.10) created by joel.sherrill (bfin tools as defined in RSB have GCC ICE during build) http://devel.rtems.org/ticket/2282 2015-03-05T17:15:21 [trac] #2283/defect (v:4.10) created by joel.sherrill (lm32 gcc ICE building libdl) http://devel.rtems.org/ticket/2283 2015-03-05T17:15:33 [trac] #2283/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2283 2015-03-05T17:15:47 [trac] #2282/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2282 2015-03-05T17:16:03 Sorry.. real issues that needed to be tracked. :( 2015-03-05T17:18:59 kiwichris FYI m32r warning: gdb-7.8.2.tar.xz: no hash found 2015-03-05T17:19:21 Ok. 2015-03-05T17:19:31 Just missing the hash line 2015-03-05T17:19:33 I can file a ticket if you prefer. :) 2015-03-05T17:25:29 [trac] #2282/defect (new) updated by joel.sherrill (The toolset I have installed reports this: bfin-rtems4.11-gcc --version ...) http://devel.rtems.org/ticket/2282 2015-03-05T18:31:22 *** jpc has joined #rtems 2015-03-05T18:33:55 hey jpc 2015-03-05T18:35:52 hello 2015-03-05T19:42:21 *** gedare has joined #rtems 2015-03-05T19:42:21 *** ChanServ sets mode: +o gedare 2015-03-05T19:52:52 *** gedare has quit IRC (Ping timeout: 264 seconds) 2015-03-05T19:54:28 welcome to rtems jpc 2015-03-05T20:02:12 thanks! 2015-03-05T20:02:41 sorry for the delayed response. I've been making dinner. 2015-03-05T20:07:29 so is there anybody here? 2015-03-05T20:23:39 Dinner time here also. Europeans asleep. North americans winding down. :) 2015-03-05T20:23:44 see everyone tomorrow 2015-03-05T21:22:09 *** gedare has joined #rtems 2015-03-05T21:22:09 *** ChanServ sets mode: +o gedare 2015-03-05T21:33:25 *** gedare has quit IRC (Ping timeout: 255 seconds) 2015-03-05T22:36:07 *** jpc has quit IRC (Ping timeout: 246 seconds) 2015-03-05T22:46:47 [trac] TBR/UserManual/Example_Application_Compiling edited by cartho (Changed build dir pc586 to pc386 for consistency) http://devel.rtems.org/wiki/TBR/UserManual/Example_Application_Compiling 2015-03-05T22:58:38 [trac] TBR/UserManual/Example_Application_Compiling edited by cartho (If cloned from repository shown at the top of the page, directory name is "examples-v2".) http://devel.rtems.org/wiki/TBR/UserManual/Example_Application_Compiling 2015-03-05T23:44:53 *** jpc has joined #rtems 2015-03-05T23:48:42 *** monstr has joined #rtems 2015-03-05T23:48:55 *** jpc has quit IRC (Ping timeout: 246 seconds) 2015-03-05T23:51:09 *** dr__house is now known as dr__house|away 2015-03-06T00:09:36 *** andrei_ has joined #rtems 2015-03-06T00:17:05 *** ita has joined #rtems 2015-03-06T00:26:54 *** zoso has joined #rtems 2015-03-06T00:43:21 *** sgulati has joined #rtems 2015-03-06T00:55:54 *** acenos has quit IRC (Ping timeout: 272 seconds) 2015-03-06T01:13:45 *** kiwichris has quit IRC () 2015-03-06T01:23:57 *** idleguy has joined #rtems 2015-03-06T01:27:50 Hello! 2015-03-06T01:28:05 Which mailinglist should I use for introducing myself to the community 2015-03-06T01:29:11 On one hand, there's the gsoc mailinglist 2015-03-06T01:30:35 which makes sense in this context. On the other, I assume not everyone is on the gsoc mailinglist, so devel makes more sense 2015-03-06T01:30:44 ? 2015-03-06T02:03:19 *** sgulati has quit IRC (Ping timeout: 246 seconds) 2015-03-06T02:09:08 *** Davidbrcz has joined #rtems 2015-03-06T02:24:53 *** Davidbrcz has quit IRC (Ping timeout: 256 seconds) 2015-03-06T02:38:21 *** idleguy has left #rtems 2015-03-06T03:51:03 *** Hesham has joined #rtems 2015-03-06T04:24:47 [git] 81b03c6 by Sebastian Huber: [score: Add and use PRIORITY_PSEUDO_ISR] http://devel.rtems.org/changeset/81b03c6/rtems 2015-03-06T04:24:47 [git] b8a5abf by Sebastian Huber: [score: Update _Thread_Heir only if necessary Previously, the _Thread_Heir was updated unconditionally in case a new heir was determined. The _Thread_Dispatch_necessary was only updated in case the executing thread was preemptible or an internal thread was unblocked. Change this to update the _Thread_Heir and 2015-03-06T04:24:47 _Thread_Dispatch_necessary only in case t...] http://devel.rtems.org/changeset/b8a5abf/rtems 2015-03-06T04:24:47 [git] 6157743 by Sebastian Huber: [score: Simplify and fix signal delivery Deliver the POSIX signals after the thread state was updated to avoid race-conditions on SMP configurations. Update #2273.] http://devel.rtems.org/changeset/6157743/rtems 2015-03-06T04:24:47 [git] 222dc77 by Sebastian Huber: [score: Add and use _Thread_Do_dispatch() The _Thread_Dispatch() function is quite complex and the time to set up and tear down the stack frame is significant. Split this function into two parts. The complex part is now in _Thread_Do_dispatch(). Call _Thread_Do_dispatch() in _Thread_Enable_dispatch() only if necessary. This 2015-03-06T04:24:47 increases the average cas...] http://devel.rtems.org/changeset/222dc77/rtems 2015-03-06T04:24:47 [git] 1512761 by Sebastian Huber: [score: Thread dispatch dis/enable without Giant Update #2273.] http://devel.rtems.org/changeset/1512761/rtems 2015-03-06T04:24:48 [git] 43e1573 by Sebastian Huber: [score: Add ISR lock to Objects_Control This enables per-object SMP locks on SMP configurations and is the first step to support fine-grained locking. On uni-processor configuration there will be no overhead. The _Objects_Acquire() is intended to replace _Objects_Get_isr_disable(). Update #2273.] 2015-03-06T04:24:48 http://devel.rtems.org/changeset/43e1573/rtems 2015-03-06T04:24:49 [git] 5b393fa by Sebastian Huber: [score: Add thread acquire Update #2273.] http://devel.rtems.org/changeset/5b393fa/rtems 2015-03-06T04:24:49 [git] 4c8a0ac by Sebastian Huber: [score: Add thread wait flags Update #2273.] http://devel.rtems.org/changeset/4c8a0ac/rtems 2015-03-06T04:24:50 [git] 7d6e94b by Sebastian Huber: [score: Implement fine-grained locking for events Use the ISR lock of the thread object to protect the event state and use the Giant lock only for the blocking operations. Update #2273.] http://devel.rtems.org/changeset/7d6e94b/rtems 2015-03-06T04:31:23 [trac] #2273/enhancement (accepted) updated by sebastian.huber ([7d6e94b12ab94f13d3e769702d50f91be7d5884b/rtems] ...) http://devel.rtems.org/ticket/2273 2015-03-06T04:54:19 *** antgreen has joined #rtems 2015-03-06T05:18:52 *** zoso has quit IRC (Ping timeout: 240 seconds) 2015-03-06T05:48:12 Any idea how to download a patch from gitweb? https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;a=commitdiff;h=a67a65b4c300ee2b771c35e3aeaf65f7822136ad 2015-03-06T06:32:31 *** sgulati has joined #rtems 2015-03-06T06:48:56 *** gedare has joined #rtems 2015-03-06T06:48:56 *** ChanServ sets mode: +o gedare 2015-03-06T06:51:22 *** sgulati has quit IRC (Ping timeout: 246 seconds) 2015-03-06T07:04:01 [git] da06fe9 by Sebastian Huber: [score: Fix SMP lock implementation See also: https://nahratzah.wordpress.com/2012/10/12/a-trivial-fair-spinlock/ http://concurrencyfreaks.blogspot.de/2014/05/relaxed-atomics-optimizations-for.html] http://devel.rtems.org/changeset/da06fe9/rtems 2015-03-06T07:18:37 Hesham: https://git.rtems.org/rtems-source-builder/tree/source-builder/sb/download.py#n139 2015-03-06T07:18:42 Hesham: This is the doco in the code for gitweb.cgi patches 2015-03-06T07:19:44 verm__: Thanks, I'll try it. 2015-03-06T07:26:13 *** kiran4399 has joined #rtems 2015-03-06T07:33:27 andrei_ either users or devel. Doesn't matter at this point. 2015-03-06T07:33:43 verm__ I am looking at the documentation build issue 2015-03-06T07:34:00 * DrJoel_ says .. another day iced in. Should melt today. 2015-03-06T07:42:46 DrJoel: great thanks 2015-03-06T07:42:58 another day? not bad 2015-03-06T07:43:13 [git] c096dc1 by Joel Sherrill: [doc: Update Ada User Guide to match C] http://devel.rtems.org/changeset/c096dc1/rtems 2015-03-06T07:44:20 Our deck has 2 inches of what looks like snow but is so solid, walking on it doesn't leave footprints. 2015-03-06T07:44:57 verm__ docs build now. Someone added to the C Users manual and didn't update the master doc for the Ada. Hinting at the buildbot. :) 2015-03-06T07:45:04 I worry doxygen isn't any better 2015-03-06T07:45:44 if you are turning on regular doc builds, please look at the list of tools configure probes for. It needs some likely not installed to generate the figures in the manua 2015-03-06T07:52:13 doxygen built fine 2015-03-06T07:52:37 i don't want to put the old docs under buildbot 2015-03-06T07:52:44 it's monolithic and takes ages 2015-03-06T08:00:46 i might do it anyway just to have everything under buildbot 2015-03-06T08:00:55 *** rtems-bot has joined #rtems 2015-03-06T08:01:00 Ages is an exagerration 2015-03-06T08:01:01 real 0m13.063s 2015-03-06T08:01:02 user 0m5.549s 2015-03-06T08:01:02 sys 0m3.671s 2015-03-06T08:01:22 *** andrei_ has quit IRC (Ping timeout: 246 seconds) 2015-03-06T08:01:24 That's a bootstrap of doc after the clone and building it 2015-03-06T08:01:35 well, the sphinx docs takes less than a min for all of them including pdf 2015-03-06T08:01:41 so it's entirely relative 2015-03-06T08:01:49 *** rtems-bot_ has quit IRC (Ping timeout: 245 seconds) 2015-03-06T08:02:00 but the biggest issue is the fact that it's monolithic.. i know it could be splitup but that means a bootstrap for each one 2015-03-06T08:04:21 echo one being what? 2015-03-06T08:04:32 You only have to bootstrap the doc/ directory for docs 2015-03-06T08:07:24 [git] be2e60c by Sebastian Huber: [Delete unused files] http://devel.rtems.org/changeset/be2e60c/rtems-libbsd 2015-03-06T08:07:24 [git] f59a2da by Sebastian Huber: [Use C11 and C++11 This is necessary to use and .] http://devel.rtems.org/changeset/f59a2da/rtems-libbsd 2015-03-06T08:07:24 [git] 19adb93 by Sebastian Huber: [rtems-bsd-cxx.cc: Fix include order] http://devel.rtems.org/changeset/19adb93/rtems-libbsd 2015-03-06T08:07:24 [git] dd9e4ef by Sebastian Huber: [Mark header files as empty] http://devel.rtems.org/changeset/dd9e4ef/rtems-libbsd 2015-03-06T08:07:24 [git] e684326 by Sebastian Huber: [Change license to BSD] http://devel.rtems.org/changeset/e684326/rtems-libbsd 2015-03-06T08:07:25 [git] 9a50734 by Sebastian Huber: [atomic.h: Format changes] http://devel.rtems.org/changeset/9a50734/rtems-libbsd 2015-03-06T08:07:25 [git] e5a0175 by Sebastian Huber: [atomic.h: Use or for SMP] http://devel.rtems.org/changeset/e5a0175/rtems-libbsd 2015-03-06T08:07:26 [git] 8ed1b18 by Sebastian Huber: [rtems_bsd_mutex: SMP support via ISR locks] http://devel.rtems.org/changeset/8ed1b18/rtems-libbsd 2015-03-06T08:07:26 [git] e5db084 by Sebastian Huber: [ZONE(9): Enable per-processor cache for SMP This prevents a potential deadlock via the Giant lock and is a performance benefit.] http://devel.rtems.org/changeset/e5db084/rtems-libbsd 2015-03-06T08:07:27 [git] 3cafb9a by Sebastian Huber: [tests: Add PROFREPORT shell command] http://devel.rtems.org/changeset/3cafb9a/rtems-libbsd 2015-03-06T08:13:21 Hesham: have you pulled the patches out yourself and attempted to apply them to gcc-4.9.2? 2015-03-06T08:13:22 Patching gcc-4.9.2 with gitweb moxie patches failed to apply :/ I think it would be needed to by manually-generated. 2015-03-06T08:13:30 yeah that's what i wondered. 2015-03-06T08:13:57 you might consider asking antgreen if he can create a patch against 4.9.2 for the relevant changes 2015-03-06T08:14:09 By e-mail? 2015-03-06T08:14:37 otherwise i think we either try to create some patches ourselves, or release without a working toolchain for moxie until later. 2015-03-06T08:15:23 I would try to generate patches manually for now, if it was too difficult, I'll tell you 2015-03-06T08:15:28 Okay. 2015-03-06T08:16:42 DrJoel: not for doxygen, though 2015-03-06T08:17:47 Doxygen does take a while. :) 2015-03-06T08:19:20 anyway i'll eventually set it up 2015-03-06T08:19:37 for now in order to get 4.11 out i wanted to make sure i could take a 'snapshot' of the docs easily 2015-03-06T08:21:11 I worked on a project that took ~1 hour to do Doxygen without figures and over 4 to generate the figures. Only ran it at night either way 2015-03-06T08:23:48 DrJoel_: how do you feel about releasing with the handful of broken toolchains? 2015-03-06T08:24:29 rephrase: no one is going to step forward to fix them, so we should just release ;) 2015-03-06T08:26:56 DrJoel: crazy 2015-03-06T08:27:47 i think it sucks to release with any broken toolchains, however i think it's worse to delay a release because of it 2015-03-06T08:28:03 we can always fix them in a point release, it's more important to have the 4.11 'checkpoint' 2015-03-06T08:28:24 and then regularly have point releases every couple of weeks once things really slow down we can move to 4.12 2015-03-06T08:29:11 If we are talking about moxie, that's OK. Are there others? 2015-03-06T08:29:50 real 14m29.518s 2015-03-06T08:29:58 not including bootstrap :( 2015-03-06T08:30:50 [trac] #2284/defect (v:4.11) created by joel.sherrill (h8300 gets error linking dl0* tests) http://devel.rtems.org/ticket/2284 2015-03-06T08:31:58 There was discussion about parallelizing the graphics generation. I wonder if they ever got that into a released version of Doxygen. It would help us 2015-03-06T08:32:21 But 15 minutes twice a day isn't a huge deal. It is just machine time. Use it or lose it. 2015-03-06T08:32:22 we've mogint away from doxygen for building the docs 2015-03-06T08:32:25 it'll be sphinx 2015-03-06T08:32:30 doxygen -> xml -> sphinx 2015-03-06T08:32:38 sphinx handles all the image generation 2015-03-06T08:32:42 way, way faster 2015-03-06T08:32:45 And how long will that take? We have to have current documentation until then that is up to date. 2015-03-06T08:32:48 and we also get pretty docs 2015-03-06T08:33:05 I am not arguing about goal. But 0 today isn't acceptable 2015-03-06T08:33:06 i already have it working but it's not slated until the waf build comes out 2015-03-06T08:33:17 oh, i know this is for the future 2015-03-06T08:33:22 i'm not touching anything for 4.11 2015-03-06T08:35:39 strange, 'Doxyfile' isn't being created 2015-03-06T08:35:44 We just need the docs as we have them being available 2015-03-06T08:35:58 they are available, always have been just outdated 2015-03-06T08:36:23 Only since the move. They were generated twice a day before that 2015-03-06T08:37:02 correction, they're not outdated anymore 2015-03-06T08:37:08 on a positive note, I think I may have work arounds for all architectures/BSPs except moxie. 2015-03-06T08:37:35 I had about 20 BSPs which built far enough to find out they didn't have enough memory to link jffs2_fsscandir01 which is new. 2015-03-06T08:38:13 Starting another build sweep. If that goes well, I will have a patch series to post later today. 2015-03-06T08:38:20 nice 2015-03-06T08:38:29 https://docs.rtems.org/rsb/ 2015-03-06T08:38:33 And when merged, all the libdl tickets can move to 4.11.1 or close 2015-03-06T08:38:49 Once I start the build, I need to get back to work on a deadline. :( 2015-03-06T08:40:55 *** ita has quit IRC (Remote host closed the connection) 2015-03-06T08:42:04 DrJoel_: DrJoel: bfin and lm32 according to your tickets filed last night. 2015-03-06T08:43:27 lm32 should be libdl tripping a GCC ICE and if luck holds, disabling libdl addresses that 2015-03-06T08:43:57 If you read the bfin ticket, we need to drop back to gcc 4.8.x. I put the hash of the last bfin toolset I have. If you can drop it back to that, then that can be resolved 2015-03-06T08:45:08 oh. that's easy there is already a 4.8 config availabble 2015-03-06T08:45:38 i can push it to RSB now if you want... using tools/rtems-gcc-4.8.3-newlib-git-1 2015-03-06T08:45:47 DrJoel_: ^ 2015-03-06T08:45:49 Please do 2015-03-06T08:45:51 ok. 2015-03-06T08:46:14 What's the git commit to write a 00 patch toput an explanation in. I will go ahead and send the patches for review even with testing going on. It is Friday 2015-03-06T08:46:37 [git] dbb78dc by Gedare Bloom: [4.11/bfin: roll back to gcc 4.8] http://devel.rtems.org/changeset/dbb78dc/rtems-source-builder 2015-03-06T08:46:40 ok DrJoel_ pull RSB and try bfin 2015-03-06T08:46:50 --cover-letter? 2015-03-06T08:46:58 it's an argument to git-send-email 2015-03-06T08:47:07 Ahh... thanks 2015-03-06T08:47:20 rather 2015-03-06T08:47:24 git format-patch --cover-letter 2015-03-06T08:47:36 and then you can modify the 0000- file 2015-03-06T08:47:44 and then git send-email *.patch 2015-03-06T08:48:12 Yep.. just have only used that once or twice. It is good when the series is a bit large or odd 2015-03-06T08:48:33 sure... 2015-03-06T08:48:36 what are yous ubmitting? 2015-03-06T08:54:34 oh, and apparently tools won't build on mingw. 2015-03-06T08:54:36 The series of patches that has almost everything building now. Workarounds for libdl issues, some fixes, more tests to skip on some BSPs 2015-03-06T08:54:45 That's a kiwichris issue :) 2015-03-06T08:54:50 He is workign on it I think 2015-03-06T08:54:52 yeah i think he was 2015-03-06T08:55:13 ok let's release! 2015-03-06T08:55:46 it's as easy as... git checkout -b 4.11 ; git push upstream 4.11 2015-03-06T08:55:46 ;-) 2015-03-06T08:56:03 This round of patches should move us forward. Let's see what is still in the queue. 2015-03-06T08:56:12 *nods* 2015-03-06T08:56:13 Chris may not have filed an issue with gdb patches 2015-03-06T08:56:32 if it is not a problem in rtems.git, we can still cut the branch 2015-03-06T08:56:49 many of the remaining issues are external to the repo itself 2015-03-06T08:57:12 Agreed. I just want to make sure we have all issues in Trac and make a branching decision based on a known state. 2015-03-06T08:57:17 yeah. 2015-03-06T08:58:03 Patches coming. Make sure I used the right Trac annotation. :) 2015-03-06T09:09:45 i don't know if trac will pick up the number without a # or not, verm__ ? 2015-03-06T09:09:55 closes 2247. 2015-03-06T09:09:59 shouold tayht work inn a commit msg? 2015-03-06T09:10:02 it should 2015-03-06T09:10:04 ok 2015-03-06T09:10:24 # is nice as it will highlight on the trac timeline as it's wiki syntax 2015-03-06T09:10:41 I think all the ones marked "updates" mean that the milestone moves to 4.11.1 or beyond. 2015-03-06T09:10:52 gotcha. now, is there any way to also issue commands like "Milestone = 4.11.1" from git-commit? 2015-03-06T09:11:07 just curious verm__ :) 2015-03-06T09:11:14 not asking for an implementation 2015-03-06T09:12:52 SAT word in a ml post! 2015-03-06T09:13:16 which one? 2015-03-06T09:13:59 Remember a PhD is better than an STD because penicillen can't take your PhD away 2015-03-06T09:14:56 conjoin 2015-03-06T09:15:55 :) 2015-03-06T09:20:15 [git] bf9d387 by Sebastian Huber: [score: Allow NULL for SMP lock name] http://devel.rtems.org/changeset/bf9d387/rtems 2015-03-06T09:20:15 [git] 2bd71d9 by Sebastian Huber: [shell: Add PROFREPORT command] http://devel.rtems.org/changeset/2bd71d9/rtems 2015-03-06T09:21:18 I applied about 4 moxie patches to gcc-4.9.2 but the remaining can't be applying cleanly, and needs a lot of time to re-generate manually. So I think I'll have to ask antgreen about a gcc-4.9.2 patch 2015-03-06T09:22:12 antgreen: you around? 2015-03-06T09:22:39 gedare: milestone:4.1.1 will highlight it in the timeline 2015-03-06T09:22:49 all the wiki commands work in commit messages 2015-03-06T09:25:03 [git] 33b72fd by Alexander Krutwig: [testsupport: Add cascade option to parallel test] http://devel.rtems.org/changeset/33b72fd/rtems 2015-03-06T09:25:06 verm__ just on a line by itself? 2015-03-06T09:25:19 for what? 2015-03-06T09:25:33 ah iwas wondering about updating the milestone field directly 2015-03-06T09:25:41 *** sebhub has quit IRC (Quit: Konversation terminated!) 2015-03-06T09:25:43 oh 2015-03-06T09:25:43 since we'll have to bump the milestone for his patches 2015-03-06T09:25:45 that milestone command. Will it move it without me touching the web page 2015-03-06T09:26:01 i wrote something to do that a long time ago but this was before trac integrated the ticket updater 2015-03-06T09:26:05 i can look at doing it again 2015-03-06T09:26:13 but right now it won't automatically update the milestone 2015-03-06T09:26:27 OK. I don't mind doing it. Just asking 2015-03-06T09:27:11 yeah don't waste effort on it we don't do it often enough i would guess. 2015-03-06T09:27:45 I don't think it will happen much either. 2015-03-06T09:28:15 Normally we fix or just move out with no commit. This is rather unusual. I can't remember every committing a work around before 2015-03-06T09:28:57 other than the one comment i made, patches look fine.. 2015-03-06T09:29:09 but that comment applies to all the patches that touched configure.ac 2015-03-06T09:29:18 for disabling libdl 2015-03-06T09:30:26 *** zoso has joined #rtems 2015-03-06T09:31:49 Replied. Very deliberate decision to do it that way 2015-03-06T09:34:46 *** Davidbrcz has joined #rtems 2015-03-06T09:35:17 *** antgreen has quit IRC (Ping timeout: 246 seconds) 2015-03-06T09:35:23 ok idc 2015-03-06T09:35:29 commit 'em 2015-03-06T09:36:11 :) We may want to discuss h8300 being on the "won't fix" list and closing that one. But the others are possible at some point. 2015-03-06T09:36:21 I will wait until the build finishes. 2015-03-06T09:36:24 ok 2015-03-06T09:36:27 great 2015-03-06T09:45:35 ./docs.sh: line 18: 16225 Abort trap (core dumped) doxygen > ${d}/doxy.log 2>&1 2015-03-06T09:45:37 that's always good 2015-03-06T09:45:57 Doxygen dumped? 2015-03-06T09:46:49 yeah 2015-03-06T09:47:58 let me try it here and see how it goes 2015-03-06T09:48:12 i think the problem is with pango 2015-03-06T09:48:24 what's that? 2015-03-06T09:48:29 gui package 2015-03-06T09:48:36 ahhh 2015-03-06T09:48:52 rebuilding to see if it's fixed 2015-03-06T09:49:06 i'm going to bump the irc ticket to 4.11.1 2015-03-06T09:49:17 the bot is still logging the channel always has been 2015-03-06T09:49:38 Are the logs at OSU or on the old machine? 2015-03-06T09:49:44 old machine 2015-03-06T09:50:26 [trac] #2200/defect (new) updated by amar (Bumping this. The old machine has been logging all this time so we have not lost ...) http://devel.rtems.org/ticket/2200 2015-03-06T09:50:27 Jeff could open the firewall and Apache on it enough to let the IRC logs go through and then we just have to point .org to them 2015-03-06T09:50:45 i don't want to do that 2015-03-06T09:50:54 it's just as much effort to move it to its perminant location 2015-03-06T09:51:08 all the old web services have been shut off / removed 2015-03-06T09:51:31 Just offering 2015-03-06T09:51:39 i know :) 2015-03-06T09:54:25 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-06T09:55:00 Fatal error 'unlock error' at line 341 in file /usr/src/lib/libthr/thread/thr_umtx.c (errno = 0) 2015-03-06T09:55:01 oh, nice 2015-03-06T09:55:26 this drvmgr code is really bloated stuff. 2015-03-06T09:55:30 eww. 2015-03-06T09:55:51 always nice to get a threading errow hen generating documentation 2015-03-06T09:56:50 at least it consistantly breaks at the same spot 2015-03-06T09:57:02 Generating docs for compound Thread_Action_control... 2015-03-06T09:57:25 it seems to have worked ok though 2015-03-06T09:57:27 It generated on my VM OK and fairly quickly. 2015-03-06T09:57:28 real 1m21.995s 2015-03-06T09:57:28 user 0m53.091s 2015-03-06T09:57:28 sys 0m16.802s 2015-03-06T09:57:50 Chrome is chugging loading the index.html though. :( 2015-03-06T09:58:12 https://docs.rtems.org/doxygen/cpukit/html/ 2015-03-06T09:58:48 hmmm maybe it didn't work ok 2015-03-06T09:58:54 no figures in mine though. 2015-03-06T09:59:53 does my url seem fine? 2015-03-06T10:00:03 it's 94MB where the old one was 180MB 2015-03-06T10:00:14 the build seems to have finished OK other than that one threading erroor 2015-03-06T10:00:22 oh i know why it's so huge 2015-03-06T10:00:45 or not 2015-03-06T10:01:19 i guess the HTML styling is different 2015-03-06T10:01:25 https://docs.rtems.org/doxygen/cpukit/html/structPOSIX__Message__queue__Control.html 2015-03-06T10:01:27 no figures 2015-03-06T10:01:38 You need graphviz installed .. I was missing it on my VM 2015-03-06T10:02:11 FWIW warnings.log has ~7800 lines. Some won't get fixed like undocumented elements in third party code but others are formatting issues in code we own 2015-03-06T10:02:12 it's installed 2015-03-06T10:02:32 yeah i'm missing that graph, hmm 2015-03-06T10:02:36 *** Davidbrcz has quit IRC (Ping timeout: 256 seconds) 2015-03-06T10:02:39 that page I linked should have a graph 2015-03-06T10:02:51 i know 2015-03-06T10:02:59 hmm 2015-03-06T10:03:21 The directories above need to be visible because that's where my script put the Doxygen build logs. 2015-03-06T10:03:42 [joel@localhost doxygen]$ pwd 2015-03-06T10:03:42 [joel@localhost doxygen]$ ls 2015-03-06T10:03:42 cpukit doxy.log warnings.log 2015-03-06T10:08:09 i only used parts of your script 2015-03-06T10:08:27 The logs still need to be publicly accessible 2015-03-06T10:09:01 why? 2015-03-06T10:09:25 So everyone can see what to fix. Not everyone builds Doxygen but a lot of people edit it. 2015-03-06T10:09:33 oh 2015-03-06T10:09:35 ok 2015-03-06T10:12:43 what do you want doxy.log? 2015-03-06T10:13:10 where? directory above output is OK. 2015-03-06T10:13:12 ok it's fixed 2015-03-06T10:13:15 https://docs.rtems.org/doxygen/cpukit/html/structPOSIX__Message__queue__Control.html 2015-03-06T10:13:24 But doxygen has a warning and run log so there should be two files 2015-03-06T10:13:31 https://docs.rtems.org/doxygen/doxy.log 2015-03-06T10:13:34 https://docs.rtems.org/doxygen/warnings.log 2015-03-06T10:15:25 * @param[in] src_size The size of the string in bytes (inludes '\0' if any). 2015-03-06T10:15:39 produces this warning 2015-03-06T10:15:56 for the "\0" 2015-03-06T10:15:58 [trac] #2198/infra (new) updated by amar (Docs have been built but are not updating automatically. I will re-generate this ...) http://devel.rtems.org/ticket/2198 2015-03-06T10:16:01 [trac] #2199/defect (new) updated by amar (Docs have been built but are not updating automatically. I will re-generate this ...) http://devel.rtems.org/ticket/2199 2015-03-06T10:16:48 [trac] Developer/Release edited by amar (Add list of docs to be generated.) http://devel.rtems.org/wiki/Developer/Release 2015-03-06T10:17:09 grrr.. that needs to be "\\0" 2015-03-06T10:18:24 in the source? 2015-03-06T10:18:29 oh you mean in the comment? 2015-03-06T10:20:19 *** zoso has quit IRC (Ping timeout: 256 seconds) 2015-03-06T10:20:20 comment 2015-03-06T10:22:55 *** zoso has joined #rtems 2015-03-06T10:28:28 *** zoso has quit IRC (Remote host closed the connection) 2015-03-06T10:38:20 *** idleguy has joined #rtems 2015-03-06T10:39:46 [git] cc69334 by Joel Sherrill: [Fix a number of minor Doxygen formatting issues] http://devel.rtems.org/changeset/cc69334/rtems 2015-03-06T10:42:12 well.. only about 200 real Doxygen issues 2015-03-06T10:42:27 and ~7000 places it complains we need documentation. Many are third party so don't cares. 2015-03-06T10:55:00 DrJoel_: how's the re-build going? 2015-03-06T10:56:09 I have 48 BSP log files and no errors so far. Probably 40 completed so far OK. 2015-03-06T11:02:11 *** ita has joined #rtems 2015-03-06T11:02:35 55 log files now and no errors! 2015-03-06T11:03:11 building a new bfin toolset. I disabled psxclock01 because something weird happened building m68k and mips. Got a permission erorr while installing rtems-tools. after building all the other architectures 2015-03-06T11:06:31 hm. 2015-03-06T11:06:39 maybe a bad prefix. 2015-03-06T11:07:54 Same script driving the architectures in a loop. It appears to be on the version.cfg file. 2015-03-06T11:08:25 57 logs and moxie is only failure so far 2015-03-06T11:16:24 *** kiran4399 has joined #rtems 2015-03-06T11:16:34 oh, about deprecating notepadsd, we could probably generate a warning in confdefs.h 2015-03-06T11:16:57 That's a good piece 2015-03-06T11:17:14 yeah and then turn it into an error later. 2015-03-06T11:18:18 https://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html search for deprecated attribute. That's what I was trying to remember 2015-03-06T11:19:18 rtems_clock_get should have that attribute also. I think there is only one test now that uses it. I tried to clean things up to get ready to obsolete it. 2015-03-06T11:19:20 hmmm.. confdefs.h treats notepads weirdly. they are enabled by default unless the user definesCONFIGURE_DISABLE_CLASSIC_API_NOTEPADS 2015-03-06T11:19:53 That may not have even been present in 4.10. Check. We have latitude 2015-03-06T11:20:40 @gedare: So I was checking out the BSP source code for RTEMS and how to write the peripheral support for Beaglebone black... 2015-03-06T11:20:40 kiran4399: Error: "gedare:" is not a valid command. 2015-03-06T11:21:04 beng-nl ping.. beagle 2015-03-06T11:21:51 beng-nl: ping 2015-03-06T11:23:13 ok kiran4399 do you have a qeustion? 2015-03-06T11:24:08 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2015-03-06T11:24:15 yes.. so for writing the bsp pheripheral support for beaglebone black what can I do on qemu ? 2015-03-06T11:24:16 DrJoel_: it is the same in 4.10. notepads are enabled unless disabled. 2015-03-06T11:24:45 We could switch the sense. Make them specify enabled and then deprecate them so they get a warning when linking 2015-03-06T11:24:47 depends on what devices qemu can model for it. i do not know myself. 2015-03-06T11:25:50 Well, qemu is a simulator.. So I don't think simular can show the results. 2015-03-06T11:26:03 I'll be getting my cable in next 4 days.. 2015-03-06T11:26:18 So until then what can I do so that I can write a good proposal ? 2015-03-06T11:26:42 of what? qemu is great for testing a lot of things.. uarts, nics, video, someone tested canbus with it. If supported, you can get pretty far 2015-03-06T11:26:51 get bb running under qemu. write your ideas. we don't expect you to be writing code for a good proposal. 2015-03-06T11:27:08 I would expect a BB NIC could be brought up on qemu. beng-nl? 2015-03-06T11:27:11 you should have a good mental model of what you are going to be doing. 2015-03-06T11:27:18 and be able to write about it 2015-03-06T11:27:22 without writing a single line of code 2015-03-06T11:27:54 See the application template we provide and use it. Ultimately, you want to define mid-term and final goals you will be evaluated against. Along witht he steps to get there 2015-03-06T11:29:40 Does RTEMS boot on Beaglebone black ? 2015-03-06T11:30:54 *** monstr has quit IRC (Read error: Connection reset by peer) 2015-03-06T11:31:54 *** sujayraaj has joined #rtems 2015-03-06T11:33:22 kiran4399: https://devel.rtems.org/wiki/TBR/BSP/Beagleboard 2015-03-06T11:34:57 *** antgreen has joined #rtems 2015-03-06T11:39:48 kiran4399: ach 2015-03-06T11:39:50 kiran4399: ack 2015-03-06T11:40:39 kiran4399, DrJoel_, gedare: qemu only emulates the bbxm 2015-03-06T11:41:56 Does it have value for adding peripheral support which would work on black? 2015-03-06T11:43:04 *** ragunath has joined #rtems 2015-03-06T11:43:23 DrJoel: when i put doxygen under buildbot i'll have it error out on warnings and blacklist any 3rd party packages 2015-03-06T11:43:32 (for errors/warnings) 2015-03-06T11:44:21 There are 7200 warnings now. Many will never get fixed because they are in third party code which doesn't even have markup 2015-03-06T11:44:38 So erroring out isn't an option. That's the point of the log 2015-03-06T11:44:45 DrJoel_: he said he'd ignore 3rd party code. 2015-03-06T11:45:02 only error out on warnings in rtems doxygen 2015-03-06T11:45:12 if we get warning-free, this means all commits have to pass doxygen also. 2015-03-06T11:45:14 I am referring to third party code in rtems 2015-03-06T11:45:23 beng-nl: I am little confused about the support of rtems towards beaglebone-black. Not beagleboard or beaglebone white... I mean is rtems compatable with Beaglebone black ? 2015-03-06T11:45:24 zlib for example. Network stack 2015-03-06T11:45:27 gedare: yep that's the point 2015-03-06T11:45:32 md5, sha, crypt 2015-03-06T11:45:38 DrJoel: i know what you mean 2015-03-06T11:45:40 DrJoel_: i sent a patch to devel with notepad change. 2015-03-06T11:45:54 ok.. will try to look at it son 2015-03-06T11:45:55 soon 2015-03-06T11:45:56 i will blacklist any vendor code 2015-03-06T11:46:08 sure. i haven't built it.. 2015-03-06T11:46:28 I am still over 100 BSPs from finishing this sweep though 2015-03-06T11:46:31 hmm, also will have to update tests 2015-03-06T11:47:40 *** idleguy has left #rtems 2015-03-06T11:55:52 *** kiran4399 has quit IRC (Ping timeout: 246 seconds) 2015-03-06T11:56:02 73 log files and only moxie fails :) 2015-03-06T11:56:32 Only 120 more to go. There are now 193 BSP variants 2015-03-06T12:17:45 *** antgreen has quit IRC (Ping timeout: 250 seconds) 2015-03-06T12:33:27 193 now?! 2015-03-06T12:35:22 Do either of you see what's wrong with cpukit/libdl/rtl.h from a Doxygen viewpoint? There is something not right. See the log .. 2015-03-06T12:36:22 url? 2015-03-06T12:36:39 posted to mailing list.. should also be in your log.. hold on 2015-03-06T12:37:22 What's the URL for where the warnings log is on the rtems.org server 2015-03-06T12:37:52 https://lists.rtems.org/pipermail/devel/2015-March/010223.html has the cut down 2015-03-06T12:37:57 look for rtl.h 2015-03-06T12:38:10 I don't see the mismatched group in i2c.h either :( 2015-03-06T12:39:00 i posted those above earlier: https://docs.rtems.org/doxygen/doxy.log https://docs.rtems.org/doxygen/warnings.log 2015-03-06T12:41:00 ok.. the directory isn't browseable so you have to know the magic name 2015-03-06T12:41:12 yeah none of the directories are browsable 2015-03-06T12:41:21 we can link to it from the main site if we want to point to the file 2015-03-06T12:41:49 search on the page for rtl.h. I have fixed a few other problems 2015-03-06T12:42:16 I have a patch pending to fix all "unknown command" warnings 2015-03-06T12:44:26 up to 81 and still only the moxie :) 2015-03-06T12:52:15 *** Davidbrcz has joined #rtems 2015-03-06T13:26:10 *** ragunath has quit IRC (Ping timeout: 246 seconds) 2015-03-06T13:26:30 [git] e0eb07a by Joel Sherrill: [Fix more Doxygen typos] http://devel.rtems.org/changeset/e0eb07a/rtems 2015-03-06T13:34:01 nibbling at the obvious Doxygen issues 2015-03-06T13:34:10 nice 2015-03-06T13:34:46 93 logs and still only moxie fails which we know about 2015-03-06T13:35:12 We can pick on Chris.. rfs and libdl seem to have the most. :) 2015-03-06T13:36:45 hopefully anthony can chime up and give us a good toolset to use 2015-03-06T13:36:52 chime-in 2015-03-06T13:45:19 [trac] #2285/defect (v:4.11) created by joel.sherrill (_CORE_mutex_Seize as inline and body prototypes do not match) http://devel.rtems.org/ticket/2285 2015-03-06T13:58:03 :( 2015-03-06T13:58:13 you're working against the system, joel 2015-03-06T14:02:18 Sebastian tinkered. :( 2015-03-06T14:03:06 yeah. 2015-03-06T14:03:27 the macro is wrong 2015-03-06T14:03:31 As I was fixing Doxygen, I noticed the inconsistency. He should fix it. 2015-03-06T14:03:35 switch _executing and _mutx 2015-03-06T14:04:21 fwiw it doesn't matter since the arguments get passed through correctly 2015-03-06T14:04:38 I fixed it. 2015-03-06T14:04:42 as someone I know is fond of saying, "You could call it a monkey, I don't care" 2015-03-06T14:05:10 Thanks for spotting the fix. I am so focused on @param names not matching or just being missing, that it slipped by 2015-03-06T14:33:30 [git] 325a61d by Joel Sherrill: [coremuteximpl.h: Reorder macro names to match body and fix Doxygen warnings closes 2285.] http://devel.rtems.org/changeset/325a61d/rtems 2015-03-06T14:33:31 [git] a7e4de2 by Joel Sherrill: [Fix even more Doxygen issues] http://devel.rtems.org/changeset/a7e4de2/rtems 2015-03-06T14:33:51 That was a quick ticket 2015-03-06T14:35:46 *** kiwichris has joined #rtems 2015-03-06T14:38:56 ita, hi 2015-03-06T14:39:09 kiwichris: hi there 2015-03-06T14:39:56 ita, just been reading the updates to the ticket about mingw64/python. I think there change is going to be an on going issue for them. 2015-03-06T14:40:41 113 log files, only moxie has an error so far. 80 left to go. :) 2015-03-06T14:41:11 It needs to a compatible python or it is not. 2015-03-06T14:41:15 kiwichris: have you tried executing the command after setting MSYSTEM to ? 2015-03-06T14:41:36 FYI mips and m68k built today. No idea 2015-03-06T14:42:09 ita, no and I can add that to the RSB. The problem then becomes one of a user's waf experience, say when they attempt to the build RTEMS's waf version. 2015-03-06T14:42:56 ita, one of the reasons I looked at that python was to get the python devel libs installed so gdb can be built with python. The python.org .lib files are MSVC and do not link 2015-03-06T14:42:59 kiwichris: well, for that they should better call a .bat wrapper anyway? who calls python directly on windows? 2015-03-06T14:43:43 ita, no idea, I am not a big Windows users. I am thinking when inside the MSYS2 shell 2015-03-06T14:44:30 uh, then thinking about the msys shell, who will call python-mingw instead of the default msys2 python? 2015-03-06T14:45:19 The msys2 vs mingw64 python can be handled but no installing the MSYS2 one. 2015-03-06T14:46:22 kiwichris: do you mean building with the msys2 python will not work? 2015-03-06T14:46:47 Yes, because the RSB makes MSYS2 binaries 2015-03-06T14:47:00 and we want mingw64, ie native Windows binaries 2015-03-06T14:47:28 This support is present if someone wants to build in cygwin 2015-03-06T14:48:36 aha, so mingw+cygwin==ok, mingw+cmd.exe==ok but mingw+msys2==no ? 2015-03-06T14:49:46 ita, cygwin is just cygwin, no cross to mingw, mingw+cmd.exe == ok but the RSB cannot do this because it needs a shell to run gcc's configure and the last one is correct 2015-03-06T14:51:15 ita, a solution is to make python.org's python first in the path then attempt to reference the mingw64/python libs on gdb's configure command line with a path. My only concern is it being fragile. 2015-03-06T14:55:38 its, the other solution is to find a way to convert the python.org libs to a gcc format. 2015-03-06T14:58:27 *** dr__house|away is now known as dr__house 2015-03-06T14:58:37 kiwichris: fragile? what if configuration tests verify that the system is doing what you are expecting? 2015-03-06T15:03:22 ita, fragile in needing the user to install 2 different python packages and then set a path. One python does not let waf run (which can be worked around) and the other fails to build gdb as we expect. 2015-03-06T15:03:51 ita, the flow on is then having to have a special waf command to run on MSYS2 with a specific python. 2015-03-06T15:04:15 to build rtems and our examples 2015-03-06T15:04:45 * DrJoel_ waves at ita.. saw the whois and now know who you are :) 2015-03-06T15:05:31 kiwichris: i think you should report that to the mingw authors 2015-03-06T15:06:13 hi 2015-03-06T15:06:42 121 log files with only moxie failure. 2015-03-06T15:06:55 ok good 2015-03-06T15:08:02 ita, I agree this is an issue created by their change. 2015-03-06T15:09:17 verm__: 2015-03-06T15:09:17 closes 2285. 2015-03-06T15:09:17 did not get picked up in trac 2015-03-06T15:10:28 [trac] #2285/defect (new) updated by gedare ([325a61d524ca6539dd5ee0cb19e06d90e3deaca1/rtems]) http://devel.rtems.org/ticket/2285 2015-03-06T15:10:43 i guess you do need the # i didn't think you did 2015-03-06T15:10:44 verm__: can you set the default repository for changesets to be rtems in trac? 2015-03-06T15:10:54 [trac] #2285/defect (closed) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2285 2015-03-06T15:10:57 that or the case was bad. i dunno 2015-03-06T15:11:10 closed now 2015-03-06T15:11:58 kiwichris fyi rtl/libdl and rfs have most of the remaining Doxygen warnings that are not for "not documented" 2015-03-06T15:12:15 Haha 2015-03-06T15:12:28 Who wrote that code ? 2015-03-06T15:12:33 :) 2015-03-06T15:12:47 rtl.h has some weird group issues that I couldn't figure out 2015-03-06T15:12:54 I think I fixed some typos but most are real 2015-03-06T15:15:05 * DrJoel_ is moving away from the keyboad .. build has a couple more hours and then I will push. 2015-03-06T15:15:26 unfortunately, I need to reenable psxclock01 and leave this going again. I didnt have new toolsets for all targets so left it disabled 2015-03-06T15:16:51 heh 2015-03-06T15:17:49 gedare: there is no way to do that.. the reason is if you set a default you will never be able to change it as all the old references will break 2015-03-06T15:19:42 um, do we want to change it? 2015-03-06T15:19:50 this is the rtems project after all.. 2015-03-06T15:20:00 i'm pretty sure the default repository will be rtems. 2015-03-06T15:20:13 yep, but this involves the trac project which is used by tens of thousands of projects 2015-03-06T15:20:18 erm 2015-03-06T15:20:22 my point is they haven't written support for it since it's a bad idea 2015-03-06T15:20:37 it's come up before.. 2015-03-06T15:20:37 well, when i put just a changeset without a repo, and hover over the link, it says there is no default set 2015-03-06T15:20:42 so i assumed there is a way to set one. that's all. 2015-03-06T15:20:49 example? 2015-03-06T15:20:51 sec 2015-03-06T15:21:11 i wonder if we're talking about different things... 2015-03-06T15:21:30 could be 2015-03-06T15:21:42 i'll go make an example 2015-03-06T15:23:00 [trac] #2285/defect (closed) updated by gedare ([325a61d524ca6539dd5ee0cb19e06d90e3deaca1]) http://devel.rtems.org/ticket/2285 2015-03-06T15:23:04 https://devel.rtems.org/ticket/2285#comment:3 2015-03-06T15:23:24 oh it's because you didn't add a default repo 2015-03-06T15:23:36 that's an error not a notification 2015-03-06T15:23:41 or, hmm 2015-03-06T15:23:42 one sec 2015-03-06T15:23:45 maybe they changed it 2015-03-06T15:28:39 fixed 2015-03-06T15:28:51 looks like they added support for aliases, nice 2015-03-06T15:29:13 now to make sense of the source layout, sigh 2015-03-06T15:29:29 i'll send it out for review first hopefully i get it right in the first go 2015-03-06T15:29:54 *** Davidbrcz has quit IRC (Ping timeout: 272 seconds) 2015-03-06T15:30:56 ok 2015-03-06T15:31:21 i mean.. it is not so important we can also focus on the parts that make it hard to do waf building 2015-03-06T15:31:26 and refactor/rename other parts later 2015-03-06T15:33:07 ita, the issue on MSYS2 has come down to the https://docs.python.org/3/library/os.html#os.sep doco. I am stating the doco clearly says the os.sep can determine a POSIX or Windows file system. 2015-03-06T15:44:13 verm__: i look forward to seeing the proposal for the layout reorg. 2015-03-06T15:44:23 formal proposal i guess i should say :0 2015-03-06T15:45:16 oh verm__ thanks for fixing that default repo thing. 2015-03-06T15:45:58 kiwichris: indeed 2015-03-06T15:46:23 heh 2015-03-06T15:46:23 sure 2015-03-06T15:46:29 yeah going to go back and look at the email(s) now i hope i understand everything ok 2015-03-06T15:46:31 DrJoel_: any complaint bumping #2183 (SMP x86 thread migration) to 4.11.1 now? 2015-03-06T15:46:54 yeah it's a bit of a hare's nest. 2015-03-06T15:47:42 i don't know if that's a real idiom/colloquialism, but I think I like it. 2015-03-06T15:48:27 *** sujayraaj has quit IRC (Quit: Page closed) 2015-03-06T15:50:07 wrapping up for the weekend. cya 2015-03-06T15:51:36 cya! 2015-03-06T15:54:41 *** gedare has quit IRC (Ping timeout: 264 seconds) 2015-03-06T15:57:40 [trac] #2183/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2183 2015-03-06T15:58:11 I suppose gedare wasn't there to see that. I don't think anyone will work on it soon. If they do, it will get fixed. 2015-03-06T16:26:22 [git] a84249d by Chris Johns: [Windows fixes to build with MSYS2. The path handling has been cleaned up and support for file names longer than 256 characters.] http://devel.rtems.org/changeset/a84249d/rtems-source-builder 2015-03-06T16:26:22 [git] 46ec6b5 by Chris Johns: [Windows is win32 when testing for the host.] http://devel.rtems.org/changeset/46ec6b5/rtems-source-builder 2015-03-06T16:26:22 [git] c4c6264 by Chris Johns: [Update the documentation path.] http://devel.rtems.org/changeset/c4c6264/rtems-source-builder 2015-03-06T16:26:22 [git] 88cb50a by Chris Johns: [Fix removall for Windows and POSIX file systems.] http://devel.rtems.org/changeset/88cb50a/rtems-source-builder 2015-03-06T16:26:22 [git] 55a82ac by Chris Johns: [4.11/arm: Move to gdb-7.9.] http://devel.rtems.org/changeset/55a82ac/rtems-source-builder 2015-03-06T17:00:57 159 logs and only avr and moxie failing.. 2015-03-06T17:04:59 DrJoel: I know the discussion should be on the devel list 2015-03-06T17:05:10 *** Davidbrcz has joined #rtems 2015-03-06T17:05:11 i don't know why i get so much flak for wanting to get some seed information for an idea 2015-03-06T17:05:38 you guysd have a deeper understanding of various concepts that i don't, i don't see a reason to go to the list and ask basic questions 2015-03-06T17:05:52 that just results in burnout and the people who i want to actually read stop reading the thread since it's not of any concern to them 2015-03-06T17:06:07 i'd rather go into it with an idea that's at least a little fleshed out 2015-03-06T17:11:13 Propose it with milestones and not version numbers. Email us all a clean updated copy with the comments addressed. Hopefully we can get it technically sound as a proposal and not get derailed into things that need community feedback 2015-03-06T17:12:01 those are milestones 2015-03-06T17:12:24 when you list them as versions, there is overloaded meaning. 2015-03-06T17:12:26 i don't want to whitewash what it is, each milestone needs a version number, both are incremental 2015-03-06T17:12:35 i see 2015-03-06T17:13:05 i had not thought about that, it makes sense 2015-03-06T17:13:33 yeah i wasn't trying to get derailed i should have been more explicit in what i wanted 2015-03-06T17:13:49 I'm not either. They are just steps. One thing being involved with the Open Group has taught me is that if you can present just the technical meat as meat, it goes smoother. 2015-03-06T17:14:11 Give folks less space to derail the discussion. :) 2015-03-06T17:16:15 heh i can do that, i'm not used to doing it outside of a work environmen though 2015-03-06T17:16:52 This may be worse since the virtual nature of it is easier to derail and harder to get back on track 2015-03-06T17:16:56 i like hearing the options and concerns though i don't want to give an impression that i'm unwilling to listen 2015-03-06T17:17:15 that and a lot of differing experience 2015-03-06T17:17:49 We want you to be public about things. Hell I said a couple of days ago the new quad-core Pi wouldn't be a good gsoc project due to lack of documentation. 2015-03-06T17:18:30 Then I reversed myself when I learned it is close to the same peripherals at a different address with a different CPU core. Did you see Alan Cudmore actually had some code running on it quickly? 2015-03-06T17:18:50 nope, i'm not against being public however you guys have the advantage of deeper knowledge about what i'm asking 2015-03-06T17:19:05 where i have zero, to me going into the conversation without hearing back cursory input is a waste of time 2015-03-06T17:19:11 previously i've just been harrasing chris 2015-03-06T17:19:16 i can go back to doing that i'm sure he won't mind.... 2015-03-06T17:19:33 (with all that free time he has) 2015-03-06T17:21:26 The basic rules for a reorg are: 2015-03-06T17:21:38 1) cpukit has always been cpu model and board indepdent 2015-03-06T17:22:17 2) c/src traditionally was things that violated that rule but probably the Ada and C++ library can be multilib'ed now. .I recall tool issues not code issues 2015-03-06T17:22:56 3) libchip is BSP independent HW devices. NS16550 UART can show up on any board. Can easily be reorganized into libbsp/shared using current names 2015-03-06T17:23:36 4) libcpu. cache.h probably should be in cpukit since it defines an API expected to be provided by CPU model specific code. 2015-03-06T17:23:58 Some of libcpu/CPU can be libbsp/CPU/shared. 2015-03-06T17:24:17 cpukit/score/src and cpukit/score/cpu bother me 2015-03-06T17:24:20 Other parts may only be used by a single BSP so there is a strong argument to be made (if it is a SoC device driver) for moving it into the appropriate BSP 2015-03-06T17:24:34 i want to remove all the 'src' directories, there is no point in having them 2015-03-06T17:24:38 We can argue names independent of the "buckets" 2015-03-06T17:25:12 That's a case to make in the proposal. :) 2015-03-06T17:25:13 if something is used by a single bsp it should be part of that bsp only 2015-03-06T17:25:23 unless it changes down the road then it has to go to a common area 2015-03-06T17:25:29 that's the only way to encourage re-use of code 2015-03-06T17:25:53 Sometimes you see the reuse potential when you do the first one. Especially when there aren't many BSPs in an architecture. 2015-03-06T17:25:58 But yes.. that's the general rule 2015-03-06T17:26:02 and maintain a strict disconnection otherwise you get what we have now: some bsps that cherry pick from here-and-there 2015-03-06T17:27:52 sigh collecting all the changes suggested by gedare is a pain, hmm 2015-03-06T17:30:52 libbsp/shared should be ok for all of libchip <- i don't understand this one 2015-03-06T17:31:01 c/src/libchip -> rtems/bsp/shared ? 2015-03-06T17:31:27 I was giving instructions based on the current tree.. so that sounds right in your proposal 2015-03-06T17:31:36 ok 2015-03-06T17:31:40 i figured that's what you meant 2015-03-06T17:31:43 The idea is that any of this code can be reused by any BSP. 2015-03-06T17:33:19 Although it has exceptions and suffers from history, the organization is really driven by the rules above. multilib across all BSPs, CPU model specific, HW device shared across architectures, SoC commonly reused peripherals, 2015-03-06T17:33:35 shared BSP infrastructure due to RTEMS, shared BSP code within an architecture, and BSP specific 2015-03-06T17:35:33 right 2015-03-06T17:35:45 and i really want to reflect that cleanly with the directory layout and source file names 2015-03-06T17:36:05 i also want to introduce the concept that 'rtems' is completely unhooked from 'bsp' 2015-03-06T17:36:37 that way all you need to do is build 'librtems' and you can write your own bsp 2015-03-06T17:36:42 That was the intention of the cpukit. The mirror image was bspkit but that never came together. Probably too hard in autoconf. 2015-03-06T17:36:53 there is some sharing that goes on right now from c -> cpukit 2015-03-06T17:37:19 i had to add some hacks to waf in order to add strange include dirs that didn't make any sense 2015-03-06T17:38:27 It is supposed to be cpukit defines interfaces and provides frameworks in some cases. 2015-03-06T17:38:35 *** gedare has joined #rtems 2015-03-06T17:38:35 *** ChanServ sets mode: +o gedare 2015-03-06T17:38:52 I am about to burn out for the day. Already hit the beer fridge.. Sierra Nevada Beer Camp. 2015-03-06T17:39:46 heh long couple days 2015-03-06T17:40:11 Finally above freezing but likely will refreeze tonight 2015-03-06T17:40:51 We have tickets to Peter and the Star Catchers tomorrow night (touring Broadway show) including cast party after so I really want it to be ok to get out 2015-03-06T17:41:37 *** Davidbrcz has quit IRC (Ping timeout: 265 seconds) 2015-03-06T17:41:56 166 logs and only 2 fails (avr, moxie). I have actually rebuilt m68k, mips, bfin, and now arm tools while this has been going on. 2015-03-06T17:42:26 After the last 27 BSPs finish I will push the changes. Watch the "updates" because I think all of those let the milestone bump to 4.11.1 2015-03-06T17:42:42 where do we stand on 2205 RTEMS tools don't build on MinGW due to missing (kiwichris) 2015-03-06T17:42:53 sure. 2015-03-06T17:43:48 well i just dropped by for a quick look. heading out. 2015-03-06T17:44:20 gedare only avr and moxie fails out of 166/193 :) 2015-03-06T17:48:09 *** gedare has quit IRC (Ping timeout: 245 seconds) 2015-03-06T17:55:38 awesome 2015-03-06T17:55:45 what is the problem with avr? 2015-03-06T18:00:20 avr has a gcc ICE compiling RFS 2015-03-06T18:01:08 avr doesn't seem to get much attention from gcc. I suppose I could add logic to disable rfs to see if it gets farther after this round. 2015-03-06T18:01:12 We have to do that for libdl 2015-03-06T18:04:44 avr doesn't get much attention that's surprising 2015-03-06T18:04:49 are there other compilers out there that are used? 2015-03-06T18:11:00 *** jpc has joined #rtems 2015-03-06T18:25:40 *** jpc has quit IRC (Quit: Page closed) 2015-03-06T18:28:39 Not free/open source ones AFAIK. Atmel provides a ton of patches the last time I looked. The gcc maintainer from atmel isn't there anymore. I don't know if he got replaced. 2015-03-06T18:33:34 ouch that sucks 2015-03-06T18:34:07 There are two maintainers listed. But it is low on the rtems target priority from my perspective. 2015-03-06T18:35:04 any suggestions for c/src/libchip? i had it under rtems/bsp/chip 2015-03-06T18:37:02 Are you proposing we keep shared? It may make sense to just disperse the subdirectories into rtems/bsp/shared/CHIPTYPE 2015-03-06T18:37:37 then what about c/src/lib/libcpu? 2015-03-06T18:39:34 It is going to vary by directory. Some may be able to go all the way to a specific BSP. Most are a subdir under rtems/BSP/ARCH/shared. I think at least the cache.h API file can move to cpukit. 2015-03-06T18:39:40 so a case by case 2015-03-06T18:40:29 I dispersed some of the mips not long ago. If you can identify things used by a single BSP, moving them now is no big deal. Those were the cases I was looking for. 2015-03-06T18:41:01 hmm 2015-03-06T18:41:19 not sure how to explain this in my proposal 2015-03-06T18:42:31 Just write a set of rules. There will have to be a judgment call on each one. 2015-03-06T18:43:08 If you can script looking at the libbsp/CPU/BSP/Makefile.am to see which BSPs pull in which .rel files from libcpu, then the decisions may be easier 2015-03-06T18:43:51 Move it to a nice new location, change the libcpu Makefile.am, change the BSP's Makefile.am. Seriously this is reorg that is easy enough in the current build environment and may make your life easier 2015-03-06T18:44:21 i can already move the files automatically 2015-03-06T18:44:25 without even editing a single file 2015-03-06T18:44:33 and i can prove using unit tests that the build is the same as before 2015-03-06T18:44:41 i don't want to touch auto*, period 2015-03-06T18:45:06 i really just want to get this deployed prove it's OK and forget about the past 2015-03-06T18:45:29 Each subdirectory will have to be evaluated anyway. We do need a list of libcpay.u XXX.rel is used by BSPs X, Y, and Z. If you find unused code, we can kill it to 2015-03-06T18:45:32 today 2015-03-06T18:45:59 i've already unwound all of that 2015-03-06T18:46:03 years ago 2015-03-06T18:46:11 look at the c/wscript_* files 2015-03-06T18:46:19 and in the wscript files in cpukit/ 2015-03-06T18:46:39 there are some duplicate files which are duplicated on purpose i just wanted to get it built 2015-03-06T18:46:49 But the movement of each subdirectory in libcpu requires a decision on where to put it based on what it is and how many BSPs use it. 2015-03-06T18:46:54 it's easy enough to write something in the build that will print out if two bsps have shared files 2015-03-06T18:46:58 yeah 2015-03-06T18:47:30 well i'm aiming for 90% accuracy anyway, the good thing is once waf is in moving things around will be a trivial operation 2015-03-06T18:48:03 With a little information and analysis, the first move is more likely to be the last one. :) 2015-03-06T18:50:02 i agree but now we don't have a reason to keep things where they are for legacy reasons 2015-03-06T18:50:14 in a years time we may want to move some parts around again who knows 2015-03-06T18:50:22 it's rare to get it right the first time with a move this large 2015-03-06T19:00:06 well i'll capture as much as i can from the email you and geare sent 2015-03-06T19:00:15 gedare sent and post this to the devel list in a few mins 2015-03-06T19:03:54 OK. I am doing some consistency checking on Michele's new book. :) 2015-03-06T19:04:09 is that what the beer was for? :) 2015-03-06T19:04:49 lol.. no she asked me to do it after I drank the beer. :) 2015-03-06T19:05:13 Working at the family room table isn't ideal. My office chair is WAY more comfortable 2015-03-06T19:05:29 the beer is to make the chair more comfortable? that's a good reason 2015-03-06T19:05:49 190 log files and 4 failures. I suspect I missed a couple of arm BSPs not having enough memory to link some tests. Almost done with this pass 2015-03-06T19:05:56 That's my excuse and I am sticking to it. :) 2015-03-06T19:06:20 i worked around this in waf, all the tests are built but a warning is printed if it fails due to memory constraints 2015-03-06T19:06:30 no need to unmark them we can just build them all by default 2015-03-06T19:07:26 so i should remove the version numbers and just use milestones eh 2015-03-06T19:07:29 ok then 2015-03-06T19:07:43 i'll use ordered milestones and discuss the topic of the 5.x version 2015-03-06T19:07:54 or should i leave the 'verison' information out entirely? 2015-03-06T19:18:47 Just treat them as milestones. And steps within the milestones 2015-03-06T19:19:05 ok 2015-03-06T19:19:13 i'll do that 2015-03-06T19:21:00 sent 2015-03-06T19:21:04 going to send the 2nd email shortly. 2015-03-06T19:46:54 sent 2015-03-06T19:52:45 *** jpc has joined #rtems 2015-03-06T20:04:38 [git] 8f49dda by Joel Sherrill: [lpc1768_mbed_ahb_ram.tcfg: Remove as it is a junk file This file does not have "-testsuite" in the name and is ignored by the build system.] http://devel.rtems.org/changeset/8f49dda/rtems 2015-03-06T20:04:38 [git] a87e8b8 by Joel Sherrill: [lpc1768_mbed_ahb_ram-testsuite.tcfg: Add ftp01] http://devel.rtems.org/changeset/a87e8b8/rtems 2015-03-06T20:04:38 [git] 683ebac by Joel Sherrill: [raspberrypi: Do not include default IRQ handler and BSP specific one This was tripping a linker error in the dl0[12] tests. closes 2247.] http://devel.rtems.org/changeset/683ebac/rtems 2015-03-06T20:04:38 [git] 9bef63e by Joel Sherrill: [Temporarily disable libdl for bfin There is an issue linking dl0* which has not been resolved. This issue is being tracked but is not considered a release blocker. This patch is a workaround which disables libdl for the bfin until the ticket is resolved. updates 2252.] http://devel.rtems.org/changeset/9bef63e/rtems 2015-03-06T20:04:38 [git] 1d061ee by Joel Sherrill: [Temporarily disable libdl for v850 There is an issue linking dl0* which has not been resolved. This issue is being tracked but is not considered a release blocker. This patch is a workaround which disables libdl for the v850 until the ticket is resolved. updates 2260.] http://devel.rtems.org/changeset/1d061ee/rtems 2015-03-06T20:04:38 [git] a726ca6 by Joel Sherrill: [Temporarily disable libdl for lm32 There is a GCC ICE when building libdl. This temporarily disables building libdl until that is resolved. updates 2283.] http://devel.rtems.org/changeset/a726ca6/rtems 2015-03-06T20:04:38 [git] ee113c2 by Joel Sherrill: [dltests-broken-on-this-bsp.tcfg: New file used for dl* tests work around] http://devel.rtems.org/changeset/ee113c2/rtems 2015-03-06T20:04:39 [git] e516d45 by Joel Sherrill: [Temporarily disable libdl tests on some MIPS BSPSs These BSPs are little endian while most are big endian. The dl tool does not support the -EL option yet. updates 2279.] http://devel.rtems.org/changeset/e516d45/rtems 2015-03-06T20:04:39 [git] 7af34c1 by Joel Sherrill: [Always disable networking for h8300] http://devel.rtems.org/changeset/7af34c1/rtems 2015-03-06T20:04:40 [git] dea43a0 by Joel Sherrill: [Temporarily ignore dl* tests on some PowerPC BSPs Some BSPs have issues linking the dl* tests. These issues have not been investigated. This is a temporary workaround until we have an explanation and permanent solution. updates 2258.] http://devel.rtems.org/changeset/dea43a0/rtems 2015-03-06T20:04:40 [git] 1ae60ae by Joel Sherrill: [Add fsscandir01 variants as needed to more BSPs testsuite configuration] http://devel.rtems.org/changeset/1ae60ae/rtems 2015-03-06T20:04:41 [git] 19a1f6b by Joel Sherrill: [Temporarily disable libdl for h8300 The h8300 gets a linking error for the dl0* tests. This temporarily disables libdl until that can be investigated. updates 2284.] http://devel.rtems.org/changeset/19a1f6b/rtems 2015-03-06T20:05:57 [trac] #2247/defect (closed) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2247 2015-03-06T20:06:15 [trac] #2252/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2252 2015-03-06T20:06:33 [trac] #2258/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2258 2015-03-06T20:06:47 [trac] #2260/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2260 2015-03-06T20:07:04 [trac] #2279/defect (assigned) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2279 2015-03-06T20:07:21 suck on that!!! 2015-03-06T20:07:46 woohoo! 2015-03-06T20:07:56 7 tickets left 2015-03-06T20:07:58 I think I updated all the tickets I needed to close or move the milestone. I am mentally wiped so double check 2015-03-06T20:09:09 My tools are all new as of yesterday or today except moxie (broken). 2015-03-06T20:09:30 Overnight I will built all BSPs using unmodified head, all tests, networking, etc. 2015-03-06T20:09:37 Hopefully only avr and moxie fail 2015-03-06T20:12:11 seems ok 2015-03-06T20:12:18 are notepads going to be handled in 4.11? 2015-03-06T20:13:29 Yes but marked as deprecated. Chris wants task variables in same category. 2015-03-06T20:13:49 * verm__ shivers 2015-03-06T20:13:57 Gedare and I were discussing how to mark the methods with gcc attributes so you get linker warning. 2015-03-06T20:13:57 i didn't know sebastian liked avoiding {}'s 2015-03-06T20:14:14 We are conservative. Mark them as deprecated. Remove soon after branching 2015-03-06T20:14:23 that's a good idea 2015-03-06T20:14:25 ()'s on expressions was what I asked about. 2015-03-06T20:14:43 He also had an editor setting where he wrote f( a, b, c) and I am still finding those to fix. :( 2015-03-06T20:14:47 if ( _Thread_Executing->is_preemptible || the_thread->current_priority == 0 ) _Thread_Dispatch_necessary = true; 2015-03-06T20:14:49 is really bad 2015-03-06T20:15:12 because you can get silent failures in certain cases 2015-03-06T20:15:39 i know you were talking about parentheses i just noticed that in the patch 2015-03-06T20:15:44 It would take a lot of work to switch to all single line true only if's to have braces 2015-03-06T20:16:19 there are some nice free static checkers to look that specifically :) 2015-03-06T20:19:11 something like that would be in the category of "make a decision and make GCI tasks" 2015-03-06T20:19:34 I also have finish the psxhdrs directory so it has all methods. Another GCI possibility 2015-03-06T20:19:49 all the doxygen file headers will need to be updated 2015-03-06T20:20:04 for what? 2015-03-06T20:20:30 i've never looked but RTEMS doesn't use the doxygen commands to describe / categorise source files? 2015-03-06T20:20:56 I don't think so. What commands? 2015-03-06T20:21:45 @brief 2015-03-06T20:21:45 verm__: Error: "brief" is not a valid command. 2015-03-06T20:21:49 @ingroup list,of,groups 2015-03-06T20:21:49 verm__: Error: "ingroup" is not a valid command. 2015-03-06T20:21:59 @see 2015-03-06T20:21:59 verm__: Error: "see" is not a valid command. 2015-03-06T20:22:17 eg if there is file.c and file.h in file.c you do @see file.h and in file.h @see file.c 2015-03-06T20:22:38 also @file name.x 2015-03-06T20:22:54 that way you can go @see anywhere in the source and doxygen will know where to find it, it requires unique names though 2015-03-06T20:23:02 that's where grouping comes in handy 2015-03-06T20:23:06 a nice way to organise source 2015-03-06T20:27:25 We do those but I am not sure it has worked all the way through the tests 2015-03-06T20:27:58 The psxhdrs test is just a lot of simple C files that ensure you can invoke a method as the man page says. Exact .h files. Lots of missing cases 2015-03-06T20:28:59 once waf is in and settled i'll be moving onto the test suite 2015-03-06T20:29:06 i have a huge proposal for that one 2015-03-06T22:39:41 *** idleguy has joined #rtems 2015-03-06T22:51:04 [trac] #2208/enhancement (new) updated by amar (I wrote this in python using pygit2 it gives the following output The full log is ...) http://devel.rtems.org/ticket/2208 2015-03-06T22:51:37 [trac] rtems_log.py (ChangeLog generator.) attached to #2208 by amar http://devel.rtems.org/ticket/2208 2015-03-06T22:52:10 [trac] #2208/enhancement (accepted) updated by amar (empty comment) http://devel.rtems.org/ticket/2208 2015-03-06T22:53:31 *** jpc has quit IRC (Ping timeout: 246 seconds) 2015-03-06T22:55:13 [trac] #2205/defect (new) updated by amar (Formatting.) http://devel.rtems.org/ticket/2205 2015-03-06T22:56:48 [trac] Developer/Release edited by amar (Add ChangeLog step.) http://devel.rtems.org/wiki/Developer/Release 2015-03-06T22:58:36 [trac] Developer/Release edited by amar (trac.) http://devel.rtems.org/wiki/Developer/Release 2015-03-07T00:05:14 *** dr__house is now known as dr__house|away 2015-03-07T00:23:32 *** jhollister has joined #rtems 2015-03-07T01:33:19 *** idleguy has quit IRC (Quit: Page closed) 2015-03-07T04:33:59 *** idleguy has joined #rtems 2015-03-07T04:51:53 *** Davidbrcz has joined #rtems 2015-03-07T04:52:10 *** Hesham has joined #rtems 2015-03-07T05:39:41 *** Davidbrcz has quit IRC (Ping timeout: 264 seconds) 2015-03-07T06:06:39 *** gedare has joined #rtems 2015-03-07T06:06:40 *** ChanServ sets mode: +o gedare 2015-03-07T06:18:29 *** gedare has quit IRC (Ping timeout: 246 seconds) 2015-03-07T07:30:59 *** Davidbrcz has joined #rtems 2015-03-07T07:51:16 *** Davidbrcz has quit IRC (Ping timeout: 255 seconds) 2015-03-07T08:13:41 *** idleguy has left #rtems 2015-03-07T08:15:18 *** kiran4399 has joined #rtems 2015-03-07T08:17:07 Can anyone tell me how to fill the table in https://devel.rtems.org/wiki/GSoC/2015 ? 2015-03-07T08:17:18 beng-nl: ping 2015-03-07T08:31:24 *** ragunath has joined #rtems 2015-03-07T08:48:33 kiran4399: ack 2015-03-07T08:51:43 So how can I fill the tables in above mentioned link ? 2015-03-07T08:55:03 it's a wiki page 2015-03-07T08:55:13 so you register yourself for the wiki, log in 2015-03-07T08:55:40 and then you can click 'edit this page' at the bottom 2015-03-07T09:05:41 *** waleed has joined #rtems 2015-03-07T09:08:03 kiran4399: ^ 2015-03-07T09:09:40 ragunath: nice that you're also interested! 2015-03-07T09:09:59 beng-nl: i have compiled and run beagle bone images ... would like to share the proofs 2015-03-07T09:10:13 aha, on hardware? 2015-03-07T09:10:24 yes i have a beaglebone black 2015-03-07T09:10:38 great! 2015-03-07T09:10:53 with a serial cable i take it 2015-03-07T09:10:59 or you wouldn't see much :) 2015-03-07T09:11:53 ragunath: do any of the suggested projects interest you? 2015-03-07T09:11:54 yes with a serial cable PL2303 USB 2015-03-07T09:12:03 ragunath: good 2015-03-07T09:12:28 yes i am interested in adding more peripheral supportt... guess that has the highest priority 2015-03-07T09:12:59 beng-nl: just like me... :-) 2015-03-07T09:13:34 beng-nl: Can me and ragunath work on the same project ? 2015-03-07T09:13:36 kiran4399: yes :) we won't forget you 2015-03-07T09:14:05 ethernet support is important i guess 2015-03-07T09:15:50 kiran4399: it's possible but i think unlikely we'll have 2 beagle gsoc students 2015-03-07T09:17:13 ragunath: the only thing that's *officially* required is to run a modified hello world on the beagle bsp 2015-03-07T09:17:29 unofficially, the more you can show, the better 2015-03-07T09:17:46 beng-nl: So you mean to say both can apply for the project but only one can get through ? 2015-03-07T09:17:48 personally my suggestion for the beagle bsp is to debug a failing test 2015-03-07T09:18:22 kiran4399: it's not really my decision and it depends on some external factors (slots, other projects) but personally i think it's a bad idea to accept more than one yes 2015-03-07T09:19:18 beng-nl: shall I change to Raspberry-pi ? 2015-03-07T09:19:20 kiran4399, ragunath: the hello world project is described at https://devel.rtems.org/wiki/GSoC/GettingStarted 2015-03-07T09:20:16 [trac] GSoC/2015 edited by kiran4399 (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-03-07T09:20:52 kiran4399: i'd say it's premature to change projects, just because there's another interested person. it's up to you though. 2015-03-07T09:21:53 [trac] GSoC/2015 edited by kiran4399 (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-03-07T09:23:26 beng-nl: Well then I'll follow your advice.. I'll stick to this. The reason why I preferred Beaglebone black over RPi was becuase I worked on the AM335x processor and its subsystems. But anyway I'll work on this.. 2015-03-07T09:23:35 beng-nl: i ran the tests on qemu (bbxm) ... only one failure but 11 timeouts.... but that failure test ran successfully on hardware (BBB) 2015-03-07T09:23:42 kiran4399: that is a good reason to stay :-) 2015-03-07T09:24:03 ben 2015-03-07T09:24:10 rag 2015-03-07T09:25:23 beng-nl: i will explore more on the test cases and let you know 2015-03-07T09:26:43 beng-nl: Do you know the PRU subsystem on the BBB ? 2015-03-07T09:26:43 ragunath: great idea. honestly i'm asking a lot as a submission project but it would be very valuable/impressive if you could solve one or even just show us how far you got. solving/diagnosing one on the bbxm emulator would also be great. 2015-03-07T09:27:21 kiran4399: only in very general terms 2015-03-07T09:29:39 beng-nl: It is a programmable real time unit which is a seperate submicroprocessor independent of the OS activity in the main processor. It is basically for real-time purposes to get fast input ouput from the GPIO pins. Implementing this in the BSP would be great !! 2015-03-07T09:29:59 yep, that much i knew, it's on the suggested list actually for htis reason 2015-03-07T09:30:01 this* 2015-03-07T09:33:02 [trac] Developer/Projects/Open/ImproveBeagleBSP edited by beng (empty comment) http://devel.rtems.org/wiki/Developer/Projects/Open/ImproveBeagleBSP 2015-03-07T09:33:09 oops, forgot comment 2015-03-07T09:34:55 beng-nl:So the approach I would like to have for the project is: 1) GPIO 2)ADC 3)PRU.. What do you think ? 2015-03-07T09:36:21 [trac] Developer/Projects/Open/ImproveBeagleBSP edited by beng (softened the required change to official hello world + debug test suggestion) http://devel.rtems.org/wiki/Developer/Projects/Open/ImproveBeagleBSP 2015-03-07T09:39:38 *** Davidbrcz has joined #rtems 2015-03-07T09:46:01 kiran4399: not sure. 2015-03-07T09:46:53 kiran4399: which peripherals should be done first depend more on the expertise & motivation of the student, weighted by how rtems is used i suppose 2015-03-07T09:47:15 kiran4399: personally i think MMC has a higher value/work tradeoff than e.g. pru 2015-03-07T10:00:22 beng-nl: Alright I'll buy your advice... I think I'll keep MMC as first priority.. 2015-03-07T10:00:45 +1 MMC and NIC would be high on my list. 2015-03-07T10:00:55 What other peripherals are left to support Ben? 2015-03-07T10:01:08 beng-nl: Then I'll go to GPIO becuase it is really very important and basic for Embedded boards 2015-03-07T10:01:15 DrJoel_: i made a list of suggestions at https://devel.rtems.org/wiki/Developer/Projects/Open/ImproveBeagleBSP 2015-03-07T10:01:25 kiran4399: ok 2015-03-07T10:02:35 beng-nl: i can see a i2c file in beagle directory ... is i2c already supported? 2015-03-07T10:02:35 DrJoel_: i expect you'll know best what would be most appreciated by rtems users 2015-03-07T10:02:39 OK.. just woke up and still fuzzy 2015-03-07T10:02:44 beng-nl: is usb gadget support available? 2015-03-07T10:03:07 Is the i2c using the new or old API? I suspect it predates adding the Linux compatible API 2015-03-07T10:03:24 ragunath: for the bbxm possibly, i never tested it, i wouldn't bet too much on it 2015-03-07T10:03:34 oh, usb support is missing 2015-03-07T10:03:39 verm__: yes 2015-03-07T10:03:50 it is? nice 2015-03-07T10:03:55 that's easily the best feature of the BB 2015-03-07T10:04:00 use it all the time 2015-03-07T10:04:08 no sorry.. eh i meant yes usb is missing 2015-03-07T10:04:20 oh, heh 2015-03-07T10:04:32 usb gadget == usb otg == usb device mode? 2015-03-07T10:04:46 sigh.. force avr to skip RFS and it immediately gets a GCC ICE in dosfs 2015-03-07T10:04:50 yeah device mode, so you can make it into a USB key 2015-03-07T10:05:04 verm__: aha right 2015-03-07T10:05:09 i took a BB and made it act like a USB key with wireless instead of having to take a usb key to saras sewing machine i have wireless access to it now 2015-03-07T10:05:11 pretty awesome 2015-03-07T10:05:17 drag and drop files to it 2015-03-07T10:05:25 verm__: wow, super cool 2015-03-07T10:05:46 yep, any legacy device that only supports usb keys can be made wireless 2015-03-07T10:05:51 do we support for musb core in rtems? 2015-03-07T10:06:02 BB really coveted the market with that, no support in any other device that i know for host mode 2015-03-07T10:06:33 DrJoel: :( 2015-03-07T10:07:57 DrJoel: i sent the email(s) out last night and removed the 'version' numbers using milestones only 2015-03-07T10:08:12 should i send a seperate email about the versioning or leave it for now? 2015-03-07T10:08:16 i'd really like to start opening tickets 2015-03-07T10:09:36 beng-nl: why is that we cannot boot rtems directly on Beaglebone black just like we do for debian and angstrom ? 2015-03-07T10:11:29 *** Davidbrcz has quit IRC (Ping timeout: 265 seconds) 2015-03-07T10:12:05 kiran4399: hm, you can boot rtems directly on BBB, with an sdcard or netbooting, i've done it many times 2015-03-07T10:12:11 using uboot that is 2015-03-07T10:12:22 kiran4399: what do you mean by boot rtems directly? 2015-03-07T10:12:58 I mean booting it and ssh into it.. 2015-03-07T10:13:03 aha i see 2015-03-07T10:13:18 that would work 2015-03-07T10:13:32 if network would work & there was an ssh daemon 2015-03-07T10:13:58 so there is no ssh daemon running in the init.d directory ? 2015-03-07T10:14:00 kiran4399: but if you don't have console output, it's impractical to make any lowlevel changes 2015-03-07T10:15:34 beng-nl: means ? 2015-03-07T10:16:58 kiran4399: rtems apps are single-image, so a single purpose app linked with rtems system code running as a unit. so the notion of a seperately started ssh daemon doesn't fly, it's not general-purpose like that. there is telnet support (started as a thread that you could enable in the app at build time) and conceivably ssh could work the same way, it's not in base afaik though 2015-03-07T10:18:00 beng-nl: is rtems uniprocess os ? 2015-03-07T10:18:23 kiran4399: i mean making low-level changes (like peripheral support) is impractical with only network access (i.e. without console access) 2015-03-07T10:18:43 kiran4399: no, there is SMP support (in fact the M stands for multiprocessor) 2015-03-07T10:19:07 oh, well, unless you mean the other thing, then yes, RTEMS is a single process and doesn't fork or exec 2015-03-07T10:21:20 You download each statically linked executable individually. It is an application with just the OS capabilities it needs to perform application mission 2015-03-07T10:21:36 You will debug with JTAG or some other remote gdb setup 2015-03-07T10:21:59 beng-nl: So ssh can also started as a thread ? 2015-03-07T10:22:17 kiran4399: yes i expect that is possible though there is no code in base for it 2015-03-07T10:22:22 telnetd does 2015-03-07T10:22:56 You could build a base image with dynamic loading like Alan Cudmore does. But without networking, how would you get the module to load? 2015-03-07T10:23:54 DrJoel_: By using JTAG ? 2015-03-07T10:24:42 DrJoel_: is someone looking at #2205 and #2245? 2015-03-07T10:24:45 DrJoel_: dropbear has an mit license, would that hypothetically be ok in base? 2015-03-07T10:24:52 what are those? 2015-03-07T10:25:11 socket.h header file issue is next on my list 2015-03-07T10:25:16 https://devel.rtems.org/wiki/Release/4.11 top two on there RTEMS tools don't build on MinGW due to missing , IPV4 sys/socket.h should include 2015-03-07T10:25:23 ah, ok 2015-03-07T10:25:28 The other must be Chris 2015-03-07T10:25:43 dropbear is a mini ssh-server.. 2015-03-07T10:28:58 [trac] #2208/enhancement (closed) updated by amar (Closing this ticket it has been linked from [wiki:Developer/Release]. The script ...) http://devel.rtems.org/ticket/2208 2015-03-07T10:31:23 Can anyone tell me the difference between libcpu and libbsp ? 2015-03-07T10:32:13 [trac] #2217/enhancement (closed) updated by amar (Closing this as the current behaviour is expected. Anyone willing to add support ...) http://devel.rtems.org/ticket/2217 2015-03-07T10:32:20 verm__ ? 2015-03-07T10:32:26 ? 2015-03-07T10:33:27 Can I get a double check on a ticket? I think I actually fixed https://devel.rtems.org/ticket/2245 and forgot to close the ticket. 2015-03-07T10:33:49 Does testsuites/psxtests/psxhdrs now have subirectories based on .h name? 2015-03-07T10:34:00 And did I add sys/socket with ALL the methods from socket.h? 2015-03-07T10:34:03 If so, it is done. 2015-03-07T10:34:14 yeah you fixed it 2015-03-07T10:34:19 a month ago 2015-03-07T10:35:02 [trac] #2245/defect (closed) updated by amar (Fixed in [547c2282eea7a2e78341413f805a7781775b292c]) http://devel.rtems.org/ticket/2245 2015-03-07T10:35:22 5 tickets.. 2015-03-07T10:36:08 FWIW that test is the one I mentioned which GCI could add 100s more cases to. It would ensure our APIs adhere to the man page 2015-03-07T10:36:45 ok.. libbsp is strictly board dependent code. Things shared across multiple boards can go in a shared directory (with an architecture or across all) 2015-03-07T10:37:41 libcpu contains infrastructure needed because of the CPU model that are not needed by RTEMS OS proper. Originally conceived to handle things like cache, on-cpu DMA, 2015-03-07T10:38:11 With more SoC's, it now has peripherals that show up on multiple SOCs within a family. Say Freescale uses the same clock on all their CPU models. We want to share it. 2015-03-07T10:38:51 libcpu and libchip will end up redistributed in the future to mostly libbsp as appropriate 2015-03-07T10:38:54 i'm hoping i can get my test harness system up before GCI if i can we can have thousands of tasks for students to do 2015-03-07T10:39:06 Just need 100s of mentors. :) 2015-03-07T10:39:19 heh, well the nice thing is it is a web UI 2015-03-07T10:39:24 Last year we were 4 mentors down and had to decline 2015-03-07T10:39:40 that will be open to the world for submission and post-review 2015-03-07T10:39:43 so it will be easy to manage 2015-03-07T10:40:08 the other GCI tasks will be for populating the new website 2015-03-07T10:40:08 [trac] #2283/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2283 2015-03-07T10:40:16 i'm going to post an email about that shortly 2015-03-07T10:40:22 and open it up for people to look at 2015-03-07T10:41:22 [trac] #2282/defect (new) updated by joel.sherrill (For the purposes of 4.11, Gedare moved us back to a gcc 4.8 based toolset until this ...) http://devel.rtems.org/ticket/2282 2015-03-07T10:41:32 [trac] #2282/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2282 2015-03-07T10:41:57 Update https://devel.rtems.org/wiki/Release/4.11 now 2015-03-07T10:42:06 Unless something changes, I think I am off the hook 2015-03-07T10:42:25 should notepads be moved to gedare? 2015-03-07T10:42:36 who is handling moxie? 2015-03-07T10:42:53 [trac] #2265/enhancement (assigned) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2265 2015-03-07T10:43:16 Sneaky of gedare to raise the issue, work on it, and assign it to me. :) 2015-03-07T10:43:32 Hesham was trying to run down the moxie with Anthony Green's assistance 2015-03-07T10:43:44 i don't think he's talked to anthony has he? 2015-03-07T10:43:50 Hesham ...? 2015-03-07T10:44:10 i'm pretty sure hesham, was working on it by himself since anthony is MIA 2015-03-07T10:44:11 No news, I did mention him here, but no reply 2015-03-07T10:44:30 i sent him an email a few days ago asking if he he wanted to grab a pint next week 2015-03-07T10:44:34 no response 2015-03-07T10:45:21 I didn't try to build with my ~4 applied patches 2015-03-07T10:45:30 Let my try now, hopefully it may work 2015-03-07T10:45:46 Do you need a "just-working" patch? 2015-03-07T10:45:54 OK.. going away from keyboard. I thought i still had the issue 2015-03-07T10:45:57 or set.. yes 2015-03-07T10:48:53 DrJoel_: before you go 2015-03-07T10:49:01 should i hold off on an email about the versioning for the waf migration? 2015-03-07T10:49:41 i don't want to introduce that overload you were talking about...i also want to continue on with my work 2015-03-07T10:53:01 DrJoel_: So my work will mostly be in libbsp .. right ? 2015-03-07T10:53:07 Hold off. Better to get the tree format worked through. Plus it is a weekend. Posting on Friday is often a good way to get no response. 2015-03-07T10:53:21 kiran4399 most likely. 2015-03-07T10:53:32 ok 2015-03-07T10:53:38 yeah i just wanted to get it out there 2015-03-07T10:53:40 Except if you work on USB or NIC. That COULD end up in the new TCP/IP stack code base. It contains all the drivers 2015-03-07T10:54:02 if i had waited something would have inevitably come up to dlay it yet again 2015-03-07T10:54:41 *** waleed has quit IRC (Ping timeout: 264 seconds) 2015-03-07T10:55:22 I know. That's why I just pushed through all the libdl issues. It was a blocker. I wasn't going to allow a release with so many architectures failing to build and link all tests. 2015-03-07T10:55:33 and it was clear that Chris wasn't going to get to it. 2015-03-07T10:55:50 It is also possible that some of the architectures may never get attention to get it working. 2015-03-07T10:56:30 yeah 2015-03-07T10:56:43 well this will be solved very quickly with phabricator and pre commit building 2015-03-07T10:56:59 if developers can't get their commits in due to breakage then we'll have to re-evalulate the BSPs 2015-03-07T10:57:21 the biggest reason the source layout will be handy is we can 'peel off' BSPs into their own packages 2015-03-07T10:57:25 This is the worst one I recall. libdl was cross architecture and architecture dependent 2015-03-07T10:57:26 that can be maintained seperately 2015-03-07T10:57:57 The new stuff will help with BSP issues but this wasn't one. Chris committed code which had to be ported to every architecture and simply couldn't do it without getting the code to more people 2015-03-07T10:57:58 still, it's pretty late in the game to be fixing it, it shouldn't have landed with as many issues as it has -- i'm not laying blame 2015-03-07T10:58:07 this is specifically where phabricator and buildbot will help us 2015-03-07T10:58:33 It should help for adding a test which some BSPs can't fit into memory. I am constantly fixing those because no one builds all BSPs. 2015-03-07T10:58:51 yeah i ran into lots of those :( 2015-03-07T10:58:58 they won't make the build fail anymore 2015-03-07T10:59:05 waf automatically captures the output and passes 2015-03-07T10:59:14 so we can have a warning log of tests that are too big 2015-03-07T11:00:10 *** kiran4399 has quit IRC (Quit: Page closed) 2015-03-07T13:01:26 *** dr__house|away is now known as dr__house 2015-03-07T13:06:22 The current partial moxie patch didn't work 2015-03-07T13:08:22 time to try a gcc snapshot? 2015-03-07T13:10:43 This will need a new RSB .cfg file or there is an existing one? 2015-03-07T13:10:47 Let my check 2015-03-07T13:11:59 me* 2015-03-07T13:14:02 5-20150301 ? 2015-03-07T13:14:18 Try it .. I am guessing a new bset file 2015-03-07T13:15:15 bset? Maybe a new .cfg file at tools/ just for moxie, makes sense? 2015-03-07T13:15:29 And refer to it from .bset 2015-03-07T13:22:12 Do the least possible. Chris or Geare will review it. :) 2015-03-07T13:35:21 Ok, I understand that it's a temp work-around 2015-03-07T13:41:59 *** ragunath has left #rtems 2015-03-07T13:43:37 Building... Fingers crossed 2015-03-07T13:47:42 I don't know if there is a GCC PR filed for the bfin issue either. 2015-03-07T13:50:15 If it was fixed in the latest snapshot, I can edit the bfin .bset file too 2015-03-07T13:59:07 Well, gcc builds finally. No the problem is that rtems-tools fail to build 2015-03-07T13:59:31 Weird python error by any chance? 2015-03-07T14:01:05 It's something related to waf I think 2015-03-07T14:01:13 I'll pull and try again 2015-03-07T14:01:34 I had the mips and m68k fail with something similar and then just did it again and it worked 2015-03-07T14:02:14 Weird, hopefully this second try would make it. It just takes a lot of time 2015-03-07T14:02:53 Yeah.. with 8 bsps building, it took 75 minutes to do the m68k toolset. Lots of multilibs and sngle threaded 2015-03-07T14:07:21 *** namaste_ has joined #rtems 2015-03-07T14:11:35 hello, I am final year B.Tech student of Electronics and would like to work on the project beagle bsp improvements. So I have referred http://www.shrike-systems.com/beagleboard-xm-beaglebone-black-and-everything-else-rtems-on-the-beagles.html 2015-03-07T14:11:51 Now anyone guide how to proceed further? 2015-03-07T14:19:37 how far did you get? 2015-03-07T14:19:48 beng-nl: is the author of that BSP and that's his website 2015-03-07T14:23:00 DrJoel Finally it's built! 2015-03-07T14:24:45 Can you build RTEMS? That's always the rub. 2015-03-07T14:25:10 Great news though. That is one of the two tickets listed as blockers before branching 2015-03-07T14:25:36 namaste_: Do you have a BB? 2015-03-07T14:25:43 ya I do haave 2015-03-07T14:25:48 *have 2015-03-07T14:25:58 Did you run ticker? hello? 2015-03-07T14:26:30 Actually I gone through the procedure 2015-03-07T14:26:39 and I am going to build it 2015-03-07T14:27:09 but I want to ask about the further aspects of this project 2015-03-07T14:27:21 OK. After you do the hello world project, please put your name in this list https://devel.rtems.org/wiki/GSoC/2015 2015-03-07T14:27:39 Sure ask away. I assume beng-nl is having a nice night in Amsterdam and not really online 2015-03-07T14:29:45 DrJoel: configure: error: C compiler cannot create executables :( 2015-03-07T14:30:01 try compiling "int main() {return 0;} 2015-03-07T14:30:16 Ok 2015-03-07T14:30:42 @DrJoel : Peripheral support for BBB is the final goal of this project. right? 2015-03-07T14:30:42 namaste_: Error: "DrJoel" is not a valid command. 2015-03-07T14:31:16 DrJoel : Peripheral support for BBB is the final goal of this project. right? 2015-03-07T14:31:36 DrJoel: bin/ld crt0.o: Relocations in generic ELF (EM: 223) 2015-03-07T14:31:44 rt0.o: error adding symbols: File in wrong format 2015-03-07T14:32:25 on BBB. yes.. we want more peripherals supported. I know the MMC and NIC currently have no drivers 2015-03-07T14:32:41 Hesham: add a -v to that command line and see where crt0 came from 2015-03-07T14:33:58 DrJoel: i am here as it happens :) 2015-03-07T14:34:39 namaste_: how far are you? 2015-03-07T14:35:24 didn't get your question 2015-03-07T14:35:41 beng-nl 2015-03-07T14:36:10 namaste_: you were asking how to proceed further; does that mean you have rtems running on the bbxm emulator and/or beaglebone hardware? 2015-03-07T14:37:20 DrJoel: Not sure, does this relate to ctr0? > /tmp/ccrTgNJZ.o -lgcc -lc -lgcc 2015-03-07T14:38:46 beng-nl: Till now I haven't built rtems on BBB. But I cross compiled vanilla kernel for BBB and it was running successfully. So I do have the experience about all the steps 2015-03-07T14:38:48 no, crt0.o should be explicitly mentioned, like /usr/lib/crt0.o (but not the native location ofcourse) (but that might be the problem) 2015-03-07T14:39:06 and it should be on the ld command line somewhere, and that should be printed indeed when gcc invoked with -v 2015-03-07T14:39:18 namaste_: okay 2015-03-07T14:39:36 namaste_: yeah i meant crossbuilding 2015-03-07T14:39:44 and I do have basic knowledge of embeded systems 2015-03-07T14:39:57 namaste_: so you could do the hello world project as described by the rtems gsoc getting started page? 2015-03-07T14:40:13 ya 2015-03-07T14:40:16 I will 2015-03-07T14:41:00 I was also trying to have xenomi kernel for BBB 2015-03-07T14:41:29 * xenomai 2015-03-07T14:41:50 are they a gsoc project this year also? 2015-03-07T14:42:25 Don't have idea 2015-03-07T14:42:40 As I my project is on quadrotor 2015-03-07T14:43:00 I have come to know about BeaglePilot 2015-03-07T14:43:11 So tried that stuff 2015-03-07T14:44:03 :) 2015-03-07T14:45:19 beng-nl: Reference http://dev.ardupilot.com/wiki/building-the-code/building-for-beaglebone-black-on-linux/ 2015-03-07T14:46:34 /home/hesham/development/rtems/lib/gcc/moxie-rtems4.11/5.0.0/../../../../moxie-rtems4.11/lib/crt0.o 2015-03-07T14:46:55 lib/gcc ? 2015-03-07T14:49:12 I have to go home now, will come back within an hour or so 2015-03-07T14:49:56 namaste_: ok, so the first steps are to build rtems for bbxm and/or bbb. you can see my webpage & the rtems wiki page for steps. 2015-03-07T14:50:15 my page you already have; the rtems wiki page is https://devel.rtems.org/wiki/Developer/Projects/Open/ImproveBeagleBSP 2015-03-07T14:50:36 it tells you step by step what to do to prepare for a gsoc proposal 2015-03-07T14:50:50 ya already opened 2015-03-07T14:50:54 okay 2015-03-07T14:50:59 * DrJoel is looking forward to http://peterandthestarcatcher.com/ tonight :) 2015-03-07T14:51:07 following that blog 2015-03-07T14:51:17 namaste_: cool. let me know if you need help. 2015-03-07T14:51:28 sure 2015-03-07T14:52:04 it is a prequel to Peter Pan.. seeing touring Broadway show tonight! :) 2015-03-07T14:52:12 wow 2015-03-07T14:52:20 a RL play, great 2015-03-07T14:53:45 yeah! We have had season tickets to the local Broadway Theater League for over 15 years. Tonight we get a before show meet and greet along with an after the play dinner/party with case. 2015-03-07T14:53:55 *** Hesham has quit IRC (Ping timeout: 256 seconds) 2015-03-07T14:54:04 cast. 2015-03-07T14:55:55 oh wow.. fantastic 2015-03-07T15:03:15 we donate a bit each year to ensure good seats, reserved area for parking, before show reception, ... :) 2015-03-07T15:03:23 I love the shows 2015-03-07T15:03:51 I have seen Spamalot three times now. Twice during one weekend with Richard Chamberlain as Arthur 2015-03-07T15:04:16 *** namaste_ has left #rtems 2015-03-07T15:18:59 *** namaste_ has joined #rtems 2015-03-07T15:22:10 beng-nl https://www.rtems.org/ftp/pub/rtems/people/chrisj/source-builder/source-builder.html#_host_setups : page not found 2015-03-07T15:26:40 heading away from the keyboard 2015-03-07T16:12:18 namaste_: hm-hm 2015-03-07T17:02:14 *** Hesham has joined #rtems 2015-03-07T17:02:54 Sorry I met a fiend and spent ah hour talking. 2015-03-07T17:03:06 Any hints for the ctr0.o issue? 2015-03-07T17:05:12 add a -v to the moxie gcc invocation and see where it is getting it. What as it is using, etc 2015-03-07T17:05:29 FYI this is the warnings report for today without tests or networking: 2015-03-07T17:05:30 Unique Warnings : 99 2015-03-07T17:05:31 BSPs : 190 2015-03-07T17:05:31 BSPs with Zero : 1 2015-03-07T17:05:31 BSPs with only in shared: 180 2015-03-07T17:06:09 My last thing to send here where the place it gets crt0 from 2015-03-07T17:06:12 I think I have a couple of patches to knock some more out 2015-03-07T17:06:18 /home/hesham/development/rtems/lib/gcc/moxie-rtems4.11/5.0.0/../../../../moxie-rtems4.11/lib/crt0.o 2015-03-07T17:06:28 sorry... missed that 2015-03-07T17:06:31 do a file on it. 2015-03-07T17:06:44 do a moxie-rtems4.11-gcc -c on the main.c and see what type of file is it 2015-03-07T17:07:07 Ok 2015-03-07T17:07:34 And check what symbols are in the crt0.o vs your or1k tools. We stub a lot of symbols to make autoconf happy during probes. So if it is the real rtems one, it will have the same symbols 2015-03-07T17:07:40 about to leave for play. 2015-03-07T17:12:16 Class: ELF32 2015-03-07T17:15:14 Which library should it crt0 be at? 2015-03-07T17:15:18 libgcc.a ? 2015-03-07T17:17:03 I can't find crt0 standalone file 2015-03-07T17:17:30 Ah that's it 2015-03-07T17:20:12 A lot of system calls symbols and RTEMS stubs are out there 2015-03-07T17:28:16 *** namaste_ has quit IRC (Ping timeout: 246 seconds) 2015-03-07T17:32:14 mmm, I have to sleep now, will work on it tomorrow. P.S, I can't see the IRC logs.. 2015-03-07T17:32:24 Good night everyone! 2015-03-07T18:03:02 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2015-03-07T18:37:19 *** Davidbrcz has joined #rtems 2015-03-07T19:24:38 *** Davidbrcz has quit IRC (Ping timeout: 244 seconds) 2015-03-07T19:30:50 *** altairpearl has joined #rtems 2015-03-07T19:36:00 verm__, the feature set in libdl was not the issue 2015-03-07T19:36:10 verm__, at the time of putting in a number of tool sets did not build 2015-03-07T19:36:48 Hesham, I think a new version will need a new .cfg file in rtems/tools. 2015-03-07T19:37:12 Hesham, just catching up, what is being fixed ? 2015-03-07T19:37:32 A complete tools test build can be run with '../source-builder/sb-set-builder --prefix=/no-where-real --regression 4.11/rtems-all' 2015-03-07T19:38:14 Back in a few hours ... 2015-03-07T19:54:00 he's trying to fix the moxie tools 2015-03-07T19:54:13 there are two tickets owned by you that need fixing, #2205 and #2254 -- gedare has a fix for #2265 already 2015-03-07T19:54:21 after that all our tickets will be done! 2015-03-07T21:06:13 *** jhollister has quit IRC (Ping timeout: 256 seconds) 2015-03-08T00:12:57 *** jarie has joined #rtems 2015-03-08T01:38:58 *** dr__house is now known as dr__house|away 2015-03-08T01:51:06 *** jarie has quit IRC (Quit: nah nah nah nah, nah nah nah nah, hey hey hey, goodbye) 2015-03-08T01:52:32 *** jarie_ has joined #rtems 2015-03-08T01:57:03 *** jarie_ has quit IRC (Ping timeout: 252 seconds) 2015-03-08T03:06:38 *** mkhoory has joined #rtems 2015-03-08T03:09:35 *** mkhoory has quit IRC (Read error: Connection reset by peer) 2015-03-08T03:10:01 *** mkhoory has joined #rtems 2015-03-08T03:20:11 *** altairpearl has quit IRC (Remote host closed the connection) 2015-03-08T04:09:03 *** altairpearl has joined #rtems 2015-03-08T05:07:20 *** Hesham has joined #rtems 2015-03-08T05:17:32 [trac] #2208/enhancement (closed) updated by chrisj (How do we make the 4.11 release ? Can this script be used and so should be put in ...) http://devel.rtems.org/ticket/2208 2015-03-08T05:21:13 *** Davidbrcz has joined #rtems 2015-03-08T05:24:49 Interesting, tried to build moxie tools again, and it fails with the new assembly code. 2015-03-08T05:46:46 *** Davidbrcz has quit IRC (Ping timeout: 244 seconds) 2015-03-08T06:34:42 *** gedare has joined #rtems 2015-03-08T06:34:42 *** ChanServ sets mode: +o gedare 2015-03-08T06:53:20 *** gedare has quit IRC (Ping timeout: 272 seconds) 2015-03-08T07:38:29 *** ragunath has joined #rtems 2015-03-08T07:41:06 *** waleed has joined #rtems 2015-03-08T07:56:58 *** Davidbrcz has joined #rtems 2015-03-08T08:45:00 *** Davidbrcz_ has joined #rtems 2015-03-08T09:16:33 [trac] #2208/enhancement (closed) updated by amar (I was planning on doing that once we decided it produced a good result.) http://devel.rtems.org/ticket/2208 2015-03-08T09:52:16 *** Hesham has quit IRC (Ping timeout: 252 seconds) 2015-03-08T10:25:37 *** Hesham has joined #rtems 2015-03-08T10:40:41 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2015-03-08T11:32:22 *** ragunath has quit IRC (Ping timeout: 246 seconds) 2015-03-08T11:45:06 *** altairpearl has quit IRC (Ping timeout: 246 seconds) 2015-03-08T11:47:03 *** altairpearl has joined #rtems 2015-03-08T11:56:37 *** jarie_ has joined #rtems 2015-03-08T12:12:03 *** jarie_ has quit IRC (Ping timeout: 246 seconds) 2015-03-08T12:34:45 *** altairpearl has quit IRC (Read error: Connection reset by peer) 2015-03-08T12:52:19 *** altairpearl has joined #rtems 2015-03-08T13:06:00 *** Hesham has joined #rtems 2015-03-08T13:06:47 [trac] #2286/defect (v:4.10) created by joel.sherrill (cpustdatomic.h on 16 and 64 bit architectures) http://devel.rtems.org/ticket/2286 2015-03-08T13:09:33 [trac] #2286/defect (new) updated by joel.sherrill (empty comment) http://devel.rtems.org/ticket/2286 2015-03-08T13:48:56 *** ragunath has joined #rtems 2015-03-08T13:50:55 *** ragunath has left #rtems 2015-03-08T14:08:47 *** Davidbrcz has quit IRC (Ping timeout: 252 seconds) 2015-03-08T14:08:49 *** Davidbrcz_ has quit IRC (Ping timeout: 264 seconds) 2015-03-08T14:30:41 *** dr__house|away is now known as dr__house 2015-03-08T15:56:58 *** kiwichris has quit IRC (Ping timeout: 255 seconds) 2015-03-08T15:59:33 *** namaste_ has joined #rtems 2015-03-08T16:05:43 *** namaste_ has quit IRC (Ping timeout: 246 seconds) 2015-03-08T16:14:03 *** kiwichris has joined #rtems 2015-03-08T16:14:49 *** waleed has quit IRC (Ping timeout: 264 seconds) 2015-03-08T16:15:31 *** Davidbrcz has joined #rtems 2015-03-08T16:16:01 *** Davidbrcz_ has joined #rtems 2015-03-08T16:26:32 [trac] #2287/defect (v:4.11) created by joel.sherrill (RTEMS printf warnings - newlib's inttypes.h and gcc's newlib-stdint.h) http://devel.rtems.org/ticket/2287 2015-03-08T16:36:04 [trac] #2288/defect (v:4.11) created by joel.sherrill (Use of assert()) http://devel.rtems.org/ticket/2288 2015-03-08T16:37:56 *** mkhoory has quit IRC (Read error: Connection reset by peer) 2015-03-08T16:38:20 *** mkhoory has joined #rtems 2015-03-08T16:45:53 *** mkhoory has quit IRC (Read error: Connection reset by peer) 2015-03-08T16:46:43 *** mkhoory has joined #rtems 2015-03-08T16:55:06 *** titankiller has joined #rtems 2015-03-08T17:04:17 kiwichris warnings are dropping fast. The inttypes.h/stdint-newlib.h issue is now looking like one of the biggest sources on 32 bit targets 2015-03-08T17:04:54 I tagged gedare for a cpustdatomic.h issue on 16 and 64 bit targets. One issue that puts >900 warnings in a build log 2015-03-08T17:07:13 *** Hesham has quit IRC (Ping timeout: 244 seconds) 2015-03-08T17:22:53 [trac] #2200/defect (new) updated by chrisj (Can we rsync those logs over to the servers and at least have something available ?) http://devel.rtems.org/ticket/2200 2015-03-08T17:35:32 *** titankiller has quit IRC (Ping timeout: 272 seconds) 2015-03-08T17:46:26 *** Davidbrcz has quit IRC (Ping timeout: 256 seconds) 2015-03-08T17:46:37 *** Davidbrcz_ has quit IRC (Ping timeout: 264 seconds) 2015-03-08T17:54:18 DrJoel_, nice work 2015-03-08T17:54:43 DrJoel, I have been sorting through a number of things I have had on hold 2015-03-08T17:55:42 Thanks.. I think we need to step back and evaluate if the tickets captured all the pre-branching/release activity. It is close but we have played catch up 2015-03-08T18:12:32 kiwichris can you take a look at this warning? I am guessing you will know what to do based on other architectures 2015-03-08T18:12:34 log/m32r-m32rsim.log:../../../../../../rtems/c/src/../../cpukit/libdl/rtl-mdreloc-m32r.c:72:44: warning: suggest parentheses around arithmetic in operand of '|' [-Wparentheses] 2015-03-08T18:14:11 But our coding standard says too many paren's are bad ;) 2015-03-08T18:14:24 lol 2015-03-08T18:15:33 not enough is also bad. ;) 2015-03-08T18:15:40 under 90 warnings now 2015-03-08T18:15:58 I do not agree with that coding standard 2015-03-08T18:16:12 no parens? 2015-03-08T18:16:15 on expressions? 2015-03-08T18:16:33 you can always pipe up on the thread. :) 2015-03-08T18:16:47 I feel explicit is better as it describes the intention rather than assume the programmed intended the default precidence 2015-03-08T18:17:12 Been too busy it enter into debates about coding standards 2015-03-08T18:17:17 Bottomless pits 2015-03-08T18:17:19 yeah.. people reading later don't know if you were smart or lucky 2015-03-08T18:17:35 Agreed 2015-03-08T18:17:39 I am just trying to plow through the all building, revisit warnings pass 2015-03-08T18:17:43 I will always add them 2015-03-08T18:21:53 The warning is valid and the (tmp & 0xfff....) should have parens 2015-03-08T18:22:20 see line 92 below 2015-03-08T18:24:36 thanks. I will likely push a handful of warning fixes in the next few hours 2015-03-08T18:26:52 Thanks for this. 2015-03-08T18:27:23 no big deal. 2015-03-08T18:28:04 we need to review the getting started doc for 4.11. Probably needs gutting but we need something 2015-03-08T19:20:43 DrJoel_, still about ? 2015-03-08T19:21:10 for not much longer.. was proofreading Michele's book and dinner is close 2015-03-08T19:21:12 but shoort 2015-03-08T19:26:26 We have found some interesting issues in mongoose in 4.11 related to the http auth handling. 2015-03-08T19:27:39 If it can't be fixed easily, monkey may be an option. THe licensing is a deadend on Mongoose unless you want to get a license 2015-03-08T19:28:20 Monkey for 4.11 ? 2015-03-08T19:28:51 We have mg in RTEMS 4.11 with known problems. It is about the state of what we provide in a release. 2015-03-08T19:29:07 We are looking into a solution. 2015-03-08T19:30:43 If possible we should look at a patch to fix mg to just clean it up and have it work as expected. It is a case of one browser not working as expected and it turns out mg is at fault 2015-03-08T19:32:02 :( we need to move o another httpd due to licensing but the timing is never great 2015-03-08T19:32:11 Hpefully you can fix it. 2015-03-08T19:32:13 dinner time 2015-03-08T19:32:57 cya 2015-03-08T20:36:03 [trac] #2200/defect (new) updated by amar (I'm going to move all of them soon enough. Again rsyncing will be just as much work ...) http://devel.rtems.org/ticket/2200 2015-03-08T20:41:42 *** jarie has joined #rtems 2015-03-08T20:52:42 Hello! Browsed through projects, Beagle BSP improvements caught my interest, currently trying to perform the prerequisites. 2015-03-08T20:54:12 i.e. trying to accomplish the 'Getting started as a GSOC student' section 2015-03-08T22:10:00 *** jarie__ has joined #rtems 2015-03-08T22:13:10 *** jarie__ has quit IRC (Remote host closed the connection) 2015-03-08T22:55:43 [trac] #2205/defect (new) updated by chrisj (I have run the rtems-tools from git in MSYS2 with a python.org python and it all ...) http://devel.rtems.org/ticket/2205 2015-03-08T23:00:15 [trac] #2254/defect (new) updated by chrisj (Where is a list of patches we need ? Can the patches be used with a valid gcc ...) http://devel.rtems.org/ticket/2254 2015-03-08T23:51:25 *** dr__house is now known as dr__house|away