2012-05-21T01:23:34 *** arvind_khadri has joined #rtems 2012-05-21T01:23:34 *** arvind_khadri has joined #rtems 2012-05-21T01:49:51 *** xiangfu has quit IRC (Ping timeout: 246 seconds) 2012-05-21T01:53:10 *** sebhub has joined #rtems 2012-05-21T01:53:48 good morning 2012-05-21T02:08:09 *** xiangfu has joined #rtems 2012-05-21T02:17:07 *** xian9fu has joined #rtems 2012-05-21T02:20:38 *** xiangfu has quit IRC (Ping timeout: 246 seconds) 2012-05-21T02:20:59 *** xian9fu is now known as xiangfu 2012-05-21T02:57:41 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-05-21T03:50:42 *** arvind_khadri has joined #rtems 2012-05-21T04:34:14 *** lcpfnvcy has joined #rtems 2012-05-21T05:24:06 *** xiangfu has quit IRC (Remote host closed the connection) 2012-05-21T07:42:00 *** jennifer has joined #rtems 2012-05-21T08:06:26 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-05-21T08:12:45 *** QingPei has joined #rtems 2012-05-21T08:26:21 hi 2012-05-21T08:30:08 hi 2012-05-21T08:54:38 *** weiY has joined #rtems 2012-05-21T09:00:50 *** weiY has quit IRC (Ping timeout: 260 seconds) 2012-05-21T09:08:28 *** weiY has joined #rtems 2012-05-21T09:47:03 *** weiY has quit IRC (Ping timeout: 252 seconds) 2012-05-21T09:48:55 *** arvind_khadri has joined #rtems 2012-05-21T09:54:54 *** weiY has joined #rtems 2012-05-21T10:47:50 *** arvind_khadri has quit IRC (Ping timeout: 256 seconds) 2012-05-21T10:51:11 *** sebhub has quit IRC (Read error: Operation timed out) 2012-05-21T10:58:29 *** weiY has quit IRC (Ping timeout: 245 seconds) 2012-05-21T10:59:21 *** verm__ has quit IRC (Read error: Connection reset by peer) 2012-05-21T11:16:51 *** QingPei has quit IRC (Ping timeout: 246 seconds) 2012-05-21T11:27:20 *** arvind_khadri has joined #rtems 2012-05-21T11:29:06 *** arvind_khadri has joined #rtems 2012-05-21T11:30:46 *** arvind_khadri has joined #rtems 2012-05-21T11:32:47 *** arvind_khadri has joined #rtems 2012-05-21T11:35:03 *** arvind_khadri has joined #rtems 2012-05-21T11:37:08 *** arvind_khadri has joined #rtems 2012-05-21T11:39:05 *** arvind_khadri has joined #rtems 2012-05-21T11:41:16 *** arvind_khadri has joined #rtems 2012-05-21T12:07:57 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-05-21T12:22:39 *** arvind_khadri has joined #rtems 2012-05-21T12:27:56 *** DrJoel has joined #rtems 2012-05-21T12:27:57 *** DrJoel has joined #rtems 2012-05-21T12:27:57 *** ChanServ sets mode: +o DrJoel 2012-05-21T12:31:34 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-05-21T13:07:55 *** verm__ has joined #rtems 2012-05-21T13:07:58 *** verm__ has joined #rtems 2012-05-21T15:28:22 DrJoel: nice thought with the gdb version.. it'd be interesting if that is the primary reason for their observed speedup. 2012-05-21T17:06:41 *** lcpfnvcy has quit IRC (Ping timeout: 252 seconds) 2012-05-21T18:07:52 gedare, thanks for following the atomic emails. 2012-05-21T19:10:11 *** deb_ has joined #rtems 2012-05-21T19:18:44 *** lcpfnvcy has joined #rtems 2012-05-21T20:01:33 *** xiangfu has joined #rtems 2012-05-21T20:48:26 kiwichris: glad to help students who seek help. :) 2012-05-21T21:12:02 *** deb_ has left #rtems ("Leaving") 2012-05-21T21:12:34 *** Deb has joined #rtems 2012-05-21T21:35:47 *** arvind_khadri has joined #rtems 2012-05-21T21:35:47 *** arvind_khadri has joined #rtems 2012-05-21T22:14:30 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-05-21T22:16:22 *** arvind_khadri has joined #rtems 2012-05-21T22:17:57 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-21T22:18:23 *** arvind_khadri has joined #rtems 2012-05-21T22:18:27 *** arvind_khadri has joined #rtems 2012-05-21T22:20:20 *** arvind_khadri has joined #rtems 2012-05-21T23:03:09 *** arvind_khadri has quit IRC (Ping timeout: 260 seconds) 2012-05-22T00:13:37 *** Deb has quit IRC (Quit: Leaving) 2012-05-22T00:58:03 *** sevikkk has quit IRC (Ping timeout: 244 seconds) 2012-05-22T01:00:16 *** arvind_khadri has joined #rtems 2012-05-22T01:03:59 *** sevikkk has joined #rtems 2012-05-22T01:09:16 *** arvind_k has joined #rtems 2012-05-22T01:13:37 *** arvind_k is now known as arvind_khadri 2012-05-22T01:13:54 *** arvind_khadri has joined #rtems 2012-05-22T01:52:47 *** sebhub has joined #rtems 2012-05-22T01:53:40 good morning 2012-05-22T02:21:07 *** arvind_khadri has quit IRC (Quit: Leaving) 2012-05-22T03:14:38 *** arvind_khadri has joined #rtems 2012-05-22T03:14:38 *** arvind_khadri has joined #rtems 2012-05-22T03:23:17 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-05-22T04:07:26 *** cdcs has joined #rtems 2012-05-22T04:27:05 *** arvind_khadri has joined #rtems 2012-05-22T05:08:43 *** arvind_khadri has quit IRC (Quit: Leaving) 2012-05-22T05:10:52 *** arvind_khadri has joined #rtems 2012-05-22T05:12:13 *** arvind_khadri has joined #rtems 2012-05-22T05:56:51 *** xiangfu has quit IRC (Ping timeout: 246 seconds) 2012-05-22T07:49:22 *** xiangfu has joined #rtems 2012-05-22T07:53:45 *** xiangfu has quit IRC (Ping timeout: 246 seconds) 2012-05-22T07:55:16 *** antgreen has joined #rtems 2012-05-22T08:05:38 *** xiangfu has joined #rtems 2012-05-22T08:28:45 *** lcpfnvcy has quit IRC (Ping timeout: 246 seconds) 2012-05-22T08:30:42 *** lcpfnvcy has joined #rtems 2012-05-22T08:45:43 * DrJoel realizes that he left yesterday without killing irc.. didn't feel well and went home 2012-05-22T09:06:17 *** gedare_ has joined #rtems 2012-05-22T09:10:12 *** QingPei has joined #rtems 2012-05-22T09:21:51 --reply-all: I vehemently disagree with this email. It is not only non-helpful, it is broken. 2012-05-22T09:24:45 *** xiangfu has quit IRC (Ping timeout: 246 seconds) 2012-05-22T09:28:56 broken is the C programming language since is has no templates ;-) 2012-05-22T09:29:08 heh 2012-05-22T09:30:02 in C11 we have generic macros for more fun with the pre-processor 2012-05-22T09:30:19 i haven't studed c11 yet 2012-05-22T09:30:22 *studied 2012-05-22T09:30:55 it made some common compiler extensions the standard 2012-05-22T09:38:14 gedare: +1 2012-05-22T09:38:40 I vehemently disagree with your post. And your mother dresses you funny 2012-05-22T10:21:08 *** arvind_khadri has quit IRC (Ping timeout: 240 seconds) 2012-05-22T10:27:12 ok, FINALLY the updater is working like it should 2012-05-22T10:27:24 we have realtime updates to github 2012-05-22T10:27:55 i enver did find out what the problem was, i ended up having to add some extra things to make it work on redhat.. not sure what's so different than other operating systems 2012-05-22T10:31:03 yes, looks pretty up to date, nice work 2012-05-22T10:31:33 it'll update after every commit now 2012-05-22T10:34:31 i'll create the teams shortly 2012-05-22T10:35:12 I'm curious... The sync was set with a post-receive-hook on git.rtems repo? 2012-05-22T10:35:50 yes and a custom async daemon i wrote using twisted 2012-05-22T10:36:16 post-receive-hook pings the daemon which runs the update, it can run in parallel as well for commits across different repositories at the same time 2012-05-22T10:38:05 hum... i thought that a "git push --quiet user@github.com/repo.git &" was enough for that 2012-05-22T10:38:34 not sure, i'd never do it that way anyway 2012-05-22T10:39:05 you shouldn't launch background processes from a commit there's no way for you to control what it does 2012-05-22T10:41:29 in theory it should work, but i understand that having no output or control over the background process is a liability. 2012-05-22T10:43:41 hmm i've never tried it but i'm skeptical that it will work properly 2012-05-22T10:44:01 launching a background process from a non-pty shell underneath git seems like asking for trouble 2012-05-22T10:44:26 well the pty isn't nessicary but underneath git.. 2012-05-22T10:49:31 don't know yet. i will try it to sync my local repo with a remote bitbucket for backup purposes 2012-05-22T10:50:40 *** sebhub has quit IRC (Read error: Operation timed out) 2012-05-22T10:50:48 i would think that it will hang the git process until it finishes 2012-05-22T10:51:02 which can be a long time when github is laggy 2012-05-22T10:51:58 i only use the post-receive to run a python script to sync the repo with TRAC, but since it is a local service it doesn't have those lag issues 2012-05-22T10:02:08 *** rtemsLogger has joined #rtems 2012-05-22T11:15:45 yeah, as long as your trac instance isn't busy that works 2012-05-22T11:19:11 *** cdcs has quit IRC (Ping timeout: 245 seconds) 2012-05-22T11:23:55 *** Hesham has joined #rtems 2012-05-22T11:25:20 hi gedare 2012-05-22T12:06:12 *** alseh has joined #rtems 2012-05-22T12:06:28 hi hesham 2012-05-22T12:06:36 i was at lunch :) 2012-05-22T12:06:54 hope you enjoyed it :) 2012-05-22T12:07:02 i am working on the Doc now 2012-05-22T12:07:59 ok 2012-05-22T12:08:45 let's chat in about 5 min? 2012-05-22T12:16:58 Ok 2012-05-22T12:23:44 Hesham: ready when you are 2012-05-22T12:24:16 i edited the introduction part according to your comments 2012-05-22T12:24:26 it's looking much better don't you think? 2012-05-22T12:24:34 sure 2012-05-22T12:24:47 do you have a clearer picture of what you will be doing? or were you pretty clear before? 2012-05-22T12:25:23 yeah but need to define the attributes 2012-05-22T12:26:04 to make the picture clearer 2012-05-22T12:26:16 i could work on the interfaces now 2012-05-22T12:26:17 that can come. i think there was an attempt at it in the gsoc2011 repository, and then there was some feedback in rtems-devel ml suggesting we use some standard terms (I forget exactly) 2012-05-22T12:26:21 ok super 2012-05-22T12:26:43 attributes is just a set of bits / flags, pretty simple 2012-05-22T12:27:03 there are at least 5 useful attributes that I'm aware... r/w/x, cacheable, shareable 2012-05-22T12:27:12 shareable has to do with SMP code. 2012-05-22T12:27:19 rather, SMP code accessing shared memory 2012-05-22T12:27:58 5 access attributes then 2012-05-22T12:28:20 yeah but don't be limited we can use an integer or enum type 2012-05-22T12:28:31 it should be easy to add more attributes if we find we need to do so 2012-05-22T12:28:34 could be extended later easily 2012-05-22T12:28:37 ya 2012-05-22T12:28:40 good 2012-05-22T12:29:00 do you have other questions from the changes that you made so far? 2012-05-22T12:29:15 yeah 2012-05-22T12:29:35 Should the part of attached tasks be deleted ? 2012-05-22T12:30:06 You said there is no need to notifies attached tasks to an Arena that it's deleted 2012-05-22T12:30:17 notify* 2012-05-22T12:30:24 um, notify has a precise meaning usually 2012-05-22T12:30:40 it indicates that the task gets some (a)synchronous message telling it to do something 2012-05-22T12:30:50 what's the proper term to use ? 2012-05-22T12:31:09 there is no reason to "notify", but for some operations there is a need to update the task's state due to changes in arenas 2012-05-22T12:31:34 for example if an arena is deleted you should be able to update all of it's attached tasks so that they no longer have the arena in their TCB->arenas list. 2012-05-22T12:31:51 so that the context switch code does not attempt to (de)activate a deleted arena, right? 2012-05-22T12:31:52 good 2012-05-22T12:32:04 yeah 2012-05-22T12:32:13 that would be a specific requirement that necessitates arena to have some way to iterate its attached tasks 2012-05-22T12:32:14 i will edit this on the doc 2012-05-22T12:32:17 ok good 2012-05-22T12:32:35 did you reviewed it ? 2012-05-22T12:33:07 i skimmed through it. the contents are much improved. I will help you a little with formatting and english usage before you send it out 2012-05-22T12:33:46 we will send the introduction only or what ? 2012-05-22T12:34:05 i think you can remove everything between the link [1] through the start of interfaces. the proposed interface will be good to include 2012-05-22T12:34:23 move the other material into some notes section that won't get sent out 2012-05-22T12:34:46 i think that material in the middle is mostly redundant now 2012-05-22T12:34:59 or is more detail than necessary for the first description 2012-05-22T12:35:04 yeah i did not intend to send it anyway 2012-05-22T12:35:18 sounds good. i think the interface will be good to send 2012-05-22T12:35:33 once you get a chance to clean it up and add some more description about each function 2012-05-22T12:35:45 it was some of previous initial work when the picture was not clear right now 2012-05-22T12:35:52 * gedare_ nodes 2012-05-22T12:35:55 * gedare_ nods too 2012-05-22T12:36:27 i will edit it now 2012-05-22T12:37:35 ok. any other questions / comments? 2012-05-22T12:38:13 i think there is no technical questions 2012-05-22T12:38:20 but when we will send it ? 2012-05-22T12:40:02 after you revise the interface and we clean up the intro for english usage. depends on how long that takes 2012-05-22T12:41:39 i am working on it now 2012-05-22T12:55:19 *** rtemsLogger` has joined #rtems 2012-05-22T12:55:34 *** rtemsLogger has quit IRC (Ping timeout: 245 seconds) 2012-05-22T13:10:38 Hesham: I have to go to another meeting 2012-05-22T13:11:07 ok i will work on the doc & interfaces , review them when you are free 2012-05-22T13:11:25 i'll check your changes later. and i added more notes for the intro. i'd like you to break it into 3 paragraphs, i marked them 2012-05-22T13:11:45 ok 2012-05-22T13:21:41 *** QingPei has left #rtems 2012-05-22T14:32:54 *** antgreen has quit IRC (Read error: Connection reset by peer) 2012-05-22T14:49:26 Hesham: i gave you a few more comments. headed home. i will read it more thoroughly tonight. maybe you can send it out tomorrow. 2012-05-22T14:49:48 *** gedare_ has quit IRC (Quit: Ex-Chat) 2012-05-22T16:14:38 *** alseh has quit IRC (Remote host closed the connection) 2012-05-22T16:42:38 *** Hesham has quit IRC (Quit: Leaving.) 2012-05-22T16:51:25 Morning 2012-05-22T16:57:42 DrJoel, about ? 2012-05-22T19:08:57 morning kiwichris 2012-05-22T19:14:43 gedare, hi 2012-05-22T19:15:09 how's it going 2012-05-22T19:15:28 Good. I have the runtime loader in a personal git repo now 2012-05-22T19:15:53 Almost have the allocator changes done and just the external symbols to sort 2012-05-22T19:16:12 nice 2012-05-22T19:16:31 good progress. can't wait to see it get in the tree 2012-05-22T19:16:32 With the externals sorted we can load dependent object files. 2012-05-22T19:16:43 and not have holes in the heap 2012-05-22T19:17:17 The allocator supports READ, READ_WRITE, and READ_EXECUTE requests 2012-05-22T19:17:23 i'm not particularly aware of the problems as they exist ;) 2012-05-22T19:18:04 If object A needs object B and B needs A you have to load one with an unresolved external 2012-05-22T19:18:11 hmm, this is to allocate shared objects like libs? or general heap? 2012-05-22T19:18:33 The loader needs to keep the symbol around to check if a new obj resolves it. 2012-05-22T19:18:35 ahhh, I see. 2012-05-22T19:18:46 Once resolved the external can be thrown away 2012-05-22T19:19:22 It is an interface in the loader that can be hooked and a custom allocator used 2012-05-22T19:19:35 The default uses the C heap 2012-05-22T19:20:05 interesting 2012-05-22T19:20:24 http://git.rtems.org/chrisj/rtl.git/tree/rtl-allocator.h 2012-05-22T19:20:43 This is old now cause I have changed the tags and added new features 2012-05-22T19:21:06 i see the shape of things 2012-05-22T19:21:57 this reminds me of a request for a feature in the memory protection work... something about mapping objects with names + protection 2012-05-22T19:22:33 pretty sure it can be handled separately, probably something like this rtl approach (even simpler i bet) coupled with a flexible "Arena" ;) 2012-05-22T19:23:27 I think so. I have removed the strings tag cause I can handle them better with a single table per module. 2012-05-22T19:23:37 i had my student looking at ways to make allocation more uniform in rtems, but abandoned it in order to focus on the arena design for nwo 2012-05-22T19:23:38 The externals will be indirect allocations 2012-05-22T19:24:01 ah. 2012-05-22T19:24:20 It will be a simple page type allocation then broken up and compacted etc 2012-05-22T19:24:43 The number should be low 2012-05-22T19:25:02 but the aim is no holes in the heap once all the code is loaded. 2012-05-22T19:25:10 well hole added by the loader 2012-05-22T19:25:51 Worth cloning and building just to see how waf handles an RTEMS app 2012-05-22T19:25:54 yeah it sounds nice, efficient. 2012-05-22T19:26:13 yeah i glanced at earlier versions you had, i spent a day catching up on waf 2012-05-22T19:27:03 There is also an mmap implementation in the code 2012-05-22T19:29:17 nice. there was a bit of a port in the mmu work before but i'm not sure about it. 2012-05-22T19:32:29 I tested this version with the NetBSD link editor 2012-05-22T19:57:13 *** xiangfu has joined #rtems 2012-05-22T20:23:20 *** gedare has quit IRC (Quit: Leaving) 2012-05-22T22:12:12 *** QingPei has joined #rtems 2012-05-23T01:48:39 *** sebhub has joined #rtems 2012-05-23T01:50:00 good morning 2012-05-23T02:50:38 *** xiangfu has quit IRC (Ping timeout: 240 seconds) 2012-05-23T03:03:11 *** xiangfu has joined #rtems 2012-05-23T03:40:58 *** xiangfu has quit IRC (Ping timeout: 276 seconds) 2012-05-23T03:52:58 *** xiangfu has joined #rtems 2012-05-23T04:00:58 *** QingPei has left #rtems 2012-05-23T04:13:49 *** A0Sheds has quit IRC (Ping timeout: 244 seconds) 2012-05-23T04:26:46 *** A0Sheds has joined #rtems 2012-05-23T05:27:08 *** sebhub has quit IRC (Ping timeout: 245 seconds) 2012-05-23T06:09:14 *** Deb has joined #rtems 2012-05-23T06:32:45 *** Deb has quit IRC (Quit: This computer has gone to sleep) 2012-05-23T06:48:11 *** arvind_khadri has joined #rtems 2012-05-23T06:59:38 *** xiangfu has quit IRC (Ping timeout: 240 seconds) 2012-05-23T07:01:59 *** xiangfu has joined #rtems 2012-05-23T08:22:44 *** zw_yao has joined #rtems 2012-05-23T08:31:08 *** cdcs has joined #rtems 2012-05-23T08:32:06 * DrJoel really needs to remember to log off before going home 2012-05-23T08:36:55 *** weiY has joined #rtems 2012-05-23T08:52:52 *** QingPei has joined #rtems 2012-05-23T09:04:24 Hi all. i encounter some error when i git clone the rep from github 2012-05-23T09:04:42 Cloning into rtems... 2012-05-23T09:04:46 The authenticity of host 'github.com (207.97.227.239)' can't be established. 2012-05-23T09:04:46 RSA key fingerprint is 16:27:ac:a5:76:28:2d:36:63:1b:56:4d:eb:df:a6:48. 2012-05-23T09:04:46 Are you sure you want to continue connecting (yes/no)? yes 2012-05-23T09:04:46 Warning: Permanently added 'github.com,207.97.227.239' (RSA) to the list of known hosts. 2012-05-23T09:04:46 Permission denied (publickey). 2012-05-23T09:04:47 fatal: The remote end hung up unexpectedly 2012-05-23T09:04:59 anyone can help me? 2012-05-23T09:12:59 *** gedare has joined #rtems 2012-05-23T09:13:11 Hello weiY 2012-05-23T09:13:21 git clone git://github.com/RTEMS/rtems.git 2012-05-23T09:13:25 is working for me 2012-05-23T09:14:21 are you in China? It could be the "Great Firewall" 2012-05-23T09:15:05 No, the GFW does not block github 2012-05-23T09:15:46 yes i am from china. but i think i seems the problem of ssh 2012-05-23T09:17:03 Maybe your certificates? 2012-05-23T09:17:25 it ask me passwd for openssh 2012-05-23T09:17:41 what is the passwd? 2012-05-23T09:19:12 awkward... What is the command you are executing? just git clone git://github.com/RTEMS/rtems.git? 2012-05-23T09:19:16 the passwd of github user? 2012-05-23T09:19:29 it seems that you are trying to access your private repo 2012-05-23T09:19:37 yeah, it is really the passwd of my user of github 2012-05-23T09:19:48 haha, not i git clone the repo which i fork from github/rtems 2012-05-23T09:19:56 that makes sense :) 2012-05-23T09:20:00 i think 2012-05-23T09:20:00 yeah 2012-05-23T09:20:18 but i can not git clone using ssh 2012-05-23T09:20:55 *** xiangfu has quit IRC (Ping timeout: 276 seconds) 2012-05-23T09:20:55 using http is ok for me 2012-05-23T09:21:29 yeah your were cloning from your personal repo. The password is the one you specified when you created the ssh key 2012-05-23T09:22:06 you were* 2012-05-23T09:22:10 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2012-05-23T09:22:49 when i created the ssh key? sign up github? 2012-05-23T09:23:45 the github is so slow. my speed is just 7KiB/s 2012-05-23T09:24:02 No. to acess your repos in github you need to add one ssh key to your github account 2012-05-23T09:24:07 http://help.github.com/linux-set-up-git/ 2012-05-23T09:24:21 read the "Next: Set Up SSH Keys" part 2012-05-23T09:26:15 the only thing i can do is staying 2012-05-23T09:27:53 weiY, Out of curiousity, How fast is the connection to rtems.org for you? 2012-05-23T09:29:11 joel, wait for a moment, i will test clone a repo from rtems.org 2012-05-23T09:30:52 cdcs, thank you. i will read this part 2012-05-23T09:38:02 Hi joel. i think the result is disappointed. until now it is still compressing objects 2012-05-23T09:39:15 remote: Counting objects: 404624, done. 2012-05-23T09:39:15 remote: Compressing objects: 53% (39366/74274) 2012-05-23T09:39:15 remote: Compressing objects: 100% (74274/74274), done. 2012-05-23T09:39:15 it display this picture all along 2012-05-23T09:39:43 weiY, No problem. I tried and that problem occurs when you try to clone your personal repo and you don't have a sshkey added to your github account. So just add one. 2012-05-23T09:41:41 weiY, very slow here too(I'm also in China). about 10KiB/s, I think I'll finish the clone after sleep. 2012-05-23T09:42:18 ok, i just read the help materials you gived me. i said that i clone from rtems.org is still slow 2012-05-23T09:43:03 Just to compare: clone from github 360 Kib/s; clone from git.rtems.org ~600Kib/s 2012-05-23T09:43:04 zw_yao, yes. i will hung it till tommorrow 2012-05-23T09:44:32 so envy your internet 2012-05-23T09:44:32 you are an aircraft and i am just a snails 2012-05-23T09:44:41 FWIW I am inside the network so on a LAN with git.rtems.org 2012-05-23T09:44:57 Receiving objects: 100% (404624/404624), 53.76 MiB | 2.56 MiB/s, done. 2012-05-23T09:45:43 you are an beam 2012-05-23T09:47:03 i clone from github is 7-10KiB/s, clone from rtems.org is 1KiB/s. so slow 2012-05-23T09:48:00 You guys' speed is amazing~ 2012-05-23T09:48:00 i donnot know whether i will clone completely until i am wake up tommorrow 2012-05-23T09:49:30 *** zw_yao has quit IRC (Remote host closed the connection) 2012-05-23T09:52:23 *** zw_yao has joined #rtems 2012-05-23T10:05:11 *** zw_yao has quit IRC (Quit: Leaving) 2012-05-23T10:16:58 *** RayX has joined #rtems 2012-05-23T10:47:13 *** Hesham has joined #rtems 2012-05-23T10:49:54 gedare: Could i send the design to rtems-devel ml now or what ? 2012-05-23T10:50:01 hi Hesham 2012-05-23T10:50:05 hi gedare 2012-05-23T10:50:29 fix the duplication of Arena_Create in your interface 2012-05-23T10:50:52 what is Get_boundary? boundary is a new term... 2012-05-23T10:51:02 update every variable named "attrib" to be attribute 2012-05-23T10:51:41 i have re-named the next one to be Arena_Create_with_allocation 2012-05-23T10:51:52 i see, that should go under optional? 2012-05-23T10:52:00 yeah 2012-05-23T10:52:05 ok 2012-05-23T10:52:08 *** weiY has quit IRC (Ping timeout: 240 seconds) 2012-05-23T10:52:28 for internal api we don't need to return status code. status code would be used for an "rtems_arena_xxx" api 2012-05-23T10:52:31 Get_boundary just get the starting address and Arena size 2012-05-23T10:52:47 split that into two functions, one to get the start address the other to get the size 2012-05-23T10:52:56 ok 2012-05-23T10:53:05 and return those values directly instead of via pointers 2012-05-23T10:53:17 ok 2012-05-23T10:53:30 same with the other functions, decide what they return, if nothing then make them void 2012-05-23T10:53:31 and functions that does not have return value would be void ? 2012-05-23T10:53:39 ok 2012-05-23T10:53:55 and give a one-line comment for each function to describe what it does 2012-05-23T10:54:18 i intend to do that 2012-05-23T10:54:25 *** jennifer has quit IRC (Quit: Leaving) 2012-05-23T10:54:37 why did you revert _Arena_Detach_all_tasks? 2012-05-23T10:55:43 maybe that function is not even necessary 2012-05-23T10:56:09 all you do it ... for tcb in acb->tasks: arena_detach(acb, tcb) 2012-05-23T10:56:23 could be used before deleting Arena 2012-05-23T10:56:37 i'm not sure it needs to be done more than just in delete itself 2012-05-23T10:56:45 anyways thats fine 2012-05-23T10:56:53 i don't see it as ever making it into the public interface though 2012-05-23T10:57:05 seems like a private helper function for your implementation, at best 2012-05-23T10:57:11 yeah 2012-05-23T10:57:38 i will edit it now and come back 2012-05-23T10:58:46 let's try to keep the interface simple, things that are exposed to the outside world or might make it into a "classic" API 2012-05-23T10:58:46 ok 2012-05-23T11:08:55 Hesham: please include a draft version of what Arena_Control struct might look like 2012-05-23T11:09:16 ok 2012-05-23T11:22:04 gedare: if we used linked list we would use chain manager right ? 2012-05-23T11:22:42 Probably. Chains make it easy to implement linked list in rtems. 2012-05-23T11:23:17 yeah 2012-05-23T11:23:25 we have to figure out what objects go on what chains. it might be that we use an array of tcbs in the acb, so that won't use a chain 2012-05-23T11:23:35 but in the tcb there might be a chain of acb's 2012-05-23T11:23:43 but the problem then is sharing 2012-05-23T11:24:10 when two acb's can be on more than one tcb's chain, the linked list will break 2012-05-23T11:24:49 wait, i think i have it backwards 2012-05-23T11:25:05 the tcb might use an array of acbs 2012-05-23T11:25:10 so that context switch is fast 2012-05-23T11:25:14 and bounded by array size 2012-05-23T11:25:14 yeah 2012-05-23T11:25:24 but one acb can be on any number of tcbs 2012-05-23T11:25:32 right 2012-05-23T11:25:33 so there will be a chain of tcbs in the acb 2012-05-23T11:25:43 but there still is a problem of duplication 2012-05-23T11:25:55 not sure how to solve it yet. 2012-05-23T11:26:04 do you know what the problem is? 2012-05-23T11:26:16 i do not think it's a problem 2012-05-23T11:27:26 could you illustrate that point? 2012-05-23T11:28:39 yeah 2012-05-23T11:29:05 the way the chain works is that each "node" is embedded in some struct 2012-05-23T11:29:17 you put the struct into a linked list by adding it to the chain 2012-05-23T11:29:22 you traverse the chain by walking through the nodes 2012-05-23T11:29:39 but you cannot have one struct on two linked lists 2012-05-23T11:29:47 you would need the struct to contain two nodes 2012-05-23T11:29:56 yeah 2012-05-23T11:29:58 i got it 2012-05-23T11:30:00 or, generalizing, n nodes for the struct to be on n lists 2012-05-23T11:30:06 that's a problem 2012-05-23T11:31:19 could not we create two structs for one node ? 2012-05-23T11:31:40 i don't think i follow your thought 2012-05-23T11:32:25 a pointer to node maybe ? 2012-05-23T11:32:34 in order to iterate through the tcbs, you need to have a chain_node in the tcb that the acb->chain_control points to 2012-05-23T11:32:57 perhaps you can draw your idea to make it more clear. 2012-05-23T11:33:31 if the time of operations that require traversing the list of tcbs from one acb is not a problem, then you could couple the node struct with the acb's pointer in the tcb->acbs[] array. that might work... 2012-05-23T11:34:44 that might not even increase the traversal time substantially. hmm 2012-05-23T11:36:03 i can think about that problem 2012-05-23T11:37:25 but for now attached tasks are kept in Thread_Control* attached_tasks[max_arena_number] ; in TCB ? 2012-05-23T11:38:49 more like Arena_Control* attached_arenas[max_arenas_per_task]; right? 2012-05-23T11:39:40 *** QingPei has left #rtems 2012-05-23T11:40:05 sorry i mean in ACB 2012-05-23T11:40:21 that would be linked list 2012-05-23T11:40:53 but for TCB that's right it should be Arena_Control* attached_arenas[max_arenas_per_task]; 2012-05-23T11:42:28 yeah in ACB I'd expect to see something like... Chain_Control attached_tasks_lists; 2012-05-23T11:42:36 or something similar to that 2012-05-23T11:43:12 yeah right 2012-05-23T11:44:05 we'll have to answer the question of how to efficiently get the tcb->attached_arenas[] == acb for arena_attach and arena_detach. 2012-05-23T11:44:31 maybe. i'm not sure if the time of those operations matter too much 2012-05-23T11:45:14 the most important issue is the time of the context switch. 2012-05-23T11:45:21 you said it does not matter when we were discussing about how to create a relationship between Arena/Task 2012-05-23T11:45:31 ok. probably it does not 2012-05-23T11:45:38 but i'm not infallible :) 2012-05-23T11:46:17 Lets send it to the community and get comments ? 2012-05-23T11:46:29 i don't see the definition of the Arena_Control struct yet 2012-05-23T11:46:43 and there are still some of my comments for you to resolve 2012-05-23T11:47:24 and a few of these functions still return status_code 2012-05-23T11:47:42 not finished with it yet 2012-05-23T11:47:43 but you're really close! 2012-05-23T11:47:52 well then no, we can't send it yet ;) 2012-05-23T11:48:36 add comments in the structure to briefly describe each field 2012-05-23T11:48:43 ok 2012-05-23T11:48:47 i added it 2012-05-23T11:49:05 i think there should be an Object_Control as the first field in the Arena_Control structure. 2012-05-23T11:49:29 that way arenas can be managed by object manager and we can deal with configuration of it in a way similar to other "objects" 2012-05-23T11:50:58 That's why i was thinking of "embeding" object manager 2012-05-23T11:54:22 think the arena_status has to be a 32-bits or is it just a bool "is_active"? 2012-05-23T11:54:22 We could add other interfaces during implementation stage , that's only the basic interfaces that should exist 2012-05-23T11:54:37 yeah but we want it to be as complete and featureful as we can 2012-05-23T11:54:51 so that we get enough early feedback about our direction as possible and detect problems sooner rather than later 2012-05-23T11:54:53 i made it uint32 because we might think of new status of arena 2012-05-23T11:54:58 so it could be extended 2012-05-23T11:55:15 i think those fields can be added later separately probably 2012-05-23T11:56:06 ok then it should be bool 2012-05-23T11:56:16 sounds good 2012-05-23T11:57:22 make the uint32 all uint32_t 2012-05-23T11:57:34 i renamed starting_address to start_address 2012-05-23T11:57:44 i noticed that 2012-05-23T11:59:16 i eliminated some functions that i don't see as necessary to understand the design 2012-05-23T11:59:28 you can add them back if you disagree :) 2012-05-23T12:01:46 ok i would also like you to propose what the attached_arenas will look like inside of the Thread_Control structure 2012-05-23T12:01:49 i added _Arena_Is_task_attached() to check if that task attached to a given Arena before it can access it 2012-05-23T12:01:51 what do you think 2012-05-23T12:01:52 ? 2012-05-23T12:02:09 i don't think so.. who is going to do this check? 2012-05-23T12:02:32 the context switch just walks through the attached_arenas[] updating directly.. 2012-05-23T12:02:40 doesn't have to do any checking at all 2012-05-23T12:03:36 So the hardware will deal with the fault of forbidden access an Arena ? 2012-05-23T12:03:56 to an Arena * 2012-05-23T12:04:20 yeah 2012-05-23T12:04:37 i think _Arena_Attach can fail when there are already too many attached arenas on a task. it might be worthwhile eventually to define an Arena_Status type that will map into rtems_status_codes later. something to mull over 2012-05-23T12:05:26 can i edit interfaces to have Arena_Status now ? 2012-05-23T12:05:50 change name of the "arena_status" field of Arena_Control to be "is_active" 2012-05-23T12:06:15 bu Arena_Status done 2012-05-23T12:06:17 i think you can have these functions returning either an Arena_Status or an int 2012-05-23T12:06:35 Arena_Status would be an int 2012-05-23T12:07:01 i mean as a return status code 2012-05-23T12:07:21 yeah that's fine. we can fill in details later 2012-05-23T12:09:51 We did not agree what _Arena_Initialize should do yet 2012-05-23T12:11:09 *** arvind_khadri has joined #rtems 2012-05-23T12:15:09 correct 2012-05-23T12:15:18 that can be dealt with later 2012-05-23T12:16:04 ok 2012-05-23T12:16:06 need to explain a bit about Arena_Create 2012-05-23T12:16:18 and then add some idea of what will be stored in Thread_Control structure for the Arena management 2012-05-23T12:16:45 ok 2012-05-23T12:27:28 what do you think you need to put in Thread_Control? 2012-05-23T12:32:17 an array of attached_arenas 2012-05-23T12:33:30 Arena_Control* attached_arenas[max_arenas_per_task]; 2012-05-23T12:39:47 ok, but what about for iterating a list of tcbs? 2012-05-23T12:39:51 from an arena 2012-05-23T12:40:43 This would happens to update attached_arenas list 2012-05-23T12:41:21 in the tcbs 2012-05-23T12:42:04 i.e when deleting an Arena , this function will iterate the list of tcb to delete this Arena from attaches_arenas list 2012-05-23T12:42:05 yeah, but what about for functions like Arena_Delete that have to traverse all the tasks 2012-05-23T12:42:05 right ? 2012-05-23T12:42:14 what is the list of tcbs though... 2012-05-23T12:42:38 if it is an rtems chain then there needs to be a chain_node in a tcb for each list that the tcb will be on, right? 2012-05-23T12:43:02 yeah 2012-05-23T12:43:29 so in the TCB there will be... 2012-05-23T12:43:37 Arena_Control* attached_arenas[max_arenas_per_task]; 2012-05-23T12:43:54 Chain_Node arena_list_node[max_arenas_per_task]; 2012-05-23T12:43:54 ? 2012-05-23T12:43:58 yeah 2012-05-23T12:44:39 or maybe a structure is defined... struct Arena_Per_task { Chain_Node node; Arena_Control *attached_arena; } 2012-05-23T12:44:59 and Arena_Per_task attached_arenas[max_arenas_per_task]; 2012-05-23T12:45:11 or something like that hmm? 2012-05-23T12:45:19 they all works 2012-05-23T12:45:23 i think 2012-05-23T12:45:27 some designs will be better than others 2012-05-23T12:45:43 so this is something you'll have to consider and maybe present some options for 2012-05-23T12:45:55 now ? 2012-05-23T12:46:27 it would be best to provide a mostly complete picture 2012-05-23T12:47:01 or, if you aren't sure what to do, just say that it is part of your design you are still working on. but I think it can be fleshed out with just a bit more work and presented with everything 2012-05-23T12:47:26 it's important to understand how the tasks - arenas relationship work, and how they fit in each others structures 2012-05-23T12:48:43 i will review chain manager again and consider some options 2012-05-23T12:49:14 it's mostly important to understand that you need a unique Chain_Node for every Chain that a structure will belong to 2012-05-23T12:49:37 so if a tcb is going to be on 8 arena's attached_tasks chains, then the tcb has to have 8 chain_nodes 2012-05-23T12:50:27 it makes sense in my mind to couple the chain_node with the pointer to the acb... that way you can walk through the chain starting from the arena, and at each task immediately be able to access its pointer back to the acb of the list being walked 2012-05-23T12:51:39 i think i got it 2012-05-23T12:51:42 there may also be some useful extra fields that could be stored in the Arena_Per_task structure, such as the index in the array of attached_arenas 2012-05-23T12:52:09 *** Jennifer has joined #rtems 2012-05-23T12:52:13 now, instead of storing this as an array of these Per_task structures, you can store just a pointer in the TCB 2012-05-23T12:52:21 this way if arena is not used, the cost is just a single null pointer 2012-05-23T12:52:41 and when an arena is attached to a task, check if the pointer is null and if so allocate the array on-demand 2012-05-23T12:52:41 right 2012-05-23T12:53:11 when the last arena is detached you could deallocate the array and set the pointer to null also, to release resources 2012-05-23T12:53:20 this would be expected to happen when a task is deleted I guess. 2012-05-23T12:54:08 Would i have to illustrate that in the design or what ? 2012-05-23T13:04:04 allocation of the array would make use of heap manager 2012-05-23T13:05:46 heap or workspace. depends on how we want to configure the space. can worry about that part later 2012-05-23T13:06:02 i think you should sketch the changes to the Thread_Control structure... 2012-05-23T13:06:23 like show how it will have a new pointer to an array of structs, and what those structs are 2012-05-23T13:06:57 ok i will do it now 2012-05-23T13:07:38 ok i'm taking a break for lunch. i'll come check back in about 30-40 minutes 2012-05-23T13:07:50 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-05-23T13:08:34 take your time , i will work hard to send the design today 2012-05-23T13:20:14 *** arvind_khadri has joined #rtems 2012-05-23T13:28:57 *** Hesham has quit IRC (Quit: Leaving.) 2012-05-23T13:29:02 *** cdcs has quit IRC (Quit: Page closed) 2012-05-23T13:36:48 *** Hesham has joined #rtems 2012-05-23T13:47:07 *** arvind_khadri has quit IRC (Ping timeout: 260 seconds) 2012-05-23T14:02:25 hey Hesham i'm back. any questions so far? 2012-05-23T14:02:41 ah, looking at your new picture. 2012-05-23T14:03:36 ok, it should probably be a pointer Arena_Per_thread *attached_arenas 2012-05-23T14:04:21 i think it makes decent sense. try to write it up in terms of struct definitions 2012-05-23T14:06:06 what should be a pointer 2012-05-23T14:06:08 ? 2012-05-23T14:06:30 just a single pointer in the Thread_Control 2012-05-23T14:06:41 struct Thread_Control { 2012-05-23T14:06:42 ... 2012-05-23T14:06:45 ah, you mean instead of the array 2012-05-23T14:06:50 Arena_Per_thread *attached_arenas; 2012-05-23T14:06:51 ... 2012-05-23T14:06:52 } 2012-05-23T14:06:54 ya 2012-05-23T14:06:56 ok 2012-05-23T14:07:04 it will be allocated to an array dynamically if arenas are used 2012-05-23T14:07:40 i understand 2012-05-23T14:08:09 ok. write it up in your design with the new struct and show like I did how it will be pointed to from thread_Control 2012-05-23T14:08:37 check it 2012-05-23T14:10:02 yeah something like that. good 2012-05-23T14:10:28 what else ? 2012-05-23T14:10:42 the primary disadvantage I see with this approach is that you can't get the tcb when iterating the list... 2012-05-23T14:11:16 but it may be is unnecessary. 2012-05-23T14:11:27 you mean attached_arenas list ? 2012-05-23T14:11:56 say you are deleting an arena... 2012-05-23T14:12:01 you start at the ACB 2012-05-23T14:12:23 and you walk through its attached_tasks list 2012-05-23T14:12:56 for each node on the list, all you can do is access entries in the Arena_Per_task structure 2012-05-23T14:13:16 you can't get at the thread_control structure itself, unless we embed a pointer to it. 2012-05-23T14:13:33 (inside of the per_task struct). 2012-05-23T14:13:50 but, I don't know if there is a need to access the tcb 2012-05-23T14:13:55 so for now we won't worry about it 2012-05-23T14:14:03 it can be tacked on later if necessary 2012-05-23T14:14:33 add the sruct Arena_Per_thread {...} to your design doc 2012-05-23T14:14:59 and say that the Thread_Control will have a pointer to an array of that struct. 2012-05-23T14:15:10 ok 2012-05-23T14:15:45 i don't know if the index is necessary right now 2012-05-23T14:15:51 you might be able to omit it 2012-05-23T14:15:53 but i think when deleting an Arena we should get the attached tasks TCBs to update attached_arenas list 2012-05-23T14:16:08 all you have to do is ... 2012-05-23T14:16:31 remove it from the chain 2012-05-23T14:16:37 for each Arena_Per_task in attached_tasks: 2012-05-23T14:16:49 rtems_chain_remove( apt.node ); 2012-05-23T14:16:54 apt.attached_arena = NULL; 2012-05-23T14:17:02 aha 2012-05-23T14:17:30 when you attach a task you search for the first non-null apt.attached_arena 2012-05-23T14:17:44 when you detach it is like deleting it 2012-05-23T14:17:56 except just for the one task's arena_per_task 2012-05-23T14:18:41 i added the struct now 2012-05-23T14:20:58 add some comments 2012-05-23T14:24:31 what do we need "is_active" flag for? 2012-05-23T14:26:16 kinda of (optional ) status , think it would be useful later 2012-05-23T14:27:58 let's scratch it for now 2012-05-23T14:28:04 keep the design a little simpler 2012-05-23T14:28:33 deleted 2012-05-23T14:28:36 i agree it might be useful for some architectures that can selectively load/unload TLB entries 2012-05-23T14:28:54 then you can use the is_active flag for arenas that are in both the current and next task during context switch 2012-05-23T14:29:16 yeah 2012-05-23T14:29:21 so i could keep it ? 2012-05-23T14:30:18 don't send it in email 2012-05-23T14:30:30 scratch it for now. add it in later if it makes sense 2012-05-23T14:30:41 remember it but get rid of it :) 2012-05-23T14:31:09 ok i will :) 2012-05-23T14:33:26 ok I think it looks good to send.. up through the optional interfaces, leave those off for now 2012-05-23T14:33:41 and without is_active 2012-05-23T14:33:47 i deleted the index field already from Arena_Per_task 2012-05-23T14:34:18 finally i could send it ? 2012-05-23T14:34:27 yeah 2012-05-23T14:35:54 gedare: Should i add anything else or just send it as it's ? 2012-05-23T14:36:43 i think it stands pretty well as it is... You should put an intro line saying this is an intro to your design and that you would appreciate any feedback 2012-05-23T14:36:57 ok 2012-05-23T14:37:17 you might want to email it to yourself first, to see how gmail cuts off the lines to make sure the formatting looks ok 2012-05-23T14:50:58 gedare: sent ! 2012-05-23T14:51:16 yay 2012-05-23T14:51:19 how do you feel about it? 2012-05-23T14:52:19 the doc ? 2012-05-23T14:52:27 or sending it ? :) 2012-05-23T14:52:57 heh, both 2012-05-23T14:53:00 the doc is pretty good / organized / introduced / every thing :) 2012-05-23T14:53:20 so what is next? 2012-05-23T14:53:32 plan to Code 2012-05-23T14:53:48 first i have to embed libmmu 2012-05-23T14:54:21 should i download libmmu.* only and run make ? 2012-05-23T14:54:54 yeah it might be a good idea to try getting the libmmu stuff in a working shape 2012-05-23T14:55:00 shake off the cobwebs 2012-05-23T14:55:35 and the low-level work is needed also right ? 2012-05-23T14:55:46 *** Jennifer has left #rtems 2012-05-23T14:55:59 yes in order to enforce memory attributes... but you should just work with stubs for now i think 2012-05-23T14:56:27 unless you have a lot of motivation for the low-level coding 2012-05-23T14:57:03 ask tuxmaniac if the powerpc support works in the psim simulator 2012-05-23T14:57:36 if it does, then you can do functional testing with enforcement using powerpc (might have to update the low-level code a little to make it fit the libmmu software architecture) 2012-05-23T14:58:10 otherwise i would say to adapt the libmmu testsuites to use arenas instead of interacting directly with libmmu 2012-05-23T15:00:21 i do not have experience with powerpc architecture 2012-05-23T15:00:43 but i only wanna test my code 2012-05-23T15:00:48 what do you suggest ? 2012-05-23T15:01:54 implement nop stubs that offer no support for memory attributes in the sparc32 2012-05-23T15:02:51 it would be nice to get the powerpc code working since most of the memory protection enforcement has been implemented already 2012-05-23T15:03:15 i will try 2012-05-23T15:03:29 you could work on the ARM mmu portion, but i worry about how long it will take in order to have confidence you got it right so that you can implement higher-level code relying on it 2012-05-23T15:04:08 i'd rather see progress on the high-level arena layers 2012-05-23T15:04:23 even if they only are tested with stubbed libmmu code. 2012-05-23T15:04:54 i will try to got it work with the current implementation 2012-05-23T15:04:57 if it's not 2012-05-23T15:05:01 i will use stubs 2012-05-23T15:05:03 but .. 2012-05-23T15:05:10 all i need is just paste the source code the proper rtems place and run make or what ? 2012-05-23T15:05:24 to the proper * 2012-05-23T15:06:37 when you add new code into rtems it can be a little complicated, the build system is complex 2012-05-23T15:07:36 mainly, you will have to add the new files, and you have to merge some files too if they already exist, like the Makefile.am and configure.ac files 2012-05-23T15:07:52 and then you have to run from rtems top directory... ./bootstrap -p 2012-05-23T15:07:55 then ./bootstrap 2012-05-23T15:08:18 then you do the configure and make like when you built RTEMS for your hello world samples 2012-05-23T15:08:34 although you'll need to --enable-tests=all in order to build the libmmu tests 2012-05-23T15:08:40 (when you configure RTEMS). 2012-05-23T15:09:33 please could your write the ./configure command that the previous libmmu project worked on 2012-05-23T15:10:25 i don't know exactly, here is one that should work. 2012-05-23T15:10:50 ../../rtems/configure --target=sparc-rtems4.11 --enable-rtemsbsp=sis --enable-tests --enable-cxx --enable-maintainer-mode 2012-05-23T15:10:58 or just ../rtems...depends on your build tree layout 2012-05-23T15:11:12 and where your rtems sources are located 2012-05-23T15:11:13 i want just the target and bsp 2012-05-23T15:11:15 thanks 2012-05-23T15:11:32 for the powerpc it would be uhmmm 2012-05-23T15:11:42 ../rtems/configure --target=powerpc-rtems4.11 --enable-rtemsbsp=psim 2012-05-23T15:12:03 the low-level implementation of libmmu works on both ? 2012-05-23T15:12:19 and then you can run it using powerpc-rtems4.11-gdb or something..there are helpers for these scripts in the rtems-testing.git 2012-05-23T15:12:30 i don't think anyone implemented stubs for sparc yet 2012-05-23T15:12:49 just an implementation for powerpc and for sparc64 2012-05-23T15:13:08 (sparc != sparc64) 2012-05-23T15:13:11 i will try with powerpc 2012-05-23T15:13:13 i know 2012-05-23T15:13:19 32 bit 2012-05-23T15:13:34 ok send an email to aanjhan if you need some help he might have advice since he has experience with it 2012-05-23T15:13:54 i will 2012-05-23T15:14:01 now that i think about it, psim might not be the bsp... so i'm not sure! 2012-05-23T15:14:35 you can also ask peter dufault and quanming shi..if you have their emails, or i can give them to you 2012-05-23T15:14:41 they worked on the gsoc2011 project 2012-05-23T15:15:15 Could not i just ask on the devel ml ? 2012-05-23T15:15:47 it would be good if you sent me their emails 2012-05-23T15:18:47 gedare: i have to go now , i will send you any progress i will do , see you later :) 2012-05-23T15:18:50 *** Hesham has quit IRC (Quit: Leaving.) 2012-05-23T15:25:28 *** gedare has quit IRC (Quit: Leaving) 2012-05-23T17:19:52 *** jennifer has joined #rtems 2012-05-23T17:43:34 *** DrJoel has quit IRC (Quit: Leaving) 2012-05-23T18:56:05 *** Deb has joined #rtems 2012-05-23T18:59:53 *** xiangfu has joined #rtems 2012-05-23T20:26:47 *** jennifer has quit IRC (Quit: Page closed) 2012-05-23T22:40:21 *** QingPei has joined #rtems 2012-05-23T23:40:29 *** arvind_khadri has joined #rtems 2012-05-24T01:10:06 *** sebhub has joined #rtems 2012-05-24T01:10:54 good morning 2012-05-24T01:12:50 morning 2012-05-24T01:14:20 sebhub, hi 2012-05-24T01:15:32 Just saw the reverts from Ralf. Was a little concerned when I first saw them. 2012-05-24T01:20:01 which reverts in the tool chain? 2012-05-24T01:32:01 No in the rtems report to the generated files you put in. 2012-05-24T01:32:22 libtest/complex 2012-05-24T01:39:25 ah the double reverts 2012-05-24T01:39:41 Hmm yes. Still not sure what is going on here. 2012-05-24T01:40:00 I saw no reply on the +1s from Joel and I from Ralf 2012-05-24T01:40:11 so thought this was finished 2012-05-24T01:40:42 maybe ralf indents to fix it in differently 2012-05-24T01:40:46 intend 2012-05-24T01:41:18 I think keeping the whole thing in C is a better idea. Generated source like this is always a pain. 2012-05-24T01:45:18 these couple of files were the only exception that prevented a read-only source tree 2012-05-24T02:06:36 *** alseh has joined #rtems 2012-05-24T02:08:04 Do you know why Ralf prefers make generated files ? 2012-05-24T02:08:35 Also what is the requirement for read-only source tree based on ? 2012-05-24T02:09:08 some source control systems need this (like Synergy) 2012-05-24T02:10:09 I do not follow. How do you edit a file ? 2012-05-24T02:10:35 you have to mark it for edit like in RCS 2012-05-24T02:11:34 Oh a Rational product. I remember their bit sliced Ada boxes. 2012-05-24T02:12:11 yes, some of our customers use it 2012-05-24T02:14:06 Can you create a directory in the tree ? 2012-05-24T02:14:27 i don't know the details 2012-05-24T02:14:36 Could you please find out and let me know ? 2012-05-24T02:14:48 ok 2012-05-24T02:14:51 Thanks 2012-05-24T03:01:49 *** xiangfu has quit IRC (Ping timeout: 245 seconds) 2012-05-24T03:03:46 *** xiangfu has joined #rtems 2012-05-24T03:38:08 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-05-24T03:44:55 *** kristianpaul has joined #rtems 2012-05-24T04:03:59 *** QingPei has left #rtems 2012-05-24T04:27:51 *** arvind_k has joined #rtems 2012-05-24T04:30:46 *** arvind_khadri has quit IRC (Ping timeout: 256 seconds) 2012-05-24T04:31:18 *** arvind_k is now known as arvind_khadri 2012-05-24T04:31:35 *** arvind_khadri has joined #rtems 2012-05-24T04:32:58 *** claas_ziemke has joined #rtems 2012-05-24T04:55:39 *** alseh has quit IRC (Remote host closed the connection) 2012-05-24T05:13:47 *** alseh has joined #rtems 2012-05-24T05:38:29 *** claas_ziemke has quit IRC (Ping timeout: 245 seconds) 2012-05-24T05:53:18 *** claas_ziemke has joined #rtems 2012-05-24T07:09:49 *** jennifer has joined #rtems 2012-05-24T07:48:09 *** sebhub has quit IRC (Remote host closed the connection) 2012-05-24T07:55:47 *** claas_ziemke has quit IRC (Read error: No route to host) 2012-05-24T08:02:53 *** gedare has joined #rtems 2012-05-24T08:05:27 *** claas_ziemke has joined #rtems 2012-05-24T08:08:22 *** arvind_khadri has quit IRC (Ping timeout: 256 seconds) 2012-05-24T08:23:54 *** ppisa has joined #rtems 2012-05-24T08:24:32 *** zw_yao has joined #rtems 2012-05-24T08:26:25 pavel: hi how r u? good to see u again :-) 2012-05-24T08:28:06 *** sebhub has joined #rtems 2012-05-24T08:34:09 *** ppisa has quit IRC (Ping timeout: 252 seconds) 2012-05-24T08:39:09 *** QingPei has joined #rtems 2012-05-24T08:54:19 *** claas_ziemke has quit IRC (Ping timeout: 245 seconds) 2012-05-24T08:54:22 *** claas_ziemke1 has joined #rtems 2012-05-24T08:54:56 is the gsoc meeting here? 2012-05-24T08:56:16 hope so :-) 2012-05-24T08:56:30 sebhub: yes should be in about an hour 2012-05-24T08:57:21 i thought now... didnt u say 15:00 utc? 2012-05-24T08:58:08 maybe i got my clocks wrong! :o 2012-05-24T08:58:43 hm, is it 14:00 utc right now? 2012-05-24T08:59:29 ooops my bad... 2012-05-24T08:59:43 i always get confused with summer time etc... 2012-05-24T08:59:46 maybe summer time + normal +1 2012-05-24T09:00:08 @sebhub: 100% true 2012-05-24T09:00:08 claas_ziemke1: Error: "sebhub:" is not a valid command. 2012-05-24T09:00:10 classic off-by-one error :) 2012-05-24T09:00:17 jupp 2012-05-24T09:01:15 anyway since i am here and you are too... can u give me a hint on where to start setting up the basic files for the bsp? 2012-05-24T09:02:24 use the stm32f bsp as a template 2012-05-24T09:02:27 ii have a working arm-none-linux-gnueabi compiler and can boot a bare c programm... but i have some problems with the linkercmds script... 2012-05-24T09:02:29 ok 2012-05-24T09:02:43 and please don't put anything into libcpu 2012-05-24T09:03:03 i dont have any intention to mess around there :-) 2012-05-24T09:03:07 hm, the linkcmds.base should work 2012-05-24T09:03:27 you should use the rtems-armeabi4.11 tool chain 2012-05-24T09:04:02 i know... just wanted to mention because now i know my qemu is working... 2012-05-24T09:04:46 this is very good, to have a working qemu 2012-05-24T09:06:42 *** ppisa has joined #rtems 2012-05-24T09:07:42 the rtems toolchain is working in gdb arm sim but not in qemu... i compiled a gdbarmsim bsp hello and flashed it on nand... <- i know this aint gonna work, just wanted to play around... i always get a: qemu: fatal: trying to execute outside RAM or ROM at 0x..... i guess that has to do with linkcmds right? 2012-05-24T09:12:14 *** DrJoel has joined #rtems 2012-05-24T09:12:14 *** ChanServ sets mode: +o DrJoel 2012-05-24T09:13:01 gedare.. irc status meeting? We didn't announce it this week 2012-05-24T09:13:31 hi joel... at least he annoenced it to me :-) 2012-05-24T09:13:58 ok 2012-05-24T09:14:13 * DrJoel was taking a long weekend and forgot to pass the baton to lead the meeting .. LOL 2012-05-24T09:14:52 but gedare was already ahead of me and copied the status template to 5/24 :) 2012-05-24T09:14:54 qemu is not a simple instruction set simulator, it can simulate a particular hardware with its peripherals 2012-05-24T09:15:39 *** claas_ziemke1 has quit IRC (Read error: No route to host) 2012-05-24T09:16:15 *** claas_ziemke has joined #rtems 2012-05-24T09:16:20 My birthday cake http://sphotos.xx.fbcdn.net/hphotos-ash3/577062_10150856327490914_540805913_9690046_1585598980_n.jpg 2012-05-24T09:16:33 sorry my laptop is a little freaky today... did i miss anything? 2012-05-24T09:17:08 qemu is not a simple instruction set simulator, it can simulate a particular hardware with its peripherals 2012-05-24T09:17:31 claas_ziemke: Someone submitted a Beagleboard BSP while you were away.. :) 2012-05-24T09:17:38 mahah 2012-05-24T09:18:26 joel nice cake... looks like mafe by a greedy dentist... :-) 2012-05-24T09:19:14 sebhub: i know... so i guess i have to read the OMAP3530 TRM again and figure out the right linkcmds script then? 2012-05-24T09:20:17 http://git.rtems.org/rtems/tree/c/src/lib/libbsp/arm/shared/startup 2012-05-24T09:20:25 :) Michele makes a fondant with marshmallow's. The cake is delicious.. layered with butter cream icing. His arm are rice krispy treats :) 2012-05-24T09:20:28 do you start with the Cortex-A or Cortex-M? 2012-05-24T09:20:51 http://git.rtems.org/rtems/tree/c/src/lib/libbsp/arm/stm32f4/startup/linkcmds.stm32f4 2012-05-24T09:21:03 i read those files... (mostly) omap3530 is a cortex-a8 2012-05-24T09:21:34 has the Cortex-A the classic ARM modes? 2012-05-24T09:21:40 i think i have to change the RAM_INT to 0x80000000 because thats the base address to the sdram... 2012-05-24T09:22:08 sorry for being stupid, what do u mean with "classic"? 2012-05-24T09:22:59 ARM7TDMI, ARM9 cores 2012-05-24T09:23:08 ok, i will check that 2012-05-24T09:23:19 So cool cake. Is it made in home? 2012-05-24T09:23:33 how many modes has the Cortex-A ? 2012-05-24T09:23:44 Deb: Yep. My wife made it. :) 2012-05-24T09:24:44 *** weiY has joined #rtems 2012-05-24T09:26:28 www.arm.com says that cortex-a8 is "Binary compatible with ARM926, ARM1136, and ARM1176 Processors" 2012-05-24T09:27:42 claas_ziemke: I would read that with a grain of sand.. probably a superset of user mode instructions if you dug deeper 2012-05-24T09:27:55 interrupt processing could be completely different 2012-05-24T09:27:59 ok 2012-05-24T09:28:25 its armv7 though... 2012-05-24T09:28:43 armv7 is to vague 2012-05-24T09:28:54 true... 2012-05-24T09:28:59 armv7-m is very special for example 2012-05-24T09:29:23 how many processor modes has the cortex-a? 2012-05-24T09:29:38 uses it the cp-15 cache and mmu? 2012-05-24T09:29:48 *** medivhc has joined #rtems 2012-05-24T09:30:27 have to evaluate that... 2012-05-24T09:30:41 its armv7-a as far as i know 2012-05-24T09:30:52 yes 2012-05-24T09:34:50 hm... now i am a little bit confused... does the toochain support the cortex-a8 at all? 2012-05-24T09:38:40 my planned next steps where to create a beagle folder in the libbsp folder (basically a copy of the stm32f4 folder... and to addapt the linkcmds file... now i guess thats way to naive... :-) 2012-05-24T09:39:02 no please stop with the stm32f4 as a template 2012-05-24T09:39:10 ok 2012-05-24T09:39:40 the cortex-a8 should use the armv7 thumb multilib or the armv4 multilib 2012-05-24T09:39:50 not optimal, but it should work 2012-05-24T09:40:18 ok... 2012-05-24T09:40:28 the processor mode count is important to decide which is the best template 2012-05-24T09:40:53 ok i will try to find out... 2012-05-24T09:41:07 i guess it should be in the processor trm right? 2012-05-24T09:42:25 in the armv7-a manual 2012-05-24T09:42:37 *** Hesham has joined #rtems 2012-05-24T09:42:51 ok 2012-05-24T09:45:48 you mean the "Processor operating states"? 2012-05-24T09:46:41 if so, it has 3: ARM state, Thumb State amd ThumbEE state... 2012-05-24T09:46:51 no, the modes 2012-05-24T09:47:07 ok sorry :-) 2012-05-24T09:47:13 abort, supervisor, undef, user for example 2012-05-24T09:48:38 ah ok... now i know what you mean... (i hope :-)) 2012-05-24T09:53:02 please update https://docs.google.com/document/d/12CRKukPcNB9QeZOanJY-Rvd2lrlPycwA_0pWSKo5xek/edit with your presence and students can jot some notes in there for what you plan to report 2012-05-24T09:54:51 gedare: i will... 2012-05-24T09:54:52 yeah, i have done 2012-05-24T09:55:14 sebhub: according to the armv7-a manual it has 8 modes.. 2012-05-24T09:55:30 can you enumerate them please 2012-05-24T09:55:43 User, FIQ, IRQ, Supervisor, Monitor, Abort, Undefinded and System 2012-05-24T09:55:57 ok, great this is the classic ARM 2012-05-24T09:56:16 if you say its great than its great :-) 2012-05-24T09:56:18 next step: cache and mmu, does it use the CP-15? 2012-05-24T09:56:25 one sec. 2012-05-24T09:56:38 great means, we don't have to implement it from scratch 2012-05-24T09:56:57 cp-15 == co-processor 15 2012-05-24T09:57:27 ************************ 5 min until meeting starts. i'm just going to go down the list of attending students in order so prepare yourselves accordingly 2012-05-24T09:59:31 "The ARMv7-A memory system incorporates a Memory Management Unit (MMU), 2012-05-24T09:59:31 controlled by CP15 registers. The memory system supports virtual addressing, with the 2012-05-24T09:59:32 MMU performing virtual to physical address translation, in hardware, as part of program 2012-05-24T09:59:32 execution." 2012-05-24T10:00:14 sounds good 2012-05-24T10:00:26 you can use the lpc32xx as a template 2012-05-24T10:00:32 http://git.rtems.org/rtems/tree/c/src/lib/libbsp/arm/lpc32xx 2012-05-24T10:00:51 ok 2012-05-24T10:02:07 * DrJoel has 10:01 :) Gedare.. 2012-05-24T10:02:14 yep! 2012-05-24T10:02:16 gedare: who do you want to go first? 2012-05-24T10:02:18 first up, Deb 2012-05-24T10:02:34 Hi Deb are you here? 2012-05-24T10:03:08 Hello to all and checking my connection. 2012-05-24T10:03:20 *** WikL has joined #rtems 2012-05-24T10:03:22 Yes 2012-05-24T10:03:24 Sorry 2012-05-24T10:03:30 * dr__house raises hand. Present 2012-05-24T10:03:46 It's ok. Deb what is your status, how is the project starting, do you have any questions or concerns? 2012-05-24T10:03:48 *** weiY has quit IRC (Ping timeout: 245 seconds) 2012-05-24T10:04:05 Yes, the project has not been fully started 2012-05-24T10:04:17 I am still thinking about the framework 2012-05-24T10:04:29 I have not done any coding yet 2012-05-24T10:04:44 That's expected at this stage. What would be great is to get some of your thinking onto "paper" and get some feedback 2012-05-24T10:04:47 deb do I understand correctly that you are evaluating existing frameworks? 2012-05-24T10:05:36 I am doing research on the functionality of the framework should implement 2012-05-24T10:05:59 And the difference with the testing of a operating system 2012-05-24T10:06:12 especially which is run on emulators 2012-05-24T10:06:20 ok.. are you looking at existing frameworks? Just trying to think if it made sense to put together an evaluation of ones you were looking at 2012-05-24T10:06:37 a table of features.. which frameworks have which.. even if you discount one for not having a major feature 2012-05-24T10:06:40 Not with operating systes 2012-05-24T10:06:58 OK 2012-05-24T10:07:00 Thank you 2012-05-24T10:07:18 Don't get hung up on testing an OS. These are just applications which test libraries 2012-05-24T10:07:21 also try not to get hung up...yes 2012-05-24T10:07:33 OK 2012-05-24T10:07:34 what he said ;) 2012-05-24T10:08:05 any other issues to report now, or what you plan to do next? 2012-05-24T10:08:22 deb.. think review or trade study at this point. You know some requirements you want.. will discover others.. just write and make tables :) It helps 2012-05-24T10:08:41 thank you 2012-05-24T10:08:56 I will ask more question on the mailing lists. 2012-05-24T10:09:08 OK Deb thank you, please update your project status in the google doc. next up is WikL 2012-05-24T10:09:25 hi WikL 2012-05-24T10:10:08 * gedare pokes WikL 2012-05-24T10:10:13 *** weiY has joined #rtems 2012-05-24T10:10:22 here 2012-05-24T10:10:36 hi :) how is your project starting? 2012-05-24T10:11:11 I'm doing research around how to go around interfacing RTEMS to POK 2012-05-24T10:11:20 a lot of questions came up 2012-05-24T10:11:30 on both parts 2012-05-24T10:11:36 ok. yes i think that would be expected. are you writing them down / collecting them? 2012-05-24T10:11:41 I am 2012-05-24T10:11:58 have them written down on paper though 2012-05-24T10:12:01 ok good. it will be a nice artifact, and please feel free to report on your progress in the rtems-devel ml 2012-05-24T10:12:17 should I use rtems-devel or the mailing list I've set up? 2012-05-24T10:12:20 ah, well you might consider storing them in a gdoc that you share with your mentors 2012-05-24T10:12:24 don't want to spam everybody 2012-05-24T10:13:05 it's ok to ask valid questions about design and implementation in rtems-devel 2012-05-24T10:13:16 Please share them on rtems-devel.. even if it is just to point to a doc 2012-05-24T10:13:17 they can be quite helpful for someone who has to use your project later 2012-05-24T10:13:30 I see 2012-05-24T10:14:15 and .. if you find your own answers feel free to respond to your own questions :) 2012-05-24T10:14:17 isn't Julien joining the meeting? 2012-05-24T10:14:26 sure 2012-05-24T10:14:29 i don't see him so maybe he is still busy again this week 2012-05-24T10:14:36 k 2012-05-24T10:14:45 have you two been able to communicate effectively? 2012-05-24T10:15:05 yes, he gave me advice about the problem I've been having with the SPARC emulator 2012-05-24T10:15:08 ok super 2012-05-24T10:15:24 so what are you doing now / next? 2012-05-24T10:15:34 and do you have any requests or sticky points? 2012-05-24T10:15:38 WikL: Keep a list of questions and answers along with write ups and decisions. This is a complicated project with a lot of details. I call this a "What do we know" document. Keeps from revisiting decisions later. Captures them and thoughts. No real format. Just decisions. Open questions, etc 2012-05-24T10:16:07 yes. this knowledge base needs to be shared and available 2012-05-24T10:16:26 i encourage all students to do this in fact :) 2012-05-24T10:16:27 right now I need to find answers to some problems, later I will get to designing the adaptation layer properly 2012-05-24T10:17:04 ok well..ask for help if you are struggling with some problems 2012-05-24T10:17:17 there are a couple problem points, but I think they might be too specific for this meeting; I'll post them to the list 2012-05-24T10:17:22 great 2012-05-24T10:17:35 *** arvind_khadri has joined #rtems 2012-05-24T10:17:38 thank you. next up is weiY 2012-05-24T10:17:44 ok 2012-05-24T10:17:46 hi weiY 2012-05-24T10:17:55 hi good night 2012-05-24T10:18:01 what is your status? 2012-05-24T10:18:20 until now i have made a report about atomic implementations at the freebsd kernel 2012-05-24T10:18:43 good. do you have any problems right now, or what is next? 2012-05-24T10:18:49 and analyses where and what types of atomic operations were used 2012-05-24T10:19:15 now i have to decide where can i place the atomic API in the source code 2012-05-24T10:19:21 libcpu or score/cpu 2012-05-24T10:19:39 i believe the atomics should be generic architecture-specific assembly code mainly? 2012-05-24T10:19:44 so it should be score/cpu 2012-05-24T10:20:00 on some architectures, you may need to know more than the multilib variant to implement them. :( 2012-05-24T10:20:14 we should answer that question :) 2012-05-24T10:20:19 yeah, i also intend to place them in score/cpu 2012-05-24T10:20:24 but, on those architecutres you can split them out to libcpu 2012-05-24T10:20:32 as the need arises 2012-05-24T10:20:33 The API needs to be in score for sure with an rtems_ wrapper. If possible the implementatino should be in score/cpu but there should be a backup location in case it can't be implemented there based upon the cpu model knowledge. 2012-05-24T10:20:48 yeah, most of code was assembly 2012-05-24T10:20:49 For example, you can't know the CPU core number on sparc v7 without using a CPU model specific features 2012-05-24T10:21:33 so your next step is to determine the code placement, and present an intitial interface design? 2012-05-24T10:21:41 yeah, if there are some specific features i will post them to list 2012-05-24T10:21:45 the sparc has no atomic support in its instruction set? 2012-05-24T10:21:48 yes 2012-05-24T10:22:11 ok weiY anything else to report? 2012-05-24T10:22:27 now i am mainly foucs on the powerpc or arm or i386 2012-05-24T10:22:38 v9 does but i don't know about v7/8 2012-05-24T10:23:03 id' have to go read a manual :) 2012-05-24T10:23:27 ok.. please identify what platforms you will be testing your code with. the more the merrier 2012-05-24T10:23:34 next up: Hesham 2012-05-24T10:23:34 no, recently i will setup the github. some questions , but i can solve 2012-05-24T10:23:55 ah,good speaking of github i hope everyone is figuring it out 2012-05-24T10:24:11 gedare: thx... it was a great help... already forked.. 2012-05-24T10:24:13 Hesham: what is your status? 2012-05-24T10:24:29 I have finished the high-level design of my project to introduce Arena manager and sent it to rtems-devel ml with initially  proposed interfaces to Arena API to get some feedbacks. 2012-05-24T10:24:48 great. 2012-05-24T10:24:52 what do you plan next? 2012-05-24T10:25:11 Now i am getting previous libmmu to work on powerpc to test my high-level part of my project. 2012-05-24T10:25:42 I have some problems with merging libmmu work with rtems code , i am trying to fix it 2012-05-24T10:25:48 super. i noticed two things with your email to the ml. first it is formatted with html, for all students I think we will prefer plain text 2012-05-24T10:26:06 and second is that there is no public API for the arenas, so you will want to consider that 2012-05-24T10:26:16 the libmmu work is outdated so that is no surprise 2012-05-24T10:26:32 work with your mentors and the prior students to resolve your issues there 2012-05-24T10:26:36 i will consider re-sending the design in plain text 2012-05-24T10:26:46 plain text is easier to archive/search 2012-05-24T10:27:15 i will contact them for sure 2012-05-24T10:27:19 ok. please update your status in the googledoc https://docs.google.com/document/d/12CRKukPcNB9QeZOanJY-Rvd2lrlPycwA_0pWSKo5xek/edit 2012-05-24T10:27:35 anything else? 2012-05-24T10:27:47 no i think that's all 2012-05-24T10:27:49 next up : zw_yao 2012-05-24T10:27:49 thanks 2012-05-24T10:28:04 zw_yao: hi 2012-05-24T10:28:15 OK 2012-05-24T10:28:19 I'm here. 2012-05-24T10:28:29 ok. what's your status 2012-05-24T10:28:51 Now I've done a summary of the design, and send it to the list. 2012-05-24T10:28:58 good. i saw it 2012-05-24T10:29:34 what's next, do you need to wait for feedback or will you proceed with one or more of your ideas? 2012-05-24T10:29:48 However, I feel I'm behind of my work. I haven't determined which approach should be selected now. 2012-05-24T10:30:15 Which is expected done before this meeting. 2012-05-24T10:30:40 there's always alternatives to pick from, and you'll have to convince yourself of the right one, and then convince everyone else 2012-05-24T10:31:03 that will have to be done either by reasoning through the decision, or by implementing something and showing that it works well (enough). 2012-05-24T10:31:14 OK, I see:) 2012-05-24T10:31:43 since you have multiple competing designs, just make sure you have enough details on each so you are able to evaluate them analytically, and implement them 2012-05-24T10:31:51 any other questions? 2012-05-24T10:32:09 I'll make the decision before next meeting as soon as possible. And try to do some coding this week. 2012-05-24T10:32:26 ok sounds good. next up: alseh 2012-05-24T10:32:38 zw_yao: please update your status with your todo 2012-05-24T10:32:43 OK, that's all. thanks! 2012-05-24T10:32:47 *** xiangfu has quit IRC (Remote host closed the connection) 2012-05-24T10:32:50 alseh: Hi 2012-05-24T10:32:54 hello 2012-05-24T10:33:11 what's your status? 2012-05-24T10:33:27 I have cloned the github versions, and will shortly make my forks 2012-05-24T10:33:58 I discussed with pavel that MWindows should be kept separate from the project .... in a separate repository 2012-05-24T10:34:22 ok have you and ppisa been having good discussions then? 2012-05-24T10:34:45 yes, he is very helpfull 2012-05-24T10:34:48 also, I have failed to build microwindows yet 2012-05-24T10:35:01 what BSP do work up till now ? .... 2012-05-24T10:35:02 it would be good to have you send some of your questions/decisions to rtems-devel so they are archived 2012-05-24T10:35:09 DrJoel: do you know? 2012-05-24T10:35:16 *** weiY has quit IRC (Ping timeout: 265 seconds) 2012-05-24T10:35:27 the wiki suggested skyeye simulator for ARM gumstix ? 2012-05-24T10:35:30 DrJoel: What BSP should microwindows (and rtems graphics toolkit) work on / was tested with? 2012-05-24T10:35:59 anyway that is a good question for rtems-devel as well, since others in the future may want to find out 2012-05-24T10:36:07 the i386 works on qemu with microwindows 2012-05-24T10:36:21 *** weiY has joined #rtems 2012-05-24T10:36:30 ok, I have qemu working, so I will continue on that path .... 2012-05-24T10:36:44 thanks 2012-05-24T10:36:51 ok. anything else today? 2012-05-24T10:37:08 also, I have finished yesterday wit university projects, so now my time will be dedicated to RTEMS 2012-05-24T10:37:10 your project involves a lot of moving parts, getting all those projects to build and then integrate, so keep at it 2012-05-24T10:37:13 super 2012-05-24T10:37:16 that's all, thank you 2012-05-24T10:37:25 next up: claas_ziemke 2012-05-24T10:37:31 ok... hi all 2012-05-24T10:37:39 so what have i done so far? 2012-05-24T10:37:40 hi. how's your project coming? 2012-05-24T10:38:03 i read a lot of dokumentation about hardware and got to know th e bsp and rtems directory layout... 2012-05-24T10:38:25 i built some scripts helping me to build nand flash images for beagkeboard/qemu 2012-05-24T10:38:42 ok good. i encourage you to keep some notes as you go / questions, which might help to improve the documentation too. 2012-05-24T10:39:00 also i verified my compilers and qemu, so everything seems to work nicely 2012-05-24T10:39:04 ok 2012-05-24T10:39:10 ok, and you have beagleboard/qemu working for a raw binary? 2012-05-24T10:39:35 what's next? 2012-05-24T10:39:42 yes... bare c programms work like a charm with arm-none-linux-gnueabi- toolchain 2012-05-24T10:40:01 also i forked on github and cloned from there... 2012-05-24T10:40:07 ok great 2012-05-24T10:40:41 next will be to make an initial "BSP" from the lpc32xx as a template as sebhub just told me prior to this meetin... thx, its a great help 2012-05-24T10:40:48 also more documentation reading... 2012-05-24T10:41:04 i also hope to finally mess around in the code next week... 2012-05-24T10:41:11 ok. sounds like a good plan 2012-05-24T10:41:17 and documentation/notes writing... 2012-05-24T10:41:47 i think thats mostly all for now... any questions? 2012-05-24T10:41:51 or suggestions? 2012-05-24T10:42:15 nothing i can think of. you're on the right track for a bsp it's mostly about tools, copying a similar one, and then adapting any boot code, console code, and then interrupt code 2012-05-24T10:42:32 get through boot, then get hello world to work ;) 2012-05-24T10:42:40 then turn on interrupts and cross your fingers 2012-05-24T10:42:46 *** cdcs has joined #rtems 2012-05-24T10:42:51 next up medivhc 2012-05-24T10:42:54 ok... since i got a bare c program to talk to uart i think thats quite doable... 2012-05-24T10:42:58 ok... thx 2012-05-24T10:43:09 claas_ziemke: when you get hello working, feel free to submit it for review and merge 2012-05-24T10:43:09 medivhc: hi 2012-05-24T10:43:22 drjoel: ok i will 2012-05-24T10:43:31 ok I have post my first proposal. 2012-05-24T10:43:48 ok yes, regarding the stats 2012-05-24T10:44:11 do you have a good idea where to go from there? 2012-05-24T10:44:19 and I will make some change as sebhub says 2012-05-24T10:44:28 yes 2012-05-24T10:44:31 last week we had a few action items from you 2012-05-24T10:44:35 Bdbuf statistics definition 2012-05-24T10:44:39 Bdbuf benchmark/reference scenarios 2012-05-24T10:44:45 use and verify correctness of bdbuf command in shell 2012-05-24T10:44:50 are these addressed now? 2012-05-24T10:45:24 the command may not be updated 2012-05-24T10:46:07 ok. also there was a comment from kiwichris that the bdbufs should be modular, do you have some idea whether that is useful or not? 2012-05-24T10:46:15 sorry, bdbuf stats should be modular^ 2012-05-24T10:47:23 bdbuf stats will not be modular I think 2012-05-24T10:47:47 they can be if it is a set of plugins for actions at a particular event 2012-05-24T10:47:51 see malloc family 2012-05-24T10:47:52 the primary concern is the cache replacement strategy, these performance related stats are a bonus work 2012-05-24T10:48:08 i see. well we should be certain that the stats gathering is easy to turn on/off or that it is "low impact" 2012-05-24T10:48:28 yeah i see the stats as a good place to be starting though it makes sense to get familiar ;) 2012-05-24T10:48:35 so .. try not to spend too much time on the stats 2012-05-24T10:48:47 ok 2012-05-24T10:48:56 do you have ideas for the cache replacement algorithm design and implementation already? 2012-05-24T10:49:38 no very clear 2012-05-24T10:49:42 as you work through the stats implementation you should work on clarifying that 2012-05-24T10:49:53 so that you can move on to implementing new cache replacement algorithms 2012-05-24T10:50:00 I think so. 2012-05-24T10:50:18 and have some way to evaluate them 2012-05-24T10:50:27 ok, anything else? 2012-05-24T10:50:57 no other students remaining I think 2012-05-24T10:50:59 That's all 2012-05-24T10:51:35 DrJoel: jennifer have you been in contact with your students for OMAP4, TCP/IP? 2012-05-24T10:52:25 I emailed with Kevin this morning he is ramping up on the project 2012-05-24T10:52:26 We have had some contact with Kevin. He is getting started now that exams are over. Jennifer asked if he had built a sim BSP for each target. We want to make sure he can run something on all targets. Haven't gotten reply yet 2012-05-24T10:52:40 On the OMAP4 .. need to email him and ask for detailed status 2012-05-24T10:52:54 ok please follow up on them, and see if this time is bad for them to meet 2012-05-24T10:53:09 **** meeting adjourned **** 2012-05-24T10:53:34 gedare: Does this look like me? http://sphotos.xx.fbcdn.net/hphotos-ash2/74725_10150855460731589_539166588_9984762_93126124_n.jpg 2012-05-24T10:53:35 just wanted to say thank u for the very nice and welcoming atmosphere here and on the ml to all the mentors and RTEMS geeks... :-) 2012-05-24T10:53:36 any students with other questions or technical questions feel free to ask :) 2012-05-24T10:53:51 DrJoel: Nerdy and crudely drawn 2012-05-24T10:53:54 gedare: did we update our status on action items? gitub moved :) 2012-05-24T10:53:58 haha i like... 2012-05-24T10:54:00 DrJoel: yes 2012-05-24T10:54:08 I have another rather simple question ... what is the correct target for arm ? arm-rtems4.11-eabi or arm-rtemseabi4.11 ? 2012-05-24T10:54:19 claas_ziemke: we're trying to put forth a good foot :) 2012-05-24T10:54:25 heh 2012-05-24T10:54:34 alseh: wow you stepped into it. :) 2012-05-24T10:54:40 indeed 2012-05-24T10:54:54 i think it is arm-rtemseabi4.11.....but it SHOULD be arm-rtems4.11 2012-05-24T10:54:56 start with arm-rtemseabi. With any luck there will be new tools shortly which rename it to arm-rtems 2012-05-24T10:55:13 good night . i have to bed .byb 2012-05-24T10:55:13 unfortunately we have some...administrative difficulties with that toolchain 2012-05-24T10:55:14 I thought it was arm-os-eabi .... but I saw arm-rtemseabi4.11 somewhere 2012-05-24T10:55:17 late rwe 2012-05-24T10:55:19 later weiY 2012-05-24T10:55:43 if the target does not follow the pattern CPU-*-rtems*, then it is NOT the right toolset :) 2012-05-24T10:56:04 (the middle * is omitted vendor) 2012-05-24T10:56:15 erm, 2012-05-24T10:56:18 yea 2012-05-24T10:56:26 oh, now I understand, I will rebuild my toolchain 2012-05-24T10:56:47 please make sure you use the patches in rtems-crossrpms.git and build newlib 2012-05-24T10:57:24 ok... as i understand thats kind of upstream for me :-) i use arm-rtemseabi4.11 2012-05-24T10:57:38 I've questions, however, it's late here. I'll discuss them on the list. And good night, bye. 2012-05-24T10:58:01 yeah if you are using the RPMs it should "just work"... we're trying to get better at making it easy to build tools from source but the going is slow 2012-05-24T10:58:07 ok night zw_yao 2012-05-24T10:58:42 i dont use rpms as i use gentoo... i build everything from source ... just a habbit :-) 2012-05-24T10:59:01 I try to eat the RPM dog food :) 2012-05-24T10:59:12 But I also test gcc and friends on the upstream 2012-05-24T10:59:22 by the way... are there any gentoo users in the RTEMS crowd? 2012-05-24T10:59:38 I use ArchLinux, and I built from sources, can not use rpms 2012-05-24T10:59:44 * DrJoel is heading out for his birthday 2012-05-24T10:59:47 *** DrJoel has quit IRC (Quit: Leaving) 2012-05-24T10:59:51 Hello Joel, please, can you add section to http://wiki.rtems.org/wiki/index.php/RTEMSGraphicsToolkit for test results and list on which BSPs you or last version has been tested? 2012-05-24T10:59:59 OK, late 2012-05-24T11:00:04 to late.. :-) 2012-05-24T11:00:43 i am just asking about gentoo because i might write some "ebuilds" for rtems in the future because i get tired of building it manually :-) 2012-05-24T11:02:29 can't you use rpmbuild on gentoo? 2012-05-24T11:02:58 i maybe could... but i wont... 2012-05-24T11:03:08 claas_ziemke: ask on the rtems-users ml 2012-05-24T11:03:25 btw i used to build from source with a script, but got sick of it ;) 2012-05-24T11:03:30 ok i will 2012-05-24T11:04:05 ebuilds are not much more as a build script, which can also resolve dependencies... 2012-05-24T11:04:29 that's good. there are some people working on better tools for automated building of toolchains 2012-05-24T11:04:42 but that work is not public right now 2012-05-24T11:04:54 for political reasons ;) 2012-05-24T11:04:54 ok... 2012-05-24T11:04:59 is the polite way to put it 2012-05-24T11:05:33 i only know there is a script called "crossdev" for gentoo which can churn out any possible crossdev comiler toolchain in no time :-) 2012-05-24T11:05:43 it should be quite easy to build the tools with rpmbuild on nearly every linux 2012-05-24T11:05:59 sebhub: true... 2012-05-24T11:06:37 the nice thing about writing a ebauild for me would be that it would integrate flawlessly with my distros package managment system... 2012-05-24T11:07:12 nuff said about that... 2012-05-24T11:08:29 *** WikL has quit IRC () 2012-05-24T11:09:38 *** medivhc has quit IRC (Quit: Leaving) 2012-05-24T11:10:30 *** weiY has quit IRC (Ping timeout: 250 seconds) 2012-05-24T11:15:37 *** Hesham has quit IRC (Quit: Leaving.) 2012-05-24T11:16:49 bye eveyone... 2012-05-24T11:17:55 bye 2012-05-24T11:18:03 *** sebhub has quit IRC (Remote host closed the connection) 2012-05-24T11:38:28 *** cdcs has quit IRC (Quit: Page closed) 2012-05-24T11:41:20 bye guys 2012-05-24T11:41:23 *** alseh has quit IRC (Quit: Leaving) 2012-05-24T11:48:53 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-05-24T11:51:16 *** claas_ziemke has left #rtems 2012-05-24T12:09:23 *** ppisa has quit IRC (Ping timeout: 252 seconds) 2012-05-24T13:12:22 *** antgreen has joined #rtems 2012-05-24T13:26:41 *** rtemsLogger has joined #rtems 2012-05-24T13:26:41 *** jennifer has quit IRC (Ping timeout: 245 seconds) 2012-05-24T14:28:08 *** QingPei has left #rtems 2012-05-24T15:39:13 *** Hesham has joined #rtems 2012-05-24T15:43:09 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-05-24T15:55:55 hi gedare 2012-05-24T17:44:22 hey hesham 2012-05-24T17:44:32 still around? i was out for a spin on my bicycle 2012-05-24T17:44:54 i see you sent plain text that is nice 2012-05-24T17:52:15 push bike ? 2012-05-24T17:52:53 gedare: i am here 2012-05-24T17:53:38 gedare: i have added a question about merging libmmu code into question doc 2012-05-24T18:07:00 ok i'll take a look 2012-05-24T18:07:56 kiwichris: i guess so if that's what you call them in australia :) 2012-05-24T18:08:01 http://en.wikipedia.org/wiki/Bicycle 2012-05-24T18:08:26 peddle pike etc etc 2012-05-24T18:08:34 pike -> bike 2012-05-24T18:08:41 yeah 2012-05-24T18:09:51 it was a nice ride. i saw a hawk with a snake (not sure if it killed or scavenged), and a beaver, and I beat the storms home. 2012-05-24T18:14:40 Hesham: did you create the libmmu directory in cpukit/ ? 2012-05-24T18:14:45 or copy it 2012-05-24T18:14:48 either way... 2012-05-24T18:15:09 anyway it would be easier for me if you listed all the steps you took and commands you ran 2012-05-24T18:15:19 otherwise it is hard for me to know what you have done 2012-05-24T18:15:36 how can i know if you missed something if i can't see everything you did :) 2012-05-24T18:19:12 gedare: yeah i created libmmu 2012-05-24T18:19:29 i will list all what i did 2012-05-24T18:33:15 gedare: i listed all the steps in the doc 2012-05-24T18:45:57 Hesham: did you successfully build rtems for psim before applying libmmu changes? 2012-05-24T18:49:31 Hesham: does there need to be a change to cpukit/configure.ac? 2012-05-24T18:49:59 http://code.google.com/p/gsoc2011-rtems-mmu-support-project/source/browse/trunk/rtems/cpukit/rtems-mmu-cpukit.diff 2012-05-24T18:50:43 you might also have to modify testsuites/libtests/Makefile.am and configure.ac to add those mmutests to the build 2012-05-24T18:51:12 i guess you just missed some files :) 2012-05-24T18:51:18 what ? 2012-05-24T18:51:36 what files ? 2012-05-24T18:51:39 and you also might need this one: http://code.google.com/p/gsoc2011-rtems-mmu-support-project/source/browse/trunk/rtems/cpukit/wrapup/Makefile.am 2012-05-24T18:51:45 cpukit/configure.ac needs a change 2012-05-24T18:52:02 and i think you need to add libmmu.a into wrapup in order to get it to build into librtemscpu 2012-05-24T18:52:22 i see 2012-05-24T18:52:26 and you may need to figure out how to add the mmutest1 and mmutest2 to the libtests/Makefile.am and configure.ac 2012-05-24T18:52:37 in order for those tests to actually build 2012-05-24T18:52:52 but! you should verify you can build rtems for psim anyway.. 2012-05-24T18:53:04 it's always best to start from a "known good" state 2012-05-24T18:53:07 editing configure.ac will be manual right ? 2012-05-24T18:53:07 and iterate in the badness :) 2012-05-24T18:53:19 yeah there's a diff to give you the idea of what to change 2012-05-24T18:53:31 all those build files should be manually fixed since they might have changed since this was update 2012-05-24T18:53:38 part of the problem with CVS ;) 2012-05-24T18:53:46 it was hard to track these kinds of things 2012-05-24T18:53:58 no problem i will work on it 2012-05-24T18:54:07 see, there was no way i would have known this was your problem if you hadn't told me EVERYTHING :) 2012-05-24T18:54:20 by the way, isn't it pretty late there ;) 2012-05-24T18:54:57 hope that i could finish this work soon , and discuss about coding plan tomorrow :) 2012-05-24T18:55:33 Dr joel advised me to put out .h file with doxygen before doing any coding 2012-05-24T18:57:02 that's a good idea 2012-05-24T18:57:14 gedare, 2012-05-24T18:57:17 hey kiwichris 2012-05-24T18:57:50 Just been look at http://www.ti.com/tool/sw-drl 2012-05-24T18:57:59 This is a package from TI for these devices. 2012-05-24T18:58:20 I downloaded the package but did not install after reading the license agreement. 2012-05-24T18:58:37 heh. free as in beer 2012-05-24T18:58:41 It is very against open source 2012-05-24T18:59:10 i don't see a link to the license on the page so hardly surprising 2012-05-24T18:59:24 It is hidden in a Windows installer. 2012-05-24T18:59:35 Taken up the issue with TI. 2012-05-24T18:59:47 niiice. I see their disclaimer though "royalty-free use (for use with Stellaris MCUs)" 2012-05-24T19:00:00 good luck 2012-05-24T19:00:28 That is fine. The actual wording is much much worse on the covering click thru license.The source has ... 2012-05-24T19:00:31 You may not combine this software with "viral" open-source 2012-05-24T19:00:31 / software in order to form a larger program. 2012-05-24T19:00:48 http://code.google.com/p/openec-32/ 2012-05-24T19:00:52 check that out instead :) 2012-05-24T19:01:36 i think Google has some package released too that is bsd licensed 2012-05-24T19:01:43 but i'm not sure where to find it exactly without hunting 2012-05-24T19:02:11 oops, that project is empty. i found the wrong thing 2012-05-24T19:03:16 There are plenty of other ARM processors to use 2012-05-24T19:03:35 yeah 2012-05-24T19:03:55 anyway, you should yell at them for sure 2012-05-24T19:04:05 I never yell ;) 2012-05-24T19:04:17 EVER 2012-05-24T19:04:22 heh 2012-05-24T19:05:05 I should release an open source license that lets people use the software at the risk of losing their eternal soul to the code's authors. 2012-05-24T19:05:27 How to you police it ? 2012-05-24T19:05:33 to -> do 2012-05-24T19:05:50 presumably i don't have to. eventually i'll just reap souls. ;) 2012-05-24T19:06:43 see it works better than viral open source, because it relies on mythical / spiritual prices to be paid, so I don't have to enforce it "in this life time" 2012-05-24T19:06:44 ;-) 2012-05-24T19:07:17 ohhh open source after life debt 2012-05-24T19:07:23 exactly! 2012-05-24T19:07:51 not sure if that is actually free though since it is somewhat of an in-kind trade 2012-05-24T19:07:58 i'll have to work on it. 2012-05-24T19:08:03 And if you pay a fee can you buy your way out of it a little like selling your soul 2012-05-24T19:08:04 in order to get it OSI approved 2012-05-24T19:08:30 anyways. i'm off to make/eat dinner. 2012-05-24T19:08:44 I am off out as well 2012-05-24T19:08:48 ok see you 2012-05-24T19:11:44 *** Hesham has quit IRC (Quit: Leaving.) 2012-05-24T19:26:46 *** xiangfu has joined #rtems 2012-05-24T19:37:37 *** kuzew has quit IRC (Ping timeout: 265 seconds) 2012-05-24T19:37:39 *** zw_yao has quit IRC (Ping timeout: 265 seconds) 2012-05-24T19:37:39 *** lcpfnvcy has quit IRC (Ping timeout: 265 seconds) 2012-05-24T19:37:42 *** zw_yao has joined #rtems 2012-05-24T19:37:48 *** lcpfnvcy_ has joined #rtems 2012-05-24T19:38:29 *** xiangfu has quit IRC (*.net *.split) 2012-05-24T19:38:29 *** A0Sheds has quit IRC (*.net *.split) 2012-05-24T19:39:04 *** kiwichris has quit IRC (*.net *.split) 2012-05-24T19:39:05 *** verm__ has quit IRC (*.net *.split) 2012-05-24T19:39:08 *** Deb has quit IRC (*.net *.split) 2012-05-24T19:39:09 *** panzon has quit IRC (*.net *.split) 2012-05-24T19:39:16 *** ChanServ has quit IRC (*.net *.split) 2012-05-24T19:42:49 *** xiangfu has joined #rtems 2012-05-24T19:42:50 *** A0Sheds has joined #rtems 2012-05-24T19:44:11 *** kuzew_ has joined #rtems 2012-05-24T19:44:12 *** Deb has joined #rtems 2012-05-24T19:44:12 *** verm__ has joined #rtems 2012-05-24T19:44:12 *** kiwichris has joined #rtems 2012-05-24T19:44:12 *** panzon has joined #rtems 2012-05-24T19:44:12 *** ChanServ has joined #rtems 2012-05-24T19:44:12 *** holmes.freenode.net sets mode: +o ChanServ 2012-05-24T19:44:38 *** kristianpaul has joined #rtems 2012-05-24T19:44:38 *** kristianpaul has joined #rtems 2012-05-24T19:44:56 *** xiangfu has quit IRC (*.net *.split) 2012-05-24T19:44:56 *** A0Sheds has quit IRC (*.net *.split) 2012-05-24T19:50:56 *** xiangfu has joined #rtems 2012-05-24T19:50:57 *** A0Sheds has joined #rtems 2012-05-24T20:05:41 *** lcpfnvcy_ has quit IRC (Ping timeout: 252 seconds) 2012-05-24T20:13:30 *** zw_yao has quit IRC (Ping timeout: 260 seconds) 2012-05-24T21:00:48 doh, i forgot to say i wouldn't be at the meeting 2012-05-24T21:35:47 *** kuzew_ is now known as kuzew 2012-05-24T22:02:08 *** QingPei has joined #rtems 2012-05-24T22:17:03 *** gedare has quit IRC (Quit: Leaving) 2012-05-24T23:25:29 *** lcpfnvcy has joined #rtems 2012-05-24T23:33:29 *** lcpfnvcy has quit IRC (Ping timeout: 245 seconds) 2012-05-24T23:41:08 *** arvind_khadri has joined #rtems 2012-05-24T23:41:15 *** lcpfnvcy has joined #rtems 2012-05-24T23:41:41 *** xiangfu has quit IRC (Ping timeout: 256 seconds) 2012-05-24T23:43:09 *** xiangfu has joined #rtems 2012-05-24T23:56:50 *** xiangfu has quit IRC (Ping timeout: 244 seconds) 2012-05-25T00:04:27 *** Deb has left #rtems ("Leaving") 2012-05-25T00:10:26 *** xiangfu has joined #rtems 2012-05-25T00:12:34 *** arvind_khadri has quit IRC (Ping timeout: 265 seconds) 2012-05-25T00:13:49 *** arvind_khadri has joined #rtems 2012-05-25T00:29:12 *** lcpfnvcy has quit IRC (Ping timeout: 252 seconds) 2012-05-25T00:42:53 *** antgreen has quit IRC (Ping timeout: 252 seconds) 2012-05-25T01:28:33 *** lcpfnvcy has joined #rtems 2012-05-25T03:25:08 *** arvind_khadri has quit IRC (Ping timeout: 240 seconds) 2012-05-25T03:38:54 *** xiangfu has quit IRC (Ping timeout: 245 seconds) 2012-05-25T08:10:45 *** Deb has joined #rtems 2012-05-25T08:21:41 *** alexs has joined #rtems 2012-05-25T08:22:07 *** alexs is now known as Guest36029 2012-05-25T08:22:24 *** alexs__ has joined #rtems 2012-05-25T08:22:54 *** alexs__ has joined #rtems 2012-05-25T08:24:02 *** alexs__ has quit IRC (Client Quit) 2012-05-25T08:24:10 *** alexs__ has joined #rtems 2012-05-25T08:24:55 *** alexs__ is now known as alseh 2012-05-25T08:25:22 *** alseh has joined #rtems 2012-05-25T08:25:54 *** alseh has quit IRC (Client Quit) 2012-05-25T08:26:13 *** alseh has joined #rtems 2012-05-25T08:34:23 *** alseh has quit IRC (Remote host closed the connection) 2012-05-25T09:02:33 *** antgreen has joined #rtems 2012-05-25T09:49:23 *** gedare has joined #rtems 2012-05-25T10:01:27 *** Deb has quit IRC (Quit: Leaving) 2012-05-25T10:53:47 *** Hesham has joined #rtems 2012-05-25T10:54:47 *** Hesham1 has joined #rtems 2012-05-25T11:03:31 hi gedare 2012-05-25T11:05:32 hey Hesham1 2012-05-25T11:05:57 ready for the meeting ? :) 2012-05-25T11:07:17 yes 2012-05-25T11:07:27 how's it going 2012-05-25T11:07:48 good 2012-05-25T11:08:14 did you have any luck with libmmu? 2012-05-25T11:08:14 Successfully built rtems for powepc after merging libmmu work 2012-05-25T11:08:33 good... did the mmutests work? I never tested them after I modified libmmu 2012-05-25T11:09:06 so you might end up wanting to revert the libmmu back to before I made my changes... you'd have to dig through the svn changes to find the revision before I made my changes 2012-05-25T11:09:19 i did not merge mmutest yet , but all other parts are successfully built 2012-05-25T11:09:35 one issue i have 2012-05-25T11:09:50 that when trying to run hello example in gdb 2012-05-25T11:09:57 i got that error 2012-05-25T11:10:01 Unknown PVR value of 0x0. Please add it to 2012-05-25T11:11:08 so you don't see Hello World? 2012-05-25T11:11:17 tuxmaniac: are you here? 2012-05-25T11:11:37 no 2012-05-25T11:11:41 i did not 2012-05-25T11:11:51 also ticker ex has the same problem 2012-05-25T11:12:30 i see. try building a "vanilla" rtems for powerpc / psim and see if that has the same problem 2012-05-25T11:12:53 if so please report your configure/build and problem to rtems-devel 2012-05-25T11:13:44 vanilla ? 2012-05-25T11:13:58 plain. no modifications 2012-05-25T11:14:03 check out fresh git 2012-05-25T11:14:13 bootstrap it, configure/build it for powerpc psim 2012-05-25T11:14:16 aha 2012-05-25T11:14:20 ok 2012-05-25T11:14:31 i had some problems myself with using psim before.... 2012-05-25T11:14:33 i used a forked version from git-hub 2012-05-25T11:14:41 that should be fine 2012-05-25T11:14:52 i can't remember, but i think i used the "psim" script in the rtems-testing.git module 2012-05-25T11:15:05 i think i had trouble with using powerpc-gdb directly 2012-05-25T11:15:22 i will try using it 2012-05-25T11:15:39 but can we work in coding plan along with this problem ? 2012-05-25T11:16:06 on** 2012-05-25T11:16:07 yes. wht is your plan 2012-05-25T11:16:19 a header file for Arena ? 2012-05-25T11:18:35 ok that sounds like a good idea 2012-05-25T11:18:40 straightforward from what you already have 2012-05-25T11:18:46 you'll also need to rewrite some of libmmu i think 2012-05-25T11:18:53 so work on a plan for that 2012-05-25T11:19:12 i don't think we want libmmu to have a public-facing interface 2012-05-25T11:19:45 which part in libmmu has to be re-written ? 2012-05-25T11:20:48 well, some of it should get moved into arena code 2012-05-25T11:21:02 and some of it should just convert to an internal implementation that can be used from arena code 2012-05-25T11:21:22 imo "libmmu" should just disappear 2012-05-25T11:21:43 but where Arena code should reside ? 2012-05-25T11:24:25 i'm not sure yet. 2012-05-25T11:24:44 probably cpukit/libarena or something. 2012-05-25T11:25:04 I would suggest you transform libmmu into libarena 2012-05-25T11:25:12 that will be an easy path forward for you 2012-05-25T11:25:39 yeah 2012-05-25T11:30:54 i should be here tuesday, but probably not next week friday 2012-05-25T11:30:57 just fyi 2012-05-25T11:32:05 it's ok 2012-05-25T11:39:42 anything else today? 2012-05-25T11:42:51 no that's all 2012-05-25T11:44:21 only one thing else , is that i built another vanilla version and it has the same problem 2012-05-25T11:47:33 ok i suggest you send an email then 2012-05-25T11:47:59 ok i will 2012-05-25T12:00:43 *** Hesham1 has quit IRC (Quit: Leaving.) 2012-05-25T12:23:42 *** QingPei has left #rtems 2012-05-25T14:13:47 *** gedare has quit IRC (Quit: Ex-Chat) 2012-05-25T15:38:02 *** antgreen has quit IRC (Remote host closed the connection) 2012-05-25T18:29:06 *** gedare has joined #rtems 2012-05-25T18:52:52 *** gedare has quit IRC (Quit: Leaving) 2012-05-25T19:20:27 *** Deb has joined #rtems 2012-05-25T19:22:28 *** Deb has quit IRC (Client Quit) 2012-05-25T20:58:50 *** xiangfu has joined #rtems 2012-05-25T21:40:13 *** xiangfu has quit IRC (Quit: Leaving) 2012-05-25T22:08:15 *** Deb has joined #rtems 2012-05-26T00:14:21 *** QingPei has joined #rtems 2012-05-26T01:30:13 *** QingPei1 has joined #rtems 2012-05-26T01:33:29 *** QingPei has quit IRC (Ping timeout: 245 seconds) 2012-05-26T04:20:50 *** alseh has joined #rtems 2012-05-26T04:39:35 *** alseh has quit IRC (Remote host closed the connection) 2012-05-26T05:41:23 *** alseh has joined #rtems 2012-05-26T05:59:23 *** A0Sheds has quit IRC (Read error: Operation timed out) 2012-05-26T06:12:45 *** A0Sheds has joined #rtems 2012-05-26T06:12:45 *** A0Sheds has joined #rtems 2012-05-26T06:19:32 *** A0Sheds has quit IRC (Ping timeout: 246 seconds) 2012-05-26T06:32:19 *** A0Sheds has joined #rtems 2012-05-26T06:32:19 *** A0Sheds has joined #rtems 2012-05-26T07:26:37 *** Deb has quit IRC (Quit: This computer has gone to sleep) 2012-05-26T08:42:47 *** QingPei1 has left #rtems 2012-05-26T10:02:40 *** QingPei has joined #rtems 2012-05-26T12:20:15 *** Hesham has joined #rtems 2012-05-26T12:49:20 *** arvind_khadri has joined #rtems 2012-05-26T12:52:14 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-05-26T13:04:38 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2012-05-26T13:25:32 *** QingPei has left #rtems 2012-05-26T13:32:36 *** panzon has quit IRC (Ping timeout: 244 seconds) 2012-05-26T13:46:55 *** panzon has joined #rtems 2012-05-26T13:47:39 *** Hesham has joined #rtems 2012-05-26T14:00:52 *** panzon has quit IRC (Ping timeout: 245 seconds) 2012-05-26T14:16:06 *** panzon has joined #rtems 2012-05-26T14:23:55 *** Hesham has quit IRC (Quit: Leaving.) 2012-05-26T14:35:56 *** Hesham has joined #rtems 2012-05-26T19:24:40 *** Deb has joined #rtems 2012-05-26T19:58:08 *** kiwichris has quit IRC (Ping timeout: 240 seconds) 2012-05-26T20:03:11 *** kiwichris has joined #rtems 2012-05-26T20:28:38 *** Deb has quit IRC (Quit: This computer has gone to sleep) 2012-05-27T01:05:08 *** A0Sheds has quit IRC (Ping timeout: 246 seconds) 2012-05-27T01:10:41 *** Deb has joined #rtems 2012-05-27T01:14:57 *** Deb has left #rtems 2012-05-27T01:17:03 *** A0Sheds has joined #rtems 2012-05-27T01:17:03 *** A0Sheds has joined #rtems 2012-05-27T01:58:12 *** alseh has quit IRC (Remote host closed the connection) 2012-05-27T07:19:10 *** xiangfu has joined #rtems 2012-05-27T08:12:09 *** xiangfu has quit IRC (Remote host closed the connection) 2012-05-27T09:55:03 *** QingPei has joined #rtems 2012-05-27T10:08:44 *** arvind_khadri has joined #rtems 2012-05-27T10:10:28 *** arvind_khadri has joined #rtems 2012-05-27T10:12:21 *** arvind_khadri has joined #rtems 2012-05-27T10:14:18 *** arvind_khadri has joined #rtems 2012-05-27T10:15:59 *** arvind_khadri has joined #rtems 2012-05-27T10:17:26 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:18:05 *** arvind_khadri has joined #rtems 2012-05-27T10:19:01 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:19:32 *** arvind_khadri has joined #rtems 2012-05-27T10:20:52 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:21:19 *** arvind_khadri has joined #rtems 2012-05-27T10:22:25 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:22:56 *** arvind_khadri has joined #rtems 2012-05-27T10:24:17 *** arvind_khadri has joined #rtems 2012-05-27T10:26:44 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:27:22 *** arvind_khadri has joined #rtems 2012-05-27T10:28:29 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:28:50 *** Hesham has joined #rtems 2012-05-27T10:29:03 *** arvind_khadri has joined #rtems 2012-05-27T10:30:24 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:30:55 *** arvind_khadri has joined #rtems 2012-05-27T10:32:38 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:33:09 *** arvind_khadri has joined #rtems 2012-05-27T10:34:31 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-05-27T10:35:06 *** arvind_khadri has joined #rtems 2012-05-27T10:44:12 *** arvind_k has joined #rtems 2012-05-27T10:45:29 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2012-05-27T10:46:02 *** arvind_k has quit IRC (Max SendQ exceeded) 2012-05-27T10:46:38 *** arvind_k has joined #rtems 2012-05-27T10:55:03 *** arvind__ has joined #rtems 2012-05-27T10:56:33 *** arvind_k has quit IRC (Ping timeout: 265 seconds) 2012-05-27T11:10:47 *** arvind__ has quit IRC (Ping timeout: 252 seconds) 2012-05-27T11:11:14 *** arvind__ has joined #rtems 2012-05-27T11:31:07 *** arvind_k has joined #rtems 2012-05-27T11:32:21 *** arvind__ has quit IRC (Read error: Operation timed out) 2012-05-27T11:33:05 *** arvind_k has quit IRC (Max SendQ exceeded) 2012-05-27T11:33:48 *** arvind_k has joined #rtems 2012-05-27T11:44:28 *** arvind_k has quit IRC (Read error: Operation timed out) 2012-05-27T11:50:32 *** QingPei has left #rtems 2012-05-27T11:56:09 *** arvind_k has joined #rtems 2012-05-27T11:57:42 *** arvind_k has quit IRC (Max SendQ exceeded) 2012-05-27T11:58:31 *** arvind_k has joined #rtems 2012-05-27T11:59:58 *** arvind_k has quit IRC (Max SendQ exceeded) 2012-05-27T12:00:32 *** arvind_k has joined #rtems 2012-05-27T12:05:57 *** arvind_k has quit IRC (Ping timeout: 248 seconds) 2012-05-27T12:38:45 *** Hesham has quit IRC (Quit: Leaving.) 2012-05-27T13:17:43 *** arvind_khadri has joined #rtems 2012-05-27T13:41:47 *** Hesham has joined #rtems 2012-05-27T13:49:02 *** Hesham has quit IRC (Quit: Leaving.) 2012-05-27T14:10:11 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-05-27T16:28:34 *** Hesham has joined #rtems 2012-05-27T16:33:42 *** Hesham1 has joined #rtems 2012-05-27T16:35:47 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-05-27T16:58:09 *** Hesham1 has quit IRC (Quit: Leaving.) 2012-05-27T17:04:08 *** panzon has quit IRC (Ping timeout: 252 seconds) 2012-05-27T17:33:52 *** panzon has joined #rtems 2012-05-27T19:18:12 *** xiangfu has joined #rtems 2012-05-27T21:03:11 *** xiangfu has quit IRC (Ping timeout: 246 seconds) 2012-05-27T21:05:49 *** xiangfu has joined #rtems 2012-05-27T21:11:28 *** xiangfu has quit IRC (Ping timeout: 250 seconds) 2012-05-27T21:17:18 *** xiangfu has joined #rtems 2012-05-27T21:30:34 *** xiangfu has quit IRC (Quit: Leaving) 2012-05-27T21:30:47 *** xiangfu has joined #rtems 2012-05-27T23:50:35 *** arvind_khadri has joined #rtems