2012-06-11T00:01:43 *** arvind_khadri has joined #rtems 2012-06-11T01:13:27 *** sebhub has joined #rtems 2012-06-11T01:13:37 good morning 2012-06-11T02:36:49 *** kiwichris has joined #rtems 2012-06-11T02:49:37 *** kiwichris has quit IRC (Quit: Leaving) 2012-06-11T02:49:56 *** kiwichris has joined #rtems 2012-06-11T02:50:13 *** kiwichris has quit IRC (Client Quit) 2012-06-11T02:51:22 *** kiwichris has joined #rtems 2012-06-11T02:52:52 *** kiwichris has joined #rtems 2012-06-11T03:34:32 *** kiwichris has quit IRC (Ping timeout: 244 seconds) 2012-06-11T03:56:08 *** zw_yao has joined #rtems 2012-06-11T04:06:09 *** xiangfu has quit IRC (Remote host closed the connection) 2012-06-11T04:28:59 *** zw_yao has quit IRC (Remote host closed the connection) 2012-06-11T04:36:50 *** zw_yao has joined #rtems 2012-06-11T04:54:44 *** zw_yao has quit IRC (Read error: Connection reset by peer) 2012-06-11T05:12:00 *** kiwichris has joined #rtems 2012-06-11T06:13:38 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2012-06-11T06:14:50 *** arvind_khadri has joined #rtems 2012-06-11T06:24:26 *** QingPei has joined #rtems 2012-06-11T07:32:47 *** zw_yao has joined #rtems 2012-06-11T07:53:47 *** zw_yao has quit IRC (Read error: Connection reset by peer) 2012-06-11T07:56:09 *** HalfMadDad has joined #rtems 2012-06-11T08:09:53 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2012-06-11T08:56:31 *** xiangfu has joined #rtems 2012-06-11T09:19:49 hi xiang 2012-06-11T09:22:41 *** xiangfu has quit IRC (Remote host closed the connection) 2012-06-11T09:43:38 *** weiY has joined #rtems 2012-06-11T09:44:41 *** antgreen has joined #rtems 2012-06-11T09:49:27 *** xiangfu has joined #rtems 2012-06-11T09:55:09 *** xiangfu has quit IRC (Ping timeout: 245 seconds) 2012-06-11T10:45:21 *** sebhub has quit IRC (Remote host closed the connection) 2012-06-11T10:58:10 *** weiY has quit IRC (Ping timeout: 265 seconds) 2012-06-11T12:53:25 *** QingPei has quit IRC (Ping timeout: 252 seconds) 2012-06-11T13:07:06 *** QingPei has joined #rtems 2012-06-11T14:12:07 *** QingPei has left #rtems 2012-06-11T17:43:03 *** deb has joined #rtems 2012-06-11T18:15:31 *** kiwichris_ has joined #rtems 2012-06-11T18:17:15 *** kiwichris has quit IRC (Ping timeout: 260 seconds) 2012-06-11T19:41:52 *** xiangfu has joined #rtems 2012-06-11T19:48:20 *** xiangfu has quit IRC (Remote host closed the connection) 2012-06-11T20:22:59 *** xiangfu has joined #rtems 2012-06-11T20:37:38 *** xiangfu has quit IRC (Ping timeout: 240 seconds) 2012-06-11T23:17:04 *** antgreen has quit IRC (Read error: Operation timed out) 2012-06-11T23:36:21 *** deb has quit IRC (Quit: Leaving) 2012-06-11T23:51:00 *** arvind_khadri has joined #rtems 2012-06-12T01:48:08 *** arvind_khadri has quit IRC (Ping timeout: 240 seconds) 2012-06-12T02:08:09 *** arvind_khadri has joined #rtems 2012-06-12T04:00:22 *** Hesham has joined #rtems 2012-06-12T04:16:41 *** Hesham has quit IRC (Quit: Leaving.) 2012-06-12T04:51:56 *** kiwichris_ has quit IRC (Quit: This computer has gone to sleep) 2012-06-12T05:04:31 *** kiwichris has joined #rtems 2012-06-12T05:07:16 *** kiwichris has quit IRC (Client Quit) 2012-06-12T07:13:53 *** sebhub has joined #rtems 2012-06-12T07:13:56 hi 2012-06-12T07:25:40 *** kiwichris has joined #rtems 2012-06-12T08:17:39 *** weiY has joined #rtems 2012-06-12T09:18:05 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-06-12T10:17:49 *** weiY has quit IRC (Ping timeout: 256 seconds) 2012-06-12T10:43:53 *** Hesham has joined #rtems 2012-06-12T10:46:56 *** antgreen has joined #rtems 2012-06-12T11:00:34 *** deb has joined #rtems 2012-06-12T11:01:10 *** antgreen has quit IRC (Read error: Connection reset by peer) 2012-06-12T11:08:35 *** Hesham1 has joined #rtems 2012-06-12T11:11:20 *** Hesham has quit IRC (Ping timeout: 244 seconds) 2012-06-12T11:30:01 *** sebhub has quit IRC (Remote host closed the connection) 2012-06-12T11:55:45 *** Hesham1 has quit IRC (Ping timeout: 252 seconds) 2012-06-12T12:07:01 *** gedare has joined #rtems 2012-06-12T13:12:03 *** deb has quit IRC (Quit: Leaving) 2012-06-12T14:02:13 *** Hesham has joined #rtems 2012-06-12T14:12:16 Hi gedare, i was here on time but you was not here, anyway i can send you an e-mail with updates & problems if you're busy now 2012-06-12T14:12:36 you can update me quick.. i'm working on something but I can spare a few minutes 2012-06-12T14:12:46 i had to do something earlier. 2012-06-12T14:13:30 ok, i have many problems and updates, i will send you detailed message with all things 2012-06-12T14:13:45 ok 2012-06-12T14:13:52 Also i had pushed merged work to my repo 2012-06-12T14:14:06 nice i'll try to check it out 2012-06-12T14:19:00 *** Hesham1 has joined #rtems 2012-06-12T14:21:54 *** Hesham has quit IRC (Read error: Connection reset by peer) 2012-06-12T14:29:32 *** Hesham1 has quit IRC (Ping timeout: 246 seconds) 2012-06-12T14:34:47 *** Hesham has joined #rtems 2012-06-12T14:40:23 *** mikeg has joined #rtems 2012-06-12T14:41:06 Hello, does anyone know where I can find recent information about claas's attempts to make a BSP for Beagle? 2012-06-12T14:46:15 *** Hesham has quit IRC (Ping timeout: 252 seconds) 2012-06-12T14:53:28 mikeg: there might be some info on the rtems-gsoc mailing list, otherwise you might consider asking on rtems-users (or rtems-gsoc) 2012-06-12T14:54:18 Thanks, I didn't see the rtems-gsoc list. I might just have to email him I guess 2012-06-12T14:54:32 probably best. I don't think he has made very much progress quite yet 2012-06-12T14:55:52 Mhmm. I'm shopping for a few boards in that price range for RTOS use (especially w/ Ada), so the only thing I'm looking to hear from him is which board he think he'll have done in a few months 2012-06-12T14:56:31 ahh. there are some special considerations for Ada support, but those mostly are on the tools side I believe.... not sure how the Arm support is for RTEMS Ada. 2012-06-12T14:56:37 (You might remember me trying to fix up GNAT to work on powerpc-rtems a few years ago) 2012-06-12T14:56:40 ahhh 2012-06-12T14:57:17 It was successful haha ;) 2012-06-12T14:57:27 well that's good at least 2012-06-12T14:59:24 *** mikeg_ has joined #rtems 2012-06-12T15:01:41 *** mikeg has quit IRC (Ping timeout: 245 seconds) 2012-06-12T15:08:21 gedare: one other thing, in your document about the IRC GSOC meeting you state that the IRC logs are available, but the links for June are all broken 2012-06-12T15:08:27 (not sure if you control that) 2012-06-12T15:21:46 *** mikeg_ has quit IRC (Quit: Page closed) 2012-06-12T16:10:16 *** gedare has quit IRC (Quit: Leaving) 2012-06-12T16:13:36 *** deb has joined #rtems 2012-06-12T16:18:06 *** Hesham has joined #rtems 2012-06-12T16:27:39 *** Hesham has quit IRC (Ping timeout: 265 seconds) 2012-06-12T18:19:56 * kiwichris just missed Gedare 2012-06-12T18:54:03 *** mikeg has joined #rtems 2012-06-12T19:21:19 *** deb has quit IRC (Quit: Leaving) 2012-06-12T20:01:51 *** xiangfu has joined #rtems 2012-06-12T21:34:44 *** AndyKing has joined #rtems 2012-06-12T22:00:26 *** HalfMadDad has left #rtems ("Leaving") 2012-06-12T22:19:00 *** arvind_khadri has joined #rtems 2012-06-12T22:25:59 *** AndyKing has quit IRC (Ping timeout: 244 seconds) 2012-06-12T22:26:52 *** kiwichris has quit IRC (Read error: Connection reset by peer) 2012-06-12T22:27:18 *** kiwichris has joined #rtems 2012-06-13T00:23:25 *** QingPei has joined #rtems 2012-06-13T00:33:57 *** AndyKing has joined #rtems 2012-06-13T00:43:21 *** AndyKing has quit IRC (Ping timeout: 246 seconds) 2012-06-13T00:45:32 *** AndyKing has joined #rtems 2012-06-13T01:32:30 *** xiangfu has quit IRC (Ping timeout: 244 seconds) 2012-06-13T01:34:18 *** sebhub has joined #rtems 2012-06-13T01:35:08 good morning 2012-06-13T01:35:33 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-06-13T02:18:05 *** QingPei has quit IRC (Ping timeout: 244 seconds) 2012-06-13T02:18:17 sebhub, hi 2012-06-13T02:26:38 hi 2012-06-13T02:57:15 *** QingPei has joined #rtems 2012-06-13T03:26:43 *** xiangfu has joined #rtems 2012-06-13T03:31:26 *** xiangfu has quit IRC (Ping timeout: 256 seconds) 2012-06-13T03:43:24 *** xiangfu has joined #rtems 2012-06-13T04:02:17 *** arvind_khadri has joined #rtems 2012-06-13T04:02:51 *** QingPei has left #rtems 2012-06-13T04:16:03 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2012-06-13T04:16:22 sebhub, should the atomic API be using 'unsigned int' etc or should we use specific types ? I tend to think specific types. 2012-06-13T04:17:05 i think specific types are better since this adds some self documentation to the code and highlights atomic variables 2012-06-13T04:18:12 *** arvind_khadri has joined #rtems 2012-06-13T04:18:47 *** AndyKing has quit IRC (Ping timeout: 252 seconds) 2012-06-13T04:40:48 *** AndyKing has joined #rtems 2012-06-13T05:15:54 *** arvind_khadri has quit IRC (Read error: Operation timed out) 2012-06-13T07:02:19 *** xiangfu has quit IRC (Quit: Leaving) 2012-06-13T07:36:29 *** xiangfu has joined #rtems 2012-06-13T07:53:46 *** weiY has joined #rtems 2012-06-13T08:43:27 *** antgreen has joined #rtems 2012-06-13T09:26:45 *** xiangfu has quit IRC (Quit: Leaving) 2012-06-13T09:45:40 *** AndyKing has quit IRC (Quit: Leaving) 2012-06-13T10:06:07 *** sebhub has quit IRC (Remote host closed the connection) 2012-06-13T10:57:23 *** weiY has quit IRC (Ping timeout: 245 seconds) 2012-06-13T12:02:13 *** antgreen has quit IRC (Ping timeout: 248 seconds) 2012-06-13T12:16:26 *** gedare has joined #rtems 2012-06-13T14:45:34 *** gedare has quit IRC (Quit: Ex-Chat) 2012-06-13T15:27:44 *** antgreen has joined #rtems 2012-06-13T16:31:38 *** antgreen has quit IRC (Ping timeout: 240 seconds) 2012-06-13T20:16:20 *** xiangfu has joined #rtems 2012-06-13T23:36:34 *** antgreen has joined #rtems 2012-06-13T23:45:24 *** kuzew has quit IRC (Read error: Operation timed out) 2012-06-13T23:47:11 *** kuzew has joined #rtems 2012-06-14T00:01:02 *** arvind_khadri has joined #rtems 2012-06-14T00:23:35 *** antgreen has quit IRC (Ping timeout: 265 seconds) 2012-06-14T01:02:36 *** xiangfu has quit IRC (Ping timeout: 244 seconds) 2012-06-14T01:24:09 *** AndyKing has joined #rtems 2012-06-14T01:52:38 *** sebhub has joined #rtems 2012-06-14T01:53:52 good morning 2012-06-14T01:59:11 sebhub, I agree we should define specific types. 2012-06-14T02:01:29 hi 2012-06-14T02:02:09 ok, i think wei removed them in the latest version 2012-06-14T02:11:54 Yes he did in respond to what Ralf said. In his current code he has used unsigned int. I asked him to check bds and it uses int_u etc which are not sized. 2012-06-14T02:13:33 I think we should use a specific atomic type. which maps to the arch's specific type 2012-06-14T02:13:45 We should not use sizes. 2012-06-14T02:15:26 we should also address the use case for this atomic api 2012-06-14T02:16:14 i think an atomic flag, int and pointer should be sufficient 2012-06-14T02:17:22 had you time to look at the coldfire interrupt patch? 2012-06-14T02:41:19 *** xiangfu has joined #rtems 2012-06-14T07:43:55 *** AndyKing has quit IRC (Read error: Connection reset by peer) 2012-06-14T07:44:34 *** AndyKing has joined #rtems 2012-06-14T08:07:51 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2012-06-14T08:25:24 *** arvind_khadri has joined #rtems 2012-06-14T08:46:45 *** AndyKing has quit IRC (Quit: Leaving) 2012-06-14T08:52:51 *** zw_yao has joined #rtems 2012-06-14T08:59:49 *** zw__yao has joined #rtems 2012-06-14T08:59:49 *** zw_yao has quit IRC (Read error: Connection reset by peer) 2012-06-14T09:09:38 *** sebhub has quit IRC (Ping timeout: 240 seconds) 2012-06-14T09:18:52 *** arvind_khadri has quit IRC (Remote host closed the connection) 2012-06-14T09:19:08 *** arvind_khadri has joined #rtems 2012-06-14T09:19:08 *** arvind_khadri has joined #rtems 2012-06-14T09:23:21 *** xiangfu has quit IRC (Remote host closed the connection) 2012-06-14T09:24:35 *** jennifer has joined #rtems 2012-06-14T09:39:57 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-06-14T09:44:44 *** arvind_khadri has quit IRC (Ping timeout: 265 seconds) 2012-06-14T09:46:27 *** kristianpaul has joined #rtems 2012-06-14T09:46:27 *** kristianpaul has joined #rtems 2012-06-14T09:58:48 *** ppisa has joined #rtems 2012-06-14T10:07:30 *** zw__yao has quit IRC (Quit: Bye) 2012-06-14T11:16:52 *** ppisa has left #rtems ("Kopete 0.12.7 : http://kopete.kde.org") 2012-06-14T14:31:37 *** jennifer has quit IRC (Quit: Leaving) 2012-06-14T14:41:39 *** Hesham has joined #rtems 2012-06-14T14:52:05 *** gedare has joined #rtems 2012-06-14T14:53:05 hi Hesham 2012-06-14T14:53:11 hi gedare 2012-06-14T14:53:20 i have a couple minutes 2012-06-14T14:53:25 thought i'd check in on you 2012-06-14T14:53:41 yeah i have added a question to my doc 2012-06-14T14:54:05 you can check it whenever you are free 2012-06-14T14:54:08 i see it. 2012-06-14T14:54:35 And ? 2012-06-14T14:54:39 i'm not sure if we should include domain in libmm or not... 2012-06-14T14:54:51 you have made a pretty good analysis of the choices 2012-06-14T14:55:15 i think it's good for allocation 2012-06-14T14:55:31 although regardless of allocation strategy, i think it should be 1:1 mapping between mpe and arena... 2012-06-14T14:55:38 not domain and arena 2012-06-14T14:55:47 yeah that's what i intended to do 2012-06-14T14:55:49 a domain is the set of arenas that a task has active 2012-06-14T14:56:10 but managing that set requires a lot more than the simple list-based approach taken by libmm domains 2012-06-14T14:56:16 so it should be possible to use mpe's without domains 2012-06-14T14:56:35 domain is a useful concept.. but the way it was implemented in libmm is insufficient 2012-06-14T14:56:47 but when i am implementing _Arena_Create i mapped it to rtems_mm_create_entry 2012-06-14T14:56:56 that needs a domain by default 2012-06-14T14:57:05 yes i see the problem 2012-06-14T14:57:17 we need to decouple mpe's and domains 2012-06-14T14:57:19 in libmm. 2012-06-14T14:57:38 yeah 2012-06-14T14:57:40 because we should perhaps like to share mpe's across multiple tasks/domains 2012-06-14T14:57:58 the linked-list approach taken by libmm is not good enough for multitask protection 2012-06-14T14:58:11 we could delete domain part and thinking of a new way if needed 2012-06-14T14:58:13 it might be ok for statically defined protection, but not for when you have context switching or sharing 2012-06-14T14:58:39 i agree. eliminate domain entirely and re-implement the concept inside of arena 2012-06-14T14:58:51 i will map arena to mpe 2012-06-14T14:58:57 the libmm will still be useful for both trnaslation to hw, and also can work for setting up a static memory map 2012-06-14T14:59:02 and edit the mmu work 2012-06-14T14:59:22 ok good. 2012-06-14T14:59:41 good for now 2012-06-14T15:00:01 you can adopt the mmutest1 also if possible. or write your own, or find the original version that existed before i wrote the one currently in the svn 2012-06-14T15:00:22 ok i will check it 2012-06-14T15:00:35 the original version relied on the memory layout of powerpc, and on the page size and other hw-specific features that should not exist in a generic test of libmm 2012-06-14T15:00:45 so it is not quite useful, but can be a good enough starting point I guess 2012-06-14T15:00:50 aha 2012-06-14T15:00:54 another thing 2012-06-14T15:00:59 otherwise you should design a test that creates some static memory map 2012-06-14T15:01:14 you might be able to use variables exported by the linker for that...maybe 2012-06-14T15:01:31 i may use alignment implementation according to page size for Arenas 2012-06-14T15:01:45 but i would look for the old version of mmutest1 first, because that can give you some idea bout whether your powerpc code actually works 2012-06-14T15:01:54 alignment should be specified by BSP 2012-06-14T15:01:58 since it knows the requirements 2012-06-14T15:02:15 because page size is hw specific feature 2012-06-14T15:02:30 yeah, it's implementation would be at BSP layer 2012-06-14T15:02:32 or more generally alignment requirements for mpe 2012-06-14T15:02:47 or we may import BSP page size 2012-06-14T15:02:49 keep the terminology "page"-clean, meaning avoid talking aobut pages 2012-06-14T15:03:15 i think. at least for now. 2012-06-14T15:03:20 ok 2012-06-14T15:03:29 you can use minimum alignment, minimum size, something like these field names would be fine 2012-06-14T15:03:36 make snese? 2012-06-14T15:03:44 yeah it make sense 2012-06-14T15:04:39 would you let updated with any new work 2012-06-14T15:04:56 i will let you updated with any new work ** 2012-06-14T15:05:42 ok. send out an email describing what the BSP provides, and how it does so 2012-06-14T15:06:04 powerpc or in general ? 2012-06-14T15:06:05 to rtems-devel, following up from the prior email where ric suggested you let the BSP define the layout of the attribute structure 2012-06-14T15:06:11 in general 2012-06-14T15:06:20 you can use powerpc as an example if you have sufficient details 2012-06-14T15:06:33 that's ok 2012-06-14T15:08:14 all right. i ought to be around tomorrow if you have any more questions or progress. 2012-06-14T15:08:45 do you have your code on a branch on your rtems fork on github? 2012-06-14T15:08:56 not yet 2012-06-14T15:09:11 ok you should work on doing that too please 2012-06-14T15:09:13 i have pushed merged work only 2012-06-14T15:09:19 okay 2012-06-14T15:09:33 remove your push to the master... perhaps by re-forking 2012-06-14T15:09:41 and push your work to a branch instead 2012-06-14T15:10:01 it will be easier to track branches and for you to get updates from upstream if necessary 2012-06-14T15:10:04 i made a new branch 2012-06-14T15:10:16 but it's located at my machine only 2012-06-14T15:10:48 push it to github when you can. and also undo the commit you made to your rtems master 2012-06-14T15:11:14 i commited it to my forked repo only 2012-06-14T15:11:29 yeah i know but i would prefer that you only push your changes in branches, not in master 2012-06-14T15:11:48 how could i create a branch at github ? 2012-06-14T15:12:00 i know how to do it on my machine 2012-06-14T15:12:02 make a branch locally, and then git push 'branch' 2012-06-14T15:12:15 only that 2012-06-14T15:12:17 replacing 'branch' with the name of your local branch 2012-06-14T15:12:26 ok i will 2012-06-14T15:14:42 ok. with that, i'm out 2012-06-14T15:14:46 *** gedare has quit IRC (Quit: Leaving) 2012-06-14T15:22:21 *** Hesham has quit IRC (Ping timeout: 252 seconds) 2012-06-14T15:45:04 *** antgreen has joined #rtems 2012-06-14T16:38:21 *** Hesham has joined #rtems 2012-06-14T16:45:43 *** Hesham has quit IRC (Ping timeout: 265 seconds) 2012-06-14T16:53:37 *** Hesham has joined #rtems 2012-06-14T17:41:22 *** Hesham has quit IRC (Quit: Leaving.) 2012-06-14T20:16:51 *** xiangfu has joined #rtems 2012-06-14T23:31:30 *** antgreen has quit IRC (Ping timeout: 244 seconds) 2012-06-14T23:59:35 *** arvind_khadri has joined #rtems 2012-06-15T01:33:01 *** sebhub has joined #rtems 2012-06-15T01:35:22 good morning 2012-06-15T03:12:59 *** WikL has joined #rtems 2012-06-15T07:14:29 *** WikL has quit IRC (Ping timeout: 260 seconds) 2012-06-15T07:26:47 *** weiY has joined #rtems 2012-06-15T07:27:11 hi chris 2012-06-15T07:27:17 are you here 2012-06-15T07:35:11 Hi sebhub, are you around? 2012-06-15T09:41:20 *** gedare has joined #rtems 2012-06-15T09:49:27 *** xiangfu has quit IRC (Quit: Leaving) 2012-06-15T10:15:16 *** arvind_khadri has quit IRC (Ping timeout: 244 seconds) 2012-06-15T10:17:28 *** Hesham has joined #rtems 2012-06-15T10:24:34 hey Hesham how's it going? 2012-06-15T10:24:54 currently reviewing svn changes 2012-06-15T10:25:01 i have a concern too 2012-06-15T10:25:34 When creating an Arena it should have MPE 2012-06-15T10:25:58 MPEs could be allocated dynamically 2012-06-15T10:26:04 at creation time of arena 2012-06-15T10:26:06 or 2012-06-15T10:26:33 When initializing Arena, It allocated MPEs for maximum numbers of Arenas 2012-06-15T10:26:50 preallocated i think. 2012-06-15T10:27:02 And keep it in a chain ? 2012-06-15T10:27:23 hmmmmm 2012-06-15T10:27:41 actually not preallocated, but included in the Arena? 2012-06-15T10:27:45 and preallocate the arena? 2012-06-15T10:27:59 yeah maybe 2012-06-15T10:28:12 Arena_Control { ... MM_MPE the_mpe; ... }; 2012-06-15T10:28:18 that would make Arena_Control has rtems_mm_entr 2012-06-15T10:28:18 y 2012-06-15T10:28:19 btw need to rename MPE 2012-06-15T10:28:25 yeah 2012-06-15T10:28:28 ok good you already did that 2012-06-15T10:28:46 so arena_control "composes" (includes) the mm_entry 2012-06-15T10:28:46 but what about static configuration 2012-06-15T10:29:03 i think it still will work 2012-06-15T10:29:28 ok 2012-06-15T10:29:43 we might also eventually need to consider how to "hand-off" the control of memory from bootloader to rtems 2012-06-15T10:30:17 right now every bsp that needs to use mmu does a static approach i think, so the bootstrap of rtems just sets up the memory map 2012-06-15T10:30:23 and rtems will remap memory 2012-06-15T10:30:33 yeah but probably needs to know the original map 2012-06-15T10:30:43 how could i know it ? 2012-06-15T10:30:46 dunno yet 2012-06-15T10:30:47 from assembly code ? 2012-06-15T10:31:08 i was setting it at Assembly stage 2012-06-15T10:31:10 well linkcmds will tell you the intended mapping 2012-06-15T10:31:18 ok 2012-06-15T10:31:23 well just something to think about 2012-06-15T10:31:25 and be aware of 2012-06-15T10:31:39 there are a lot of problems with taking over the mmu hw 2012-06-15T10:32:03 ok, for now i will initializing Arenas and eliminating domain management 2012-06-15T10:32:07 ok 2012-06-15T10:32:21 i will push my code today, hope you will be there if there were problems with github 2012-06-15T10:32:40 i may be around.. 2012-06-15T10:32:55 i want you to push a branch that has just your libmm fixes too 2012-06-15T10:33:07 and then you can have a branch where you are starting your arena code 2012-06-15T10:33:13 we need to get the libmm work into mergeable state 2012-06-15T10:33:52 when i am creating a branch, i think it create one for all the repo 2012-06-15T10:33:58 not only libmm 2012-06-15T10:35:27 yes 2012-06-15T10:35:43 that is ok it only stores the files that are different 2012-06-15T10:35:44 it that ok ? 2012-06-15T10:35:55 ok good 2012-06-15T10:36:05 yes.. spend a little time to read about version control and git maybe if you don't know a lot yet 2012-06-15T10:36:15 it will do good to understane 2012-06-15T10:36:20 i will 2012-06-15T10:36:49 i had read aanjhan suggested blog too 2012-06-15T10:39:39 ok 2012-06-15T10:46:31 *** sebhub has quit IRC (Remote host closed the connection) 2012-06-15T10:58:21 *** mikeg_ has joined #rtems 2012-06-15T11:05:18 *** weiY has quit IRC (Ping timeout: 265 seconds) 2012-06-15T11:09:34 *** antgreen has joined #rtems 2012-06-15T11:33:40 Hesham: i'm going to head home soon. i might be back again, but do you have any questions for me before I go? 2012-06-15T11:35:17 gedare: yeah just one 2012-06-15T11:35:37 it's ok if you do not have time now 2012-06-15T11:35:48 go ahead 2012-06-15T11:36:14 I am using object management for handling Arenas 2012-06-15T11:36:32 one issue is checking overlapping addresses when creating arenas 2012-06-15T11:36:47 is it possible with object manager ? 2012-06-15T11:38:06 not directly no 2012-06-15T11:38:39 the object manager has no idea about the objects implementation, so it has no idea about arena internals such as the address range 2012-06-15T11:39:14 i found it has inactive chain used when allocating objects 2012-06-15T11:40:00 so i am thinking about iterate over active chain if there were one 2012-06-15T11:40:22 object manager preallocates enough of the object type so that it can satisfy allocation requests in O(1) time 2012-06-15T11:40:33 the inactive chain is basically a free list of objects 2012-06-15T11:41:17 aha, Then i might manage arenas with a chain manager 2012-06-15T11:41:21 there is no active list, the implementation of the object (e.g. Arena Manager) can choose to use the Object.Node field in its own lists if it likes 2012-06-15T11:41:38 yeah you can put arena_control on your own chain after you allocate it from the object manager 2012-06-15T11:42:01 ok 2012-06-15T11:42:28 it's up to the implementation to return the object back to the object manager (object free or whatever) which will return the object to the inactive list 2012-06-15T11:42:42 so after allocation the object manager has no handle on the object 2012-06-15T11:43:08 aha i see 2012-06-15T11:43:35 i will initialize empty chain when initializing Arenas 2012-06-15T11:44:04 it would contains all allocated Arenas that's currently used 2012-06-15T11:44:16 or preallocated via object manager 2012-06-15T11:48:28 i don't follow your thinking quite 2012-06-15T11:49:41 you'll be putting an arena_control on some kind of chain after you allocate... we discussed before about how you might have to organize access to arena_control though 2012-06-15T11:50:18 based on threads 2012-06-15T11:50:28 if we want to tie memory mapping to threads 2012-06-15T11:50:55 that's what i mean 2012-06-15T11:51:44 "you'll be putting an arena_control on some kind of chain after you allocate" 2012-06-15T11:55:53 ok 2012-06-15T11:55:57 great 2012-06-15T11:56:16 let's set a goal for committing libmm 2012-06-15T11:57:30 ok 2012-06-15T11:57:31 if you get it pushed i should be able to review it before next tuesday... do whatever changes by end of next week... review again and maybe get public review, and then submit it the week after next? June 25-29? 2012-06-15T11:58:45 i will finish my exams at June 24 2012-06-15T11:59:15 how much work do you estimate is left on libmm before you can have me review it? 2012-06-15T11:59:45 you mean implementing all Arena API or what ? 2012-06-15T12:00:13 i can eliminate domain work and re-implementing Arenas according to mpes only 2012-06-15T12:00:44 and finishing _Arena_Create and _Arena_Initialize 2012-06-15T12:00:54 no no nothing about arena 2012-06-15T12:01:01 just the core libmm stuff 2012-06-15T12:01:17 the refactoring, renaming, attributes from bsp, etc 2012-06-15T12:01:23 the core would just contains changes of domain and names 2012-06-15T12:01:26