2012-07-09T00:49:45 *** arvind_khadri has joined #rtems 2012-07-09T02:27:06 *** A2Sheds has quit IRC (Ping timeout: 264 seconds) 2012-07-09T02:28:03 *** ppisa has joined #rtems 2012-07-09T02:28:18 *** A2Sheds has joined #rtems 2012-07-09T03:47:59 *** arvind_khadri has quit IRC (Quit: Leaving) 2012-07-09T03:50:04 *** arvind_khadri has joined #rtems 2012-07-09T04:08:25 *** hiddenpearls has joined #rtems 2012-07-09T05:26:14 *** alseh has joined #rtems 2012-07-09T06:14:41 *** hiddenpearls1 has joined #rtems 2012-07-09T06:16:48 *** hiddenpearls has quit IRC (Ping timeout: 252 seconds) 2012-07-09T06:32:55 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2012-07-09T07:29:00 *** hiddenpearls has joined #rtems 2012-07-09T07:50:45 *** xiangfu has quit IRC (Quit: Leaving) 2012-07-09T08:19:01 *** jahf has quit IRC (Ping timeout: 272 seconds) 2012-07-09T08:34:11 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-09T08:35:46 *** hiddenpearls has joined #rtems 2012-07-09T08:47:39 *** weiY has joined #rtems 2012-07-09T08:54:17 *** jahf has joined #rtems 2012-07-09T09:09:00 *** sebhub has joined #rtems 2012-07-09T09:09:05 hi 2012-07-09T09:09:18 Hi sebhub, good night 2012-07-09T09:19:50 *** deb has joined #rtems 2012-07-09T09:23:30 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-09T09:31:51 *** deb has quit IRC (Quit: Leaving) 2012-07-09T09:34:12 *** jahf has quit IRC (Ping timeout: 245 seconds) 2012-07-09T09:35:45 *** jahf has joined #rtems 2012-07-09T09:37:51 *** hiddenpearls has joined #rtems 2012-07-09T10:03:46 *** hiddenpearls has quit IRC (Ping timeout: 255 seconds) 2012-07-09T10:05:20 *** hiddenpearls has joined #rtems 2012-07-09T10:07:04 *** hiddenpearls has quit IRC (Client Quit) 2012-07-09T10:09:41 *** alseh has quit IRC (Ping timeout: 248 seconds) 2012-07-09T10:20:30 *** hiddenpearls has joined #rtems 2012-07-09T10:20:43 *** weiY has quit IRC (Ping timeout: 246 seconds) 2012-07-09T10:24:21 *** hiddenpearls1 has joined #rtems 2012-07-09T10:24:45 *** hiddenpearls has quit IRC (Read error: Connection reset by peer) 2012-07-09T10:26:53 *** alseh has joined #rtems 2012-07-09T10:36:20 *** alseh has quit IRC (Ping timeout: 248 seconds) 2012-07-09T10:45:34 *** sebhub has quit IRC (Remote host closed the connection) 2012-07-09T11:06:28 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2012-07-09T11:12:19 *** ppisa has quit IRC (Quit: Kopete 0.12.7 : http://kopete.kde.org) 2012-07-09T11:29:14 *** Hesham has joined #rtems 2012-07-09T11:46:58 *** Hesham1 has joined #rtems 2012-07-09T11:47:08 *** Hesham has quit IRC (Ping timeout: 244 seconds) 2012-07-09T12:16:47 *** Hesham1 has quit IRC (Ping timeout: 250 seconds) 2012-07-09T12:18:02 *** Hesham has joined #rtems 2012-07-09T12:46:53 *** Hesham1 has joined #rtems 2012-07-09T12:47:07 *** Hesham has quit IRC (Ping timeout: 240 seconds) 2012-07-09T13:09:18 *** Hesham1 has quit IRC (Ping timeout: 252 seconds) 2012-07-09T13:10:04 *** sevikkk1 has joined #rtems 2012-07-09T13:12:41 *** sevikkk has quit IRC (Ping timeout: 265 seconds) 2012-07-09T13:16:53 *** Hesham has joined #rtems 2012-07-09T13:22:03 *** deb has joined #rtems 2012-07-09T13:39:33 *** Hesham has quit IRC (Ping timeout: 250 seconds) 2012-07-09T14:36:50 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2012-07-09T17:19:25 *** DrJoel has joined #rtems 2012-07-09T17:19:25 *** DrJoel has joined #rtems 2012-07-09T17:19:25 *** ChanServ sets mode: +o DrJoel 2012-07-09T17:49:35 *** DrJoel has quit IRC (Quit: Leaving) 2012-07-09T19:11:21 *** hiddenpearls has joined #rtems 2012-07-09T19:19:52 *** hiddenpearls1 has joined #rtems 2012-07-09T19:22:34 *** hiddenpearls has quit IRC (Ping timeout: 252 seconds) 2012-07-09T20:19:06 *** hiddenpearls has joined #rtems 2012-07-09T20:22:47 *** hiddenpearls1 has quit IRC (Read error: Connection reset by peer) 2012-07-09T20:22:50 *** xiangfu has joined #rtems 2012-07-09T20:25:54 *** xiangfu has quit IRC (Ping timeout: 264 seconds) 2012-07-09T20:28:50 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-09T20:37:12 *** xiangfu has joined #rtems 2012-07-09T20:42:20 *** Hesham has joined #rtems 2012-07-09T21:15:20 *** Hesham1 has joined #rtems 2012-07-09T21:18:34 *** Hesham has quit IRC (Ping timeout: 244 seconds) 2012-07-09T21:27:19 *** zw_yao has joined #rtems 2012-07-09T21:33:20 *** QingPei has joined #rtems 2012-07-09T21:41:23 *** zw_yao has quit IRC (Ping timeout: 245 seconds) 2012-07-09T21:45:33 *** Hesham1 has quit IRC (Ping timeout: 245 seconds) 2012-07-09T21:46:47 *** Hesham has joined #rtems 2012-07-09T21:58:21 *** Hesham has quit IRC (Ping timeout: 244 seconds) 2012-07-09T22:31:25 *** xiangfu has quit IRC (Ping timeout: 265 seconds) 2012-07-09T22:41:04 *** xiangfu has joined #rtems 2012-07-09T23:24:38 *** arvind_khadri has joined #rtems 2012-07-10T00:02:14 *** QingPei has left #rtems 2012-07-10T00:53:26 *** hiddenpearls has joined #rtems 2012-07-10T01:04:59 *** kiwichris has joined #rtems 2012-07-10T01:08:17 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-10T01:41:26 *** WikL has joined #rtems 2012-07-10T01:54:12 *** sebhub has joined #rtems 2012-07-10T02:23:53 good morning 2012-07-10T02:31:15 *** xiangfu has quit IRC (Ping timeout: 250 seconds) 2012-07-10T02:39:12 *** xiangfu has joined #rtems 2012-07-10T03:20:02 *** hiddenpearls has joined #rtems 2012-07-10T03:23:36 *** zw_yao has joined #rtems 2012-07-10T03:43:56 *** kiwichris_ has joined #rtems 2012-07-10T03:47:05 *** kiwichris has quit IRC (Ping timeout: 250 seconds) 2012-07-10T04:28:36 *** kiwichris__ has joined #rtems 2012-07-10T04:30:08 *** kiwichris_ has quit IRC (Ping timeout: 246 seconds) 2012-07-10T04:46:48 *** xiangfu has quit IRC (Ping timeout: 245 seconds) 2012-07-10T04:51:06 *** arvind_khadri has quit IRC (Ping timeout: 264 seconds) 2012-07-10T05:08:46 *** arvind_khadri has joined #rtems 2012-07-10T05:09:52 *** arvind_khadri has joined #rtems 2012-07-10T05:11:22 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-07-10T05:11:24 *** zw_yao has quit IRC (Read error: Connection reset by peer) 2012-07-10T05:13:12 *** arvind_khadri has joined #rtems 2012-07-10T06:00:27 *** kiwichris_ has joined #rtems 2012-07-10T06:01:18 *** kiwichris__ has quit IRC (Ping timeout: 252 seconds) 2012-07-10T06:29:49 *** kiwichris__ has joined #rtems 2012-07-10T06:31:44 *** kiwichris_ has quit IRC (Ping timeout: 252 seconds) 2012-07-10T06:56:08 *** zw_yao has joined #rtems 2012-07-10T08:00:01 *** hiddenpearls1 has joined #rtems 2012-07-10T08:01:16 *** hiddenpearls has quit IRC (Ping timeout: 265 seconds) 2012-07-10T08:14:54 *** hiddenpearls has joined #rtems 2012-07-10T08:16:54 *** hiddenpearls1 has quit IRC (Ping timeout: 264 seconds) 2012-07-10T08:21:56 *** hiddenpearls has quit IRC (Ping timeout: 248 seconds) 2012-07-10T08:23:21 *** hiddenpearls1 has joined #rtems 2012-07-10T08:33:34 *** DrJoel has joined #rtems 2012-07-10T08:33:34 *** ChanServ sets mode: +o DrJoel 2012-07-10T08:36:41 sebhub, when is thomas on holiday? 2012-07-10T08:57:31 hi DrJoel 2012-07-10T08:57:31 *** zw_yao has quit IRC (Read error: Connection reset by peer) 2012-07-10T08:59:08 *** kuzew has quit IRC (Ping timeout: 252 seconds) 2012-07-10T08:59:44 *** kuzew has joined #rtems 2012-07-10T09:00:21 hi juli1 2012-07-10T09:05:37 *** kuzew has quit IRC (Read error: Operation timed out) 2012-07-10T09:15:05 *** kuzew has joined #rtems 2012-07-10T09:20:51 DrJoel: so, did you try the libbsd stuff with real hard ? 2012-07-10T09:21:30 juli1, not yet. still wanting a ping to work on qemu 2012-07-10T09:22:41 DrJoel: I understand 2012-07-10T09:22:57 on my side, I just discover that RTnet under linux is not up to date 2012-07-10T09:29:07 :( it is always something. 2012-07-10T09:35:22 no but the driver for the realtek board has not been updated since ... 2002 2012-07-10T09:35:56 that is always the problem with projects that adapt the driver for their own needs and when you do not have a broad audience 2012-07-10T09:36:08 that is why having the libbsd approach is a good way 2012-07-10T10:21:37 *** arvind_khadri has quit IRC (Ping timeout: 240 seconds) 2012-07-10T10:21:54 *** gedare has joined #rtems 2012-07-10T10:28:27 sorry to bother: I'm trying to add an include path to the compiler and adding CPU_CFLAGS=-I/path/ to the $BSP.cfg file inside $BSP/make/custom/ isn't recognized. Is this not the proper way to do this? 2012-07-10T10:30:31 CFLAGS itself might work 2012-07-10T10:30:56 won't it overwrite the original CFLAGS? 2012-07-10T10:31:35 er, oh sorry yeah. you added the -I to CPU_CFLAGS? 2012-07-10T10:31:47 yes 2012-07-10T10:32:10 not sure why that doesn't work. but you might be able to pass CFLAGS to configure that might not overwrite it... 2012-07-10T10:32:26 i'm not sure about the standard way to add an include path 2012-07-10T10:33:22 i think the CPU_CFLAGS is only used for the bsp files... 2012-07-10T10:33:54 it's been a little while since i mucked around with them though 2012-07-10T10:34:52 the file on which the compilation breaks is inside c/src/lib/libbsp/$ARCH/$BSP/ 2012-07-10T10:35:27 even when adding rubish to CFLAGS or CPU_FLAGS, it doesn't show up in the compilation command while running make 2012-07-10T10:36:27 juli1, We learned our lesson the hard way on how to deal with the TCP/IP stack. The rtems-libbsd code also includes the USB stack. 2012-07-10T10:36:39 if you just do make it won't pull the modified .cfg file... 2012-07-10T10:36:47 you have to rerun configure 2012-07-10T10:37:04 otherwise it uses the .cfg file it finds in your build tree 2012-07-10T10:37:08 rather than your source tree 2012-07-10T10:37:20 The hope is that as time goes on, we can more easily update to newer versions and possibly pull in more functionality. Wifi seems to be of interest and USB NIC adapters would be nice 2012-07-10T10:37:56 WikL, What are you trying to include? 2012-07-10T10:38:07 I'll check it again, but I suspected this and it seemed like the .cfg was pulled just with make 2012-07-10T10:38:23 trying to include the header files for POK 2012-07-10T10:38:38 What do you need besides syscalls for the BSP? 2012-07-10T10:39:37 WikL: why do you need the header ? 2012-07-10T10:39:42 just declare the extern function ! 2012-07-10T10:39:43 right now nothing, only the syscall prototypes and some typedefs which it requires 2012-07-10T10:39:58 oh my.. Centos 6 apparently got a BIG update overnight. 381 RPMs coming down now 2012-07-10T10:40:13 Personally I would add those to the BSP for now 2012-07-10T10:40:13 WikL: do the fast way : include the definition of the function and the typedef in your bsp 2012-07-10T10:40:20 and this is it 2012-07-10T10:40:22 I mean 2012-07-10T10:40:24 the prototype for the syscall requires an enum from one of the headers 2012-07-10T10:40:24 the point is 2012-07-10T10:40:30 we want to pass the mid term review 2012-07-10T10:40:41 WikL: include the enum 2012-07-10T10:40:44 We do that for syscall support for simulators. Otherwise, you need code from newlib, gdb, skyeye, qemu, etc 2012-07-10T10:41:10 hack :) 2012-07-10T10:41:20 yes 2012-07-10T10:41:22 right now for the syscall I already made I just pasted the required definitions inside the file 2012-07-10T10:41:29 but it's a dirty hack 2012-07-10T10:41:32 and ? 2012-07-10T10:41:37 yeah it works 2012-07-10T10:41:39 we do not care 2012-07-10T10:41:41 so 2012-07-10T10:41:42 ok then 2012-07-10T10:41:43 update the doc 2012-07-10T10:41:44 send 2012-07-10T10:41:54 I give me the proof we have a first demo in hands 2012-07-10T10:42:08 getting right on it 2012-07-10T10:42:40 thanks 2012-07-10T10:43:48 [rtems-libbsd commit] Makefile: updated to probably install architecture specific .h files 2012-07-10T10:43:55 gedare, desperately stuck on how to provide user space .h files for rtems-libbsd. :( 2012-07-10T10:44:04 * DrJoel laughs at own stupid comment 2012-07-10T10:44:05 going to choose whether to install or not? 2012-07-10T10:44:06 ;) 2012-07-10T10:44:21 i like that kind of build system 2012-07-10T10:44:22 I found another mistake in the install patterns this morning... sigh... 2012-07-10T10:44:30 probabilistic make! 2012-07-10T10:44:44 DrJoel: I don't quite understand the problem so i have not great insight 2012-07-10T10:44:51 the freebsd files are all under and user space is supposed to be 2012-07-10T10:45:12 Installed as not 2012-07-10T10:45:23 ok 2012-07-10T10:45:44 DrJoel: and ? 2012-07-10T10:45:52 DrJoel: Add a -Ifreebsd/ 2012-07-10T10:45:57 and this is it, no ? 2012-07-10T10:45:57 Plus some magic RTEMS .h files to rename stuff.. kernel code builds OK. 2012-07-10T10:46:27 I hope so.. if it doesn't conflict with include other prequisites... 2012-07-10T10:46:50 seems like it shouldn't if the kernel include file works 2012-07-10T10:47:12 otherwise would #include "freebsd/file.h" work? 2012-07-10T10:48:59 seems to be an issue with /home/joel/newbsd/install/include/freebsd/machine/_types.h:27:2: error: #error "the header file must be included first" 2012-07-10T10:49:02 Even if I do this: 2012-07-10T10:49:15 #include 2012-07-10T10:49:15 #include 2012-07-10T10:49:25 I guess I can keep digging.. 2012-07-10T10:52:32 odd. 2012-07-10T10:53:11 I'll keep looking.. it is always hard to sort out conflicts between BSD .h files and newlib/RTEMS .h files. Weird side-effects 2012-07-10T10:58:11 Full path works. I think when you add -I.../freebsd it ends up picking up BSD version of .h files instead of ours. 2012-07-10T10:58:40 something to discuss with Sebastirn 2012-07-10T11:09:03 *** WikL has quit IRC (Ping timeout: 250 seconds) 2012-07-10T11:09:45 *** WikL has joined #rtems 2012-07-10T11:27:29 *** gedare has quit IRC (Quit: Ex-Chat) 2012-07-10T11:47:22 *** sebhub has quit IRC (Remote host closed the connection) 2012-07-10T12:11:43 *** Hesham has joined #rtems 2012-07-10T13:06:13 *** Hesham has quit IRC (Ping timeout: 248 seconds) 2012-07-10T13:23:40 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2012-07-10T13:23:53 *** lcpfnvcy has quit IRC (Ping timeout: 245 seconds) 2012-07-10T13:27:58 *** hiddenpearls has joined #rtems 2012-07-10T13:33:40 *** Hesham1 has joined #rtems 2012-07-10T13:39:14 *** WikL has quit IRC () 2012-07-10T13:42:26 *** Hesham1 has quit IRC (Ping timeout: 250 seconds) 2012-07-10T13:42:33 *** Hesham has joined #rtems 2012-07-10T14:01:59 *** Hesham has quit IRC (Ping timeout: 255 seconds) 2012-07-10T14:58:24 *** deb has left #rtems ("Leaving") 2012-07-10T15:24:44 *** lcpfnvcy has joined #rtems 2012-07-10T16:14:10 hi all 2012-07-10T16:18:17 hey chris.. 2012-07-10T16:18:40 learning that the Python import script fixes things for kernel space and isn't good for user space. 2012-07-10T16:19:24 Some .h files like are shared with kernel and got freebsd/ stuck in front. Trying to find a path through user space now. 2012-07-10T16:19:30 *** hiddenpearls1 has joined #rtems 2012-07-10T16:19:49 Ah that is not great. I suppose simpler to sort now you have things working. 2012-07-10T16:21:49 Jennifer is sick today but had stack quiescent after fxp_attach. Now trying to get the user space side together enough to set IP address, gateway, etc 2012-07-10T16:22:12 Great to see. 2012-07-10T16:22:31 http://www.facebook.com/photo.php?fbid=10151900846570046&set=a.10151900845920046.876450.100681630045&type=1 2012-07-10T16:22:44 Finally rain here. That's downtown Chattanooga a couple of hours ago. 2012-07-10T16:23:13 click through to see more pictures. Pounding here. 2012-07-10T16:23:27 *** hiddenpearls has quit IRC (Ping timeout: 265 seconds) 2012-07-10T16:24:43 Some great shots. 2012-07-10T16:25:46 :) 2012-07-10T16:59:32 DrJoel, This one is super https://www.facebook.com/photo.php?fbid=10151900846360046&set=a.10151900845920046.876450.100681630045&type=1&permPage=1 2012-07-10T17:02:22 :) 2012-07-10T17:02:27 *** DrJoel has quit IRC (Quit: Leaving) 2012-07-10T17:04:05 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2012-07-10T17:25:08 *** hiddenpearls has joined #rtems 2012-07-10T17:43:09 *** sevikkk1 has quit IRC (Remote host closed the connection) 2012-07-10T17:50:02 *** sevikkk has joined #rtems 2012-07-10T18:31:31 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-10T18:52:11 *** hiddenpearls has joined #rtems 2012-07-10T18:52:50 *** hiddenpearls has left #rtems 2012-07-10T18:53:04 *** kiwichris__ has quit IRC (Quit: This computer has gone to sleep) 2012-07-10T18:53:46 *** kiwichris__ has joined #rtems 2012-07-10T19:02:05 *** kiwichris__ has quit IRC (Quit: This computer has gone to sleep) 2012-07-10T19:02:55 *** kiwichris__ has joined #rtems 2012-07-10T19:05:32 *** kiwichris has joined #rtems 2012-07-10T20:11:21 *** xiangfu has joined #rtems 2012-07-10T20:16:05 *** xiangfu has quit IRC (Ping timeout: 248 seconds) 2012-07-10T20:28:59 *** kiwichris has quit IRC (Read error: Connection reset by peer) 2012-07-10T20:29:14 *** xiangfu has joined #rtems 2012-07-10T20:30:16 *** kiwichris has joined #rtems 2012-07-10T20:34:21 *** kiwichris_ has joined #rtems 2012-07-10T20:35:37 *** kiwichris has quit IRC (Ping timeout: 240 seconds) 2012-07-10T21:53:10 *** zw_yao has joined #rtems 2012-07-10T22:19:43 *** zw_yao has quit IRC (Ping timeout: 245 seconds) 2012-07-10T23:19:17 *** xiangfu has quit IRC (Ping timeout: 255 seconds) 2012-07-10T23:19:41 *** xiangfu has joined #rtems 2012-07-10T23:22:07 *** kiwichris__ has joined #rtems 2012-07-10T23:22:07 *** kiwichris_ has quit IRC (Read error: Connection reset by peer) 2012-07-10T23:42:42 *** xiangfu has quit IRC (Ping timeout: 264 seconds) 2012-07-11T00:00:32 *** xiangfu has joined #rtems 2012-07-11T00:20:54 *** Hesham has joined #rtems 2012-07-11T00:22:06 *** Hesham1 has joined #rtems 2012-07-11T00:23:45 *** Hesham2 has joined #rtems 2012-07-11T00:25:27 *** Hesham has quit IRC (Ping timeout: 250 seconds) 2012-07-11T00:26:47 *** Hesham1 has quit IRC (Ping timeout: 265 seconds) 2012-07-11T00:29:37 *** Hesham2 has quit IRC (Ping timeout: 240 seconds) 2012-07-11T00:49:28 *** arvind_khadri has joined #rtems 2012-07-11T01:04:17 *** kiwichris has joined #rtems 2012-07-11T01:04:38 *** kiwichris__ has quit IRC (Ping timeout: 255 seconds) 2012-07-11T01:31:41 *** sebhub has joined #rtems 2012-07-11T01:38:39 good morning 2012-07-11T01:45:16 *** WikL has joined #rtems 2012-07-11T02:11:39 *** zw_yao has joined #rtems 2012-07-11T02:26:29 *** shineworld has joined #rtems 2012-07-11T02:34:57 *** WikL has quit IRC () 2012-07-11T02:52:08 *** WikL has joined #rtems 2012-07-11T02:53:14 *** shineworld has left #rtems ("Sto andando via") 2012-07-11T03:28:09 *** kristianpaul has quit IRC (Read error: Connection reset by peer) 2012-07-11T03:39:17 *** kristianpaul has joined #rtems 2012-07-11T03:58:40 *** kristianpaul has quit IRC (Remote host closed the connection) 2012-07-11T03:58:57 *** kristianpaul has joined #rtems 2012-07-11T04:08:41 *** zw_yao has quit IRC (Read error: Connection reset by peer) 2012-07-11T04:44:28 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-11T04:45:14 *** kristianpaul has joined #rtems 2012-07-11T04:54:32 *** kristianpaul has quit IRC (Remote host closed the connection) 2012-07-11T05:05:32 *** kristianpaul has joined #rtems 2012-07-11T05:11:19 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-07-11T05:12:28 *** kristianpaul has joined #rtems 2012-07-11T05:31:03 *** kristianpaul has quit IRC (Remote host closed the connection) 2012-07-11T05:37:12 *** kristianpaul has joined #rtems 2012-07-11T06:29:09 *** kiwichris_ has joined #rtems 2012-07-11T06:31:38 *** kiwichris has quit IRC (Ping timeout: 246 seconds) 2012-07-11T06:55:40 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-11T06:56:13 *** kristianpaul has joined #rtems 2012-07-11T07:09:11 *** kristianpaul has quit IRC (Remote host closed the connection) 2012-07-11T07:09:19 *** kristianpaul has joined #rtems 2012-07-11T07:17:04 *** gedare has joined #rtems 2012-07-11T07:18:27 *** kristianpaul has quit IRC (Remote host closed the connection) 2012-07-11T07:36:38 *** cdcs has joined #rtems 2012-07-11T07:43:00 *** kristianpaul has joined #rtems 2012-07-11T08:05:29 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-07-11T08:05:44 *** kristianpaul has joined #rtems 2012-07-11T08:51:59 *** Guest70904 has quit IRC (Quit: Leaving) 2012-07-11T09:12:49 *** xiangfu has quit IRC (Remote host closed the connection) 2012-07-11T09:54:16 *** arvind_khadri has quit IRC (Quit: Leaving) 2012-07-11T09:05:12 *** rtemsLogger has joined #rtems 2012-07-11T10:21:52 *** WikL has quit IRC () 2012-07-11T10:47:31 *** sebhub has quit IRC (Remote host closed the connection) 2012-07-11T12:01:46 *** gedare has quit IRC (Quit: Leaving) 2012-07-11T12:45:24 *** cdcs has quit IRC (Quit: Page closed) 2012-07-11T13:26:39 *** DrJoel has joined #rtems 2012-07-11T13:26:39 *** DrJoel has joined #rtems 2012-07-11T13:26:39 *** ChanServ sets mode: +o DrJoel 2012-07-11T14:20:25 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-07-11T14:26:19 *** kristianpaul has joined #rtems 2012-07-11T16:57:57 *** ppisa has joined #rtems 2012-07-11T17:29:49 *** kiwichris has joined #rtems 2012-07-11T17:30:33 *** kiwichris_ has quit IRC (Ping timeout: 245 seconds) 2012-07-11T17:35:32 *** ppisa has left #rtems ("Kopete 0.12.7 : http://kopete.kde.org") 2012-07-11T17:55:30 *** DrJoel has quit IRC (Quit: Leaving) 2012-07-11T18:22:46 *** kiwichris_ has joined #rtems 2012-07-11T18:24:40 *** kiwichris has quit IRC (Ping timeout: 246 seconds) 2012-07-11T18:41:52 *** kiwichris__ has joined #rtems 2012-07-11T18:41:52 *** kiwichris_ has quit IRC (Read error: Connection reset by peer) 2012-07-11T19:41:56 *** xiangfu has joined #rtems 2012-07-11T21:23:52 *** zw_yao has joined #rtems 2012-07-11T21:24:44 *** zw_yao has joined #rtems 2012-07-11T23:38:28 *** kiwichris_ has joined #rtems 2012-07-11T23:40:28 *** kiwichris__ has quit IRC (Ping timeout: 248 seconds) 2012-07-12T00:03:59 *** zw_yao has quit IRC (Remote host closed the connection) 2012-07-12T00:05:26 *** A2Sheds has quit IRC (Quit: puff of smoke) 2012-07-12T00:25:07 *** L84Supper has joined #rtems 2012-07-12T00:58:46 *** arvind_khadri has joined #rtems 2012-07-12T01:03:41 *** zw_yao has joined #rtems 2012-07-12T01:52:28 *** sebhub has joined #rtems 2012-07-12T01:53:09 good morning 2012-07-12T02:20:01 *** alseh has joined #rtems 2012-07-12T02:57:55 *** alseh has quit IRC (Ping timeout: 244 seconds) 2012-07-12T03:42:14 *** kiwichris__ has joined #rtems 2012-07-12T03:43:59 *** kiwichris_ has quit IRC (Ping timeout: 246 seconds) 2012-07-12T06:29:00 *** kiwichris has joined #rtems 2012-07-12T06:31:07 *** kiwichris__ has quit IRC (Ping timeout: 240 seconds) 2012-07-12T06:43:41 *** alseh has joined #rtems 2012-07-12T06:48:26 *** sebhub has quit IRC (Ping timeout: 246 seconds) 2012-07-12T07:04:22 *** arvind_khadri has quit IRC (Ping timeout: 244 seconds) 2012-07-12T07:18:14 *** zw_yao has quit IRC (Ping timeout: 250 seconds) 2012-07-12T07:24:44 *** arvind_khadri has joined #rtems 2012-07-12T07:25:16 *** WikL has joined #rtems 2012-07-12T07:32:12 *** zw_yao has joined #rtems 2012-07-12T07:34:21 *** zw__yao has joined #rtems 2012-07-12T07:36:44 *** zw_yao has quit IRC (Ping timeout: 248 seconds) 2012-07-12T07:42:40 *** antgreen has joined #rtems 2012-07-12T07:44:37 *** antgreen has quit IRC (Remote host closed the connection) 2012-07-12T07:45:01 *** antgreen has joined #rtems 2012-07-12T08:02:13 *** zw__yao has quit IRC (Read error: Connection reset by peer) 2012-07-12T08:16:04 *** jennifer has joined #rtems 2012-07-12T08:16:28 *** jennifer is now known as Guest54582 2012-07-12T08:18:27 *** QingPei has joined #rtems 2012-07-12T08:18:32 *** jenniferA has joined #rtems 2012-07-12T08:23:16 *** DrJoel has joined #rtems 2012-07-12T08:23:16 *** ChanServ sets mode: +o DrJoel 2012-07-12T08:27:05 *** sebhub has joined #rtems 2012-07-12T08:30:59 *** xiangfu has quit IRC (Ping timeout: 246 seconds) 2012-07-12T08:40:20 *** weiY has joined #rtems 2012-07-12T08:40:59 *** ppisa has joined #rtems 2012-07-12T08:42:33 I see one mentor evaluation outstanding 2012-07-12T08:42:47 And one student evaluation 2012-07-12T08:43:07 nothing else ? 2012-07-12T08:43:54 Hi 2012-07-12T08:44:04 hi kristianpaul 2012-07-12T08:45:29 juli1, that's all I see that has not been done. I can name names :) 2012-07-12T08:46:04 I was looking an rtems app that uses semphores and event flags.. 2012-07-12T08:46:04 Need to learn by example.. 2012-07-12T08:46:26 WikL, Have you figured out how to get memory address and length for RTEMS Workspace from Pok? 2012-07-12T08:48:07 Hi joel, you see just one student evaluation? 2012-07-12T08:48:36 I see one missing student evaluation and one missing mentor evaluation 2012-07-12T08:48:48 DrJoel: I think I did all the required work regarding the mid term review, no ? 2012-07-12T08:49:00 you did. :) 2012-07-12T08:49:14 Joel, how about my evaluation, i have submitted my evaluation 2012-07-12T08:49:50 weiY, It is in. We only see that much 2012-07-12T08:50:58 Ok 2012-07-12T08:51:34 There is about one hour left to meeting 2012-07-12T08:53:32 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-12T08:59:26 *** kristianpaul has joined #rtems 2012-07-12T09:02:40 *** antgreen has quit IRC (Read error: Connection reset by peer) 2012-07-12T09:13:20 *** gedare has joined #rtems 2012-07-12T09:22:13 morning 2012-07-12T09:26:03 gedare, do you see one missing student eval and one missing mentor eval? 2012-07-12T09:26:20 yep 2012-07-12T09:26:24 i pinged the mentor 2012-07-12T09:26:52 and the student has committed to doing it today 2012-07-12T09:35:16 *** zw_yao has joined #rtems 2012-07-12T09:38:20 *** QingPei has quit IRC (Ping timeout: 244 seconds) 2012-07-12T09:43:19 *** tuxmaniac has joined #rtems 2012-07-12T09:45:28 *** arvind_khadri has quit IRC (Quit: Leaving) 2012-07-12T09:45:34 *** panzon has quit IRC (Remote host closed the connection) 2012-07-12T09:53:25 fwiw this meeting is quick.. everyone sent in mid-term reports.. focus on any hurdles you need mentor to help address. Action items 2012-07-12T09:55:06 It would worth to discuss about rtems-graphics-toolkit upadates, Makefiles etc, when time allows. 2012-07-12T09:55:28 ppisa, that would be fine. We just don't need a point by point status from anyone. :) 2012-07-12T09:56:20 ppisa, the library ports were pretty easy initially so updates should be painless.. but he has a primary goal of NXLib running, right? 2012-07-12T09:56:53 * tuxmaniac announces his presence 2012-07-12T09:56:56 \o/ 2012-07-12T09:57:18 actually, I have NXLib running through FLTK 2012-07-12T09:57:35 hi tuxmaniac 2012-07-12T09:57:55 *** QingPei has joined #rtems 2012-07-12T09:58:01 alseh, :) Then you are due for some code submission, review and merge.. 2012-07-12T09:58:20 you probably have time for updating libraries and maybe a cool demo of some sort 2012-07-12T09:59:21 Yes there is time for that, but we should be in clean state enough to push it to RTEMS version. 2012-07-12T09:59:49 I have still to clean up code a little, as ppisa is saying 2012-07-12T09:59:53 For now, some tests from fltk work, because external image or html files are needed. 2012-07-12T09:59:55 hi 2012-07-12T10:00:36 work or work not? 2012-07-12T10:00:39 NXlib is not ./configure based, so do you and Gedare, others agree to add rtems-graphics-toolkit/nxlib/Makefile.rtems 2012-07-12T10:00:43 for now I use the filesystem object from microwindows. However, I think it should be external to microwindows, to allow user to add more easily files 2012-07-12T10:00:44 getting started joel? 2012-07-12T10:01:09 i'm not particular about how a thing is built, as long as the build system is maintainable... but others may have complaints 2012-07-12T10:01:14 DrJoel: I'm still looking into how to get the memory information 2012-07-12T10:01:21 OK can we reserve slot with Joel and Gedare after meeting. 2012-07-12T10:01:25 rtems has the ability to add tar files into the IMFS 2012-07-12T10:01:28 yes gedare sorry 2012-07-12T10:01:30 i'll be around 2012-07-12T10:01:38 I guess alseh and ppisa are first 2012-07-12T10:02:09 yes i exclusively use tar files and imfs to get large data in to applications 2012-07-12T10:02:17 but that's because my environment is evil. 2012-07-12T10:02:35 nfs is also nice for development 2012-07-12T10:02:35 * DrJoel doesn't want to rewrite some other project's build infrastructure. The graphics toolkit build system is a simple wrapper to try to hide the differences and give you "all, clean, install" 2012-07-12T10:03:18 *** medivhc has joined #rtems 2012-07-12T10:03:34 thet's why I created Makefile.rtems, however, chamges to Makefile were also needed 2012-07-12T10:03:44 NFS is nice but not on every target. The test cases should not require network connection for files and should not require MicroWindows for files unless they are really using MicroWindows for graphics 2012-07-12T10:03:52 Do not introduce unnecessary dependencies 2012-07-12T10:03:52 *** soh_cah_toa has joined #rtems 2012-07-12T10:03:53 Yes, but this is how Microwin is done now rtems-graphics-toolkit/microwin/src/Makefile.rules 2012-07-12T10:04:14 Agree files should go into embedded tar. 2012-07-12T10:04:33 yes, i would prefer that also ;-) 2012-07-12T10:04:52 The access is same as for usual FS. 2012-07-12T10:05:05 so this is the right waty. 2012-07-12T10:05:34 Is the graphics driver the next thing to review and merge? Did the questions get addressed? 2012-07-12T10:05:55 No response for driver yet. 2012-07-12T10:06:00 waiting on licensing 2012-07-12T10:06:05 I try ask persons again. 2012-07-12T10:06:49 If no reply, then Alex can try to write two problematic functions according to the specs. 2012-07-12T10:06:59 Is there something equivalent from BSD land that can replace it? 2012-07-12T10:07:15 I could check 2012-07-12T10:07:43 Either way, once it is clean, we want to review and merge it 2012-07-12T10:07:46 I have looked but it seems that BSD is not able to use QEMU cirrus directly. But I have spent only short trime on that. 2012-07-12T10:08:10 ok ppisa .. then a clean re-write might be the only path 2012-07-12T10:09:09 have you all updated any libraries in the toolkit at this point that need merging? 2012-07-12T10:09:23 * DrJoel hates to get deluged with code at the end of the summer. :) 2012-07-12T10:09:41 *** cdcs has joined #rtems 2012-07-12T10:10:21 Yes the code should be merged. The question is, if we should keep NXlib separate same as Microwindows or integrate it into rtems-graphics-toolkit GIT 2012-07-12T10:10:54 I have written more thoughts in my yesterday e-mails. 2012-07-12T10:11:00 maybe we can use git sub-modules 2012-07-12T10:11:37 i think it is a fine idea. 2012-07-12T10:11:40 That is what I have written about. I like them much, but I know, that it is nightmare for GIT novice users. 2012-07-12T10:11:47 if the patches are merged upstream and it is a simple checkout, then no need to mirror. Greg Haer was good at merging Roxana's changes last time so we didn't have a delta. and it is a SLOW project 2012-07-12T10:11:52 if the instructions are clear it should not matter 2012-07-12T10:12:04 hmm, or that. 2012-07-12T10:12:07 or hidden in a script :) 2012-07-12T10:12:28 we should move to another student 2012-07-12T10:12:37 come back to this later 2012-07-12T10:12:45 OK 2012-07-12T10:12:51 fine by me :) 2012-07-12T10:13:44 I'm always OK to be next:) 2012-07-12T10:13:58 alseh and ppisa do you two have a "todo" list google doc which can be shared? This project has a lot of pieces and they are largely independent. 2012-07-12T10:14:12 zw_yao, What's your next challenge.. status is not needed 2012-07-12T10:14:36 Alex initial proposal but we should create another one, what about Wiki? 2012-07-12T10:14:52 It is the most transparent and accesible. 2012-07-12T10:15:25 ppisa, that's good. Just want something to track coding, patches upstream, issues, review, merge, on pieces 2012-07-12T10:15:36 I think my problem is I don't know how to test my one_rbtree approach to ensure it work correctly. 2012-07-12T10:16:08 Do all current tests pass? 2012-07-12T10:16:15 Yes. 2012-07-12T10:16:37 gedare, does a coverage run make sense next? 2012-07-12T10:16:49 And I also add another 3 tests to test the multi-pthreads multi-keys condition. 2012-07-12T10:17:30 coverage run may help.. 2012-07-12T10:17:33 Have you added a key test with unlimited posix threads with low allocation unit so the old code would break? 2012-07-12T10:17:42 i am also interested in coverage reports due to my file system changes 2012-07-12T10:17:47 And another for Classic API tasks instead of threads 2012-07-12T10:17:53 general coverage run would be good ;) 2012-07-12T10:18:07 i've never managed to run it myself successfully 2012-07-12T10:18:11 sebhub, duly noted.. will try to get some sparc results posted ASAP .. today is my anniversary though :) 2012-07-12T10:18:19 nice 2012-07-12T10:18:48 those two test cases would be good 2012-07-12T10:18:55 especially if they show breakage in the old api 2012-07-12T10:19:00 err.. does coverage refer to all current tests? 2012-07-12T10:19:34 it's a run of the rtems tests to determine how much of the code paths of rtems are "covered" 2012-07-12T10:19:47 zw_yao, yes 2012-07-12T10:19:55 basically it shows if something is implemented in rtems that is not being used 2012-07-12T10:19:55 Oh, I see. 2012-07-12T10:19:59 so a baseline run would be what is in git 2012-07-12T10:20:24 then I would apply your patches, run again and we would see where your changes didn't get executed 2012-07-12T10:20:24 but... that is less an issue for you until it actually gets run. we farm that task out to drjoel ;) 2012-07-12T10:20:54 so for you zw_yao i think i'd like to see a test case using unlimited keys..that would break the old api -- as mentioned 2012-07-12T10:21:07 does classic api tasks matter? 2012-07-12T10:21:15 this is a posix-y thing... 2012-07-12T10:21:25 OK, I'll think about that. 2012-07-12T10:21:34 the two unlimited test cases should let each task run, set it, get it, sleep and then delete themselves.. need as many tasks/threads in system as the BSP's memory allows, then let them each do some key stuff, 2012-07-12T10:21:53 May want two more variations where the tasks do not exit but you delete the key. 2012-07-12T10:23:11 So that's 4 new tests if you don't have them already. Those are "unlimited" Classic API tasks and POSIX threads as one variable, key delete versus thread delete self. That gives 4 cases. 2012-07-12T10:23:38 *** kuzew has quit IRC (Ping timeout: 252 seconds) 2012-07-12T10:23:56 OK, I get it. I'll discuss the detail on ml. 2012-07-12T10:23:57 If you have tests already which can run against the current implementation, then we should review and merge those -- even if they highlight failures 2012-07-12T10:24:08 ok.. who's got the next challenge? 2012-07-12T10:24:21 I can go 2012-07-12T10:24:41 *** Hesham has joined #rtems 2012-07-12T10:24:49 if I may 2012-07-12T10:24:53 I have none such test yet. 2012-07-12T10:24:54 sure.. we have beaten you up in email. :) Do you know the next steps? 2012-07-12T10:25:19 Did you begin the meeting ? 2012-07-12T10:25:21 *** kuzew has joined #rtems 2012-07-12T10:25:24 beating deserved 2012-07-12T10:25:49 well I have to figure out the memory thing 2012-07-12T10:25:54 how to get bsp_get_work_area working 2012-07-12T10:26:00 zw_yao, we can discuss with gedare how I can get a coverage run for you. If you are on github, I can probably just check it and do it 2012-07-12T10:26:22 yes Hesham we're on wikl 2012-07-12T10:26:40 I don't know how to get that info in Pok. Hopefully juli1 can tell us and it is a simple syscall or symbols in linkcmds. 2012-07-12T10:26:43 I see there's a structure in POK containing information about partitions, including memory 2012-07-12T10:26:47 All of my code has been push to github. 2012-07-12T10:27:09 Can i be next as i have to travel today ? 2012-07-12T10:27:16 ok zw_yao after I publish baseline on head, email gedare and I when you think you want me to run it 2012-07-12T10:27:24 WikL, ok with you? 2012-07-12T10:27:27 Hesham, go 2012-07-12T10:27:31 sure sure 2012-07-12T10:27:36 thanks DrJoel 2012-07-12T10:27:41 OK, I see. WikL go head. 2012-07-12T10:28:12 well Hesham now :) 2012-07-12T10:28:32 Well, Now i have two versions of MMU work (Quanming) and Gedare modified version 2012-07-12T10:29:06 Quanming works well and raise mmu exceptions when access violation occues 2012-07-12T10:29:19 But modified work does not raise exceptions 2012-07-12T10:29:23 zw_yao: I've been yielded 2012-07-12T10:30:04 i have debugged that problem and find that searching for valid PTE at pagetable always fails 2012-07-12T10:30:20 that's the main challenge. so Hesham has to back off and redo the libmm conversion so that mutliple cpu/archs can benefit from libmm.... but use the quanming version as a baseline since he can test it for correctness on ppc now 2012-07-12T10:30:23 since he has ppc working 2012-07-12T10:30:43 yeah 2012-07-12T10:30:49 sounds like a solvable problem.. just tedious 2012-07-12T10:30:51 unfortunate but we could not detect the problem until now because of trouble with psim 2012-07-12T10:30:52 yea 2012-07-12T10:31:17 so that's the main thing to do now... and resolve how to design the libmm a little better 2012-07-12T10:31:26 anything else Hesham 2012-07-12T10:31:33 without working on Arena ? 2012-07-12T10:31:42 can't do anything with arena until libmm is fixed. 2012-07-12T10:31:45 no point to it 2012-07-12T10:31:55 right? 2012-07-12T10:31:57 Arena just uses pte 2012-07-12T10:32:40 and there is global ALUT corresponds to domain which PTEs resides 2012-07-12T10:33:23 i think we could implement Arena making use creating PTEs along with refactoring Mid-level API 2012-07-12T10:33:35 making use of ** 2012-07-12T10:34:54 ok it can go in parallel i suppose 2012-07-12T10:35:06 but focus on evolving libmm so that more architectures can be added 2012-07-12T10:35:13 while maintaining the ppc works 2012-07-12T10:35:55 DrJoel: i think we can move on 2012-07-12T10:36:02 ok thanks 2012-07-12T10:36:17 Hesham,sounds good 2012-07-12T10:36:20 WikL, back up 2012-07-12T10:36:37 so there is a Pok structure with this information. Is it accessible inside a partition? 2012-07-12T10:36:40 *** Hesham has left #rtems 2012-07-12T10:36:58 don't know yet, just found it 2012-07-12T10:37:16 juli1, is this information accessible within a partition? 2012-07-12T10:37:26 WikL, are you printing now with printk? 2012-07-12T10:37:32 yes 2012-07-12T10:38:16 then this memory info and the first context switch are the next challenges. I expect if you get the memory info righ, you will correctly run to the first context switch point. 2012-07-12T10:38:30 Once the first context switch works, you should have hello world 2012-07-12T10:39:29 yep 2012-07-12T10:40:11 WikL, ask Juli and other mentors via email best way to get that information. That will move you quickly to context switch and clock tick. cdcs expects clock tick to be hardest 2012-07-12T10:40:21 so move quickly through mem info and context switch 2012-07-12T10:40:27 who's next 2012-07-12T10:40:33 me. joel 2012-07-12T10:40:36 rigt 2012-07-12T10:40:39 go weiY 2012-07-12T10:40:43 *right 2012-07-12T10:41:12 I have a problem now about how to design the test case for atomic operations with memory barrier 2012-07-12T10:41:25 I have to think the failing scene of not using memory barrier with atomic operations. This involves 2012-07-12T10:41:26 runnig several tasks to access share date on mult-core which was influenced by memory barrier option. 2012-07-12T10:41:49 And concurrencykit memory barrier test case is not easy to ported to rtems. Because concurrencykit have different atomic API design. 2012-07-12T10:42:33 gedare, ideas? 2012-07-12T10:42:47 its hard to test for consistent failures like that 2012-07-12T10:43:07 first provide coverage tests 2012-07-12T10:43:11 yeah, so i have to think many scence to test it 2012-07-12T10:43:34 try not to get blocked on correctness test. it's important but can be hard 2012-07-12T10:43:34 ok, then to test is function? 2012-07-12T10:43:40 Please, read http://en.wikipedia.org/wiki/Memory_ordering for memory ordering and 2012-07-12T10:44:08 think ppisa, i have read this wiki 2012-07-12T10:44:28 and got some usefull information about memory order 2012-07-12T10:44:59 and next is great article from Paul McKenney http://www.rdrop.com/users/paulmck/scalability/paper/ordering.2007.09.19a.pdf 2012-07-12T10:45:42 If I know well your task it is to add missing GCC atomic functions 2012-07-12T10:45:58 There is example implementation on GCC pages. 2012-07-12T10:46:04 Is this paper is talking about cache coherence? 2012-07-12T10:46:49 Yes, i have add the atomic implementation but now i have to design the test case to test them 2012-07-12T10:46:56 Yes abot all possible aspect from loose Alpha to tight x86 and all necessary barriers optimized fro minimal overhead. 2012-07-12T10:47:35 first make sure you have test cases that at least call the atomics.. 2012-07-12T10:47:37 To find mistakes it allmost impossible by testing there. The time windows are extremelly short. 2012-07-12T10:47:42 ya 2012-07-12T10:47:49 fabricating the failure scenarios is hard 2012-07-12T10:48:00 yeah, so i think first step is to think all the case which may be influced by memory barrier 2012-07-12T10:48:09 impossible in cross platform tests I would imagine 2012-07-12T10:48:14 But in RTEMS case, do we tend for SMP or UP now. 2012-07-12T10:48:17 do you have test cases that just execute the atomic ops? 2012-07-12T10:48:40 In UP there is no problem with cache, only with interrupts. 2012-07-12T10:49:08 yeah, until now i have add seven test to run atomic operations with no memory barrier option 2012-07-12T10:49:24 and all test cases are passed 2012-07-12T10:50:02 does it make sense to have target independent UP version which is just interrupt disable? This could be default 2012-07-12T10:50:04 yeah, using atomic operation maybe slow the cpu 2012-07-12T10:50:21 Then this is something that is added as part of adding SMP support for an architecture 2012-07-12T10:50:24 on up 2012-07-12T10:50:44 Yes it has reason to have generic UP version with IRQ disable. 2012-07-12T10:50:47 freebsd uses this on some arm systems (interrupt disable for atomic ops) 2012-07-12T10:51:00 It is much, much cheaper. 2012-07-12T10:51:35 one rtems design philosophy is to provide same capabilities across all targets. With interrupt disable version by default with SMP not enabled, all existing BSPs get support instantly 2012-07-12T10:51:48 But even on UP, if cpu allows that, the load, mark , save check or cmpxchg version is fater. 2012-07-12T10:52:48 yes, i think there should add a SMP related macor 2012-07-12T10:52:52 GCC tries to use these instruction oin targets which have implemented them. It only calls helpers if machine isntructions are missing. 2012-07-12T10:53:14 to indicate run SMP or UP 2012-07-12T10:53:37 ARM above v5, all PowerePC, x86 (except old pentimum for long long) has them. 2012-07-12T10:53:41 weiY, if that makes sense, can you provide something similar? 2012-07-12T10:53:52 *** zw__yao has joined #rtems 2012-07-12T10:54:26 Joel , you mean GCC atomic way? 2012-07-12T10:55:10 You need barriers on SMP. You should read Paul's article what are required classes. Then each arch needs to define what it not necessary for given arch and how is implemented the rest. 2012-07-12T10:55:58 weiY, conceptually.. yes.. give default for each.. let architecture override individual methods or use default. 2012-07-12T10:56:07 But I have felling that you have to provide functions even when GCC can generate instructions for indirect calls, but can be made simple. 2012-07-12T10:56:27 *** AndChat-637364 has joined #rtems 2012-07-12T10:57:21 any students left? 2012-07-12T10:57:27 hi 2012-07-12T10:57:39 hi 2012-07-12T10:57:45 here 2012-07-12T10:57:46 hi medivhc . any other students? medivhc can go now i think 2012-07-12T10:57:56 ok, i will study it 2012-07-12T10:57:58 wrap up / send rtems-devel this discussion 2012-07-12T10:58:07 about the smp vs up, and barrier vs isr disable 2012-07-12T10:58:08 kevin .. which mips bsp are you trying now 2012-07-12T10:58:25 *** zw__yao has quit IRC (Ping timeout: 244 seconds) 2012-07-12T10:58:29 medivhc: what are your challenges/obstacles? 2012-07-12T10:59:36 I don't know how to describe the problem. The clock need a cirular queue. 2012-07-12T11:00:04 * DrJoel is heading to another meeting.. will try to watch 2012-07-12T11:00:26 i saw your fsbenchmark code... what other code are you working on? 2012-07-12T11:00:27 DrJoel: csb350 2012-07-12T11:01:44 I try to design the api. and test it using simple clock. 2012-07-12T11:02:21 was the design sent to rtems-devel ml? i don't remember it but maybe i missed seeing it 2012-07-12T11:02:46 Not yet. 2012-07-12T11:03:14 I talk some ideas on the ml 2012-07-12T11:04:00 medivhc, repost if needed .. make the subject clear that you need help 2012-07-12T11:04:05 my first thought is that state change from or to CACHED state will be the API. 2012-07-12T11:04:43 *** weiY has quit IRC () 2012-07-12T11:05:08 well i don't think i can get enough detail here to give any idea, you would need to send some details and questions to rtems-devel 2012-07-12T11:05:46 if the buffer is in access_cached state and extract from that cirular queue. The order will lost. That means the clock must maintain the all buffer in the cirular queue. 2012-07-12T11:06:19 Ok I will . 2012-07-12T11:07:06 you can additional nodes or other data structures to the rtems_bdbuf_buffer to maintain such a circular queue 2012-07-12T11:08:06 yes. but will it be inefficient? 2012-07-12T11:08:27 i don't know, what are the alternatives? 2012-07-12T11:09:01 you can use a global pointer table, but this leads to problems once you delete a buffer due to recycling 2012-07-12T11:09:03 The get release api make it complex. 2012-07-12T11:09:44 ok, can you please provide an overview about this on the mailing list 2012-07-12T11:10:07 because i don't know what the "get release api" is 2012-07-12T11:10:39 I mean the bdbuf api. 2012-07-12T11:11:04 the bdbuf api must not change, this should all work internally 2012-07-12T11:11:10 i think we need a detailed explanation of the problem and issues about solving it. 2012-07-12T11:11:35 I know. 2012-07-12T11:11:40 fwiw this seems late in coming 2012-07-12T11:12:16 hmm.. soh_cah_toa you're on deck. 2012-07-12T11:12:24 medivhc: anything else? 2012-07-12T11:12:31 Ok for me . 2012-07-12T11:12:43 ok. soh_cah_toa any major challenges to report on? 2012-07-12T11:13:37 we have been chatting privately.. sounds like keeping his branch up to date with the fast development is an issue 2012-07-12T11:13:38 if you s/major/minor/ then yes ;) 2012-07-12T11:13:57 yeah, since i'm using a fork 2012-07-12T11:14:56 i'm off for a bit to another meeting 2012-07-12T11:15:03 k 2012-07-12T11:15:08 i see. in your case you might like to just track the repo yourself 2012-07-12T11:15:33 *** medivhc has quit IRC (Quit: Leaving) 2012-07-12T11:15:41 i don't know can you add the primary repo as a remote and pull from it? 2012-07-12T11:15:46 seems like it should work 2012-07-12T11:16:06 he may be better off just cloning, doing a little, us merging it, then cloning again 2012-07-12T11:16:44 I added 116 files myself yesterday .. how can he track something that is moving that fast with a github mirror 2012-07-12T11:16:45 i can add git://git.rtems.org/rtems-libbsd.git as pull origin and git://github.com/soh-cah-toa/rtems-libbsd.git as a push origin 2012-07-12T11:16:55 and the github mirror hadn't been updating . 2012-07-12T11:16:56 call incoming 2012-07-12T11:16:57 but the that kinda messes up the logs 2012-07-12T11:17:00 k 2012-07-12T11:17:04 *** WikL has quit IRC () 2012-07-12T11:17:56 hmm.. ok 2012-07-12T11:18:29 i can see that. not sure what to tell you.. except the messed up logs on your github are mostly just a nuisance 2012-07-12T11:18:48 alright... 2012-07-12T11:18:48 yeah 2012-07-12T11:19:04 any other students? my head is starting to hurt... dentist numbed me too much i suspect its in my brain ;) 2012-07-12T11:19:14 but yeah, i posted my status to the rtems-gsoc ml last night. i'll do my midterm eval after i have some breakfast after the meeting 2012-07-12T11:19:22 great 2012-07-12T11:19:24 joel and jennifer and i will figure it out afterward 2012-07-12T11:19:25 you're the last student ;) 2012-07-12T11:19:29 ah ok 2012-07-12T11:19:34 k I'll be here 2012-07-12T11:19:46 i think that's the end of the meeting.. 2012-07-12T11:19:55 i'm sticking around a wee bit longer if anyone has some questions for me 2012-07-12T11:20:07 k cool 2012-07-12T11:20:09 ppisa: alseh: did you still have some issues to discuss? 2012-07-12T11:20:35 If you have some time to discuss, then yes 2012-07-12T11:21:27 well, the integration of NXLib is a topic ... 2012-07-12T11:21:28 i have to leave for the train, bye 2012-07-12T11:21:30 i'll try :) 2012-07-12T11:21:33 later sebhub 2012-07-12T11:21:34 The first, should FLTG go directly to RGT (rtems-graphic-toolkit) 2012-07-12T11:21:53 jenniferA, DrJoel: ok so i'm gonna go have some breakfast since i'm starving. ;) i'll be back in a bit and we can figure out the fork vs. branch issue 2012-07-12T11:22:00 what's the advantages/disadvantages? 2012-07-12T11:22:14 one sec rereading the email you sent 2012-07-12T11:23:40 ok... what is the speed of updates in fltk? 2012-07-12T11:24:15 if the project tracks slowly and releases are infrequent then we can just incorporate it directly... 2012-07-12T11:24:33 otherwise i think it is a good opportunity to try out git submodules and show how to use them. might be handy for other projects too 2012-07-12T11:24:36 as I see on their site, it's weekly snapshot 2012-07-12T11:24:36 The FLTK is quickly moving target, but we cannot keep with that. 2012-07-12T11:24:58 I used last stable version 1.3 and they are at 3.0 development 2012-07-12T11:25:25 then i think we must have either a script to fetch and update, or use submodule 2012-07-12T11:25:30 and just make clear documentation 2012-07-12T11:25:36 should I try to make a script to automate the porting process ? 2012-07-12T11:25:37 Yes, 1.3 should be used for now. 3.0 is in many ears redesing state. 2012-07-12T11:25:50 *** AndChat-637364 has quit IRC (Quit: Bye) 2012-07-12T11:26:06 It sould work. 2012-07-12T11:26:31 *** sebhub has quit IRC (Remote host closed the connection) 2012-07-12T11:26:38 it should be only including a few libs in configure.in and different makefile for tests 2012-07-12T11:26:56 ok that sounds like a decent idea to me 2012-07-12T11:27:09 also, the same problem with NXLib :) .... 2012-07-12T11:27:24 I think I can make a Makefile.rtems and leave the package intact 2012-07-12T11:27:27 just want some way that it should be minimal effort for someone to push some buttons and hopefully get a newer version as needed. 2012-07-12T11:27:36 i'm fine with that 2012-07-12T11:27:57 i don't like to mess around with build systems myself... 2012-07-12T11:28:16 and should I send it to NXLib maintainer, or leave it in toolkit ? 2012-07-12T11:28:19 so i take the lazy approach ;) 2012-07-12T11:28:36 do they haev similar things already in nxlib? 2012-07-12T11:28:59 no, they don't 2012-07-12T11:29:11 then we should probably be the ones maintaining it 2012-07-12T11:29:16 No, but NXlib is same source as Microwindows and holds RTEMS support. 2012-07-12T11:29:21 hmmm 2012-07-12T11:29:35 you can ask if tehy are interested then 2012-07-12T11:29:37 microwindows has a Makefile.rtems, there was where I got the idea from 2012-07-12T11:29:51 The No is to "do they haev similar things already in nxlib?" 2012-07-12T11:30:06 yes, they don't have other ports in nxlibs 2012-07-12T11:30:12 there is the no 2012-07-12T11:30:17 it would be better to get it tracked by them. 2012-07-12T11:31:13 also, I included 3 headers in X11 headers in NXLib, that could also be a problem 2012-07-12T11:32:11 probably you would have to work that out with them? 2012-07-12T11:32:31 I think that there should be RTEMS clone of Microwindows and NXlib on rtems.org and then politelly ask for merge. 2012-07-12T11:33:47 this would be good to ask about on rtems-devel. the good thing about having our own clone is we can track at our own speed... 2012-07-12T11:33:59 the bad thing is bitrotting and losing track of something 2012-07-12T11:35:16 Then what we should use for taking CFLAGS, LDFLAGS etc from 2012-07-12T11:35:16 include $(RTEMS_MAKEFILE_PATH)/Makefile.inc 2012-07-12T11:35:56 Ie this is correct 2012-07-12T11:35:56 include $(RTEMS_CUSTOM) 2012-07-12T11:35:56 include $(CONFIG.CC)or makefile leaf.inc 2012-07-12T11:36:14 is preffered? 2012-07-12T11:37:07 i don't know the answer to that 2012-07-12T11:38:06 OK, we can go to mailing list with these, I would suggest public rtems-devel 2012-07-12T11:38:43 Then if we want VESA graphic driver, we need to add functions to manipulate GDT entries. 2012-07-12T11:39:39 Again, I can write the summary but somebody from RTEMS main maintainers needs to coordinate approve that. 2012-07-12T11:40:32 ok. i don't have enough experience with this to help a lot 2012-07-12T11:41:46 OK, so we move on mailing list with these. 2012-07-12T11:42:05 *** kuzew has quit IRC (Ping timeout: 246 seconds) 2012-07-12T11:42:17 *** kuzew has joined #rtems 2012-07-12T11:42:42 great 2012-07-12T11:43:54 than, I think that's all for now, right ppisa ? 2012-07-12T11:44:10 So thanks for discussion. Alex, have you something more to Gedare now< 2012-07-12T11:44:57 seems good to me i'm encouraged by the progress of this project. i look forward to seeing commits coming :) 2012-07-12T11:45:03 nothing more from me, although, I wold like a pointer to what should be my priority now :) 2012-07-12T11:45:06 soon as these issues get resolved 2012-07-12T11:45:34 check bsd sources for driver code that can be more readily used 2012-07-12T11:46:00 and clean up the code you have so far, see if we can get any of it committed 2012-07-12T11:46:00 that's why we asked for VESA, because BSD had a vesa driver 2012-07-12T11:46:21 ok, clean up, than aditional libraries :) 2012-07-12T11:46:29 i see. ok make that clear in email 2012-07-12T11:47:17 another thing, sebhub updated the rtems graphic toolkit in the official tree, but the github did not update, shouldn't it ? 2012-07-12T11:49:36 verm__: how's the updater doing lately? 2012-07-12T11:50:07 alseh: the updates from rtems git repos to github have been flaky 2012-07-12T11:50:39 you should be able to set the rtems gtk as a remote and pull from it if the update doesn't go through 2012-07-12T11:51:24 I will , I just wondered why they aren't updated :) 2012-07-12T11:51:34 you can use 2012-07-12T11:51:34 git remote add rtems-org git://.... 2012-07-12T11:51:34 then 2012-07-12T11:51:34 git pull rtems-org 2012-07-12T11:52:18 thanks, I got that, the only thing it was that I was not sure why they did't update :) 2012-07-12T11:53:09 the script just doesn't work right 2012-07-12T11:53:47 ok than, thanks for all the advices :) 2012-07-12T11:57:00 ok. good luck :) 2012-07-12T12:00:57 *** ppisa has left #rtems ("Kopete 0.12.7 : http://kopete.kde.org") 2012-07-12T12:22:47 *** hiddenpearls1 has joined #rtems 2012-07-12T12:23:13 *** arvind_khadri has joined #rtems 2012-07-12T12:25:28 *** gedare has quit IRC (Quit: Leaving) 2012-07-12T12:32:33 *** WikL has joined #rtems 2012-07-12T12:46:13 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2012-07-12T13:04:08 *** hiddenpearls has joined #rtems 2012-07-12T13:06:37 *** hiddenpearls1 has quit IRC (Ping timeout: 240 seconds) 2012-07-12T13:13:25 *** arvind_khadri has joined #rtems 2012-07-12T13:18:10 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2012-07-12T13:21:49 *** cdcs has quit IRC (Quit: Page closed) 2012-07-12T13:24:39 *** soh_cah_toa has quit IRC (Quit: [(--)]ZZzzz...) 2012-07-12T13:27:14 *** soh_cah_toa has joined #rtems 2012-07-12T13:27:48 well, my neice woke up so i had to go watch her for a bit. i'm back now though 2012-07-12T13:27:58 booting up my vm now 2012-07-12T13:28:39 ok 2012-07-12T13:29:58 *** arvind_khadri has joined #rtems 2012-07-12T13:30:13 as far as using a fork vs. a branch, i'll try just set it up to pull from git.rtems.org/rtems-libbsd.git but push to soh-cah-toa/rtems-libbsd.git and see if that fixes anything 2012-07-12T13:30:54 the data flow will be a bit strange but it'll have to do in this case 2012-07-12T13:31:12 *** WikL has quit IRC () 2012-07-12T13:31:19 the tree is changing so fast. It may be easier to work directly from git.rtems.org, make small changes on a local branch, rebase frequently 2012-07-12T13:31:23 not sure though 2012-07-12T13:31:56 yeah, it would accomplish the same behavior w/o the obscure hack of mixing repos 2012-07-12T13:32:16 it would be a bit more centralized 2012-07-12T13:33:51 In my mind, you solve one issue, submit a patch or two, we merge, you update. repeat 2012-07-12T13:34:42 which, essentially, is what branches are exactly designed for 2012-07-12T13:34:57 yeah cant argue 2012-07-12T13:35:29 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-07-12T13:50:23 soh_cah_toa, how goes it? 2012-07-12T13:52:59 trying to figure out how i can make sure i can make a remote tracking branch w/o actually pushing anything 2012-07-12T13:53:10 i wanna make sure i have the right permissions 2012-07-12T13:53:26 normally, i'd just do git push -u origin 2012-07-12T13:53:34 How about you just pull the git.rtems.org master and see if you can build to the same state I get. 2012-07-12T13:53:59 yeah, i suppose that can come after. let me see 2012-07-12T13:54:45 I am more worried about your building than me being able to track your changes. If you submit frequently, then a simple branch for one fix will work fine. Submit patch, we commit, you update, verify it is OK, then make a temporary working branch 2012-07-12T13:54:59 if you figure out more with git great, but not critical. Just make small frequent patches 2012-07-12T13:55:05 right, ok 2012-07-12T13:56:42 yup, it was a sync issue. now i just get the cpufunc.h error you had before 2012-07-12T13:58:11 great. and the solution is to update the .py script to bring in the cpufunc.h for every supported architecture so we don't see this again. :) 2012-07-12T13:58:24 When we get outside that set, we can make them empty files. 2012-07-12T14:01:37 so should it be moved to the list for addHeaderFiles() or addEmptyHeaderFiles()? 2012-07-12T14:07:03 it is CPU specific.. let me find an example 2012-07-12T14:07:27 devNic.addCPUDependentHeaderFiles( 2012-07-12T14:07:36 that's what i was just looking at 2012-07-12T14:07:49 search for that. You can just add the other cpufunc.h files there. Try just mips first, then add the others. 2012-07-12T14:07:50 i'd do that for each arch w/ just cpufunc.h in the list? 2012-07-12T14:07:55 Yep 2012-07-12T14:08:19 Be warned.. i386 had a few methods we ifndef __rtems__ around because RTEMS or a BSP had a method with the same name 2012-07-12T14:08:26 but pretty straight forward if no name conflicts. 2012-07-12T14:08:36 ok so addHeaderFiles() isn't a "these files are for all architectures" method? 2012-07-12T14:08:45 yes 2012-07-12T14:09:11 but if it has a CPU directory in the name and is a .h, it goes in the CPUDependentHeaderFiles list 2012-07-12T14:09:35 addCPUDependentSourceFiles is for .c files which are CPU dependent. That also impacts the Makefile 2012-07-12T14:09:50 right 2012-07-12T14:10:15 if you get past cpufunc.h on mips, add them all and send a patch. I will merge it 2012-07-12T14:10:23 so freebsd/machine is for all architectures then? 2012-07-12T14:10:45 yep. 2012-07-12T14:10:47 k 2012-07-12T14:12:15 but if the file is freebsd/machine/cpufunc.h, do i change it to freebds/mips/cpufunc.h? 2012-07-12T14:12:33 s/freebds/freebsd/ 2012-07-12T14:12:51 in the freebsd tree, they are here: 2012-07-12T14:12:56 ./amd64/include/cpufunc.h 2012-07-12T14:12:56 ./powerpc/include/cpufunc.h 2012-07-12T14:12:56 ./sun4v/include/cpufunc.h 2012-07-12T14:12:56 ./sparc64/include/cpufunc.h 2012-07-12T14:12:56 ./ia64/include/cpufunc.h 2012-07-12T14:12:57 ./pc98/include/cpufunc.h 2012-07-12T14:12:59 ./arm/include/cpufunc.h 2012-07-12T14:13:01 ./i386/include/cpufunc.h 2012-07-12T14:13:03 ./mips/include/cpufunc.h 2012-07-12T14:13:18 The only one we have is for i386 and it is ./freebsd/i386/include/freebsd/machine/cpufunc.h which is where the script put it. 2012-07-12T14:13:37 The include patch (-I dir) includes the freebsd/$(CPU)/include directory 2012-07-12T14:15:23 ok that makes sense, but the error is complaing about freebsd/machine/cpufunc.h not freebsd/mips/cpufunc.h 2012-07-12T14:16:20 When -I ./freebsd/i386/include, the path to the file remaining is freebsd/machine/cpufunc.h 2012-07-12T14:16:24 replace i386 with mips 2012-07-12T14:18:15 *** deb has joined #rtems 2012-07-12T14:19:36 well, i'm still kinda confused when/where it gets translated but i'll just go ahead try it out for mips 2012-07-12T14:20:29 From the Makefile 2012-07-12T14:20:35 CFLAGS += -I . 2012-07-12T14:20:56 adds the directory which lets generic freebsd files get found on include path 2012-07-12T14:21:01 CFLAGS += -I freebsd/$(RTEMS_CPU)/include 2012-07-12T14:21:17 does the same for CPU specific directories like the i386 where cpufunc.h is located 2012-07-12T14:21:22 CFLAGS += -I rtemsbsd/$(RTEMS_CPU)/include 2012-07-12T14:21:23 CFLAGS += -I freebsd/$(RTEMS_CPU)/include 2012-07-12T14:21:33 ignore those two .. bad copy 2012-07-12T14:21:35 CFLAGS += -I rtemsbsd 2012-07-12T14:21:35 CFLAGS += -I rtemsbsd/$(RTEMS_CPU)/include 2012-07-12T14:21:40 is the same for things in rtemsbsd 2012-07-12T14:23:26 yeah, i can see that when it builds. but then when systm.h asks to #include freebsd/machine/cpufunc.h wouldn't that path actually be freebsd/mips/include/freebsd/machine/cpufunc.h? 2012-07-12T14:23:52 obivously, that path is just stupid but whatever the file asks for get search for in freebsd//include 2012-07-12T14:24:39 yes to first .. 2012-07-12T14:25:11 right.. the set of files needed to build a complete set is a mix of generic BSD, cpu specific BSD, generic rtems BSD support, CPU specific BSD support 2012-07-12T14:31:36 oh wait. i forgot out the -I . 2012-07-12T14:32:52 Those are all in the Makefile so you don't have to touch that 2012-07-12T14:33:11 Should just be a matter of adding the mips cpufunc.h to *.py, running the script and building 2012-07-12T14:34:06 yeah, i know that. i'm just trying to understand all this magic 2012-07-12T14:35:33 but still, there really is no freebsd/machine/cpufunc.h which systm.h is asking for. i'm confused where gcc or make says "when a file wants to include something in freebsd/machine, instead look for it freebsd/mips" 2012-07-12T14:35:40 yeah, there's CFLAGS += -I freebsd/$(RTEMS_CPU)/include 2012-07-12T14:36:05 but systm.h never asked for freebsd/mips/cpufunc.h 2012-07-12T14:36:10 systm.c 2012-07-12T14:36:39 Think of it as a set union of the contents of directories... look for this file in these places. The paths are relative so you effectively get the union 2012-07-12T14:38:10 oh wait, there really is a freebsd/mips/include/freebsd/machine directory 2012-07-12T14:39:04 *** QingPei has left #rtems 2012-07-12T14:39:23 and then if freebsd/mips/include/freebsd/machine/foo.h doesn't exist, gcc then searches ./freebsd/machine/foo.h 2012-07-12T14:40:02 depends on the order of the directories but effectively yes. 2012-07-12T14:40:07 right 2012-07-12T14:42:57 *** DrJoel has quit IRC (Remote host closed the connection) 2012-07-12T14:43:29 ok, now *that* makes sense. i was assuming freebsd/$(RTEMS_CPU)/include/freebsd/machine didn't exist and that somehow/somewhere the dir freebsd/machine got translated to freebsd/mips 2012-07-12T14:43:40 got it, *now* i can edit the script :) 2012-07-12T14:44:17 *** DrJoel has joined #rtems 2012-07-12T14:44:17 *** DrJoel has joined #rtems 2012-07-12T14:44:17 *** ChanServ sets mode: +o DrJoel 2012-07-12T14:44:28 back 2012-07-12T14:44:33 k 2012-07-12T14:45:17 desktop froze and I have to kill session :( 2012-07-12T14:45:35 np 2012-07-12T14:46:43 *** DrJoel has quit IRC (Read error: No route to host) 2012-07-12T14:47:01 *** DrJoel has joined #rtems 2012-07-12T14:47:01 *** DrJoel has joined #rtems 2012-07-12T14:47:01 *** ChanServ sets mode: +o DrJoel 2012-07-12T14:54:12 soh_cah_toa, building for you now? 2012-07-12T14:54:20 http://pastebin.com/FqQ6iJPr 2012-07-12T14:54:24 strange. i added cpufunc.h just like that ^ right after the i386 equivalent and ran the script as ./freebsd-to-rtems.py -f /home/rtems/code/src/rtems-freebsd-8.2/ -r /home/rtems/code/src/rtems-libbsd/ and it said 8 files were changed yet the error doesn't disappear 2012-07-12T14:55:37 hm, i did a forward operation so i know that's right so i must've not added the right thing to the script 2012-07-12T14:55:52 *** jenniferA has quit IRC (Quit: Leaving) 2012-07-12T14:56:42 is the install point removed? 2012-07-12T14:57:30 you mean the value of INSTALL_BASE? 2012-07-12T14:58:04 yes but also did you end up with cpufunc.h in your source tree? 2012-07-12T14:58:25 I suspect with 8 files changed you should have done a -R before adding mips ... 2012-07-12T14:58:26 no, that's the problem 2012-07-12T14:59:00 i never modified anything though, there's no need to move anything back to the freebsd tree 2012-07-12T15:00:55 But jennifer and I may have changes to files in rtems-libbsd that need to be moved back to your master freebsd-8.2 source. With 8 files changed, that's a hint that you picked up more than you wanted. 2012-07-12T15:01:10 Did you only add the mips cpufunc.h? Then only Makefile and the one new file 2012-07-12T15:01:28 oh wright 2012-07-12T15:01:36 dug 2012-07-12T15:01:41 let me see 2012-07-12T15:03:03 0 files were changed 2012-07-12T15:03:05 that makes no sense 2012-07-12T15:06:31 hm... 2012-07-12T15:10:23 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-12T15:19:55 *** sbahra has joined #rtems 2012-07-12T15:20:03 any progress? 2012-07-12T15:20:13 oh wait, the error is for cpuregs.h 2012-07-12T15:20:19 i'll just do the same thing for it 2012-07-12T15:20:33 Hi all. Out of curiosity, is the person working on "Atomic Operations for RTEMS" around here? 2012-07-12T15:20:59 I'd like to know what it would take to get http://concurrencykit.org to a stage where it would be an acceptable option for RTEMS. It has the most complete set of atomic operations of the libraries that are out there. 2012-07-12T15:21:53 don't see him.. best to send email to rtems-devel or him personally 2012-07-12T15:22:53 Cool. Where can I learn more about the project? 2012-07-12T15:22:59 (As well as find his contact information) 2012-07-12T15:23:17 that would be WeiY 2012-07-12T15:23:24 i'll try to find his email 2012-07-12T15:23:38 (name is Deng Hengyi) 2012-07-12T15:24:20 He posted status to the rtems-gsoc list this week for mid-term. rtems.org/mailman/rtems-gsoc. He also discussed it on irc a few hours ago so there should be some info in today's IRC log 2012-07-12T15:25:12 sbahra: it's wei.a.yang@gmail.com 2012-07-12T15:25:38 Sounds good. Thank you. 2012-07-12T16:01:32 soh_cah_toa, how is it going? 2012-07-12T16:01:47 building right now 2012-07-12T16:01:52 seems to be going good so far 2012-07-12T16:02:05 adding cpuregs.h was fine 2012-07-12T16:02:22 I expect you will do some linkcmds tinkering on mips before being able to link testsuite/link01 :) 2012-07-12T16:02:39 But when things build, submit any patches required for misp 2012-07-12T16:02:41 mips 2012-07-12T16:02:54 ok 2012-07-12T16:06:46 hm, local/port_before.h can't find freebsd/bsd.h. i feel like i've seen this before as well 2012-07-12T16:11:06 well, freebsd/bsd.h is in rtemsbsd which should be in the include path 2012-07-12T16:11:49 though, oddly enough, even though it's added in CFLAGS i'm not seeing it here when it's actually executed 2012-07-12T16:12:59 make install at the top. I haven't done a great job of integrating the user space stuff. :( 2012-07-12T16:13:01 it is new 2012-07-12T16:15:25 you mean -I $(INSTALL_BASE)? 2012-07-12T16:15:58 that doesn't sound right 2012-07-12T16:16:48 everything in freebsd and rtemsbsd is in "kernel space". freebsd-userspace is built on top of that to run in "user space". For FreeBSD the difference is real.. for us, it is a namespace issue 2012-07-12T16:17:34 right 2012-07-12T16:20:57 adding INSTALL_BASE does nothing. though that makes sense 2012-07-12T16:21:54 It was already added. If there is nothing there, then it won't help. 2012-07-12T16:22:00 yeah, i just noticed that 2012-07-12T16:22:03 try make install first. 2012-07-12T16:22:16 oh that's what you meant 2012-07-12T16:22:17 FWIW $(INSTALL_BASE)/include is in the top level Makefile and I am considering that may be a mistake 2012-07-12T16:22:30 I am heading home. 2012-07-12T16:22:40 ok 2012-07-12T16:22:45 today is my wedding anniversary and I should have already taken off earlier 2012-07-12T16:22:49 :) 2012-07-12T16:22:56 I will be back on in the morning 2012-07-12T16:22:59 *** DrJoel has quit IRC (Quit: Leaving) 2012-07-12T16:23:00 uh yeah, you should probably head home :) 2012-07-12T16:27:07 *** kuzew has quit IRC (Ping timeout: 240 seconds) 2012-07-12T16:33:59 *** kuzew has joined #rtems 2012-07-12T17:46:13 *** sbahra has quit IRC (Quit: sbahra) 2012-07-12T17:59:08 *** kuzew has quit IRC (Ping timeout: 248 seconds) 2012-07-12T18:00:34 *** kuzew has joined #rtems 2012-07-12T18:30:31 *** peerst has quit IRC (Read error: Operation timed out) 2012-07-12T18:46:29 *** peerst has joined #rtems 2012-07-12T19:23:48 *** alseh has quit IRC (Remote host closed the connection) 2012-07-12T19:42:38 *** zw_yao has quit IRC (Quit: Leaving) 2012-07-12T20:00:50 *** xiangfu has joined #rtems 2012-07-12T20:37:24 *** soh_cah_toa has quit IRC (Quit: hasta la pasta!) 2012-07-12T20:56:27 *** zw_yao has joined #rtems 2012-07-12T21:02:04 *** QingPei has joined #rtems 2012-07-12T21:51:20 *** kiwichris has quit IRC (Ping timeout: 265 seconds) 2012-07-12T21:54:11 *** kiwichris has joined #rtems 2012-07-12T21:59:33 *** kiwichris has quit IRC (Ping timeout: 265 seconds) 2012-07-12T22:00:13 *** kiwichris has joined #rtems 2012-07-12T22:05:22 *** deb has left #rtems ("Leaving") 2012-07-12T22:13:55 *** arvind_khadri has joined #rtems 2012-07-12T22:19:25 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-07-12T22:23:38 *** kiwichris has quit IRC (Read error: Connection reset by peer) 2012-07-12T22:29:00 *** kuzew has quit IRC (Ping timeout: 248 seconds) 2012-07-12T22:29:31 *** kuzew has joined #rtems 2012-07-12T22:35:30 *** kiwichris has joined #rtems 2012-07-12T22:41:56 *** arvind_khadri has joined #rtems 2012-07-12T22:42:21 *** kiwichris_ has joined #rtems 2012-07-12T22:43:00 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2012-07-12T22:45:17 *** kiwichris has quit IRC (Ping timeout: 246 seconds) 2012-07-12T22:52:13 *** arvind_khadri has joined #rtems 2012-07-12T22:53:45 *** arvind_khadri has quit IRC (Remote host closed the connection) 2012-07-12T22:53:45 *** kiwichris_ has quit IRC (Read error: Connection reset by peer) 2012-07-12T22:54:55 *** arvind_khadri has joined #rtems 2012-07-12T22:57:21 *** kiwichris_ has joined #rtems 2012-07-12T23:02:21 *** arvind_khadri has quit IRC (Ping timeout: 255 seconds) 2012-07-12T23:48:21 *** QingPei has left #rtems 2012-07-13T00:50:15 *** zw_yao has quit IRC (Ping timeout: 244 seconds) 2012-07-13T01:35:16 *** sebhub has joined #rtems 2012-07-13T01:37:37 *** arvind_khadri has joined #rtems 2012-07-13T01:50:51 *** WikL has joined #rtems 2012-07-13T02:19:39 *** WikL has quit IRC (Ping timeout: 244 seconds) 2012-07-13T03:04:31 *** WikL has joined #rtems 2012-07-13T03:31:10 good morning 2012-07-13T04:36:03 *** WikL has quit IRC (Ping timeout: 244 seconds) 2012-07-13T05:59:01 *** xiangfu has quit IRC (Ping timeout: 265 seconds) 2012-07-13T06:29:58 *** kiwichris__ has joined #rtems 2012-07-13T06:30:50 *** kiwichris_ has quit IRC (Ping timeout: 246 seconds) 2012-07-13T07:39:20 *** weiY has joined #rtems 2012-07-13T07:41:31 *** kiwichris__ has quit IRC (Ping timeout: 244 seconds) 2012-07-13T08:47:14 *** gedare has joined #rtems 2012-07-13T09:45:33 *** mwalle has quit IRC (Quit: WeeChat 0.3.2) 2012-07-13T10:04:48 *** sbahra has joined #rtems 2012-07-13T10:24:48 *** arvind_khadri has quit IRC (Ping timeout: 255 seconds) 2012-07-13T10:27:07 hi 2012-07-13T11:02:47 *** tuxmaniac has quit IRC (Quit: leaving) 2012-07-13T11:06:44 *** sebhub has quit IRC (Remote host closed the connection) 2012-07-13T11:15:02 *** Hesham has joined #rtems 2012-07-13T11:26:47 hi Hesham done traveling? 2012-07-13T11:30:30 yeah, i have reached yesterday 2012-07-13T11:30:53 how are you gedare ? 2012-07-13T11:31:07 good. you settled in and getting productive? 2012-07-13T11:31:55 i am traveling each week from & to my home town 2012-07-13T11:32:05 As i am working at Cairo 2012-07-13T11:32:50 i see what are you doing in Cairo? 2012-07-13T11:33:31 I am working at a Company called SilMinds as a Software Engineer 2012-07-13T11:34:20 SilMinds offers innovative solutions for Decimal Floating Point Arithmetic (DFPA) hardware acceleration 2012-07-13T11:34:46 I see. sounds nice. you are not having too much trouble splitting your time between that and your GSoC obligations? 2012-07-13T11:35:19 *** antgreen has joined #rtems 2012-07-13T11:35:26 GSoC is my first priority now 2012-07-13T11:35:44 Also Ramadan is coming and working hours would be reduced 2012-07-13T11:36:12 that will allow me to work 8+ hours at GSoC 2012-07-13T11:36:17 per day 2012-07-13T11:36:40 and i have two days vacation, Friday and Saturday 2012-07-13T11:36:58 Ok. well, i will say you passed your midterm, but much more progress is necessary in the second half now 2012-07-13T11:37:27 i know there were issues mostly because of the powerpc problems, hopefully those have been mostly resolved and you can focus on getting code into submittable shape 2012-07-13T11:37:45 yeah i will try to finish all high-level/mid-level work this month 2012-07-13T11:38:02 what is your outline of work to do? 2012-07-13T11:38:23 two weeks for libmm and Arena 2012-07-13T11:38:35 and the rest for ARM work 2012-07-13T11:38:50 about libmm 2012-07-13T11:38:53 ok. i think libmm should take hardly much time at all, except for dealing with the BSP attributes. 2012-07-13T11:39:14 actually i have implemented a way at modified code 2012-07-13T11:39:23 but when i tried to test it, it failed 2012-07-13T11:39:28 oh sad 2012-07-13T11:39:38 you might like to consider replying to the email by Ric that mentioned having the BSP define the attributes and see if you can get some input on ways to do it 2012-07-13T11:39:47 and maybe present the way you tried and failed 2012-07-13T11:40:40 i think it failed because MMU did not raise an exception 2012-07-13T11:41:11 one challenge i have faced when trying to implement this way is ... 2012-07-13T11:41:31 where attributes should be defined at BSP code ? 2012-07-13T11:43:04 and if it was just declared at high-level code, should it be a pointer or dummy struct containing boolean fields for possible features ? 2012-07-13T11:43:24 these are good questions ;) 2012-07-13T11:44:16 i have defined it as 32 bit variable at high-level code 2012-07-13T11:44:22 that's a good start 2012-07-13T11:44:37 these bits are completely handled by BSP code only 2012-07-13T11:44:41 but what if the target uses 64-bit registers? 2012-07-13T11:45:01 we may create a struct containing two 32bit variables 2012-07-13T11:45:12 "we" who? 2012-07-13T11:45:18 me :) 2012-07-13T11:46:14 But actually in powerpc or arm , MMU attributes bits does not exceed 8 bits 2012-07-13T11:47:03 see the problem is ... we need a 'type' defined that can be handled at the high-level (in libmm) without relying too much on low-level 2012-07-13T11:47:20 for example , powerpc has 2 bits for protection 2012-07-13T11:48:14 mmm 2012-07-13T11:48:49 but low-level will handle setting / un-setting attributes withing this type 2012-07-13T11:49:27 correct 2012-07-13T11:49:42 the type has to be opaque at the high-level, but 'known' type... 2012-07-13T11:49:59 So, when a user calls set_read for example, BSP code will know what to set 2012-07-13T11:50:26 yep 2012-07-13T11:50:37 Yeah i have implemented this way 2012-07-13T11:50:50 But one question .... 2012-07-13T11:51:36 If the user has to set write attributes and cache related, does he has to call set_write , set_cache_though_back 2012-07-13T11:51:46 every time he want this ? 2012-07-13T11:52:03 and what if user want to un-set this attributes 2012-07-13T11:52:18 that means there must be a functions doing that 2012-07-13T11:52:27 the attributes should persist until something changes them 2012-07-13T11:53:03 You mean a function to change them ? 2012-07-13T11:53:23 yes.. i'm not sure yet what makes sense, but probably you can just do set/clear for each attribute, or just define a function that can be called with the attribute field, type of attribute (which might be BSP defined?), and true/false for set/clear 2012-07-13T11:53:26 yes 2012-07-13T11:53:46 a user should be able to set the attributes on a range of memory.. and have those attributes persist on that memory until something else changes those attributes 2012-07-13T11:54:29 *** mumptai_ has joined #rtems 2012-07-13T11:54:36 hi 2012-07-13T11:54:43 either because we use some notion of 'domain' (e.g. arena), or because the user explicitly sets the attributes for the same range of memory to something different 2012-07-13T11:54:47 hello mumptai_ 2012-07-13T11:55:12 i see 2012-07-13T11:55:16 hello mumptai_ 2012-07-13T11:55:37 *** antgreen` has joined #rtems 2012-07-13T11:56:11 ok, function eliminates errors maybe produced by wrongly setting attributes bits by the user 2012-07-13T11:56:44 but it costs a call ( for a function ) for every single attribute to be changed 2012-07-13T11:57:07 *** antgreen has quit IRC (Ping timeout: 240 seconds) 2012-07-13T11:57:15 right ? 2012-07-13T11:59:04 Hesham: i think maybe there is a way to define a set of attributes to set for one function call, but it is kind of irrelevant, the important performance consideration is to have the attributes stored in such a way that the hw is quickly loaded with the attributes 2012-07-13T11:59:24 you think... the user sets attributes for each memory region one time 2012-07-13T12:00:18 then if the region is used by more than one arena, every time the arena swap happens the attributes that were previously set need to be loaded to the hw 2012-07-13T12:00:32 this is the time-critical piece 2012-07-13T12:01:10 does that make sense? 2012-07-13T12:01:49 not totally clear 2012-07-13T12:02:47 *** Hesham1 has joined #rtems 2012-07-13T12:02:47 *** Hesham has quit IRC (Read error: Connection reset by peer) 2012-07-13T12:03:06 ok i can try again .. 2012-07-13T12:03:42 lets say i want to give two sets of attributes to the same memory region 2012-07-13T12:03:46 You mean two Arenas can have the same region (MPE) at one time ? 2012-07-13T12:03:52 one set for use at a certain time, another set for use at a different time 2012-07-13T12:04:05 yeah but with different attributes 2012-07-13T12:04:05 i see yeah 2012-07-13T12:04:10 same memory range 2012-07-13T12:04:23 But both can not be loaded at the same time of course 2012-07-13T12:04:30 for example... task1 has a stack it can read/write, maybe task 2 can read task1 stack but not write 2012-07-13T12:04:42 I got it 2012-07-13T12:04:44 so while task1 executes its stack is r/w, but when task2 takes over the task1.stack is ro 2012-07-13T12:04:46 ok 2012-07-13T12:05:00 we have to make that switch efficient 2012-07-13T12:05:15 That's maps to installing/uninstalling MPEs 2012-07-13T12:05:46 yes 2012-07-13T12:06:14 Installing searches for valid pte and change its attributes 2012-07-13T12:06:24 that's what currently happens 2012-07-13T12:06:26 at least on powerpc 2012-07-13T12:06:30 yeah 2012-07-13T12:06:36 might be different in other architecutres 2012-07-13T12:06:39 but uninstalling is not implemented 2012-07-13T12:07:13 might have to 'dump' the entire hw for some architecture, and then re-install each individual one you want enforced... it depends on the hw a lot 2012-07-13T12:07:28 yeah i know 2012-07-13T12:07:37 so.. it's important that the attributes in the MPE can be quickly put into the hw registers 2012-07-13T12:07:50 I think that's why you create install/uinstall MPE 2012-07-13T12:08:17 is not it quickly process in powerpc with installing function ? 2012-07-13T12:09:08 this means the BSP / low-level should define the structure of attributes... and also why the low-level has to implement the install/uninstall... 2012-07-13T12:09:11 what do you mean? 2012-07-13T12:10:00 I mean for powepc implementation now, attributes can be simply loaded by install function 2012-07-13T12:10:27 doesn't it go through a 2012-07-13T12:10:31 translate function? 2012-07-13T12:10:51 to convert the attributes? 2012-07-13T12:10:52 yeah but i have edited it not to go through translation 2012-07-13T12:11:20 i guess i'd have to look at what you have done. but then the question would be whether your approach would work for other architectures. 2012-07-13T12:12:10 I can illustrate what i tried to implement and how and sent it to rtems-devel ml to get feedbacks 2012-07-13T12:12:20 send * 2012-07-13T12:12:20 that would be useful 2012-07-13T12:12:38 note that it was previously rejected to be using a 32-bit integer for storing the attributes 2012-07-13T12:13:02 re-read what Ric sent you about your previous design, ok? 2012-07-13T12:13:45 ok, i will re-read it and sent what's my suggestion if we used 32-bit integer and list some other options 2012-07-13T12:13:56 ok. 2012-07-13T12:14:40 good, anything else? 2012-07-13T12:15:06 no.. just focus on this libmm work. 2012-07-13T12:15:32 i go back and forth about the later arena work. whether it should be a separate thing, or an optional extension of libmm 2012-07-13T12:16:22 but we will cross that bridge when we get there 2012-07-13T12:16:26 first thing to get right is libmm 2012-07-13T12:16:31 me too, i was thinking that it's just a light layer above libmm 2012-07-13T12:16:48 is your branch on github working code? 2012-07-13T12:17:07 yeah but does not raise MMU exceptions 2012-07-13T12:17:14 (modified work) 2012-07-13T12:17:27 ah, 2012-07-13T12:17:35 ok you need to start over with the conversion 2012-07-13T12:17:48 since i think you prefer not to try to debug the whole change 2012-07-13T12:18:03 if you implement each modification at a time then you should be able to test / verify it as you go 2012-07-13T12:18:24 although it's entirely possible that the problem exists in the mmutest itself... hard to say for sure though 2012-07-13T12:18:24 it's kinda both ( conversion and debugging ) 2012-07-13T12:18:28 yea 2012-07-13T12:18:41 i suggest creating a fresh branch to post your work to 2012-07-13T12:18:42 i will adapt your changes step by step 2012-07-13T12:18:53 and add a commit each time you make a 'step' 2012-07-13T12:18:58 and test it each step 2012-07-13T12:19:00 and try to provide useful commit messages 2012-07-13T12:19:16 ok 2012-07-13T12:19:19 like... "libmmu: change name of xxx to yyy" or something 2012-07-13T12:19:45 ok i will 2012-07-13T12:19:51 identify the component that was modified, and a brief description of the change, you can also use the longer commit area for more deteail about the change 2012-07-13T12:20:02 but hopefully each change is small and concise, and you can explain it simply 2012-07-13T12:20:41 i will commit "each" step that is successfully working with test cases 2012-07-13T12:23:06 gedare: i have to go now, anything important else ? 2012-07-13T12:23:48 nah 2012-07-13T12:23:57 get some work done :) 2012-07-13T12:24:21 i will :) 2012-07-13T12:24:48 thanks, bye 2012-07-13T12:24:52 *** Hesham1 has quit IRC (Quit: Leaving.) 2012-07-13T13:05:59 *** mumptai_ has quit IRC (Ping timeout: 246 seconds) 2012-07-13T13:53:30 *** deb has joined #rtems 2012-07-13T14:48:05 *** weiY has quit IRC (Ping timeout: 248 seconds) 2012-07-13T15:01:47 *** gedare has quit IRC (Quit: Ex-Chat) 2012-07-13T15:05:51 *** antgreen` has quit IRC (Remote host closed the connection) 2012-07-13T15:06:17 *** antgreen` has joined #rtems 2012-07-13T16:12:23 *** kiwichris has joined #rtems 2012-07-13T16:39:47 *** antgreen` has quit IRC (Remote host closed the connection) 2012-07-13T17:48:53 *** kuzew has quit IRC (Ping timeout: 248 seconds) 2012-07-13T17:49:44 *** kuzew has joined #rtems 2012-07-13T18:16:22 *** deb has quit IRC (Quit: This computer has gone to sleep) 2012-07-13T18:29:35 *** weiY has joined #rtems 2012-07-13T18:39:25 *** weiY has quit IRC () 2012-07-13T18:45:56 *** kuzew has quit IRC (Ping timeout: 248 seconds) 2012-07-13T18:46:32 *** kuzew has joined #rtems 2012-07-13T18:50:27 *** sbahra has quit IRC (Quit: sbahra) 2012-07-13T19:40:36 *** deb has joined #rtems 2012-07-13T22:14:59 *** deb has left #rtems ("Leaving") 2012-07-13T23:58:34 *** kiwichris has quit IRC (Read error: Connection reset by peer) 2012-07-13T23:59:30 *** kiwichris has joined #rtems 2012-07-14T00:05:10 *** kiwichris_ has joined #rtems 2012-07-14T00:07:56 *** kiwichris has quit IRC (Ping timeout: 246 seconds) 2012-07-14T04:36:29 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-14T05:20:29 *** WikL has joined #rtems 2012-07-14T05:23:29 *** kristianpaul has joined #rtems 2012-07-14T06:31:00 *** kiwichris__ has joined #rtems 2012-07-14T06:31:40 *** kiwichris_ has quit IRC (Ping timeout: 248 seconds) 2012-07-14T15:06:22 *** arvind_khadri has joined #rtems 2012-07-14T16:51:17 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-07-14T16:58:29 *** WikL has quit IRC () 2012-07-14T20:31:34 *** kiwichris__ has quit IRC (Quit: Leaving) 2012-07-15T00:10:40 *** arvind_khadri has joined #rtems 2012-07-15T01:39:56 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-07-15T02:52:30 *** arvind_khadri has joined #rtems 2012-07-15T05:20:01 *** WikL has joined #rtems 2012-07-15T08:20:26 *** hiddenpearls has joined #rtems 2012-07-15T09:14:15 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-15T09:43:04 *** hiddenpearls has joined #rtems 2012-07-15T10:26:50 *** hiddenpearls1 has joined #rtems 2012-07-15T12:15:19 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2012-07-15T12:31:43 *** hiddenpearls has joined #rtems 2012-07-15T12:52:17 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-15T12:57:39 *** hiddenpearls has joined #rtems 2012-07-15T13:30:49 *** hiddenpearls has quit IRC (Quit: Leaving.) 2012-07-15T13:37:46 *** jkridner has joined #rtems 2012-07-15T13:37:57 I see we've reached mid-terms on GSoC... 2012-07-15T13:38:06 anyone know the status of RTEMS on BeagleBone? 2012-07-15T13:47:15 jkridner: you can contact the student 2012-07-15T13:47:24 I think he can keep you updated with the status 2012-07-15T14:09:39 *** hiddenpearls1 has joined #rtems 2012-07-15T14:11:20 do you know how best to contact him? 2012-07-15T14:11:40 I have his name as Claas Ziemke 2012-07-15T14:11:48 but, that is about it. 2012-07-15T14:12:47 http://wiki.rtems.org/wiki/index.php/User_talk:Claas_Ziemke doesn't give any useful contact info 2012-07-15T14:13:08 nor http://www.rtems.com/wiki/index.php/User:Claas_Ziemke 2012-07-15T14:14:56 * jkridner finds http://www.rtems.com/ml/rtems-users/2012/june/msg00060.html 2012-07-15T14:15:36 *** hiddenpearls1 has quit IRC (Quit: Leaving.) 2012-07-15T14:50:58 *** jkridner has quit IRC (Quit: jkridner) 2012-07-15T15:17:25 *** jkridner has joined #rtems 2012-07-15T15:17:25 *** jkridner has joined #rtems 2012-07-15T16:48:44 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-07-15T16:57:45 *** WikL has quit IRC () 2012-07-15T17:01:12 *** jkridner has quit IRC (Read error: Connection reset by peer) 2012-07-15T17:16:27 *** jkridner has joined #rtems 2012-07-15T17:16:27 *** jkridner has joined #rtems 2012-07-15T17:29:50 *** kiwichris has joined #rtems 2012-07-15T19:28:08 *** xiangfu has joined #rtems