2012-06-25T02:14:09 *** sebhub has joined #rtems 2012-06-25T02:30:21 good morning 2012-06-25T02:59:11 *** arvind_khadri has joined #rtems 2012-06-25T04:53:23 *** arvind_k has joined #rtems 2012-06-25T04:53:25 *** arvind_khadri has quit IRC (Ping timeout: 260 seconds) 2012-06-25T05:06:08 *** arvind_k has quit IRC (Ping timeout: 240 seconds) 2012-06-25T07:27:08 *** sevikkk has joined #rtems 2012-06-25T09:21:00 *** gedare has joined #rtems 2012-06-25T10:49:38 *** sebhub has quit IRC (Remote host closed the connection) 2012-06-25T11:49:00 *** gedare has quit IRC (Quit: Leaving) 2012-06-25T12:21:51 *** antgreen has joined #rtems 2012-06-25T12:44:05 *** vidhoon has joined #rtems 2012-06-25T12:44:23 *** vidhoon has left #rtems 2012-06-25T13:06:29 *** antgreen has quit IRC (Remote host closed the connection) 2012-06-25T14:51:56 *** sevikkk has quit IRC (Remote host closed the connection) 2012-06-25T15:03:49 *** sevikkk has joined #rtems 2012-06-25T18:50:38 *** sevikkk has quit IRC (Ping timeout: 240 seconds) 2012-06-25T18:53:13 *** sevikkk has joined #rtems 2012-06-25T21:42:02 *** QingPei has joined #rtems 2012-06-26T00:18:26 *** arvind_k has joined #rtems 2012-06-26T00:21:14 *** arvind_k is now known as arvind_khadri 2012-06-26T00:21:29 *** arvind_khadri has joined #rtems 2012-06-26T01:56:24 *** sebhub has joined #rtems 2012-06-26T02:19:32 *** sebhub has quit IRC (Write error: Broken pipe) 2012-06-26T02:19:58 *** sebhub has joined #rtems 2012-06-26T02:21:22 sebhub, hi 2012-06-26T02:30:22 hi 2012-06-26T02:36:41 *** sevikkk has quit IRC (*.net *.split) 2012-06-26T02:36:43 *** kiwichris_ has quit IRC (*.net *.split) 2012-06-26T03:04:56 *** kiwichris_ has joined #rtems 2012-06-26T03:05:58 *** sevikkk has joined #rtems 2012-06-26T03:28:40 *** QingPei has left #rtems 2012-06-26T03:49:59 sebhub, just been looking over the off_t thread on the new lib list. 2012-06-26T03:50:22 I made the change to a 64bit off_t and RTEMS supports this. 2012-06-26T03:51:14 I also think all RTEMS targets should support this. 2012-06-26T03:52:08 yes, the 64-bit offsets are good 2012-06-26T03:52:22 4GiB is not much these days 2012-06-26T03:52:59 The compiler seems to handle 64bit ok and a good implementation should limit the places maths are done. 2012-06-26T03:53:06 i have to fix the ftello some time in the future, but now other things have priority 2012-06-26T03:53:35 The *64 functions in newlib seem a poor idea. 2012-06-26T03:53:41 addition is quite ok for architectures with carry flags 2012-06-26T03:54:00 I assume that API comes from Linux. 2012-06-26T03:54:04 the division on the other hand is VERY expensive on some architectures like ARMv4T 2012-06-26T03:54:30 yes, ABI backward compatibility 2012-06-26T03:55:46 for rtems these *64() functions make no sense 2012-06-26T04:20:50 *** dr__house has quit IRC (Ping timeout: 272 seconds) 2012-06-26T04:23:55 *** dr__house has joined #rtems 2012-06-26T04:23:55 *** dr__house has joined #rtems 2012-06-26T06:26:25 *** zw_yao has joined #rtems 2012-06-26T06:30:49 *** zw_yao has quit IRC (Ping timeout: 250 seconds) 2012-06-26T07:05:12 *** deb has joined #rtems 2012-06-26T07:17:32 *** QingPei has joined #rtems 2012-06-26T08:03:02 *** weiY has joined #rtems 2012-06-26T08:18:26 Hi chris 2012-06-26T08:18:33 are you aroud? 2012-06-26T08:20:10 sebhub, are you aroud? 2012-06-26T08:21:09 yes 2012-06-26T08:24:37 sebhub, hi good night. do you now how to run the rtems smp test for i386 2012-06-26T08:24:59 i run them on the vmware. there are some errors on all the smp test 2012-06-26T08:25:47 never worked with the smp tests so far 2012-06-26T08:26:19 ok, do you konw who can help me about this question? 2012-06-26T08:26:43 the mailing list hopefully 2012-06-26T08:28:28 ok, i will 2012-06-26T08:44:05 *** weiY has quit IRC () 2012-06-26T08:52:41 *** weiY has joined #rtems 2012-06-26T09:04:39 *** alseh has joined #rtems 2012-06-26T09:23:14 *** gedare has joined #rtems 2012-06-26T09:34:57 *** weiY has quit IRC () 2012-06-26T10:48:06 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-06-26T11:01:52 *** Hesham has joined #rtems 2012-06-26T11:02:41 hey gedare 2012-06-26T11:02:46 hi Hesham 2012-06-26T11:02:57 can we chat now ? 2012-06-26T11:03:45 yea 2012-06-26T11:03:49 how's it going? 2012-06-26T11:04:15 i have asked quanming about the problem of isr 2012-06-26T11:04:27 he told me that he did not face it 2012-06-26T11:04:37 that's a problem! 2012-06-26T11:04:51 he told me you may know it 2012-06-26T11:04:57 i guess the issue must have started with my refactoring or rewriting of the test case 2012-06-26T11:05:09 did you try to start over with smaller, simpler test case? 2012-06-26T11:05:11 Aanjhan was in vacation 2012-06-26T11:05:50 i have extracted all quanming files before your refactoring 2012-06-26T11:05:57 and compiled it 2012-06-26T11:06:07 the problem is in mmutest1 2012-06-26T11:06:15 ./../../../../psim/lib/librtemscpu.a(dummy.o):(.data+0xd8): undefined reference to `main' 2012-06-26T11:06:25 i asked quanming about that problem 2012-06-26T11:06:35 he told me he had faced it before 2012-06-26T11:06:48 but did not remember how he solved it 2012-06-26T11:07:16 so i asked on the devel m-l 2012-06-26T11:07:23 but i got no reply till now 2012-06-26T11:07:45 probably you won't 2012-06-26T11:08:15 gotta find out where that dummy.o is coming from.... 2012-06-26T11:08:30 when i was trying to build my own test suite 2012-06-26T11:08:38 i faced the same problem of that dummy.o 2012-06-26T11:08:46 i told you about it 2012-06-26T11:09:15 but i avoided that problem with adding test case to the working mmutest1 2012-06-26T11:09:19 yeah. that dummy.o does not happen for regular samples? 2012-06-26T11:09:26 no 2012-06-26T11:09:45 only for mmutest1 or when i am trying to write my own test suite 2012-06-26T11:10:07 for your email you did not give nearly enough information for someone to help you 2012-06-26T11:10:52 i will post another one with sufficient details about the problem 2012-06-26T11:11:16 i also tried to delete all isr functions 2012-06-26T11:11:20 but i had the same error 2012-06-26T11:11:39 try a test without the clock driver enabled 2012-06-26T11:11:50 that's the only thing i can think is causing an isr. 2012-06-26T11:11:54 but the real problem isn't the isr... 2012-06-26T11:12:03 it is that the mmu registers are not being configured properly i would guess 2012-06-26T11:12:48 at code ? 2012-06-26T11:13:33 i'm not sure. send me instructions on how to build your project and what you are running on psim and i will try to replicate your work 2012-06-26T11:14:10 i have two versions now 2012-06-26T11:14:11 where can i see your test suite for mmu? 2012-06-26T11:14:45 the one recently one or quanming's ? 2012-06-26T11:14:55 *** QingPei has left #rtems 2012-06-26T11:15:10 the one recently 2012-06-26T11:16:26 it's at https://github.com/heshamelmatary/rtems/tree/arena_test 2012-06-26T11:17:10 where is it in the tree? 2012-06-26T11:18:00 i asked you to start with a simpler test case... you should copy something like hello world and then add some simple mmu calls 2012-06-26T11:18:06 testsuites/libtests/mmutest1/ 2012-06-26T11:18:24 instead of trying to reengineer mmutest1 2012-06-26T11:18:24 since you haven't gotten mmutest1 to work.. 2012-06-26T11:18:25 trying to go from something broken to a different thing working is hard 2012-06-26T11:18:54 when i tried to create my own test suite i got dummy.o error 2012-06-26T11:19:16 that's fine we need to reduce the test case to something that is simple and causes the problem 2012-06-26T11:19:17 i will try with hello sample again 2012-06-26T11:19:22 you get the dummy.o error with mmutest1 2012-06-26T11:19:30 yeah and with new test case 2012-06-26T11:19:34 so if you make a simple test case that produces dummy.o it will be easier to pin-point the problem 2012-06-26T11:19:55 this is called test case reduction 2012-06-26T11:20:01 and is important for debugging 2012-06-26T11:20:16 i want to see the minimal test case that causes the problem you have 2012-06-26T11:21:06 should i try to make use of hello sample or re-write new one ? 2012-06-26T11:21:13 make use of hello sample 2012-06-26T11:21:17 that is almost as simple as you can get 2012-06-26T11:21:31 you can make it simpler by removing the printf and disabling console support 2012-06-26T11:22:00 ok i will try it now 2012-06-26T11:27:51 if i want to try your code what branch should i use? 2012-06-26T11:27:55 i checked out your github 2012-06-26T11:32:53 you mean test cases ? 2012-06-26T11:33:09 i think there were some errors in code which i have fixed but did not pushed it 2012-06-26T11:33:42 i have tried to include 2012-06-26T11:33:42 #include "system.h" 2012-06-26T11:33:42 #include 2012-06-26T11:33:42 #include 2012-06-26T11:33:51 at hello sample 2012-06-26T11:34:02 i got the same dummy.o error 2012-06-26T11:34:21 then i deleted the #include "system.h" part 2012-06-26T11:34:27 it worked 2012-06-26T11:34:46 i'm looking at libmm branch right now. 2012-06-26T11:36:13 you could perhaps move the define CONFIGURE_INIT and include system.h to after the other includes 2012-06-26T11:36:21 sometimes that may make a difference.. 2012-06-26T11:37:00 hello sample does not have system.h 2012-06-26T11:37:11 i included it 2012-06-26T11:37:24 i see... did you delete the configure variables from the end of the init.c file? 2012-06-26T11:37:32 you might like to just try including all of them in init.c 2012-06-26T11:37:36 instead of using system.h file 2012-06-26T11:38:02 ok i will try 2012-06-26T11:40:54 is the prototype of rtems_task_init resides at system.h ? 2012-06-26T11:43:56 hmmm that should pull in with #include or similar 2012-06-26T11:44:14 but... you should not need to create a task in order to test mmu? 2012-06-26T11:45:51 How could i receive status of mmu HW ? 2012-06-26T11:46:03 config.status: error: cannot find input file: `Makefile.in' 2012-06-26T11:46:03 configure: error: ../../../../../../../../rtems/c/src/lib/libcpu/powerpc/configure failed for lib/libcpu/powerpc 2012-06-26T11:46:03 gmake[2]: *** [psim] Error 1 2012-06-26T11:46:14 bootstrap is not working right on libmm branch. checking this 2012-06-26T11:46:23 "receive status"? 2012-06-26T11:47:03 Makefile.am:228: blank line following trailing backslash 2012-06-26T11:47:22 simple fix 2012-06-26T11:47:23 i mean when initialize mmu hardware, i have to check status variable for success or failure and print it 2012-06-26T11:48:25 which makefile ? 2012-06-26T11:48:26 i'm confused about what you are asking. Add what you need to init.c so that you can make calls to libmm 2012-06-26T11:48:31 c/src/lib/libcpu/powerpc 2012-06-26T11:48:59 probably just #include 2012-06-26T11:49:15 to hello sample ? 2012-06-26T11:49:45 yeah 2012-06-26T11:49:59 trying to create a simple test of mmu 2012-06-26T11:50:42 i would just add a call to mm_initialize() and then exit..make sure that builds and runs 2012-06-26T11:51:01 ok 2012-06-26T11:52:59 i think your libmm branch is too broke for me to figure out right now 2012-06-26T11:54:07 libmmu.* ? 2012-06-26T11:54:24 what are you asking me? 2012-06-26T11:55:12 you mean that there are errors with libmmu.c and libmmu.h or what ? 2012-06-26T11:55:16 yeah 2012-06-26T11:55:46 I have compiled them successfully before 2012-06-26T11:56:49 powerpc-rtems4.11-gcc --pipe -B../../../../.././lib/ -B../../../../.././psim/lib/ -specs bsp_specs -qrtems -DPACKAGE_NAME=\"rtems-c-src-lib-libcpu-powerpc\" -DPACKAGE_TARNAME=\"rtems-c-src-lib-libcpu-powerpc\" -DPACKAGE_VERSION=\"4.10.99.0\" -DPACKAGE_STRING=\"rtems-c-src-lib-libcpu-powerpc\ 4.10.99.0\" -DPACKAGE_BUGREPORT=\"http://www.rtems.org/bugzilla\" -DPACKAGE_URL=\"\" -I. -I../../../../../../../../rtems/c/src/lib/libcpu/powerpc -isystem ../../../.. 2012-06-26T11:56:49 /.././psim/lib/include -mcpu=603e -Dppc603e -O2 -g -fno-keep-inline-functions -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -MT mpc6xx/mmu/mpc6xx_mmu_rel-mmu_support.o -MD -MP -MF mpc6xx/mmu/.deps/mpc6xx_mmu_rel-mmu_support.Tpo -c -o mpc6xx/mmu/mpc6xx_mmu_rel-mmu_support.o `test -f 'mpc6xx/mmu/mmu_support.c' || echo '../../../../../../../../rtems/c/src/lib/libcpu/powerpc/'`mpc6xx/mmu/mmu_support.c 2012-06-26T11:56:50 In file included from ../../../../../../../../rtems/c/src/lib/libcpu/powerpc/mpc6xx/mmu/mmu_support.c:23:0: 2012-06-26T11:56:52 ../../../../.././psim/lib/include/libcpu/memoryprotection.h:41:5: error: unknown type name 'rtems_memory_protection_permission' 2012-06-26T11:58:44 i think i did not pushed my work only to cpukit 2012-06-26T11:58:52 that's after renaming functions 2012-06-26T11:59:13 you have to pull another branch 2012-06-26T11:59:44 https://github.com/heshamelmatary/rtems/tree/mid/low-level_Work 2012-06-26T12:02:40 ok will try that one 2012-06-26T12:05:30 the hello sample test fails the same way when i included rtems_memory_protect_init ( ); ( quanming version ) 2012-06-26T12:05:51 with dummy.o ? 2012-06-26T12:05:59 or with the isr problem? 2012-06-26T12:06:13 isr problem 2012-06-26T12:06:31 *** MMU ALUT TEST 1 BEGINS *** 2012-06-26T12:06:31 initialize the memory protect manager 2012-06-26T12:06:31 cpu 1, cia 0x25b8: double interrupt - MSR[RI] bit clear when attempting to deliver interrupt, cia=0x25b8, msr=0x40; srr0=0x7c48(cia), srr1=0x8a072(msr); trap-vector=0x700, trap-msr=0x40 2012-06-26T12:08:12 that's the output of your modified hello world? 2012-06-26T12:08:37 copy-paste the body of your init() function here 2012-06-26T12:08:42 should be short.. 2012-06-26T12:08:43 yeah 2012-06-26T12:08:51 ok 2012-06-26T12:09:23 rtems_status_code status; 2012-06-26T12:09:23 puts( "\n\n*** MMU ALUT TEST 1 BEGINS ***\n" ); 2012-06-26T12:09:23 puts( "initialize the memory protect manager\n"); 2012-06-26T12:09:23 status = rtems_memory_protect_init ( ); 2012-06-26T12:09:23 if(status != RTEMS_SUCCESSFUL) 2012-06-26T12:09:24 { 2012-06-26T12:09:24 printf("Failed:initialize the memory protect manager failed; status = %d\n",status); 2012-06-26T12:09:25 exit(0); 2012-06-26T12:09:25 } 2012-06-26T12:09:26 printf("ALUT created\n"); 2012-06-26T12:09:26 printf( "\n\n*** HELLO WORLD TEST ***\n" ); 2012-06-26T12:09:27 printf( "Hello World\n" ); 2012-06-26T12:09:47 rtems_status_code status; 2012-06-26T12:09:47 puts( "\n\n*** MMU ALUT TEST 1 BEGINS ***\n" ); 2012-06-26T12:09:47 puts( "initialize the memory protect manager\n"); 2012-06-26T12:09:47 status = rtems_memory_protect_init ( ); 2012-06-26T12:09:47 if(status != RTEMS_SUCCESSFUL) 2012-06-26T12:09:48 { 2012-06-26T12:09:48 printf("Failed:initialize the memory protect manager failed; status = %d\n",status); 2012-06-26T12:09:49 exit(0); 2012-06-26T12:09:49 } 2012-06-26T12:10:31 good 2012-06-26T12:10:46 mmutest1 isn't building on this branch by the way. 2012-06-26T12:11:01 what's the error ? 2012-06-26T12:11:31 powerpc-rtems4.11-gcc -B../../../../../psim/lib/ -specs bsp_specs -qrtems -DHAVE_CONFIG_H -I. -I../../../../../../../../rtems/c/src/../../testsuites/libtests/mmutest1 -I.. -mcpu=603e -Dppc603e -O2 -g -fno-keep-inline-functions -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -MT init.o -MD -MP -MF .deps/init.Tpo -c -o init.o ../../../../../../../../rtems/c/src/../../testsuites/libtests/mmutest1/init.c 2012-06-26T12:11:31 ../../../../../../../../rtems/c/src/../../testsuites/libtests/mmutest1/init.c: In function 'Init': 2012-06-26T12:11:31 ../../../../../../../../rtems/c/src/../../testsuites/libtests/mmutest1/init.c:45:3: error: unknown type name 'rtems_memory_protection_entry' 2012-06-26T12:11:43 you have to pull the testing branch too 2012-06-26T12:11:50 ? 2012-06-26T12:12:11 https://github.com/heshamelmatary/rtems/tree/arena_test 2012-06-26T12:12:15 hmmmmmmmm 2012-06-26T12:12:20 *** sebhub has quit IRC (Ping timeout: 260 seconds) 2012-06-26T12:12:23 that's the last one 2012-06-26T12:12:26 i think you are using git some funny way 2012-06-26T12:12:59 you told me to create branches for libmm 2012-06-26T12:13:16 yeah. create a branch for each "topic" you work on 2012-06-26T12:13:28 ok so... 2012-06-26T12:13:40 is arena_test the combination of libmm and mid/low-level? 2012-06-26T12:14:06 no it's just testsuites/mmutest1 work 2012-06-26T12:14:16 it works on them 2012-06-26T12:14:32 a git branch stores all of the files in the tree based on where you branch from 2012-06-26T12:14:46 in the entire tree, not just a subtree 2012-06-26T12:15:26 i see 2012-06-26T12:15:32 go read more about gi 2012-06-26T12:15:33 git 2012-06-26T12:15:59 arena_test only has mmutest1 changes 2012-06-26T12:16:47 no it also has "Getting libmmu working on powerpc psim target" and "Implementing Arena_Initialize and Arena_Create without domain manager…" 2012-06-26T12:17:11 because you branched it from mid/low-level_Work 2012-06-26T12:17:17 which is fine, you just need to be clear about that 2012-06-26T12:18:00 i will delete them and make just one branch for libmm 2012-06-26T12:18:32 read http://git-scm.com/book so you understand 2012-06-26T12:18:57 I'd like you to provide me with libmm containing all the changes needed to use the result from previous gsoc stuff 2012-06-26T12:19:09 ok 2012-06-26T12:19:13 nothing to do with arenas 2012-06-26T12:19:22 just the mmu handling code and some test cases 2012-06-26T12:20:01 i did that on my local machine but i did not push it yet because of test cases error 2012-06-26T12:20:18 well I'd like something that I can replicate your work with 2012-06-26T12:20:18 as you told me 2012-06-26T12:20:23 right now i'm not sure what to use... 2012-06-26T12:20:57 In file included from ../../../../../../../../rtems/c/src/lib/libcpu/powerpc/../shared/src/memoryprotection_manager.c:12:0: 2012-06-26T12:20:57 ../../../../.././psim/lib/include/libcpu/memoryprotection.h:41:5: error: unknown type name 'rtems_memory_protection_permission' 2012-06-26T12:21:00 that's in arena_test branch 2012-06-26T12:22:16 the problem of renaming functions 2012-06-26T12:24:59 you may pull Mid-low level branch and edit/mmutest1/init.c according to the commit of that file at arena_test branch 2012-06-26T12:25:06 just to work for now 2012-06-26T12:25:21 and i will fix that error 2012-06-26T12:31:02 i'll wait until you fix it and push a working version (that compiles anyway) 2012-06-26T12:31:19 ok 2012-06-26T12:31:25 i have two versions now 2012-06-26T12:31:36 quanming version 2012-06-26T12:31:52 and the edited one 2012-06-26T12:32:00 which one to push ? 2012-06-26T12:32:40 i suggest that you push quanming version, and then push the updates you make 2012-06-26T12:32:50 so we have the history available 2012-06-26T12:33:23 my guess about the MSR bit is that the mmu exception handler is generating an mmu exception 2012-06-26T12:33:25 which is bad news 2012-06-26T12:33:54 basically it means the mmu exception handler for powerpc is using code or data that is not mapped by the mmu 2012-06-26T12:34:04 cascaded mmu exceptions = death 2012-06-26T12:34:51 But why that error arises now although it just worked fine with quanming ? 2012-06-26T12:35:45 i changed the code quite a bit to reorganize and make it more generic 2012-06-26T12:36:00 probably initialization of the mmu hw does something differently now than it did before with quanming's code 2012-06-26T12:37:12 aha 2012-06-26T12:37:30 What may be the solution for that problem ? 2012-06-26T12:37:49 one thing to do is to hook up gdb and step through initialize() until the error 2012-06-26T12:39:02 the gdb produces the error of PVR value 2012-06-26T12:39:17 i asked quanming about it too and he did not know 2012-06-26T12:39:24 and i asked about it at devel m-l 2012-06-26T12:39:42 what line does it? 2012-06-26T12:40:06 you should set a breakpoint at _CPU_Memory_protection_Initialize and step through 2012-06-26T12:41:11 PVR is the Processor Version Register and is read-only 2012-06-26T12:41:26 Unknown PVR value of 0x0. Please add it to 2012-06-26T12:42:00 that's the error even with hello sample 2012-06-26T12:42:10 ah, broken powerpc-gdb then 2012-06-26T12:42:50 you were not clear about that in your email 2012-06-26T12:43:35 i asked that question two times at devel ml 2012-06-26T12:43:48 anyway 2012-06-26T12:44:11 i will create just one branch for all my work 2012-06-26T12:44:13 is it ok ? 2012-06-26T12:44:21 did you use psim-gdb from rtems-testing repo 2012-06-26T12:44:30 yeah sure 2012-06-26T12:44:52 aha i used it 2012-06-26T12:45:20 it works just fine now 2012-06-26T12:46:17 that's what drjoel suggested you try when you reported this issue last month. 2012-06-26T12:46:31 but it has the same isr error 2012-06-26T12:46:41 type bt 2012-06-26T12:46:44 when it breaks there 2012-06-26T12:47:10 Unknown PVR value of 0x0. Please add it to 2012-06-26T12:47:32 you mean after it breaks i should type bt ? 2012-06-26T12:47:35 yea] 2012-06-26T12:47:37 yeah 2012-06-26T12:47:37 No stack. 2012-06-26T12:47:45 ok it errors out completely... 2012-06-26T12:48:16 so instead you should set a breakpoint and step until it breaks 2012-06-26T12:48:28 that should give you an idea of what line of code causes the problem 2012-06-26T12:48:33 ok i will try it now 2012-06-26T13:07:25 gedare: _CPU_cache_flush_1_data_line 2012-06-26T13:07:51 progress... 2012-06-26T13:07:55 what should be the range of addresses of that chache buffer ? 2012-06-26T13:08:16 I guess that is coming from rtems_cache_flush_multiple_data_lines 2012-06-26T13:08:47 quanming set that to the entire memory range (RamBase, RamBase+RamSize) in order to flush the whole data cache 2012-06-26T13:08:58 i am stepping until i reached that point of code 2012-06-26T13:09:14 ok what is the range of the entire memory ? 2012-06-26T13:10:18 when you reach that call do backtrace 2012-06-26T13:10:23 and you can see where it came from 2012-06-26T13:10:42 but my guess is pagetable_.c: _CPU_Memory_protection_Initialize 2012-06-26T13:11:06 yeah i set a break point there 2012-06-26T13:11:17 and just stepping from that point 2012-06-26T13:11:32 until i reached that part 2012-06-26T13:11:57 it just increment addr and calls that function again 2012-06-26T13:11:58 addr=0x197c40 2012-06-26T13:12:11 addr=0x36c840 2012-06-26T13:12:24 do you know which address happens when it dies? 2012-06-26T13:12:24 is this reasonable address ? 2012-06-26T13:12:30 not yet 2012-06-26T13:12:37 RamSize is 16M 2012-06-26T13:12:42 and I think RamBase is 0 2012-06-26T13:12:53 (c/src/lib/libbsp/powerpc/psim/startup/linkcmds) 2012-06-26T13:13:22 so ? 2012-06-26T13:14:34 1 meg is 0x100000, so 16M is 0x1000000 > 0x36c840 2012-06-26T13:14:46 so that address is certainly within the ram... 2012-06-26T13:15:05 hmmm 2012-06-26T13:17:50 need to figure out why an interrupt occurs while flushing the cache 2012-06-26T13:18:23 the problem is that the IR bit is cleared, but when an exception occurs during the cleared IR bit this causes the death 2012-06-26T13:19:49 i am trying to reach the end of addresses and find what happens after that 2012-06-26T13:20:13 ok i could see that taking awhile :) 2012-06-26T13:20:23 by the way is this on a branch you can push so i can see? or not clean yet? 2012-06-26T13:20:58 you could try to comment out the 'rtems_cache_flush_multiple_data_lines' ... I'm not sure what will happen if the cache is not flushed though 2012-06-26T13:21:18 ok i will try it 2012-06-26T13:21:33 this is quanming version which i am debugging 2012-06-26T13:21:57 when i am creating a branch 2012-06-26T13:22:11 where should my current working directory be ? 2012-06-26T13:22:35 doesn't matter. the branch will include everything in the repository 2012-06-26T13:23:24 then when i push, where should i stay ? 2012-06-26T13:23:59 everything that is committed will get pushed 2012-06-26T13:24:34 aha 2012-06-26T13:24:46 can i delete the current branches ? 2012-06-26T13:25:00 from github & my local machine ? 2012-06-26T13:25:11 yes its a little weird for github 2012-06-26T13:25:19 for a local branch its just ... git branch -d 2012-06-26T13:25:25 or git branch -D to force it 2012-06-26T13:25:38 for remote branch it is ... git push : 2012-06-26T13:25:50 i think that is the way 2012-06-26T13:26:02 go read about git ;) 2012-06-26T13:26:54 i must 2012-06-26T13:27:29 i just wanna figure the currently error and solve it :( 2012-06-26T13:28:03 yeah i know. you've been stuck for awhile 2012-06-26T13:28:05 i will set a beak point after flushing cache and step from there 2012-06-26T13:28:42 good 2012-06-26T13:45:01 gedare: it reaches the end of the function and then produces the isr error 2012-06-26T13:51:04 end of what function 2012-06-26T13:51:30 Hesham: ^ 2012-06-26T13:52:23 _CPU_Pagetable_Initialize 2012-06-26T13:52:34 (quanming) 2012-06-26T13:52:51 ok 2012-06-26T13:53:03 does it return from that function?/ 2012-06-26T13:53:23 no 2012-06-26T13:53:35 what instruction does it die on? 2012-06-26T13:53:53 } 2012-06-26T13:54:27 it dies when it reaches "}" 2012-06-26T13:54:33 ok maybe it is the CLRBAT() assembly 2012-06-26T13:55:22 where is it and what should i do ? 2012-06-26T13:56:31 at this point i would say to respond to quanming and aanjhan with your findings 2012-06-26T13:57:49 ok, can i work on another part until they reply ? 2012-06-26T13:58:30 yeah umm 2012-06-26T13:58:55 arena ? 2012-06-26T13:59:16 well, fix your git issues for one thing 2012-06-26T13:59:44 ah 2012-06-26T13:59:45 and then you can work on arena in parallel to getting libmm to work properly 2012-06-26T14:00:08 i will create new branch after deleting previous branches 2012-06-26T14:00:18 then push quanming work 2012-06-26T14:00:19 ok. put up the quanming version too as your first commit 2012-06-26T14:00:22 great 2012-06-26T14:00:33 then the recently one 2012-06-26T14:00:33 i will try that out to see if it fails for me the same way 2012-06-26T14:00:42 if it does then we can bug them for more help. 2012-06-26T14:00:54 how much have you read / thought about the low-level ARM mmu code? 2012-06-26T14:01:45 initializing the MMU and creating pagetables with attributes 2012-06-26T14:02:10 i wonder how hard it would be to do the arm support code? 2012-06-26T14:02:48 the arena code does not make a lot of sense without libmm in place 2012-06-26T14:02:55 i noticed there is already a source file that handles some mmu operations 2012-06-26T14:02:57 but libmm can't be finished without a working prototype 2012-06-26T14:03:29 yeah 2012-06-26T14:03:41 powerpc or arm 2012-06-26T14:04:12 for now we are trying with powerpc 2012-06-26T14:04:54 you can spend some time on either low-level arm or arena while you wait for answers about the powerpc problem 2012-06-26T14:05:38 and fix your git :) 2012-06-26T14:05:46 i will :) 2012-06-26T14:05:57 but what you suggest working on arm or arena ? 2012-06-26T14:07:49 arena 2012-06-26T14:08:07 even if all you have are stubs for libmm you can still implement arena 2012-06-26T14:08:38 so put together the libmm branch, and then branch for arena work and add a stub implementation of libmm for some architecture to use for testing 2012-06-26T14:08:45 either use arm or sparc32 with stubs 2012-06-26T14:09:16 sparc32 might be better, since you can also make a branch of libmm wherre you start to implement the arm libmm support 2012-06-26T14:09:48 and not have many conflicts among your branches 2012-06-26T14:10:06 so after i create libmm branch 2012-06-26T14:10:30 i will create another one for sparc32 ? 2012-06-26T14:11:22 yeah you can just make a branch from your libmm branch 2012-06-26T14:11:35 add sparc32 stubs, and get to work on arena code. 2012-06-26T14:12:57 would i chose the place of these stubs ? 2012-06-26T14:15:26 should be like the implementations for powerpc and sparc64 are laid out 2012-06-26T14:15:40 i mean, you really are just going to use "no_memoryprotection.c" 2012-06-26T14:15:48 it doesn't really take any work... 2012-06-26T14:15:57 just pointing the makefile at the right file 2012-06-26T14:16:01 the stubs already exis 2012-06-26T14:16:40 and what about the simulator for sparc32 ? 2012-06-26T14:16:51 gdb ? 2012-06-26T14:17:54 brb meeting 2012-06-26T14:17:59 sis 2012-06-26T14:18:23 ok thanks for your time & help 2012-06-26T14:21:31 *** Hesham has quit IRC (Quit: Leaving.) 2012-06-26T15:28:13 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-06-26T15:55:21 *** gedare has quit IRC (Quit: Ex-Chat) 2012-06-26T18:25:50 *** kristianpaul has joined #rtems 2012-06-26T18:25:50 *** kristianpaul has joined #rtems 2012-06-26T22:17:01 *** alseh has quit IRC (Read error: Operation timed out) 2012-06-26T22:31:16 *** alseh has joined #rtems 2012-06-26T22:41:54 *** deb has left #rtems ("Leaving") 2012-06-26T23:41:05 *** arvind_khadri has joined #rtems 2012-06-27T01:50:29 *** sebhub has joined #rtems 2012-06-27T01:54:44 good morning 2012-06-27T02:03:53 morning 2012-06-27T02:11:31 *** QingPei has joined #rtems 2012-06-27T03:31:19 *** weiY has joined #rtems 2012-06-27T03:32:35 Hi chris, are you around? 2012-06-27T03:33:46 Hi sebhub, are you around? 2012-06-27T03:34:02 hi 2012-06-27T03:34:39 now i have encounter some problem about qemu to run pcp4 rtems test 2012-06-27T03:35:03 do you now the command to run the application on qemu for pcp4? 2012-06-27T03:35:09 know 2012-06-27T03:35:21 is pcp4 a bsp name? 2012-06-27T03:36:14 yes, its a bsp name (i386 architecture i think) 2012-06-27T03:36:42 i see the readme under the smptest, it seems that only leon3 and pcp4 bsp support smptest 2012-06-27T03:38:05 hm, there is no script in the rtems-testing module 2012-06-27T03:40:34 and any script to run pc386 bsp on qemu? 2012-06-27T03:40:42 yes, but this is single core 2012-06-27T03:41:31 http://git.rtems.org/rtems-testing 2012-06-27T03:41:54 maybe you can ask joel to add a pcp4 script 2012-06-27T03:42:27 ok, thank you, i will study it. and joel have no time and i send mail to him not got reply 2012-06-27T03:45:24 sebhub, which file is the script to run pc386 under the rtems-testing? 2012-06-27T04:05:46 *** QingPei has left #rtems 2012-06-27T04:07:07 they are all in sim-scripts and have the bsp name 2012-06-27T04:07:18 there is also a README 2012-06-27T05:39:06 *** weiY has quit IRC () 2012-06-27T07:04:14 *** cdcs has joined #rtems 2012-06-27T08:52:22 *** alseh has quit IRC (Read error: Connection reset by peer) 2012-06-27T08:55:16 *** alseh has joined #rtems 2012-06-27T10:03:11 *** alseh has quit IRC (Read error: Operation timed out) 2012-06-27T10:10:45 *** alseh has joined #rtems 2012-06-27T10:33:56 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-06-27T10:49:21 *** sebhub has quit IRC (Ping timeout: 252 seconds) 2012-06-27T11:34:04 *** Hesham has joined #rtems 2012-06-27T11:55:40 *** tuxmaniac has joined #rtems 2012-06-27T11:55:41 *** tuxmaniac has joined #rtems 2012-06-27T11:56:01 Hesham: around? 2012-06-27T11:56:18 yeah i am here 2012-06-27T12:12:02 *** cdcs has quit IRC (Quit: Page closed) 2012-06-27T13:15:46 *** tuxmaniac has quit IRC (Quit: tuxmaniac) 2012-06-27T14:23:31 *** panzon has quit IRC (Ping timeout: 264 seconds) 2012-06-27T14:36:15 *** panzon has joined #rtems 2012-06-27T14:56:06 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-06-27T20:52:56 *** alseh_ has joined #rtems 2012-06-27T20:55:36 *** alseh has quit IRC (Ping timeout: 265 seconds) 2012-06-27T21:08:28 *** xiangfu has joined #rtems 2012-06-27T21:23:08 *** kiwichris__ has joined #rtems 2012-06-27T21:25:53 *** kiwichris_ has quit IRC (Ping timeout: 244 seconds) 2012-06-27T22:14:21 *** xiangfu has quit IRC (Quit: Leaving) 2012-06-27T22:16:46 *** xiangfu has joined #rtems 2012-06-27T22:17:46 *** alseh_ has quit IRC (Ping timeout: 265 seconds) 2012-06-27T22:40:57 *** alseh has joined #rtems 2012-06-27T23:04:23 *** xiangfu has quit IRC (*.net *.split) 2012-06-27T23:04:24 *** kiwichris__ has quit IRC (*.net *.split) 2012-06-27T23:09:57 *** xiangfu has joined #rtems 2012-06-27T23:09:57 *** kiwichris__ has joined #rtems 2012-06-27T23:22:18 *** xiangfu has quit IRC (*.net *.split) 2012-06-27T23:22:18 *** kiwichris__ has quit IRC (*.net *.split) 2012-06-27T23:28:17 *** xiangfu has joined #rtems 2012-06-27T23:28:17 *** kiwichris__ has joined #rtems 2012-06-27T23:57:58 *** arvind_khadri has joined #rtems 2012-06-28T01:09:58 *** sebhub has joined #rtems 2012-06-28T01:10:54 good morning 2012-06-28T01:36:19 *** alseh has quit IRC (Remote host closed the connection) 2012-06-28T01:39:33 *** aanjhan has joined #rtems 2012-06-28T03:31:35 *** cdcs has joined #rtems 2012-06-28T03:45:29 *** aanjhan has quit IRC (Quit: aanjhan) 2012-06-28T03:52:58 *** aanjhan has joined #rtems 2012-06-28T03:59:01 *** aanjhan has quit IRC (Quit: aanjhan) 2012-06-28T04:37:18 *** aanjhan has joined #rtems 2012-06-28T04:54:02 *** aanjhan has quit IRC (Quit: aanjhan) 2012-06-28T06:21:30 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-06-28T07:11:38 *** aanjhan has joined #rtems 2012-06-28T07:23:22 *** aanjhan has quit IRC (Quit: aanjhan) 2012-06-28T08:03:33 Hello everybody 2012-06-28T08:18:24 *** zw_yao has joined #rtems 2012-06-28T08:18:36 *** xiangfu has quit IRC (Quit: Leaving) 2012-06-28T08:19:06 *** zw_yao has quit IRC (Client Quit) 2012-06-28T08:34:31 *** sebhub has quit IRC (Remote host closed the connection) 2012-06-28T08:46:42 *** arvind_khadri has joined #rtems 2012-06-28T08:59:09 *** alseh has joined #rtems 2012-06-28T09:23:35 *** WikL has joined #rtems 2012-06-28T09:24:21 good evening all 2012-06-28T09:27:01 hi 2012-06-28T09:27:12 cdcs: are you here ? 2012-06-28T09:27:18 WikL: shall we start now ? 2012-06-28T09:30:08 *** WikL has quit IRC (Ping timeout: 240 seconds) 2012-06-28T09:30:23 *** WikL has joined #rtems 2012-06-28T09:31:00 WikL: are you there ? 2012-06-28T09:31:23 yes 2012-06-28T09:31:25 ok 2012-06-28T09:31:26 *** jennifer has joined #rtems 2012-06-28T09:31:32 we need claudio then 2012-06-28T09:31:32 got disconnected for a second there 2012-06-28T09:31:40 cdcs: are you there ? 2012-06-28T09:32:05 Hello 2012-06-28T09:32:40 sry. My IRC doesn't give me activity notifications 2012-06-28T09:33:40 ok 2012-06-28T09:33:43 no problem 2012-06-28T09:33:44 ok 2012-06-28T09:33:46 so 2012-06-28T09:33:51 I think we can start ? 2012-06-28T09:34:02 WikL: please tell us updates about your status 2012-06-28T09:34:45 I finally secured a stable compilation and linking on RTEMS part 2012-06-28T09:34:58 with using a parameter passed to configure 2012-06-28T09:35:17 turns out there were not only conflicts in the cpukit directory 2012-06-28T09:35:56 but also in the shared directory in libcpu 2012-06-28T09:35:56 ok 2012-06-28T09:35:59 sounds great 2012-06-28T09:36:13 which are always included while compiling for i386 2012-06-28T09:37:11 it's a bit of a mess and sometimes it requires issuing a make clean while making changes inside libbsp 2012-06-28T09:37:24 but works 2012-06-28T09:39:42 right now I'm looking into the compilation process of POK to make it stick together 2012-06-28T09:40:27 at this point i think i would leave that for later. 2012-06-28T09:40:42 ok 2012-06-28T09:40:46 ok, so 2012-06-28T09:40:54 do you need more support or just time to make the work ? 2012-06-28T09:40:56 I think they can be separate build processes. You just need to include RTEMS libs on POK 2012-06-28T09:41:46 and not POK libs on RTEMS? 2012-06-28T09:42:03 since the "BSP" will be calling POK functions 2012-06-28T09:42:42 Those symbols will be left undefined by the compiler 2012-06-28T09:43:12 They will be linked you when link everything together: POK+RTEMS+UserApplication 2012-06-28T09:44:35 right 2012-06-28T09:44:54 ok so 2012-06-28T09:45:00 I think cdcs is right 2012-06-28T09:45:12 probably you will need some linker script magic (linkcmds as named by rtems) but nothing to complicated. 2012-06-28T09:45:12 you can just say : I want to integrate RTEMS on top of RTEMS 2012-06-28T09:45:16 on top of POK 2012-06-28T09:45:31 and see your RTEMS bin as a black box 2012-06-28T09:45:45 cdcs: right and by the way, it would be sufficient for the mid-term review 2012-06-28T09:45:53 *** ppisa has joined #rtems 2012-06-28T09:46:53 as a side note: my semester ends at the beginning of the next week and from then on I will have all of my time at the project's disposal 2012-06-28T09:47:14 ok 2012-06-28T09:47:18 ok 2012-06-28T09:47:23 but it does not change the mid term review date 2012-06-28T09:47:27 *** deb has joined #rtems 2012-06-28T09:47:34 for the mid term, we just expect a sample application with partitions 2012-06-28T09:47:52 and due to the constraints, it is obvious that some tricks can be used 2012-06-28T09:47:57 as the linker scripts, etc ... 2012-06-28T09:48:04 you should focus on "making thinks work" 2012-06-28T09:48:10 right 2012-06-28T09:48:13 so both parts will have unresolved symbols 2012-06-28T09:48:22 ? 2012-06-28T09:48:24 yes. 2012-06-28T09:48:24 no 2012-06-28T09:48:27 I mean 2012-06-28T09:48:31 On RTEMS 2012-06-28T09:48:36 you build your system 2012-06-28T09:48:48 and then, on POK, you build POK for one or two partitions 2012-06-28T09:49:00 and then, you have a particular linkscript for this particular example 2012-06-28T09:49:19 and the linkscript put the partitions in a known memory area 2012-06-28T09:49:28 so that you can assume where the partitions are 2012-06-28T09:49:29 I mean 2012-06-28T09:49:35 this would be sufficient for a first shot 2012-06-28T09:49:38 a first demo 2012-06-28T09:50:50 hallo ? 2012-06-28T09:50:53 that sounds clear ? 2012-06-28T09:50:55 you can do something like: memory: | POK Kernel | P1: RTEMS+User_App1 | P2: RTEMS+User_App2 | 2012-06-28T09:51:05 I'm here, just thinking it through 2012-06-28T09:51:11 cdcs: exactly 2012-06-28T09:51:24 WikL: ok but I have a teleconf in 5 minutes 2012-06-28T09:51:33 so, if you have questions, please do so now :-) 2012-06-28T09:51:51 Anyway, we are still available by mail of course 2012-06-28T09:52:10 I understand the idea, just have to figure out the implementation 2012-06-28T09:52:43 just have to look into linker scripts a bit, that's all 2012-06-28T09:53:30 your RTEMS compilation will have the POK system calls undefined and the User Entry point/ Configuration undefined 2012-06-28T09:53:32 ok 2012-06-28T09:53:40 In POK, you can have a look at pok/misc/ldscripts/x86/x86-qemu 2012-06-28T09:53:51 you have the ldscript we use for the kernel and the partitions 2012-06-28T09:54:12 the the partitions, we usually put them after the ++po_end symbol 2012-06-28T09:54:25 and thus, when the kernel loads, it loads elf files starting from this symbol 2012-06-28T09:54:53 And we also add in the binary an information about the size of each ELF file embedded in the __po_end symbol 2012-06-28T09:55:32 looks clear 2012-06-28T09:55:48 do you have other questions ? 2012-06-28T09:56:19 not right now 2012-06-28T09:56:37 Yes and now your partition will contain your modified RTEMS plus the user app. 2012-06-28T09:56:48 cdcs: right 2012-06-28T09:56:49 ok 2012-06-28T09:57:00 WikL: so, I advise you to update the shared google doc from joel 2012-06-28T09:57:07 I have already 2012-06-28T09:57:13 and then, come back to us when you have questions 2012-06-28T09:57:20 will update it according to what we established now 2012-06-28T09:57:23 We stay available for you 2012-06-28T09:57:30 thanks 2012-06-28T09:57:31 and hope to see the demo for the mid-term review 2012-06-28T09:57:34 :-) 2012-06-28T09:57:37 Good luck ! 2012-06-28T09:57:48 thanks again, looking forward to the demo too 2012-06-28T09:57:50 * juli1 is going for another meeting now :D 2012-06-28T09:58:00 bye juli1 2012-06-28T09:58:58 should I stay in and report to Joel too? 2012-06-28T09:59:22 your partition entry point will now be boot_card i think. 2012-06-28T09:59:38 Joel is in another meeting this morning and I think I am managing the meeting 2012-06-28T10:00:09 So you should report to jennifer ;) 2012-06-28T10:00:26 :) 2012-06-28T10:01:11 Please modify the google doc: https://docs.google.com/document/d/1RFWt8k3dD6hn_hI9hlxQgS7AD5goZk-3DiGxWVJ-8Kk/ed 2012-06-28T10:01:25 "Sorry, the page (or document) you have requested does not exist." 2012-06-28T10:01:42 https://docs.google.com/document/d/1RFWt8k3dD6hn_hI9hlxQgS7AD5goZk-3DiGxWVJ-8Kk 2012-06-28T10:01:59 thanks :) 2012-06-28T10:02:05 https://docs.google.com/document/d/1RFWt8k3dD6hn_hI9hlxQgS7AD5goZk-3DiGxWVJ-8Kk/edit 2012-06-28T10:02:53 Who wants to start today? 2012-06-28T10:03:34 *** zw_yao has joined #rtems 2012-06-28T10:05:07 I suggest to use oder in Attending Students Google doc 2012-06-28T10:05:08 *** claas_ziemke has joined #rtems 2012-06-28T10:05:23 Ok 2012-06-28T10:05:54 claas_ziemke? 2012-06-28T10:05:57 hi 2012-06-28T10:06:01 yes i am here 2012-06-28T10:06:08 How is it going? 2012-06-28T10:06:18 going good :-) 2012-06-28T10:06:53 Any problems? Are you ready for midterm? 2012-06-28T10:07:12 just wasnt tooo productuve last week, because i constantly had to play soccer witm my nephews... u know... soccer euro championship... :-) 2012-06-28T10:07:28 yeah i think so... what do i have to provide? 2012-06-28T10:08:05 You are evaluated on what you have done so far on your project 2012-06-28T10:08:18 You should be about half way through your schedule 2012-06-28T10:08:53 ok, i update my schedule on the google docs document... 2012-06-28T10:09:00 ment will update 2012-06-28T10:09:09 great :) 2012-06-28T10:09:43 did i miss the meeting? 2012-06-28T10:10:18 This is this weeks meeting 2012-06-28T10:10:39 ah ok, i thought i was on time :-) 2012-06-28T10:11:23 If you don't have anything else to report or any problems we can move on to WikL 2012-06-28T10:11:45 ok, i have something to report 2012-06-28T10:11:51 ok 2012-06-28T10:12:07 i found git://gitorious.org/+freebsd-omap-team/freebsd/freebsd-omap.git which actually is the freebsd port for beagleboard 2012-06-28T10:12:31 so i start reading their driver code in order to see how i can reuse some of it for the rtems port 2012-06-28T10:12:49 as i understand freebsd license should work with rtems 2012-06-28T10:13:01 good! How it it looking for reuse 2012-06-28T10:13:36 cant say till now, i am not very familiar with freebsd kernel, so i have to learn a little bit about it before i can say if we can reuse the code 2012-06-28T10:13:43 but it looks promising 2012-06-28T10:14:22 second thing i did is to read a lot of hardware documentation so i have an idea how to start implementeing the uart driver now 2012-06-28T10:14:30 todo for next week: 2012-06-28T10:14:41 i will start implementing the driver at the weekend 2012-06-28T10:15:07 and update documentation in order to get tready for the midterm evaluation 2012-06-28T10:15:11 thats it 2012-06-28T10:15:30 great 2012-06-28T10:15:36 any questions? 2012-06-28T10:15:54 Let us know if you have any question on the rtems-libbsd code 2012-06-28T10:16:00 ok 2012-06-28T10:16:23 Sounds like you are making good progress 2012-06-28T10:16:32 thant thats all from my side... ah before i forget, joel wanted to send me an email with some tutorial stuff on rtems drivers 2012-06-28T10:16:40 Next WikL 2012-06-28T10:16:47 here 2012-06-28T10:16:49 i didnt receive anything, so plz remind him, should be helpfull 2012-06-28T10:16:56 now im done 2012-06-28T10:16:59 I'll make sure that gets sent 2012-06-28T10:17:39 thx 2012-06-28T10:17:55 well I managed to get passed a problem I've been struggling with for a while now concerning the compilation on a somewhat modified RTEMS rtee 2012-06-28T10:17:57 your welcome :) 2012-06-28T10:18:17 good 2012-06-28T10:18:18 *** soh_cah_toa has joined #rtems 2012-06-28T10:18:29 and put to work what was established on the previous meeting 2012-06-28T10:18:40 there are some dirty hacks but it works 2012-06-28T10:19:25 Do you have that in your repository ready for review? 2012-06-28T10:19:25 I just discussed with juli1 and cdcs what has to be done next, and at the moment I'm clear about the immediate goals 2012-06-28T10:19:35 great! 2012-06-28T10:19:38 in repository yes 2012-06-28T10:20:13 but I should look at it one more time to see if anything needs cleening up 2012-06-28T10:20:32 good idea 2012-06-28T10:21:33 If you don't have anything else we can move on to alseh 2012-06-28T10:21:36 *** gedare has joined #rtems 2012-06-28T10:21:49 yes, that is all 2012-06-28T10:22:27 ok, hi all 2012-06-28T10:22:46 so, I now have NXLib building with RTEMS enviroment 2012-06-28T10:22:59 *** QingPei has joined #rtems 2012-06-28T10:23:00 and curently am using their X library subset 2012-06-28T10:23:52 I still have a minor problem with linking a test program, but it should be passed soon 2012-06-28T10:24:08 *** arvind_khadri has quit IRC (Ping timeout: 240 seconds) 2012-06-28T10:24:33 Have you looked at that DJGPP packages integration? It should be quite similar to RTEMS needs. 2012-06-28T10:24:57 yes, have looked at them 2012-06-28T10:25:15 I could not work on this too much because I have an exam tomorow 2012-06-28T10:26:13 what I am unclear now is that there is quite a difference from their X header files to my linux ones (as I compared them) 2012-06-28T10:26:24 how could I decide what is needed subset ? 2012-06-28T10:27:32 I would expet that DJGPP one is older but selected to be sufficient to compile FLTK. So I expet that it worth to try them. 2012-06-28T10:27:59 ok, so I will use them and FTLK as reference 2012-06-28T10:28:37 and in paralel, for the following libraries I wanted to upgrade, I was thinking of creating a script. I think NXLib should also be included in that script 2012-06-28T10:28:50 Then it would worth to test current version. 2012-06-28T10:32:27 I mean after starting with DJGPP one. 2012-06-28T10:33:12 current version of what ? .. I don't get it 2012-06-28T10:34:14 The current version of the X headers from Linux. 2012-06-28T10:34:26 oh, right :) 2012-06-28T10:34:31 The DJGPP first and then actual. 2012-06-28T10:34:36 understand 2012-06-28T10:35:28 Any other problems/questions/comments? 2012-06-28T10:35:52 From my side no. 2012-06-28T10:35:55 should I start some documentation ? ... update the RTEMS Graphical Toolkit wiki before midterm ? 2012-06-28T10:36:38 and nothing else from my side, I just need some more time to fix my problems and check things 2012-06-28T10:37:11 Yes update as much as possible. Setup, testing, pointers to the code, etc. 2012-06-28T10:37:35 ok, thank you. than that's all 2012-06-28T10:37:45 Ok... Next up is Deb 2012-06-28T10:37:56 hello 2012-06-28T10:38:03 I have been doing well 2012-06-28T10:38:29 I talked to my mentor about the expected outcome for midterm 2012-06-28T10:38:44 I think I am making progress for it. 2012-06-28T10:38:50 Not ready for review though 2012-06-28T10:39:15 Just have one question. 2012-06-28T10:39:24 ok 2012-06-28T10:40:13 If certain tests, not hardware involved, are working on sis. Is it guarantteed to work on other BSPs? 2012-06-28T10:40:23 no 2012-06-28T10:40:26 FP and special features are disabled 2012-06-28T10:40:29 OK. 2012-06-28T10:40:56 How to ensure that it works well on other platforms? 2012-06-28T10:40:58 there is a probability it will work, but there can always be a problem on another bsp 2012-06-28T10:41:23 You can test on all the available simulated platforms pretty easily 2012-06-28T10:42:06 That question ties in to the scripts available in rtems-testing: making testing across multiple simulators easier could be another goal for you. 2012-06-28T10:42:07 what are the list of available simulated platforms? 2012-06-28T10:42:12 The hardware ones, we have to rely on people that have the hardware for acual testing... and just make sure it compiles and links without problems 2012-06-28T10:42:27 thank you gedare 2012-06-28T10:42:41 http://wiki.rtems.org/wiki/index.php/Emulator 2012-06-28T10:42:50 OK 2012-06-28T10:42:55 that page is probably outdated but gives some idea 2012-06-28T10:43:35 OK 2012-06-28T10:43:45 but you should work with your mentors to determine if you should focus effort on generalizing your framework for multiple platforms, or on getting it to work for sis and then expanding it out 2012-06-28T10:43:55 both approaches have benefits 2012-06-28T10:44:11 OK. I am working for the latter one currently 2012-06-28T10:44:33 I am choosing the framework to use basing on testing it on sis 2012-06-28T10:45:19 I have no more questions. 2012-06-28T10:45:40 Ok next up ... zw_yao 2012-06-28T10:46:07 thank you jennifer 2012-06-28T10:46:17 Hi, jennifer. My status: I have added gedare's idea (in the last mail) to my implementation, and pass the 5 key tests in current testsuites. 2012-06-28T10:47:20 great .. and have you ran into any problems 2012-06-28T10:47:42 Excepting the critical region issues,my first version of POSIX key have been nearly done. 2012-06-28T10:48:06 zw_yao: you sent an email offlist and i responded to it. have you incorporated my suggestions there also for thread deletion and key deletion? 2012-06-28T10:48:36 Yeah, I've added them to my implementation now. 2012-06-28T10:48:39 you should send a brief report to rtems-devel, and prepare a branch on your github with the commits 2012-06-28T10:48:49 Pushed to github. 2012-06-28T10:48:55 OK, I will. 2012-06-28T10:49:03 ok send link to the branch to the ml 2012-06-28T10:49:11 along with your brief report of what you did 2012-06-28T10:49:23 it should be easy to pull your branch and build it and run it 2012-06-28T10:49:29 Ok, I see. 2012-06-28T10:49:49 what's next? 2012-06-28T10:50:03 However, the critical region related haven't been considered.. 2012-06-28T10:51:07 I'll dicuss this later on list. 2012-06-28T10:51:09 ok so you need to resolve that issue. also do the test cases include tests for key deletion when multiple tasks have the same pthread_key_t, or for thread deletion/creation while using tasks? 2012-06-28T10:51:59 Should I write the test? 2012-06-28T10:52:13 Or just run the current test in the testsuites? 2012-06-28T10:52:40 evaluate if the available tests are testing everything... 2012-06-28T10:53:00 i have a feeling the tests do not deal with the two issues... key delete with multiple threads, and thread delete with keys 2012-06-28T10:53:13 but i have not looked at it myself 2012-06-28T10:53:53 OK, I'll also read the test, and make sure those tests cover my code. 2012-06-28T10:54:47 Does any doc need to be done before midterm? 2012-06-28T10:55:33 evaluate what you proposed to have done...see how close you are; i think that submitting test cases can be a really good contribution to make 2012-06-28T10:56:51 I see. I was also expected to add tests in my proposal. 2012-06-28T10:57:07 sounds like you know what you have to do then? 2012-06-28T10:58:03 I remember Joel said I should also write the test in my proposal's comments.. 2012-06-28T10:58:19 did i miss the meeting? 2012-06-28T10:58:27 still in progress 2012-06-28T10:58:29 verm__: more or less. deb went just before 2012-06-28T10:58:30 great 2012-06-28T10:59:00 Although, I've not much experience about code testing, anyway I try my best:) 2012-06-28T10:59:15 zw_yao: look at the tests, see if they are testing the conditions of key delete w/ multithreading, or thread delete with keys 2012-06-28T10:59:30 gedare: ok, i see that i'll add one note: the goal of her midterm is to produce a report based on the various testing frameworks she has been trying out 2012-06-28T10:59:31 if they are not then you will get the opportunity to write such tests, ask questions as needed. 2012-06-28T10:59:52 OK, I will. 2012-06-28T10:59:53 thank you verm__ 2012-06-28T10:59:56 that data will go up on the wiki.. things are still on target for that happening at which point we'll take input from the community on the mailing lists about thoughts/ideas 2012-06-28T11:00:05 great. 2012-06-28T11:00:11 to both of you :) 2012-06-28T11:00:13 zw_yao: anything else? 2012-06-28T11:00:20 I'm done. 2012-06-28T11:00:21 * gedare pokes soh_cah_toa 2012-06-28T11:00:45 soh_cah_toa: are you here? name isn't listed and no status in the gdoc 2012-06-28T11:00:55 i'm here 2012-06-28T11:01:02 i added my status though a while ago 2012-06-28T11:01:06 oh wait, for today? 2012-06-28T11:01:19 yes for today 2012-06-28T11:01:29 ah ok, nvm then 2012-06-28T11:01:32 * gedare cedes the floor to jennifer and soh_cah_toa 2012-06-28T11:01:45 ooc, is joel not here today? 2012-06-28T11:01:54 no joel is in another meeting 2012-06-28T11:02:32 ok 2012-06-28T11:02:56 have there been any pci changes recently? i ask b/c i cloned a new copy of rtems-libbsd and the Makefile is different since there are new files pci_pci.c and pci_bus.c 2012-06-28T11:03:19 that was added a couple of weeks ago 2012-06-28T11:03:33 I am working on another pci patch that will be added shortly 2012-06-28T11:03:55 ok 2012-06-28T11:04:01 it shouldn't be an issue with the arm patch you are working on though 2012-06-28T11:04:09 how is that progressing? 2012-06-28T11:04:12 no 2012-06-28T11:04:16 well like i said, i got a new local copy since i was getting testsuite errors for usb01 on several arm bsp's (the csb's and edb's). fortunately, that seemed to solve most of the problems. except for one, that is 2012-06-28T11:04:23 it seems the link01 can't find “freebsd/bsd.h” that's #include'd by test.c. 2012-06-28T11:04:46 btw, removing in_cksum_arm.S fixed everything from last meeting 2012-06-28T11:05:04 here's what i think: 2012-06-28T11:05:08 did you replace it with the appropriate c file? 2012-06-28T11:05:31 yeah, i left it untouched 2012-06-28T11:05:56 so here's what my config.inc looks like: 2012-06-28T11:06:00 RTEMS_MAKEFILE_PATH = ${HOME}/newbsd/bsp-install/arm-rtems4.11/csb336/ 2012-06-28T11:06:00 INSTALL_BASE = ${HOME}/newbsd/install 2012-06-28T11:06:10 b/c of that, arm-rtems4.11 adds ~/newbsd/install/include to the search path w/ -I which is odd since it doesn't even exist 2012-06-28T11:06:25 “freebsd/bsd.h” is in rtemsbsd/freebsd but since there's so much Makefile magic going on, i have no idea how to add it to the search path 2012-06-28T11:07:49 I can work with you after the meeting and get the bsd.h problem solved... I'm not seeing it so it is probably a path issue 2012-06-28T11:08:04 ok 2012-06-28T11:08:19 besides that, where is the github repo i should push changes to? 2012-06-28T11:08:58 you should fork your own from the RTEMS ones 2012-06-28T11:09:07 and then you can push to your fork from your local clone of your fork 2012-06-28T11:09:15 and let others know about your work on your fork 2012-06-28T11:09:26 (it's weird that work and fork don't rhyme.) 2012-06-28T11:09:59 ah, found it 2012-06-28T11:10:21 supposedly the RTEMS repos there are getting updated although last i heard the updater was broke 2012-06-28T11:10:23 hadn't needed it until now 2012-06-28T11:10:29 but that gives you a place to post your code 2012-06-28T11:10:49 you can also add the git.rtems.org repo as an upstream to pull changes yourself 2012-06-28T11:11:02 into your local clone and then push to your fork 2012-06-28T11:11:13 good idea 2012-06-28T11:11:36 is there some documentation on using the new network stack? 2012-06-28T11:11:55 in rtems-libbsd there is a txt file with info 2012-06-28T11:12:00 on the tree 2012-06-28T11:12:05 libbsd.txt 2012-06-28T11:12:13 soh_cah_toa: did you find that has sufficient details? 2012-06-28T11:12:49 it certainly has a lot of details ;) 2012-06-28T11:12:53 eh, it's confused me on several occasions (don't remember why) but had whatever it was cleared up on the ml 2012-06-28T11:14:06 let us know if it can be made clearer or you have any suggested hints to add 2012-06-28T11:14:21 sure 2012-06-28T11:14:26 just one more question: 2012-06-28T11:14:51 i'm also curious as why libbsd tests (usb01) are in rtems-libbsd? i know they're both from freebsd but aren't they two separate things? 2012-06-28T11:15:05 s/libbsd/libusb/ 2012-06-28T11:15:26 "whatever freebsd usb stack is called" 2012-06-28T11:16:23 rtems-libbsd is to be the rtems version of the libbsd tree ... pieces are being added and it is under development :) 2012-06-28T11:16:24 *** zw_yao has quit IRC (Read error: Connection reset by peer) 2012-06-28T11:17:05 ah, so it's not just for the tcp/ip stack? anything that's ported from freebsd lives there? 2012-06-28T11:17:16 correct 2012-06-28T11:17:22 gotcha, ok 2012-06-28T11:18:02 do you have the gdoc url so i can update it? 2012-06-28T11:18:17 https://docs.google.com/document/d/1RFWt8k3dD6hn_hI9hlxQgS7AD5goZk-3DiGxWVJ-8Kk/edit?pli=1 2012-06-28T11:18:22 ty 2012-06-28T11:18:30 Any other issues for the meeting? 2012-06-28T11:18:47 i guess that's it then 2012-06-28T11:19:03 Any one else thats not on the gdoc? 2012-06-28T11:20:35 Ok ... meeting over then 2012-06-28T11:20:47 Bye 2012-06-28T11:20:55 *** zw_yao has joined #rtems 2012-06-28T11:20:58 soh_cah_toa: jennifer: how much work to use the network stack right now? 2012-06-28T11:21:18 *** ppisa has left #rtems ("Kopete 0.12.7 : http://kopete.kde.org") 2012-06-28T11:21:33 gedare: well, afaik, it can't be "used" just yet 2012-06-28T11:21:43 the sysinit stuff hasn't been fully complete yet 2012-06-28T11:22:07 dunno who's working on that though 2012-06-28T11:22:31 I've got fxp almost up on an i386 2012-06-28T11:22:55 oh wow 2012-06-28T11:23:37 *** claas_ziemke has left #rtems 2012-06-28T11:23:40 i see. what's the hard 'port-specific' work? 2012-06-28T11:23:50 I've got a few more pieces on the pci bus to commit but it should be in within a couple of days 2012-06-28T11:24:13 hmm... 2012-06-28T11:25:18 I keep getting mired down in where to break the freebsd to rtems... 2012-06-28T11:25:22 *** alseh has quit IRC (Remote host closed the connection) 2012-06-28T11:25:33 i would imagine the sysinit stuff. dunno if it's the hardest but i think it's the most critical 2012-06-28T11:26:44 I've got to step out for a minute... I'll be right back 2012-06-28T11:26:47 ok thanks.. i'll have to consider jumping in for the sparc64 support at some point. 2012-06-28T11:26:48 sure 2012-06-28T11:27:04 gedare: ok cool 2012-06-28T11:27:49 *** WikL has quit IRC () 2012-06-28T11:28:32 ok .. back ... sorry 2012-06-28T11:28:51 ok 2012-06-28T11:29:21 I think I'm starting to get a handle on the sysinit stuff at least for the net port 2012-06-28T11:29:25 we've been keen for networking stack upgrade because the old stack did not work at all for us :) 2012-06-28T11:30:51 Most of the other network chips are compiling, but a couple aren't linking at the moment, and I haven't tested any of the others yet 2012-06-28T11:31:36 well, considering all the pieces that each person is working on, i'd like to imagine that you'd be able to have a very minimum "alpha" for i386 or maybe psim ready by the end of the summer 2012-06-28T11:31:37 ok. well i'll try to keep my eyes on it... busy these days so no idea if / when i might get a chance to try it out. 2012-06-28T11:32:21 great :) 2012-06-28T11:32:38 I could use another set of eyes on it 2012-06-28T11:32:57 awesome 2012-06-28T11:34:35 ok so i gotta go have breakfast now. give me no more than an hour to eat and get dressed and everything 2012-06-28T11:35:28 we are hitting a point where we want to discuss on the list what the structure for adding the network tests should be. At the moment I'm leaning toward pulling it all into the rtems-libbsd, but Joel and I were just talking about this yesterday 2012-06-28T11:35:52 k 2012-06-28T11:36:01 ok, bbiab 2012-06-28T11:36:06 I'll go get some lunch and meet back here in an hour or so? 2012-06-28T11:36:12 yup 2012-06-28T11:36:14 tests should be at 'rtems' level though, not the bsd level 2012-06-28T11:36:39 it may be easier to build them directly in rtems-libbsd, but it does not make a lot of sense from an interface coupling level :) 2012-06-28T11:36:59 except those tests that violate api 2012-06-28T11:37:08 my $0.02 2012-06-28T11:37:47 I'm sure it will be beat around some before a solution is decided upon :) 2012-06-28T11:39:28 I need to get lunch... bbib 2012-06-28T11:40:34 *** jennifer is now known as Guest66896 2012-06-28T11:56:26 *** soh_cah_toa has quit IRC (Ping timeout: 246 seconds) 2012-06-28T12:09:06 *** Guest66896 has quit IRC (Quit: Leaving) 2012-06-28T12:09:22 *** jenniferA has joined #rtems 2012-06-28T12:24:43 *** DrJoel has joined #rtems 2012-06-28T12:24:43 *** DrJoel has joined #rtems 2012-06-28T12:24:52 *** ChanServ sets mode: +o DrJoel 2012-06-28T12:32:59 *** cdcs has quit IRC (Quit: Page closed) 2012-06-28T12:38:21 *** soh_cah_toa has joined #rtems 2012-06-28T12:50:43 soh_cah_toa: I will ned to leave for an appointment in a little over an hour so whenever you are ready to work on the path issue let me know. 2012-06-28T12:50:58 yup just booting up my vm now 2012-06-28T12:53:47 RTEMS_MAKEFILE_PATH is where the rtems install occurs, where as INSTALL_BASE is where the libbsd code is installed... after you build the INSTALL_BASE should have the file .../newbsd/install/include/freebsd/bsd.h. 2012-06-28T12:54:24 ok 2012-06-28T12:54:59 I think you may be missing a step or skipping something when you update that keeps causing you this path issue 2012-06-28T12:58:01 well, i built arm csb336 as usual using --prefix=${HOME}/newbsd/bsp-install 2012-06-28T12:58:37 then cloned a new copy of rtems-libbsd and only commented out line +416 to ignore the in_cksum_arm.S file 2012-06-28T12:58:49 you did a make all install, which should have putt the rtems code into the bsp-install directory then 2012-06-28T12:59:12 yeah, it's there 2012-06-28T12:59:20 *** deb has left #rtems ("Leaving") 2012-06-28T12:59:41 it's got: 2012-06-28T12:59:43 after you cloned did you do a reverse script to get the new data into the freebsd directory? 2012-06-28T13:00:11 i don't get it. what new data? 2012-06-28T13:01:10 whatever new that has been added to rtems-libbsd tree since you last cloned may have freebsd code that has additions that need to be added to the freebsd tree 2012-06-28T13:01:31 if you miss this step when you run the script you will overwrite the cloned tree 2012-06-28T13:02:04 i cloned last night again, left than 12 hours ago 2012-06-28T13:02:16 wouldn't just a git pull --rebase do the same thing? 2012-06-28T13:02:20 if there were changes 2012-06-28T13:03:48 the git pull --rebase only pulls down the rtems-libbsd tree, the script is pull stuff from your freebsd tree, modifies it, and re-writes it to the rtems-libbsd tree 2012-06-28T13:04:48 *** ChanServ has quit IRC (shutting down) 2012-06-28T13:05:33 *** ChanServ has joined #rtems 2012-06-28T13:05:33 *** verne.freenode.net sets mode: +o ChanServ 2012-06-28T13:06:50 this is what i don't get about all this. the freebsd "tree" (just whatever's in the freebsd/ directory) is in the rtems-libbsd repo 2012-06-28T13:07:43 The purpose behind what we are doing is to automate changes to the freebsd code so that we can upgrade as freebsd upgrades 2012-06-28T13:07:52 thats why the two trees 2012-06-28T13:08:37 freebsd tree is what we will try to submit to freebsd eventually and will often get out of sync in git with the rtems-libbsd tree which is being worked on 2012-06-28T13:10:06 rtems-libbsd tree is what we will track as what works with rtems at the moment 2012-06-28T13:11:07 and the files aren't identical... if you compare two same named files you will see automated changes that the script makes to the file 2012-06-28T13:13:09 What the reverse script does is put the rtems specific line addtions back into your working version of the freebsd tree 2012-06-28T13:13:26 i don't understand what you mean by "the script is pull stuff from your freebsd tree". my freebsd directory is whatever is in the master rtems-libbsd repo. to pull something means to download something from a remote repo. it's impossible to pull something local 2012-06-28T13:14:10 i always thought the script pulled in whatever changes the freebsd project has made to their code into rtems-libbsd so you can commit it 2012-06-28T13:15:58 Not exactly... it is doing a copy/modify of your working freebsd tree, which is the tree we are tracking source to be submitted back to freebsd in 2012-06-28T13:16:03 so if i make a change to something in the freebsd directory (which i did by deleting in_cksum_arm.S), i have to tell the python script somehow so that it doesn't pull in that file anymore? 2012-06-28T13:16:28 You never make changes directly in the freebsd tree 2012-06-28T13:17:23 You should take the in_cksum_arm.S file out of the .py file and delete it from the rtems_libbsd tree 2012-06-28T13:18:14 the script will not copy the asm file any longer and will modify the makefile not to use it 2012-06-28T13:19:20 what do you mean the rtems_libbsd tree? there's only on in_cksum_arm.S file and it's in the freebsd/ directory which is in the rtems-libbsd repo 2012-06-28T13:19:27 s/on/one/ 2012-06-28T13:21:24 There is a freebsd-8.2 tree that you pulled down that is the original freebsd code 2012-06-28T13:21:41 it is not in the rtems-libbsd repo 2012-06-28T13:22:37 i kinda remember talking about that but don't think i ever needed it. i only removed in_cksum_arm.S from rtems-libbsd 2012-06-28T13:23:14 Joel here. :) 2012-06-28T13:24:06 When we say freebsd tree, we are referring to the FreeBSD source that is SVN revision 222496 2012-06-28T13:24:56 the baseline version of this is in the tarball http://www.rtems.org/ftp/pub/rtems/people/joel/freebsd/ 2012-06-28T13:25:43 The directory rtems-libbsd/freebsd (and rtems-libbsd/contrib) are FreeBSD source code that has been processed by the freebsd-to-rtems.py and are in the "RTEMS TCP/IP Stack" build tree\ 2012-06-28T13:26:07 and that's where the py script pulls in and modies the source from? 2012-06-28T13:26:12 ok 2012-06-28T13:26:15 Yep. 2012-06-28T13:26:33 Jennifer had a picture on her whiteboard at one point showing a flow... is it still there Jennifer? 2012-06-28T13:26:44 If so, I can take a pic with my phone and send it to him 2012-06-28T13:27:10 so yeah, i can't even modify that by mistake since i don't have commit privileges nor do i use svn 2012-06-28T13:27:12 I've not removed anything, but I'm not seeing a pic of it 2012-06-28T13:27:28 ok.. may just be in my head. :) 2012-06-28T13:27:49 correct. And neither do we. It is an anonymous checkout. 2012-06-28T13:28:04 so i can safely remove in_cksum_arm.S from the freebsd/ directory and modify the py script to either 1) no longer pull it in or 2) pull it in but don't build it in the Makefile 2012-06-28T13:28:15 Changes needed to get that to build on RTEMS are only allowed as (1) additions of lines and (2) modifications by the .py script 2012-06-28T13:28:27 right 2012-06-28T13:28:51 I would (1) remove it from git and the Makefile along with any .h files that were starting to be needed 2012-06-28T13:28:55 But doesn't he have to have a local copy of the freebsd to run the script correctly? 2012-06-28T13:29:03 submit that as a patch. Clean it out first 2012-06-28T13:29:48 the py script doesn't know how to just svn checkout revision 222496? 2012-06-28T13:29:51 yes Jennifer 2012-06-28T13:30:03 No. You have to have a local checkout. The script assumes you did that already 2012-06-28T13:30:38 then add code to pull in libbsd/freebsd-8.2/arm/arm/in_cksum.c to the rtems-libbsd build 2012-06-28T13:30:57 svn co http://svn.freebsd.org/base/releng/8.2/sys/ -r222496 freebsd-8.2 according to the libbsd.txt instructions 2012-06-28T13:31:11 You could certainly check that out yourself as a baseline. it should be cool. 2012-06-28T13:31:46 But you will need to run a -R after that to propagate changes we have made to files (and track with git) in rtems-libbsd back to the tree. That's how the system works. 2012-06-28T13:31:53 FreeBSD checkout gives virgin source 2012-06-28T13:32:19 rtems-libbsd has subset of those files with modifications plus rtemsbsd specific implementations as needed 2012-06-28T13:32:26 alright, i'll do that. but imo wouldn't it be easier to just change the py script to perform a svn checkout instead of having to do it manually? 2012-06-28T13:32:51 clone it into a temp directory, modify whatever, the copy it over to rtems-libbsd/freebsd directory, the remove the temp directory 2012-06-28T13:33:05 You don't do it very often.. we did it once for all the work we have done. 2012-06-28T13:33:13 No need to be abusive to the FreeBSD svn repository. 2012-06-28T13:34:09 well, not sure why a simple checkout is abusive but i'll just follow the rules then 2012-06-28T13:34:12 *** QingPei has left #rtems 2012-06-28T13:34:15 how do i tell the py script where the freebsd source is? 2012-06-28T13:34:20 It probably would be good for us to go back to tarball of a clean checkout. The one on the ftp site is actually Sebastian's and has some additions and changes for the usb stack 2012-06-28T13:34:27 oh wait, -f i think 2012-06-28T13:34:40 yeah, it's -f 2012-06-28T13:34:42 ./freebsd-to-rtems.py -? 2012-06-28T13:34:42 freebsd-to-rtems.py [args] 2012-06-28T13:34:42 -?|-h|--help print this and exit 2012-06-28T13:34:42 -d|--dry-run run program but no modifications 2012-06-28T13:34:42 -D|--diff provide diff of files between trees 2012-06-28T13:34:43 -e|--early-exit evaluate arguments, print results, and exit 2012-06-28T13:34:44 -m|--makefile just generate Makefile 2012-06-28T13:34:46 -R|--reverse default FreeBSD -> RTEMS, reverse that 2012-06-28T13:34:48 -r|--rtems RTEMS Libbsd directory 2012-06-28T13:34:50 -f|--freebsd FreeBSD SVN directory 2012-06-28T13:34:52 -v|--verbose enable verbose output mode 2012-06-28T13:35:21 After a git update or fresh checkout, you want to run a "-R" to move the preferred RTEMS changes back to the FreeBSD source tree. 2012-06-28T13:35:52 After that, you move forward to add new source. When you have modified the new source to compile, you move it back with the -R 2012-06-28T13:36:17 back to the freebsd svn checkout? i thought touching that was off limits? 2012-06-28T13:36:24 At all times, the "builds for RTEMS" version of FreeBSD code is in rtems-libbsd. But the original is in the freebsd-8.2 directory 2012-06-28T13:36:33 right 2012-06-28T13:37:00 committing to them is off-limits. We can hack on that directory all we want. Eventually we will have a working stack and will then begin to reduce the diff and deal with getting it fed upstream 2012-06-28T13:37:46 You should never manually change anything in the FreeBSD source tree though, only let the script do that 2012-06-28T13:37:57 +1 on Jennifer 2012-06-28T13:38:33 so then the freebsd tarball isn't an exact r222496 if the py script changes it? 2012-06-28T13:39:13 the script changes your working copy, it doesn't submit anything 2012-06-28T13:39:17 no.. but you can diff it back to that if you like.. or check your own out 2012-06-28T13:39:56 but then all out freebsd trees will be out of sync 2012-06-28T13:40:11 i'll be pushing changes i made to rtems-libbsd that will only work on my machine 2012-06-28T13:40:23 s/out/our/ 2012-06-28T13:40:57 That is why you always run the -R version of the script after a pull 2012-06-28T13:41:40 There are two views of the world ... (1) "works with RTEMS" which is what we care most about 2012-06-28T13:41:57 (2) how much did we change FreeBSD r222496 2012-06-28T13:42:14 Right now we care more about (1) and that's always our primary focus 2012-06-28T13:42:37 (2) will matter when we have a workign TCP/IP stack and begin to wade through the changes to get them fed (where acceptable) upstream 2012-06-28T13:48:22 ay yi yi...so if i removed in_cksum_arm.S from the freebsd/ directory in rtems-libbsd and modified the Makefile so that it won't add it to CFILES, my next step is this: 2012-06-28T13:48:38 first checkout the freebsd r222496 revision, then do a reverse -R operation w/ the py script pointing it to the directory where i just checked out freebsd, then modify the script so that it automatically generates a Makefile that doesn't add in_cksum_arm.S to CFILES? 2012-06-28T13:49:39 the script modifys the CFILE parameter in the makefile... it generates the makefile 2012-06-28T13:50:10 You never have to modify the makefile at all 2012-06-28T13:50:31 yeah, i know. i just did that so i know what changes i need to make to the py script 2012-06-28T13:51:09 All you have to do is remove the one line reference to in_cksum_arm.S 2012-06-28T13:51:12 in the py file 2012-06-28T13:51:18 yeah 2012-06-28T13:52:00 correct 2012-06-28T13:52:00 Remove all references to the in_cksum_asm.S file first. 2012-06-28T13:52:38 That would be a good patch to submit and let one of us verify.. guessing me since I have more BSPs built at hand. :) 2012-06-28T13:52:46 k 2012-06-28T13:57:27 just to double check, i should fork rtems/rtems-libbsd.git, add soh-cah-toa/rtems-libbsd.git as a push url, push my changes there, then submit that work as a pull request right? 2012-06-28T14:00:40 *** alexs_ has joined #rtems 2012-06-28T14:00:41 correct ... 2012-06-28T14:01:13 *** alexs_ is now known as alseh 2012-06-28T14:02:16 I've got to run... but Joel should be around to answer questions this afternoon... If you have problems after he leaves email me and I'll check it once I get home 2012-06-28T14:02:27 ok 2012-06-28T14:02:47 *** jenniferA has quit IRC (Quit: Leaving) 2012-06-28T14:49:11 DrJoel: when running the py script w/ -R and -f pointing to the freebsd tree, it complains that the rtems directory was not found. however the usage message refers to the rtems libbsd directory, not the rtems directory. do i just use `-r .`? it really doesn't understand that the current working directory actually is the rtems-libbsd directory? 2012-06-28T15:10:08 no .. the -r needs a full path. 2012-06-28T15:10:29 The -r refers to the rtems-libbsd directory 2012-06-28T15:11:37 ./freebsd-to-rtems.py -R -f /home/joel/newbsd/libbsd/freebsd-8.2/ -r /home/joel/newbsd/git/rtems-libbsd 2012-06-28T15:11:43 for me after a git update 2012-06-28T15:11:53 b/c i did -r . and it seemed to work fine 2012-06-28T15:12:08 accident more than design :) 2012-06-28T15:14:10 would it be worth adding the cwd as a default case for RTEMS_DIR? after all, it's run as ./freebsd-to-rtems.py and is more intuitive 2012-06-28T15:21:13 This isn't a command to run lightly. 2012-06-28T15:21:25 Assuming anything like cwd is bad to me 2012-06-28T15:22:55 is there a reason someone would run it anywhere outside of the cwd though? 2012-06-28T15:23:19 I sometimes did when I was working on it. I had two trees with different changes. :) 2012-06-28T15:23:47 ah, yeah that's practical 2012-06-28T15:26:14 i also noticed the lines starting at +107: 2012-06-28T15:26:17 elif o in ("-r", "--rtems"): 2012-06-28T15:26:18 RTEMS_DIR = a 2012-06-28T15:26:24 appear twice right next to each other. could that just a copy/paste error perhaps or does it actually serve some not readily apparent purpose? 2012-06-28T15:38:01 that's a typo. :( 2012-06-28T15:38:07 Is an option on the list not processed 2012-06-28T15:38:42 It's a cut and paste. :( 2012-06-28T15:38:51 i figured as much 2012-06-28T15:38:59 and --verbose is not listed as a valid option. 2012-06-28T15:39:07 You want to submit a patch for that one? 2012-06-28T15:39:23 yeah, i'll commit to my fork 2012-06-28T15:39:37 * DrJoel just ordered a Nexus 7 2012-06-28T15:42:00 neat, those tablets are incredible 2012-06-28T15:44:21 I sprung for the 16GB version and a case. Seemed not worth it otherwise. 2012-06-28T15:44:45 I use my G2 like a tablet anyway so this will be easier on me at home. 2012-06-28T15:50:14 yeah, i always go for the extra capacity. i think i'll never need it but sure enough, it fills up in no time 2012-06-28T16:10:21 since you can't add it later .. seems best to max it out now 2012-06-28T16:18:47 * DrJoel waves bye.. still under the weather a bit and heading home 2012-06-28T16:18:51 *** DrJoel has quit IRC (Quit: Leaving) 2012-06-28T18:12:55 *** alseh has quit IRC (Ping timeout: 252 seconds) 2012-06-28T18:29:22 *** gedare has quit IRC (Quit: Leaving) 2012-06-28T20:17:33 *** xiangfu has joined #rtems 2012-06-28T20:43:22 *** kristianpaul has quit IRC (Quit: leaving) 2012-06-28T20:46:46 *** kristianpaul has joined #rtems 2012-06-28T20:46:46 *** kristianpaul has joined #rtems 2012-06-28T22:19:10 *** QingPei has joined #rtems 2012-06-29T00:18:21 *** arvind_khadri has joined #rtems 2012-06-29T00:44:59 *** QingPei has left #rtems 2012-06-29T03:11:08 *** lcpfnvcy has joined #rtems 2012-06-29T05:06:17 *** zw_yao has quit IRC (Ping timeout: 255 seconds) 2012-06-29T05:21:33 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-06-29T05:40:15 *** arvind_khadri has joined #rtems 2012-06-29T06:58:29 *** Hesham has joined #rtems 2012-06-29T08:17:49 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-06-29T08:23:27 *** kristianpaul has joined #rtems 2012-06-29T08:37:57 *** xiangfu has quit IRC (Ping timeout: 248 seconds) 2012-06-29T09:18:08 *** QingPei has joined #rtems 2012-06-29T09:44:39 *** gedare has joined #rtems 2012-06-29T09:45:22 Hesham: i'm about to go home.. will you be around in about 2 hours? 2012-06-29T09:47:08 meanwhile... i saw your debugging output. one other thing you might do is to make a table (in a spreadsheet) containing the values of the variables in the function and how they change 2012-06-29T09:52:56 *** gedare has quit IRC (Quit: Ex-Chat) 2012-06-29T10:10:59 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2012-06-29T11:23:42 . 2012-06-29T11:49:25 *** neo1987___ has joined #rtems 2012-06-29T11:53:17 *** neo1987___ has joined #rtems 2012-06-29T13:02:51 *** soh_cah_toa has quit IRC (Ping timeout: 246 seconds) 2012-06-29T13:24:01 *** gedare has joined #rtems 2012-06-29T13:24:29 hi Hesham 2012-06-29T13:24:54 hi gedare 2012-06-29T13:25:17 i saw your email.. find out what the values in MSR means 2012-06-29T13:25:27 sounds like you're chasing down the right track though 2012-06-29T13:26:17 aanjhan told me to find MSR[RI] 2012-06-29T13:26:24 it's bit 26 2012-06-29T13:26:58 but MSR at info register is 16 bit 2012-06-29T13:27:37 interesting.... 2012-06-29T13:27:47 print out the value stored in the 'msr' variable during that function 2012-06-29T13:27:57 is the function storing the msr incorrectly maybe? 2012-06-29T13:27:58 i had sent all of this too aanjhan waiting his reply 2012-06-29T13:29:04 i have already created a watchpoint at msr 2012-06-29T13:29:15 ok 2012-06-29T13:29:17 but i do not know what should be the "correct value" 2012-06-29T13:29:41 heh find the powerpc manual and look for the definition of the msr register:) 2012-06-29T13:31:19 you might try moving the write_msr to after the clrbats otherwise i don't know 2012-06-29T13:31:25 working on anything else? 2012-06-29T13:31:41 I have it right now, that how i found that MSR[RI] is bit 26 2012-06-29T13:31:49 well 2012-06-29T13:32:00 i think it will not cause the problem 2012-06-29T13:32:24 at quanming code there were no CLRBAT after write_msr 2012-06-29T13:33:22 i see 2012-06-29T13:33:49 i thought that was causing the problem ....... 2012-06-29T13:34:05 Privileged instruction—A privileged instruction type program exception is generated when the execution of a privileged instruction is attempted and the MSR user privilege bit, MSR[PR], is set. This exception is also generated for mtspr or mfsprwith an invalid SPR field if spr[0] = 1 and MSR[PR] = 1. 2012-06-29T13:34:52 but as i told you MSR[PR] is bit 17 and info registers shows only 16 bit for MSR 2012-06-29T13:35:27 are you referring to the correct powerpc manual? 2012-06-29T13:36:04 well you would assume the upper bits are 0 2012-06-29T13:37:16 it's not entirely manual, it's part of it http://web.eecs.umich.edu/~stever/373/ch6-int.pdf 2012-06-29T13:38:24 i found 2012-06-29T13:38:32 http://www.freescale.com/files/product/doc/MPCFPE32B.pdf 2012-06-29T13:38:56 that's probably close to the proper one... not exactly right still 2012-06-29T13:38:59 i got it not :) 2012-06-29T13:39:04 you can see section 2.3.1 2012-06-29T13:39:08 Machine State Register (MSR) 2012-06-29T13:39:16 this one is at Final report for quanming 2012-06-29T13:39:19 the endianness of the register may matter.... 2012-06-29T13:40:55 the figure i'm looking at shows that bits 0-11 are unused... but has them at the little-endian format 2012-06-29T13:41:14 you are right 2012-06-29T13:41:24 i found that 2012-06-29T13:41:28 ok 2012-06-29T13:41:29 i will check again now 2012-06-29T13:41:33 ok 2012-06-29T13:43:38 how could i view spr register ? 2012-06-29T13:43:55 it's not listed when i write info registers ! 2012-06-29T13:46:41 gedare: one thing else, how could i build rtems for sparc64 ? " the configuration params " 2012-06-29T13:51:53 um, just target=sparc64-rtems4.11 and bsp=usiii (or niagara) 2012-06-29T13:52:05 but i don't know that it can be executed anywhere useful right now 2012-06-29T13:53:48 Should i use it for stubs or changing to another target ? 2012-06-29T13:54:18 *** QingPei has left #rtems 2012-06-29T13:54:50 use sparc for stubs.... 2012-06-29T13:54:57 sparc64 != sparc 2012-06-29T13:55:10 to build for sparc you do target=sparc-rtems4.11 and rtems-bsp=sis 2012-06-29T13:55:45 aha 2012-06-29T13:56:18 but where should memoryprotection_manager.c ( including stubs ) should reside at sparct ? 2012-06-29T13:56:22 sparc ** 2012-06-29T14:07:15 well i think the stubs already exist... you just have to specify to use the "no_memoryprotection.c" file in c/src/lib/libcpu/sparc/Makefile.am 2012-06-29T14:11:39 i will continue implementing Arena API, Any other issues to take care of ? 2012-06-29T14:12:09 none i can think of 2012-06-29T14:12:31 you have your code on github now? 2012-06-29T14:12:42 with usable branches? did you go learn about how to use git? 2012-06-29T14:12:53 yeah i have pushed it two day ago 2012-06-29T14:13:01 both versions 2012-06-29T14:13:18 you can check them 2012-06-29T14:14:07 i just did not deal with branches before, i read about it and understood how it works 2012-06-29T14:17:18 ok 2012-06-29T14:17:26 i hope to get some time this weekend to try 2012-06-29T14:21:39 later hes 2012-06-29T14:21:40 Hesham, 2012-06-29T14:21:49 yeah here 2012-06-29T14:21:55 i'm taking off 2012-06-29T14:22:08 *** gedare has quit IRC (Quit: Ex-Chat) 2012-06-29T14:22:08 ok thanks gedare 2012-06-29T14:45:17 *** Hesham has quit IRC (Quit: Leaving.) 2012-06-29T21:31:28 *** mikeg has quit IRC (Ping timeout: 244 seconds) 2012-06-29T23:20:43 *** arvind_khadri has joined #rtems 2012-06-30T00:03:33 *** deb has joined #rtems 2012-06-30T00:06:40 *** deb has quit IRC (Client Quit) 2012-06-30T00:20:58 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2012-06-30T00:29:55 *** arvind_khadri has joined #rtems 2012-06-30T00:51:02 *** alseh has joined #rtems 2012-06-30T01:04:38 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-06-30T01:08:58 *** arvind_khadri has joined #rtems 2012-06-30T01:34:20 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2012-06-30T01:40:17 *** peerst has joined #rtems 2012-06-30T09:06:07 *** deb has joined #rtems 2012-06-30T09:21:03 *** kristianpaul has quit IRC (Quit: leaving) 2012-06-30T09:22:39 *** kristianpaul has joined #rtems 2012-06-30T09:22:39 *** kristianpaul has joined #rtems 2012-06-30T09:43:53 *** kristianpaul has quit IRC (Quit: Reconnecting) 2012-06-30T09:44:02 *** kristianpaul has joined #rtems 2012-06-30T09:44:02 *** kristianpaul has joined #rtems 2012-06-30T10:22:04 *** kristianpaul has quit IRC (Remote host closed the connection) 2012-06-30T10:22:20 *** kristianpaul has joined #rtems 2012-06-30T10:34:40 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-06-30T10:37:37 *** kristianpaul has joined #rtems 2012-06-30T10:37:37 *** kristianpaul has joined #rtems 2012-06-30T10:59:24 *** mikeg has joined #rtems 2012-06-30T11:05:09 *** alseh has quit IRC (Remote host closed the connection) 2012-06-30T14:17:23 *** Hesham has joined #rtems 2012-06-30T14:29:11 *** dr__house has quit IRC (Remote host closed the connection) 2012-06-30T15:31:51 *** panzon has quit IRC (Ping timeout: 246 seconds) 2012-06-30T15:44:52 *** panzon has joined #rtems 2012-06-30T16:06:30 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-06-30T22:06:53 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-06-30T22:45:33 *** kristianpaul has joined #rtems 2012-06-30T22:45:33 *** kristianpaul has joined #rtems 2012-06-30T22:50:54 *** kristianpaul has quit IRC (Quit: leaving) 2012-06-30T22:56:48 *** kristianpaul has joined #rtems 2012-06-30T22:56:48 *** kristianpaul has joined #rtems 2012-06-30T23:38:26 *** kristianpaul has quit IRC (Quit: leaving) 2012-06-30T23:38:55 *** kristianpaul has joined #rtems 2012-06-30T23:38:55 *** kristianpaul has joined #rtems 2012-07-01T00:00:33 *** Hesham has joined #rtems 2012-07-01T00:03:12 *** deb has quit IRC (Quit: Leaving) 2012-07-01T00:29:47 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-07-01T05:11:08 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-01T05:12:15 *** kristianpaul has joined #rtems 2012-07-01T05:37:43 *** kristianpaul has quit IRC (Remote host closed the connection) 2012-07-01T05:37:56 *** kristianpaul has joined #rtems 2012-07-01T05:47:56 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-01T05:48:58 *** kristianpaul has joined #rtems 2012-07-01T05:55:56 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-01T05:56:52 *** kristianpaul has joined #rtems 2012-07-01T06:08:44 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-01T06:25:57 *** kristianpaul has joined #rtems 2012-07-01T06:35:48 *** weiY has joined #rtems 2012-07-01T06:39:40 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-01T06:48:55 *** kristianpaul has joined #rtems 2012-07-01T06:48:55 *** kristianpaul has joined #rtems 2012-07-01T07:06:52 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-07-01T07:08:07 *** kristianpaul has joined #rtems 2012-07-01T07:19:21 *** kristianpaul has quit IRC (Remote host closed the connection) 2012-07-01T07:19:37 *** kristianpaul has joined #rtems 2012-07-01T07:19:37 *** kristianpaul has joined #rtems 2012-07-01T07:38:34 *** QingPei has joined #rtems 2012-07-01T07:38:47 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-07-01T07:44:04 *** kristianpaul has joined #rtems 2012-07-01T07:44:04 *** kristianpaul has joined #rtems 2012-07-01T08:23:07 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-07-01T08:23:56 *** kristianpaul has joined #rtems 2012-07-01T08:35:09 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-07-01T08:51:23 *** kristianpaul has joined #rtems 2012-07-01T08:55:44 *** kristianpaul has quit IRC (Client Quit) 2012-07-01T08:56:42 *** kristianpaul has joined #rtems 2012-07-01T08:56:42 *** kristianpaul has joined #rtems 2012-07-01T09:02:18 *** kristianpaul has quit IRC (Quit: leaving) 2012-07-01T09:03:08 *** kristianpaul has joined #rtems 2012-07-01T09:03:08 *** kristianpaul has joined #rtems 2012-07-01T09:10:17 *** weiY has quit IRC () 2012-07-01T10:51:38 *** Hesham has joined #rtems 2012-07-01T11:10:23 *** deb has joined #rtems 2012-07-01T12:41:33 *** kiwichris__ has quit IRC (Ping timeout: 255 seconds) 2012-07-01T13:54:45 *** QingPei has left #rtems 2012-07-01T15:13:27 *** panzon has quit IRC (Ping timeout: 246 seconds) 2012-07-01T15:15:59 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-07-01T15:26:04 *** panzon has joined #rtems 2012-07-01T15:31:24 *** Hesham has joined #rtems 2012-07-01T15:40:04 *** Hesham1 has joined #rtems 2012-07-01T15:42:14 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-07-01T16:23:40 *** deb has quit IRC (Quit: This computer has gone to sleep) 2012-07-01T17:03:47 *** Hesham1 has quit IRC (Ping timeout: 246 seconds) 2012-07-01T17:43:34 *** kiwichris has joined #rtems 2012-07-01T19:03:17 *** deb has joined #rtems 2012-07-01T19:51:50 *** xiangfu has joined #rtems 2012-07-01T20:39:20 *** ChanServ has quit IRC (*.net *.split) 2012-07-01T21:10:28 *** arvind_khadri has joined #rtems 2012-07-01T21:36:45 *** deb has left #rtems ("Leaving") 2012-07-01T22:00:54 *** ChanServ has joined #rtems 2012-07-01T22:00:54 *** verne.freenode.net sets mode: +o ChanServ 2012-07-01T23:17:38 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2012-07-01T23:37:39 *** dr__house has joined #rtems 2012-07-01T23:37:39 *** dr__house has joined #rtems 2012-07-01T23:51:22 *** Hesham has joined #rtems