2012-07-30T00:08:28 *** kiwichris has joined #rtems 2012-07-30T00:08:29 *** xiangfu has joined #rtems 2012-07-30T00:14:17 *** xiangfu has quit IRC (Ping timeout: 272 seconds) 2012-07-30T00:38:36 *** xiangfu has joined #rtems 2012-07-30T00:43:25 *** xiangfu has quit IRC (Ping timeout: 272 seconds) 2012-07-30T00:58:14 *** xiangfu has joined #rtems 2012-07-30T01:02:22 *** arvind_khadri has joined #rtems 2012-07-30T01:02:59 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-07-30T01:03:03 *** xiangfu has quit IRC (Ping timeout: 272 seconds) 2012-07-30T01:03:51 *** arvind_khadri has joined #rtems 2012-07-30T01:04:42 *** arvind_khadri has joined #rtems 2012-07-30T01:32:59 *** sebhub has joined #rtems 2012-07-30T01:38:25 good morning 2012-07-30T02:27:40 sebhub, hi 2012-07-30T02:27:50 what format is the arm obj file ? 2012-07-30T02:28:01 ELF 2012-07-30T02:28:22 elf32-littlearm or elf32-bigarm, elf32-little or elf32-big ? 2012-07-30T02:28:35 see objdump --help 2012-07-30T02:29:40 arm-rtems4.11-objcopy --info 2012-07-30T02:30:24 normally elf32-littlearm 2012-07-30T02:30:34 ok 2012-07-30T02:30:36 thanks 2012-07-30T02:30:38 i would not use arm-rtems4.11 2012-07-30T02:30:49 use arm-rtemseabi4.11 instead 2012-07-30T02:31:10 for the binutils it doesn't matter, but the gcc is different 2012-07-30T02:31:11 It is an EABI toolset 2012-07-30T02:31:39 I do not use i386-rtemsia32-gcc 2012-07-30T02:31:49 I do not use i386-rtemsia324.11-gcc 2012-07-30T02:32:46 I have a relocator for arm building so I hope you can play and test for me :) 2012-07-30T02:33:04 i am not sure what the difference between elf32-littlearm and elf32-little is 2012-07-30T02:33:26 Same 2012-07-30T02:43:04 What does this mean ... 2012-07-30T02:43:07 ? 2012-07-30T02:43:13 /Users/chris/Development/rtems/build/4.11/arm-rtems4.11/lpc2362/lib/linkcmds:69 cannot move location counter backwards (from 000000004000a85c to 0000000040008000) 2012-07-30T02:43:44 this bsp has a very limited rom size 2012-07-30T02:43:48 your code is to big 2012-07-30T02:43:55 :( 2012-07-30T02:44:07 No RTL loader for this device. 2012-07-30T02:44:13 What should I test ? 2012-07-30T02:44:28 lpc32xx_mzx 2012-07-30T02:45:04 Thanks 2012-07-30T02:45:06 did you chose the lpc2362 at random? 2012-07-30T02:45:51 Yeap. The waf build tool picks up all installed bsps 2012-07-30T03:04:12 I do not have a pc32xx_mzx cause all the fileio sample does not build and the build breaks 2012-07-30T03:04:29 lpc32xx_mzx 2012-07-30T03:06:19 Yeah that one. A BSP before that one does not build the fileio (I think) 2012-07-30T03:06:34 Just playing with --enable-tests=no 2012-07-30T03:06:43 probably the lpc32xx_mzx_stage_1 2012-07-30T03:06:53 No sure. Wiped the output. 2012-07-30T03:07:16 If it is this hard for me to get this stuff to build I do feel for our new users 2012-07-30T03:07:57 the problem is that not all tests build for every bsp due to code size limitations 2012-07-30T03:09:07 Yeap and the small ram hack has out lived its welcome 2012-07-30T03:09:21 It is lpc2362 2012-07-30T03:10:21 need to wipe the build tree cause changing the options even with maintainer mode does not refresh/rebuld the 'make' configure phases. 2012-07-30T03:50:22 *** xiangfu has joined #rtems 2012-07-30T06:10:09 *** lcpfnvcy has quit IRC (Read error: Connection reset by peer) 2012-07-30T06:20:44 *** mumptai_ has joined #rtems 2012-07-30T06:20:46 hi 2012-07-30T06:25:46 *** lcpfnvcy has joined #rtems 2012-07-30T07:02:28 *** alseh has joined #rtems 2012-07-30T07:05:20 *** weiY has quit IRC (Ping timeout: 240 seconds) 2012-07-30T07:05:39 *** alseh_ has joined #rtems 2012-07-30T07:06:52 *** alseh has quit IRC (Ping timeout: 246 seconds) 2012-07-30T07:13:27 *** weiY has joined #rtems 2012-07-30T08:03:35 *** jahf has quit IRC (Ping timeout: 248 seconds) 2012-07-30T08:04:45 *** jahf has joined #rtems 2012-07-30T08:30:09 *** weiY has quit IRC (Ping timeout: 246 seconds) 2012-07-30T08:30:31 *** weiY has joined #rtems 2012-07-30T08:37:17 *** alseh_ has quit IRC (Read error: Operation timed out) 2012-07-30T08:42:44 *** mumptai_ has quit IRC (Ping timeout: 248 seconds) 2012-07-30T09:11:58 *** antgreen has joined #rtems 2012-07-30T09:33:05 *** alseh has joined #rtems 2012-07-30T10:25:27 *** arvind_khadri has quit IRC (Ping timeout: 272 seconds) 2012-07-30T10:31:39 *** weiY has quit IRC (Ping timeout: 240 seconds) 2012-07-30T10:40:45 *** QingPei has joined #rtems 2012-07-30T11:05:10 *** sebhub has quit IRC (Remote host closed the connection) 2012-07-30T11:34:02 *** deb has joined #rtems 2012-07-30T12:36:24 *** xiangfu has quit IRC (Ping timeout: 252 seconds) 2012-07-30T12:43:57 *** QingPei has left #rtems 2012-07-30T12:48:19 *** xiangfu has joined #rtems 2012-07-30T12:51:31 *** xiangfu has quit IRC (Client Quit) 2012-07-30T14:33:52 *** WikL has joined #rtems 2012-07-30T15:05:56 *** alseh has quit IRC (Read error: Operation timed out) 2012-07-30T15:21:04 *** alseh has joined #rtems 2012-07-30T15:32:42 *** antgreen has quit IRC (Remote host closed the connection) 2012-07-30T16:00:27 *** alseh has quit IRC (Remote host closed the connection) 2012-07-30T17:37:59 *** WikL has quit IRC (Ping timeout: 240 seconds) 2012-07-30T17:40:05 *** WikL_ has joined #rtems 2012-07-30T18:01:57 *** WikL_ has quit IRC (Remote host closed the connection) 2012-07-30T21:09:34 *** xiangfu has joined #rtems 2012-07-30T21:53:02 *** xian9fu has joined #rtems 2012-07-30T21:56:48 *** xiangfu has quit IRC (Ping timeout: 246 seconds) 2012-07-31T01:32:21 *** sebhub has joined #rtems 2012-07-31T01:32:28 good morning 2012-07-31T01:35:38 *** arvind_khadri has joined #rtems 2012-07-31T01:38:50 *** arvind_khadri has quit IRC (Client Quit) 2012-07-31T01:43:40 *** arvind_khadri has joined #rtems 2012-07-31T01:55:12 sebhub, hi 2012-07-31T01:56:50 hi 2012-07-31T02:01:36 I have pushed arm changes to the RTL project. It should build all arm bsps it can. If you look in the top of wscript there is a list of those I have filtered out 2012-07-31T02:01:54 Can you please pull and see if it builds for you ? 2012-07-31T02:04:14 for powerpc i get this: /home/sh/git-rtems-rtl/libbsd/include/sys/cdefs.h:59:27: fatal error: machine/cdefs.h: No such file or directory 2012-07-31T02:04:25 ~/Downloads/waf-1.7.0/waf configure build --rtems-archs=powerpc 2012-07-31T02:16:22 *** mikeg has quit IRC (Quit: Leaving) 2012-07-31T02:18:18 *** arvind_khadri has joined #rtems 2012-07-31T02:29:43 hehe, looks like powerpc is also needed. 2012-07-31T02:29:52 how did arm go ? 2012-07-31T02:30:29 what is the command line for arm? does it work with arm-rtemseabi4.11? 2012-07-31T02:31:04 If you use just --rtems does it break on powerpc ? 2012-07-31T02:32:02 waf: error: --rtems option requires an argument 2012-07-31T02:32:56 the path to your installed RTEMS 2012-07-31T02:33:08 not the built RTEMS, but the installed RTEMS 2012-07-31T02:33:39 --rtems-archs=arm works for me 2012-07-31T02:34:02 therefore 'waf --rtems= --rtems-archs=arm' 2012-07-31T02:34:12 ~/Downloads/waf-1.7.0/waf configure build --rtems-archs=arm 2012-07-31T02:34:17 and you can --rtems-tools=path1,path2 2012-07-31T02:34:18 this still builds the powerpc 2012-07-31T02:34:31 i have everything installed in the standard prefix 2012-07-31T02:34:34 not build 2012-07-31T02:34:38 just the configure 2012-07-31T02:34:42 no build that is 2012-07-31T02:34:57 waf configure --rtems-archs=arm 2012-07-31T02:35:50 ~/Downloads/waf-1.7.0/waf configure --rtems-archs=powerpc 2012-07-31T02:35:53 seems to work 2012-07-31T02:36:03 ~/Downloads/waf-1.7.0/waf configure --rtems-archs=arm 2012-07-31T02:36:04 Ah ok 2012-07-31T02:36:04 Setting top to : /home/sh/git-rtems-rtl 2012-07-31T02:36:06 Setting out to : /home/sh/git-rtems-rtl/build 2012-07-31T02:36:07 Could not find any architectures 2012-07-31T02:36:09 (complete log in /home/sh/git-rtems-rtl/build/config.log) 2012-07-31T02:36:25 You need to add --rtems= to the installed bsps 2012-07-31T02:37:01 waf --rtems=/home/sh/rtems/4.11 --rtems-archs=powerpc,arm 2012-07-31T02:37:02 the powerpc bsps are detected without this option 2012-07-31T02:37:22 The --rtems-archs or --rtems ? 2012-07-31T02:37:48 ~/Downloads/waf-1.7.0/waf configure --rtems-archs=powerpc 2012-07-31T02:37:51 Are the tools in your path ? 2012-07-31T02:37:54 yes 2012-07-31T02:38:28 sparc works also 2012-07-31T02:39:09 sparc build works also 2012-07-31T02:39:14 but powerpc fails 2012-07-31T02:39:22 I get "Could not find any architectures" if I do not have any powerpc bsps installed 2012-07-31T02:39:48 What happens if you just 'waf configure' does that fail ? 2012-07-31T02:39:49 does it cope with the rtemseabi special case? 2012-07-31T02:40:02 The could be correct. 2012-07-31T02:40:22 I have not butchered the script to handle that case. I suppose I should 2012-07-31T02:41:02 I am not sure what works and does not work. 2012-07-31T02:41:23 this special case has to go away, but currently i don't have the time to push this forward 2012-07-31T02:41:49 Understood. I have just symlinked to arm-rtems4.11-gcc :) 2012-07-31T02:42:15 Waf: Entering directory `/home/sh/git-rtems-rtl/build/powerpc-rtems4.11-mpc5674fevb_spe' 2012-07-31T02:42:16 [ 2/27] c: dlfcn.c -> build/powerpc-rtems4.11-mpc5674fevb_spe/dlfcn.c.1.o 2012-07-31T02:42:18 [ 2/27] c: dlfcn-shell.c -> build/powerpc-rtems4.11-mpc5674fevb_spe/dlfcn-shell.c.1.o 2012-07-31T02:42:19 [ 3/27] c: rtl.c -> build/powerpc-rtems4.11-mpc5674fevb_spe/rtl.c.1.o 2012-07-31T02:42:21 [ 4/27] c: rtl-alloc-heap.c -> build/powerpc-rtems4.11-mpc5674fevb_spe/rtl-alloc-heap.c.1.o 2012-07-31T02:42:22 [ 5/27] c: rtl-allocator.c -> build/powerpc-rtems4.11-mpc5674fevb_spe/rtl-allocator.c.1.o 2012-07-31T02:42:24 [ 6/27] c: rtl-chain-iterator.c -> build/powerpc-rtems4.11-mpc5674fevb_spe/rtl-chain-iterator.c.1.o 2012-07-31T02:42:25 In file included from /home/sh/git-rtems-rtl/libbsd/include/dlfcn.h:36:0, 2012-07-31T02:42:27 from ../../dlfcn.c:19: 2012-07-31T02:42:28 /home/sh/git-rtems-rtl/libbsd/include/sys/cdefs.h:59:27: fatal error: machine/cdefs.h: No such file or directory 2012-07-31T02:42:30 compilation terminated. 2012-07-31T02:42:31 In file included from /opt/rtems-4.11/lib/gcc/powerpc-rtems4.11/4.7.1/../../../../powerpc-rtems4.11/include/string.h:12:0, 2012-07-31T02:42:33 from ../../dlfcn-shell.c:25: 2012-07-31T02:42:34 /home/sh/git-rtems-rtl/libbsd/include/sys/cdefs.h:59:27: fatal error: machine/cdefs.h: No such file or directory 2012-07-31T02:42:36 compilation terminated. 2012-07-31T02:42:37 In file included from /opt/rtems-4.11/lib/gcc/powerpc-rtems4.11/4.7.1/../../../../powerpc-rtems4.11/include/string.h:12:0, 2012-07-31T02:42:39 from ../../rtl.c:24: 2012-07-31T02:42:40 /home/sh/git-rtems-rtl/libbsd/include/sys/cdefs.h:59:27: fatal error: machine/cdefs.h: No such file or directory 2012-07-31T02:42:42 compilation terminated. 2012-07-31T02:42:43 In file included from /opt/rtems-4.11/lib/gcc/powerpc-rtems4.11/4.7.1/../../../../powerpc-rtems4.11/include/string.h:12:0, 2012-07-31T02:42:45 from /opt/rtems-4.11/powerpc-rtems4.11/mpc5674fevb_spe/lib/include/rtems/score/cpu.h:41, 2012-07-31T02:42:46 from /opt/rtems-4.11/powerpc-rtems4.11/mpc5674fevb_spe/lib/include/rtems/score/percpu.h:20, 2012-07-31T02:42:48 from /opt/rtems-4.11/powerpc-rtems4.11/mpc5674fevb_spe/lib/include/rtems/system.h:21, 2012-07-31T02:42:49 from /opt/rtems-4.11/powerpc-rtems4.11/mpc5674fevb_spe/lib/include/rtems/chain.h:23, 2012-07-31T02:42:51 from /home/sh/git-rtems-rtl/rtl-chain-iterator.h:23, 2012-07-31T02:42:52 Yeah I need to add support for powerpc. It does not relocation and machine dependent support. 2012-07-31T02:42:52 from ../../rtl-chain-iterator.c:22: 2012-07-31T02:42:54 /home/sh/git-rtems-rtl/libbsd/include/sys/cdefs.h:59:27: fatal error: machine/cdefs.h: No such file or directory 2012-07-31T02:42:55 compilation terminated. 2012-07-31T02:42:57 In file included from /home/sh/git-rtems-rtl/libbsd/include/link.h:40:0, 2012-07-31T02:42:58 from /home/sh/git-rtems-rtl/rtl.h:21, 2012-07-31T02:43:00 from ../../rtl-allocator.c:18: 2012-07-31T02:43:01 /home/sh/git-rtems-rtl/libbsd/include/link_elf.h:12:33: fatal error: machine/elf_machdep.h: No such file or directory 2012-07-31T02:43:03 compilation terminated. 2012-07-31T02:43:04 In file included from /opt/rtems-4.11/lib/gcc/powerpc-rtems4.11/4.7.1/../../../../powerpc-rtems4.11/include/string.h:12:0, 2012-07-31T02:43:06 from /opt/rtems-4.11/powerpc-rtems4.11/mpc5674fevb_spe/lib/include/rtems/score/cpu.h:41, 2012-07-31T02:43:07 from /opt/rtems-4.11/powerpc-rtems4.11/mpc5674fevb_spe/lib/include/rtems/score/percpu.h:20, 2012-07-31T02:43:09 from /opt/rtems-4.11/powerpc-rtems4.11/mpc5674fevb_spe/lib/include/rtems/system.h:21, 2012-07-31T02:43:10 from /opt/rtems-4.11/powerpc-rtems4.11/mpc5674fevb_spe/lib/include/rtems/chain.h:23, 2012-07-31T02:43:12 from /home/sh/git-rtems-rtl/rtl-indirect-ptr.h:24, 2012-07-31T02:43:13 from /home/sh/git-rtems-rtl/rtl-allocator.h:21, 2012-07-31T02:43:15 from ../../rtl-alloc-heap.h:19, 2012-07-31T02:43:16 from ../../rtl-alloc-heap.c:18: 2012-07-31T02:43:17 I am using github's gist for pastes :) 2012-07-31T02:43:18 /home/sh/git-rtems-rtl/libbsd/include/sys/cdefs.h:59:27: fatal error: machine/cdefs.h: No such file or directory 2012-07-31T02:43:19 compilation terminated. 2012-07-31T02:43:59 I will add powerpc. I have not tested the relocation code arm. 2012-07-31T02:44:04 ok 2012-07-31T02:44:58 It is easy to test. Just set a break point then get gdb to disassemble the instructions at the location and then step the fix up and check after 2012-07-31T02:45:26 Fix up code is in rtl-mdreloc-arm. 2012-07-31T02:45:30 c 2012-07-31T02:46:18 The specific issue in arm is the PC24 support 2012-07-31T02:46:28 There needs to be a way to test this. 2012-07-31T02:47:14 Thumb support is present 2012-07-31T02:51:16 ok, i hope i find some time to test this 2012-07-31T02:52:35 Thanks. 2012-07-31T02:53:05 There is an init script in the source which is built into a tar file and embedded in the executable 2012-07-31T02:53:54 Also when you are ready I need to walk you though a couple of things related to the base image symbol loading. 2012-07-31T02:54:37 If your target has a disk or network access and you can access it you can load the base image symbols at runtime else there is an awkward path to get them embedded 2012-07-31T02:54:51 I hope to get this sorted with a little from Amar at some stage 2012-07-31T03:05:03 *** xian9fu has quit IRC (Quit: Leaving) 2012-07-31T03:05:04 i have a lot to do before the holiday in less than four weeks, so it will be probably after the holiday 2012-07-31T03:05:16 *** xian9fu has joined #rtems 2012-07-31T03:05:38 maybe this is some work for your new collegue as a training excersise 2012-07-31T03:12:50 *** deb has quit IRC (Quit: This computer has gone to sleep) 2012-07-31T03:17:42 *** mumptai_ has joined #rtems 2012-07-31T04:40:49 *** arvind_khadri has joined #rtems 2012-07-31T04:48:23 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2012-07-31T05:06:24 *** arvind_khadri has joined #rtems 2012-07-31T05:23:00 *** WikL_ has joined #rtems 2012-07-31T05:25:46 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2012-07-31T06:26:52 *** QingPei has joined #rtems 2012-07-31T07:31:03 *** deb has joined #rtems 2012-07-31T08:46:49 *** WikL_ has quit IRC (Ping timeout: 255 seconds) 2012-07-31T08:48:42 *** WikL has joined #rtems 2012-07-31T09:09:28 *** weiY has joined #rtems 2012-07-31T09:12:46 *** WikL has quit IRC (Remote host closed the connection) 2012-07-31T09:18:07 *** arvind_khadri has joined #rtems 2012-07-31T09:27:43 *** gedare has joined #rtems 2012-07-31T09:35:54 *** xian9fu has quit IRC (Ping timeout: 252 seconds) 2012-07-31T10:01:53 *** Hesham has joined #rtems 2012-07-31T10:02:53 hi gedare, how are you ? 2012-07-31T10:03:17 hi Hesham i am well, yourself? 2012-07-31T10:03:34 fine thanks 2012-07-31T10:03:48 can we discuss now ? 2012-07-31T10:04:14 yes 2012-07-31T10:04:38 great 2012-07-31T10:04:48 About ppc code ... 2012-07-31T10:05:15 I have defined incomplete pointer at high-level code .... 2012-07-31T10:05:37 and defined its struct at libcpu code as your example 2012-07-31T10:06:05 and i tried to easily use type name at high-level code to access struct members 2012-07-31T10:06:10 but that failed 2012-07-31T10:06:19 you can't dereference the pointer 2012-07-31T10:06:25 all you can do is pass around the pointer to incomplete type 2012-07-31T10:06:28 that's why i used struct libcpu_mmu_mpe 2012-07-31T10:06:39 because the compiler does not know what the structure is like when compiling libcpu 2012-07-31T10:06:54 or score i mean 2012-07-31T10:07:03 yes i know 2012-07-31T10:07:05 *** jenniferA has joined #rtems 2012-07-31T10:07:07 ok.. 2012-07-31T10:07:12 but you commented at some part which is ..... 2012-07-31T10:07:41 pt_entry = (struct libcpu_mmu_mpe) mpe->cpu_mpe; 2012-07-31T10:07:44 *** jenniferA has quit IRC (Client Quit) 2012-07-31T10:07:59 and asked whether casting is necessary 2012-07-31T10:08:34 in that cases yes it is necessary 2012-07-31T10:09:03 Ok 2012-07-31T10:09:14 that's fine I just wasn't sure while reading the code. 2012-07-31T10:09:20 i can typedef struct libcpu_mmu_mpe "any type name" but i was just experimenting 2012-07-31T10:09:26 did you tested it ? 2012-07-31T10:09:33 test * 2012-07-31T10:09:33 i did not try to run it yet 2012-07-31T10:09:38 i just read some of the code 2012-07-31T10:09:56 trying to get the code to look right first ): 2012-07-31T10:09:58 :_ 2012-07-31T10:10:02 can't make a smile. 2012-07-31T10:10:02 ok, the remaining work would be cleaning the code 2012-07-31T10:10:08 exactly. 2012-07-31T10:10:16 i rather have clean code first :) 2012-07-31T10:10:49 otherwise i have to read the code too many times. 2012-07-31T10:11:04 once when i test it, and again after cleaning 2012-07-31T10:11:12 Ok i will clean it today 2012-07-31T10:11:26 The other issue 2012-07-31T10:11:36 thanks. i had a deadline today.. hopefully i might have more time to test it soon 2012-07-31T10:11:55 good 2012-07-31T10:12:05 about struct libcpu_mmu_mpe 2012-07-31T10:12:18 it's only defined at BSP code 2012-07-31T10:12:46 so i left it as it is, as its name is self illustrated 2012-07-31T10:13:02 but at high-level code i will rename it to libcpu_mpe 2012-07-31T10:13:06 is that ok ? 2012-07-31T10:13:17 ;hmmmm 2012-07-31T10:13:19 no 2012-07-31T10:13:25 i think... 2012-07-31T10:13:36 well what will it look like on another BSP that uses mpu? 2012-07-31T10:13:48 would it still be libcpu_mmu_mpe in the BSP? 2012-07-31T10:13:52 at high-level ? 2012-07-31T10:14:00 no 2012-07-31T10:14:02 at BSP code 2012-07-31T10:14:05 it would be 2012-07-31T10:14:13 libcpu_mpu_mpe 2012-07-31T10:14:18 hmm. 2012-07-31T10:14:24 every BSP define it 2012-07-31T10:14:33 according its processor 2012-07-31T10:14:33 why bother with making every BSP define it with a different name? 2012-07-31T10:15:12 could it be libbsp_mpe? 2012-07-31T10:15:40 or actually .... rtems_bsp_mm_mpe 2012-07-31T10:15:41 sure, but what if some one read the code and wanted to know whether this BSP uses mpe or mmu ? 2012-07-31T10:15:45 to be more suitable 2012-07-31T10:16:16 i don't know that matters so much 2012-07-31T10:16:40 it matters more i think to be consistent across BSPs.. 2012-07-31T10:16:52 if it does not matter i would change the name 2012-07-31T10:16:56 that every BSP that is going to support rtems_mm defines the same structure type 2012-07-31T10:17:16 it's ok 2012-07-31T10:17:25 this way it is easy to understand across bsps.... 2012-07-31T10:17:30 and also for something like ARM.. 2012-07-31T10:17:50 where you have some BSP using mpu, and some using mmu, it may simplify some shared code 2012-07-31T10:18:19 i see 2012-07-31T10:18:54 i asked about lm3s6965 board and which code it imports from libcpu 2012-07-31T10:19:28 sebastian replied that it has arm7m processor which does only have mpu 2012-07-31T10:20:13 and there is an existing code for mpu at cpukit/.../score/.../arm7v 2012-07-31T10:20:32 good 2012-07-31T10:20:42 you should be able to use it then? 2012-07-31T10:20:45 But regions are statically defined at startup stage 2012-07-31T10:21:14 of course... that is probably how any CPU that has any kind of MM will do it in RTEMS 2012-07-31T10:21:31 that's the reason to have this project in the first place 2012-07-31T10:23:13 so what is next? 2012-07-31T10:25:24 *** Hesham has quit IRC (Ping timeout: 255 seconds) 2012-07-31T10:27:16 *** weiY has quit IRC (Ping timeout: 248 seconds) 2012-07-31T10:28:56 *** Hesham has joined #rtems 2012-07-31T10:29:05 sorry disconnected, did i miss anything ? 2012-07-31T10:29:13 But regions are statically defined at startup stage 2012-07-31T10:29:13 of course... that is probably how any CPU that has any kind of MM will do it in RTEMS 2012-07-31T10:29:13 that's the reason to have this project in the first place 2012-07-31T10:29:13 so what is next? 2012-07-31T10:29:25 i wrote the design without testing anyway 2012-07-31T10:29:35 I am thinking to come back to an earlier RTEMS version that one simulator work with 2012-07-31T10:29:41 i thought armulator would be great 2012-07-31T10:29:46 but i discovered it was removed from RTEMS 2012-07-31T10:29:50 What do you suggest ? 2012-07-31T10:30:30 doesn't gdb support one of the ARM target that use the CP15 MMU ? 2012-07-31T10:30:49 arm920? 2012-07-31T10:30:55 i tried all of them 2012-07-31T10:31:07 arm920 code is at csb337 BSP 2012-07-31T10:31:34 I tried to get it working but i could not 2012-07-31T10:31:46 would i get an earlier version of RTEMS and try ? 2012-07-31T10:32:15 what about gdbarmsim/arm920 ? 2012-07-31T10:32:30 not a good idea to use earlier versions of RTEMS to do development work... 2012-07-31T10:32:44 have to expend more effort merging 2012-07-31T10:33:14 i tried gdbarmsim too but it does not work 2012-07-31T10:34:45 running csb337/arm-rtems4.11/c/csb337/testsuites/samples/hello/hello.exe for maximum 60 seconds... 2012-07-31T10:34:51 and it hangs 2012-07-31T10:34:55 you never got any ARM target to work on gdb? 2012-07-31T10:35:18 in RTEMS, NO 2012-07-31T10:35:30 only lm3s6965 on QEMU 2012-07-31T10:35:58 did you try to run the executables directly in gdb? 2012-07-31T10:36:02 or only through using run scripts? 2012-07-31T10:37:27 i tried arm-rtems4.11-gdb too with all of them 2012-07-31T10:37:46 at the first place, i am trying arm-rtems4.11-gdb 2012-07-31T10:37:52 and if it fails 2012-07-31T10:38:05 i try the ones at sim-scripts 2012-07-31T10:38:45 and i tried skyeye also 2012-07-31T10:39:12 yeah you got some suggestions from sebhub and rayx about your efforts 2012-07-31T10:39:54 did you use the skyeye version 1.2.2? 2012-07-31T10:40:01 with the patch from the ml? 2012-07-31T10:40:30 i did tried to reach the path but it seems to be broken 2012-07-31T10:40:42 at the site 2012-07-31T10:40:54 i see 2012-07-31T10:40:57 patch** 2012-07-31T10:41:09 try asking xi yang for the patch directly then 2012-07-31T10:41:17 he is a gsoc student this year also so he should be available 2012-07-31T10:41:32 What i was trying today to compile skyeye 1.2.2 again with gcc 3.3 2012-07-31T10:41:59 if skyeye worked with csb337 that will be great 2012-07-31T10:42:10 i will ask him now 2012-07-31T10:42:12 Ok 2012-07-31T10:42:39 otherwise you should detail all that you have tried and send an email 2012-07-31T10:42:46 something should be working... 2012-07-31T10:43:04 at any rate you can use qemu with lm3s6965 2012-07-31T10:43:34 yes it is now working and sebastian told me that everything is ok 2012-07-31T10:46:45 ok. spend a little more time to try to get something else working so you get some breadth 2012-07-31T10:46:59 it is important to have multiple target cpus and bsps 2012-07-31T10:47:12 *** sebhub has quit IRC (Remote host closed the connection) 2012-07-31T10:47:19 i sent xi yang an email now 2012-07-31T10:47:59 I designed the ARM code for MMU to be easily shared between Most arm targets 2012-07-31T10:48:46 So we may create an upper layer for ARM targets that have similar features related to MMU 2012-07-31T10:49:13 yeah well the trick will be to see whether when your design meets implementation if targets using mpu and mmu will work 2012-07-31T10:49:21 there are also ppc targets that use mpu 2012-07-31T10:49:32 The process of installing/unstalling mpe would be very quickly/easy 2012-07-31T10:51:22 Okay... 2012-07-31T10:51:23 i just hope to get skyeye working with CSB337 as i studied all mmu code for arm920 2012-07-31T10:51:30 yeah 2012-07-31T10:52:06 gdb really should work too though... i am surprised, and i think if you provide some more detail about what you are trying and what is going wrong maybe you can elicit more help from the ml... 2012-07-31T10:52:35 you might try rtems-users to see if anyone actively uses one of the arm gdb variants.. 2012-07-31T10:52:54 when you do ask your questions make sure you provide plenty of supporting details. you tend to send emails to the ml that are overly terse 2012-07-31T10:53:51 ok i will sent an email with enough details with what i try to do and how, hoping to fine a help 2012-07-31T10:54:47 *** mikeg has joined #rtems 2012-07-31T10:55:35 ok. when you clean the code i will give it a try for powerpc 2012-07-31T10:56:01 i will clean it and update you today so you can try it 2012-07-31T10:56:13 give me a clean branch with a single commit that integrates all of the changes from Quanming's code 2012-07-31T10:57:13 and has everything you've done to rename and refactor that code :) 2012-07-31T10:57:27 That means i have to get Quanming work back and commit it at a new branch then add my modifications and push it again right ? 2012-07-31T10:58:10 you can probably just make a new branch from your current one and use git rebase -i on it 2012-07-31T10:58:22 that will let you squash commits together and modify commit messages 2012-07-31T10:59:04 but yes basically... i want fewer commits for code review because that makes it easier for me to find what changes actually matter 2012-07-31T10:59:06 ok i will try that 2012-07-31T10:59:19 it's good to develop with lots of commits... 2012-07-31T10:59:33 but then for review it's easier to see just the set of everything that changes all at once 2012-07-31T11:00:03 although actually I guess github shows me that too 2012-07-31T11:00:03 it is ok, i will do it 2012-07-31T11:00:07 OK 2012-07-31T11:00:36 i will notify you when i do it 2012-07-31T11:00:40 ok thanks. 2012-07-31T11:01:03 most welcome :) 2012-07-31T11:17:04 *** arvind_khadri has quit IRC (Ping timeout: 260 seconds) 2012-07-31T11:20:50 *** arvind_khadri has joined #rtems 2012-07-31T11:41:38 *** Hesham has quit IRC (Ping timeout: 240 seconds) 2012-07-31T11:44:08 *** Hesham has joined #rtems 2012-07-31T11:48:21 *** verm__ has quit IRC (Read error: Connection reset by peer) 2012-07-31T11:58:31 *** verm__ has joined #rtems 2012-07-31T12:19:36 *** antgreen has joined #rtems 2012-07-31T12:20:29 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-07-31T12:44:28 *** gedare has quit IRC (Quit: Ex-Chat) 2012-07-31T12:56:40 *** antgreen has quit IRC (Read error: Connection reset by peer) 2012-07-31T13:19:55 *** alseh has joined #rtems 2012-07-31T13:24:52 *** mumptai_ has quit IRC (Ping timeout: 248 seconds) 2012-07-31T13:38:11 *** Hesham has joined #rtems 2012-07-31T13:41:26 *** antgreen has joined #rtems 2012-07-31T14:01:16 *** hizzle has joined #rtems 2012-07-31T14:10:18 I was thinking about porting rtems to the microblaze. Has any one ventured into this realm? 2012-07-31T14:14:59 *** Hesham has quit IRC (Ping timeout: 260 seconds) 2012-07-31T14:16:57 *** Hesham has joined #rtems 2012-07-31T14:32:13 *** krzysiekm13 has joined #rtems 2012-07-31T14:37:29 *** jhirata has joined #rtems 2012-07-31T14:41:29 *** krzysiekm13 has left #rtems 2012-07-31T14:43:45 *** hizzle has quit IRC (Ping timeout: 245 seconds) 2012-07-31T14:48:00 *** Hesham1 has joined #rtems 2012-07-31T14:49:08 what would be the best way to provide drivers for xilinx peripherals? They can be used with ppc, microblaze, and arm on the new zynq processors 2012-07-31T14:49:19 I mean, where should they go in the source tree 2012-07-31T14:51:39 *** Hesham has quit IRC (Ping timeout: 250 seconds) 2012-07-31T14:57:03 *** Hesham has joined #rtems 2012-07-31T14:59:06 *** Hesham1 has quit IRC (Ping timeout: 255 seconds) 2012-07-31T15:01:28 *** antgreen has quit IRC (Ping timeout: 246 seconds) 2012-07-31T15:02:42 *** Hesham1 has joined #rtems 2012-07-31T15:04:48 *** Hesham has quit IRC (Ping timeout: 252 seconds) 2012-07-31T15:24:13 *** Hesham has joined #rtems 2012-07-31T15:25:30 *** alseh has quit IRC (Remote host closed the connection) 2012-07-31T15:27:00 *** Hesham1 has quit IRC (Ping timeout: 248 seconds) 2012-07-31T15:44:03 *** Hesham has quit IRC (Quit: Leaving.) 2012-07-31T17:02:06 *** verm__ has quit IRC (Changing host) 2012-07-31T17:02:08 *** verm__ has joined #rtems 2012-07-31T18:37:30 *** Hesham has joined #rtems 2012-07-31T18:40:04 *** jhirata has left #rtems ("Leaving") 2012-07-31T18:41:16 *** jhirata has joined #rtems 2012-07-31T19:11:20 *** Hesham has quit IRC (Quit: Leaving.) 2012-07-31T19:35:43 *** weiY has joined #rtems 2012-07-31T19:36:24 *** QingPei has left #rtems 2012-07-31T20:51:26 *** weiY has quit IRC (Ping timeout: 246 seconds) 2012-07-31T21:03:08 9 2012-07-31T22:11:00 10 2012-07-31T22:45:36 1:45 2012-07-31T23:24:37 *** deb has left #rtems ("Leaving") 2012-08-01T00:04:06 *** arvind_khadri has quit IRC (Ping timeout: 250 seconds) 2012-08-01T00:36:20 *** lcpfnvcy has quit IRC (Ping timeout: 248 seconds) 2012-08-01T00:41:43 *** lcpfnvcy has joined #rtems 2012-08-01T00:50:54 *** lcpfnvcy has quit IRC (Ping timeout: 250 seconds) 2012-08-01T00:56:50 *** lcpfnvcy has joined #rtems 2012-08-01T01:03:03 *** lcpfnvcy has quit IRC (Ping timeout: 246 seconds) 2012-08-01T01:09:08 *** lcpfnvcy has joined #rtems 2012-08-01T02:10:29 *** sebhub has joined #rtems 2012-08-01T02:36:21 good morning 2012-08-01T03:43:46 *** mumptai_ has joined #rtems 2012-08-01T04:23:28 *** juli1 has quit IRC (Ping timeout: 245 seconds) 2012-08-01T04:33:47 *** arvind_khadri has joined #rtems 2012-08-01T06:58:15 *** krzysiekm13 has joined #rtems 2012-08-01T07:00:30 *** arvind_khadri has quit IRC (Ping timeout: 249 seconds) 2012-08-01T07:17:42 *** arvind_khadri has joined #rtems 2012-08-01T07:21:08 *** krzysiekm13 has left #rtems 2012-08-01T07:22:24 *** mumptai_ has quit IRC (Ping timeout: 246 seconds) 2012-08-01T07:34:50 *** mumptai_ has joined #rtems 2012-08-01T08:02:47 *** arvind_khadri has joined #rtems 2012-08-01T08:17:52 *** deb has joined #rtems 2012-08-01T08:30:38 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-08-01T08:38:16 *** krzysiekm13 has joined #rtems 2012-08-01T08:38:22 hi 2012-08-01T08:39:07 mi name is Krzysiek, i'm selected by ESA as this year ESA SOCIS participant 2012-08-01T08:39:27 i'll be working to improve rtems test coverage 2012-08-01T08:41:31 so i'd like to say "Hello" and i hope to be part of community :) 2012-08-01T08:03:12 *** rtemsLogger has joined #rtems 2012-08-01T09:06:02 *** alseh has joined #rtems 2012-08-01T09:14:37 *** krzysiekm13 has left #rtems 2012-08-01T09:16:11 *** gedare has joined #rtems 2012-08-01T09:16:56 *** alseh has quit IRC (Ping timeout: 240 seconds) 2012-08-01T10:12:42 *** lcpfnvcy has quit IRC (Ping timeout: 264 seconds) 2012-08-01T10:18:22 *** lcpfnvcy has joined #rtems 2012-08-01T10:26:39 *** lcpfnvcy has quit IRC (Ping timeout: 252 seconds) 2012-08-01T10:32:33 *** lcpfnvcy has joined #rtems 2012-08-01T10:37:07 *** weiY has joined #rtems 2012-08-01T10:38:37 *** lcpfnvcy has quit IRC (Ping timeout: 240 seconds) 2012-08-01T10:45:44 *** sebhub has quit IRC (Remote host closed the connection) 2012-08-01T10:46:08 *** lcpfnvcy has joined #rtems 2012-08-01T10:56:30 *** mumptai_ has quit IRC (Ping timeout: 264 seconds) 2012-08-01T11:04:36 *** weiY has quit IRC (Ping timeout: 240 seconds) 2012-08-01T12:24:22 *** mumptai_ has joined #rtems 2012-08-01T12:42:06 *** mumptai_ has quit IRC (Ping timeout: 264 seconds) 2012-08-01T15:31:07 *** alseh has joined #rtems 2012-08-01T16:14:51 *** gedare has quit IRC (Quit: Leaving) 2012-08-01T19:13:55 *** alseh has quit IRC (Ping timeout: 240 seconds) 2012-08-01T19:14:40 *** alseh has joined #rtems 2012-08-01T19:23:18 *** alseh has quit IRC (Ping timeout: 246 seconds) 2012-08-01T21:33:02 *** ChanServ has quit IRC (*.net *.split) 2012-08-01T21:50:46 *** ChanServ has joined #rtems 2012-08-01T21:50:46 *** cameron.freenode.net sets mode: +o ChanServ 2012-08-01T21:51:39 *** alseh has joined #rtems 2012-08-01T23:09:01 *** alseh has quit IRC (Ping timeout: 244 seconds) 2012-08-01T23:10:06 *** deb has left #rtems ("Leaving") 2012-08-01T23:20:48 *** alseh has joined #rtems 2012-08-02T00:51:42 *** lcpfnvcy has quit IRC (Ping timeout: 264 seconds) 2012-08-02T01:02:59 *** lcpfnvcy has joined #rtems 2012-08-02T03:46:37 *** jhirata has quit IRC (Ping timeout: 240 seconds) 2012-08-02T03:58:13 *** alseh has quit IRC (Remote host closed the connection) 2012-08-02T04:09:22 *** ChanServ has quit IRC (*.net *.split) 2012-08-02T04:27:10 *** ChanServ has joined #rtems 2012-08-02T04:27:10 *** cameron.freenode.net sets mode: +o ChanServ 2012-08-02T05:02:51 *** vlasta has joined #rtems 2012-08-02T05:24:14 *** vlasta has left #rtems ("Odcházím") 2012-08-02T06:11:10 *** jhirata has joined #rtems 2012-08-02T06:33:20 *** alseh has joined #rtems 2012-08-02T06:45:31 *** arvind_khadri has joined #rtems 2012-08-02T07:17:16 *** arvind_khadri has quit IRC (Ping timeout: 244 seconds) 2012-08-02T07:29:05 *** arvind_khadri has joined #rtems 2012-08-02T07:40:50 *** zw_yao has joined #rtems 2012-08-02T07:58:11 *** weiY has joined #rtems 2012-08-02T08:18:10 *** gedare has joined #rtems 2012-08-02T08:20:13 zw_yao: you can ask your question about Objects_Open_u32 to rtems-devel. also your indentation problem is likely due to using tabs instead of spaces 2012-08-02T08:21:20 ok i see you did ask your question, good 2012-08-02T08:21:56 is the irc meeting happening today? 2012-08-02T08:22:47 gedare, Yeah, I've asked this question on ml. 2012-08-02T08:22:55 i was hoping to be around but i have a meeting to go to now... deb should be around 2012-08-02T08:23:15 *** WikL has joined #rtems 2012-08-02T08:23:21 And I'll check the tabs later. 2012-08-02T08:23:29 *** WikL__ has joined #rtems 2012-08-02T08:25:46 verm__: yes it should though no announcement was made yet 2012-08-02T08:49:45 *** weiY has quit IRC (Ping timeout: 244 seconds) 2012-08-02T08:51:55 *** cdcs has joined #rtems 2012-08-02T08:54:52 *** zw_yao has quit IRC (Read error: Connection timed out) 2012-08-02T08:55:40 *** zw_yao has joined #rtems 2012-08-02T09:01:15 *** tuxmaniac has joined #rtems 2012-08-02T09:19:13 *** krzysiekm13 has joined #rtems 2012-08-02T09:25:40 *** Hesham has joined #rtems 2012-08-02T09:26:41 *** Hesham has quit IRC (Client Quit) 2012-08-02T09:39:36 *** Hesham has joined #rtems 2012-08-02T09:43:55 https://docs.google.com/document/d/1gQtnb1wJ6o84CRd28T1xNpBdbCR8b9fEMl4Lu4StpwI/edit# 2012-08-02T09:44:00 ^ gdoc for the meeting 2012-08-02T09:51:11 *** deb has joined #rtems 2012-08-02T09:51:55 *** medivhc has joined #rtems 2012-08-02T09:57:39 gdoc for meeting: https://docs.google.com/document/d/1gQtnb1wJ6o84CRd28T1xNpBdbCR8b9fEMl4Lu4StpwI/edit# 2012-08-02T09:59:34 okie let's start 2012-08-02T09:59:54 Hesham: you're up 2012-08-02T09:59:57 yes 2012-08-02T10:00:00 what's your status 2012-08-02T10:00:09 I have moved all attributes handling to BSP code and avoided any translations. 2012-08-02T10:00:22 Also i have embedded a pointer to libbsp_mm_mpe to high-level mpe Control block, this way, the process of searching for particular PTE in a pagetable is eliminated.when changing page attributes. 2012-08-02T10:00:47 and wrote code for arm920t mmu 2012-08-02T10:01:08 wasn't that code already in the tree? i thought there was some. 2012-08-02T10:01:36 yes but it is just initializing master page table 2012-08-02T10:01:38 *** julien__ has joined #rtems 2012-08-02T10:01:44 *** julien__ is now known as juli1 2012-08-02T10:01:51 ah i see.. so you added more features? 2012-08-02T10:01:54 with fault entries (ox00) 2012-08-02T10:01:54 *** zw_yao has quit IRC (Quit: Leaving) 2012-08-02T10:02:01 ok 2012-08-02T10:02:07 yes i added a control block for page table 2012-08-02T10:02:13 and region 2012-08-02T10:02:16 *** zw_yao has joined #rtems 2012-08-02T10:02:36 you have prepped a branch for review for committing? 2012-08-02T10:02:47 Not yet 2012-08-02T10:02:49 ok 2012-08-02T10:03:04 i will write some macros/functions and push it today/tomorrow 2012-08-02T10:03:05 it would be good to try to get your libmm modifications reviewed and committed 2012-08-02T10:03:19 where ? 2012-08-02T10:03:20 i thought you already had cleaned that libmm code and it was ready to review 2012-08-02T10:03:54 i cleaned it and waiting your review/testing 2012-08-02T10:03:57 iok 2012-08-02T10:04:10 i thought so. your libmm_working branch? 2012-08-02T10:04:30 yes it is 2012-08-02T10:04:31 can you make a pull request on github for it? 2012-08-02T10:04:52 How? 2012-08-02T10:04:57 you'll figure it out :) 2012-08-02T10:05:00 ok anything else? 2012-08-02T10:05:21 no thanks :) 2012-08-02T10:05:30 next: WikL 2012-08-02T10:05:36 here 2012-08-02T10:05:52 btw i'm going in the order listed on the gdoc so if you haven't put your name in i will miss you until the end :) 2012-08-02T10:06:00 update please! 2012-08-02T10:06:01 so I started to implement the clock driver 2012-08-02T10:06:09 * gedare wakes juli1 and cdcs 2012-08-02T10:06:21 but am not sure if it's the proper way to go about it 2012-08-02T10:06:31 well, does it work? 2012-08-02T10:06:36 Hello! 2012-08-02T10:06:52 it's got a bug somewhere on the POK side 2012-08-02T10:07:20 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-02T10:07:23 ok, it's not "somewhere", I know where it is. just can't figure out why it's breaking the code 2012-08-02T10:07:27 hi all 2012-08-02T10:07:29 i see.. well that's a start 2012-08-02T10:07:40 yet 2012-08-02T10:07:47 what is the bug? how it is manifested? 2012-08-02T10:08:39 I've added a system call which passes a pointer to a function which should be called at every clock tick 2012-08-02T10:08:39 from the clock interrupt handler 2012-08-02T10:09:19 and ? 2012-08-02T10:09:21 but the function passed by pointer is never called, it's like the very pointer is wrong somehow 2012-08-02T10:09:22 does that works ? 2012-08-02T10:09:24 does pok invoke partition code directly? 2012-08-02T10:09:24 *** weiY has joined #rtems 2012-08-02T10:09:44 yes 2012-08-02T10:09:57 and that's what I was going to ask about, since it doesn't seem proper 2012-08-02T10:10:04 indeed it does not to me 2012-08-02T10:10:58 can it be somehow registered as an event, called from the kernel but residing in partition? 2012-08-02T10:11:35 and it has to be called at every clock tick interrupt 2012-08-02T10:11:38 i'm not familiar enough with pok to give advice.. but normally, partition code should not be invoked directly by a partition kernel 2012-08-02T10:11:49 it makes sense. you are registering a partition interrupt handler 2012-08-02T10:11:50 so i think it is likely a design flaw to pass a function pointer directly 2012-08-02T10:12:45 the function should only be called after every POK critical code in the clock tick interrupt handler has finished 2012-08-02T10:12:53 right 2012-08-02T10:13:20 yeah that makes sense... WikL are you using some method in pok to register the interrupt handler? 2012-08-02T10:13:41 well right now the handler is a macro 2012-08-02T10:13:47 I mean originaly 2012-08-02T10:14:04 you cannot let the partition's specific handler destroy the time keeping in POK, and therefore the time partitioning 2012-08-02T10:14:11 calling pok_shed() and increasing the tick counter 2012-08-02T10:15:11 I've added my own function to that, which checks if the registered pointer is null, and if not, calling the function 2012-08-02T10:15:23 but yes, it's partition code and can contain anything 2012-08-02T10:15:31 so it's evidently not safe 2012-08-02T10:16:03 ok 2012-08-02T10:16:22 write up your report and try to debug i guess... 2012-08-02T10:16:25 but I didn't find any other way to call partition code at each interrupt 2012-08-02T10:16:52 somehow pok must provide a way to register interrupt handlers 2012-08-02T10:17:00 whatever that is will be the mechanism to use I would think 2012-08-02T10:17:18 but I cannot overwrite the current handler 2012-08-02T10:17:33 yeah... 2012-08-02T10:17:44 pok should offer some way to register additional handlers for interrupt sources 2012-08-02T10:18:02 and iterate through the partition handlers 2012-08-02T10:18:11 unless i'm missing something this is how I would expect to find things 2012-08-02T10:18:38 exactly. but you will always have the POK handler calling the Partitions handler. 2012-08-02T10:18:38 that's what I was looking for at the beginning, but didn't find a way 2012-08-02T10:18:56 (unless you have some kind of hardware virtualization) 2012-08-02T10:19:49 ok well i think you should dig deeper on this 2012-08-02T10:19:57 before trying to debug your current approach 2012-08-02T10:20:03 ok 2012-08-02T10:20:07 make sure your approach is "right" 2012-08-02T10:20:12 understood 2012-08-02T10:20:12 i suspect you missed something 2012-08-02T10:20:31 anything else? 2012-08-02T10:20:47 I'll post a problem description on the list and will keep looking into it myself 2012-08-02T10:20:52 nope, that's it 2012-08-02T10:20:57 ok 2012-08-02T10:21:06 next up is deb 2012-08-02T10:21:14 hi 2012-08-02T10:21:19 hi. what's your status? 2012-08-02T10:21:33 I was working on the file system plan 2012-08-02T10:21:41 I have some questions 2012-08-02T10:22:00 feel free to ask your questions through rtems-devel ml also..don't have to wait for these meetings 2012-08-02T10:22:06 especially now we are getting to crunch time 2012-08-02T10:22:08 go ahead 2012-08-02T10:22:33 I tried to separate the layers of file system to fit in the four groups 2012-08-02T10:22:48 hardware-independent, CPU, BSP, and peripheral 2012-08-02T10:23:08 I read the doc on file system development for RTEMS 2012-08-02T10:23:27 It seems that you have divide it into two layers stuff 2012-08-02T10:23:34 First layer is API 2012-08-02T10:23:54 Second layer is filesystem specific handlers and file type specific handlers 2012-08-02T10:24:06 that is about right.. there is also device access below that 2012-08-02T10:24:31 yeah 2012-08-02T10:24:57 But I don't know where to find the device access layer and the interface between the two 2012-08-02T10:25:22 The dependencies between handler layer and the device access layer 2012-08-02T10:25:36 ask on rtems-devel. :) i think the bdbuf might be it though 2012-08-02T10:25:46 OK 2012-08-02T10:26:31 Where could I find dependencies between the source files? 2012-08-02T10:26:33 but.. this is all getting quite complicated, i thought we were going to see some test framework already, maybe using the current test suite code... 2012-08-02T10:26:48 *** zw_yao has quit IRC (Quit: Leaving) 2012-08-02T10:27:01 *** zw_yao has joined #rtems 2012-08-02T10:27:14 http://www.rtems.org/onlinedocs/doxygen/cpukit/html/group__rtems__bdbuf.html 2012-08-02T10:27:34 you might like to talk to medivhc he is working with bdbuf on his gsoc 2012-08-02T10:27:45 and he has experience from the past in writing test cases 2012-08-02T10:27:58 I need to make it clear... 2012-08-02T10:28:17 Last week, I thought I was working on test case writing 2012-08-02T10:28:20 It seems not 2012-08-02T10:28:40 the only thing i could think otherwise would be to use the build system to figure out file dependencies, but that seems even harder 2012-08-02T10:28:58 I have to layout the ondisk hierarchy 2012-08-02T10:29:10 for which tool? 2012-08-02T10:29:38 not tool specific, for example, the ondisk hierarchy for file system 2012-08-02T10:29:41 i see 2012-08-02T10:29:55 ask on rtems-devel lay out all of your findings and your problems... 2012-08-02T10:30:15 I wonder whether or not block device buffer supports all file systems? or just disk? 2012-08-02T10:30:16 we need some more details and more time to dig than i can give in the few minutes here 2012-08-02T10:30:23 OK 2012-08-02T10:30:30 especially i am not expert in the filesystems 2012-08-02T10:30:52 realistically we need to get some code out of you soon too 2012-08-02T10:31:05 I asked amar about this too 2012-08-02T10:31:16 good.. 2012-08-02T10:31:23 He said what I was doing was study 2012-08-02T10:31:40 The GSOC is not all about code 2012-08-02T10:32:11 this is true, but we must have something by the end. try to clarify with him what he thinks that will be 2012-08-02T10:32:22 OK 2012-08-02T10:32:37 it's fine to study through the midterm..we even expect it for ambitious projects like yours.. 2012-08-02T10:32:47 but in the end we must have some code to contribute 2012-08-02T10:33:10 OK 2012-08-02T10:33:16 i understand you are probably frustrated, and i hope we can help you out... post to the rtems-devel and keep asking questions and probing 2012-08-02T10:33:49 OK 2012-08-02T10:33:52 it would be good to summarize all you learned so far through studying. and try to form a plan with verm__ . 2012-08-02T10:33:53 OK 2012-08-02T10:34:09 i'll wait for your email for more. 2012-08-02T10:34:17 OK 2012-08-02T10:34:21 *** JenniferA has joined #rtems 2012-08-02T10:34:23 moving on... zw_yao 2012-08-02T10:34:35 here. 2012-08-02T10:34:42 status please 2012-08-02T10:35:24 Hi, last week, I finish the 4 test cases, and post my pull request. 2012-08-02T10:35:48 I think the one rbtree implementation is ready to review. 2012-08-02T10:36:37 super. 2012-08-02T10:36:46 (Sorry, my network is unstable...) 2012-08-02T10:36:52 what do you have next? 2012-08-02T10:37:15 And next, I'm thinking about the one rbtree per thread approach. 2012-08-02T10:38:18 However, I think that would not be much work, because it is a little similiar to current one rbtree approach. 2012-08-02T10:38:22 Do you have any way to evaluate the current implementation? I mean, I assume you can make both approaches "work", so how would you say which one is better? 2012-08-02T10:39:48 there is 2 time test in tmtest, which can evalute the runtime of the approach. 2012-08-02T10:40:11 however, I've not much other idea about evaluation.. 2012-08-02T10:41:07 okay. i just don't know if it is worthwhile to expend effort on another implementation 2012-08-02T10:41:16 unless there is a clear advantage 2012-08-02T10:42:01 I see.. 2012-08-02T10:42:37 *** DrJoel_ has joined #rtems 2012-08-02T10:42:49 * DrJoel_ is lurking.. really in a meeting 2012-08-02T10:43:01 i think you should report on your design to rtems-devel and solicit feedback on next steps. maybe provide a few options 2012-08-02T10:43:25 And at first, I also think about the hash approach, I'm still not clear if that approach is totally unacceptable... 2012-08-02T10:43:42 write a report 2012-08-02T10:43:42 OK. 2012-08-02T10:43:55 next up: medivhc 2012-08-02T10:44:04 hi 2012-08-02T10:44:18 hi medivhc what is your status? 2012-08-02T10:44:38 I am working on clock-pro using the policy api. 2012-08-02T10:45:34 I prefer to use the the approximation here http://linux-mm.org/PeterZClockPro2 2012-08-02T10:46:07 use two clocks 2012-08-02T10:46:49 Ok 2012-08-02T10:46:53 But still not solve how the hand hot move in the two clock. 2012-08-02T10:47:12 I plan to read the code in that page. 2012-08-02T10:47:27 will you be using any of the code there? 2012-08-02T10:47:37 no 2012-08-02T10:47:42 ok. it looks like a lot :) 2012-08-02T10:48:16 ok well you have a lot to do.. if you run into trouble send email to rtems-devel and CC your mentor 2012-08-02T10:48:22 I just don't figure out how the hand_hot move in the non_resident clock. 2012-08-02T10:49:32 i'm confused about using 2 clocks 2012-08-02T10:49:41 did your mentor suggest it? 2012-08-02T10:49:49 the non_resident buff need hash table or something to search. But I am not sure. 2012-08-02T10:50:33 ok.. well send an email to rtems-devel stating what you will implement and why... 2012-08-02T10:50:52 ok 2012-08-02T10:51:05 and why not to implement the original ClockPro 2012-08-02T10:51:34 next up is weiY 2012-08-02T10:51:41 OK 2012-08-02T10:52:13 medivhc: make sure you will implement something that you can do in chunks... or you risk running out of time before reaching a point you have code that can commit. 2012-08-02T10:52:17 This week my main work is to port the powerpc atomic operations to rtems from FreeBSD kernel 2012-08-02T10:52:32 ok 2012-08-02T10:52:36 I see 2012-08-02T10:52:44 and now most of code is ready, but still a little work 2012-08-02T10:52:54 and need to test it 2012-08-02T10:53:01 good how was it to add another architecture? 2012-08-02T10:53:26 do you think it is possible to script the effort? make it easier to maintain or pull in updated code from BSD sources? 2012-08-02T10:53:52 medivhc: If you can find a research paper on these algorithms or write up, please make sure those get passed along and included in your documentation. Especially as you are deciphering the algorithm, it would help if we could read any papers describing the algorithm 2012-08-02T10:53:54 i'm not saying that is necessary but you may want to think about how that might work or not. 2012-08-02T10:54:10 i think it will be hard to just a script. because there are so many name changes 2012-08-02T10:54:21 DrJoel_: the primary papers are already linked from the project's wiki page (for original LIRS and ClockPro) 2012-08-02T10:54:27 yeah, i understand what you say 2012-08-02T10:54:41 * DrJoel_ blushes from ignorance :( 2012-08-02T10:54:42 ok weiY. do you think you will look at synchronization primitives next? 2012-08-02T10:55:46 i think i need some time to test the powerpc atomic 2012-08-02T10:55:57 ok. do you have powerpc target working? 2012-08-02T10:56:20 i have a p1020 target 2012-08-02T10:56:33 it might be nice to provide at least some kind of locking mechanism using the new atomics 2012-08-02T10:56:35 but i do not test rtems on this target 2012-08-02T10:56:57 i don't know if we have bsp support for it or not. 2012-08-02T10:57:13 and rtems do not support SMP on powerpc architecture 2012-08-02T10:57:15 psim would work for non-SMP 2012-08-02T10:57:20 correct 2012-08-02T10:57:25 so i don't know that more effort should go into itt 2012-08-02T10:57:43 and how about rtems support for p1020 bsp 2012-08-02T10:57:44 *** medivhc has quit IRC (Quit: Leaving) 2012-08-02T10:57:53 i'd say either you do the sparc32, or you implement some kind of SMP synchronization primitive, or you port a lock-free data structure. 2012-08-02T10:58:01 yo'ud have to ask about that on rtems-devel 2012-08-02T10:58:28 the qoriq bsp might support it but i have no idea 2012-08-02T10:59:14 ok, after i test all the case for powerpc i will do next step 2012-08-02T10:59:32 but this test maybe only on non-SMP mode 2012-08-02T10:59:36 ok it should not take long to do testing 2012-08-02T10:59:42 yeah 2012-08-02T11:00:07 next i will send a plan about next step to list 2012-08-02T11:00:09 that will be good though.. easy to reproduce non-SMP powerpc testing using psim 2012-08-02T11:00:11 good 2012-08-02T11:00:21 next up: alseh 2012-08-02T11:00:23 ok. 2012-08-02T11:00:35 hi all 2012-08-02T11:00:35 *** zw_yao has quit IRC (Quit: Leaving) 2012-08-02T11:00:38 hi alseh 2012-08-02T11:00:53 my status is a pause from solving the bug with the keyboard 2012-08-02T11:01:01 and solved the library updates 2012-08-02T11:01:05 ok 2012-08-02T11:01:15 i'm guessing the other libraries were easy 2012-08-02T11:01:29 quite easy ... small modifications 2012-08-02T11:01:34 I believe there is a bsp for that.. see the qoriq BSP. Sebastian should be able to comment.. ask on rtems-devel 2012-08-02T11:01:45 but it wouldn't have SMP support.. just use your hardwrae 2012-08-02T11:01:49 do you have the ports on your github? 2012-08-02T11:02:20 if you like you can make a pull request for updating these libraries 2012-08-02T11:02:22 not yet because I had some problems with git 2012-08-02T11:02:24 ok 2012-08-02T11:02:42 I will make it when they are ok ... won't take long 2012-08-02T11:02:46 super. 2012-08-02T11:02:59 what's next? 2012-08-02T11:03:07 I also read a littel about VBE ... 2012-08-02T11:03:22 the first thing I want to solve the keyboard bug in nxlib 2012-08-02T11:03:30 how bad is the bug 2012-08-02T11:03:40 it does not work at all 2012-08-02T11:03:45 oh: ) 2012-08-02T11:04:00 in microwindows kbd works, and the scancode is not sent forward 2012-08-02T11:04:08 ok 2012-08-02T11:04:49 then, what should be next ? ... recreate cirrus driver , or creating the VBE-based driver ? 2012-08-02T11:04:55 *** weiY has quit IRC (Ping timeout: 248 seconds) 2012-08-02T11:04:55 did you get any idea if the cirrus driver can be remade? 2012-08-02T11:05:18 *** weiY has joined #rtems 2012-08-02T11:05:20 yes, as ppisa said at the last meeting, there are 2 functions that require modifications 2012-08-02T11:05:30 do these drivers go into rtems, or into some other repo? 2012-08-02T11:05:47 rtems ... they are framebuffer drivers in the BSP 2012-08-02T11:06:10 ok. see if you can find the documentation you need to recreate the troublesome functions 2012-08-02T11:07:03 I have also the Linux driver (with license problem), and the cirrus MIT driver for documentation, that should be ok 2012-08-02T11:07:07 you'll need a specification of the features 2012-08-02T11:07:20 mit driver should be good. 2012-08-02T11:07:26 just have to avoid reading the linux driver too much 2012-08-02T11:07:56 MIT driver is actually more *rtems-friendly* 2012-08-02T11:08:04 can be adapted quite well 2012-08-02T11:08:04 mit driver might be usable directly 2012-08-02T11:08:16 well, the mit license is safe 2012-08-02T11:08:21 so if it is using the mit license that is good 2012-08-02T11:08:48 I'll see what the problems are and ask on the mailinglist 2012-08-02T11:09:02 OK great 2012-08-02T11:09:15 that it? 2012-08-02T11:09:23 yes, that's it for now 2012-08-02T11:09:30 ok. class dismissed 2012-08-02T11:09:30 thank you :) 2012-08-02T11:09:51 i look forward to seeing reports on rtems-devel 2012-08-02T11:25:11 *** weiY has quit IRC (Ping timeout: 248 seconds) 2012-08-02T11:25:41 *** weiY has joined #rtems 2012-08-02T11:30:26 *** gedare has quit IRC (Quit: Leaving) 2012-08-02T11:33:25 *** DrJoel_ has quit IRC (Quit: Page closed) 2012-08-02T11:36:36 *** weiY has quit IRC (Ping timeout: 248 seconds) 2012-08-02T11:38:29 *** WikL has quit IRC (Ping timeout: 255 seconds) 2012-08-02T11:38:44 *** WikL__ has quit IRC (Ping timeout: 248 seconds) 2012-08-02T11:40:05 *** WikL has joined #rtems 2012-08-02T11:40:33 *** WikL__ has joined #rtems 2012-08-02T11:40:42 *** JenniferA has quit IRC (Quit: Page closed) 2012-08-02T11:43:42 *** beng-nl has joined #rtems 2012-08-02T11:46:23 *** jhirata has left #rtems ("Leaving") 2012-08-02T12:09:09 *** Hesham has joined #rtems 2012-08-02T12:18:12 *** krzysiekm13 has left #rtems 2012-08-02T12:20:52 *** Hesham has quit IRC (Ping timeout: 248 seconds) 2012-08-02T12:53:50 *** tuxmaniac has quit IRC (Quit: Lost terminal) 2012-08-02T12:54:34 *** alseh has quit IRC (Remote host closed the connection) 2012-08-02T12:54:49 *** cdcs has quit IRC (Quit: Page closed) 2012-08-02T14:30:28 *** gedare has joined #rtems 2012-08-02T14:34:17 *** gedare has quit IRC (Client Quit) 2012-08-02T15:13:34 *** Hesham has joined #rtems 2012-08-02T15:17:33 lo 2012-08-02T15:29:26 *** Hesham1 has joined #rtems 2012-08-02T15:32:20 *** Hesham has quit IRC (Ping timeout: 248 seconds) 2012-08-02T15:33:08 *** WikL__ has quit IRC (Remote host closed the connection) 2012-08-02T15:33:08 *** WikL has quit IRC (Remote host closed the connection) 2012-08-02T15:49:37 *** Hesham1 has quit IRC (Quit: Leaving.) 2012-08-02T16:17:06 *** rokka has quit IRC (Ping timeout: 265 seconds) 2012-08-02T16:18:39 *** rokka has joined #rtems 2012-08-02T19:14:46 *** Hesham has joined #rtems 2012-08-02T20:08:04 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-02T20:31:55 *** arvind_khadri has quit IRC (Ping timeout: 252 seconds) 2012-08-02T20:54:58 *** arvind_khadri has joined #rtems 2012-08-02T21:01:43 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-08-02T21:14:24 *** arvind_khadri has joined #rtems 2012-08-02T22:02:17 *** xiangfu has joined #rtems 2012-08-02T22:45:58 *** xiangfu has quit IRC (Ping timeout: 245 seconds) 2012-08-02T22:47:28 *** xiangfu has joined #rtems 2012-08-02T23:30:30 *** deb has left #rtems ("Leaving") 2012-08-03T00:13:29 *** krzysiekm13 has joined #rtems 2012-08-03T00:53:05 *** krzysiekm13 has quit IRC (Quit: Leaving.) 2012-08-03T00:53:53 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2012-08-03T00:53:58 *** krzysiekm13 has joined #rtems 2012-08-03T01:00:52 *** xiangfu has quit IRC (Ping timeout: 248 seconds) 2012-08-03T01:01:27 *** arvind_khadri has joined #rtems 2012-08-03T01:12:39 *** xiangfu has joined #rtems 2012-08-03T01:40:32 *** sebhub has joined #rtems 2012-08-03T01:41:52 good morning 2012-08-03T01:58:53 *** arvind_khadri has quit IRC (Ping timeout: 245 seconds) 2012-08-03T02:08:28 *** xiangfu has quit IRC (Ping timeout: 245 seconds) 2012-08-03T02:15:55 *** arvind_khadri has joined #rtems 2012-08-03T03:08:41 *** Hesham has joined #rtems 2012-08-03T03:40:08 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-03T03:48:07 *** WikL has joined #rtems 2012-08-03T03:51:07 *** WikL_ has joined #rtems 2012-08-03T03:51:08 *** WikL__ has joined #rtems 2012-08-03T03:51:41 /users/ 2012-08-03T03:51:56 sry 2012-08-03T03:52:19 *** WikL__ has quit IRC (Client Quit) 2012-08-03T03:52:53 *** WikL_ has quit IRC (Client Quit) 2012-08-03T03:52:53 *** WikL has quit IRC (Client Quit) 2012-08-03T03:53:13 *** WikL has joined #rtems 2012-08-03T04:03:02 *** weiY has joined #rtems 2012-08-03T04:37:22 *** WikL has quit IRC (Remote host closed the connection) 2012-08-03T05:29:30 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-08-03T05:31:18 *** arvind_khadri has joined #rtems 2012-08-03T05:41:07 lo sebhub 2012-08-03T05:41:15 * beng-nl came here inspired by the rtems-on-curiosity news 2012-08-03T05:41:28 that is some pretty amazing stuff 2012-08-03T05:42:51 *** WikL has joined #rtems 2012-08-03T05:50:00 *** WikL has quit IRC (Read error: Connection reset by peer) 2012-08-03T05:52:46 *** sh has joined #rtems 2012-08-03T05:53:45 *** krzysiekm13 has quit IRC (Ping timeout: 244 seconds) 2012-08-03T05:53:45 *** juli1 has quit IRC (Ping timeout: 244 seconds) 2012-08-03T05:53:55 *** sebhub has quit IRC (Read error: Connection reset by peer) 2012-08-03T05:53:58 *** arvind_khadri has quit IRC (Ping timeout: 240 seconds) 2012-08-03T05:54:14 *** WikL has joined #rtems 2012-08-03T05:55:33 *** sh is now known as Guest78628 2012-08-03T05:58:38 *** juli1 has joined #rtems 2012-08-03T06:34:07 *** Hesham has joined #rtems 2012-08-03T07:01:09 *** Hesham1 has joined #rtems 2012-08-03T07:02:10 *** arvind_khadri has joined #rtems 2012-08-03T07:03:38 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-08-03T07:10:15 *** WikL_ has joined #rtems 2012-08-03T07:11:32 *** WikL has quit IRC (Ping timeout: 248 seconds) 2012-08-03T07:16:43 *** Hesham1 has quit IRC (Quit: Leaving.) 2012-08-03T07:23:06 *** arvind_khadri has quit IRC (Ping timeout: 248 seconds) 2012-08-03T07:23:54 *** arvind_khadri has joined #rtems 2012-08-03T08:19:53 *** sevikkk has quit IRC (Ping timeout: 240 seconds) 2012-08-03T08:37:23 *** xiangfu has joined #rtems 2012-08-03T08:54:18 *** WikL_ has quit IRC (Ping timeout: 248 seconds) 2012-08-03T08:55:30 *** beng-nl is now known as beng-nl-1984 2012-08-03T08:55:39 *** beng-nl-1984 is now known as beng-nl-1984g 2012-08-03T08:56:08 *** WikL has joined #rtems 2012-08-03T08:56:56 *** antgreen has joined #rtems 2012-08-03T09:05:58 *** weiY has quit IRC (Ping timeout: 255 seconds) 2012-08-03T09:26:20 *** gedare has joined #rtems 2012-08-03T09:48:24 *** gedare has quit IRC (Quit: Leaving) 2012-08-03T10:16:49 *** Hesham has joined #rtems 2012-08-03T10:36:23 *** antgreen has quit IRC (Ping timeout: 245 seconds) 2012-08-03T10:42:07 *** WikL has quit IRC (Ping timeout: 240 seconds) 2012-08-03T10:43:57 *** WikL has joined #rtems 2012-08-03T10:44:56 *** Guest78628 has quit IRC (Remote host closed the connection) 2012-08-03T10:53:07 *** WikL has quit IRC (Ping timeout: 240 seconds) 2012-08-03T10:55:20 *** WikL has joined #rtems 2012-08-03T11:05:33 *** sevikkk has joined #rtems 2012-08-03T11:05:48 *** QingPei has joined #rtems 2012-08-03T11:13:05 *** beng-nl-1984g is now known as beng-nl-1984g-94 2012-08-03T11:13:12 *** beng-nl-1984g-94 is now known as beng-1984g-94kg 2012-08-03T11:25:29 *** beng-1984g-94kg is now known as beng-1984g 2012-08-03T11:26:19 *** Hesham1 has joined #rtems 2012-08-03T11:28:33 *** Hesham has quit IRC (Ping timeout: 246 seconds) 2012-08-03T11:30:23 *** sevikkk has quit IRC (Ping timeout: 244 seconds) 2012-08-03T11:30:48 *** Hesham1 has quit IRC (Client Quit) 2012-08-03T11:39:49 *** krzysiekm13 has joined #rtems 2012-08-03T12:22:14 *** deb has joined #rtems 2012-08-03T12:51:36 *** krzysiekm13 has left #rtems 2012-08-03T12:54:44 *** Hesham has joined #rtems 2012-08-03T12:55:25 *** Hesham1 has joined #rtems 2012-08-03T12:57:22 *** Guest6452 has joined #rtems 2012-08-03T12:58:53 *** Hesham has quit IRC (Ping timeout: 245 seconds) 2012-08-03T13:08:03 *** Guest6452 has quit IRC (Ping timeout: 245 seconds) 2012-08-03T13:32:39 *** peerst has quit IRC (*.net *.split) 2012-08-03T13:32:40 *** ChanServ has quit IRC (*.net *.split) 2012-08-03T13:32:40 *** juli1 has quit IRC (*.net *.split) 2012-08-03T13:32:40 *** rokka has quit IRC (*.net *.split) 2012-08-03T13:32:41 *** Hesham1 has quit IRC (*.net *.split) 2012-08-03T13:32:42 *** verm__ has quit IRC (*.net *.split) 2012-08-03T13:32:42 *** deb has quit IRC (*.net *.split) 2012-08-03T13:32:43 *** arvind_khadri has quit IRC (*.net *.split) 2012-08-03T13:32:43 *** lcpfnvcy has quit IRC (*.net *.split) 2012-08-03T13:32:44 *** QingPei has quit IRC (*.net *.split) 2012-08-03T13:32:44 *** WikL has quit IRC (*.net *.split) 2012-08-03T13:32:45 *** kristianpaul has quit IRC (*.net *.split) 2012-08-03T13:32:45 *** dr__house has quit IRC (*.net *.split) 2012-08-03T13:32:46 *** kuzew has quit IRC (*.net *.split) 2012-08-03T13:32:48 *** jahf has quit IRC (*.net *.split) 2012-08-03T13:32:48 *** kiwichris has quit IRC (*.net *.split) 2012-08-03T13:32:48 *** mikeg has quit IRC (*.net *.split) 2012-08-03T13:51:47 *** Hesham has joined #rtems 2012-08-03T13:51:47 *** peerst has joined #rtems 2012-08-03T13:51:47 *** verm__ has joined #rtems 2012-08-03T13:51:47 *** lcpfnvcy has joined #rtems 2012-08-03T13:51:47 *** ChanServ has joined #rtems 2012-08-03T13:51:47 *** rokka has joined #rtems 2012-08-03T13:51:47 *** juli1 has joined #rtems 2012-08-03T13:51:47 *** arvind_khadri has joined #rtems 2012-08-03T13:51:47 *** deb has joined #rtems 2012-08-03T13:51:47 *** QingPei has joined #rtems 2012-08-03T13:51:47 *** WikL has joined #rtems 2012-08-03T13:51:47 *** kuzew has joined #rtems 2012-08-03T13:51:47 *** dr__house has joined #rtems 2012-08-03T13:51:47 *** kristianpaul has joined #rtems 2012-08-03T13:51:47 *** kiwichris has joined #rtems 2012-08-03T13:51:47 *** jahf has joined #rtems 2012-08-03T13:51:47 *** mikeg has joined #rtems 2012-08-03T13:51:47 *** cameron.freenode.net sets mode: +o ChanServ 2012-08-03T13:56:46 *** WikL has quit IRC (Ping timeout: 255 seconds) 2012-08-03T13:58:38 *** WikL has joined #rtems 2012-08-03T14:07:14 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-03T14:42:42 *** Hesham has joined #rtems 2012-08-03T14:52:11 *** WikL has quit IRC (Remote host closed the connection) 2012-08-03T15:26:05 *** antgreen has joined #rtems 2012-08-03T16:06:21 *** krzysiekm13 has joined #rtems 2012-08-03T16:06:28 *** krzysiekm13 has left #rtems 2012-08-03T16:11:54 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-03T16:19:28 *** Hesham has joined #rtems 2012-08-03T19:16:59 *** kristianpaul has quit IRC (Quit: leaving) 2012-08-03T19:29:45 *** kristianpaul has joined #rtems 2012-08-03T19:29:45 *** kristianpaul has joined #rtems 2012-08-03T20:13:40 *** antgreen has quit IRC (Ping timeout: 246 seconds) 2012-08-03T22:22:03 *** Hesham1 has joined #rtems 2012-08-03T22:23:35 *** Hesham1 has quit IRC (Client Quit) 2012-08-03T22:25:43 *** Hesham has quit IRC (Ping timeout: 272 seconds) 2012-08-03T23:39:52 *** deb has left #rtems ("Leaving") 2012-08-03T23:44:03 *** arvind_khadri has quit IRC (Ping timeout: 244 seconds) 2012-08-04T00:18:35 *** arvind_khadri has joined #rtems 2012-08-04T03:55:48 *** Hesham has joined #rtems 2012-08-04T04:40:29 *** Hesham1 has joined #rtems 2012-08-04T04:42:10 *** Hesham has quit IRC (Ping timeout: 244 seconds) 2012-08-04T05:12:06 *** Hesham1 has quit IRC (Ping timeout: 264 seconds) 2012-08-04T05:17:25 *** Hesham has joined #rtems 2012-08-04T06:06:35 *** Hesham1 has joined #rtems 2012-08-04T06:08:30 *** Hesham has quit IRC (Ping timeout: 264 seconds) 2012-08-04T06:29:35 *** arvind_khadri has quit IRC (Ping timeout: 272 seconds) 2012-08-04T06:41:24 *** arvind_khadri has joined #rtems 2012-08-04T07:15:19 *** sevikkk has joined #rtems 2012-08-04T08:22:27 *** Hesham1 has quit IRC (Quit: Leaving.) 2012-08-04T10:58:43 *** Hesham has joined #rtems 2012-08-04T11:47:38 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-04T11:59:24 *** Hesham has joined #rtems 2012-08-04T12:31:30 *** QingPei has quit IRC (Quit: Leaving.) 2012-08-04T12:31:41 *** QingPei has joined #rtems 2012-08-04T12:54:15 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-04T12:58:13 *** Hesham has joined #rtems 2012-08-04T13:23:49 *** kuzew has quit IRC (Ping timeout: 252 seconds) 2012-08-04T13:25:27 *** kuzew has joined #rtems 2012-08-04T13:46:46 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-04T14:55:33 *** krzysiekm13 has joined #rtems 2012-08-04T15:08:06 *** Hesham has joined #rtems 2012-08-04T16:17:46 *** krzysiekm13 has left #rtems 2012-08-04T16:51:14 *** Hesham has quit IRC (Quit: Leaving.) 2012-08-04T17:35:54 lo all 2012-08-04T17:42:09 hi 2012-08-04T17:43:47 kiwichris! 2012-08-04T17:43:53 kiwichris: how nice to run into you here 2012-08-04T17:44:04 kiwichris: i am minix-ben 2012-08-04T17:58:25 hehe Hi 2012-08-04T17:58:41 How are you ? 2012-08-04T17:58:44 kiwichris: i came here inspired by the impending MSL landing 2012-08-04T17:59:05 Yeah it is getting close. 2012-08-04T17:59:07 kiwichris: on which i am reliably informed RTEMS is running .. on the radio(s) 2012-08-04T17:59:14 Must be about a day 2012-08-04T17:59:30 There are some posts on the mailing lists from Joel about this 2012-08-04T17:59:44 yeah, it's about 8am monday my time.. so.. in.. 31ish hours? 2012-08-04T17:59:48 It is similar to the set up in orbit around mars now 2012-08-04T18:00:12 kiwichris: i am okay thank you! 2012-08-04T18:00:18 how are you and/or RTEMS? 2012-08-04T18:00:30 *** beng-1984g is now known as beng-nl 2012-08-04T18:00:38 I am good working away on various things 2012-08-04T18:00:51 sounds familiar :) 2012-08-04T18:01:00 i'm watching the john carmack quakecon keynote 2012-08-04T18:01:00 I have been watching Minix. Well done on the shared lib support 2012-08-04T18:01:09 kiwichris: thank you! 2012-08-04T18:01:25 it's really nice to have that in now.. a big step (although we need more of them) 2012-08-04T18:01:26 That is a big thing for you. 2012-08-04T18:01:35 it is! 2012-08-04T18:01:51 I understand. Is threading being worked on ? 2012-08-04T18:02:36 it's just for the functionality now; but, i'm going to be working on more buffer cache / vm unification, and that combined with making the binaries dynamic will bring a multiplied number of benefits i believe :) 2012-08-04T18:03:03 We have an active file cache development. 2012-08-04T18:03:22 It is called bdbuf in RTEMS. A GSoC project is based around it 2012-08-04T18:03:43 The clock alg was added 2012-08-04T18:03:52 for ageing? 2012-08-04T18:04:06 Yes 2012-08-04T18:04:27 ARC is nice but has an IBM patient (see postgress for details) 2012-08-04T18:04:35 hmm i see 2012-08-04T18:04:59 Talked the postgress guy about this that at the lunch outside where the pic where taken last year 2012-08-04T18:05:11 at gsoc 2012-08-04T18:05:11 :) 2012-08-04T18:05:34 There is another alg that postgress uses that is ok 2012-08-04T18:05:39 hmmm 2012-08-04T18:06:05 Our cache is thread based with multiple output worked threads 2012-08-04T18:06:55 ARC looks good 2012-08-04T18:07:11 Yeah. Simple and effecting but cannot be used :( 2012-08-04T18:07:20 effective 2012-08-04T18:07:28 hmpf 2012-08-04T18:07:57 don't all software projects infringe on zillions of patents, realistically and also hypothetically speaking? 2012-08-04T18:08:01 http://wiki.rtems.org/wiki/index.php/Bdbuf_improvements 2012-08-04T18:08:41 Maybe but not intentionally 2012-08-04T18:09:19 the evaluation of the current prefetch strategy rings familiar ;-) 2012-08-04T18:10:20 LIRS, hmm 2012-08-04T18:10:31 Yeah that is the one in postgress 2012-08-04T18:10:48 i will read about it, i have an opportunity to use an eviction algorithm in the near future ;) 2012-08-04T18:11:00 which is to say, i have to choose a good one 2012-08-04T18:11:23 That would be nice. In RTEMS all this needs to be fast and small 2012-08-04T18:11:35 We have an AVL tree to hold the blocks 2012-08-04T18:12:41 This code is running in the MMS satellite 2012-08-04T18:13:16 * beng-nl secretly learns about what the MMS satellite is in order to pretend he naturally knows all about what it is 2012-08-04T18:13:35 ahhhh 2012-08-04T18:13:59 wow, yes, i'd read about it, it's an exciting measurement to do 2012-08-04T18:14:08 entirely magical how it would work 2012-08-04T18:15:47 kiwichris: any chance you will go to gsoc summit this year? 2012-08-04T18:15:55 it seems likely i will 2012-08-04T18:18:34 Yes I should be heading over 2012-08-04T18:19:17 great. 2012-08-04T18:19:33 The MMS has a large ram disk and runs this cache and the RFS file system on a 32MHz processor 2012-08-04T18:19:51 ah, RFS is yours right :) 2012-08-04T18:19:54 yeah 2012-08-04T18:20:13 wow that's amazing man. 2012-08-04T18:20:18 Getting it to perform that the key aim 2012-08-04T18:20:30 The cache is a major part of this 2012-08-04T18:20:52 is the cache behaviour real-timable? 2012-08-04T18:21:26 Realtime caches is what Joel did his PHD on, but I might sneak an answer in here :) 2012-08-04T18:22:28 A cache is not really real-time. It is threaded and that lets the user handle prorities how-ever the issues come down to the file syste 2012-08-04T18:22:41 and the driver structures 2012-08-04T18:23:01 For example a read request queued just after a write has started 2012-08-04T18:23:11 The disk is serial 2012-08-04T18:24:52 umm 2012-08-04T18:26:32 ok, i get that the disk serializes all i/o requests 2012-08-04T18:27:36 i was just imagining that the cache can be in so many states it's hard to make its performance predictable 2012-08-04T18:27:41 which is what i imagine real-time is all about 2012-08-04T18:27:52 as you can tell i'm not very versed in it ;) 2012-08-04T18:29:14 I will find the link with the statemachine pic, hang on 2012-08-04T18:29:27 ohh now we're talking 2012-08-04T18:33:52 http://www.rtems.org/onlinedocs/doxygen/cpukit/html/group__rtems__bdbuf.html#details 2012-08-04T18:34:23 * beng-nl clicks 2012-08-04T18:35:18 The cache allows a single memory pool to be used for different buffer sizes. For example mounting a disk may use 512 byte blocks and the file system will switch 4096 byte blocks 2012-08-04T18:35:31 ohhh 2012-08-04T18:36:10 the 2nd law of thermodynamics defied ;) 2012-08-04T18:36:15 The cache reuses the same memory. In RTEMS is important because the 512 byte blocks may only used for a moment then never again and having a dedicated pool is a waste 2012-08-04T18:36:39 :) 2012-08-04T19:08:41 Sunday morn here and about to head. Great to catch up. Drop by anytime. 2012-08-04T19:09:44 righto, it's roughly bedtime for me, thanks for chatting! 2012-08-04T19:09:49 i'll be around for a bit :) 2012-08-04T19:13:19 off to Jack Russell racing, yeap we race little mad dogs, not sure who is madder !! 2012-08-04T19:13:28 :-) 2012-08-04T19:13:38 i'll tell one of my buddies, he's nuts about jack russells 2012-08-04T20:22:42 *** Hesham has joined #rtems 2012-08-04T20:24:13 *** Hesham has quit IRC (Client Quit) 2012-08-04T23:51:57 *** peerst has quit IRC (Ping timeout: 260 seconds) 2012-08-04T23:52:52 *** peerst has joined #rtems 2012-08-05T02:09:42 *** arvind_khadri has quit IRC (Ping timeout: 244 seconds) 2012-08-05T02:25:10 *** QingPei has quit IRC (Quit: Leaving.) 2012-08-05T03:33:08 *** kristianpaul has quit IRC (Read error: Connection reset by peer) 2012-08-05T03:51:28 *** WikL has joined #rtems 2012-08-05T04:49:35 *** kristianpaul has joined #rtems 2012-08-05T04:49:35 *** kristianpaul has joined #rtems 2012-08-05T04:58:08 *** sevikkk has quit IRC (Ping timeout: 244 seconds) 2012-08-05T05:09:18 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-08-05T05:42:03 *** kristianpaul has joined #rtems 2012-08-05T05:42:03 *** kristianpaul has joined #rtems 2012-08-05T05:49:53 *** kristianpaul has quit IRC (Ping timeout: 272 seconds) 2012-08-05T06:07:37 *** WikL has quit IRC (Ping timeout: 272 seconds) 2012-08-05T06:08:55 *** WikL has joined #rtems 2012-08-05T07:17:28 *** sevikkk has joined #rtems 2012-08-05T07:32:49 *** arvind_khadri has joined #rtems 2012-08-05T07:38:31 *** arvind_khadri has quit IRC (Ping timeout: 246 seconds) 2012-08-05T09:07:08 *** kristianpaul has joined #rtems 2012-08-05T09:07:08 *** kristianpaul has joined #rtems 2012-08-05T09:27:51 *** kristianpaul has quit IRC (Read error: Connection reset by peer) 2012-08-05T09:30:36 *** kristianpaul has joined #rtems 2012-08-05T09:30:36 *** kristianpaul has joined #rtems 2012-08-05T10:10:32 *** arvind_khadri has joined #rtems 2012-08-05T10:12:08 *** arvind_khadri has joined #rtems 2012-08-05T10:13:21 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-08-05T10:13:44 *** arvind_khadri has joined #rtems 2012-08-05T10:15:26 *** arvind_khadri has quit IRC (Max SendQ exceeded) 2012-08-05T10:15:52 *** arvind_khadri has joined #rtems 2012-08-05T10:17:59 *** arvind_khadri has quit IRC (Client Quit) 2012-08-05T10:23:36 *** WikL has quit IRC (Remote host closed the connection) 2012-08-05T10:40:54 *** beng-nl has quit IRC (Ping timeout: 264 seconds) 2012-08-05T13:04:17 *** mikeg has quit IRC (*.net *.split) 2012-08-05T13:10:30 *** 50UABD8AM <50UABD8AM!~mikeg@unaffiliated/mikeg> has joined #rtems 2012-08-05T13:10:40 *** 50UABD8AM is now known as mikeg 2012-08-05T15:29:36 *** Hesham1 has joined #rtems 2012-08-05T16:12:51 *** Hesham1 has quit IRC (Ping timeout: 246 seconds) 2012-08-05T16:19:13 *** Hesham1 has joined #rtems 2012-08-05T16:33:06 *** Hesham1 has quit IRC (Ping timeout: 264 seconds) 2012-08-05T16:34:03 *** Hesham1 has joined #rtems 2012-08-05T17:13:12 *** Hesham1 has quit IRC (Quit: Leaving.) 2012-08-05T17:41:50 *** kristianpaul has quit IRC (Ping timeout: 248 seconds) 2012-08-05T17:43:30 *** kristianpaul has joined #rtems 2012-08-05T17:43:30 *** kristianpaul has joined #rtems 2012-08-05T20:05:33 *** xiangfu has joined #rtems 2012-08-05T23:22:54 Reminder to tune into NASA TV.