2010-08-02T00:01:29 *** kiwichris has quit IRC 2010-08-02T00:34:18 *** kiwichris has joined #rtems 2010-08-02T01:02:49 *** kiwichris has quit IRC 2010-08-02T02:05:53 *** sebhub has joined #rtems 2010-08-02T02:06:00 good morning 2010-08-02T02:11:02 *** projectgus has quit IRC 2010-08-02T02:30:59 *** mwalle has quit IRC 2010-08-02T02:33:24 *** mwalle has joined #rtems 2010-08-02T02:46:32 *** kiwichris has joined #rtems 2010-08-02T03:49:16 *** exception13 has joined #rtems 2010-08-02T05:00:32 *** madrazr has joined #rtems 2010-08-02T06:09:25 good noon 2010-08-02T06:10:24 *** methril_work has joined #rtems 2010-08-02T06:38:50 *** exception13 has quit IRC 2010-08-02T07:37:28 *** madrazr has quit IRC 2010-08-02T08:04:47 *** exception13 has joined #rtems 2010-08-02T08:09:13 *** exception13 has left #rtems 2010-08-02T08:21:11 *** kiwichris has quit IRC 2010-08-02T09:02:30 Hi 2010-08-02T09:02:43 can i build RTEMS bsp pc386 with my classic x86 toolchain ? 2010-08-02T09:02:56 or do i need rtems-4.11-386 toolchain ? 2010-08-02T09:07:29 *** zwj has joined #rtems 2010-08-02T09:13:06 I think you need the special toolchain 2010-08-02T09:17:03 ok 2010-08-02T09:17:05 compiling ... 2010-08-02T09:17:20 btw I think you can also start to have a look at USB 2010-08-02T09:17:31 it's a pain, but also a priority 2010-08-02T09:17:48 with USB we can have mouse + keyboard to be really able to use and develop flickernoise 2010-08-02T09:20:46 *** exception13 has joined #rtems 2010-08-02T09:34:22 does not compile ... 2010-08-02T09:35:21 http://pastebin.com/46FxMshk 2010-08-02T09:35:22 lekernel: an idea ? 2010-08-02T09:36:34 try to use the rpms 2010-08-02T09:36:47 humm each time i tried by the past, it always failed 2010-08-02T09:36:54 or ask on #opencores, they defend that autoconf is good 2010-08-02T09:36:59 i always compiled the toolchain after that 2010-08-02T09:37:28 compiling for lm32 is ok, for i386-rtems4.11 is not 2010-08-02T09:37:36 use the rpms... 2010-08-02T09:37:45 don't waste time on autocrap 2010-08-02T09:37:58 i wasted a lot of time with rpms too :) 2010-08-02T09:38:09 use rpm2cpio... it's easy 2010-08-02T09:38:14 which host os do you use? 2010-08-02T09:38:23 ubuntu 2010-08-02T09:39:15 what was the problem with the rpms? missing or incompatible shared libraries? 2010-08-02T09:39:47 just didn't install itself 2010-08-02T09:40:00 some of them was ok for installing, others won't 2010-08-02T09:40:05 don't remember the exact problem 2010-08-02T09:40:12 use rpm2cpio 2010-08-02T09:40:34 yes, the rtems toolchain installation is only copy from a to b 2010-08-02T09:41:12 lekernel: ok i will try 2010-08-02T09:43:12 ok binutils installed correctly 2010-08-02T09:52:59 configure: error: C compiler cannot create executables 2010-08-02T09:53:02 with the rpm one 2010-08-02T09:53:58 *** exception13 has left #rtems 2010-08-02T09:56:29 fallen@tablet:~/toolchains$ i386-rtems4.11-gcc main.c -o main 2010-08-02T09:56:30 /opt/rtems-4.11/libexec/gcc/i386-rtems4.11/4.5.1/cc1: error while loading shared libraries: libcloog.so.0: cannot open shared object file: No such file or directory 2010-08-02T09:58:07 simply install libcloog.so.0 2010-08-02T10:02:20 yes 2010-08-02T10:02:25 need a symbolic link 2010-08-02T10:03:35 *** exception13 has joined #rtems 2010-08-02T10:04:22 ok works better ): 2010-08-02T10:04:24 :) 2010-08-02T10:04:58 your normal libcloog.so did not have these links? 2010-08-02T10:05:12 it had several links 2010-08-02T10:05:17 but not this one 2010-08-02T10:05:22 there was several for -ppl 2010-08-02T10:06:09 maybe your libcloog has not the right version 2010-08-02T10:06:19 dunno, but it works now 2010-08-02T10:06:33 i didn't need this btw for lm32-gcc 2010-08-02T10:06:43 i guess i could have compiled without libcloog support 2010-08-02T10:25:12 *** gedare has joined #rtems 2010-08-02T10:26:10 sebhub:hi 2010-08-02T10:35:28 hi 2010-08-02T10:37:44 when do you have free time? we can plan to make a IRC meeting to discuss the project of system event api 2010-08-02T10:39:03 i am online this week from 10:00 up to 18:00 utc 2010-08-02T10:46:13 i have to leave now, bye 2010-08-02T10:46:35 *** sebhub has quit IRC 2010-08-02T10:58:47 *** zwj has quit IRC 2010-08-02T11:00:28 *** gedare has quit IRC 2010-08-02T11:40:51 *** exception13 has left #rtems 2010-08-02T11:55:43 *** madrazr has joined #rtems 2010-08-02T12:01:15 *** madrazr has joined #rtems 2010-08-02T12:07:00 *** madrazr has quit IRC 2010-08-02T12:15:05 *** madrazr has joined #rtems 2010-08-02T13:19:07 *** bubaflub has joined #rtems 2010-08-02T13:23:42 *** Alison_Chaiken has joined #rtems 2010-08-02T13:24:28 *** DrJoel has joined #rtems 2010-08-02T13:24:28 *** ChanServ sets mode: +o DrJoel 2010-08-02T13:34:38 *** madrazr has quit IRC 2010-08-02T14:00:30 *** Alison_Chaiken has quit IRC 2010-08-02T15:22:43 *** antgreen has joined #rtems 2010-08-02T15:55:09 *** antgreen has left #rtems 2010-08-02T15:57:03 *** gedare has joined #rtems 2010-08-02T16:02:08 hi drjoel 2010-08-02T16:03:19 hey gedare .. have a nice weekend? 2010-08-02T16:03:36 pretty nice, got a minute to PM? 2010-08-02T16:05:57 sure 2010-08-02T16:48:58 *** exception13 has joined #rtems 2010-08-02T17:04:31 *** bubaflub has quit IRC 2010-08-02T17:11:43 *** gedare has quit IRC 2010-08-02T17:31:38 *** DrJoel has quit IRC 2010-08-02T17:49:38 *** projectgus has joined #rtems 2010-08-02T18:35:16 *** exception13 has left #rtems 2010-08-02T18:41:12 *** methril_work has quit IRC 2010-08-02T22:50:53 *** kiwichris has joined #rtems 2010-08-02T23:13:43 *** methril_ has quit IRC 2010-08-02T23:25:58 *** methril_ has joined #rtems 2010-08-02T23:35:34 *** methril__ has joined #rtems 2010-08-02T23:37:24 *** methril_ has quit IRC 2010-08-03T01:32:37 *** sebhub has joined #rtems 2010-08-03T01:34:41 *** madrazr has joined #rtems 2010-08-03T01:42:04 good morning 2010-08-03T02:01:55 *** ckennelly has joined #rtems 2010-08-03T02:23:35 *** projectgus has quit IRC 2010-08-03T02:23:53 *** madrazr has quit IRC 2010-08-03T03:47:56 *** kiwichris has quit IRC 2010-08-03T04:10:48 *** exception13 has joined #rtems 2010-08-03T04:14:32 *** kiwichris has joined #rtems 2010-08-03T05:07:19 *** bazinski has joined #rtems 2010-08-03T05:41:19 *** bazinski has quit IRC 2010-08-03T05:45:50 *** bazinski has joined #rtems 2010-08-03T06:16:29 *** madrazr has joined #rtems 2010-08-03T06:46:46 *** methril_work has joined #rtems 2010-08-03T08:05:30 *** exception13 has left #rtems 2010-08-03T08:06:56 kiwichris, ping 2010-08-03T08:07:05 Hi 2010-08-03T08:07:33 i'm taking a look to the drivers code 2010-08-03T08:07:39 of the bdm tool 2010-08-03T08:07:46 Ok 2010-08-03T08:08:01 all OSs versions are updated? 2010-08-03T08:08:14 Sorry, I am not sure I understand 2010-08-03T08:09:25 I see there are freebsd, ioperm, linux, linux-2.6, openbsd, sco and win directories 2010-08-03T08:09:39 each one is for one OS (except ioperm) 2010-08-03T08:09:53 all them are working? 2010-08-03T08:09:59 Yes and I think all use the ioperms method 2010-08-03T08:10:12 the structure of the OSBDM drivers are so different 2010-08-03T08:10:32 I have no idea and I have only currently support ioperms. Other maybe be testing the other drivers. You would need to ask on the mailing list 2010-08-03T08:10:48 I have not look 2010-08-03T08:10:53 what's the main difference? 2010-08-03T08:10:56 do you know? 2010-08-03T08:11:03 Between what ? 2010-08-03T08:11:18 Anyone familiar with psim? I have a bug on i386-pc-solaris2.10 host to powerpc-rtems4.11 target where RAM_SIZE isn't detected properly when making the device tree. 2010-08-03T08:11:42 sorry, between for example ioperm and linux2.6 driver 2010-08-03T08:11:48 mikeg, Joel is the best person to ask about this. 2010-08-03T08:12:19 the problem is tests/psim looks for a line like "0x01000000 A ", but powerpc-rtems4.11-nm returns "0x01000000 A RamSize" 2010-08-03T08:12:23 okie kiwichris 2010-08-03T08:12:27 The ioperm is direct access to the hardware from user land. The driver hits the hardware from the kernel 2010-08-03T08:12:35 been difficult to catch him on irc recently :) 2010-08-03T08:13:06 ioperm is easiest to debug then :) 2010-08-03T08:13:15 methril_work, yes 2010-08-03T08:13:26 mikeg, I would send an email to the list. 2010-08-03T08:13:29 it would be nice to unify the usb part? 2010-08-03T08:13:51 yeah, working on it ;) 2010-08-03T08:13:56 methril_work, no idea. The libusb is used. 2010-08-03T08:14:17 Is that what OSBDM uses ? 2010-08-03T08:14:24 I am off for the night. Bye all 2010-08-03T08:14:25 *** kiwichris has quit IRC 2010-08-03T08:14:37 they check for OSBDM interfaces, and TBLCF 2010-08-03T08:20:18 *** zwj has joined #rtems 2010-08-03T08:21:24 For those interested, the problem was /usr/bin/awk on Solaris doesn't like the command "\" whereas /usr/xpg4/bin/awk works ok 2010-08-03T08:31:57 I guess that probably isn't worth a PR :) "Solaris2.10 awk is broken and 3 years old. Workaround: install GNU awk or use usable awk in /usr/xpg4/bin/awk" 2010-08-03T08:42:40 *** DrJoel has joined #rtems 2010-08-03T08:42:40 *** ChanServ sets mode: +o DrJoel 2010-08-03T08:43:18 hiya DrJoel 2010-08-03T08:45:51 i was saying earlier that the psim script doesn't work with /usr/bin/awk in Solaris2.10 (RAM_SIZE isn't detected), but using /usr/xpg4/bin/awk or installing gawk work 2010-08-03T08:46:33 psim uses awk?!? Who added something that uses awk? I don't speak awk and wrote that script LOL 2010-08-03T08:47:01 RAM_SIZE=`$NM ${1} | awk '/\/{print "0x"$1}'` 2010-08-03T08:47:44 oh... I wonder who added that and when.. can you rewrite it as a sed? I would be happy to apply that as a patch. No reason to assume awk is present. 2010-08-03T08:47:57 For that matter, you could probably do that with grep and cut 2010-08-03T08:48:14 okie 2010-08-03T08:48:52 Does simplifying how that is done make sense to you? Doesn't requiring awk seem heavy for just that use? 2010-08-03T08:49:46 I mean if I'm the only one that's ever run into an issue on not having a valid awk install.. 2010-08-03T08:49:50 and it was an easy workaround 2010-08-03T08:50:54 No.. better to fix it. I have fixed a number of places where the default shell on a system is not bash. 2010-08-03T08:51:19 My thing is that finding what was broken was the issue 2010-08-03T08:52:12 That's another issue I had actually. /bin/sh didn't work with ./boostrap, I had to use bash 2010-08-03T08:53:01 That would be a Ralf issue.. feel free to file a PR. We are all pretty picky about not wanting host issues like that to cause problems but since most of us are on Linux systems, it is hard to be perfect 2010-08-03T08:53:28 BSD users like peerst find issues. And Ubuntu doesn't have bash as default shell 2010-08-03T09:03:47 *** bubaflub has joined #rtems 2010-08-03T09:05:31 anyways, DrJoel, here's a simple patch for grep/cut.. sorry i don't have a CVS tree to diff against http://pastie.org/1072960 2010-08-03T09:10:13 ohhhh.. you are using the version in libbsp.. did you know about the rtems-testing CVS module? It includes a lot of helpful scripts including much improved versions of all similulator scripts 2010-08-03T09:10:24 it looks like Strauman checked that in last September 2010-08-03T09:10:41 yes i've used that module.. didn't know it had new versions of psim though 2010-08-03T09:11:11 everything in there is fairly new (or improved) to make it easier to run automated tests. Will fix this in CVS anyway 2010-08-03T09:11:57 Yeah, the only thing I've used from rtems-testing was gcc/do_one to learn to how compile RTEMSAda 2010-08-03T09:13:56 (which I'm now in the middle of a j_gccada equivalent for gcc-4.5.1 RC). i was getting an rtems/endian.h error earlier on gcc-4.4.4. fingers crossed this build works :) 2010-08-03T09:14:01 I collected all the simulator scripts into one framework under sim-scripts so they had the same interface across all simulators.. makes it easier to automate testing 2010-08-03T09:14:23 * DrJoel needs to file a PR for gcc/Ada to get that errno issue committed 2010-08-03T09:29:02 DrJoel: if i want to use Ada with both psim and beatnik, can i just use --enable-rtemsbsp="beatnik psim" then CFLAGS_FOR_TARGET="-B${PREFIX}/${TARGET}/beatnik/lib/ -B${PREFIX}/${TARGET}/psim/lib/" ? 2010-08-03T09:29:40 The --enable-rtemsbsp is OK. But you only build for one BSP at a time. Where is the CFLAGS going? 2010-08-03T09:29:56 You have to build a BSP when you actually compile and link an executable 2010-08-03T09:30:09 for j_gccada, when I'm building GCC. do you mean i need to build two GCC executables? 2010-08-03T09:32:00 no.. you shouldn't have to build two gcc's... 2010-08-03T09:32:19 I mean when I'm actually making an RTEMS binary, I would use RTEMS_MAKEFILE_PATH to link to only one BSP. But, when I'm building GNAT, it needs to find the RTEMS headers 2010-08-03T09:32:35 You should only specify one BSP on a CFLAGS_FOR_TARGET.. it will give conflicting sets of libraries... 2010-08-03T09:33:02 the headers should be there from the multilib build. We should only need the bsp to link tests.. (I could be surprised). 2010-08-03T09:33:12 Does it matter which one? And if I use, for example, beatnik in CFLAGS_FOR_TARGET, could I use the resulting gnatmake to compile a psim binary? 2010-08-03T09:33:26 I am hoping this is an artifact of the old old way of building 2010-08-03T09:33:49 you should be able to as long as you don't end up forcing some -mcpu type option down to the library build which is incompatible 2010-08-03T09:34:08 Ok, I'll get back to you on the results 2010-08-03T09:59:39 *** zwj has quit IRC 2010-08-03T10:01:50 DrJoel: I investigated the error I was getting earlier with powerpc-rtems4.11-gnatlink (Exception name: SYSTEM.ASSERTIONS.ASSERT_FAILURE Message: make.adb:7473). For some reason -mcpu=7400 is specified twice, and if I remove one it compiles fine 2010-08-03T10:02:47 And the binary works fine on my MVME6100, and a psim build, when the same fix is applied, works as well 2010-08-03T10:03:27 grrr... that's a stupid error.. file that with gcc... should be able to duplicate that one on Linux. Sounds like a generic error.. even if it is acceptable to flag it as an error, it should have a better message 2010-08-03T10:04:05 * mikeg is scared :X never filed a GCC PR 2010-08-03T10:04:05 Wonderful for working executable! Did you try to run the ACATS? 2010-08-03T10:04:27 no, i'll be happy to test it out when i can though :) 2010-08-03T10:05:24 don't be scared.. just give clear explanation and reproducibility. If you can make it fail on linux, then it is WAY MORE LIKELY to get addressed. cc me on it.. 2010-08-03T10:05:54 Ok, will do 2010-08-03T10:32:37 *** sebhub has quit IRC 2010-08-03T11:22:04 anyone knows how to register a isr in powerpc new-exceptions? 2010-08-03T11:23:55 peerst: there should be multiple examples in the libbssp/powerpc/shared/clock* 2010-08-03T11:28:50 In /c/src/lib/libbsp/powerpc/shared/clock/clock.c ? 2010-08-03T11:29:21 Hmm it looks like its using the decrementer for clock 2010-08-03T11:30:53 But it looks like tha call ppc_exc_set_handler() registers it ... grepping for this ... 2010-08-03T11:42:25 hmm it somehow looks like the different mpc types have different isr registering functions?? 2010-08-03T11:43:11 like mpc55xx_interrupt_handler_install() ... 2010-08-03T11:44:35 right.. on different cpu models, it can use different clock sources. So you are (I think) looking for a non-decrementer example 2010-08-03T11:47:22 yes 2010-08-03T11:48:22 in my last project we still used the old exception handling 2010-08-03T11:56:12 *** bubaflub has quit IRC 2010-08-03T11:57:43 I hope it has finally gone away :D I hate the terms old and new 2010-08-03T11:58:10 since the model varies some by CPU arch, I am trying to encourage folks to call them "simple vectored" and "pic vectored". 2010-08-03T11:58:20 But the powerpc took SO-O-O long to convert. :( 2010-08-03T11:59:25 *** madrazr has quit IRC 2010-08-03T12:13:07 yeah it took a while 2010-08-03T12:14:37 and the project that uses the "old" exception handling is still active (its having its own bsp and hacked the rtems build into something like bare bsp) 2010-08-03T12:23:36 *** bubaflub has joined #rtems 2010-08-03T12:24:25 RTEMS projects never die. We actually offer support for 4.6 and I have done bug fix releases for legacy support customers 2010-08-03T12:26:49 Yeah, and porting to newer versions is only done rarely, only if features of newer versions are needed 2010-08-03T12:44:37 or if you update hardware and need a newer compiler 2010-08-03T12:44:55 but we know that and offer support/bug fixes 2010-08-03T12:54:20 *** tuxmaniac has joined #rtems 2010-08-03T13:29:54 updating hardware might actually happen soon in this project, then we have to give up all the bare bsp ish stuff and move on 2010-08-03T13:34:16 is it a cpu model that is already supported? 2010-08-03T13:34:40 yeah probably 2010-08-03T13:34:53 its not selected yet 2010-08-03T13:34:54 but might be some arm 2010-08-03T13:35:09 thomas will be building the hardware 2010-08-03T13:35:18 and selecting the cpu 2010-08-03T14:07:17 well that narrows it down a little... they have been doing arm work lately 2010-08-03T14:14:08 *** tuxmaniac has quit IRC 2010-08-03T14:18:15 *** ckennelly has left #rtems 2010-08-03T14:25:50 *** bubaflub has quit IRC 2010-08-03T14:25:56 *** bubaflub has joined #rtems 2010-08-03T15:05:52 DrJoel: so I'm running the testsuite with psim, and the command line run_all_rtems.sh uses is `psim -C -l 180 $exe` which fails because -C is not an option, and the -l 180 causes it to hang. running the testsuite with those disabled is generating "PASS"'s 2010-08-03T15:06:47 Actually, I was using /opt/rtems-4.11/powerpc-rtems4.11/psim/tests/psim instead of ~/rtems-testing/bin/psim 2010-08-03T15:07:13 psim takes those when you use the psim from rtems-testing. :-D 2010-08-03T15:08:56 Do you know Rob Pettit? i was explaining that i was almost finished, but that a gcc dev wanted me to run the testsuite, and he asked if it was Joel Sherril of Mitre :p 2010-08-03T15:17:20 actually I do some work with their robotics group and two from Mitre and I are doing a tutorial at SIGAda this year. I was asked if he could ask me questions. I haven't heard any from him though. Are you workign with him? 2010-08-03T15:22:42 he's my manager :) 2010-08-03T15:33:07 LOL.. I had a feeling it was a small world.. well I was asked if he could contact me after you and I had already made a lot of progress 2010-08-03T16:24:56 *** DrJoel has quit IRC 2010-08-03T16:36:23 *** bubaflub has quit IRC 2010-08-03T16:42:46 *** bubaflub has joined #rtems 2010-08-03T16:52:57 *** kiwichris has joined #rtems 2010-08-03T17:08:08 *** kiwichris has quit IRC 2010-08-03T17:25:23 *** kiwichris has joined #rtems 2010-08-03T17:26:13 *** methril_work has quit IRC 2010-08-03T17:26:39 *** methril_work has joined #rtems 2010-08-03T17:30:03 *** kiwichris has quit IRC 2010-08-03T17:35:32 *** kiwichris has joined #rtems 2010-08-03T17:59:53 *** kiwichris has quit IRC 2010-08-03T18:21:08 *** methril__ has quit IRC 2010-08-03T18:21:19 *** methril__ has joined #rtems 2010-08-03T18:49:06 *** projectgus has joined #rtems 2010-08-03T20:17:59 *** bazinski has quit IRC 2010-08-03T20:24:10 *** kiwichris has joined #rtems 2010-08-03T20:28:15 *** bubaflub has quit IRC 2010-08-03T20:35:46 *** bazinski has joined #rtems 2010-08-03T21:19:49 *** DrJoel has joined #rtems 2010-08-03T21:19:49 *** ChanServ sets mode: +o DrJoel 2010-08-03T21:20:00 * DrJoel waves hi 2010-08-03T21:24:39 *** kiwichris has quit IRC 2010-08-03T21:25:34 *** kiwichris has joined #rtems 2010-08-03T21:25:40 Hi Joel 2010-08-03T21:25:59 hello.. 2010-08-03T21:26:04 got a minute? 2010-08-03T21:26:07 Yeap 2010-08-03T21:50:48 *** DrJoel has quit IRC 2010-08-03T21:58:57 *** bubaflub has joined #rtems 2010-08-03T22:33:08 Hi bob 2010-08-03T22:48:23 hello kiwichris 2010-08-03T22:48:31 how ya doing? 2010-08-03T22:48:37 Looks like you are making great progress. 2010-08-03T22:48:49 thanks. i'm a bit nervous that i won't hit the end goal before the GSoC term is up, though. 2010-08-03T22:49:01 I am good. I can create scripts that should build RTEMS tools on MacOS 2010-08-03T22:49:07 kiwichris: nice! 2010-08-03T22:49:21 i was looking at some stuff that automates VirtualBox development environments 2010-08-03T22:49:25 I can generate the scripts from spec files, but they have bugs in them 2010-08-03T22:49:47 Which bit of virtual box are you looking at ? 2010-08-03T22:50:15 i used Virtual Box for my dev environment, i was looking at a program called Vagrant that automates the replication and deploying of Virtual Box environments 2010-08-03T22:50:59 Ah ok. Does this makes a image than you install and you have a setup working ? 2010-08-03T22:52:21 On the build system changes, the relative vs absolute src path can effect debug info and finding the source within a debugger 2010-08-03T22:56:53 ok, yeah 2010-08-03T22:56:56 right now i have absolute paths 2010-08-03T22:56:58 should i use relative? 2010-08-03T22:57:05 and Vagrant would automate setting up dev images 2010-08-03T22:57:14 it'd be cool to have some stuff where the BSPs and RTEMS are already running 2010-08-03T22:57:18 lowers the barrier to entry 2010-08-03T22:57:37 I would attempt to mimic the path to configure.pl if you can 2010-08-03T22:57:46 ok, i can 2010-08-03T22:57:56 Great 2010-08-03T23:48:10 *** kiwichris has quit IRC 2010-08-04T00:25:03 *** bubaflub has quit IRC 2010-08-04T00:30:44 *** kiwichris has joined #rtems 2010-08-04T00:44:08 *** bubaflub has joined #rtems 2010-08-04T00:45:12 *** bubaflub has quit IRC 2010-08-04T01:33:35 *** bubaflub has joined #rtems 2010-08-04T01:40:18 *** bubaflub has quit IRC 2010-08-04T01:42:50 *** th_d has joined #rtems 2010-08-04T02:05:05 *** sebhub has joined #rtems 2010-08-04T02:05:20 good morning 2010-08-04T02:18:28 *** projectgus has quit IRC 2010-08-04T02:59:47 *** bazinski has quit IRC 2010-08-04T03:22:59 *** bazinski has joined #rtems 2010-08-04T03:40:33 *** exception13 has joined #rtems 2010-08-04T04:03:32 *** lcpfnvc has joined #rtems 2010-08-04T04:22:44 sebhub, hi 2010-08-04T04:22:56 hi chris :-) 2010-08-04T04:23:06 hi Thomas. 2010-08-04T04:23:18 Wondering is a bus_space chat could happen. 2010-08-04T04:23:25 Wondering if a bus_space chat could happen. 2010-08-04T04:23:56 The technical part is mainly covered by Sebastian, but as always I have my opinions ;-) 2010-08-04T04:24:54 My view is this is not a big deal. I may be wrong but lets have a look 2010-08-04T04:26:34 The busspace concept is very versatile. It tries to solve the same things the libchip abstration tries to do, but is more general 2010-08-04T04:26:45 abstraction... 2010-08-04T04:27:31 Yes I think we should have something in RTEMS in the score/cpu level that can be mapped to the bus_space API or the parts we need 2010-08-04T04:28:17 Hm, bus space must be adaptable in the BSP level. 2010-08-04T04:28:18 e.g 2010-08-04T04:28:19 For every multilib variant there is a common way IO is accessed (for getting the i386 io space) 2010-08-04T04:28:28 *** kuzew has quit IRC 2010-08-04T04:28:30 no! 2010-08-04T04:28:35 wait... 2010-08-04T04:29:12 in many situations, all HW will access a peripheral in the same way, e.g. IO-mapped with proper, non-gapped addresses 2010-08-04T04:29:34 but the bus space concept allows the BSP to provide modified access rules 2010-08-04T04:30:08 like accessing I/O registers at non-standard starting addresses, with non-standard spacing between registers 2010-08-04T04:30:46 or even indirect accesses like accessing registers over a different/serial bus system 2010-08-04T04:30:55 Sure, but I see this as a sparse array of the define IO unit 2010-08-04T04:31:14 The device size is the sum of all addresses on the bus used. 2010-08-04T04:31:19 although not all boards will use this flexibility, we should provide a way to wkeep it available 2010-08-04T04:31:46 and ideally all would collapse to simple I/O accesses in the "standard cases". 2010-08-04T04:31:58 Yes 2010-08-04T04:32:03 Hm, not sure whether my comments are clear eanough??? 2010-08-04T04:32:20 I get some but I suspect not all 2010-08-04T04:32:56 *** kuzew has joined #rtems 2010-08-04T04:33:09 *** bazinski has quit IRC 2010-08-04T04:33:36 I am not seeing how the BSP vs the CPU method of accessing a specific bus can vary. 2010-08-04T04:34:03 Hm, an artifical example: 2010-08-04T04:34:14 Assume you have a nasty 16550 serial device 2010-08-04T04:34:19 Yeap 2010-08-04T04:34:20 With its standard register set 2010-08-04T04:34:24 Yeap 2010-08-04T04:34:33 and a driver for that, quite similar to the libchip one 2010-08-04T04:34:49 Hang here 2010-08-04T04:34:50 Now if the device is not directly attached to the CPU bus 2010-08-04T04:35:13 but only accessible over a communication channel (e.g. via special USB messages) 2010-08-04T04:35:14 I assume we are not including struct mapped devices ? 2010-08-04T04:36:11 then with the abstrace "bus space" model, you can use the existing driver, and the BSP simply provides functions to read/write registers, no matter whether they are I/O mapped or must be passed via USB to the external hardware. 2010-08-04T04:36:23 struct mapped devices are possible aswell. 2010-08-04T04:36:43 brb 2010-08-04T04:36:46 Ups, no they are not 2010-08-04T04:38:52 They might be, in a ways. 2010-08-04T04:39:23 If the accesses to the struct fields are with specialbus space function calls. 2010-08-04T04:39:34 on the phone :) 2010-08-04T04:39:39 ok:-) 2010-08-04T04:43:12 splitting the usb stack into two directory trees (and thus diverges from the freebsd directory layout) will make update from freebsd more difficult 2010-08-04T04:47:44 Sorry about that. 2010-08-04T04:48:49 Thomas, the struct comments confuse me. A mapped struct is compiler controlled. Any other access is a directed access. 2010-08-04T04:50:34 I think we need to aim at a solution that covers most devices. Knowing hardware engineers they dream up special cases but we I think a full generic solution is difficult. 2010-08-04T04:50:47 sorry, phone ;-) 2010-08-04T04:50:51 :) 2010-08-04T04:51:44 sebhub, why would we split the code ? 2010-08-04T04:52:05 you cannot use in cpukit 2010-08-04T04:52:14 and device drivers need 2010-08-04T04:53:00 Is the driver layer the host controller code ? 2010-08-04T04:53:33 although this is only a minor problem, since the dependent files are not part of freebsd 2010-08-04T04:53:57 Yes and I was working towards this :) 2010-08-04T04:54:19 I do not agree with the bsp.h being need by drivers. This is something in RTEMS that could be changed. 2010-08-04T04:54:56 If we consider the "typical" bus interfaces only and leave the special cases to one side. 2010-08-04T04:55:10 the big issue here is bus_space(9) and bus_dma(9) 2010-08-04T04:55:43 Yes and if we can agree on a solution then this smooths the path. 2010-08-04T04:56:42 The DMA interface is a little harder. 2010-08-04T04:56:45 it is very difficult to avoid BSP dependent suff in drivers, in the end you will end up with APIs like in freebsd for resource managment 2010-08-04T04:57:20 Yes and if this gives us the ability to include a USB stack then that is great. 2010-08-04T04:57:57 If we make RTEMS need BSP inside the tree to build then we have moved backwards 2010-08-04T04:58:14 you can include the usb stack in the current structure outside the cpukit 2010-08-04T04:58:59 Lets ask a different questions. 2010-08-04T04:59:20 What CPU is the stack running on ? 2010-08-04T04:59:29 sh@eb0011:~/svn/rtems-apps/libusb > gg '.*include.*bsp' . 2010-08-04T04:59:31 ./rtems/freebsd/machine/rtems-bsd-cache.h:#include 2010-08-04T04:59:32 ./rtems/freebsd/dev/usb/controller/ehci_mpc83xx.c:#include 2010-08-04T04:59:33 ./rtems/freebsd/dev/usb/controller/ehci_mpc83xx.c:#include 2010-08-04T04:59:35 ./rtems/freebsd/dev/usb/controller/ohci_lpc32xx.c:#include 2010-08-04T04:59:37 ./rtems/freebsd/dev/usb/controller/ohci_lpc32xx.c:#include 2010-08-04T04:59:38 ./rtems/freebsd/dev/usb/controller/ohci_lpc32xx.c:#include 2010-08-04T04:59:40 ./rtems/freebsd/dev/usb/controller/ohci_lpc24xx.c:#include 2010-08-04T04:59:41 ./rtems/freebsd/dev/usb/controller/ohci_lpc24xx.c:#include 2010-08-04T04:59:42 ./rtems/freebsd/dev/usb/controller/ohci_lpc24xx.c:#include 2010-08-04T04:59:44 ./rtems/freebsd/dev/usb/controller/ohci_lpc24xx.c:#include 2010-08-04T04:59:46 ./rtems/freebsd/dev/usb/controller/ohci_lpc32xx-tyco-2010-07-22.c:#include 2010-08-04T04:59:47 ./rtems/freebsd/dev/usb/controller/ohci_lpc32xx-tyco-2010-07-22.c:#include 2010-08-04T04:59:48 ./rtems/freebsd/dev/usb/controller/ohci_lpc32xx-tyco-2010-07-22.c:#include 2010-08-04T04:59:50 these are the parts that depend on the BSP 2010-08-04T04:59:52 ARM and PowerPC 2010-08-04T05:00:02 Great and thanks. 2010-08-04T05:00:26 all these files are not part of freebsd 2010-08-04T05:00:43 If I have 20 different boards that support this device, how many would require different bus_space mapping code ? 2010-08-04T05:00:53 and the include file is only used by bus_dam(9) 2010-08-04T05:00:55 On the same processor. 2010-08-04T05:01:52 My point is the point of abstraction of what is needed can be more common and related to the CPU than a BSP. 2010-08-04T05:01:53 the bus_space can be used to map pci space to cpu space for example (usb host controller on pci bus) 2010-08-04T05:02:02 Yes. 2010-08-04T05:02:38 There is BSP specific code to set the hardware up but the actual access instructions would be the same. 2010-08-04T05:04:44 ok, i have no particular tendency on this issue, but some has to work out all the details, implement it, etc. this needs a budget 2010-08-04T05:05:14 Yeap. We need a possible technical path. 2010-08-04T05:06:03 If we assume a suitable technical solution would the rest of the code go into cpukit ok > 2010-08-04T05:06:05 If we assume a suitable technical solution would the rest of the code go into cpukit ok ? 2010-08-04T05:06:56 if we assume we have a bus_space and bus_dma inside the cpukit everything else can go inteo the cpukit too 2010-08-04T05:07:12 except the low level device drivers 2010-08-04T05:07:15 Ok. That is great. It bound the problem. 2010-08-04T05:07:29 Sure 2010-08-04T05:08:35 If this is the case then may I say well done on the port. 2010-08-04T05:09:35 the freebsd people did it very well 2010-08-04T05:10:32 80% of the usb stack related work went into pure rtems issues 2010-08-04T05:11:02 Are these issues fixed in CVS ? 2010-08-04T05:11:04 like timer server, interrupt server, bdbuf, file system 2010-08-04T05:11:33 i am mostly in sync with the cvs head 2010-08-04T05:11:43 i did not fix all issues 2010-08-04T05:11:58 like unmount of file systems in use 2010-08-04T05:12:24 Ah yes, this is a known problem 2010-08-04T05:12:45 And there are other related issues. 2010-08-04T05:12:52 a careless unmount can corrupt the system in subtile ways, very hard to debug 2010-08-04T05:13:19 *** bazinski has joined #rtems 2010-08-04T05:13:47 have to take a meal 2010-08-04T05:13:58 need to check home work 2010-08-04T05:24:07 *** bazinski has quit IRC 2010-08-04T05:28:57 *** bazinski has joined #rtems 2010-08-04T05:39:23 *** bazinski has quit IRC 2010-08-04T05:42:12 hi, what are the options to run the hello.exe for pc386 bsp ? 2010-08-04T05:42:24 i tried with and without -nographic and -serial stdio 2010-08-04T05:42:31 i never got the "hello world" message 2010-08-04T05:42:57 but in graphic mode i can see the "initialized console port CONSOLE." 2010-08-04T05:43:10 (iam using qemu) 2010-08-04T05:45:34 you may have a look at the rtems-testing module 2010-08-04T05:45:40 cvs -d :pserver:anoncvs@www.rtems.com:/usr1/CVS co -P rtems-testing 2010-08-04T05:51:52 *** Guest34745 has joined #rtems 2010-08-04T05:54:22 ok thanks 2010-08-04T06:26:48 *** madrazr has joined #rtems 2010-08-04T06:41:31 sebhub: what can i find in this source tree to help me ? 2010-08-04T06:45:03 there is a sim-scripts directory, you can use these files to run various bsps on a simulator 2010-08-04T06:47:13 ok thanks :) 2010-08-04T06:48:21 hum ok seems i must generate an image out of the .exe 2010-08-04T06:56:06 sebhub: i have found the parameters to pass to qemu in the pc386.in script 2010-08-04T06:56:10 but not how to generate the image 2010-08-04T06:56:16 * Fallenou it still searching in it 2010-08-04T06:58:51 *** lekernel has quit IRC 2010-08-04T07:07:51 *** lekernel has joined #rtems 2010-08-04T07:08:24 the bsps should generate a hello.ralf 2010-08-04T07:08:46 use this instead of hello.exe, look at the makefile output for the hello sample 2010-08-04T07:19:49 *** exception13 has left #rtems 2010-08-04T07:24:19 ok i got this : http://www.rtems.com/wiki/index.php/QEMU :) 2010-08-04T07:38:45 *** madrazr has quit IRC 2010-08-04T07:40:21 *** kiwichris has quit IRC 2010-08-04T07:41:25 *** lekernel has quit IRC 2010-08-04T07:47:12 ok it works, i run the hello :) 2010-08-04T07:49:14 Can someone check if this is the accepted testresults format? First testsuite email. http://gcc.gnu.org/ml/gcc-testresults/2010-08/msg00334.html 2010-08-04T07:50:13 *** antgreen has joined #rtems 2010-08-04T07:52:45 *** lekernel has joined #rtems 2010-08-04T08:08:11 *** Guest34745 has quit IRC 2010-08-04T08:09:33 *** antgreen has quit IRC 2010-08-04T08:28:18 *** antgreen has joined #rtems 2010-08-04T08:37:12 *** tuxmaniac has joined #rtems 2010-08-04T08:47:44 *** antgreen has quit IRC 2010-08-04T08:57:32 *** bubaflub has joined #rtems 2010-08-04T09:06:41 *** |th_d| has joined #rtems 2010-08-04T09:11:07 *** th_d has quit IRC 2010-08-04T09:21:25 *** zwj has joined #rtems 2010-08-04T09:37:17 *** _th_d_ has joined #rtems 2010-08-04T09:39:43 *** |th_d| has quit IRC 2010-08-04T09:51:15 *** sebhub has quit IRC 2010-08-04T10:50:43 *** madrazr has joined #rtems 2010-08-04T10:54:48 *** antgreen has joined #rtems 2010-08-04T11:30:29 *** zwj has quit IRC 2010-08-04T11:44:49 *** Alison_Chaiken has joined #rtems 2010-08-04T11:52:19 *** _th_d_ has quit IRC 2010-08-04T11:58:43 *** th_d has joined #rtems 2010-08-04T12:41:25 *** aniceberg has joined #rtems 2010-08-04T13:00:59 *** madrazr has quit IRC 2010-08-04T13:18:23 *** aniceberg has quit IRC 2010-08-04T13:22:17 i am using rtems-testing scripts to run an application of pc386 bsp and i get this error message : 2010-08-04T13:22:24 isa irq 9 already assigned 2010-08-04T13:22:39 i am following this "how to" : http://www.rtems.com/wiki/index.php/QEMU 2010-08-04T13:23:20 i am trying to run a TCP/IP application, i guess this isa irq has to do with the network adaptater :) 2010-08-04T13:29:50 *** aniceberg has joined #rtems 2010-08-04T13:34:49 *** madrazr has joined #rtems 2010-08-04T13:45:58 *** madrazr has quit IRC 2010-08-04T13:48:49 *** tuxmaniac has quit IRC 2010-08-04T14:09:57 it seems the bsp pc386 compiles with the DEC21 nic by default, how to compile with ne2000 ? 2010-08-04T15:57:44 *** th_d has quit IRC 2010-08-04T16:02:31 *** bubaflub has quit IRC 2010-08-04T16:32:31 *** aniceberg has quit IRC 2010-08-04T18:34:50 *** projectgus has joined #rtems 2010-08-04T18:40:23 *** krzabr has joined #rtems 2010-08-04T19:25:02 *** alinrus has joined #rtems 2010-08-04T19:35:59 *** Alison_Chaiken has quit IRC 2010-08-04T20:14:21 *** antgreen has quit IRC 2010-08-04T20:20:21 *** alinrus has quit IRC 2010-08-04T20:23:45 *** krzabr has quit IRC 2010-08-04T20:31:08 *** antgreen has joined #rtems 2010-08-04T20:46:49 *** alinrus has joined #rtems 2010-08-04T21:23:30 *** alinrus has quit IRC 2010-08-04T21:49:30 *** methril_ has joined #rtems 2010-08-04T21:50:08 *** methril_work has quit IRC 2010-08-04T21:50:33 *** methril_work has joined #rtems 2010-08-04T21:50:44 *** methril__ has quit IRC 2010-08-05T00:22:58 *** antgreen has quit IRC 2010-08-05T00:44:50 *** kiwichris has joined #rtems 2010-08-05T01:44:11 *** methril__ has joined #rtems 2010-08-05T01:45:37 *** methril_ has quit IRC 2010-08-05T01:56:15 *** sebhub has joined #rtems 2010-08-05T01:56:20 good morning 2010-08-05T02:18:24 *** projectgus has quit IRC 2010-08-05T02:27:33 *** madrazr has joined #rtems 2010-08-05T02:38:49 *** madrazr has quit IRC 2010-08-05T02:43:03 *** madrazr has joined #rtems 2010-08-05T02:43:53 *** alinrus has joined #rtems 2010-08-05T02:52:15 *** madrazr has quit IRC 2010-08-05T03:07:23 *** madrazr has joined #rtems 2010-08-05T03:30:30 sebhub, hi 2010-08-05T04:07:48 *** kiwichris has quit IRC 2010-08-05T04:33:04 *** alinrus1 has joined #rtems 2010-08-05T04:33:04 *** alinrus has quit IRC 2010-08-05T04:42:02 *** madrazr has quit IRC 2010-08-05T04:42:04 *** madrazr1 has joined #rtems 2010-08-05T04:52:38 *** alinrus has joined #rtems 2010-08-05T04:52:38 *** alinrus1 has quit IRC 2010-08-05T05:46:35 *** aniceberg has joined #rtems 2010-08-05T05:57:14 *** exception13 has joined #rtems 2010-08-05T06:11:03 *** alinrus1 has joined #rtems 2010-08-05T06:11:03 *** alinrus has quit IRC 2010-08-05T06:23:00 *** madrazr1 has quit IRC 2010-08-05T06:31:37 *** madrazr has joined #rtems 2010-08-05T06:34:58 *** antgreen has joined #rtems 2010-08-05T08:22:49 *** zwj has joined #rtems 2010-08-05T08:23:34 *** alinrus1 has quit IRC 2010-08-05T08:30:56 *** alinrus has joined #rtems 2010-08-05T08:38:33 sebhub:hi 2010-08-05T09:06:53 hi 2010-08-05T09:07:43 can you ask Joel about what to do with the sysinit api patches? 2010-08-05T09:12:30 *** tuxmaniac has joined #rtems 2010-08-05T09:16:17 ok, i will send email to him 2010-08-05T09:16:59 and about the system event, what do you think about the chris' email 2010-08-05T09:30:52 *** exception13 has left #rtems 2010-08-05T09:38:46 ok, thanks 2010-08-05T09:39:14 if think the mail from chris is a good starting point 2010-08-05T09:44:24 ok, it is really a good idea, but there is some confused part about event domain 2010-08-05T09:45:21 what is the cause of the confusement? 2010-08-05T09:47:42 he means that system event api replace the pending_event field with a array. and the array contains the event domain. if the different event domain can have the same event bit? 2010-08-05T09:50:32 i think he suggests to replace the 32-bit event registers with a table of 32-bit values with at least one entry 2010-08-05T09:50:38 each entry is an event domain 2010-08-05T09:51:34 hm, not really a table, some sort of storage 2010-08-05T09:53:09 if i can think that every domain contain a 32-bit values event sets, the difference is the entry 2010-08-05T09:53:45 currently each task has some storage for events (RTEMS_API_Control) 2010-08-05T09:54:19 if we create an event domain we create storage for events independent of tasks 2010-08-05T09:58:55 now the events set is allocated to a particular system task? sorry,i donont really understand your second message 2010-08-05T10:00:25 to process the event send/receive you need some storage for your state, this storage is currently bound to the task control block (Thread_Control) 2010-08-05T10:01:53 yeah, i see . you mean replace the storage from task control block to event domain 2010-08-05T10:04:42 *** alinrus has left #rtems 2010-08-05T10:06:22 yes, at least this is how i interpret the mail from chris 2010-08-05T10:06:36 *** tuxmaniac has quit IRC 2010-08-05T10:06:56 i have to leave now, i am back tomorrow, good evening 2010-08-05T10:07:02 but the event domain also in task control blok 2010-08-05T10:07:07 ok, see you 2010-08-05T10:07:15 *** sebhub has quit IRC 2010-08-05T10:37:24 *** exception13 has joined #rtems 2010-08-05T11:25:25 *** zwj has quit IRC 2010-08-05T11:29:25 *** koss has joined #rtems 2010-08-05T11:29:39 *** koss is now known as Guest9523 2010-08-05T11:34:27 *** Alison_Chaiken has joined #rtems 2010-08-05T12:13:49 *** madrazr has quit IRC 2010-08-05T12:16:37 *** Guest9523 has quit IRC 2010-08-05T12:32:35 *** madrazr has joined #rtems 2010-08-05T12:46:32 *** kiwichris has joined #rtems 2010-08-05T12:52:57 *** alinrus has joined #rtems 2010-08-05T12:55:43 *** tuxmaniac has joined #rtems 2010-08-05T13:18:16 *** aniceberg has quit IRC 2010-08-05T13:30:40 *** aniceberg has joined #rtems 2010-08-05T13:33:04 *** tuxmaniac has quit IRC 2010-08-05T13:34:12 *** madrazr1 has joined #rtems 2010-08-05T13:34:16 *** madrazr has quit IRC 2010-08-05T13:59:59 *** DrJoel has joined #rtems 2010-08-05T13:59:59 *** ChanServ sets mode: +o DrJoel 2010-08-05T14:20:15 *** alinrus1 has joined #rtems 2010-08-05T14:23:28 *** alinrus has quit IRC 2010-08-05T14:31:35 *** madrazr1 has quit IRC 2010-08-05T15:51:33 *** alinrus1 has quit IRC 2010-08-05T15:55:07 *** alinrus has joined #rtems 2010-08-05T15:57:25 *** antgreen has quit IRC 2010-08-05T15:57:44 *** antgreen has joined #rtems 2010-08-05T18:29:42 *** projectgus has joined #rtems 2010-08-05T18:31:13 *** DrJoel has quit IRC 2010-08-05T18:40:20 *** alinrus has quit IRC 2010-08-05T18:55:28 *** exception13 has left #rtems 2010-08-05T19:42:01 *** Alison_Chaiken has quit IRC 2010-08-05T20:20:10 *** aniceberg has quit IRC 2010-08-06T01:13:17 *** aniceberg has joined #rtems 2010-08-06T01:13:32 *** antgreen has quit IRC 2010-08-06T01:17:52 *** aniceberg has quit IRC 2010-08-06T01:50:25 *** alinrus has joined #rtems 2010-08-06T01:53:31 *** alinrus has quit IRC 2010-08-06T02:32:19 *** projectgus has quit IRC 2010-08-06T02:36:50 *** sebhub has joined #rtems 2010-08-06T02:36:55 good morning 2010-08-06T02:37:05 *** alinrus has joined #rtems 2010-08-06T02:42:49 *** kiwichris has quit IRC 2010-08-06T03:43:25 *** exception13 has joined #rtems 2010-08-06T03:48:24 *** alinrus has quit IRC 2010-08-06T04:00:57 *** kiwichris has joined #rtems 2010-08-06T04:14:40 *** madrazr has joined #rtems 2010-08-06T04:27:56 *** madrazr has left #rtems 2010-08-06T04:29:02 *** alinrus has joined #rtems 2010-08-06T05:59:45 *** kiwichris has quit IRC 2010-08-06T06:42:28 *** antgreen has joined #rtems 2010-08-06T06:43:16 *** alinrus has quit IRC 2010-08-06T06:52:46 *** exception13 has left #rtems 2010-08-06T06:55:28 *** tuxmaniac has joined #rtems 2010-08-06T06:57:43 *** methril_work has quit IRC 2010-08-06T07:03:26 *** methril_work has joined #rtems 2010-08-06T07:14:35 *** tuxmaniac has quit IRC 2010-08-06T07:35:32 *** dr__house has joined #rtems 2010-08-06T08:04:28 *** dr__house has quit IRC 2010-08-06T08:23:51 *** antgreen has quit IRC 2010-08-06T08:46:59 *** DrJoel has joined #rtems 2010-08-06T08:46:59 *** ChanServ sets mode: +o DrJoel 2010-08-06T09:01:33 *** bubaflub has joined #rtems 2010-08-06T09:01:39 *** bubaflub has left #rtems 2010-08-06T09:18:32 *** madrazr has joined #rtems 2010-08-06T09:21:06 *** bubaflub has joined #rtems 2010-08-06T09:33:00 *** alinrus has joined #rtems 2010-08-06T09:33:19 *** dr__house has joined #rtems 2010-08-06T09:36:13 *** madrazr has quit IRC 2010-08-06T09:40:03 *** madrazr has joined #rtems 2010-08-06T09:44:23 *** madrazr has quit IRC 2010-08-06T10:00:25 *** madrazr has joined #rtems 2010-08-06T10:02:37 *** sebhub has quit IRC 2010-08-06T10:13:14 *** madrazr has quit IRC 2010-08-06T10:23:45 *** zwj has joined #rtems 2010-08-06T10:48:52 DrJoel:hi 2010-08-06T10:49:24 zwj: hello how are things going? 2010-08-06T10:50:04 first project is basiclly done, and so i want to know what to do with these patches 2010-08-06T10:50:49 https://www.rtems.org/bugzilla/show_bug.cgi?id=1593 2010-08-06T10:51:00 make sure the PR is up to date, all tests on at least one BSP pass. Add instructions on what we have to do to other BSPs. 2010-08-06T10:51:39 make sure Chris Johns is on the PR. Then email the user list that your code is ready for review and incorporation. That makes it open to public review 2010-08-06T10:51:39 *** exception13 has joined #rtems 2010-08-06T10:51:51 yeah, all patch is based on sis bsp. and it passed 2010-08-06T10:52:06 Once the review phase is over and comments addressed, either Chris or Sebastian can commit it. 2010-08-06T10:52:21 ok.. than just email the list and start the review phase 2010-08-06T10:53:14 yeah, chris is also on PR. and Sebastian also think it should be on the revies phase 2010-08-06T10:53:30 email to where? use-list? 2010-08-06T10:54:38 but my patch is made based on the CVS about 2010.06.25 2010-08-06T11:06:07 rtems-users@... 2010-08-06T11:06:22 Then update your tree and verify the patch is still OK. This is part of the final review.. 2010-08-06T11:25:49 *** madrazr has joined #rtems 2010-08-06T11:27:42 ok, i will make new patches 2010-08-06T11:28:28 *** Alison_Chaiken has joined #rtems 2010-08-06T11:37:49 *** zwj has quit IRC 2010-08-06T11:41:53 *** zwj has joined #rtems 2010-08-06T12:14:12 *** DrJoel has quit IRC 2010-08-06T12:21:37 *** exception13 has left #rtems 2010-08-06T12:30:14 *** zwj has quit IRC 2010-08-06T12:49:46 *** kristianpaul has joined #rtems 2010-08-06T12:56:23 *** exception13 has joined #rtems 2010-08-06T13:07:09 *** aniceberg has joined #rtems 2010-08-06T13:18:24 *** aniceberg has quit IRC 2010-08-06T13:30:44 *** aniceberg has joined #rtems 2010-08-06T13:47:42 *** Alison_Chaiken has quit IRC 2010-08-06T14:14:11 *** madrazr has quit IRC 2010-08-06T14:56:44 *** antgreen has joined #rtems 2010-08-06T14:59:11 *** madrazr has joined #rtems 2010-08-06T15:05:55 *** madrazr has quit IRC 2010-08-06T15:37:13 *** dr__house has quit IRC 2010-08-06T15:59:51 *** bubaflub has quit IRC 2010-08-06T16:29:26 *** kiwichris has joined #rtems 2010-08-06T16:46:47 *** lekernel has quit IRC 2010-08-06T16:52:07 *** aniceberg has quit IRC 2010-08-06T17:12:11 *** DrJoel has joined #rtems 2010-08-06T17:12:11 *** ChanServ sets mode: +o DrJoel 2010-08-06T17:12:22 DrJoel, hi 2010-08-06T17:12:35 hey chris.. how's the weekend looking there? 2010-08-06T17:12:45 Fine for a change which is nice, 2010-08-06T17:12:53 still very hot here.. 2010-08-06T17:12:57 always busy with kids sport 2010-08-06T17:13:08 I should have brought this up before.. what is left for 4.10? 2010-08-06T17:13:47 I am not sure. Are all PR's correctly labelled 4.10 for those that effect it ? 2010-08-06T17:13:59 If so a search for 4.10 will show you. 2010-08-06T17:14:24 I don't know.. AFAIK there is just the object initialization one (sp70) that I would like to see fixed (unfortunately up to you ) and Ralf's "warnings to fix" 2010-08-06T17:15:08 Oh I had forgotten about that one. I will take a look this weekend. 2010-08-06T17:15:15 according to that, there are 50 PRs.. that's not right at all. :( 2010-08-06T17:15:29 No, they are not correctly labelled. 2010-08-06T17:19:12 1526 is Ralf's PR 2010-08-06T17:19:29 *** methril_work has quit IRC 2010-08-06T17:20:01 Just looking 2010-08-06T17:20:18 Some look like they should be fixed in 4.10.. others need to be moved to 4.11 2010-08-06T17:21:05 They should be moved to HEAD or something else. This way we do not have to move them again 2010-08-06T17:21:24 Or to 4.11.99 2010-08-06T17:21:32 then this label can be changed 2010-08-06T17:21:35 Add HEAD as a category and make it the default. Even 4.11.9 will be wrong one day 2010-08-06T17:22:00 The only problem with HEAD is users do not understand it. I have been down this path before 2010-08-06T17:22:10 Ahh.. ok.. 4.11.99 it is 2010-08-06T17:22:29 1574 is the sp70/coverity one 2010-08-06T17:22:33 Ok This way what the code prints is present 2010-08-06T17:22:45 The object init one ? 2010-08-06T17:22:53 yep 2010-08-06T17:22:57 Ok. 2010-08-06T17:24:05 Hmmm, it is the milestone that is the problem 2010-08-06T17:24:22 What does this get called ? 2010-08-06T17:24:38 what does what get called? 2010-08-06T17:25:09 The milestone for something we have as a bug that has no specific release to be set 2010-08-06T17:25:16 OPEN ? 2010-08-06T17:25:27 On going 2010-08-06T17:25:31 HELP :) 2010-08-06T17:25:53 That's what I was suggesting either "none" or "head" 2010-08-06T17:26:07 ping on 1505 2010-08-06T17:26:08 But HEAD is not a milestone :) 2010-08-06T17:26:21 It is a version 2010-08-06T17:26:25 it is rtems_libio_is_valid_perms being useless 2010-08-06T17:26:28 then not assigned 2010-08-06T17:26:38 Great and done. 2010-08-06T17:27:29 ok.. i will try to triage some this weekend .. if you do the same, maybe we can have a candidate set to close for 4.10.0 2010-08-06T17:28:01 I am still getting MacOS tools working :( Very close. 2010-08-06T17:28:31 Maybe 1hr of work and I should have a clean build of any arch on MacOS 2010-08-06T17:28:52 I will work on the PRs and then start on them. 2010-08-06T17:29:10 Once I have builds running I have time to sort through them. 2010-08-06T17:29:28 Please email any lists you want to focus on and with any priority. 2010-08-06T17:29:41 I have 1574 as top, then 1505. 2010-08-06T17:29:46 https://www.rtems.org/bugzilla/show_bug.cgi?id=1320 2010-08-06T17:30:03 Has a milestone of "2". I can't change it 2010-08-06T17:30:14 Ok 2010-08-06T17:30:37 I was thinking of closing it and asking Ralf to reopen another if still applicable. Any other suggestion? 2010-08-06T17:31:41 What do we want with Tool milestones ? The same as RTEMS ? 2010-08-06T17:31:49 I think so.. seems right 2010-08-06T17:32:42 Try now 2010-08-06T17:34:16 Thanks. 2010-08-06T17:35:33 I have to head off and get ready for kids sport 2010-08-06T17:35:43 *** kiwichris has quit IRC 2010-08-06T17:42:46 *** DrJoel has quit IRC 2010-08-06T18:01:46 *** kiwichris has joined #rtems 2010-08-06T18:11:01 *** kiwichris has quit IRC 2010-08-06T18:37:34 *** alinrus has quit IRC 2010-08-06T18:53:49 *** alinrus has joined #rtems 2010-08-06T19:51:00 *** exception13 has left #rtems 2010-08-06T22:06:09 *** antgreen has quit IRC 2010-08-06T22:20:09 *** kiwichris has joined #rtems 2010-08-06T22:34:08 *** mikeg has quit IRC 2010-08-06T22:34:15 *** mikeg has joined #rtems 2010-08-06T23:05:38 *** alinrus has quit IRC 2010-08-07T01:21:38 *** methril__ has quit IRC 2010-08-07T01:24:01 *** methril__ has joined #rtems 2010-08-07T01:38:12 *** madrazr has joined #rtems 2010-08-07T01:47:43 *** madrazr has quit IRC 2010-08-07T02:05:25 *** madrazr has joined #rtems 2010-08-07T02:05:41 *** madrazr1 has joined #rtems 2010-08-07T02:10:02 *** madrazr has quit IRC 2010-08-07T02:11:08 *** tuxmaniac has joined #rtems 2010-08-07T03:06:13 *** koss has joined #rtems 2010-08-07T03:06:30 *** koss is now known as Guest68388 2010-08-07T03:06:36 *** tuxmaniac has quit IRC 2010-08-07T03:08:21 *** kiwichris has quit IRC 2010-08-07T03:15:22 *** lekernel has joined #rtems 2010-08-07T03:20:34 *** Guest68388 has quit IRC 2010-08-07T03:21:56 *** methril_ has joined #rtems 2010-08-07T03:22:28 *** methril__ has quit IRC 2010-08-07T04:01:05 *** madrazr has joined #rtems 2010-08-07T04:08:20 *** madrazr has quit IRC 2010-08-07T06:12:57 *** exception13 has joined #rtems 2010-08-07T08:23:37 *** kiwichris has joined #rtems 2010-08-07T08:31:11 *** kiwichris has joined #rtems 2010-08-07T08:33:24 *** kiwichris has joined #rtems 2010-08-07T08:59:23 *** DrJoel has joined #rtems 2010-08-07T08:59:23 *** ChanServ sets mode: +o DrJoel 2010-08-07T08:59:30 DrJoel, hi 2010-08-07T08:59:54 helllo.. just waking up here.. slept late on this Saturday and just started the coffee 2010-08-07T09:20:26 *** kiwichris has quit IRC 2010-08-07T09:55:17 lekernel: your student needs to start submitting their work as a PR for review and merge, updating the wiki, etc 2010-08-07T09:55:43 Fallenou: sorry for not seeing you there.. file a PR with your BSP and let's get a review underway 2010-08-07T09:57:11 hi :) 2010-08-07T09:57:31 ok will do the PR this afternoon 2010-08-07T09:57:37 getting near pencils down :) 2010-08-07T09:57:47 please make sure you have ChangeLog entry 2010-08-07T09:58:01 there is a wiki entry to describe how to use the bsp : http://www.milkymist.org/wiki/index.php?title=Run_RTEMS_on_the_Milkymist_One 2010-08-07T09:58:14 do i need to do another one in rtems wiki ? or just put a link to this one 2010-08-07T09:58:28 There needs to be one in the RTEMS Wiki. We have a page for each supported BSP and a master page which lists all. 2010-08-07T09:58:44 It is OK if the page references that or cuts and pastes that one 2010-08-07T09:59:31 Hi, Q: about using chains, this code http://ideone.com/kuejt does it make sense? 2010-08-07T09:59:35 ok 2010-08-07T10:00:11 for having a chain that sends an event when it was empty and is not anymore? 2010-08-07T10:01:17 peerst: we need to meet again for Chinese food. :-D 2010-08-07T10:01:52 Oh yeah! You are here in August aren't you? 2010-08-07T10:01:59 peerst: I think it makes sense but is the code protected in anyway? 2010-08-07T10:02:26 hmm no, I thought the chain ops are protected? 2010-08-07T10:02:35 no unfortunately not.. there was a class scheduled for July which didn't make the minimum.. need to talk to Thomas about the fall but don't want to be a burden given what's happening 2010-08-07T10:02:55 yeah I understand 2010-08-07T10:03:12 individually with interrupts but not between statements.. how about this 2010-08-07T10:03:13 disable 2010-08-07T10:03:26 if empty sendEvent = true 2010-08-07T10:03:31 append_unprotected 2010-08-07T10:03:34 enable 2010-08-07T10:03:45 if sendEvent rtems_event_send 2010-08-07T10:04:01 I think empty check is cheaper than for 1 node 2010-08-07T10:04:22 but could be wrong.. probably both are 2 memory references and compares 2010-08-07T10:05:04 yeah disabling through the 2 calls would be playing safe 2010-08-07T10:05:04 why the "iobm" name.. I can see this being useful.. but even more so with the "system events" not tied to a task that should be there after GSOC ends 2010-08-07T10:05:36 the file is called io-buf-mgr.c => prefix iobm 2010-08-07T10:05:41 I think those would be great operations to add to the User Level Chain .. but pass in a task id 2010-08-07T10:06:07 we could call it append_with_notification( chain, node, tid, event ) .. very generic for use 2010-08-07T10:06:37 and only notify if it gets from empty to nonempty? 2010-08-07T10:07:01 and non-empty to empty... basically you have the core of a producer/consumer :-D 2010-08-07T10:07:03 for my application I have a fixed setup of tasks with one or mor input buffer queues 2010-08-07T10:07:22 yeah sure that what I'm using it for 2010-08-07T10:07:24 that's cool.. just thinking in terms of adding it as part of User Chain services 2010-08-07T10:07:43 btw I like the chains 2010-08-07T10:07:46 have you seen PR1647? I think your review would be welcomed 2010-08-07T10:08:15 Thanks .. they have been there since the beginning of RTEMS. I think the overlapping nodes in the chain_control is unusual 2010-08-07T10:08:16 did overlook that ther is a API for using them, did only know them from debugging in the core 2010-08-07T10:08:33 the overlapping nodes trick is way cool 2010-08-07T10:08:38 it is a recent addition in the history of RTEMS. 2010-08-07T10:08:53 yeah there werent there when I last looked ;-) 2010-08-07T10:09:08 sebhub: brought them to my attention 2010-08-07T10:09:32 I wish I could find the articile I first saw a ref to that in. It was Byte back in the late 80's/each 90s and didn't have code.. just had a sentence or two on Amiga or Commodore and said they did that.. I wrote our version based upon that paragraph 2010-08-07T10:10:13 do you think the disable chechempty append sendevent enable is faster performancewise? 2010-08-07T10:10:16 PR1647 refactors thread to add Scheduler handler to core. And adds support for pluggable schedulers -- current, EDF, 2010-08-07T10:11:04 empty versus check one is probably the same but using unprotected is better... If it doesn't say "unprotected" then it probably disables interrupts... so at least you only do that once 2010-08-07T10:11:15 where do i find the PR's on the website? 2010-08-07T10:11:24 https://www.rtems.org/bugzilla/show_bug.cgi?id=1647 2010-08-07T10:11:42 bugzilla.rtems.org or there should be an obvious link on the top page now that we are using Drupal 2010-08-07T10:11:42 ah thanks 2010-08-07T10:12:09 this is an important step to having an SMP RTEMS since you need an alternate scheduler implementation on SMP 2010-08-07T10:12:18 I already added per_cpu structure to RTEMS 2010-08-07T10:12:59 cool 2010-08-07T10:13:12 I'm using rtems currently on a mpc5517 2010-08-07T10:13:21 which in theory has two cores 2010-08-07T10:13:41 using only one currently 2010-08-07T10:13:41 SMP is coming along slowly... a lot of the infrastructure is in place in a private tree of mine.. just a matter of time.. which getting paid helps to create :) 2010-08-07T10:14:39 Moving to SMP is always a big uphill way for an OS ... watched several BSD's doing it 2010-08-07T10:15:18 does the toolchain support compressed opcodes for powerpc? 2010-08-07T10:15:43 I can already initialize 4 core Leon3 system and qemu with 1-32 cores... have first cut at OS->BSP iface.. need scheduler and SMP safe critical section implementation 2010-08-07T10:16:30 not the last time I checked.. We were asked about supporting that a year or so ago and I said "not a problem if there are tools" and it turned out.. no gcc at that point 2010-08-07T10:16:50 Ask at Freescale to see if they have a GNU tools port.. then push on them to submit 2010-08-07T10:17:43 DrJoel: the thing is i used rtems cvs head of the time when i started the port, and created a git repo inside of it, and dumped all the CVS files 2010-08-07T10:18:15 i didn't know you had a git mirror 2010-08-07T10:18:32 i tried several time to use it and it was not working recently 2010-08-07T10:18:51 i think that will make the patch generation a bit tricky isn't it ? 2010-08-07T10:18:57 its no hurry with the secod core for me, its for future performance needs if the arise 2010-08-07T10:19:11 s/the/they/ 2010-08-07T10:19:13 Fallenou: Then (I think) you need to check the head out again. Make a single patch from your local git and get that applied to CVS checkout. You will have to check the merge as it progresses and that is the best way 2010-08-07T10:19:30 Fallenou: does our git work now? 2010-08-07T10:19:38 hum don't know 2010-08-07T10:19:41 what is the address again ? 2010-08-07T10:19:47 DrJoel: btw you scared my girlfriend almost to fleeing the house 2010-08-07T10:20:03 Fallenou: Can't you do a diff from baseline to end of your development in git? 2010-08-07T10:20:28 hum i guess yes, i don't know how to do that but i can search over the web =) 2010-08-07T10:20:31 peerst: when? At the last dinner? I am harmless.. how did I scare her? I am just the "short grey haired American" as Ralf describes me 2010-08-07T10:20:37 we were just entering the house and you were "speaking" with an eerie voice from the cellar 2010-08-07T10:20:43 about 20min ago 2010-08-07T10:21:12 Hmmm... my voice... RTEMS Moodle? 2010-08-07T10:21:12 I was playing with the voice option of my irc client 2010-08-07T10:21:26 this morning 2010-08-07T10:21:27 oooo.. .now that would make me creepy 2010-08-07T10:21:42 DrJoel: but in my git history i have commits that remove CVS files 2010-08-07T10:21:50 i guess that would hurt =) 2010-08-07T10:21:51 thought I was setting a announcement that you are here 2010-08-07T10:22:12 but it transcribed your irc lines instead 2010-08-07T10:22:36 so a alien computer voice was speaking out of the cellar :-) 2010-08-07T10:22:57 Fallenou: so... git tracks your changes against real files.. so you should be able to do that.. 2010-08-07T10:23:23 Fallenou: worst case.. tar up the libbsp/lm32/milkymist and move it to a fresh CVS checkout.. bootstrap, compile, and deal with issues by hand 2010-08-07T10:23:37 yes i think i will go for the last solution 2010-08-07T10:23:43 Fallenou: if you would use mercurial instead of git; you could use the mq extension for patch submission 2010-08-07T10:23:48 most of your changes should be under libcpu or libbsp in a contained way 2010-08-07T10:24:36 yes they are, i just changed some makefiles , one thing in the libcpu , all the other stuff is new files in lm32/shared and lm32/milkymist 2010-08-07T10:25:16 and new tests in testsuites/samples but i guess you do not want them 2010-08-07T10:25:19 or at least not there 2010-08-07T10:25:21 Fallenou: ok.. then should be easy to merge by hand.. that's our goal for BSPs. .. make sure no tabs in .[ch], no printf's, etc 2010-08-07T10:25:31 What are they 2010-08-07T10:25:51 tests, to test framebuffer driver, and network driver 2010-08-07T10:26:01 like painting the screen different colours 2010-08-07T10:26:17 *** alinrus has joined #rtems 2010-08-07T10:26:18 sending an arp frame, doing a mini telnet server 2010-08-07T10:26:25 s/arp/udp/ 2010-08-07T10:26:36 specific to the lm32 or generic? 2010-08-07T10:26:50 generic 2010-08-07T10:27:15 file a PR with the tests and a description of each. We can discuss what to do with them there 2010-08-07T10:27:25 alinrus: feeling better? 2010-08-07T10:27:27 ok 2010-08-07T10:28:10 actually yes 2010-08-07T10:28:20 alinrus: if I may ask, what was wrong? 2010-08-07T10:30:21 did you managed to look over the code? 2010-08-07T10:30:35 i hope i can get the notification this weekend 2010-08-07T10:31:43 I meant to email you but haven't yet.. sorry.. please file a PR with AIO changes and another with LIO changes.. Then send email to RTEMS Users list that the code is in PR X and Y to review.. I will try to get to it. Reviewing Gedare's code now. 2010-08-07T10:32:34 ok, will do 2010-08-07T10:32:58 will send the lio changes after i finish the notification 2010-08-07T10:34:16 should i attach the patches or just point people to the repo? 2010-08-07T10:34:26 not a problem.. I think we all are a bit overwhelmed as GSOC winds down.. I ended up being very involved in about 1/2 of the projects and feel bad that you haven't gotten the time you deserve. Please hang in there with us. We will get your code reviewed and merged. 2010-08-07T10:35:10 Please attach the patches in a diff form against the cvs head. They need to be easy to apply. Incldue ChangeLog entries, tests, etc. When merged, we will want 100% test coverage. So we will iterate a bit 2010-08-07T10:38:21 *** alinrus has quit IRC 2010-08-07T11:00:46 *** alinrus has joined #rtems 2010-08-07T11:32:05 *** KV has joined #rtems 2010-08-07T11:40:18 *** tuxmaniac has joined #rtems 2010-08-07T12:09:27 *** tuxmaniac has quit IRC 2010-08-07T12:26:52 https://www.rtems.org/bugzilla/show_bug.cgi?id=1656 2010-08-07T13:23:39 *** exception13 has left #rtems 2010-08-07T13:41:48 *** madrazr has joined #rtems 2010-08-07T13:42:50 *** madrazr has quit IRC 2010-08-07T14:00:57 *** KV has quit IRC 2010-08-07T15:21:49 DrJoel: https://www.rtems.org/bugzilla/show_bug.cgi?id=1657 2010-08-07T15:22:45 right after you alinrus ;) 2010-08-07T16:03:35 * DrJoel feels like he is getting buried in patches to review LOL... 2010-08-07T16:03:45 Saturday evening here.. .give me some time .. will get them 2010-08-07T17:46:07 hehe :) 2010-08-07T17:46:20 sorry to give you such a hard time ! 2010-08-07T19:06:01 no problem.. just trying to be efficient, thorough and fair 2010-08-07T19:42:23 hello 2010-08-07T19:43:04 There is a study case of about usin rtems in a mobile device, like PDA or soemthing similar 2010-08-07T19:43:07 ? 2010-08-07T19:45:42 DrJoel, i understand that because of the single process model the notion of processes is a bit fuzzy on rtems, so for example if i have to use a sigqueue which takes the argument as pid_t. can i use pthread_self() rather than getpid() and obtain the same result? 2010-08-07T19:50:16 wow rtems had been since 1988 :O ? 2010-08-07T19:57:08 *** alinrus has quit IRC 2010-08-07T20:10:56 *** alinrus has joined #rtems 2010-08-07T21:37:11 *** alinrus has quit IRC 2010-08-07T21:51:59 *** alinrus has joined #rtems 2010-08-07T22:22:53 *** alinrus has quit IRC 2010-08-07T22:38:25 *** alinrus has joined #rtems 2010-08-07T22:57:49 *** alinrus has quit IRC 2010-08-07T23:03:44 *** methril_ has quit IRC 2010-08-07T23:05:26 *** methril_ has joined #rtems 2010-08-07T23:12:02 *** alinrus has joined #rtems 2010-08-07T23:57:35 *** alinrus has quit IRC 2010-08-08T00:11:46 *** alinrus has joined #rtems 2010-08-08T00:28:17 *** alinrus has quit IRC 2010-08-08T00:42:54 *** alinrus has joined #rtems 2010-08-08T01:01:17 *** alinrus has quit IRC 2010-08-08T01:16:40 *** alinrus has joined #rtems 2010-08-08T01:31:36 *** alinrus has quit IRC 2010-08-08T01:32:57 *** madrazr has joined #rtems 2010-08-08T01:46:24 *** alinrus has joined #rtems 2010-08-08T02:01:34 *** alinrus has quit IRC 2010-08-08T02:06:51 *** madrazr has quit IRC 2010-08-08T02:16:24 *** alinrus has joined #rtems 2010-08-08T02:31:09 *** alinrus has quit IRC 2010-08-08T02:45:49 *** alinrus has joined #rtems 2010-08-08T03:01:01 *** alinrus has quit IRC 2010-08-08T03:44:39 *** exception13 has joined #rtems 2010-08-08T04:18:21 DrJoel: do i need to mail about the PR somewhere ? 2010-08-08T05:00:58 *** alinrus has joined #rtems 2010-08-08T06:44:51 *** lekernel has quit IRC 2010-08-08T08:08:54 *** alinrus has quit IRC 2010-08-08T08:23:26 *** alinrus has joined #rtems 2010-08-08T08:56:17 *** methril_ is now known as methril 2010-08-08T09:11:15 alinrus: no.. even though we only support one process, the thread ids are very different. The pid is the node number. In single CPU and SMP systems, it will always be 1. 2010-08-08T09:11:37 Fallenou: why? Usually the PR itself is the forum for discussion 2010-08-08T09:35:30 *** exception13 has quit IRC 2010-08-08T09:46:57 ok, so pthread_kill it is :) 2010-08-08T09:47:14 :) We try very hard to be posix compliant even when it is a stretch 2010-08-08T09:47:51 i see but you can go only so far 2010-08-08T09:48:18 Right.. but we go farther than you would expect.. we even enforce uid/gid on file operations :-D 2010-08-08T09:49:09 makes it easier to port software 2010-08-08T09:49:44 Much.. in the class a coup[le of weeks ago, the person was interested in the GNU Scientific Library. I had never heard of it but within 30 minutes we had run the first demo 2010-08-08T09:51:00 wow, impressive. i have used gsl a lot for school, it's very useful for some engineering tasks 2010-08-08T09:51:57 I have wanted to do "application domain kits" for RTEMS. Graphics, Scientific, etc. The core is a script to help compile all the libraries, etc that would be nice in that domain. 2010-08-08T10:16:13 *** lekernel has joined #rtems 2010-08-08T10:22:03 *** alinrus has quit IRC 2010-08-08T10:55:03 *** alinrus has joined #rtems 2010-08-08T10:57:49 *** lekernel has quit IRC 2010-08-08T11:53:12 *** madrazr has joined #rtems 2010-08-08T13:58:47 DrJoel, what do you mean by "Can't compile with warning". The rest of your comments were addressed 2010-08-08T14:01:42 *** lekernel has joined #rtems 2010-08-08T14:03:23 ~/win 11 2010-08-08T14:13:04 *** alinrus has quit IRC 2010-08-08T14:27:31 *** alinrus has joined #rtems 2010-08-08T14:31:37 alinrus: no warnings when compiling. Since I didn't compile, I didn't know but wanted to say that. 2010-08-08T14:39:48 *** alinrus has quit IRC 2010-08-08T14:39:52 *** alinrus_ has joined #rtems 2010-08-08T14:57:36 http://www.fpaste.org/BaNm/ i get some errors in imfs with yesterday\s cvs 2010-08-08T14:59:29 You must need to "preinstall" or rebuild from scratch. That macro is in imfs.h. I can't see why that would fail. :( 2010-08-08T15:04:48 there's an error in imfs.h 2010-08-08T15:04:51 a typo 2010-08-08T15:05:18 if rtems_debug is defined you have 2010-08-08T15:05:22 #ifdef RTEMS_DEBUG 2010-08-08T15:05:22 #include 2010-08-08T15:05:22 #define IMFS_assert(_x) assert(_X) 2010-08-08T15:05:29 but i guess it should be lower case 2010-08-08T15:05:30 x 2010-08-08T15:05:34 assert(_x) 2010-08-08T15:06:52 ok after i fixed that i also have some errors in imfs_eval 2010-08-08T15:06:53 http://www.fpaste.org/PKZ7/ 2010-08-08T15:10:21 you forgot 2 semicolons in imfs_eval.c at lines 176 and 177 2010-08-08T15:10:45 now it works 2010-08-08T15:19:41 *** lekernel has quit IRC 2010-08-08T15:25:16 alinrus_: I just committed the fix in cvs.. thanks for the good eye. 2010-08-08T16:06:24 *** madrazr has left #rtems 2010-08-08T17:53:23 *** DrJoel has quit IRC 2010-08-08T18:13:38 if 16 Kbits will be anough to run rtems ... it seems really really small 2010-08-08T18:13:46 Hello btw 2010-08-08T18:14:05 *** projectgus has joined #rtems 2010-08-08T18:26:52 *** kiwichris has joined #rtems 2010-08-08T19:17:27 *** n3oo3n has quit IRC 2010-08-08T19:18:01 *** n3oo3n has joined #rtems 2010-08-08T20:50:39 *** kuzew has quit IRC 2010-08-08T21:09:16 *** kuzew has joined #rtems 2010-08-08T22:59:11 *** alinrus_ has quit IRC 2010-08-08T23:47:13 *** exception13 has joined #rtems 2010-08-08T23:57:59 *** exception13 has left #rtems