2015-06-22T00:08:55 *** gedare has quit IRC 2015-06-22T00:12:59 [git] 83d97da by Karel Gardas: [add basic support for OpenBSD] http://devel.rtems.org/changeset/83d97da/rtems-source-builder 2015-06-22T00:15:16 [git] e9dfd95 by Chris Johns: [Revert "add basic support for OpenBSD" This reverts commit 83d97daa070d67d177b90796ed22e4044ec9012c. It contains a GPL license.] http://devel.rtems.org/changeset/e9dfd95/rtems-source-builder 2015-06-22T00:18:42 *** mumptai has quit IRC 2015-06-22T00:18:58 [git] dcc0e5c by Chris Johns: [Change license to BSD.] http://devel.rtems.org/changeset/dcc0e5c/rtems-source-builder 2015-06-22T00:43:58 [git] 82830ce by Chris Johns: [Revert "Return the default TX and RX hardware checksum for the Cadence driver to off." This reverts commit 54fb9f01f86286ecea3093f6f524afec86d6313f. This change adds white space changes. A better solution is to revert the original change.] http://devel.rtems.org/changeset/82830ce/rtems-libbsd 2015-06-22T00:43:58 [git] 0d003d7 by Chris Johns: [Revert "if_cgem: Enable hardware checksum assist" This reverts commit dec6b678bba05624335c87b73dc160b54d845cbd. This default for TXCSUM and RXCSUM should be used and a user should turn on TXCSUM and RXCSUM.] http://devel.rtems.org/changeset/0d003d7/rtems-libbsd 2015-06-22T00:43:58 [git] b9fb1ef by Chris Johns: [Update libbsd.txt with Xilinx Zynq TXCSUM and RXCSUM controls.] http://devel.rtems.org/changeset/b9fb1ef/rtems-libbsd 2015-06-22T01:23:40 *** sujayraaj has quit IRC 2015-06-22T02:28:43 *** javamonn has joined #rtems 2015-06-22T02:47:39 *** sujayraaj has joined #rtems 2015-06-22T02:47:48 *** sujayraaj has left #rtems 2015-06-22T04:03:48 *** javamonn has quit IRC 2015-06-22T04:52:42 *** monstr has joined #rtems 2015-06-22T05:08:37 *** crazy3k has joined #rtems 2015-06-22T05:37:17 *** shelf is now known as timothyh 2015-06-22T05:53:20 [trac] GSoC/2015/NestedMutex edited by gadia (empty comment) http://devel.rtems.org/wiki/GSoC/2015/NestedMutex 2015-06-22T05:54:40 [trac] GSoC/2015/NestedMutex edited by gadia (empty comment) http://devel.rtems.org/wiki/GSoC/2015/NestedMutex 2015-06-22T06:27:39 *** hesham has quit IRC 2015-06-22T06:30:27 *** sebhub has joined #rtems 2015-06-22T06:34:13 good morning 2015-06-22T06:43:39 *** Davidbrcz has joined #rtems 2015-06-22T06:53:08 *** Davidbrcz has quit IRC 2015-06-22T06:53:29 *** Davidbrcz has joined #rtems 2015-06-22T07:04:45 *** javamonn has joined #rtems 2015-06-22T07:09:28 *** javamonn has quit IRC 2015-06-22T07:43:53 hi sebhub 2015-06-22T08:00:36 *** ketul has joined #rtems 2015-06-22T08:10:07 hi beng-nl :) 2015-06-22T08:15:23 ketul: hi 2015-06-22T08:15:46 beng-nl: hello have you seen mail from andre ? 2015-06-22T08:16:14 he updated it. 2015-06-22T08:16:15 https://github.com/asuol/rtems/blob/GPIO_API/c/src/lib/libbsp/shared/include/gpio.h 2015-06-22T08:16:42 oh very nice 2015-06-22T08:16:57 looks familiar :-) 2015-06-22T08:17:02 yep 2015-06-22T08:18:22 so now what do you think ? should we do some minor changes to our API 2015-06-22T08:18:37 beng-nl: BTW I updated as per last mail 2015-06-22T08:18:38 https://gist.github.com/ketul93/d717555951174a74c8b4 2015-06-22T08:18:40 yes, we should change ours to theirs 2015-06-22T08:18:43 ketul: ok 2015-06-22T08:18:44 ketul: let me chekc 2015-06-22T08:18:46 ketul: let me check 2015-06-22T08:19:31 ketul: looks better 2015-06-22T08:19:52 ketul: if (gpio_pin_state[gpio_pin_assign->pin_number] = GPIO_PIN_STATE_DIGITAL_OUT), ouch :-) 2015-06-22T08:20:29 *** Davidbrcz has quit IRC 2015-06-22T08:20:51 ohh 2015-06-22T08:20:51 ohh 2015-06-22T08:21:21 updated 2015-06-22T08:22:14 ketul: you took out the mutex? 2015-06-22T08:22:56 beng-nl: yes . 2015-06-22T08:23:05 do you think it is needed ? 2015-06-22T08:23:58 yes, but not a pthreads mutix 2015-06-22T08:23:59 yes, but not a pthreads mutex 2015-06-22T08:24:52 ok. then Semaphore ? 2015-06-22T08:25:12 yes, i think that's safe 2015-06-22T08:25:41 ohk will use binary semaphore 2015-06-22T08:26:01 good! 2015-06-22T08:26:19 And what about namings of the functions and others? Should I change as per andre's work 2015-06-22T08:27:46 yes please 2015-06-22T08:28:17 unless it might still be changed 2015-06-22T08:28:39 yep ! then I think gpio.h will be from andre's work. I should change the gpio.c and others right? 2015-06-22T08:28:59 as gpio.h is in /shared 2015-06-22T08:29:06 yes 2015-06-22T08:29:48 and ya beng-nl semaphores would be implemented in gpio.c right ? 2015-06-22T08:30:01 yes :) 2015-06-22T08:30:58 beng-nl: and you can also see :- https://github.com/asuol/rtems/blob/GPIO_API/c/src/lib/libbsp/arm/raspberrypi/gpio/gpio-interfaces.c 2015-06-22T08:33:31 i see 2015-06-22T08:45:18 beng-nl: then if this code is necessary then for BB I think we need to write for > 50 pins (may be more if we take digital in ) 2015-06-22T08:52:40 i don't think they need to be defined 2015-06-22T08:53:00 i think they are just used to turn jtag mode on/off 2015-06-22T08:53:07 which needn't concern us now 2015-06-22T08:53:20 beng-nl: every struct member should be defined 2015-06-22T08:53:21 ok 2015-06-22T08:56:23 *** jkridner|work has quit IRC 2015-06-22T08:57:07 *** jkridner has joined #rtems 2015-06-22T09:03:54 ketul: what do you think of andre's API? 2015-06-22T09:05:17 beng-nl: I think first I should submit patch to devel-ml. and let community should decide. What do you think ?:) 2015-06-22T09:05:39 of your current state you mean? 2015-06-22T09:05:48 yep 2015-06-22T09:05:53 good idea 2015-06-22T09:06:25 beng-nl: :) 2015-06-22T09:07:00 beng-nl: need your help for semaphore implementation. 2015-06-22T09:07:12 ok! 2015-06-22T09:07:57 beng-nl: is this example is the right one that we are looking : https://github.com/RTEMS/examples-v2/blob/a0c21e019ce0d218fd89b644ee6770d995ed0fdc/led/sem_server/init.c 2015-06-22T09:08:06 rtems_semaphore_create to create one.. 2015-06-22T09:08:23 ok 2015-06-22T09:08:43 yep 2015-06-22T09:08:46 like in Init 2015-06-22T09:09:00 rtems_semaphore_obtain is lock, rtems_semaphore_release is unlock 2015-06-22T09:10:22 rtems_semaphore_create should be done at the top of code once right. Then every time "rtems_semaphore_obtain is lock, rtems_semaphore_release is unlock " in the function should be used 2015-06-22T09:10:59 yes :-) 2015-06-22T09:11:27 depending on what kind of work you do in the critical section you may also use interrupt locks 2015-06-22T09:11:54 ok :) Thanks beng-nl 2015-06-22T09:12:10 that was sebhub, but ok 2015-06-22T09:13:34 sebhub: how to decide the critical section is there :) 2015-06-22T09:15:56 I mean how to decide when to use interrupt locks . Sorry if silly question :) 2015-06-22T10:05:26 *** crazy3k has quit IRC 2015-06-22T10:09:07 the critical section should execute in a short and bounded time interval 2015-06-22T10:10:22 [git] f9090ac by Sebastian Huber: [tmtests/tm27: Use scheduler lock] http://devel.rtems.org/changeset/f9090ac/rtems 2015-06-22T10:10:22 [git] cdf30f0 by Sebastian Huber: [rtems: Add rtems_interrupt_local_disable|enable() Add rtems_interrupt_local_disable|enable() as suggested by Pavel Pisa to emphasize that interrupts are only disabled on the current processor. Do not define the rtems_interrupt_disable|enable|flash() macros and functions on SMP configurations since they don't ensure system 2015-06-22T10:10:22 wide mutual exclusion.] http://devel.rtems.org/changeset/cdf30f0/rtems 2015-06-22T10:26:49 *** zoso has joined #rtems 2015-06-22T10:36:58 beng-nl: can you review it ? https://gist.github.com/ketul93/d717555951174a74c8b4#file-gpio-c 2015-06-22T10:38:45 beng-nl: I am getting assertion "status == RTEMS_SUCCESSFUL" failed: at 109 line. probably sebhub can also help 2015-06-22T10:43:11 *** monstr has quit IRC 2015-06-22T10:43:12 *** kulp has quit IRC 2015-06-22T10:43:12 *** cdesai_ has quit IRC 2015-06-22T10:43:12 *** rokka has quit IRC 2015-06-22T10:43:30 *** cdesai has joined #rtems 2015-06-22T10:43:36 *** monstr has joined #rtems 2015-06-22T10:44:09 *** kulp has joined #rtems 2015-06-22T10:58:26 *** zoso has quit IRC 2015-06-22T11:07:54 let's see 2015-06-22T11:11:58 beng-nl: yep pls :) 2015-06-22T11:12:30 was rtems_gpio_initialize() called? 2015-06-22T11:13:07 to create the semaphore 2015-06-22T11:13:41 ohh very silly one :( actually forgot to call because that function was defined after words 2015-06-22T11:13:47 ok 2015-06-22T11:13:50 let me check 2015-06-22T11:14:43 beng-nl: other ways is it good ? 2015-06-22T11:47:30 *** ketul has quit IRC 2015-06-22T11:58:16 *** ketul has joined #rtems 2015-06-22T12:31:26 *** kiwichris has joined #rtems 2015-06-22T12:31:41 sebhub, hi 2015-06-22T12:36:09 jhi 2015-06-22T12:41:15 I fixed the libbsd reverting those changes. 2015-06-22T12:41:35 yes, looks good now 2015-06-22T12:42:43 Great. Have you built libbsd with waf ? 2015-06-22T12:43:12 i didn't update the libbsd since i am still hunting the memory corruption bug 2015-06-22T12:43:26 Ah ok. 2015-06-22T12:43:28 and i work currently on other things 2015-06-22T12:43:53 there is still a lot to do in the smp area 2015-06-22T12:56:52 *** gedare has joined #rtems 2015-06-22T12:56:52 *** ChanServ sets mode: +o gedare 2015-06-22T13:00:38 hi ketul 2015-06-22T13:00:47 hi gedare 2015-06-22T13:00:49 ketul: have you fixed your issue with the rtems mailing lists? 2015-06-22T13:01:01 are you subscribed to the users@rtems mailing list? 2015-06-22T13:01:13 gedare: yep I got mail in bunch 2015-06-22T13:01:24 I mean 3 4 msgs in a mail 2015-06-22T13:01:29 ah ok I personally can't handle digest-form myself 2015-06-22T13:01:36 anyway, there was one email: Beaglebone Black GPIO Interrupt, and minor JTAG/gdb issue 2015-06-22T13:02:04 ah let me read once more 2015-06-22T13:03:21 yep got he even mentioned my name :) 2015-06-22T13:03:43 why are the gpio descriptors not const? 2015-06-22T13:03:52 yes and he attached some code, and you should follow-up with him since he has requirements about how to use the gpio support 2015-06-22T13:03:58 the : style initializers are not standard C 2015-06-22T13:04:25 good morning sebhub nice to see you. 2015-06-22T13:04:31 hi 2015-06-22T13:04:55 ok gedare I will do it 2015-06-22T13:05:18 ok ketul and see the comments above too. 2015-06-22T13:05:24 yep 2015-06-22T13:06:19 sebhub: if you have time it will be quite helpful if you can provide suggestions for "GSoC 2015 RPi USB Support" and "rtems-libbsd sample testsuite explanation" I have not used libbsd enough to have any advice 2015-06-22T13:06:40 even if your suggestion is, "need more information" :) 2015-06-22T13:08:15 ok, the problem is that I didn't follow the recent libbsd updates 2015-06-22T13:11:29 gedare: sorry but I can't see the attached code there. 2015-06-22T13:13:55 ketul: I will forward to you, and i suggest you switch from digest-only in your settings at https://lists.rtems.org/mailman/listinfo 2015-06-22T13:14:50 *** javamonn has joined #rtems 2015-06-22T13:16:19 ok 2015-06-22T13:22:47 gedare: as he mentioned I didn't touch on IRQ yet. But I will try it to be solved. may take some time. 2015-06-22T13:23:22 will refer TRM 2015-06-22T13:30:39 *** javamonn has quit IRC 2015-06-22T13:34:26 *** javamonn has joined #rtems 2015-06-22T13:39:09 ketul: did you also see the gpio patches from Andre? 2015-06-22T13:39:16 on devel@rtems ml 2015-06-22T13:51:26 [trac] SOCIS/2015 edited by salpha (empty comment) http://devel.rtems.org/wiki/SOCIS/2015 2015-06-22T13:52:57 *** monstr has quit IRC 2015-06-22T14:14:22 *** dougk__ has quit IRC 2015-06-22T14:32:09 *** dougk_ has joined #rtems 2015-06-22T14:44:53 *** sujayraaj has joined #rtems 2015-06-22T15:14:33 *** jrcatbagan has joined #rtems 2015-06-22T15:15:09 hi gedare, hi beng-nl 2015-06-22T15:19:12 hello jrcatbagan 2015-06-22T15:22:58 *** igutek has joined #rtems 2015-06-22T15:23:35 hi jrcatbagan 2015-06-22T15:23:38 and gedare :) 2015-06-22T15:24:55 and beng-nl 2015-06-22T15:25:03 :) 2015-06-22T15:26:17 gedare, beng-nl: Some of the Umon files in BBB port were taken from the template directory where the files from this directory do not have header comments. Is it considered to be ok to put myself as the author for the files I replace? How would I provide proper attribution in cases where the header comments are missing? 2015-06-22T15:27:38 Is the new Umon under BSD license? or RTEMS license? 2015-06-22T15:37:21 *** Davidbrcz has joined #rtems 2015-06-22T15:41:25 I was just curious so I could put the proper License header in the BBB files. 2015-06-22T15:46:19 *** Davidbrcz has quit IRC 2015-06-22T15:48:47 (1) Yes. (2) Ask Ed. (3) APL. 2015-06-22T15:50:43 ok, thank you for the clarification 2015-06-22T16:10:46 *** jrcatbagan has quit IRC 2015-06-22T16:15:15 *** jrcatbagan has joined #rtems 2015-06-22T16:17:10 *** dr__house|away is now known as dr__house 2015-06-22T18:13:13 *** yangqiao has joined #rtems 2015-06-22T18:23:14 *** yangqiao has quit IRC 2015-06-22T19:03:41 *** ketul has quit IRC 2015-06-22T19:16:33 *** gedare has quit IRC 2015-06-22T19:18:42 *** jrcatbagan has quit IRC 2015-06-22T19:22:55 *** crazy3k has joined #rtems 2015-06-22T19:40:55 *** jrcatbagan has joined #rtems 2015-06-22T19:44:48 *** javamonn has quit IRC 2015-06-22T19:45:23 *** javamonn has joined #rtems 2015-06-22T19:49:29 *** mumptai has joined #rtems 2015-06-22T20:31:46 *** gedare has joined #rtems 2015-06-22T20:31:46 *** ChanServ sets mode: +o gedare 2015-06-22T20:57:25 *** gedare has quit IRC 2015-06-22T21:04:18 *** javamonn has quit IRC 2015-06-22T21:05:09 *** javamonn has joined #rtems 2015-06-22T21:14:45 *** Davidbrcz has joined #rtems 2015-06-22T21:41:28 *** jkridner has quit IRC 2015-06-22T21:42:18 *** jkridner has joined #rtems 2015-06-22T21:43:45 *** yangqiao has joined #rtems 2015-06-22T21:50:26 *** javamonn has quit IRC 2015-06-22T22:01:29 *** yangqiao has quit IRC 2015-06-22T22:07:32 *** javamonn has joined #rtems 2015-06-22T22:13:49 [trac] GSoC/2015 edited by yangqiao (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-06-22T22:24:01 *** javamonn has quit IRC 2015-06-22T22:35:43 *** yangqiao has joined #rtems 2015-06-22T22:43:03 *** jrcatbagan has quit IRC 2015-06-22T22:45:22 *** yangqiao has joined #rtems 2015-06-22T22:47:49 *** jrcatbagan has joined #rtems 2015-06-22T22:52:25 *** jrcatbagan has quit IRC 2015-06-22T23:53:54 *** sujayraaj has quit IRC 2015-06-22T23:54:16 *** sujayraaj has joined #rtems 2015-06-23T00:38:49 *** kiwichris has quit IRC 2015-06-23T01:00:47 *** kiwichris has joined #rtems 2015-06-23T01:04:56 *** Davidbrcz has quit IRC 2015-06-23T01:44:26 *** jrcatbagan has joined #rtems 2015-06-23T01:54:22 *** jrcatbagan has quit IRC 2015-06-23T01:56:07 *** sujayraaj has quit IRC 2015-06-23T02:14:54 *** javamonn has joined #rtems 2015-06-23T02:38:16 *** javamonn has quit IRC 2015-06-23T02:44:51 *** javamonn has joined #rtems 2015-06-23T03:31:03 *** mumptai has quit IRC 2015-06-23T03:35:17 *** javamonn has quit IRC 2015-06-23T03:57:47 *** monstr has joined #rtems 2015-06-23T04:12:51 *** dr__house is now known as dr__house|away 2015-06-23T06:40:15 *** zoso has joined #rtems 2015-06-23T06:43:42 ita, about ? 2015-06-23T06:57:41 sebhub, hi 2015-06-23T06:58:47 sebhub, with cmdchown is there any C type interface to manage ownership or perms ? 2015-06-23T06:59:48 [git] b22ad9f by Sebastian Huber: [rtems-bsd-mutex: Fix resource count updates] http://devel.rtems.org/changeset/b22ad9f/rtems-libbsd 2015-06-23T07:15:12 [git] b5cdb95 by Sebastian Huber: [condvar01: Initialize mutex] http://devel.rtems.org/changeset/b5cdb95/rtems-libbsd 2015-06-23T07:16:29 rtems_shell_lookup_cmd() + direct access to the fields 2015-06-23T07:16:37 maybe we should add accessor functions 2015-06-23T07:17:37 did you test the libbsd + waf on mingw/msys/cygwin? 2015-06-23T07:21:46 Yes accessor functions would be good. 2015-06-23T07:22:06 No I did not. I will try it soon. I need to build some new Windows tools. 2015-06-23T07:22:12 Does it work ? 2015-06-23T07:22:25 *** kiwichris has quit IRC 2015-06-23T07:29:21 *** mkhoory-2 has joined #rtems 2015-06-23T07:33:16 *** mkhoory-3 has quit IRC 2015-06-23T07:34:17 *** kiwichris has joined #rtems 2015-06-23T07:38:27 i didn't test on windows 2015-06-23T07:41:55 Ok. I sort sone windows tools out and try. 2015-06-23T07:47:08 ok, thanks, in case this works, then we should remove the makefile 2015-06-23T07:51:58 Sure. 2015-06-23T07:52:13 Have you tried the waf build ? 2015-06-23T07:52:22 yes, for the recent fixes 2015-06-23T07:52:29 Great 2015-06-23T07:52:36 the problem visible via telnet01 was quite silly 2015-06-23T07:52:52 *** cdesai has quit IRC 2015-06-23T07:52:57 ? 2015-06-23T07:53:18 the resource_count updates were wrong 2015-06-23T07:53:18 hmm, MSYS2 defaults to Python3 and that breaks the RSB. 2015-06-23T07:53:35 Ah ok. I saw the patch. 2015-06-23T07:55:08 *** cdesai has joined #rtems 2015-06-23T08:05:49 *** kiwichris has quit IRC 2015-06-23T09:34:58 *** yangqiao has joined #rtems 2015-06-23T10:01:37 *** rokka has joined #rtems 2015-06-23T10:30:41 *** asuolen has joined #rtems 2015-06-23T10:31:30 *** new2rtos has joined #rtems 2015-06-23T10:31:49 hello 2015-06-23T10:33:35 I got assertion while running status = rtems_semaphore_create(rtems_build_name('A', 'S', 'E', ' '), 1, RTEMS_DEFAULT_ATTRIBUTES, 0, &lock ); assert (status == RTEMS_SUCCESSFUL); 2015-06-23T10:41:27 I think asuolen can help. seen your code og github 2015-06-23T10:41:32 :) 2015-06-23T10:45:43 have you accounted for the semaphore in confdefs? You may have more info on the error if you check the actual value of status. 2015-06-23T10:47:03 confdefs? didn't get it 2015-06-23T10:49:55 its a table used to configure your rtems application 2015-06-23T10:50:21 asuolen: where it is written? can you point me to the link? 2015-06-23T10:50:51 you have some documentation on https://docs.rtems.org/releases/rtemsdocs-4.10.1/share/rtems/html/c_user/c_user00415.html 2015-06-23T10:51:32 maybe you need to adjust the CONFIGURE_MAXIMUM_SEMAPHORES option 2015-06-23T10:52:15 but this is just a possibility. You should check what the actual value of status to know exacly what went wrong. 2015-06-23T10:52:58 ok 2015-06-23T10:54:14 asuolen: https://github.com/asuol/rtems/blob/GPIO_API/c/src/lib/libbsp/shared/gpio.c don't have any options like CONFIGURE_MAXIMUM_SEMAPHORES 2015-06-23T10:56:59 the confdefs only apply to an application. You can see sample applications at the testsuites/samples directory at the root of the rtems code base 2015-06-23T10:57:09 ohk 2015-06-23T10:57:52 thanks a lot asuolen 2015-06-23T11:00:35 asuolen: sorry if m asking more . but https://github.com/asuol/rtems/blob/GPIO_API/testsuites/samples/LIBGPIO_MULTI_TEST/init.c is also not having 2015-06-23T11:00:47 like CONFIGURE_MAXIMUM_SEMAPHORES 2015-06-23T11:03:27 the default value is 5, and in the Pi case that is enough. Only if you need more than that should you need to set that option. 2015-06-23T11:04:24 have you checked the value of status on your code? 2015-06-23T11:04:33 how do you decide the value? 2015-06-23T11:04:41 should be 5 ? 2015-06-23T11:05:13 I mean is it in RPi datasheet ? 2015-06-23T11:05:21 In case you are using my gpio code, it requires a semaphore per bank of pins. 2015-06-23T11:06:06 but why only 1 is semaphore is not sufficient for that. I think at a time 1 pin is addressed 2015-06-23T11:06:38 That is why I said for you to check the actual value of status :) It may be another reason 2015-06-23T11:07:16 you can print the value to see it 2015-06-23T11:07:27 ohk let me check 2015-06-23T11:07:42 rtems takes too much time for compilation 2015-06-23T11:09:07 if you use windows, then i would use a vm with linux for rtems development 2015-06-23T11:10:39 sebhub: m using ubuntu 2015-06-23T11:11:05 then it should only task 2-3 minutes to build it 2015-06-23T11:11:13 from scratch 2015-06-23T11:11:30 no mine is taking 8-9 mins 2015-06-23T11:11:41 may b because my pc has i3 :) 2015-06-23T11:16:01 difference btw : CONFIGURE_MAXIMUM_TASKS and CONFIGURE_MAXIMUM_TASKS ? 2015-06-23T11:16:17 *CONFIGURE_MAXIMUM_SEMAPHORES 2015-06-23T11:19:58 its in the name: the maximum number of rtems tasks (you may see it as threads), and the maximum number of semaphores. They are not related 2015-06-23T12:20:59 hi asuolen sorry for replying after a long time. I got "(null)","0" (%s,%d) as a output of status=rtems_semaphore_create( 2015-06-23T12:22:21 thats why getting assertion I think but what is the meaning of "0" ? 2015-06-23T12:57:54 sebhub: can you help? 2015-06-23T12:59:23 i don't understand the problem 2015-06-23T13:00:23 status = rtems_semaphore_create( 2015-06-23T13:00:23 rtems_build_name('A', 'B', 'C', ' '), 2015-06-23T13:00:23 1, 2015-06-23T13:00:23 RTEMS_DEFAULT_ATTRIBUTES, 2015-06-23T13:00:23 0, 2015-06-23T13:00:24 &lock 2015-06-23T13:00:24 ); 2015-06-23T13:00:25 printf("%s %d\n",status,status ); 2015-06-23T13:00:26 is my code 2015-06-23T13:00:56 and getting "(null) 0 2015-06-23T13:00:56 " 2015-06-23T13:01:16 yes, and where is the problem? 2015-06-23T13:01:30 [trac] GSoC/2015 edited by gtament (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-06-23T13:01:41 apart from status being an enum and not a string 2015-06-23T13:02:31 sebhub: while writting "assert (status == RTEMS_SUCCESSFUL);" I get assertion 2015-06-23T13:04:04 RTEMS_SUCCESSFUL is 0, so this doesn't correspond to the output 2015-06-23T13:08:24 *** javamonn has joined #rtems 2015-06-23T13:08:34 [trac] GSoC/2015/RaspberryPiUSBSupport created http://devel.rtems.org/wiki/GSoC/2015/RaspberryPiUSBSupport 2015-06-23T13:12:00 sebhub: then how I am getting assertion if it is 0 ? 2015-06-23T13:12:19 i would ask the debugger 2015-06-23T13:13:11 :) 2015-06-23T13:13:36 I mean theoretically I should not get right ? 2015-06-23T13:14:27 the assertion and the print output contradict each other 2015-06-23T13:15:28 that is written is same file. then can you think what kind of prob may b der? 2015-06-23T13:15:38 *in 2015-06-23T13:18:24 its easy to figure this out by means of a debugger 2015-06-23T13:19:03 but how I dont have any idea related to that 2015-06-23T13:19:25 related to what? 2015-06-23T13:19:41 "its easy to figure this out by means of a debugger " 2015-06-23T13:20:25 sorry, i don't understand your question 2015-06-23T13:20:52 I mean how to do debugging ? 2015-06-23T13:21:08 ok, you know what a debugger is? 2015-06-23T13:21:34 are you talking for hw debugger or sw one? 2015-06-23T13:21:54 since rtems is software, i talk about a software debugger 2015-06-23T13:22:37 ohk I know about debugger but don't have idea how and which to use for rtems 2015-06-23T13:23:04 this depends on your target, i use Lauterbach or GDb 2015-06-23T13:23:34 ok 2015-06-23T13:24:01 then without debugger should I solve my prob? if possible 2015-06-23T13:24:22 developing software without a debugger is a waste of time 2015-06-23T13:24:52 ok I want to debug the .exe file that is generated after make 2015-06-23T13:26:13 which target? 2015-06-23T13:26:42 *** javamonn has quit IRC 2015-06-23T13:30:25 *** javamonn has joined #rtems 2015-06-23T13:54:09 *** javamonn has quit IRC 2015-06-23T13:57:16 *** javamonn has joined #rtems 2015-06-23T13:58:38 *** ragunath has joined #rtems 2015-06-23T14:01:28 [trac] GSoC/2015 edited by krohini (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-06-23T14:06:27 [trac] GSoC/2015 edited by ragunath (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-06-23T14:22:18 *** DrJoel has joined #rtems 2015-06-23T14:22:19 *** DrJoel has joined #rtems 2015-06-23T14:22:19 *** ChanServ sets mode: +o DrJoel 2015-06-23T14:32:25 *** hermann_19829 has joined #rtems 2015-06-23T14:42:28 *** gtament has joined #rtems 2015-06-23T14:43:54 *** krohini has joined #rtems 2015-06-23T14:43:56 Good time of day) 2015-06-23T14:44:51 hello everybody! 2015-06-23T14:45:22 *** jrcatbagan has joined #rtems 2015-06-23T14:46:00 yangqiao: Hi) How is the mailbox? 2015-06-23T14:46:09 [trac] SOCIS/2015 edited by hermann19829 (empty comment) http://devel.rtems.org/wiki/SOCIS/2015 2015-06-23T14:46:26 hi 2015-06-23T14:47:09 I've proposed a new structure and macro to construct the buffer, as sugguested by gedare. 2015-06-23T14:47:19 *** gedare has joined #rtems 2015-06-23T14:47:20 *** ChanServ sets mode: +o gedare 2015-06-23T14:47:54 But I've got a little problem with memory allocation. you need to set the power state if I remembered right? 2015-06-23T14:48:41 exactly 2015-06-23T14:48:51 bcm2835_mbox_set_power_state(dev, BCM2835_MBOX_POWER_ID_USB_HCD, TRUE); 2015-06-23T14:48:53 I can give you a piece of code which worked for test 2015-06-23T14:49:16 Because I still need to discuss with gedare and others 2015-06-23T14:49:52 ok, I would like to help you with testing, Let's contact on email) 2015-06-23T14:50:05 ok. 2015-06-23T14:50:35 *** new2rtos has left #rtems 2015-06-23T14:52:41 *** sgworks has joined #rtems 2015-06-23T14:53:18 *** ketul has joined #rtems 2015-06-23T15:00:17 Good morning everyone 2015-06-23T15:00:28 hello all 2015-06-23T15:00:34 morning! 2015-06-23T15:00:34 hello everyone 2015-06-23T15:00:35 hi 2015-06-23T15:00:36 Hi, Gedare 2015-06-23T15:00:51 hello 2015-06-23T15:01:00 I have a few announcements before we begin. 2015-06-23T15:01:01 hi all 2015-06-23T15:01:18 1) The midterm is coming up, so make sure you are getting any deliverables ready. 2015-06-23T15:01:18 2) For the midterm, I'd like you to post a "midterm report" on your blog. 2015-06-23T15:01:18 3) I noticed some of you haven't put anything on your blog, or the link doesn't work from the trac wiki. Please also write an "introduction" post if you haven't yet. 2015-06-23T15:01:18 4) Ensure that the link to your github points to the right repository, and if you have primarily been using a branch, you might want to link to that branch instead. 2015-06-23T15:01:21 5) The best way to ensure you pass the midterm is to push code, even if it is just a test case or something unrelated to your main project such as a bug you fixed locally. 2015-06-23T15:01:24 6) Next week's meeting will start half an hour late and will focus on ensuring everyone has submitted their midterm deliverables and student evaluations. 2015-06-23T15:01:27 7) The meeting after that, on July 7th, is cancelled. 2015-06-23T15:02:09 We'll start today with sgworks. Hi sgworks, javamonn. 2015-06-23T15:02:43 hello all 2015-06-23T15:03:29 *** sgworks_ has joined #rtems 2015-06-23T15:03:41 sgworks_: you're up first 2015-06-23T15:03:55 So RTEMS-JPF model is up 2015-06-23T15:03:55 sgworks_: did you see the announcements, or should I email them to you? 2015-06-23T15:04:07 *** beng-nl-droid has joined #rtems 2015-06-23T15:04:12 and yesterday I added validation functionality to verify 2015-06-23T15:04:17 Hi all 2015-06-23T15:04:21 what you mentioned in threads 2015-06-23T15:04:47 So now yesterday Cyrille aslo mentioned that validation do get the unbounded priority inheritance model 2015-06-23T15:04:56 *problem 2015-06-23T15:05:01 so we are good on that 2015-06-23T15:05:08 I will cross-verify that 2015-06-23T15:05:15 adn for solution as well 2015-06-23T15:05:35 2. Test the code for corrected model. 2015-06-23T15:05:35 3. Implement the review changes suggested by guides. 2015-06-23T15:05:35 4. Try adding properties to JPF. 2015-06-23T15:05:35 5. Prepare the mid-term evaluation materials. 2015-06-23T15:05:45 "validation code" 2015-06-23T15:05:48 good. We'll have to figure out what to do with your model for the long-term 2015-06-23T15:06:03 "properties"? 2015-06-23T15:06:18 yes 2015-06-23T15:06:20 what properties 2015-06-23T15:06:24 hi beng-nl-droid 2015-06-23T15:06:25 its good model to verify 2015-06-23T15:06:32 *** sgworks has quit IRC 2015-06-23T15:06:42 What is involved in #4 I guess. 2015-06-23T15:06:46 some JPF properties I am also not sure on this what they are 2015-06-23T15:06:53 but will talk with cyrille 2015-06-23T15:06:58 link https://github.com/saurabhgadia4/lock-model 2015-06-23T15:07:09 yeah 2015-06-23T15:07:17 *** sujayraaj has joined #rtems 2015-06-23T15:07:18 but I guess we are good with mid-term deliverables 2015-06-23T15:07:24 ask him if he wants you to pull request back into his original model 2015-06-23T15:07:28 ANd its very slow 2015-06-23T15:07:31 heh 2015-06-23T15:07:36 so can't have more messy 2015-06-23T15:07:38 tests 2015-06-23T15:07:46 ok sure. ps great commit message there, "Fed up with Java"./ 2015-06-23T15:07:53 yeah 2015-06-23T15:07:56 1am 2015-06-23T15:08:05 so wanted to take break 2015-06-23T15:08:13 morning python and evening java 2015-06-23T15:08:14 so 2015-06-23T15:08:15 ok, thanks sgworks_ so what are you planning to get done by the midterm? 2015-06-23T15:08:20 totally opposite 2015-06-23T15:08:33 validation of both model 2015-06-23T15:08:37 okay 2015-06-23T15:08:45 How far off do you see getting the algorithm into RTEMS? 2015-06-23T15:08:47 python? for some other project you are doing? 2015-06-23T15:08:48 which shows you RTEMS has unbounded priority inheritance problem 2015-06-23T15:08:54 and its solution solves 2015-06-23T15:08:56 this 2015-06-23T15:09:01 sorry gedare for being late, I was stuck somewhere. 2015-06-23T15:09:06 So in former we will get assertion error 2015-06-23T15:09:13 sujayraaj: ok i'll email you some announcements. 2015-06-23T15:09:18 and later model will complete execution successfully 2015-06-23T15:09:21 EVERYONE.. if you have patches outstanding, make sure they are getting merged and not accidentally dropped. It is summer and there is a lot of activity. 2015-06-23T15:09:24 ok thanks sgworks_ that sounds good 2015-06-23T15:09:26 yeah for all rest 2015-06-23T15:09:34 ok 2015-06-23T15:09:43 next up will be yangqiao, anything else? 2015-06-23T15:09:50 nope 2015-06-23T15:09:55 ok, hi yangqiao 2015-06-23T15:10:01 Thanks a lot. We are on good note 2015-06-23T15:10:15 And yes wiki is updated but not blog 2015-06-23T15:10:20 hi 2015-06-23T15:10:25 so will provide link of wiki on blog 2015-06-23T15:10:26 sgworks_: Yes, much better now you are back on track. 2015-06-23T15:10:46 YES "sgworks" 2015-06-23T15:10:57 I've just sent an mail about my problem. 2015-06-23T15:11:06 ok i haven't seen it yet 2015-06-23T15:11:14 i mean, it's in my inbox stiill. 2015-06-23T15:11:31 I've tried out the zero length array , it works well. But we cannot malloc in kernel space 2015-06-23T15:12:13 I've wrote an scratch, the only problem is the allocation. 2015-06-23T15:12:31 Where precisely would the allocation get callled from? 2015-06-23T15:12:33 ok there are a few possible solutions 2015-06-23T15:12:40 i'll address them by email 2015-06-23T15:13:07 ok. 2015-06-23T15:13:09 And 2015-06-23T15:13:45 If you think the outch is accepted. I'll push the refactored code for review 2015-06-23T15:13:57 i'll take a look at it 2015-06-23T15:14:20 ok. I've looked into the lbjpeg , integrated in rsb 2015-06-23T15:14:26 and I'm working on others 2015-06-23T15:14:44 [trac] GSoC/2015 edited by gadia (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-06-23T15:14:58 alright, that is a good task to be done on the side 2015-06-23T15:15:04 *** sgworks_ has quit IRC 2015-06-23T15:15:12 anything else yangqiao ? 2015-06-23T15:15:20 *** salpha has joined #rtems 2015-06-23T15:15:29 nothing else. we'll talk about the allocation later 2015-06-23T15:15:43 yep, ok great. gtament you're up now 2015-06-23T15:15:57 hi every one! 2015-06-23T15:16:23 I completely moved to latest libbsd revision, so now I'm building with waf. I repushed driver because of that. Also I figured out almost all questions with testsuites. And got new ones, which I described on devel, but after three days, I still got no feedback. While I'm waiting I started working on a code, that will list USB devices, which I can test on any arch. 2015-06-23T15:17:09 ok, unfortunately i haven't been using libbsd so i can't provide much insight on your technical problems 2015-06-23T15:17:21 so you just have to dig yourself or hope someone that has got the experience and time can pitch in 2015-06-23T15:17:53 you may send an update/ping on the message, and add any more info you might have learned in those 3 days 2015-06-23T15:18:28 I did, I added all mentors and Chris Johns 2015-06-23T15:18:37 ok 2015-06-23T15:18:55 I am trying to bring libbsd up on pc386 so battling similar issues but many test failures at the moment. 2015-06-23T15:20:14 I got very strange bug (if the root of a bug is not my typos), when probe() method of the driver isn't called 2015-06-23T15:20:22 thanks gtament hopefully we can help you figure out your problem 2015-06-23T15:20:40 I hope so too. Thanks, Gedare 2015-06-23T15:21:03 if you need advice on debugging you may like to write a blog post about what you ahve been doing, where you can get into more details 2015-06-23T15:21:16 ok, up next is ketul 2015-06-23T15:21:17 hi ketul 2015-06-23T15:21:26 hi gedare 2015-06-23T15:21:31 hi all this week I did work for gpio, coordinated with andre. I got some problems for the gpio implementation and resolved it using "arm-rtems4.11-objdump -d gpio.exe" . I recently mailed the latest patch on devel-ml, so will wait for the community review and will hope the common gpio api running on every board. and for ADC i need some specificatio 2015-06-23T15:21:31 ns that I did coordination with beng-nl and will start shortly making some functions for it. 2015-06-23T15:21:50 I am gonna post it how to solve any problem on hardware thi week 2015-06-23T15:22:19 gedare: recently posted a patch 2015-06-23T15:23:12 ok i see it. have you tried to get git-send-email to work? 2015-06-23T15:23:31 no I didn't 2015-06-23T15:23:52 ok that is a preferred way to review patches, but i'll download that one you sent and have a look 2015-06-23T15:24:24 ok thanks for that. gedare is it good for mid term according to you? 2015-06-23T15:24:33 It's what you have. 2015-06-23T15:25:19 ok :) . still trying for ADC but it is complex compare to gpio 2015-06-23T15:25:42 I'll reserve judgment until later, but for the midterm we take into consideration the entire body of work done, effort, time put in, interaction with community, written correspondence, attendance at IRC, and progress toward goals. 2015-06-23T15:26:29 ok thats great. will post the problems and solutions on blog 2015-06-23T15:26:40 which I got on BBB 2015-06-23T15:26:44 ok, thanks. i'll try to review code today so you all can resubmit again soon, maybe even get merged before midterm and sneak it into the 4.11 release...... :) 2015-06-23T15:26:55 alright, thanks ketul. up next will be ragunath 2015-06-23T15:27:06 hi all 2015-06-23T15:27:06 thaks a lot too gedare 2015-06-23T15:27:15 Basic networking was working with BBB last week. 2015-06-23T15:27:19 This week I tested sample test cases available in libbsd. 2015-06-23T15:27:26 I was able to test most of the test cases. Most of the tests passed successfully 2015-06-23T15:27:31 specifically i was able to telnet to the machine and execute commands 2015-06-23T15:27:35 I was able to ftp to the machine and get put files 2015-06-23T15:28:00 most other tests also passed 2015-06-23T15:28:07 did you resolve the issue with the 4 you posted on the ml? 2015-06-23T15:28:10 [trac] SOCIS/2015 edited by salpha (empty comment) http://devel.rtems.org/wiki/SOCIS/2015 2015-06-23T15:28:14 I could not find out how to execute foobar client/server , media and ppp01 tests. 2015-06-23T15:28:19 ok 2015-06-23T15:28:20 i figured out arphole test 2015-06-23T15:28:22 and it passed 2015-06-23T15:28:31 nice 2015-06-23T15:28:32 don't spin too much on that. 2015-06-23T15:28:35 media tests require memory devices 2015-06-23T15:28:45 it does not need any networking 2015-06-23T15:28:51 Please write up notes and at least blog on testing this. I am asking those questions now 2015-06-23T15:29:04 did you happen to see gtament's post about his trouble with libbsd on the pi? 2015-06-23T15:29:09 sure DrJoel 2015-06-23T15:29:17 ./waf 2015-06-23T15:29:26 lol wrong window 2015-06-23T15:29:46 yes usb is not enabled in BBB 2015-06-23T15:30:06 ah, ok 2015-06-23T15:30:07 i will speak with him if i can help with anything 2015-06-23T15:30:12 great 2015-06-23T15:30:21 I can give an image so that some other person can test the changes. 2015-06-23T15:30:30 n the mean time, I am planning to prepare the patch and send it to devel. 2015-06-23T15:30:32 thanks ragunath let's get your code merged then 2015-06-23T15:30:32 *in 2015-06-23T15:30:48 okay 2015-06-23T15:30:51 anything else? 2015-06-23T15:30:55 we need to two work arounds done to make this work 2015-06-23T15:31:06 explain them with the patches 2015-06-23T15:31:22 ok i will send them to devel 2015-06-23T15:31:36 thats it from me today 2015-06-23T15:31:36 thanks. up next is krohini 2015-06-23T15:31:46 hi gedare 2015-06-23T15:32:01 hello 2015-06-23T15:32:10 I did get the speed up like I had posted to the mailing list 2015-06-23T15:32:37 *** beng-nl-droid has quit IRC 2015-06-23T15:32:52 That was an impressive speedup :) 2015-06-23T15:32:55 Have sent a patch too. Rather they are the changes that were required 2015-06-23T15:32:55 *** beng-nl-droid has joined #rtems 2015-06-23T15:33:09 ok good. the # is now right in the range of the linux performance reports 2015-06-23T15:33:13 DrJoel: Yeah :) 2015-06-23T15:33:24 do you think it will also improve rpi performance? 2015-06-23T15:33:30 gedare: Yes it is 2015-06-23T15:33:31 because that was also off by an order of magnitude 2015-06-23T15:33:31 ok 2015-06-23T15:33:43 Ya I was going to add that 2015-06-23T15:33:48 ok great 2015-06-23T15:33:59 so what is your next steps? 2015-06-23T15:34:35 I am looking to get the patch finalized. There has been some discussion on the patch 2015-06-23T15:34:55 *** beng-nl-droid has quit IRC 2015-06-23T15:35:11 *** beng-nl-droid has joined #rtems 2015-06-23T15:35:18 ok 2015-06-23T15:35:30 Have started with SMP work. but I need more discussion there 2015-06-23T15:35:52 I have posted some identified areas on the blog 2015-06-23T15:36:12 ok well bring it back up when you feel ready to work on it. email that sits for more than a few days probably isn't going to get answered, and your blog is going to be read only occasionally (and I might not pick up on questions embedded there) 2015-06-23T15:36:42 thanks krohini anything else? 2015-06-23T15:36:46 hermann_19829: you'll be next 2015-06-23T15:36:57 Ya I did post on the mailing list. It also introduced the no_cpu.h suggestion 2015-06-23T15:37:31 hi! 2015-06-23T15:37:31 Last week I had some issues with different byte size of the executable info obtained from Qemu and the symbols read from the analyzed object files. 2015-06-23T15:37:31 It seems to be working incorrect because 10 of the 11 sample tests yield 'invalid' as result and then the traces provided by Qemu are somehow inconsistent. 2015-06-23T15:37:35 So okay, I will bring it up again. And now that the cache issue is solved, I can steadily focus on SMP 2015-06-23T15:38:10 krohini: you can also ping the previous email in case you weren't satisfied with the responses 2015-06-23T15:38:18 I haven't seen any recent email hermann_19829. 2015-06-23T15:38:21 I am currently investigating why these test cases are invalid. 2015-06-23T15:38:22 I also pinged Krysztof and asked him for help, but he did not answer yet. 2015-06-23T15:38:29 Did you improve the diagnostic output? 2015-06-23T15:38:36 gedare. ok. Thanks! 2015-06-23T15:38:48 *** anandkp92 has joined #rtems 2015-06-23T15:38:52 yes I dumped the executable info 2015-06-23T15:39:02 and added some debug outputs 2015-06-23T15:39:17 In covoar, did you manage to get it to print out which two executables had conflicting sizes? 2015-06-23T15:39:29 *** beng-nl-droid has quit IRC 2015-06-23T15:39:37 I was hoping to help you analyze what got it confused. It has happened before. 2015-06-23T15:39:42 *** beng-nl-droid has joined #rtems 2015-06-23T15:40:27 hi. hope i am not too late 2015-06-23T15:40:52 I have the executable info which provides an input for the coverage map and the object files which are analyzed 2015-06-23T15:41:03 hi anandkp92 we're still going. 2015-06-23T15:41:12 the difference is between these two 2015-06-23T15:41:33 hermann_19829: DrJoel is asking if you have produced code modifications to help make debugging covoar easier 2015-06-23T15:42:00 gedare yes I tried to answer his questions :) 2015-06-23T15:42:20 *** beng-nl-droid has quit IRC 2015-06-23T15:42:27 *** beng-nl-droid has joined #rtems 2015-06-23T15:42:27 I want to have those changes so I can run it and help debug it. :) 2015-06-23T15:42:29 sorry to interrupt . some more reviews on the patch will be helpful. The change to the configuration that has been made,I suppose requires some discussion (it was something I did not expect). 2015-06-23T15:42:48 *** zoso has quit IRC 2015-06-23T15:42:49 krohini: i'll get a look to it soon i haven't gotten through the ml this morning yet. 2015-06-23T15:43:12 ok. thanks! 2015-06-23T15:43:13 krohini: discussion is part and parcel of open source development. 2015-06-23T15:43:19 DrJoel nice! Should I send you a patch? I also commited some of the changes to my github repo 2015-06-23T15:43:23 hermann_19829: yes! 2015-06-23T15:43:28 oh yes :) 2015-06-23T15:43:29 :) 2015-06-23T15:44:04 hermann_19829: we'd really like for you to merge the patches from krzysztof and your changes 2015-06-23T15:45:03 Yep. And let's get things merged into the main code. 2015-06-23T15:45:24 gedare ah you are right. since i forked krysztof's repo I also have to take care about his changes. 2015-06-23T15:45:58 Can you run tests with coverage disabled? That was broken at one point With krystof's code 2015-06-23T15:46:31 DrJoel ok! 2015-06-23T15:46:57 I have to try running the tests with coverage disabled 2015-06-23T15:46:58 I want to be testing against our main repo with limited patches. 2015-06-23T15:47:07 Not breaking is a requirement. :) 2015-06-23T15:47:34 DrJoel I think we can satisfy that :) 2015-06-23T15:47:47 :) 2015-06-23T15:48:07 Don't struggle in silence. Struggle on devel@. :) You will get help. 2015-06-23T15:48:41 *** ketul has quit IRC 2015-06-23T15:48:51 :) 2015-06-23T15:48:57 ok thanks hermann_19829 i think you should focus on cleaning up your code and submitting patches that (incrementally) improve things without breaking them. 2015-06-23T15:49:13 ok, but most of the time I do not really know what to ask 2015-06-23T15:49:26 That is a nice quote, DrJoel :) 2015-06-23T15:49:46 salpha: you're upnext 2015-06-23T15:49:51 hermann_19829: anything else from you? 2015-06-23T15:50:00 ok gedare thanks! 2015-06-23T15:50:09 hermann_19829: you can always just describe what you've been doing and where you're trying to go 2015-06-23T15:50:09 maybe one question 2015-06-23T15:50:19 ok 2015-06-23T15:50:37 do i also have to prepare something for that midterm evaluation? 2015-06-23T15:51:00 ah, DrJoel ? I don't think SOCIS has an official one, but I'd still like a report from you and salpha about your progress, and you should aim to deliver something 2015-06-23T15:51:33 our time frame is a bit different than GSoC 2015-06-23T15:52:02 yes sure gedare! just wanted to ask if there is something special to do that I haven't seen yet. 2015-06-23T15:52:07 26 June ~ 3 July is only 1 month past when we have started. 2015-06-23T15:52:37 I would prefer we follow the same patter with a different deadline. A mid-term report when it is your mid-term 2015-06-23T15:52:55 ok DrJoel 2015-06-23T15:53:11 I am not sure what we do administratively with SOCIS for mid-term. I think we do a pass/fail survey like GSOC 2015-06-23T15:53:27 yeah, so you both might have a few weeks yet 2015-06-23T15:53:35 thanks hermann_19829, salpha you're up 2015-06-23T15:53:51 in the last week, i first went through Ballista type hierarchies and the way it inject data type based faults to the system under test. 2015-06-23T15:53:58 then i updated the blog by introduction of CRASH scale. 2015-06-23T15:54:03 afterwards spent time on GRINDER source code. i have almost implemented start, stop, and restart functions of CUEAbstraction which control execution of qemu. 2015-06-23T15:54:07 many of the setup preparation for GRINDER has been done and now i can extend it to support rtems. 2015-06-23T15:54:47 i have a little technical difficulty to update GitHub but it should happen in this week. 2015-06-23T15:55:24 also there are 2 blog posts i'm going to send. one about a background on FI terminology and stuff. 2015-06-23T15:55:44 the other about tips on compiling and preparing environment for GRINDER. 2015-06-23T15:56:05 now i should implement runExperiment to pass test campaigns to qemu to be executed. 2015-06-23T15:56:26 runExperiment is a function in Qemu Abstraction of GRINDER. 2015-06-23T15:56:30 i have the 6-page GRINDER paper downloaded waiting to be read. if there are any other papers or resources to read to "catch up" on this material please refer them to me. 2015-06-23T15:56:54 github should be easy. there is extensive help online there. 2015-06-23T15:56:54 sure. 2015-06-23T15:57:06 ok good those blog posts should help 2015-06-23T15:57:29 there is a paper for taxonomy which i'll refer to, as you mentioned now. 2015-06-23T15:57:40 there are also some technical discussions for considering different options which i'll send you via email. 2015-06-23T15:57:45 are you getting close to a simple (no faults) demo of GRINDER+Qemu? 2015-06-23T15:58:05 and then getting that to run RTEMS, that is a nice milestone 2015-06-23T15:58:13 yeah 2015-06-23T15:58:20 ok thanks that will be hlepful 2015-06-23T15:58:25 i'll try to do that for midterm eval 2015-06-23T15:58:34 super 2015-06-23T15:58:39 thanks salpha anything else? 2015-06-23T15:58:48 nope, thanks gedare. 2015-06-23T15:58:50 jrcatbagan: you're up now 2015-06-23T15:59:00 hi 2015-06-23T15:59:04 *** anandkp92__ has joined #rtems 2015-06-23T15:59:15 hi jrcatbagan i skimmed your patches, and ed is working through them 2015-06-23T15:59:29 ok so I am able to reach the Umon command line already 2015-06-23T15:59:30 *** dr__house|away is now known as dr__house 2015-06-23T15:59:37 just need to finish fixing the data abort exception 2015-06-23T15:59:52 which is causing the umon commands to fail 2015-06-23T16:00:01 I already submitted the patches to umon-devel 2015-06-23T16:00:08 Ed already tested the patches 2015-06-23T16:00:26 Just need to fix the file headers, minor cleanups 2015-06-23T16:00:30 and should be ready to be merged 2015-06-23T16:00:40 *** anandkp92 has quit IRC 2015-06-23T16:00:44 In regards to the data exception I was trying to resolve 2015-06-23T16:00:53 I was talking with Ed that we should transition now to using 2015-06-23T16:00:56 the RTEMS complier 2015-06-23T16:01:02 and get Umon built that way 2015-06-23T16:01:18 also in order to avoid any problems later down the road 2015-06-23T16:01:31 if we transitioned later on 2015-06-23T16:01:41 so I am shifting gears a bit 2015-06-23T16:01:50 ok that sounds like a good idea to me 2015-06-23T16:01:53 and looking into getting Umon built using the RTEMS compiler today 2015-06-23T16:02:16 I spoke with Ed, and we were able to pinpoint why the data exception is occuring when a Umon command was being executed 2015-06-23T16:02:27 I already have in mind numerous solutions 2015-06-23T16:02:28 I think so. The other compiler is rather random and unknown from our general perspective. 2015-06-23T16:02:43 but the best solution would be to transition to building with RTEMS compiler 2015-06-23T16:02:51 and focusing the issues should any arise 2015-06-23T16:03:00 yes, it may bring a different set of problems, and might fix some issues you were facing with libc 2015-06-23T16:03:04 rather than focusing on the issues with the arm-none-eabi-gcc/toolset 2015-06-23T16:03:06 (or exacerabate them) 2015-06-23T16:03:16 as the problems might not be relevant when the transistion is over 2015-06-23T16:03:16 hmm, not a real word but i kind of like it 2015-06-23T16:03:33 yeah exactly 2015-06-23T16:03:35 Is printf() used for IO or something internal to umon? 2015-06-23T16:03:38 ok that sounds good 2015-06-23T16:03:46 Just need to finish fixing the patches I recently submitted, namely the file headers 2015-06-23T16:03:47 *** dougk_ has quit IRC 2015-06-23T16:03:52 and should be ready to be merged 2015-06-23T16:03:54 its internal to Umon 2015-06-23T16:04:09 Ed was saying about a template to follow? 2015-06-23T16:04:26 If you stick to string and memory methods, everything from libc should work. We just need to stick to a known subset of methods. 2015-06-23T16:04:36 THe template would be the file header template. 2015-06-23T16:05:06 he was saying something about doxygen? 2015-06-23T16:05:15 so I looked up the documentation at rtems.org 2015-06-23T16:05:27 should this be something I should consider incorporating into the code? 2015-06-23T16:06:01 talk with Ed about it, but I think that is the direction to go 2015-06-23T16:06:08 help him with defining a style too 2015-06-23T16:06:14 ok, cool 2015-06-23T16:06:24 maybe create a subsection of the wiki devoted to umon 2015-06-23T16:06:28 by midterm evaluations, I expect to have Umon built using the RTEMS compiler 2015-06-23T16:06:36 ok 2015-06-23T16:06:37 I already have a previous version, version 4.11 2015-06-23T16:06:45 I definitely be asking on devel if I encounter any issues 2015-06-23T16:06:54 Ill also have the patches ready to be submitted 2015-06-23T16:07:06 great. note that 4.11 version is (still) a moving target since we haven't released rtems 4.11 2015-06-23T16:07:09 Basically the patches clean up the BBB port and the Umon sources so it could be build 2015-06-23T16:07:11 *built 2015-06-23T16:07:18 ok 2015-06-23T16:07:30 ok thanks jrcatbagan anything else? 2015-06-23T16:07:37 But I'll try to get it to build with RTEMS compiler and have a patch developed 2015-06-23T16:07:37 anandkp92__: you'll be next 2015-06-23T16:07:41 to incorporate the change 2015-06-23T16:07:53 that is all for now 2015-06-23T16:08:10 ok 2015-06-23T16:08:35 hi. This week, I have the base of the gui, that is just the looks part ready. Using Amar's functions I am able to fetch all the required options. Currently I am working on the panels for the Boolean parameters, how they'd appear. After that is done I will incorporate that into the gui. 2015-06-23T16:09:02 Same would be done for the other type of options as well. 2015-06-23T16:09:22 *** yangqiao has quit IRC 2015-06-23T16:09:37 ok 2015-06-23T16:10:22 i haven't looked at your code in a bit, hopefully amar has, and i'll try to take another glance if i can 2015-06-23T16:10:34 yes. definitely. 2015-06-23T16:11:08 also, I have one small doubt. a more general git doubt, in fact 2015-06-23T16:11:30 when should i decide when to commit my changes to github? 2015-06-23T16:11:32 *** beng-nl-droid has quit IRC 2015-06-23T16:11:56 i know i can commit after what ever changes I make 2015-06-23T16:12:25 push every day 2015-06-23T16:12:40 :) 2015-06-23T16:12:46 that's the easy answer 2015-06-23T16:12:47 mm ok. thanks :) 2015-06-23T16:12:54 you should commit basically as you write code 2015-06-23T16:12:57 Use git as a backup. :) 2015-06-23T16:13:02 and push lots of little commits daily 2015-06-23T16:13:16 and when it comes time to submit, you can use 'git-rebase' to make a clean set of patches that you like 2015-06-23T16:13:25 using a branch off your working tree 2015-06-23T16:13:34 At least, that's what I do, and it works quite well for me. 2015-06-23T16:13:53 ok thanks anandkp92__ anythign else? 2015-06-23T16:13:54 i will check it out then :) thanks. 2015-06-23T16:14:02 no, that'll be all. thanks 2015-06-23T16:14:11 ok asuolen i think you're the last one 2015-06-23T16:14:25 hello 2015-06-23T16:14:35 This week I have posted patches for review regarding the RTEMS GPIO API and its rpi implementation. I have also been gathering more references/documentation to help in the I2C porting (by what I have seen only the xilinx bsp currently uses the new I2C framework). 2015-06-23T16:14:39 I am still left, gedare ? 2015-06-23T16:14:39 At this time I am correcting the issues/suggestions pointed on the GPIO API, and intend to send a new version of the patches tomorrow/thursday. I should have documentation (developer/user manual and/or blog/wiki page) posted this week. The I2C work can start right after, and at this point I am not expecting much trouble, but if it does I will post any problem on the ml regarding the new framework. 2015-06-23T16:16:38 sorry sujayraaj thanks you'll be after asuolen 2015-06-23T16:17:27 asuolen: good. I have just responded to your patch again. 2015-06-23T16:18:04 asuolen: yes the new i2c has not been used much. if you run into difficulty send email to devel ml, only sebhub has used it really so far 2015-06-23T16:18:27 will check your mail after the meeting 2015-06-23T16:18:29 but first do due diligence reading the code and any docs :) 2015-06-23T16:18:59 asuolen: did you see the email from Steve about using gpio on the bbb? 2015-06-23T16:19:00 So the Pi and BB will end up using the new GPIO API? Right. So you both are hammering on it. 2015-06-23T16:19:01 yes that is what i have been doing regarging i2c this week 2015-06-23T16:19:25 he may have some good feedback on application requirements. also ppisa had at some point given advice on it too 2015-06-23T16:19:36 yes that is the idea. From i have seen from the last patch from ketul he is including bits of the generic API, which should not be trhe case 2015-06-23T16:20:01 i told him to get it working and we can refactor out the shared parts 2015-06-23T16:20:08 since i don't want him to be held up on account of you 2015-06-23T16:20:27 more work for him maybe, but also better progress 2015-06-23T16:20:41 *** dougk_ has joined #rtems 2015-06-23T16:21:01 ok, asuolen i think there was a request for grouping pins together at some point 2015-06-23T16:21:17 and i would like a better picture of how the interrupt handling works 2015-06-23T16:21:44 yes one of the suggestions i made yesterday i think was to have the possibility of requesting multiple pins, and those pins would act as a single entity 2015-06-23T16:21:44 both conceptually, and the design, and then the API calls involved, and then a peek "under the hood" at what is done in the code 2015-06-23T16:22:00 i think that was suggested a long time ago by ppisa 2015-06-23T16:22:09 so it seems like it would be useful/needed 2015-06-23T16:22:57 Yeah.. we had an API internally years ago that would let you read >1 pin at a time. It was great when say 5 pins encoded something and you needed an atomic read 2015-06-23T16:22:58 yes it can. For instance interfaces based on gpio (such as I2C on the pi) require two pins for that 2015-06-23T16:23:39 ok 2015-06-23T16:23:56 good examples and thorough documentation then are the orders of the day :) 2015-06-23T16:24:02 thanks asuolen anything else? sujayraaj is next 2015-06-23T16:24:08 I recognise that i should explain the behaviou/working of the api better with documentation 2015-06-23T16:24:33 *** gtament has quit IRC 2015-06-23T16:24:37 maybe another more detailed blog post, as a sort of a start of a more formal documentation 2015-06-23T16:25:03 yes, and then figure out what you would like to provide in the way of doc 2015-06-23T16:25:13 either a texinfo chapter, or an asciibook, is preferred 2015-06-23T16:25:29 ok 2015-06-23T16:25:40 thanks asuolen. You're up sujayraaj. hi edsiper 2015-06-23T16:25:56 1. Successfully built monkey. Had to write a version of sendfile for myself, we don't have it in any libs, and other small hacks. I have a patch ready to be pushed upstream to monkey, and I have sent it to chris (and cc to all mentors) for review. 2015-06-23T16:26:10 2. Now I have to test the build and see to it that everything is working as it should. I am reading about IMFS and tar files to load the monkey configuration files in emulation. 2015-06-23T16:26:20 3. If possible this week I would like to get sendfile backed up from freeBSD to libc. 2015-06-23T16:26:44 *** gtament has joined #rtems 2015-06-23T16:27:18 sujayraaj: there should be some demos in examples-v2.git using imfs and tar 2015-06-23T16:27:42 Yeah. There are few in testsuites too I think. I am looking at them. 2015-06-23T16:27:48 sujayraaj: note that RTEMS uses newlib 2015-06-23T16:28:21 Yeah. I have to look at the logistics of getting sendfile in, haven't yet. will do. 2015-06-23T16:28:23 asuolen If the intention is to add it to the User's Guide, it needs to be in texinfo. There is a shell script in new_chapters to help generate a shell of a chapter. I am happy to help you when you have trouble with texinfo. It is rather arcane at this point. 2015-06-23T16:29:21 Hmmm... sendfile() is not POSIX. From the Linux man page: 2015-06-23T16:29:36 Not specified in POSIX.1-2001, or other standards. 2015-06-23T16:29:36 Other Unix systems implement sendfile() with different semantics and 2015-06-23T16:29:36 prototypes. It should not be used in portable programs. 2015-06-23T16:30:14 is there an alternative that monkey could use? 2015-06-23T16:30:20 Monkey may need to address that as a portability issue. 2015-06-23T16:30:46 I wrote a simplistic one using read/write 2015-06-23T16:31:00 I don't know. This needs to be brought up on devel@ so we can decide if Monkey should detect it missing and provide its own, consider adding it to newlib for Linux compatability, etc. 2015-06-23T16:31:03 i guess that is suboptimal 2015-06-23T16:31:27 gedare, it is indeed. 2015-06-23T16:32:12 ok. probably Monkey should be aware of the OS it targets, and have different wrappers to use sendfile, or some other efficient mechanism to copy data 2015-06-23T16:32:22 Yeah.. man page states it is all done in kernel space so more efficient because it avoids copying in and out of kernel/user space. I don't think we have that concern. 2015-06-23T16:32:33 Something to discuss. *BSD may not have it. 2015-06-23T16:32:38 regardless, we'll need something like it, and supporting sendfile in newlib (if it is not conflicting with other definitions) could be one way, if the newlib maintainers approve it. 2015-06-23T16:32:44 Or BSD may have a conflicting definition. 2015-06-23T16:33:02 gedare, yes, monkey uses preprocessors to use sendfile according to the os it runs on 2015-06-23T16:33:02 Yeah.. I lean to a Monkey issue which we have to support with a backup implementation. 2015-06-23T16:33:34 krohini: does it work for you to use MMU_CODE_CACHED in the bsp_section_text part? 2015-06-23T16:33:48 ok sujayraaj i think that is the main issue for you to resolve next 2015-06-23T16:33:58 Ya, it gives the same results 2015-06-23T16:33:58 let's not try to solve it here. Bring it up on devel@ and get it resolved. 2015-06-23T16:33:59 that involves communicating 2015-06-23T16:34:04 krohini: the same "good" results? 2015-06-23T16:34:17 yes :D The improved ones 2015-06-23T16:34:45 ok, and can you use ARMV7_CP15_START_DEFAULT_SECTIONS 2015-06-23T16:34:48 So I dont really understand why the section was defined the way it was 2015-06-23T16:34:51 sujayraaj: anything else? 2015-06-23T16:35:04 No. Thank you gedare. 2015-06-23T16:35:41 That is what I am looking at now. 2015-06-23T16:35:46 thanks sujayraaj 2015-06-23T16:35:53 ok that's the end of the meeting then! thanks everyone for coming 2015-06-23T16:35:58 ok krohini sounds good 2015-06-23T16:37:07 krohini: To be clear, if the rpi2 can't use the default shared implementation of mminit, it should provide its own c file implementation in the bsp directory. 2015-06-23T16:37:47 ok. I suppose that will be common to both the Pi variants 2015-06-23T16:38:35 well, no, the rpi2 likely has its own requirements due to smp? 2015-06-23T16:39:45 But, its just a single core running now. And the current performance on Pi 1 is not as expected 2015-06-23T16:40:15 ah 2015-06-23T16:40:17 I think testing it will be quick 2015-06-23T16:40:20 sorry yes the memory mapping 2015-06-23T16:40:23 thought you meant the mminit 2015-06-23T16:40:59 oh 2015-06-23T16:41:27 The code in shared mminit is out of place 2015-06-23T16:41:35 yeah 2015-06-23T16:41:50 will organize it soon.. Shouldnt take time 2015-06-23T16:42:38 *** salpha has quit IRC 2015-06-23T16:42:47 ok thanks krohini 2015-06-23T16:43:19 another thing 2015-06-23T16:44:22 with regard to SMP, I need to go along the lines of arm-a9mpcore-smp? 2015-06-23T16:46:01 a big difference is that there is no GIC in Pi2 2015-06-23T16:50:37 you should probably study what is available and identify the gaps 2015-06-23T16:50:43 *** Davidbrcz has joined #rtems 2015-06-23T16:52:56 I want to get the scope clear 2015-06-23T16:53:25 [trac] GSoC/2015 edited by jrcatbagan (Added status update for June 23, 2015) http://devel.rtems.org/wiki/GSoC/2015 2015-06-23T16:53:45 so that I can go about understanding accordingly 2015-06-23T16:54:04 For a start I need to get the other cores to start executing code (which will be getting their initializations done from start.S) 2015-06-23T16:54:52 Then I traced the SMP initialization from bootcard as discussed. 2015-06-23T16:56:50 yeah 2015-06-23T16:57:02 For other ARM bsps, all cores start together and have the initializations done before rtems_data_structure_initialize() 2015-06-23T16:59:05 from the bsp_start_hooks 2015-06-23T17:00:51 I need to add code for this such that this can be done for Pi 2 as well 2015-06-23T17:02:08 yeah that sounds right 2015-06-23T17:02:37 so I need to add it conditionally for Pi 2 2015-06-23T17:02:37 conceptually, the boot core gets to the point where it tells the other cores to go start executing from start.S and can wait somewhere for them to finish initializing 2015-06-23T17:02:50 and from there the code should be very similar to other targets 2015-06-23T17:02:56 um 2015-06-23T17:03:18 *** monstr has quit IRC 2015-06-23T17:03:34 you mean to distinguish rpi from rpi2? 2015-06-23T17:03:39 Also make sure your core variant has the synchronization instructions used in the context switch code. Probably just a review but best to be certain 2015-06-23T17:04:56 umm..okay 2015-06-23T17:05:31 you can also condition it on RTEMS_SMP there. 2015-06-23T17:05:54 the bsp start hooks are executed by all cores together right? for other arm bsps? 2015-06-23T17:06:22 i don't really know 2015-06-23T17:06:41 but the hooks themselves are (can be) BSP-specific 2015-06-23T17:07:18 Ya, that is right 2015-06-23T17:08:03 Where exactly does the interprocessor interrupt from boot core come into picture? 2015-06-23T17:08:36 for ARM 2015-06-23T17:09:44 sorry i can't trace into the code right now 2015-06-23T17:10:01 because from what I saw in start.S there was no condition where only n=boot code is running 2015-06-23T17:10:18 *boot core 2015-06-23T17:11:15 ok. I will resort to the mailing list and get this cleared 2015-06-23T17:11:32 yeah. if i can get to it later i'll try 2015-06-23T17:11:39 or maybe someone else will clear it up 2015-06-23T17:12:01 ok thanks. 2015-06-23T17:12:08 thanks DrJoel 2015-06-23T17:19:56 *** asuolen has quit IRC 2015-06-23T17:24:27 *** hermann_19829 has quit IRC 2015-06-23T17:35:41 *** gtament has quit IRC 2015-06-23T17:37:47 You are welcome. 2015-06-23T17:53:12 *** ragunath has quit IRC 2015-06-23T18:22:06 *** sujayraaj has quit IRC 2015-06-23T18:41:55 *** ketul has joined #rtems 2015-06-23T18:44:04 hi gedare 2015-06-23T18:47:18 hi ketul 2015-06-23T18:47:53 hi beng-nl 2015-06-23T18:48:54 *** krohini has quit IRC 2015-06-23T18:51:29 hello ketul need something? or just saying hi? 2015-06-23T18:51:55 no gedare want to discuss about your recent comments 2015-06-23T18:52:29 Can the gpio.c test case be written without having to specify the BSP? 2015-06-23T18:52:58 about this. I think it has definitions to all the functions so it would be bsp specifig 2015-06-23T18:53:20 can you link me to it in your github 2015-06-23T18:53:42 ya sure let me 2015-06-23T18:53:56 https://github.com/ketul93/rtems 2015-06-23T18:54:40 beng-nl: just for curiosity want to know will we require touchscreen on BBB running rtems :) ? 2015-06-23T18:54:47 :-) 2015-06-23T18:54:58 I mean may be in future 2015-06-23T18:55:30 or we would be targeting different sensors to be interfaced 2015-06-23T18:55:49 gee tough question 2015-06-23T18:56:05 ohh :) 2015-06-23T18:56:12 because usually the answer is "would be nice" but not necessarily the best thing to do next :) 2015-06-23T18:56:42 then again what is is a combination of what application developers would want to use and what you'd be motivated to work on 2015-06-23T18:57:08 you know when started referring TRM it doesn't have section for ADC. It is covered under TSC part :) 2015-06-23T18:57:09 but choosing between yes and no.. yes i think there'd be an application :-) 2015-06-23T18:57:53 aha :-) 2015-06-23T18:57:59 ADC is in the touch screen module? 2015-06-23T18:58:04 yep 2015-06-23T18:58:08 gee interesting 2015-06-23T18:58:57 so in BBB different configuration are there . 1 for only TSC 2. for General purpose 3. for both 2015-06-23T18:59:10 ketul: change it to bsp.h instead, and pick up beagleboneblack.h from there using CPP? 2015-06-23T18:59:30 what does TSC mean? 2015-06-23T18:59:42 Touchscreen Controller :) 2015-06-23T18:59:50 gotcha. 2015-06-23T19:00:18 i think that is a good stretch goal for your second half 2015-06-23T19:00:43 because of it has multipurpose functions looks like complex 2015-06-23T19:01:29 sure 2015-06-23T19:02:01 first you should get the adc part working, but after that it seems logical to pursue the full tsc support, to me 2015-06-23T19:02:11 +1 2015-06-23T19:02:34 gedare: https://devel.rtems.org/wiki/TBR/User/BenGras for your last question. I mean followed that approach 2015-06-23T19:02:51 ya gedare I am aiming to work only the general adc part 2015-06-23T19:02:56 as of now 2015-06-23T19:03:19 what an odd user page 2015-06-23T19:03:24 :'( 2015-06-23T19:03:34 :-) 2015-06-23T19:03:50 i made it for one purpose :) (ironically GP-IO) 2015-06-23T19:04:03 :) 2015-06-23T19:04:38 so anyway, a generic sample should not use BSP specific stuff 2015-06-23T19:05:10 but pulling in bsp.h should get you the right set of features "somehow", in my opinion 2015-06-23T19:05:30 yeah. no. i didn't like that part of Andre's design (BSP-specific #ifs in the application) - one exception where i think it's needed is board-specific stuff such as what are LEDs and the like 2015-06-23T19:05:43 because that actually is bsp-specific 2015-06-23T19:06:11 one solution to that is to use a generic macro name that gets instantiated by each supported target. 2015-06-23T19:06:37 yeah 2015-06-23T19:06:38 (and non-supported targets should still be able to run this sample, but have stub functions that return errors for everything) 2015-06-23T19:07:40 we need to be able to build every single program in the testsuites for every BSP. If the test can't be written that way, it can't get merged into RTEMS. 2015-06-23T19:07:54 There's definitely value in sorting this out though 2015-06-23T19:08:25 yeah i suppose there could be some kind of bsp-generic way to talk about "first led," "second led" if available 2015-06-23T19:08:37 so applications are still not-bsp-specific and can still do somewhat-bsp-specific stuff 2015-06-23T19:08:44 right 2015-06-23T19:08:50 at least for simple stuff this will work 2015-06-23T19:09:00 and that's all we need for testing the framework 2015-06-23T19:09:06 right 2015-06-23T19:09:21 more complex example applications can be done standalone or put into our examples repo (which already does some led stuff) 2015-06-23T19:09:24 yep and I think most of the board has usr leds for testing 2015-06-23T19:10:05 you could also consider writing functions to query the BSP for the pin numbers for specific features 2015-06-23T19:10:27 e.g. "rtems_gpio_get_led_pin(...)" I don't know how much value this has though 2015-06-23T19:10:29 I think we should make one documentation for that so you know every devel can do in that approach. 2015-06-23T19:11:14 in that way we can make the approach clear 2015-06-23T19:11:36 regardless, for the final submission of code to commit to RTEMS, make sure it adheres to the coding rules, which (should) include generic test code. 2015-06-23T19:11:46 *** yangqiao has joined #rtems 2015-06-23T19:13:06 gedare: one prob is with 80 columns rule. how can it be followed with this case "if (rtems_gpio_configure_pin_digital_out(&led_usr0_handle, BBB_LED_USR0) == GPIO_SUCCESSFUL)" 2015-06-23T19:13:30 here the function naming is the challenge. 2015-06-23T19:13:46 https://devel.rtems.org/wiki/Developer/Coding/80_characters_per_line 2015-06-23T19:15:02 gedare: already referred. In that different conditions are checked. here only 1 is there 2015-06-23T19:15:28 if ( 2015-06-23T19:15:43 rtems_gpio_configure_pin_digital_out( 2015-06-23T19:15:51 &led_usr0_handle, 2015-06-23T19:15:59 okk got 2015-06-23T19:16:01 BBB_LED_USR0 2015-06-23T19:16:11 ketul: oh yeah, you can't leave stuff that has to happen in asserts 2015-06-23T19:16:20 ) == GPIO_SUCCESSFUL 2015-06-23T19:16:22 ) { 2015-06-23T19:16:37 good catch. no side-effects in assert 2015-06-23T19:17:29 that deserves its own page in the wiki 2015-06-23T19:18:51 beng-nl: you are talking about : assert (status == RTEMS_SUCCESSFUL) ? 2015-06-23T19:19:33 that is ok 2015-06-23T19:19:44 not ok: assert(do_something() == RTEMS_SUCCESSFUL); 2015-06-23T19:19:54 because the assert() might be compiled ourt 2015-06-23T19:19:56 because the assert() might be compiled out 2015-06-23T19:20:05 and then the behaviour changes 2015-06-23T19:21:03 ketul: Use variables for the return status and check them 2015-06-23T19:21:10 yep 2015-06-23T19:21:38 think: if I comment out this assert, will the code still work? 2015-06-23T19:21:40 That also makes the code easier to fit in 80 columns, modify to print diagnostics, a variable to print in gdb. 2015-06-23T19:22:31 Also assert's should only be used for horrible inconsistencies. Normal operational issues should be normal statuses. Sometimes hard to decide which is which. 2015-06-23T19:22:55 Do we allow assert() in BSP/drivers? Or should be it the rtems internal assert controlled by debug build? 2015-06-23T19:24:59 ^ "that deserves its own page in the wiki" 2015-06-23T19:25:04 we don't have a good set of rules around it 2015-06-23T19:25:29 *** krohini has joined #rtems 2015-06-23T19:25:31 yep please 2015-06-23T19:26:06 hi gedare, 2015-06-23T19:26:14 hi 2015-06-23T19:26:15 I'm thinking that maybe I should just decalre an fix length array for the buffer, and then fill the data in. Each property has two inline function to pack/unpack data into/from the buffer at a given index, return an length so that we can move on to the next one. No structure definition for buffer nor tags. Just use functions to pack and unpack . How do you think. 2015-06-23T19:28:53 DrJoel: for "Use variables for the return status and check them" you mean to say in "if (rtems_gpio_configure_pin_digital_out(&led_usr0_handle, BBB_LED_USR0) == GPIO_SUCCESSFUL)" I should replace "GPIO_SUCCESSFUL" with a variable. Am I understanding correct? or you mean to say any other thing. 2015-06-23T19:31:55 yangqiao: sounds like you are proposing to implement a ring buffer to store mailbox data? 2015-06-23T19:32:24 or, do you anticipate it is like a stack? 2015-06-23T19:32:48 i don't have a clear picture of how the mailboxes are going to be used 2015-06-23T19:33:19 if the message ordering is queue-based, then a ring buffer makes good sense 2015-06-23T19:33:52 if messages can be removed in arbitrary order from the mailboxes, that leaves holes in any array-based buffer approach 2015-06-23T19:34:10 if message ordering is lifo, a stack should be used 2015-06-23T19:34:42 ketul: 2015-06-23T19:34:45 x = rtems_gpio_configure_pin_digital_out(...); 2015-06-23T19:35:18 if ( x != GPIO_SUCCESSFUL ) { ... } 2015-06-23T19:35:20 or whatever 2015-06-23T19:35:22 ok gedare. then I should compare x 2015-06-23T19:35:25 yaah ok 2015-06-23T19:36:30 actually truly understood but want to be clear. I think my question was not clr :) anyways got it 2015-06-23T19:38:45 :) 2015-06-23T19:40:35 Thanks gedare. Getting old and mind wanders 2015-06-23T19:41:07 I don't think assert() should be used in BSPs and drivers or an inside the rtems tree at all. We should reserve it for user applications. IMO 2015-06-23T19:41:22 tests should use the test macros to check conditions. Not assert. 2015-06-23T19:41:55 ya thats true. 2015-06-23T19:42:18 rather, we need an assert that is safe to use 2015-06-23T19:43:59 I try to explain. Before, I was always thinking about using structures to describe the buffer, and its tags. As there are many tags, and they all have common parts and different parts. It was very complicate to use a simple way to unify the describtion. You've proposed an zero length array. It do can describe all structures but we can't allocate memory dynamically in kernel driver. I've tried to allocate an array and use it for the b 2015-06-23T19:44:00 uffer structure. But I didn't make it. I don't know how can I pre allocate memory for a structure containing zero length array. While, I was thinking just now. The whole buffer can be described as an array of unsigned int. We just need to fill in the code and data in order. So I imagine , the buffer can be like an stack. We may use seperate functions to fill in data for different tags. I give you an exemple. 2015-06-23T19:50:53 gedare: something like: 2015-06-23T19:51:05 gedare : something like 2015-06-23T19:51:24 unsigned int buffer[MAX_BUFFER_LEN] __attribute__((aligned (16))); // declaration 2015-06-23T19:51:35 int index = 0; 2015-06-23T19:51:37 index += mailbox_init_buffer(_buffer) //init return the begin index for following tags 2015-06-23T19:51:44 index += mailbox_add_request_get_display_size(_buffer, _index) //fill in datas 2015-06-23T19:51:50 index += mailbox_add_request_xxxxx(_buffer, _index) 2015-06-23T19:52:05 ... // then send the buffer 2015-06-23T19:52:10 index = 0; 2015-06-23T19:52:18 index += mailbox_check_buffer_(buffer) 2015-06-23T19:52:23 index += mailbox_get_response_get_display_size(_buffer, _index, width, height) 2015-06-23T19:52:26 ... etc 2015-06-23T19:52:40 ok some questions 2015-06-23T19:52:48 what do you do when index == MAX_BUFFER_LEN? 2015-06-23T19:54:12 This is the end of buffer and we should do nothing I think. Even if we tried to get additional data from the buffer 2015-06-23T19:54:20 2) What happens if "add_request" and "get_response" are not done in the same order, like, the producer does add_request_xxx and add_request_yyy and the consumer does get_response_yyy and get_response_xxx 2015-06-23T19:55:23 yes, it's annoying It has to be in the same order... 2015-06-23T19:55:31 but, after for example mailbox_get_response_get_display_size(), isn't the buffer space that was storing that value available for reuse? 2015-06-23T19:57:56 How do we anticipate mailboxes will be used? 2015-06-23T19:58:19 After a message is read, does it go away? 2015-06-23T19:59:13 yes it becomes much complicated. 2015-06-23T20:01:36 Or , maybe we just use seperate functions for each request(tag). Do not pack them in one huge buffer as these operations are not frequentely used. 2015-06-23T20:02:22 Like the freebsd way. one function for each request. Developers won't bother the buffer things. 2015-06-23T20:04:20 Do the requests take long enough to justify queuing them up? Can they be "do X" and internally that puts a request, gets the response and has exclusive access during that period? 2015-06-23T20:06:38 In other words, is it always the case the same thread is going to do "add_request" and "get_response" soon after? 2015-06-23T20:10:24 yangqiao: If that is the case, then you can provide one function with a lock inside it, that goes the add/get, and has a pre-allocated buffer sufficient for the largest possible response. 2015-06-23T20:11:23 you can also provide a single function per request type (like bsd), and have a similar scheme with a lock inside it and a pre-allocated buffer for it 2015-06-23T20:11:24 yes, we send an request to get the response. I'm not sure about how long it would take. Afterall these interface are hardeware related. To get hardware informations/ set hardware states etc. They are mostly used for some driver to initialize, not frequently used. 2015-06-23T20:12:09 that leads me to thinking it should just be a simple "do x" interface and keep it simple under the hood. 2015-06-23T20:13:08 yeah. you can still make use of your buffer+tag structs if you like, especially if you use multiple functions where each function just defines intenrally which kind of tagged buffer it uses 2015-06-23T20:15:18 you can statically allocated zero-length arrays too it doesn't need to be dynamic. 2015-06-23T20:15:58 if you provide an initializer the size will be determined by how big the array is in the initializer 2015-06-23T20:16:09 it's explained in the gcc manual 2015-06-23T20:21:14 I may need some time to think through. Thank you gedare DrJoel . I'll be back if I can't still handle it. 2015-06-23T20:24:03 ok take your time to get it right 2015-06-23T20:26:30 *** Davidbrcz has quit IRC 2015-06-23T20:30:02 bye everyone. need to sleep :) have a good day ! 2015-06-23T20:30:11 *** yangqiao has quit IRC 2015-06-23T20:30:27 ty gedare beng-nl and DrJoel 2015-06-23T20:32:48 good night 2015-06-23T20:33:05 looked like you had a nice father's day Gedare 2015-06-23T20:34:51 yeah 2015-06-23T20:34:56 np ketul 2015-06-23T20:35:01 kind of standard weekend fare really 2015-06-23T20:35:11 same here... just added ice cream cake :) 2015-06-23T20:35:19 :) 2015-06-23T20:35:19 :-) 2015-06-23T20:35:22 nice to hear guys 2015-06-23T20:35:28 :) 2015-06-23T20:35:31 *** ketul has left #rtems 2015-06-23T20:47:50 *** jkridner has quit IRC 2015-06-23T20:48:52 *** jkridner has joined #rtems 2015-06-23T20:53:25 *** Davidbrcz has joined #rtems 2015-06-23T21:02:54 *** gedare has quit IRC 2015-06-23T21:33:24 *** yangqiao has joined #rtems 2015-06-23T21:36:24 *** yangqiao has quit IRC 2015-06-23T21:58:25 *** javamonn has quit IRC 2015-06-23T22:07:50 *** krohini has quit IRC 2015-06-23T22:15:02 *** jrcatbagan has quit IRC 2015-06-23T22:48:49 [trac] GSoC/2015 edited by Asuol (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-06-23T22:52:34 *** yangqiao has joined #rtems 2015-06-23T22:53:38 *** yangqiao has quit IRC 2015-06-23T23:00:43 *** javamonn has joined #rtems 2015-06-23T23:04:41 *** javamonn_ has joined #rtems 2015-06-23T23:07:12 *** kiwichris has joined #rtems 2015-06-23T23:07:28 *** javamonn has quit IRC 2015-06-23T23:10:39 *** kiwichri_ has joined #rtems 2015-06-23T23:11:48 *** kiwichris has quit IRC 2015-06-23T23:13:36 *** DrJoel has quit IRC 2015-06-23T23:16:17 *** Davidbrcz has quit IRC 2015-06-23T23:16:31 *** javamonn_ has quit IRC 2015-06-23T23:28:20 *** yangqiao has joined #rtems 2015-06-23T23:45:20 *** yangqiao has quit IRC 2015-06-23T23:50:10 *** javamonn has joined #rtems 2015-06-23T23:50:28 *** javamonn has joined #rtems 2015-06-24T00:01:56 *** javamonn has quit IRC 2015-06-24T00:03:26 *** javamonn has joined #rtems 2015-06-24T00:03:56 *** javamonn has quit IRC 2015-06-24T00:04:02 *** javamonn_ has joined #rtems 2015-06-24T00:18:56 *** yangqiao has joined #rtems 2015-06-24T00:19:44 *** javamonn_ has quit IRC 2015-06-24T00:21:05 *** javamonn has joined #rtems 2015-06-24T01:24:23 *** javamonn has quit IRC 2015-06-24T03:15:18 *** javamonn has joined #rtems 2015-06-24T03:20:08 *** javamonn has quit IRC 2015-06-24T03:28:21 *** javamonn has joined #rtems 2015-06-24T03:50:24 *** javamonn has quit IRC 2015-06-24T03:53:48 *** javamonn has joined #rtems 2015-06-24T04:06:54 *** crazy3k has quit IRC 2015-06-24T04:08:05 *** javamonn has quit IRC 2015-06-24T04:11:26 *** crazy3k has joined #rtems 2015-06-24T04:37:52 *** mkhoory-2 has quit IRC 2015-06-24T04:37:53 *** mkhoory-2 has joined #rtems 2015-06-24T04:47:23 *** monstr has joined #rtems 2015-06-24T05:08:52 *** anandkp92__ has quit IRC 2015-06-24T05:22:52 *** mkhoory-3 has joined #rtems 2015-06-24T05:26:23 *** mkhoory-2 has quit IRC 2015-06-24T06:27:23 *** dr__house is now known as dr__house|away 2015-06-24T06:30:38 *** yangqiao has joined #rtems 2015-06-24T06:35:14 *** ketul has joined #rtems 2015-06-24T06:48:55 [git] df8341a by Sebastian Huber: [bsp/gen5200: Simplify interrupt write support] http://devel.rtems.org/changeset/df8341a/rtems 2015-06-24T06:59:14 *** kiwichri_ has quit IRC 2015-06-24T07:17:44 *** Davidbrcz has joined #rtems 2015-06-24T07:40:04 *** Davidbrcz has quit IRC 2015-06-24T07:50:33 *** kiwichris has joined #rtems 2015-06-24T08:32:07 *** ketul has quit IRC 2015-06-24T08:42:31 *** ketul has joined #rtems 2015-06-24T09:08:20 *** yangqiao has joined #rtems 2015-06-24T09:11:45 *** zoso has joined #rtems 2015-06-24T09:39:18 *** ketul has quit IRC 2015-06-24T09:40:29 *** ketul has joined #rtems 2015-06-24T10:06:23 *** crazy3k has quit IRC 2015-06-24T10:23:18 *** ketul has quit IRC 2015-06-24T11:39:14 *** zoso has quit IRC 2015-06-24T11:47:18 *** zoso has joined #rtems 2015-06-24T11:48:01 *** ketul has joined #rtems 2015-06-24T11:52:15 *** zoso has quit IRC 2015-06-24T11:52:42 *** zoso has joined #rtems 2015-06-24T12:17:26 *** zoso has quit IRC 2015-06-24T12:24:58 [git] a3fc33f by Sebastian Huber: [score: Assert proper node size] http://devel.rtems.org/changeset/a3fc33f/rtems 2015-06-24T12:33:44 *** zoso has joined #rtems 2015-06-24T12:52:52 *** javamonn has joined #rtems 2015-06-24T13:31:45 *** yangqiao has joined #rtems 2015-06-24T13:46:02 *** gedare has joined #rtems 2015-06-24T13:46:02 *** ChanServ sets mode: +o gedare 2015-06-24T13:47:40 *** javamonn has quit IRC 2015-06-24T13:53:36 *** javamonn has joined #rtems 2015-06-24T15:00:35 *** monstr has quit IRC 2015-06-24T15:16:36 *** zoso has quit IRC 2015-06-24T15:17:19 *** yangqiao has quit IRC 2015-06-24T15:31:02 *** javamonn has quit IRC 2015-06-24T15:31:37 *** javamonn has joined #rtems 2015-06-24T15:54:02 *** Davidbrcz has joined #rtems 2015-06-24T16:04:49 *** javamonn has quit IRC 2015-06-24T16:07:02 *** javamonn has joined #rtems 2015-06-24T16:07:37 *** Davidbrcz has quit IRC 2015-06-24T16:28:14 *** Davidbrcz has joined #rtems 2015-06-24T16:42:16 *** Davidbrcz has quit IRC 2015-06-24T17:07:11 *** DrJoel has joined #rtems 2015-06-24T17:07:11 *** DrJoel has joined #rtems 2015-06-24T17:07:11 *** ChanServ sets mode: +o DrJoel 2015-06-24T17:28:24 *** Davidbrcz has joined #rtems 2015-06-24T17:41:11 *** krohini has joined #rtems 2015-06-24T17:56:53 hi.. I am getting a build error during make in my raspberry pi 2 ~/arm-rtems4.11/raspberrypi2/lib/ files. 2015-06-24T17:57:23 paste it in 2015-06-24T18:01:08 Hey everyone 2015-06-24T18:01:15 jr;;p 2015-06-24T18:01:25 or with hands on keyboard correctly .. hello 2015-06-24T18:01:26 :q! 2015-06-24T18:01:46 yeah.. I do that one also. :) 2015-06-24T18:01:55 I'm working on adding something to a BSP, mostly just to play around 2015-06-24T18:02:10 What's the best way to rebuild after changing the Makefile.am ? 2015-06-24T18:02:38 You can run bootstrap (or sb-bootstrap) from any directory with a configure.ac file. 2015-06-24T18:02:49 If you add a .h file, then you need to run bootstrap -p 2015-06-24T18:02:54 In file included from ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/system.h:23:0, 2015-06-24T18:02:56 from ../../../../../../../../rtems/c/src/../../cpukit/score/cpu/arm/cpu.c:29: 2015-06-24T18:02:57 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/score/percpu.h: At top level: 2015-06-24T18:02:59 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/score/percpu.h:316:21: error: field 'time_of_last_context_switch' has incomplete type 2015-06-24T18:03:00 Timestamp_Control time_of_last_context_switch; 2015-06-24T18:03:16 ^ 2015-06-24T18:03:16 In file included from ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/tasks.h:43:0, 2015-06-24T18:03:16 from ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/config.h:25, 2015-06-24T18:03:16 from ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/config.h:57, 2015-06-24T18:03:16 from ../../../../cpukit/../../../raspberrypi2/lib/include/rtems.h:33, 2015-06-24T18:03:16 from ../../../../../../../../rtems/c/src/../../cpukit/score/cpu/arm/cpu.c:30: 2015-06-24T18:03:17 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/score/thread.h:776:41: error: field 'cpu_time_used' has incomplete type 2015-06-24T18:03:17 Thread_CPU_usage_t cpu_time_used; 2015-06-24T18:03:18 ^ 2015-06-24T18:03:18 In file included from ../../../../cpukit/../../../raspberrypi2/lib/include/rtems.h:41:0, 2015-06-24T18:03:19 from ../../../../../../../../rtems/c/src/../../cpukit/score/cpu/arm/cpu.c:30: 2015-06-24T18:03:19 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/clock.h:74:9: error: unknown type name 'TOD_Nanoseconds_since_last_tick_routine' 2015-06-24T18:03:20 typedef TOD_Nanoseconds_since_last_tick_routine 2015-06-24T18:03:21 If you configure with --enable-maintainer-mode, then sometimes things update OK for you 2015-06-24T18:03:21 ^ 2015-06-24T18:03:23 In file included from ../../../../cpukit/../../../raspberrypi2/lib/include/rtems.h:55:0, 2015-06-24T18:03:24 from ../../../../../../../../rtems/c/src/../../cpukit/score/cpu/arm/cpu.c:30: 2015-06-24T18:03:26 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/ratemon.h:168:40: error: field 'min_cpu_time' has incomplete type 2015-06-24T18:03:28 Thread_CPU_usage_t min_cpu_time; 2015-06-24T18:03:29 ^ 2015-06-24T18:03:30 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/ratemon.h:170:40: error: field 'max_cpu_time' has incomplete type 2015-06-24T18:03:32 Thread_CPU_usage_t max_cpu_time; 2015-06-24T18:03:33 ^ 2015-06-24T18:03:35 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/ratemon.h:172:40: error: field 'total_cpu_time' has incomplete type 2015-06-24T18:03:37 Thread_CPU_usage_t total_cpu_time; 2015-06-24T18:03:38 ^ 2015-06-24T18:03:40 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/ratemon.h:175:40: error: field 'min_wall_time' has incomplete type 2015-06-24T18:03:41 Rate_monotonic_Period_time_t min_wall_time; 2015-06-24T18:03:43 ^ 2015-06-24T18:03:44 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/ratemon.h:177:40: error: field 'max_wall_time' has incomplete type 2015-06-24T18:03:44 Looks like a .h file is missing for time 2015-06-24T18:03:46 Rate_monotonic_Period_time_t max_wall_time; 2015-06-24T18:03:47 ^ 2015-06-24T18:03:49 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/ratemon.h:179:40: error: field 'total_wall_time' has incomplete type 2015-06-24T18:03:50 Rate_monotonic_Period_time_t total_wall_time; 2015-06-24T18:03:52 ^ 2015-06-24T18:03:53 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/ratemon.h:238:43: error: field 'cpu_usage_period_initiated' has incomplete type 2015-06-24T18:03:55 Thread_CPU_usage_t cpu_usage_period_initiated; 2015-06-24T18:03:56 ^ 2015-06-24T18:03:58 ../../../../cpukit/../../../raspberrypi2/lib/include/rtems/rtems/ratemon.h:244:43: error: field 'time_period_initiated' has incomplete type 2015-06-24T18:03:59 Rate_monotonic_Period_time_t time_period_initiated; 2015-06-24T18:04:01 ^ 2015-06-24T18:04:02 make[7]: *** [libscorecpu_a-cpu.o] Error 1 2015-06-24T18:04:04 make[7]: Leaving directory `/home/rck/development/rtems/4.11/b-rpi2/arm-rtems4.11/c/raspberrypi2/cpukit/score/cpu/arm' 2015-06-24T18:04:05 make[6]: *** [all-recursive] Error 1 2015-06-24T18:04:07 make[6]: Leaving directory `/home/rck/development/rtems/4.11/b-rpi2/arm-rtems4.11/c/raspberrypi2/cpukit/score/cpu' 2015-06-24T18:04:08 make[5]: *** [all-recursive] Error 1 2015-06-24T18:04:10 make[5]: Leaving directory `/home/rck/development/rtems/4.11/b-rpi2/arm-rtems4.11/c/raspberrypi2/cpukit/score' 2015-06-24T18:04:11 make[4]: *** [all-recursive] Error 1 2015-06-24T18:04:13 make[4]: Leaving directory `/home/rck/development/rtems/4.11/b-rpi2/arm-rtems4.11/c/raspberrypi2/cpukit' 2015-06-24T18:04:14 make[3]: *** [all] Error 2 2015-06-24T18:04:16 make[3]: Leaving directory `/home/rck/development/rtems/4.11/b-rpi2/arm-rtems4.11/c/raspberrypi2/cpukit' 2015-06-24T18:04:17 make[2]: *** [all-recursive] Error 1 2015-06-24T18:04:19 make[2]: Leaving directory `/home/rck/development/rtems/4.11/b-rpi2/arm-rtems4.11/c/raspberrypi2' 2015-06-24T18:04:20 make[1]: *** [all-recursive] Error 1 2015-06-24T18:04:22 make[1]: Leaving directory `/home/rck/development/rtems/4.11/b-rpi2/arm-rtems4.11/c' 2015-06-24T18:04:23 make: *** [all-recursive] Error 1 2015-06-24T18:05:16 Why is percpu.h being included from cpu.c? No port specific code includes that as far as I can tell 2015-06-24T18:05:26 Where is that command coming from? 2015-06-24T18:05:44 I'm trying to run it in c/src/lib/libbsp/arm/stm32f4 2015-06-24T18:05:52 He cut and pasted a build error. Probably should have used pastebin .. my bad 2015-06-24T18:06:24 Are you adding a .c file to Makefile.am? 2015-06-24T18:06:49 no i did not add a .c file 2015-06-24T18:07:18 Sorry that was for igutek.. 2015-06-24T18:07:31 krohini: why did you add the include of percpu.h to arm/cpu.c 2015-06-24T18:07:40 oh okay.. 2015-06-24T18:07:51 Yeah, I'm adding a c file 2015-06-24T18:08:07 I managed to get it to detect by doing a full bootstrap, configure make cycle 2015-06-24T18:08:30 I was hoping I could only bootstrap/rebuild the changed bsp 2015-06-24T18:08:47 I don't remember doing that.. I did not have to make a change to cpu.c . ok i'll quickly see that 2015-06-24T18:09:10 Umm, I made it real messy with pasting the error directly...... 2015-06-24T18:10:32 igutek: If you had configured with --enable-maintainer-mode, adding a .c file should not have required a re-bootstrap at all 2015-06-24T18:11:07 but if you need to do one, you only have to find the configure.ac located above your mod. So stmf4 directory and run bootstrap script whilethere. 2015-06-24T18:11:50 Also there is sb-bootstrap in rtems-source-builder/source-builder which parallelizes the run. It is great from the top. Runs in 45 seconds on a 6 core machine here 2015-06-24T18:12:15 DrJoel I have not included that header file. 2015-06-24T18:12:59 DrJoel: Thanks! I'm still confused. There is no bootstrap file in c/src/lib/libbsp/arm/stm32f4 2015-06-24T18:13:33 I'll try maintainer mode though, that sounds like the correct way to do things. I'll Google search what it means. Not very familiar with autotools 2015-06-24T18:16:17 The script bootstrap is at the top of the tree. I set the environment variable r to point to the top of my RTEMS source .. so $r/bootstrap. And b gets me to $r/c/src/lib/libbsp. Saves my fingers. :) 2015-06-24T18:16:33 Hopefully 4.11 is the last release with autotools. 2015-06-24T18:16:45 krohini: Use pastebin and do that again 2015-06-24T18:16:58 DrJoel: yup actually required autotools 2015-06-24T18:19:09 Awesome! That seems to be working! Thanks! 2015-06-24T18:26:27 hi igutek did it go well with the sTM folks then? 2015-06-24T18:27:45 krohini, when was the last time you updated your tools, and when was the last time you updated your rtems.git? 2015-06-24T18:28:11 there were big changes to the timestamp implementation about a month ago or so, that requires updating tools if you pulled 2015-06-24T18:28:39 oh ya..I pulled today 2015-06-24T18:28:53 yeah you need to re-build tools then with sb 2015-06-24T18:30:00 ok. i will do that then. 2015-06-24T18:35:34 *** gedare has quit IRC 2015-06-24T19:00:08 gedare: Yeah, though I haven't followed up with them regarding the licensing specifics 2015-06-24T19:23:28 *** dr__house|away is now known as dr__house 2015-06-24T19:29:43 *** ketul has quit IRC 2015-06-24T19:37:36 *** javamonn has quit IRC 2015-06-24T19:37:50 *** javamonn has joined #rtems 2015-06-24T19:41:50 *** sujayraaj has joined #rtems 2015-06-24T19:48:12 *** Davidbrcz has quit IRC 2015-06-24T19:50:40 DrJoel: Thanks very much for your help! That probably saved me 2-3 hours of time :) 2015-06-24T19:53:10 You are welcome. 2015-06-24T19:53:17 whois igutek igutek 2015-06-24T19:53:31 well that didn't work. :) 2015-06-24T19:53:58 Is my name igutek igutek? To maybe answer your query, this is Isaac Gutekunst 2015-06-24T19:54:01 if you need help witht he license discussions, just ask. 2015-06-24T19:54:12 I realized that as soon as I clicked it. 2015-06-24T19:54:17 Bug in new irc client. :( 2015-06-24T19:55:27 No worries, and thanks 2015-06-24T19:56:29 I'm going to see what ST says. All code I (we, RTEMS) would want is BSD licensed, including the HAL code and the CMSIS code 2015-06-24T19:57:46 The worry is that the top level license is less permissive, but every piece of code we'd want is BSD 3 clause 2015-06-24T19:58:21 So if there is any ambiguity about the top level license overshadowing the nested BSD licenses, it may be too risky to include 2015-06-24T20:02:09 That's basically it. http://opensource.org/licenses/BSD-2-Clause 2015-06-24T20:02:37 The owner is replaced. The 3-clause one is the one with attribution. 2015-06-24T20:04:06 We are asking questions of another semiconductor vendor where you download a kit. You have to accept an agreement about exporting, then click on an EULA for the installer, and the code is BSD licensed. The EULA has some of the worst stuff I have ever seen in it. Our outstanding question is .. how much of what is in effect? 2015-06-24T20:04:10 Here's the specific license in question: http://pastebin.com/YDGiwvgg 2015-06-24T20:04:35 *** gedare has joined #rtems 2015-06-24T20:04:35 *** ChanServ sets mode: +o gedare 2015-06-24T20:04:55 That's unfortunate. This one is at least freely available with no agreement click through 2015-06-24T20:05:08 Just the top level Release Notes 2015-06-24T20:06:32 Hmm... that reads just like the 2-clause BSD license. I have no problems with paragraph 3. You would have to be insane to think that using that code gave you the ability to claim endorsement. 2015-06-24T20:07:15 Post that and the OSI link again to the devel list. I am failing to see an issue if the code is directly downloaded with no other clickign. 2015-06-24T20:12:36 Hello DrJoel 2015-06-24T20:13:00 Will do in a few days 2015-06-24T20:14:02 Hello sujayraaj 2015-06-24T20:16:48 hi krohini 2015-06-24T20:16:56 DrJoel, about the sendfile issue, I have been reading about implementing a syscall, (which I have fairly understood) , and mbuf ( very less documentation available ). Do you know from where could I read more about mbuf and using it in an implementation? 2015-06-24T20:21:14 FreeBSD internals guide :) 2015-06-24T20:21:53 Ask on devel@. Perhaps Sebastian can point to a place in the code that is a good example of allocating mbufs and using them. 2015-06-24T20:22:33 I am prone to tell you to write a simple implementation for now and get the port working. And continue to ask questions and write an optimized version when you are able. 2015-06-24T20:25:59 DrJoel, I wrote a simple read/write version to emulate sendfile (and that version works on linux) . So monkey has built successfully for our platform. Testing it on qemu is all that is left to get the port working. 2015-06-24T20:26:15 *** Davidbrcz has joined #rtems 2015-06-24T20:43:25 *** gedare has quit IRC 2015-06-24T20:46:49 I would keep pushing working on RTEMS. 2015-06-24T20:47:23 Working before optimized. Keep asking questions and plan to implement an optimized sendfile() but optimizing that before we have a web server working seems wrong. 2015-06-24T20:47:38 This should be discussed with the other mentors though to get agreement 2015-06-24T20:55:20 Yeah. Getting my work pushed into rtems would be nice given that I am working on behalf of rtems, and most of work I have done will be pushed into monkey :) 2015-06-24T20:55:24 Nonetheless, I shall be putting this sendfile issue on hiatus for the moment, and working on getting monkey running on qemu first. If everything goes well, by the time this weekend ends we would have monkey functioning in a test environment. Then I can get back to sendfile. 2015-06-24T20:58:20 It is really beneficial to RTEMS. The Mongoose server we had got relicensed to pure GPL and we are stuck on the last version with an acceptable license. So this is a pretty big deal to us. 2015-06-24T20:58:53 The Monkey developers are awesome. We have talked at the GSOC Mentor Summit and thought this would be a good project when they were ready for it. 2015-06-24T20:59:32 I don't think they got accepted as a project this year so it was a good time to do this. 2015-06-24T20:59:46 Plus it helps them shake out porting issues. 2015-06-24T21:00:15 I actually really like GSOC projects which are porting desirable third party packages to rtems with the involvement of developers from the other project. 2015-06-24T21:07:13 Yeah. And development in monkey is happening at a really fast rate. We got it working on freebsd alright so its one more (or rather a set of) operating systems added to their tally. :) 2015-06-24T21:10:34 Yep. And RTEMS pushes on the architectures once we have one building. I am not sure they ever build on anything but x86 2015-06-24T21:13:50 I think they do have one for raspberry pi , running on Raspbian (based on debian). 2015-06-24T21:14:45 Anyways, I think I should get some sleep now. I will inform you all when I get monkey running on qemu. Thanks and bye :) 2015-06-24T21:27:50 *** sujayraaj has quit IRC 2015-06-24T21:40:11 *** javamonn has quit IRC 2015-06-24T21:41:41 *** javamonn has joined #rtems 2015-06-24T22:08:24 *** yangqiao has joined #rtems 2015-06-24T22:09:16 *** javamonn has quit IRC 2015-06-24T22:09:35 *** javamonn has joined #rtems 2015-06-24T22:13:18 *** gigetoo has quit IRC 2015-06-24T22:18:21 *** kiwichris has quit IRC 2015-06-24T22:21:51 *** kiwichris has joined #rtems 2015-06-24T22:41:35 kiwichris can't decide to come or go? 2015-06-24T23:14:24 *** gigetoo has joined #rtems 2015-06-24T23:19:59 *** krohini has quit IRC 2015-06-24T23:20:17 *** DrJoel has quit IRC 2015-06-24T23:21:23 *** javamonn has quit IRC 2015-06-24T23:29:33 *** javamonn has joined #rtems 2015-06-24T23:39:47 *** javamonn has quit IRC 2015-06-24T23:42:02 *** gigetoo has quit IRC 2015-06-24T23:44:27 *** kiwichris has quit IRC 2015-06-24T23:55:09 *** gigetoo has joined #rtems 2015-06-25T00:20:03 *** kiwichris has joined #rtems 2015-06-25T00:21:37 * kiwichris has people play with power at his house 2015-06-25T01:53:12 *** Davidbrcz has quit IRC 2015-06-25T04:52:23 *** monstr has joined #rtems 2015-06-25T05:09:15 *** sebhub has quit IRC 2015-06-25T05:13:55 *** sebhub has joined #rtems 2015-06-25T05:41:09 *** ketul has joined #rtems 2015-06-25T06:21:00 *** ketul has quit IRC 2015-06-25T07:26:48 *** ketul has joined #rtems 2015-06-25T08:17:45 *** kiwichris has quit IRC 2015-06-25T08:42:36 *** ketul has quit IRC 2015-06-25T09:29:43 [git] 8042107 by Sebastian Huber: [score: Move SMP CPU_USE_DEFERRED_FP_SWITCH check] http://devel.rtems.org/changeset/8042107/rtems 2015-06-25T11:21:36 *** javamonn has joined #rtems 2015-06-25T11:35:00 *** ketul has joined #rtems 2015-06-25T12:36:47 *** gedare has joined #rtems 2015-06-25T12:36:47 *** ChanServ sets mode: +o gedare 2015-06-25T14:29:33 hi gedare 2015-06-25T14:31:12 I got "objdump: can't disassemble for architecture UNKNOWN!" while running "objdump -d adc.exe ". What does that mean? and if possible kind of solution for the same. 2015-06-25T14:31:15 beng-nl:hi 2015-06-25T14:31:46 ketul: hi 2015-06-25T14:32:12 ketul: use the objdump specific to your target arch from the rtems toolchain 2015-06-25T14:32:54 e.g. arm-rtems4.11-objdump 2015-06-25T14:33:15 umm... let me :) 2015-06-25T14:33:18 ok ): 2015-06-25T14:33:19 ok :) 2015-06-25T14:34:29 ohh ya got. 2015-06-25T14:34:31 ok! 2015-06-25T14:34:41 hi 2015-06-25T14:35:23 written some small code for adc and trying to test but as usual it has some errors :) 2015-06-25T14:35:40 you might also like to use -dS it can help 2015-06-25T14:36:15 yup DrJoel suggested. Thanks gedare :) 2015-06-25T14:36:22 regarding your coding style, you might try to use the uncrustify tool with the configuration script available on our coding conventions page. 2015-06-25T14:36:28 ah, sure 2015-06-25T14:38:03 ya good idea gedare 2015-06-25T14:46:53 *** hermann_19829_ has joined #rtems 2015-06-25T14:49:33 beng-nl: here it is "adcvalue = rtems_adc_read(BBB_P9_39); 2015-06-25T14:49:33 800004b4: e3a00000 mov r0, #0 2015-06-25T14:49:33 800004b8: eb000048 bl 800005e0 2015-06-25T14:49:33 800004bc: e1a01000 mov r1, r0 2015-06-25T14:49:33 " and m getting LR= 0x800004BC 2015-06-25T14:49:58 *** javamonn has quit IRC 2015-06-25T14:50:31 *** javamonn has joined #rtems 2015-06-25T14:50:48 and PC? 2015-06-25T14:51:13 PC is (kinda) where the cpu was executing at trap time 2015-06-25T14:51:46 0x800005EC 2015-06-25T14:51:57 ok, and that code? 2015-06-25T14:52:12 yup about to paste :) 2015-06-25T14:52:13 void boot_card( 2015-06-25T14:52:13 const char *cmdline 2015-06-25T14:52:13 ) 2015-06-25T14:52:13 { 2015-06-25T14:52:13 800005ec: e92d4008 push {r3, lr} 2015-06-25T14:52:14 uint32_t level; 2015-06-25T14:52:14 #if defined(ARM_MULTILIB_ARCH_V4) 2015-06-25T14:52:15 uint32_t arm_switch_reg; 2015-06-25T14:53:03 hmm, that doesn't fit 2015-06-25T14:53:22 i think the cpu is executing in rtems_adc_read and LR= 0x800004BC indicates the return address 2015-06-25T14:53:41 how about adding some debug printk's? 2015-06-25T14:54:13 not added yet. I think I should add it now 2015-06-25T14:54:59 i think rtems_adc_read is reading/writing from/to the wrong place 2015-06-25T14:55:27 hmm then I should refer TRM once more .right ? 2015-06-25T14:56:00 so beng-nl problem is at mov r1, r0 ? 2015-06-25T14:57:08 i think that's just a straight register move so no, that doesn't cause a memory access 2015-06-25T14:57:52 push 2015-06-25T14:57:56 that's a memory access. 2015-06-25T14:58:03 is the SP corrupt 2015-06-25T14:59:11 ketul: what is the stack pointer going into boot_card? 2015-06-25T14:59:33 gedare: it is SP = 0x80106CF8 2015-06-25T14:59:47 is that a "good address" for it to be? 2015-06-25T15:00:22 how can I come to know whether is it or ? I mean didn't come across that 2015-06-25T15:00:28 what does the push instruction do 2015-06-25T15:01:04 (Also, what problem are you debugging, I missed it?) 2015-06-25T15:01:35 aha my adc.exe file 2015-06-25T15:02:11 push is I think for writing value to the memory 2015-06-25T15:02:43 I mean in stack 2015-06-25T15:02:56 yeah but it is taking 2 arguments so that is confusing me 2015-06-25T15:03:02 I don't really know ARM asm 2015-06-25T15:03:32 arm has a fancy way to push & pop sets of registers 2015-06-25T15:03:42 so it's pushing both r3 and lr 2015-06-25T15:04:54 got on web :- http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.dui0489e/Babefbce.html 2015-06-25T15:09:09 *** rokka_ has joined #rtems 2015-06-25T15:10:42 *** rokka has quit IRC 2015-06-25T15:11:09 ok 2015-06-25T15:11:32 well if the fault really is at that address, then the problem is likely to do with that stack address. 2015-06-25T15:12:18 ketul: you should see what section of the memory map that address is located at 2015-06-25T15:12:59 ok you mean the PC 2015-06-25T15:13:04 No, the SP 2015-06-25T15:13:08 *** krohini has joined #rtems 2015-06-25T15:13:19 hi gedare 2015-06-25T15:13:36 gedare: 2015-06-25T15:13:48 "section of the memory map" what do you mean by that? 2015-06-25T15:13:48 ketul: If you're getting a fault at a PC location where the instruction is a push, that indicates the fault is due to accessing the stack, so you should figure out what address is being accessed and why it causes a fault. 2015-06-25T15:13:49 hi krohini 2015-06-25T15:14:26 ketul: take a look in libbsp/arm/shared/startup, and libbsp/arm/beagle/startup 2015-06-25T15:14:27 i finally have a working local repo. I tried using the ARMV7_CP15_START_DEFAULT_SECTIONS. It doesn't work. 2015-06-25T15:14:47 ok got that SP is the hint for my fault 2015-06-25T15:15:20 *** hermann_19829_ has quit IRC 2015-06-25T15:15:30 krohini: you might be able to use it, plus some extra custom ones, i think that is normal 2015-06-25T15:17:42 *** ketul has quit IRC 2015-06-25T15:17:53 there must be a DEVICE memory section which points to the region for peripherals. I added this section to the DEFAULT_SECTION while trying 2015-06-25T15:18:39 hm 2015-06-25T15:19:09 and , even replacing ARMV7_MMU_READ_WRITE with ARMV7_MMU_CODE_CACHED didn't work. 2015-06-25T15:19:36 I had said earlier that it does. But I guess I may have not run the correct image while testing it 2015-06-25T15:19:39 see for example libbsp/arm/beagle/startup/bspstartmmu.c 2015-06-25T15:19:56 regarding using default sections with customization 2015-06-25T15:20:31 okay..i'll check..I saw xilinx.. 2015-06-25T15:21:44 umm..that is what I did to customize 2015-06-25T15:22:56 it hanged. 2015-06-25T15:29:28 i guess the important question is, why does the .text section need read/write? 2015-06-25T15:29:36 what is getting put in there that needs to be written? 2015-06-25T15:33:24 the other thing to fix would be raspberrypi/startup/linkcmds to define the REGION_NOCACHE as the RPI Peripheral region 2015-06-25T15:35:02 ya, that is one thing i am trying to find out. why it needs read/write there 2015-06-25T15:42:54 *** krohini has quit IRC 2015-06-25T15:55:39 *** ragunath has joined #rtems 2015-06-25T15:55:46 hi gedare 2015-06-25T15:56:03 hi ragunath 2015-06-25T15:56:19 hi ben 2015-06-25T15:56:51 beng-nl: is it ok to send the patch using git send email> 2015-06-25T15:56:52 ? 2015-06-25T15:59:32 sure 2015-06-25T16:03:58 ok thanks 2015-06-25T16:16:34 *** ragunath has quit IRC 2015-06-25T16:19:56 *** ketul has joined #rtems 2015-06-25T17:26:31 *** ragunath has joined #rtems 2015-06-25T17:26:39 hi gedare 2015-06-25T17:34:58 *** ragunath has quit IRC 2015-06-25T17:46:18 *** Davidbrcz has joined #rtems 2015-06-25T19:07:37 *** monstr has quit IRC 2015-06-25T19:14:41 *** yangqiao has joined #rtems 2015-06-25T19:30:03 *** gedare has quit IRC 2015-06-25T19:39:29 *** javamonn has quit IRC 2015-06-25T19:49:03 *** krohini has joined #rtems 2015-06-25T20:03:11 *** Davidbrcz has quit IRC 2015-06-25T20:13:57 *** krohini has quit IRC 2015-06-25T20:24:43 *** Davidbrcz has joined #rtems 2015-06-25T20:36:30 *** Davidbrcz has quit IRC 2015-06-25T20:36:38 *** Davidbrcz has joined #rtems 2015-06-25T20:36:42 *** gedare has joined #rtems 2015-06-25T20:36:42 *** ChanServ sets mode: +o gedare 2015-06-25T20:59:20 *** hermann_19829 has joined #rtems 2015-06-25T21:01:57 *** gedare has quit IRC 2015-06-25T21:28:17 *** hermann_19829 has quit IRC 2015-06-25T21:30:59 *** hermann_19829 has joined #rtems 2015-06-25T21:57:15 *** stryx` has joined #rtems 2015-06-25T22:00:17 *** ketul has quit IRC 2015-06-25T22:09:04 *** javamonn has joined #rtems 2015-06-25T22:18:54 *** kiwichris has joined #rtems 2015-06-25T22:32:46 *** jkridner has quit IRC 2015-06-25T22:44:34 *** javamonn has quit IRC 2015-06-25T22:45:21 *** javamonn has joined #rtems 2015-06-25T22:57:48 *** yangqiao has quit IRC 2015-06-25T23:20:34 *** hermann_19829 has quit IRC 2015-06-25T23:43:24 *** jkridner has joined #rtems 2015-06-25T23:48:25 *** jkridner has quit IRC 2015-06-26T00:19:58 *** javamonn has quit IRC 2015-06-26T02:01:47 *** Davidbrcz has quit IRC 2015-06-26T02:22:21 *** mkhoory-2 has joined #rtems 2015-06-26T02:26:20 *** mkhoory-3 has quit IRC 2015-06-26T04:12:16 [git] 45a6041 by Gene Smith: [modified: waf_generator.py Fix incorrect encoding of NET_CFG_GATEWAY_IP and change _IP_IP ending to _IP. Remove _IP_IP ending from NET_TAP_INTERFACE. modified: wscript Re-generated from "./freebsd-to-rtems -m" based on waf_generator.py] http://devel.rtems.org/changeset/45a6041/rtems-libbsd 2015-06-26T04:17:26 *** ketul has joined #rtems 2015-06-26T04:31:18 *** ketul has quit IRC 2015-06-26T04:38:11 *** dr__house is now known as dr__house|away 2015-06-26T05:17:26 [git] 9ee5268 by Chris Johns: [tcpdump: Allow enter/return to exit tcpdump. Implement the PCAP loop in a separate thread and block the shell thread in the stdin getchar. When a user presses enter/return call the PCAP break loop function to have it return.] http://devel.rtems.org/changeset/9ee5268/rtems-libbsd 2015-06-26T05:21:00 [git] c6e5395 by Chris Johns: [tcpdump: Return the PCAP loop return value. Increase the stack size.] http://devel.rtems.org/changeset/c6e5395/rtems-libbsd 2015-06-26T05:40:34 *** monstr has joined #rtems 2015-06-26T07:10:06 *** kiwichris has quit IRC 2015-06-26T09:23:30 [git] 0b268b8 by Sebastian Huber: [score: Move default _ISR_Is_in_progress()] http://devel.rtems.org/changeset/0b268b8/rtems 2015-06-26T09:23:31 [git] bd67d7d by Sebastian Huber: [score: Simplify Avoid Thread_Control typedef in . This helps to get rid of the include in which exposes a lot of implementation details.] http://devel.rtems.org/changeset/bd67d7d/rtems 2015-06-26T09:23:31 [git] 270ccc5 by Sebastian Huber: [score: Fix extern "C" in ] http://devel.rtems.org/changeset/270ccc5/rtems 2015-06-26T09:23:31 [git] 48fed9a by Sebastian Huber: [score: Simplify Drop the include since this file exposes a lot of implementation details.] http://devel.rtems.org/changeset/48fed9a/rtems 2015-06-26T09:23:31 [git] 96316d1 by Sebastian Huber: [score: Simplify Drop the include since this file exposes a lot of implementation details.] http://devel.rtems.org/changeset/96316d1/rtems 2015-06-26T09:23:31 [git] 5f31bbe by Sebastian Huber: [libmisc: Simplify Drop the include since this file exposes a lot of implementation details.] http://devel.rtems.org/changeset/5f31bbe/rtems 2015-06-26T09:23:31 [git] d811dac by Sebastian Huber: [score: Hide SMP lock profiling impl if disabled The problem is that empty structures have a different size in C and C++.] http://devel.rtems.org/changeset/d811dac/rtems 2015-06-26T09:23:32 [git] 27f08f5 by Sebastian Huber: [bsps/arm: Update due to API changes] http://devel.rtems.org/changeset/27f08f5/rtems 2015-06-26T09:23:32 [git] 32005a7 by Sebastian Huber: [bsp/mpc55xxevb: Fix configure script This was the only configure.ac file with bspopts.h present in AC_CONFIG_FILES(). This somehow prevented the generation of this file leading to build errors for this BSP.] http://devel.rtems.org/changeset/32005a7/rtems 2015-06-26T09:34:27 *** sujayraaj has joined #rtems 2015-06-26T09:44:49 *** sujayraaj has quit IRC 2015-06-26T10:00:22 *** ketul has joined #rtems 2015-06-26T10:39:49 *** ketul has quit IRC 2015-06-26T10:41:13 *** ketul has joined #rtems 2015-06-26T11:26:27 *** ketul has quit IRC 2015-06-26T12:25:00 *** alia has joined #rtems 2015-06-26T12:26:37 *** alia_ has joined #rtems 2015-06-26T12:26:46 *** alia_ is now known as llll 2015-06-26T12:27:12 *** llll has quit IRC 2015-06-26T12:29:06 *** ketul has joined #rtems 2015-06-26T12:29:59 *** alia has quit IRC 2015-06-26T12:59:26 beng-nl: hi :) 2015-06-26T13:13:34 [trac] Projects/GraphicsToolkit edited by ppisa (RTEMS 4.11 and VESA BIOS Extension) http://devel.rtems.org/wiki/Projects/GraphicsToolkit 2015-06-26T13:14:47 [trac] Projects/GraphicsToolkit edited by ppisa (empty comment) http://devel.rtems.org/wiki/Projects/GraphicsToolkit 2015-06-26T13:15:45 *** gedare has joined #rtems 2015-06-26T13:15:45 *** ChanServ sets mode: +o gedare 2015-06-26T13:28:33 [trac] Projects/GraphicsToolkit edited by ppisa (Actual MicroWindows repository) http://devel.rtems.org/wiki/Projects/GraphicsToolkit 2015-06-26T13:31:54 [trac] Packages/MicroWindows edited by ppisa (empty comment) http://devel.rtems.org/wiki/Packages/MicroWindows 2015-06-26T13:35:34 [trac] Packages/MicroWindows edited by ppisa (empty comment) http://devel.rtems.org/wiki/Packages/MicroWindows 2015-06-26T13:38:13 *** monstr has quit IRC 2015-06-26T13:41:04 [trac] Packages/MicroWindows edited by ppisa (empty comment) http://devel.rtems.org/wiki/Packages/MicroWindows 2015-06-26T13:42:15 [trac] Packages/MicroWindows edited by ppisa (empty comment) http://devel.rtems.org/wiki/Packages/MicroWindows 2015-06-26T13:45:01 *** hermann_19829_ has joined #rtems 2015-06-26T13:45:15 [trac] TBR/BSP/Csb336 edited by ppisa (empty comment) http://devel.rtems.org/wiki/TBR/BSP/Csb336 2015-06-26T13:51:42 [trac] TBR/Website/RTEMSReferences edited by ppisa (empty comment) http://devel.rtems.org/wiki/TBR/Website/RTEMSReferences 2015-06-26T13:52:25 [trac] SOCIS/2015 edited by hermann19829 (empty comment) http://devel.rtems.org/wiki/SOCIS/2015 2015-06-26T13:58:09 *** sebhub has quit IRC 2015-06-26T14:13:26 *** DrJoel has joined #rtems 2015-06-26T14:13:27 *** DrJoel has joined #rtems 2015-06-26T14:13:27 *** ChanServ sets mode: +o DrJoel 2015-06-26T14:14:52 hello hermann_19829_ .. what can we do quickly to run this down? 2015-06-26T14:15:36 Can you email me the two objdumps which have _Thread_Handler different. I need to visually inspect them 2015-06-26T14:15:58 hi DrJoel 2015-06-26T14:16:40 hey.. in case I didn't pass along. I am really off work today and all of next week. I may be online at times next week from hospital 2015-06-26T14:17:07 ok 2015-06-26T14:17:32 Shoot me the objdumps so I can see if what I think is wrong, is actually it 2015-06-26T14:17:40 i will be in spain next week. so not really comparably to hospital :) 2015-06-26T14:18:14 10 seconds ... 2015-06-26T14:19:32 ok they are on the road 2015-06-26T14:23:45 this is in-hospital evaluation for my son with epilepsy. Take him off meds, monitor him, and record seizures. hopefully see if there is a non-drug treatment 2015-06-26T14:24:28 ah ok. wish you all the best! 2015-06-26T14:24:47 so we are using sparc and qemu? 2015-06-26T14:25:19 And looking at _Thread_Handler for the size mismatch? 2015-06-26T14:25:39 yes we do 2015-06-26T14:26:14 and the difference is 56 bytes, which is 0x38 that occurs several times in the dump 2015-06-26T14:26:17 one point is that sparc has never AFAIK had coverage done using qemu 2015-06-26T14:27:21 but i read in the docs of the couverture project that couverture-qemu supports leon2/3 2015-06-26T14:28:55 It does but covoar has never processed sparc from qemu only tsim.. so we a couple of places that have not been exercised this way. *SPARC* is suspect 2015-06-26T14:29:34 ah i see 2015-06-26T14:31:12 ketul: hi :) 2015-06-26T14:32:36 hermann_19829_: I am pretty sure I see the problem. Now to walk you through it. 2015-06-26T14:32:52 ok 2015-06-26T14:33:23 base_sp has _Thread_Handler staart at 4000FD50 and the next method starts at 4000FDD0 2015-06-26T14:33:28 right? 2015-06-26T14:34:02 and fileio has it start at 4005B9AC and the next method is at 4005BA64 2015-06-26T14:34:20 Do you see those? I have two editors side-by-side to compare them 2015-06-26T14:35:49 yes I see them 2015-06-26T14:36:27 The problem is that fileio reports the size 14*4 bytes larger 2015-06-26T14:36:48 you mean the line 4005b8ac: 9d e3 ... in fileio right? 2015-06-26T14:38:36 4005b9ac: 9d e3 bf a0 save %sp, -96, %sp 2015-06-26T14:38:41 is the first line of the method. 2015-06-26T14:38:45 The problem is at the bottom 2015-06-26T14:39:05 4005ba24: 10 bf ff f6 b 4005b9fc <_Thread_Handler+0x50> 2015-06-26T14:39:05 4005ba28: d0 27 60 40 st %o0, [ %i5 + 0x40 ] 2015-06-26T14:39:05 4005ba2c: 40 05 ba c4 call 401ca53c <__end+0x1392ac> 2015-06-26T14:39:05 4005ba30: 40 05 bb 24 call 401ca6c0 <__end+0x139430> 2015-06-26T14:39:05 4005ba34: 40 05 bb 24 call 401ca6c4 <__end+0x139434> 2015-06-26T14:39:06 4005ba38: 40 05 bb 24 call 401ca6c8 <__end+0x139438> 2015-06-26T14:39:08 4005ba3c: 40 05 bb 24 call 401ca6cc <__end+0x13943c> 2015-06-26T14:39:12 4005ba40: 40 05 bb 24 call 401ca6d0 <__end+0x139440> 2015-06-26T14:39:14 4005ba44: 40 05 bb 24 call 401ca6d4 <__end+0x139444> 2015-06-26T14:39:16 4005ba48: 40 05 ba d0 call 401ca588 <__end+0x1392f8> 2015-06-26T14:39:18 4005ba4c: 40 05 ba dc call 401ca5bc <__end+0x13932c> 2015-06-26T14:39:20 4005ba50: 40 05 ba e8 call 401ca5f0 <__end+0x139360> 2015-06-26T14:39:22 4005ba54: 40 05 ba f4 call 401ca624 <__end+0x139394> 2015-06-26T14:39:24 4005ba58: 40 05 bb 00 call 401ca658 <__end+0x1393c8> 2015-06-26T14:39:26 4005ba5c: 40 05 bb 0c call 401ca68c <__end+0x1393fc> 2015-06-26T14:39:28 4005ba60: 40 05 bb 18 call 401ca6c0 <__end+0x139430> 2015-06-26T14:39:30 Those calls are filler at the end of the method 2015-06-26T14:40:04 beng-nl: hi doubt solved :) 2015-06-26T14:40:08 are filling the gap between this and the next method to put it on a "desirable" boundary 2015-06-26T14:40:37 (alignment) 2015-06-26T14:41:06 Yep. We want the fill pattern to be nop. 2015-06-26T14:41:23 Those are never the same instruction so are hard to recognize. 2015-06-26T14:41:47 ok i see all these lines in fileio 2015-06-26T14:41:57 Do an "info ld" and search for "FILLEXP". We want the .text section to have a fill pattern of a nop 2015-06-26T14:42:03 4000fe0c: 01 00 00 00 nop 2015-06-26T14:42:42 First .. where this breaks covoar. :) 2015-06-26T14:43:28 When reading the objdump, covoar has to through away the alignment fill at the end of the methods. You can't execute it and it causes the sizes to be different for the same code from exe to exe as you obviously see 2015-06-26T14:43:55 Target* is the code which recognizes branches, nops, etc. 2015-06-26T14:44:22 And the objdump reader has some code to start from the end of the code and knock off the nop's as don't care. 2015-06-26T14:44:32 Now back to the executable... 2015-06-26T14:45:07 We don't have a defined fill pattern for .text on the sparc linker script and this looks like it might be random contents. Not sure what it is. 2015-06-26T14:45:12 But it needs to be a nop. 2015-06-26T14:45:24 So edit c/src/lib/libbsp/sparc/shared/startup/linkcmds.base 2015-06-26T14:45:52 search for the .text section and go to the bottom (closing) brace which will say 2015-06-26T14:45:55 } >ram 2015-06-26T14:46:24 change that to } =FILLEXP(0x01000000) >ram 2015-06-26T14:46:35 which should change those to nop instructions and fix this once you relink. 2015-06-26T14:47:06 But now you know what to look for by hand when the size doesn't match. Something at the end for alignment is not being recognized as a nop by the object dump reader. 2015-06-26T14:47:15 I know I blasted that through but does it make sense? 2015-06-26T14:48:17 how could we have debugged this easierly 2015-06-26T14:48:33 alternately, where should doc be updated for this issue 2015-06-26T14:48:35 ok i think i got it 2015-06-26T14:48:46 I was just thinking of how to solve this easier in the future. :) 2015-06-26T14:48:59 guess i've been trained well. 2015-06-26T14:49:26 Mismatches are detected by covoar. I guess we can only document what the messages mean, show an example, and show the likely fixes. 2015-06-26T14:49:53 (1) is an unrecognized nop pattern. x86 has a lot of them of varying sizes so was hard to get past this initially 2015-06-26T14:50:49 (2) fill pattern needs fixing. This can be fixed in the linkcmds file but also needs to be discussed with the upstream binutils folks. I did fix this on m68k I think where there was no fill pattern defined. sparc may actually be the same 2015-06-26T14:55:17 hermann_19829_: are you trying it? I need to leave in a bit to meet someone in 35 minutes. 2015-06-26T14:55:49 yes 2015-06-26T14:56:01 is it enough to run make for relinking? 2015-06-26T15:00:19 make clean first, i doubt the linkcmds change will propagate through the dependencies 2015-06-26T15:00:49 ok thanks gedare 2015-06-26T15:00:53 The short cut on linker script changes is to remove the .exe's and make. But make clean won't take long 2015-06-26T15:01:16 gedare: Did you see the Supreme Court ruling just announced? 2015-06-26T15:01:24 no is it related to the Google case? 2015-06-26T15:01:45 Oracle vs Google one? 2015-06-26T15:02:22 no the same sex marriage. Did they issue a ruling in that also? 2015-06-26T15:03:09 A trip to north Georgia this weekend is going to be interesting between avoiding talk of the confederate battle flag and same sex marriage. Joel no talkie politics at all. :) 2015-06-26T15:03:21 i don't think they've announced if they will review it yet. 2015-06-26T15:03:27 what did they say about same sex marriage? 2015-06-26T15:03:51 is the rebel flag a hot topic now too? 2015-06-26T15:04:51 short news version is yes. Not sure about all the nuances of the ruling yet. The plaintiffs in the case were really interesting. The plaintiff for whom the suit was named just wanted to be listed as spouse on a death certificate. 2015-06-26T15:05:13 ah. 2015-06-26T15:05:49 Another was a couple with adopted kids. Where they married, they were both legally parents. They moved and suddenly were not both legally the kids' parents anymore. 2015-06-26T15:06:24 The plaintiffs had more complex issues than simple "can we get married". The devil is always in the details. 2015-06-26T15:07:17 hermann_19829_: Any report. I really need to start heading to the car 2015-06-26T15:07:34 not yet 2015-06-26T15:07:47 i will send you the new dumps via mail 2015-06-26T15:08:17 run covoar and I will check in for status when I get there but that will be on phone so won't be looking at files. 2015-06-26T15:08:24 hopefully yhis will do better. 2015-06-26T15:08:35 not logging off here .cya 2015-06-26T15:09:18 ok see you 2015-06-26T15:32:27 *** joelDroid has joined #rtems 2015-06-26T15:32:42 Any progress? 2015-06-26T15:42:44 a linker error occurs - 2015-06-26T15:43:05 "nonconstant expression for fill value" 2015-06-26T15:43:20 hi gedare hi beng-nl. 2015-06-26T15:43:34 solved prob of yesterday :) 2015-06-26T15:49:50 howdy 2015-06-26T15:50:08 hermann_19829_: it probably doesn't like the 0x 2015-06-26T15:50:34 you'll have to poke around to see how to provide a hex value 2015-06-26T15:51:10 Look in the ld info pages for an example 2015-06-26T15:51:30 yeah that or the fillexp syntax is off 2015-06-26T15:51:47 yes i did that and changed the expression to: 2015-06-26T15:52:04 > ram =0x01000000 2015-06-26T15:52:39 now it works again 2015-06-26T15:53:01 check objdump 2015-06-26T15:53:03 *** sujayraaj has joined #rtems 2015-06-26T15:53:06 gedare: adc driver working :D. When I connected dgnd to particular pin got value 0-4 :). 2015-06-26T15:53:31 good. do you have any signal generator you can use to further test it? 2015-06-26T15:54:09 umm currently I don't have. I need to go to collage for that. But I m now passed out :) 2015-06-26T15:54:48 may be testing can be done using pot (var resistor) 2015-06-26T15:55:24 sure get some rest 2015-06-26T15:55:37 actually BBB has many features about ADC. currently used the basics only :) 2015-06-26T15:55:53 and focus on cleaning up your code for midterm and get ready to fill out your evaluation form 2015-06-26T15:56:09 and participate in the discussion about the GPIO API on the devel ml 2015-06-26T15:56:45 yep sure. now I should focus on that. but m happy to see adc demo 2015-06-26T15:57:08 yeah of course. :) good job 2015-06-26T15:57:14 :) 2015-06-26T15:58:01 gedare: m not from pure ce background so may be some what poor at coding styles and all. but surely will modify it 2015-06-26T15:59:01 yeah sure will learn too much in this summer program. 2015-06-26T15:59:10 joelDroid, gedare nothing changed 2015-06-26T15:59:13 It's really about understanding the coding style rules and adhering to them. Has nothing to do with background and everything to do with reading comprehension and being detail-oriented. 2015-06-26T15:59:31 yepp thats right. 2015-06-26T15:59:44 hermann_19829_: i guess that means your change did not set the fill expression properly 2015-06-26T15:59:47 i think the command i used does not work 2015-06-26T15:59:52 hermann_19829_: you're going to have to consult the ld documentations 2015-06-26T16:00:00 gedare: will do. Thanks. 2015-06-26T16:01:42 gedare in the ld documentation I found .. } [>REGION] [AT>LMA_REGION] [:PHDR :PHDR ...] [=FILLEXP] 2015-06-26T16:02:48 Grep our tree for examples in cass you get lucky 2015-06-26T16:03:47 joelDroid for i386 I found and example .. } =0x90909090 2015-06-26T16:04:25 but for sparc it doesn't work without the "> ram" 2015-06-26T16:06:50 Yep. Has to be placed in ram 2015-06-26T16:14:31 hermann_19829_: try putting the FILL command at the start of the .text section 2015-06-26T16:14:32 https://sourceware.org/binutils/docs/ld/Output-Section-Data.html#Output-Section-Data 2015-06-26T16:14:46 hmm 2015-06-26T16:15:15 that might only work on the 'next' command, but putting it before *(.text*) 2015-06-26T16:15:19 may at least work for you 2015-06-26T16:15:57 but it's not clear why your other approach ddidn't work, with >ram =0x01000000 2015-06-26T16:16:09 it seems like that should work. 2015-06-26T16:25:19 *** joelDroid has quit IRC 2015-06-26T16:27:21 gedare you mean like .text : =FILL(0x0100000000) { .. ? 2015-06-26T16:29:57 hermann_19829_: no i think you can do something like: 2015-06-26T16:30:44 .text : 2015-06-26T16:30:44 { 2015-06-26T16:30:44 CREATE_OBJECT_SYMBOLS 2015-06-26T16:30:44 text_start = .; 2015-06-26T16:30:44 _text_start = .; 2015-06-26T16:30:45 FILL(0x01000000) 2015-06-26T16:30:47 *(.text*) 2015-06-26T16:30:59 also, check the number of 0s you are using.. 2015-06-26T16:31:36 because 0x0100000000 is not right 2015-06-26T16:32:42 This is the line I got that from: 2015-06-26T16:32:43 4000fe0c: 01 00 00 00 nop 2015-06-26T16:32:46 yes sorry. 2015-06-26T16:33:22 you also can just make it whatever pattern you want... 2015-06-26T16:33:37 even copying the x86's 2015-06-26T16:46:21 hmm.. 2015-06-26T16:46:24 .text : 2015-06-26T16:46:24 59 { 2015-06-26T16:46:24 60 CREATE_OBJECT_SYMBOLS 2015-06-26T16:46:24 61 text_start = .; 2015-06-26T16:46:24 62 _text_start = .; 2015-06-26T16:46:25 63 FILL(0x01000000) 2015-06-26T16:46:27 64 *(.text*) 2015-06-26T16:46:43 also works for linking 2015-06-26T16:46:59 but still no difference in the dump 2015-06-26T16:53:53 Hmm... that section of addresses is a jump table. Did you send fileio.exe.dmp to Gedare? 2015-06-26T16:54:15 When I search for the first address in the file, I see it referenced. I thought of this in the car. 2015-06-26T16:54:21 Now where is that in the other dmp 2015-06-26T16:54:41 *** ragunath has joined #rtems 2015-06-26T16:55:22 hi gedare 2015-06-26T16:55:58 hi ragunath 2015-06-26T16:56:01 no didn't send it to gedare yet 2015-06-26T16:56:39 so maybe it isn't really filler? 2015-06-26T16:56:43 *** sujayraaj has quit IRC 2015-06-26T16:56:54 hermann_19829_: Please send the two files. Gedare and you can double check me that that code is a jump table. I see the first address referenced. The answer may be to assume the shortest version we see if the correct size and we have unnamed code after it. 2015-06-26T16:57:13 *** ragunath_ has joined #rtems 2015-06-26T16:57:18 hi gedare 2015-06-26T16:57:40 import from freebsd needs to be in a separate commit right? 2015-06-26T16:57:40 gedare: not in this case I think. Please look at the dumps and see if you spot the address referenced like I did. covoar may have to just assume the shortest version is the correct size for now. 2015-06-26T16:57:55 This will take some discussion on gcc list 2015-06-26T16:57:58 ragunath_: yes 2015-06-26T16:58:22 Assuming smallest is correct and changing the size will be OK for now to get the merge of information to work. 2015-06-26T16:58:34 hermann_19829_: I will try to look into this but you should try to work around it 2015-06-26T16:58:45 *** Davidbrcz has joined #rtems 2015-06-26T16:59:31 *** ragunath has quit IRC 2015-06-26T16:59:51 ragunath_: and also see the directions at https://git.rtems.org/rtems-libbsd/tree/libbsd.txt#n435 2015-06-26T17:00:31 note the comments on the endiffs 2015-06-26T17:00:33 ok I sent you both the current dumps 2015-06-26T17:01:58 *** ragunath_ has quit IRC 2015-06-26T17:03:47 *** krohini has joined #rtems 2015-06-26T17:04:17 I need to leave now. hermann_19829_ please post patches and instructions. Plus if this is couverture qemu, etc. I will try to reproduce all of your setup next week if I can. 2015-06-26T17:04:23 *** DrJoel has quit IRC 2015-06-26T17:05:22 *** DrJoel has joined #rtems 2015-06-26T17:05:22 *** ChanServ sets mode: +o DrJoel 2015-06-26T17:06:16 I had the idea that the arguments to objdump might be able to be tinkered with to avoid this but don't think so now. 2015-06-26T17:07:27 I could try, but I have no idea how to tinker arguments 2015-06-26T17:08:45 got it hermann_19829_ 2015-06-26T17:09:00 I think this block might have a local symbol that isn't being dumped. I will try to poke at it next week 2015-06-26T17:09:03 it is a jump table 2015-06-26T17:09:05 It is part of this method I think 2015-06-26T17:09:06 4005ba64 : 2015-06-26T17:09:06 #include 2015-06-26T17:09:10 see address 4005ba2c 2015-06-26T17:09:27 it gets loaded at conv_c() 2015-06-26T17:09:37 4005bab4 in the objdump of fileio.exe 2015-06-26T17:09:41 and then offset into, and jmps 2015-06-26T17:09:42 see 2015-06-26T17:09:48 4005bab4: 84 10 a2 2c or %g2, 0x22c, %g2 ! 4005ba2c <_Thread_Handler+0x80> 2015-06-26T17:09:48 4005bab8: c2 00 80 01 ld [ %g2 + %g1 ], %g1 2015-06-26T17:09:48 4005babc: 81 c0 40 00 jmp %g1 2015-06-26T17:10:17 So you agree is is part of the next method before the method entry? 2015-06-26T17:10:18 so in fact you have 2 options here... 2015-06-26T17:10:30 yeah that's what it looks like 2015-06-26T17:10:42 either you ignore it, or you figure out how to measure it for coverage (better to do the measuring) 2015-06-26T17:10:49 *** sujayraaj has joined #rtems 2015-06-26T17:15:13 The problem is we have to associate it with something and it has no start marker. It it had a start marker, it would get processed correctly. 2015-06-26T17:15:39 assuming the smallest size reported is the correct one to use for the unified coverage report is OK for now'bye 2015-06-26T17:15:40 bye 2015-06-26T17:15:43 *** DrJoel has quit IRC 2015-06-26T17:20:06 *** mumptai has joined #rtems 2015-06-26T17:24:57 *** ragunath has joined #rtems 2015-06-26T17:25:16 does anybody know how to run libbsd.py in libbsd? 2015-06-26T17:33:03 *** sujayraaj has quit IRC 2015-06-26T17:41:35 ragunath: i think it gets called from freebsd-to-rtems.py 2015-06-26T17:45:41 yes but i am not sure how to run freebsd-to-rtems.py for few files 2015-06-26T17:45:56 or a single file 2015-06-26T17:51:55 *** krohini has quit IRC 2015-06-26T17:52:29 *** krohini has joined #rtems 2015-06-26T17:56:34 *** ketul has quit IRC 2015-06-26T18:19:06 ragunath: i don't know. probably you have to ask in an email for either kiwichris or sebhub to get back to you on 2015-06-26T18:32:19 ok i will check them 2015-06-26T18:33:03 yeah just send to devel one of them will probably get to it 2015-06-26T18:33:07 but maybe not until next week 2015-06-26T18:45:13 *** ragunath has quit IRC 2015-06-26T19:16:29 [trac] 0001-Create_coverage_report_incl_debug_output_for_byte_size_diff.patch (Patch containing diffs to current rtems-tools root which can be used to reproduce byte size errors.) attached to #SOCIS/2015/CoverageAnalysis by hermann19829 http://devel.rtems.org/ticket/SOCIS/2015/CoverageAnalysis 2015-06-26T19:26:54 *** ketul has joined #rtems 2015-06-26T19:28:11 [trac] couverture-qemu_patch.diff (Patch for Couverture QEMU.) attached to #SOCIS/2015/CoverageAnalysis by hermann19829 http://devel.rtems.org/ticket/SOCIS/2015/CoverageAnalysis 2015-06-26T19:28:47 [git] d84408a by Sebastian Huber: [doc: Update ARM documentation] http://devel.rtems.org/changeset/d84408a/rtems 2015-06-26T19:32:22 *** krohini has quit IRC 2015-06-26T19:39:54 [git] cb2b8f0 by Sebastian Huber: [doc: Fix interrupt level ARM documentation] http://devel.rtems.org/changeset/cb2b8f0/rtems 2015-06-26T19:46:12 *** hermann_19829_ has quit IRC 2015-06-26T20:02:58 *** Davidbrcz has quit IRC 2015-06-26T20:43:01 *** Davidbrcz has joined #rtems 2015-06-26T20:59:06 *** gedare has quit IRC 2015-06-26T21:00:06 *** ketul has quit IRC 2015-06-26T22:40:32 *** Davidbrcz has quit IRC 2015-06-27T00:33:48 *** sujayraaj has joined #rtems 2015-06-27T00:48:06 *** gedare has joined #rtems 2015-06-27T00:48:06 *** ChanServ sets mode: +o gedare 2015-06-27T00:55:11 *** gedare has quit IRC 2015-06-27T02:40:49 *** sujayraaj has quit IRC 2015-06-27T02:41:06 *** sujayraaj has joined #rtems 2015-06-27T04:19:10 *** ketul has joined #rtems 2015-06-27T05:17:35 *** ketul has quit IRC 2015-06-27T06:43:56 *** Davidbrcz has joined #rtems 2015-06-27T07:20:15 *** Davidbrcz has quit IRC 2015-06-27T09:13:55 *** sujayraaj has quit IRC 2015-06-27T13:22:25 *** gedare has joined #rtems 2015-06-27T13:22:25 *** ChanServ sets mode: +o gedare 2015-06-27T13:30:12 *** gedare has quit IRC 2015-06-27T15:06:13 *** sujayraaj has joined #rtems 2015-06-27T16:18:42 [trac] GSoC/2015 edited by sujayraaj (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-06-27T17:22:52 *** sujayraaj has quit IRC 2015-06-27T17:24:32 *** Davidbrcz has joined #rtems 2015-06-27T17:37:51 *** Davidbrcz has quit IRC 2015-06-27T18:09:42 *** ketul has joined #rtems 2015-06-27T18:46:20 *** sujayraaj has joined #rtems 2015-06-27T19:36:06 *** jkridner has joined #rtems 2015-06-27T19:41:07 *** sujayraaj has quit IRC 2015-06-27T19:44:28 beng-nl: hi :) 2015-06-27T20:14:34 beng-nl: forgot to add init.c in patch. But I updated it on https://github.com/ketul93/RTEMS-on-BBB/blob/master/ADC-TEST/patch.text 2015-06-27T21:18:34 *** sujayraaj has joined #rtems 2015-06-27T21:20:07 *** sujayraaj has quit IRC 2015-06-27T21:39:11 *** jkridner has quit IRC 2015-06-27T21:40:55 *** jkridner has joined #rtems 2015-06-27T21:42:16 *** sujayraaj has joined #rtems 2015-06-27T21:49:59 *** ketul has quit IRC 2015-06-27T22:26:38 *** kiwichris has joined #rtems 2015-06-27T22:48:45 *** kiwichris has quit IRC 2015-06-27T22:51:54 *** kiwichris has joined #rtems 2015-06-27T23:23:12 hi kiwichris 2015-06-27T23:46:39 *** sujayraaj has quit IRC 2015-06-28T00:00:20 *** sujayraaj has joined #rtems 2015-06-28T00:56:51 *** mumptai has quit IRC 2015-06-28T01:18:06 [git] b987a75 by Jarielle Catbagan: [Removed execution mode file attribute from all ASCII text files] http://devel.rtems.org/changeset/b987a75/umon 2015-06-28T01:18:07 [git] dee5246 by Jarielle Catbagan: [Duplicated csb740 directory to beagleboneblack directory for BBB port] http://devel.rtems.org/changeset/dee5246/umon 2015-06-28T01:18:07 [git] d734d80 by Jarielle Catbagan: [Updated bashrc in BBB port to target BBB] http://devel.rtems.org/changeset/d734d80/umon 2015-06-28T01:18:07 [git] 8b46c00 by Jarielle Catbagan: [Updated BBB port makefile * Changed PLATFORM and TGTDIR to match BBB target * Fixed inclusion of .make files] http://devel.rtems.org/changeset/8b46c00/umon 2015-06-28T01:18:07 [git] d871a74 by Jarielle Catbagan: [Fixed BASE variable in common.make The BASE variable was changed from $(TOPDIR)/target -> $(TOPDIR) as the "target" directory's contents were moved to the $(TOPDIR).] http://devel.rtems.org/changeset/d871a74/umon 2015-06-28T01:18:07 [git] 9049385 by Jarielle Catbagan: [Renamed linker script templates in BBB port to match BBB target * Renamed CSB740_boot.ldt -> BEAGLEBONEBLACK_boot.ldt * Renamed CSB740_ramtst.ldt -> BEAGLEBONEBLACK_ramtst.ldt] http://devel.rtems.org/changeset/9049385/umon 2015-06-28T01:18:07 [git] 38e5eef by Jarielle Catbagan: [Fixed 'boot' and 'ramtst' prerequisite makefile name in Makefile for BBB port 'boot' and 'ramtst' targets in Makefile need to have prerequisite makefile as "Makefile" instead of "makefile".] http://devel.rtems.org/changeset/38e5eef/umon 2015-06-28T01:18:08 [git] 09fa828 by Jarielle Catbagan: [Removed csb740 OMAP3530 setup in rom_reset.S in BBB port] http://devel.rtems.org/changeset/09fa828/umon 2015-06-28T01:18:08 [git] 3f3cc2b by Jarielle Catbagan: [Updated header comment in BEAGLEBONEBLACK_boot.ldt] http://devel.rtems.org/changeset/3f3cc2b/umon 2015-06-28T01:18:09 [git] 11c6677 by Jarielle Catbagan: [Replaced cpuio.c and cpuio.h in BBB port The former cpuio.* files contained code specific to CSB740. It was replaced with the template cpuio.* files with the getUartDivisor() function retained from the previous cpuio.c] http://devel.rtems.org/changeset/11c6677/umon 2015-06-28T01:18:09 [git] cf419d2 by Jarielle Catbagan: [Removed all omap3530_* peripheral source and header files from BBB port Additionally, the compilation of these files were excluded in the Makefile] http://devel.rtems.org/changeset/cf419d2/umon 2015-06-28T01:18:10 [git] 6504fc5 by Jarielle Catbagan: [Removed ad7843.* from BBB port The AD7843 Interface routines are not relevant to the BBB, therefore it is removed. ad7843.c has been removed from the compilation process in Makefile.] http://devel.rtems.org/changeset/6504fc5/umon 2015-06-28T01:18:10 [git] d9ca895 by Jarielle Catbagan: [Updated BBB config.h to target BBB * Updated header comment to reflect BBB target and the individual who who did the adaptation * CPU_* and PLATFORM_* have been set to match BBB target * ALLOCSIZE is set to 1KB * All INCLUDE_xx macros have been set to 0] http://devel.rtems.org/changeset/d9ca895/umon 2015-06-28T01:18:11 [git] c0d0e6a by Jarielle Catbagan: [Removed "ads" command from being included in the common cmdtbl.c from xcmddcl.h and xcmdtbl.h] http://devel.rtems.org/changeset/c0d0e6a/umon 2015-06-28T01:40:51 *** sujayraaj has quit IRC 2015-06-28T01:41:03 *** Davidbrcz has joined #rtems 2015-06-28T02:52:24 *** Davidbrcz has quit IRC 2015-06-28T05:02:13 *** Davidbrcz has joined #rtems 2015-06-28T05:37:23 *** Davidbrcz has quit IRC 2015-06-28T05:59:28 [trac] GSoC/2015 edited by ragunath (empty comment) http://devel.rtems.org/wiki/GSoC/2015 2015-06-28T06:33:03 *** mumptai has joined #rtems 2015-06-28T08:16:47 *** ketul has joined #rtems 2015-06-28T08:37:15 *** ketul has quit IRC 2015-06-28T08:40:29 *** ketul has joined #rtems 2015-06-28T12:09:56 [trac] GSoC/2015 edited by ketul (status updated) http://devel.rtems.org/wiki/GSoC/2015 2015-06-28T12:11:18 *** kiwichris has quit IRC 2015-06-28T12:37:19 *** yangqiao has joined #rtems 2015-06-28T12:39:46 *** yangqiao has quit IRC 2015-06-28T12:41:54 *** kiwichris has joined #rtems 2015-06-28T14:12:31 *** ketul has quit IRC 2015-06-28T16:52:38 *** sgworks has joined #rtems 2015-06-28T17:08:39 *** verm__ has quit IRC 2015-06-28T17:08:39 *** verm__ has joined #rtems 2015-06-28T17:22:44 *** sgworks has quit IRC 2015-06-28T18:11:45 *** sgworks has joined #rtems 2015-06-28T18:34:56 *** ketul has joined #rtems 2015-06-28T20:12:02 *** ita_ has joined #rtems 2015-06-28T20:13:43 *** ita has quit IRC 2015-06-28T20:52:58 *** sgworks has quit IRC 2015-06-28T21:39:15 *** sujayraaj has joined #rtems 2015-06-28T21:52:59 *** ketul has quit IRC 2015-06-28T22:08:45 *** kiwichris has quit IRC 2015-06-28T22:23:29 *** mumptai has quit IRC 2015-06-28T22:36:05 *** kiwichris has joined #rtems 2015-06-28T22:39:52 kiwichris : hi 2015-06-28T22:40:01 sujayraaj, hi 2015-06-28T22:42:37 I sent you some emails over the weeked, I put up the toolchain file, and made some changes in cmakelists so that it compiles both for linux and rtems. I also added network configuration in monkey that initializes the devices and sets up networking. Plus, since we now have working network, I tested the network demos using libbsd. 2015-06-28T22:44:05 kiwichris : The problem is with setting up the IMFS. how do we actually load the tar file in qemu? do we specify it as a fat drive or something? 2015-06-28T22:45:08 sujayraaj I saw the email and I am sorry but I had no time to answer. In the patches in github you make changes to the default sin Monkey. Why did you do that ? 2015-06-28T22:48:48 kiwichris : Ah! I forgot to put them under a conditional header. For monkey to build with rtems, there are some options that are required, like the executable should be built locally ( not installed ) , it should use the malloc memory allocator, and that presently we don't use any plugins that monkey has. edsiper told me that I should get involved with plugins later because they are in migration phase. 2015-06-28T22:49:22 kiwichris : I will add conditionals for them too. They just escaped my mind because I was using the same configuration from the beginning. 2015-06-28T22:50:35 What is a conditional header when talking about cmake ? Your changes to Monkey must only make changes that are needed. Changes to things like defaults that can be changed by the user on a command line will be rejected. 2015-06-28T22:51:22 It is ok for a user to have to get the correct set of command line options to get Monkey to build. There is also a configure type script which could hold the RTEMS defaults. 2015-06-28T22:51:38 In your cmake patch you reference another file. What is this file ? 2015-06-28T22:54:40 Its a cmake toolchain file, it sets the CMAKE_SYSTEM_NAME to rtems and forces cmake to use our compilers from rtems tools. By default cmake rejects our compiler because its not able to build a sample test case. Moreover, using a separate toolchain file helps make things easier : If a user wants to compile for rtems , he uses the toolchain file. If he doesnt want to compile for linux by default, he skips the toolchain file and ev 2015-06-28T22:55:36 I added this file to my git repo : https://github.com/sujayraaj/monkey/blob/master/rtems.toolchain 2015-06-28T22:56:02 sujayraaj, I understand the purpose but I cannot use it because I have no idea what the file contains until now. 2015-06-28T22:57:44 sujayraaj, I suggest you remove the file. The support in this file needs to be broken down and provided as separate configuration (-D??) values to cmake. 2015-06-28T22:58:01 And added to the top level cmake file in Monkey. 2015-06-28T23:00:53 Yeah. That can be done quickly. 2015-06-28T23:12:45 Excellent 2015-06-28T23:13:42 sujayraaj, I would like you to get your changes into a suitable set of minimal changes to Monkey so a pull request can be made. 2015-06-28T23:14:04 sujayraaj, getting the changes upstream would be very nice. 2015-06-28T23:14:19 It also means if upstream changes we do not need to track those changes. 2015-06-28T23:20:18 kiwichris : okay. I am doing that. By today evening I will create a set of patches that we can expect to push upstream. By the way, the other patch that I sent, with changes to C files, was that okay? 2015-06-28T23:20:38 Fantastic. 2015-06-28T23:22:47 For the files in RTEMS I suggest making a tar file that is expanded into the IMFS. There are examples about in the testsuite in RTEMS and other examples. Look in the dl01 test in RTEMS. 2015-06-28T23:23:13 Doing this means the tests and examples do not depend on external files.