2016-08-08T05:31:29 *** monstr has joined #rtems 2016-08-08T06:07:01 [trac] #2773/defect (v:4.10) created by dcb (c/src/lib/libcpu/sh/sh7032/clock/ckinit.c:210: possible bad expression ?) http://devel.rtems.org/ticket/2773 2016-08-08T06:31:47 [trac] #2774/defect (v:4.12) created by chrisj (RTEMS loaded by u-boot with dcache, icache and mmu enable does not boot.) http://devel.rtems.org/ticket/2774 2016-08-08T07:28:02 *** indy has quit IRC 2016-08-08T07:31:01 *** indy has joined #rtems 2016-08-08T07:36:21 [git] 2cb9b86 by Sebastian Huber: [posix: Fix for RTEMS_DEBUG] http://devel.rtems.org/changeset/2cb9b86/rtems 2016-08-08T07:36:21 [git] aaaf961 by Sebastian Huber: [score: Add debug support to red-black trees This helps to detect double insert and extract errors.] http://devel.rtems.org/changeset/aaaf961/rtems 2016-08-08T07:36:21 [git] e5fd901 by Alexander Krutwig: [atsam: Add atsam_rtc_get_time()] http://devel.rtems.org/changeset/e5fd901/rtems 2016-08-08T07:47:49 [git] 5c09e9d by Sebastian Huber: [sptests/splinkersets01: Fix warnings] http://devel.rtems.org/changeset/5c09e9d/rtems 2016-08-08T09:53:29 *** sebhub has joined #rtems 2016-08-08T10:23:53 *** guerby has quit IRC 2016-08-08T10:40:17 *** guerby has joined #rtems 2016-08-08T11:49:49 *** Wonjun has joined #rtems 2016-08-08T11:56:59 [git] 5d0f0de by Alexander Krutwig: [bsp/atsam: Add power support] http://devel.rtems.org/changeset/5d0f0de/rtems 2016-08-08T14:32:20 *** gedare has joined #rtems 2016-08-08T14:32:21 *** ChanServ sets mode: +o gedare 2016-08-08T15:07:03 *** jkridner has joined #rtems 2016-08-08T15:40:56 *** jkridner has quit IRC 2016-08-08T15:41:35 *** jkridner has joined #rtems 2016-08-08T17:40:46 *** dr__house|away is now known as dr__house 2016-08-08T17:49:44 *** monstr has quit IRC 2016-08-08T18:49:59 *** beng-nl_ has joined #rtems 2016-08-08T18:50:06 *** rokka_ has joined #rtems 2016-08-08T18:50:19 *** guerby has quit IRC 2016-08-08T18:50:19 *** rokka has quit IRC 2016-08-08T18:50:20 *** beng-nl has quit IRC 2016-08-08T18:50:20 *** commander has quit IRC 2016-08-08T18:50:32 *** commander has joined #rtems 2016-08-08T18:52:47 *** guerby has joined #rtems 2016-08-08T19:02:53 *** jkridner|work has joined #rtems 2016-08-08T19:06:33 *** jkridner has quit IRC 2016-08-08T19:40:11 *** jkridner has joined #rtems 2016-08-08T19:40:34 *** jkridner|work has quit IRC 2016-08-08T21:13:11 *** gedare has quit IRC 2016-08-08T21:58:48 *** dr__house is now known as dr__house|away 2016-08-09T03:30:52 *** guerby has quit IRC 2016-08-09T03:31:39 *** guerby has joined #rtems 2016-08-09T03:32:58 [trac] #2775/defect (v:4.12) created by chrisj (ARM CP15 arm_cp15_set_translation_table_entries fails if TTB in read-only memory) http://devel.rtems.org/ticket/2775 2016-08-09T05:47:55 sebhub: hi 2016-08-09T05:57:08 hi chris 2016-08-09T05:59:31 *** monstr has joined #rtems 2016-08-09T06:01:08 Thanks for notes in the BSP. Why not make this part of the default in the table? 2016-08-09T06:01:30 This would fix all BSPs 2016-08-09T06:05:36 I am not sure, it depends on your use case, without the need for debug support and dynamic loading, the read-only ttb is a bit more robust 2016-08-09T06:06:20 you can make it the default from my point of view 2016-08-09T06:06:28 Yes however it is not easy to have users learn this low level stuff to be able to make it available. 2016-08-09T06:06:37 I am not sure what the anwser is 2016-08-09T06:07:16 which question? 2016-08-09T06:07:32 To making it default or not. Beling locked down is important. 2016-08-09T06:08:15 If it is easier to use and it is documented how to handle this then maybe it can be left out as a default. 2016-08-09T06:09:46 In a zynq you need a custom MMU table anyway or you have no access to the AXI buses to the programmable logic side of the device. 2016-08-09T06:11:01 I have changes that lets a recent u-boot be used, it detects the MMU is on and turns it off before setting it up again. WIth this change I can run. 2016-08-09T06:18:31 I add the section into the MMU table and I still have an issue. 2016-08-09T06:42:28 sebhub: the short entries are 1M so does this mean there is an overlap with some entries? 2016-08-09T07:05:13 *** punitvara has joined #rtems 2016-08-09T07:08:20 sebhub: should the MMU disable on boot be in the general code or specific to the BSP? 2016-08-09T07:09:07 *** punitvara has quit IRC 2016-08-09T07:09:50 *** nwe_ has joined #rtems 2016-08-09T07:11:37 *** nwe_ has quit IRC 2016-08-09T07:12:13 *** punitvara has joined #rtems 2016-08-09T07:13:14 * punitvara 2016-08-09T07:18:21 *** punitvara has quit IRC 2016-08-09T07:24:20 [trac] #2768/defect (closed) updated by Chris Johns (In [changeset:"b0f08c83e23e69c7b19b04d38910f90b5f7af51b/rtems"]: {{{ ...) http://devel.rtems.org/ticket/2768 2016-08-09T07:24:20 [git] b0f08c8 by Chris Johns: [libmisc/untar: Set the perms to the value in the tar file. This patch parses the mode field in the tar header and sets the directory or file to the mode value in the header. Closes #2768.] http://devel.rtems.org/changeset/b0f08c8/rtems 2016-08-09T07:42:09 [trac] #2776/enhancement (v:4.12) created by AlexKrEB (SPI Framework) http://devel.rtems.org/ticket/2776 2016-08-09T07:45:39 MMU on/off depends a bit on the boot loader 2016-08-09T07:46:09 in case the MMU is on and you switch it off without some cache operations you may kill yourself 2016-08-09T07:46:27 I added the cache operation. 2016-08-09T07:46:53 If the data cache is enable and the control reg is touched it the cache is cleared 2016-08-09T07:47:19 I left the MMU control in the Zynq BSP. 2016-08-09T07:47:40 Cause I know it works and I have tested it :) 2016-08-09T07:51:30 works it with -O0 optimization level? 2016-08-09T07:51:50 ? 2016-08-09T07:52:55 in this case you use the stack a lot and your code may no longer work 2016-08-09T07:54:01 if you have stores during the cache clean and mmu cache off, then you have a problem 2016-08-09T07:54:43 It could end up a problem, I am not sure, is building and running with -O0 a requirement? 2016-08-09T07:55:14 I use it quite often for better debug support 2016-08-09T07:55:51 and its quite bad if you want to debug problem X and first have to fix the low level startup to work with -O0 2016-08-09T07:56:20 If u-boot turns it on others must have managed the switch in their kernels. 2016-08-09T07:56:46 on linux a lot of this code is in asm 2016-08-09T07:57:28 u-boot is present and exists so we need to find a suitable solution. I will take a look at the output at -O0. How do you configure for it? 2016-08-09T07:57:46 hack the .cfg? 2016-08-09T07:57:57 yes, how else could I do this? 2016-08-09T07:58:08 Just checking :) 2016-08-09T07:58:29 well one of the nice features of the build system 2016-08-09T07:58:59 Lets not start, it will ruin the day. 2016-08-09T11:39:13 *** benjamint has quit IRC 2016-08-09T11:42:18 *** benjamint has joined #rtems 2016-08-09T11:43:44 *** benjamint has quit IRC 2016-08-09T11:50:02 *** benjamint has joined #rtems 2016-08-09T13:13:00 *** monstr has quit IRC 2016-08-09T13:13:22 *** monstr has joined #rtems 2016-08-09T13:54:45 *** gedare has joined #rtems 2016-08-09T13:54:45 *** ChanServ sets mode: +o gedare 2016-08-09T15:28:12 *** dougk_ has quit IRC 2016-08-09T15:42:29 *** dougk_ has joined #rtems 2016-08-09T16:29:59 *** benjamint has quit IRC 2016-08-09T16:39:01 *** benjamint has joined #rtems 2016-08-09T17:28:36 *** dr__house|away is now known as dr__house 2016-08-09T18:52:18 *** jkridner has quit IRC 2016-08-09T18:52:24 *** jkridner|work has joined #rtems 2016-08-09T19:20:51 *** monstr has quit IRC 2016-08-09T19:43:48 *** gedare has quit IRC 2016-08-09T20:51:32 *** gedare has joined #rtems 2016-08-09T20:51:32 *** ChanServ sets mode: +o gedare 2016-08-09T20:58:52 *** gedare has quit IRC 2016-08-09T21:01:18 *** dr__house is now known as dr__house|away 2016-08-09T21:49:02 *** gedare has joined #rtems 2016-08-09T21:49:02 *** ChanServ sets mode: +o gedare 2016-08-09T21:59:13 *** gedare has quit IRC 2016-08-09T22:08:40 *** gedare has joined #rtems 2016-08-09T22:08:54 *** ChanServ sets mode: +o gedare 2016-08-09T22:20:15 *** gedare has quit IRC 2016-08-09T22:32:32 *** jkridner|work has quit IRC 2016-08-09T23:25:59 *** gedare has joined #rtems 2016-08-09T23:25:59 *** ChanServ sets mode: +o gedare 2016-08-09T23:36:35 *** gedare has quit IRC 2016-08-10T02:17:40 [trac] #2777/defect (v:4.12) created by chrisj (Remove librtems++) http://devel.rtems.org/ticket/2777 2016-08-10T05:31:17 [git] 32a3dd1 by Christian Mauderer: [pf: Add RTEMS_BSD_CONFIG_FIREWALL_xxx options.] http://devel.rtems.org/changeset/32a3dd1/rtems-libbsd 2016-08-10T05:31:17 [git] a8d016c by Christian Mauderer: [testsuite/pf01: Remove pflog-module from test. This module is not tested inside the pf01 test. Therefore remove it. It is still build in the interactive pf02 test.] http://devel.rtems.org/changeset/a8d016c/rtems-libbsd 2016-08-10T05:31:17 [git] 61e2e1b by Christian Mauderer: [rc.conf: Use more stack for the rc worker. Most of the tests run with a default stack size of 32k. If the rc worker runs with less, it uses an untested configuration.] http://devel.rtems.org/changeset/61e2e1b/rtems-libbsd 2016-08-10T05:31:17 [git] 6cfb5c2 by Christian Mauderer: [pf: Improve documentation regarding stack.] http://devel.rtems.org/changeset/6cfb5c2/rtems-libbsd 2016-08-10T05:31:17 [git] bc9e939 by Christian Mauderer: [pf: Add configuration via rc.conf.] http://devel.rtems.org/changeset/bc9e939/rtems-libbsd 2016-08-10T05:33:49 *** monstr has joined #rtems 2016-08-10T06:05:02 [git] 62c4755 by Sebastian Huber: [doc: Add FreeBSD import version section] http://devel.rtems.org/changeset/62c4755/rtems-libbsd 2016-08-10T08:00:39 [git] 21d7365 by Sebastian Huber: [sptests/splinkersets01: Fix for small-data area] http://devel.rtems.org/changeset/21d7365/rtems 2016-08-10T12:14:16 *** salpha has joined #rtems 2016-08-10T13:27:34 *** deval has joined #rtems 2016-08-10T13:46:09 *** mudit1729 has joined #rtems 2016-08-10T13:58:27 *** salpha has joined #rtems 2016-08-10T13:58:34 *** gedare has joined #rtems 2016-08-10T13:58:34 *** ChanServ sets mode: +o gedare 2016-08-10T14:00:22 Hello 2016-08-10T14:00:27 Hello 2016-08-10T14:00:45 Hi 2016-08-10T14:01:49 Hello 2016-08-10T14:01:55 *** punitvara has joined #rtems 2016-08-10T14:01:56 deval: you can start first today 2016-08-10T14:02:22 yeah 2016-08-10T14:02:24 So, I am trying to run the testsuits regarding ethernet interface of RPi. And I am facing the following problem. 2016-08-10T14:02:28 The USB runs after the BSD initialization in its own threads. But the USB device loads the network interface in the case of RPi. And when the program tries to find the interface it doesn't find one. 2016-08-10T14:02:33 Sebastian Huber suggested to see the dhcpcd example as it is using routing socket. But when I ran the example, it also gave me the same error. 2016-08-10T14:02:48 Also, meanwhile I am also resolving problems with the patch I submitted. 2016-08-10T14:04:31 I don't quite follow the problem for the loading of the network 2016-08-10T14:04:39 do you have an email or blog post describing it in detail? 2016-08-10T14:05:23 yes 2016-08-10T14:05:24 https://lists.rtems.org/pipermail/devel/2016-August/015807.html 2016-08-10T14:05:35 hi 2016-08-10T14:05:39 "you have to wait a bit " 2016-08-10T14:05:53 Sebastian suggested you to study the dhcpd example as motivation to solve your problem 2016-08-10T14:06:05 not as something that would work right away 2016-08-10T14:06:40 one quick hack to see if it is timing related is to insert a sleep call for like 5 seconds 2016-08-10T14:06:44 Okay I will go further in the example then 2016-08-10T14:07:00 but using a hard-coded timeout is not a great design 2016-08-10T14:07:28 it would be better if the USB enumeration can be "event triggered" by the completion of the BSD init and network bringup 2016-08-10T14:07:32 Yes I tried that one, but adding the sleep function caused some more errors while compiling 2016-08-10T14:07:43 true 2016-08-10T14:07:51 that is the direction for you to work 2016-08-10T14:08:09 while you clean up your other patches, and start to prep your final products 2016-08-10T14:08:15 thanks deval, anything else? 2016-08-10T14:08:17 okay understood. 2016-08-10T14:08:29 thats it from me today 2016-08-10T14:08:31 ok 2016-08-10T14:08:34 salpha: you can go now 2016-08-10T14:08:37 Thanks 2016-08-10T14:08:42 with Pavel's help, i ran his example for dynamic linking (appdl) 2016-08-10T14:08:49 i wrote an application myself to understand better how OMK works 2016-08-10T14:09:07 DL is working for me now 2016-08-10T14:09:15 i'm manually changing test case executor to dynamically load test cases 2016-08-10T14:09:22 then i need to establish a tcp connection between qemu and rtems to pass the test case name 2016-08-10T14:09:29 BTW redirecting rtems output from qemu to bash was solved with -append "--console=/dev/com1" as Pavel suggested. 2016-08-10T14:09:43 and his suggestion for passing arguments to Init task didn't work 2016-08-10T14:09:51 finally i have to integrate this process into slingshot 2016-08-10T14:10:23 salpha: make sure the qemu note is in the wiki instructions for using qemu for i386 2016-08-10T14:10:56 okay i'll put that on the blog. 2016-08-10T14:11:04 if you can handle re-compiling your application each time you run it, then you can put the arguments into the code directly 2016-08-10T14:11:06 blog or wiki you mean? 2016-08-10T14:11:14 check the RTEMS wiki page about using qemu 2016-08-10T14:11:25 okay 2016-08-10T14:11:50 no re-compiling takes lot of time. that's why i'm switching from static linking to dynamic linking. 2016-08-10T14:11:54 https://devel.rtems.org/wiki/Developer/Simulators/QEMU 2016-08-10T14:11:58 ok 2016-08-10T14:12:03 thanks for the pointer 2016-08-10T14:12:27 passing test case names via a tcp connection should work 2016-08-10T14:12:56 sure. 2016-08-10T14:12:56 i personally found using OMK much easier than WAF 2016-08-10T14:13:09 but i was thinking to write a simple application via WAF too 2016-08-10T14:13:24 as a starting point for those who want to start using libdl 2016-08-10T14:13:28 should i do that? 2016-08-10T14:13:31 good that would be helpful 2016-08-10T14:13:45 *** DrJoel has joined #rtems 2016-08-10T14:13:45 *** DrJoel has joined #rtems 2016-08-10T14:13:45 *** ChanServ sets mode: +o DrJoel 2016-08-10T14:13:49 it would be nice to have some tutorials for both and comparisons of them in some sense 2016-08-10T14:13:58 sorry.. overslept. :( 2016-08-10T14:14:05 np, hi DrJoel . 2016-08-10T14:14:21 thanks salpha anything else? 2016-08-10T14:14:29 so i'll start by sending my libdl app (OMK version) as a patch to devel mailing list 2016-08-10T14:14:46 but maybe in two weeks 2016-08-10T14:14:48 no. 2016-08-10T14:14:50 thank you 2016-08-10T14:14:52 alright 2016-08-10T14:14:59 mudit1729: you can go next 2016-08-10T14:15:07 thanks 2016-08-10T14:15:13 yeah, so 2016-08-10T14:15:22 Regarding DMA Library : Pavel has reviewed the code for DMA library and it's ready. I will be pushing a patch regarding the same by EOD today. 2016-08-10T14:15:47 It has the changes that Pavel had suggested. 2016-08-10T14:16:02 Integration to SPI/I2C interface : I have identified the code where the DMA API can be used in the I2C driver 2016-08-10T14:16:28 great. 2016-08-10T14:16:31 Regarding using DMA with anything, I have written a mail on the mailing list 2016-08-10T14:16:49 Like how probably a developer would be using it 2016-08-10T14:17:18 I have started on few blog posts that I though would be useful. Will update the mailing list as soon as possible 2016-08-10T14:17:42 I have started working on the SD card driver now, as this deliverable is a good state 2016-08-10T14:17:57 sounds good, right on target i think 2016-08-10T14:18:15 Have gotten insight regarding DMA, I identified an obvious mistake that I was making with my revious code 2016-08-10T14:18:37 mudit1729: gedare can comment but if you have an example, even if BSP specific, the examples-v2 repo can hold the code. There are examples specific to UBoot and uMon already 2016-08-10T14:19:17 examples of using DMA in a certain context you mean ? 2016-08-10T14:19:57 Could you provide me a pointer as to where I should be adding these examples ? 2016-08-10T14:20:03 *** salpha has quit IRC 2016-08-10T14:20:51 ok DrJoel can tell mudit1729 more. we'll move on. punitvara are you ready to go? hi beng-nl_ 2016-08-10T14:20:59 gedare should express an opinion but I would lead to something like examples-v2/bsp-specific/XXX_dma 2016-08-10T14:21:09 okoay 2016-08-10T14:21:17 yeah 2016-08-10T14:21:26 mudit1729: make an email to devel to discuss the issue of putting bsp-specific examples in there. 2016-08-10T14:21:31 I have completed I2C bus driver coding. Curretly code is debugging phase. It is tough time debugging. It seems interrupt service routine needs to be improved. If I try to read EEPROM in BBB then it traps inside trasmit interrupt. and then freeze in between. 2016-08-10T14:21:39 okay 2016-08-10T14:21:43 mudit Worry about the code, doc, explanation, etc first. Finding a permanent home is easy compared to that. :) 2016-08-10T14:21:54 will do that, alongside the SD card driver. 2016-08-10T14:21:57 https://github.com/punitvara/rtems/blob/new/c/src/lib/libbsp/arm/beagle/i2c/bbb-i2c.c 2016-08-10T14:22:32 Code, Explanation is ready . Doc is pending. Will update on the group with all the realted info 2016-08-10T14:22:44 That would be all from my end. 2016-08-10T14:23:09 punitvara: are you able to isolate the trap/exception that happens? 2016-08-10T14:24:05 yes 2016-08-10T14:24:10 punitvara: do you test the am335x_i2c_interrupt() in isolation to make sure it works proper 2016-08-10T14:24:37 Means? 2016-08-10T14:24:50 can you cause an interrupt and verify it is handled? 2016-08-10T14:25:34 Yeah once I enable transmit interrupt I can see enabling service routine and follow intruction 2016-08-10T14:25:37 do you have a debugger you use? or do you rely only on these printks? 2016-08-10T14:25:46 on printks 2016-08-10T14:25:49 did you ever see printk("\n inside interrupt function \n"); 2016-08-10T14:25:55 yes 2016-08-10T14:26:20 is there the possibillity to use gdb to debug it? 2016-08-10T14:26:45 Ketul was telling me to use gdb today 2016-08-10T14:26:53 all of the printks can be messing up timing 2016-08-10T14:27:06 *** mudit1729 has quit IRC 2016-08-10T14:27:14 figure out how to use gdb for it. 2016-08-10T14:27:34 I will try to debug through gdb. One more thing I am not sure whether my test case is 100% correct or not 2016-08-10T14:27:50 https://github.com/punitvara/rtems/blob/new/testsuites/samples/i2c0/init.c 2016-08-10T14:29:03 punitvara: printing a single character is useful and doesn't mess up timing quite so much. It may be too much but seeing "acba" as output is just as useful as the full lines of output. 2016-08-10T14:29:24 Are you getting a second interrupt? Do any interrupts occur after the first transmitter one? 2016-08-10T14:29:25 this is another test that would eventually be good to put to examples-v2 in a bsp-specific location. 2016-08-10T14:29:50 After one interrupt no. program stuck in one intterupt only 2016-08-10T14:29:59 that's a problem to solve.:) 2016-08-10T14:30:17 *** mudit1729 has joined #rtems 2016-08-10T14:30:38 yes that what I am finding this time. 2016-08-10T14:30:46 Can you comment on test case ? 2016-08-10T14:30:50 i don't think your test does what you think it does. 2016-08-10T14:31:21 it appears to just repeatedly read the first 32 bytes of the EEPROM? 2016-08-10T14:31:36 yes that would I want 2016-08-10T14:31:39 oh, ok 2016-08-10T14:31:54 I am enabling transmit intrrupt to send first 8 byte for memory address 2016-08-10T14:32:09 well, you should discuss the test particulars with your mentors. i'm not familiar enough with the interfaces to tell if the test itself is good. 2016-08-10T14:32:25 but the fact you only get one interrupt is a problem. 2016-08-10T14:32:30 punitvara: you are not clearing or handshaking the interrupt properly at either the device or interrupt controller 2016-08-10T14:32:32 figure out if the ISR returns 2016-08-10T14:32:39 program should come out of transmit interrupt and then receive interrupt should be set 2016-08-10T14:32:57 *** monstr has quit IRC 2016-08-10T14:33:18 But if you don't clear/handshake the first interrupt correctly, you will not get another (or a continuous stream of the same) 2016-08-10T14:33:47 anyway.. you have a handle on what to debug 2016-08-10T14:33:47 https://github.com/punitvara/rtems/blob/new/c/src/lib/libbsp/arm/beagle/i2c/bbb-i2c.c#L464 2016-08-10T14:34:01 yeah I am clearing interrupt at the end as well 2016-08-10T14:34:32 there is problem with stop condition I guess. 2016-08-10T14:35:01 I am speaking in general terms. Every device is subtly different but the patterns are the same. Verify the IRQ controller code is clearing the source if it needs to. There are likely at least two stages to getting the interrupt handled. 2016-08-10T14:35:27 ohh okay. Thanks I will look into that 2016-08-10T14:35:57 thanks punitvara. darnir are you here? 2016-08-10T14:37:13 *** punitvara has quit IRC 2016-08-10T14:37:24 alright, Wonjun you can go now 2016-08-10T14:37:31 Hello. 2016-08-10T14:37:37 I am working on patching pci, uart, apic to execute hello example. 2016-08-10T14:37:45 When I execute master rtems on qemu, I get hello message. 2016-08-10T14:37:45 On the other hand, when I execute patched rtems on qemu, I get no message. 2016-08-10T14:37:49 ok. i expected to see more email from you 2016-08-10T14:37:54 I could see only blinked qemu screen with "Booting from ROM..." messages. 2016-08-10T14:38:08 I think I should update my blog to inform my status in detail and discuss with mentors. 2016-08-10T14:38:13 I will commit my repository and update my blog about patched codes with comments. 2016-08-10T14:38:37 ok. for making changes to the pc386 bsp pci/apic you definitely need to be talking more with DrJoel and kiwichris 2016-08-10T14:39:00 There are a lot of instructions from entry point to hello world. You will just have to dig into them. For polled console output, you don't need much working hardware wise at all. 2016-08-10T14:39:34 you should be able to connect gdb to RTEMS running inside qemu though, which should help 2016-08-10T14:39:37 With gdb, you should be able to set breakpoints in the rtems image in qemu and verify you at least get to it and it is running until doing something stupid 2016-08-10T14:40:03 if you don't get to boot_card (set breakpoint), then something is horribly wrong with the memory map vs the container 2016-08-10T14:40:05 ok. in patched files. it doesn't use pci but apic it makes problem harder. 2016-08-10T14:40:16 DrJoel: i don't know if you caught it, but he is stepping back from Jailhouse and looking to improve the apic/pci/ioapic support in RTEMS directly, hopefully that can make it easier to integrate with Jaiilhouse if the native i386 support is improved. 2016-08-10T14:40:24 Turn it off. Use polled console and get that working 2016-08-10T14:40:32 Oh... missed that 2016-08-10T14:40:48 Then gdb is your friend. :) 2016-08-10T14:41:14 i hadn't considered switching to polled console for jailhouse. but i think this direction is worth him pushing on? 2016-08-10T14:41:26 already polled console because patched rtems use polled console... 2016-08-10T14:41:27 he really needs a quick win though. 2016-08-10T14:41:31 ah 2016-08-10T14:42:06 You will need to use gdb to break at your modifications and see if you even execute through them. They could be crashing. 2016-08-10T14:42:51 If you get to the first context switch, it is common to get an unexpected interrupt if you didn't set them up or clear them correctly. Common problem for new BSPs 2016-08-10T14:43:11 right. he wasnt' able to get gdb working with Jailhouse, so it seemed better to try to get the Jailhouse support patches into pc386. agree, DrJoel? 2016-08-10T14:44:04 yep. You have to test them somehow. 2016-08-10T14:44:27 If they are in small chunks, you can test them one at a time. That would be better than debugging the entire mass 2016-08-10T14:44:43 I found that patched rtems also doesn't use nested irq. 2016-08-10T14:45:18 *** mudit1729 has quit IRC 2016-08-10T14:45:26 should I change patched rtems to use nested irq? 2016-08-10T14:45:59 It is already broken. Changing more isn't going to help. 2016-08-10T14:46:15 Start with Jan's patches. One at a time, apply and test. Assuming they are independent 2016-08-10T14:47:53 Wonjun: do you have your Jailhouse configuration posted somewhere? 2016-08-10T14:48:37 yes http://wonjunhwang.blogspot.kr/ 2016-08-10T14:49:09 ok 2016-08-10T14:49:54 i am going to try to replicate some of what you have done today. i have a little bit of time i can spend on this. 2016-08-10T14:50:21 do you have any other notes that might help me other than on your blog? 2016-08-10T14:50:47 hm. if my version of ubuntu is recent enough to even run jailhouse. :) 2016-08-10T14:51:13 I learned how to build jailhouse in jailhouse git page. 2016-08-10T14:51:24 https://github.com/siemens/jailhouse 2016-08-10T14:51:27 and 2016-08-10T14:51:36 https://groups.google.com/forum/#!forum/jailhouse-dev/join 2016-08-10T14:52:20 there are some articles in jailhouse doc folder. 2016-08-10T14:52:43 ok 2016-08-10T14:53:24 and I will sends email I have contacted with jan. 2016-08-10T14:53:34 send emails 2016-08-10T14:54:29 my host side is too old. :( 2016-08-10T14:54:53 am not sure i want to upgrade! 2016-08-10T14:55:39 my hardware might not be new enough either, hm 2016-08-10T14:57:17 Gedare: When you buy stuff for children, you don't get new toys for yourself? :) 2016-08-10T14:57:41 Don't worry eventually you can buy your own toys again but you will be too old to care. LOL 2016-08-10T14:57:57 haha 2016-08-10T14:58:56 well, i have vmx, but no iommu 2016-08-10T14:59:24 i use core-2 quad 2016-08-10T14:59:37 i don't even remember when i bought this... maybe 2010? 2016-08-10T14:59:38 ::) 2016-08-10T15:00:01 I am not sure but 2016-08-10T15:00:29 i don't have EPT. that may be a problem 2016-08-10T15:00:31 I dont use iommu using intel iommu disable option 2016-08-10T15:00:59 i suspect this wouldn't work for me on this machine 2016-08-10T15:01:13 but, what i can do is try to follow what you did 2016-08-10T15:01:19 even though i can't test it out 2016-08-10T15:01:56 so you ran jailhouse with linux without a problem after you figured out all the magic commands right? 2016-08-10T15:02:47 I would like to see the git master working with Jan's patches on qemu w/o Jailhouse. Wonjun.. get the patches back to smaller individual patches and try them one at time until the regular BSP breaks. That narrows it down considerably 2016-08-10T15:02:48 yes I can run apic example on jailhouse 2016-08-10T15:02:58 and it makes the work more amenable to merge. 2016-08-10T15:03:55 yeah i think that is a good plan, to rebase those patches 2016-08-10T15:03:59 hmm 2016-08-10T15:04:06 Wonjun: you have --disable-posix in your rtems configure 2016-08-10T15:04:09 this may not be a great plan 2016-08-10T15:04:22 also you have --enable-networking, which might not be good either 2016-08-10T15:04:32 i would --enable-posix and --disable-networking for starting out 2016-08-10T15:04:45 and leave cxx enabled too. 2016-08-10T15:06:02 Yes. I also got 5 patches files. but I worried it doesn't work when I patch just one file. 2016-08-10T15:06:11 I will try 2016-08-10T15:06:38 I will change configure options too. 2016-08-10T15:08:46 because jan's student didn't use that options, I didn't use posix. 2016-08-10T15:08:54 When Jailhouse loads the inmate, does it provide a virtual memory address space? 2016-08-10T15:09:53 yes. I wrote a config file about configuring memory region and pci devices. 2016-08-10T15:11:18 *** deval has quit IRC 2016-08-10T15:13:24 Wonjun: where is your rtems config file? 2016-08-10T15:13:54 i don't understand why your git-commits are so large and modify so many files. 2016-08-10T15:15:19 I misstook. becuase I changed file permission. 2016-08-10T15:16:10 ahh. 2016-08-10T15:16:14 where is the patch? I don't see a git link in the blog. 2016-08-10T15:16:22 https://github.com/hwj1593/jailhouse-wonjun/blob/master/configs/linux-x86-demo.c 2016-08-10T15:16:27 this is config file for jailhouse 2016-08-10T15:16:36 it should be better if you created an rtems-x86-demo.c? 2016-08-10T15:16:50 I am going to have to reboot in a minute. 2016-08-10T15:17:09 I was concerned about the patches. Just going to see if I could spot why they were large 2016-08-10T15:18:20 sorry It is just name I changed linux-demo for rtems to write config file easily. 2016-08-10T15:19:09 could you give me address you watch? 2016-08-10T15:19:30 Wonjun: you should verify that the .bin file is loaded at 0x3b600000. DrJoel does the pc386 expect to be loaded at address 0x0? 2016-08-10T15:20:28 *** benjamint has quit IRC 2016-08-10T15:20:49 [ 0x3f8/8 ... 0x3ff/8] = 0, /* serial1 */+ 2016-08-10T15:20:57 *** benjamint has joined #rtems 2016-08-10T15:21:09 if you set that to something else, will it enable access to serial1 port? 2016-08-10T15:22:03 since you build rtems without SMP support, it does not know how to deal with more than 1 cpu 2016-08-10T15:22:09 you should change the config to only use 1 processor 2016-08-10T15:22:35 and there is no reason to tackle smp configurations without a single CPU config working 2016-08-10T15:22:40 brb 2016-08-10T15:22:43 yes I checked other example. I will change cpu option. 2016-08-10T15:22:44 *** DrJoel has quit IRC 2016-08-10T15:23:12 ok, and look into that serial1 option, if that enables the COM1 port then you might get to see output 2016-08-10T15:23:52 and compare the size of hello.bin with the memory allocated by the config file 2016-08-10T15:24:08 loaded address would be 0x0 because config file shows that virtual address is 0x0. 2016-08-10T15:24:29 Ok I will check agin 2016-08-10T15:32:21 could I discuss more by email? beacuse It is 12:30 am in Korea. 2016-08-10T15:53:15 yeah sorry for keeping you up so late 2016-08-10T15:53:54 Wonjun: get some sleep. :) i will not be able to test Jailhouse, but i'll try to make a little progress on the pc386 BSP if i can 2016-08-10T15:54:25 *** jkridner has joined #rtems 2016-08-10T16:10:22 [trac] #2732/defect (new) updated by Gedare Bloom (In [changeset:"709594f0fb29d4f860b99cda264ea3e74c635a72/rtems"]: {{{ ...) http://devel.rtems.org/ticket/2732 2016-08-10T16:10:22 [git] 709594f by Gedare Bloom: [posix: nanosleep: adjust elapsed time calculation Use clock_gettime before and after sleep to calculate the time spent asleep, and the amount of time remaining. updates #2732] http://devel.rtems.org/changeset/709594f/rtems 2016-08-10T17:54:25 *** dr__house|away is now known as dr__house 2016-08-10T20:55:54 *** gedare has quit IRC 2016-08-10T23:20:32 *** dr__house is now known as dr__house|away 2016-08-11T04:46:19 *** benjamint has quit IRC 2016-08-11T04:46:25 *** benjamint1 has joined #rtems 2016-08-11T04:48:46 *** benjamint1 is now known as benjamint 2016-08-11T05:41:15 *** Wonjun_ has joined #rtems 2016-08-11T05:45:58 *** rokka has joined #rtems 2016-08-11T05:48:39 *** rokka_ has quit IRC 2016-08-11T05:48:39 *** Wonjun has quit IRC 2016-08-11T05:48:39 *** darnir has quit IRC 2016-08-11T05:48:49 *** darnir has joined #rtems 2016-08-11T05:56:24 *** jkridner has quit IRC 2016-08-11T06:03:12 *** monstr has joined #rtems 2016-08-11T07:07:00 *** jkridner has joined #rtems 2016-08-11T07:12:06 *** jkridner has quit IRC 2016-08-11T07:23:09 [git] 9a4eca5 by Chris Johns: [build-system: Always enable C++ if the compiler is present. We always build a C++ compiler and building with C++ does not effect RTEMS or the runtime. This patch always enabled the support. There is no need to manually enable it any more. You can disable C++ with '--disable-cxx'. If an architecture does not have a C++ compiler 2016-08-11T07:23:09 support is automatically d...] http://devel.rtems.org/changeset/9a4eca5/rtems 2016-08-11T07:26:43 [trac] #2777/defect (closed) updated by Chris Johns (In [changeset:"03c1038edbe9b01a72d4775dcb6ffc1a03193a0c/rtems"]: {{{ ...) http://devel.rtems.org/ticket/2777 2016-08-11T07:26:44 [git] 03c1038 by Chris Johns: [librtems++: Remove from RTEMS. This is old and there are better design patterns for threading and C++. We recommend you use the new C++ standards based support. Closes #2777.] http://devel.rtems.org/changeset/03c1038/rtems 2016-08-11T07:42:51 *** Wonjun_ has quit IRC 2016-08-11T07:43:01 [git] 26a8cee by Chris Johns: [e1000phy: Provide local support for Marvell's 88E1512 PHY. This PHY is close to the standard PHY support so providing enough to allow the PHY probe to pass lets it work on a MicroZed.] http://devel.rtems.org/changeset/26a8cee/rtems-libbsd 2016-08-11T07:43:18 *** Wonjun_ has joined #rtems 2016-08-11T07:43:55 *** eZag has joined #rtems 2016-08-11T08:06:53 sebhub: hi 2016-08-11T08:07:16 sebhub: I am seeing a crash after updating to HEAD on master. 2016-08-11T08:08:42 It is down in _Chain_Initialize_empty, from _Scheduler_priority_Ready_queue_initialize, _Scheduler_priority_Initialize 2016-08-11T08:16:51 do you have a stack trace? is it with RTEMS_DEBUG? 2016-08-11T08:17:31 p _Scheduler_Table->maximum_priority 2016-08-11T08:17:38 $2 = 1431323680 2016-08-11T08:17:46 and that is before it has run 2016-08-11T08:17:53 that is any code has run 2016-08-11T08:18:37 This is the value once it crashes so it is something wrong with the loading into RAM (never had an issue) or linking 2016-08-11T08:18:46 do you use C++ and is the application configuration done in C++ 2016-08-11T08:18:58 C++ yes, config in C 2016-08-11T08:19:08 LibBSD 2016-08-11T08:19:27 But I had no problem until I updated. I do not have the hash for it being ok 2016-08-11T08:19:38 But it was recent 2016-08-11T08:19:51 maybe do a waf clean and try it again 2016-08-11T08:19:57 I did 2016-08-11T08:20:24 I will clean all parts and try again, but that is for tomorrow :) 2016-08-11T08:20:41 I worked with the libbsd just 10 minutes ago without problems 2016-08-11T08:21:00 Yes I have another Zynq (Microzed) and it is working, this is a large large C++ app 2016-08-11T08:21:18 I may have missed a clean and build step. 2016-08-11T08:22:00 By the way I have a nice stand along asm exception handler etc for ARM (ARMv7) for debugginer. It is close to working. 2016-08-11T08:22:22 anyway time to head home 2016-08-11T08:22:25 good 2016-08-11T09:29:45 *** jkridner has joined #rtems 2016-08-11T09:33:59 [git] fef3ea9 by Sebastian Huber: [score: Add _Thread_queue_Surrender() Add _Thread_queue_Surrender() to unify the mutex surrender procedures which involve a thread queue operation.] http://devel.rtems.org/changeset/fef3ea9/rtems 2016-08-11T09:33:59 [git] a827447 by Sebastian Huber: [score: Dismantle _Thread_queue_Do_extract_locked() Dismantle _Thread_queue_Do_extract_locked() into re-usable parts like _Thread_queue_MP_set_callout() and _Thread_queue_Make_ready_again(). Use them in _Thread_queue_Surrender() to propare for a new thread queue surrender operation.] 2016-08-11T09:33:59 http://devel.rtems.org/changeset/a827447/rtems 2016-08-11T09:33:59 [git] 424ffe4 by Sebastian Huber: [score: Introduce thread queue surrender operation This is an optimization for _Thread_queue_Surrender(). It helps to encapsulate the priority boosting in the priority inheritance thread queue operations.] http://devel.rtems.org/changeset/424ffe4/rtems 2016-08-11T10:32:53 *** kiwichri_ has joined #rtems 2016-08-11T10:33:19 sebhub, a full clean and build has fixed the issue. 2016-08-11T10:37:35 [git] 691ed15 by Sebastian Huber: [rtems-bsd-mutex: Use _Thread_queue_Surrender()] http://devel.rtems.org/changeset/691ed15/rtems-libbsd 2016-08-11T10:38:59 *** kiwichri_ has quit IRC 2016-08-11T10:39:38 *** kiwichri_ has joined #rtems 2016-08-11T10:49:28 *** kiwichri_ has quit IRC 2016-08-11T10:49:50 *** kiwichri_ has joined #rtems 2016-08-11T11:11:27 *** kiwichri_ has quit IRC 2016-08-11T11:11:47 *** kiwichri_ has joined #rtems 2016-08-11T11:22:23 *** kiwichri_ has quit IRC 2016-08-11T11:22:45 *** kiwichri_ has joined #rtems 2016-08-11T11:24:13 *** kiwichri_ has quit IRC 2016-08-11T11:46:09 *** salpha has joined #rtems 2016-08-11T11:57:05 [trac] SOCIS/2016 edited by salpha (empty comment) http://devel.rtems.org/wiki/SOCIS/2016 2016-08-11T11:57:31 [trac] SOCIS/2016 edited by salpha (empty comment) http://devel.rtems.org/wiki/SOCIS/2016 2016-08-11T12:05:40 *** salpha has quit IRC 2016-08-11T12:43:10 *** gedare has joined #rtems 2016-08-11T12:43:10 *** ChanServ sets mode: +o gedare 2016-08-11T13:15:57 *** gedare has quit IRC 2016-08-11T13:31:01 *** gedare has joined #rtems 2016-08-11T13:31:02 *** ChanServ sets mode: +o gedare 2016-08-11T15:00:13 *** gedare has quit IRC 2016-08-11T15:09:05 *** dougk_ has quit IRC 2016-08-11T15:22:00 *** dougk_ has joined #rtems 2016-08-11T15:59:09 *** monstr has quit IRC 2016-08-11T16:04:54 *** dougk_ has quit IRC 2016-08-11T16:15:51 *** gedare has joined #rtems 2016-08-11T16:15:51 *** ChanServ sets mode: +o gedare 2016-08-11T16:17:30 *** dougk_ has joined #rtems 2016-08-11T16:23:36 *** dr__house|away is now known as dr__house 2016-08-11T17:19:56 *** gedare has quit IRC 2016-08-11T17:33:25 *** gedare has joined #rtems 2016-08-11T17:33:26 *** ChanServ sets mode: +o gedare 2016-08-11T17:52:09 *** Wonjun_ has quit IRC 2016-08-11T18:00:38 *** gedare has quit IRC 2016-08-11T18:18:07 *** mumptai has joined #rtems 2016-08-11T18:58:38 *** dr__house is now known as dr__house|away 2016-08-11T19:25:51 *** gedare has joined #rtems 2016-08-11T19:25:51 *** ChanServ sets mode: +o gedare 2016-08-11T19:50:16 *** gedare has quit IRC 2016-08-11T21:55:18 *** gedare has joined #rtems 2016-08-11T21:55:18 *** ChanServ sets mode: +o gedare 2016-08-11T22:09:04 *** gedare has quit IRC 2016-08-11T22:24:54 *** mumptai has quit IRC 2016-08-11T23:49:20 *** bbhoss_ has joined #rtems 2016-08-11T23:52:28 *** bbhoss has quit IRC 2016-08-11T23:52:33 *** bbhoss_ is now known as bbhoss 2016-08-11T23:57:45 *** gedare has joined #rtems 2016-08-11T23:57:46 *** ChanServ sets mode: +o gedare 2016-08-12T00:05:19 *** gedare has quit IRC 2016-08-12T04:12:45 *** darnir has quit IRC 2016-08-12T04:15:30 *** darnir has joined #rtems 2016-08-12T05:39:45 *** mumptai has joined #rtems 2016-08-12T05:48:54 *** monstr has joined #rtems 2016-08-12T06:10:26 [git] c83cf5c by Chris Johns: [nfsclient: Only parse for options when present.] http://devel.rtems.org/changeset/c83cf5c/rtems-libbsd 2016-08-12T06:10:26 [git] ab7739a by Chris Johns: [rc.conf/net: Let interface option appear on a DHCP ifconfig_ line. An ifconfig_ line can have other options for the interface such as ether or hardware checksum control. Allow these to be set on the interface.] http://devel.rtems.org/changeset/ab7739a/rtems-libbsd 2016-08-12T06:40:18 sebhub: devfs has a bug 2016-08-12T07:25:17 is the devfs a real file system on FreeBSD? 2016-08-12T07:40:27 *** cdesai has quit IRC 2016-08-12T07:45:01 *** monstr has quit IRC 2016-08-12T07:45:58 *** mumptai has quit IRC 2016-08-12T07:47:26 *** cdesai has joined #rtems 2016-08-12T07:50:48 I do not think so. 2016-08-12T07:51:13 ruru rtems-libbsd.master $ df -h 2016-08-12T07:51:14 Filesystem Size Used Avail Capacity Mounted on 2016-08-12T07:51:16 /dev/gpt/root 108G 38G 62G 38% / 2016-08-12T07:51:17 devfs 1.0K 1.0K 0B 100% /dev 2016-08-12T07:51:43 Depends how you define a real fs 2016-08-12T07:52:06 ok, but it is a file system and not just some tool that adds nodes to another file system 2016-08-12T07:52:20 like we do in RTEMS 2016-08-12T07:52:34 correct, it is a fs and stands alone. 2016-08-12T07:52:58 I guess we have to add some rtems_mkdir to the devfs create 2016-08-12T07:52:59 I suppose this means /dev in my case is built at start up which is a good idea 2016-08-12T07:53:11 Nick the code from untar 2016-08-12T07:53:43 Yes I think this is the simplest solution, stat the node make sure it matches. 2016-08-12T07:54:17 When you come to make the actual node you will find out if you have gone off the IMFS and that is an error. 2016-08-12T07:54:37 It would be nice to not just assert but to manage the error messages a little better. 2016-08-12T07:55:04 yes 2016-08-12T07:56:06 however, we normally run everything in the debugger and have break points to the assert etc. functions, so we don't read the messages 2016-08-12T07:59:19 You are an exception, users need more than an assert and in this case with a file system I would like more than just an assert. 2016-08-12T07:59:38 it took me and a debugger to figure all this out 2016-08-12T08:00:08 Some one may make a mistake and mount /dev on a ram disk and this breaks devfs. 2016-08-12T08:18:02 *** monstr has joined #rtems 2016-08-12T08:20:23 *** monstr has joined #rtems 2016-08-12T09:23:13 [git] 55faa44 by Sebastian Huber: [score: Improve _RBTree_Insert_inline() Return if the inserted node is the new minimum node or not.] http://devel.rtems.org/changeset/55faa44/rtems 2016-08-12T09:31:12 [trac] #2754/defect (new) updated by chrisj (Is this to be fixed on the 4.11 branch as well?) http://devel.rtems.org/ticket/2754 2016-08-12T10:13:17 *** benjamint has quit IRC 2016-08-12T10:13:45 *** benjamint has joined #rtems 2016-08-12T10:45:28 *** benjamint has quit IRC 2016-08-12T10:45:54 *** benjamint has joined #rtems 2016-08-12T10:51:32 *** benjamint has quit IRC 2016-08-12T10:52:00 *** benjamint has joined #rtems 2016-08-12T11:46:25 *** benjamint has quit IRC 2016-08-12T12:28:55 *** benjamint has joined #rtems 2016-08-12T12:33:41 *** benjamint has quit IRC 2016-08-12T12:40:19 *** benjamint has joined #rtems 2016-08-12T13:32:37 *** monstr has quit IRC 2016-08-12T13:57:47 *** sebhub has quit IRC 2016-08-12T14:21:47 *** mumptai has joined #rtems 2016-08-12T14:22:29 *** gedare has joined #rtems 2016-08-12T14:22:29 *** ChanServ sets mode: +o gedare 2016-08-12T14:38:58 *** gedare has quit IRC 2016-08-12T14:43:59 *** jkridner has quit IRC 2016-08-12T14:53:49 *** gedare has joined #rtems 2016-08-12T14:54:00 *** ChanServ sets mode: +o gedare 2016-08-12T15:59:00 [trac] #2754/defect (new) updated by joel.sherrill (Patrick should answer from their perspective. But I think it should. But from ...) http://devel.rtems.org/ticket/2754 2016-08-12T16:32:38 *** benjamint has quit IRC 2016-08-12T17:09:04 *** dr__house|away is now known as dr__house 2016-08-12T17:33:50 *** jkridner has joined #rtems 2016-08-12T17:40:35 *** benjamint has joined #rtems 2016-08-12T21:03:12 *** gedare has quit IRC 2016-08-12T21:05:06 *** dr__house is now known as dr__house|away 2016-08-12T21:05:44 *** gedare has joined #rtems 2016-08-12T21:05:44 *** ChanServ sets mode: +o gedare 2016-08-12T21:11:04 *** gedare has quit IRC 2016-08-12T21:26:36 *** benjamint has quit IRC 2016-08-12T23:02:02 *** gedare has joined #rtems 2016-08-12T23:02:02 *** ChanServ sets mode: +o gedare 2016-08-12T23:18:40 *** gedare has quit IRC 2016-08-13T02:01:38 *** gedare has joined #rtems 2016-08-13T02:01:38 *** ChanServ sets mode: +o gedare 2016-08-13T02:12:02 hi kiwichris 2016-08-13T02:58:10 *** mumptai has quit IRC 2016-08-13T03:11:38 *** mumptai has joined #rtems 2016-08-13T03:32:32 *** gedare has quit IRC 2016-08-13T08:11:52 *** ita has quit IRC 2016-08-13T08:18:35 *** ita has joined #rtems 2016-08-13T17:25:31 *** gedare has joined #rtems 2016-08-13T17:25:32 *** ChanServ sets mode: +o gedare 2016-08-13T17:36:03 *** gedare has quit IRC 2016-08-13T18:07:54 [trac] #2754/defect (new) updated by pggauvin ([comment:2 chrisj]: > Is this to be fixed on the 4.11 branch as well? It is ...) http://devel.rtems.org/ticket/2754 2016-08-13T19:38:16 *** guillaum1 has quit IRC 2016-08-13T19:51:57 *** guillaum1 has joined #rtems 2016-08-13T23:19:52 *** mumptai has quit IRC 2016-08-14T06:44:54 *** mumptai has joined #rtems 2016-08-14T17:51:08 *** Davidbrcz has joined #rtems 2016-08-14T17:58:30 *** gedare has joined #rtems 2016-08-14T17:58:30 *** ChanServ sets mode: +o gedare 2016-08-14T18:04:52 *** gedare has quit IRC 2016-08-14T19:26:08 *** Davidbrcz has quit IRC 2016-08-14T19:55:31 *** Davidbrcz has joined #rtems 2016-08-14T20:43:54 *** Davidbrcz has quit IRC 2016-08-14T21:51:15 [trac] #2775/defect (new) updated by ppisa (I think that disabling of MMU at runtime is highly problematic. You need to stop all ...) http://devel.rtems.org/ticket/2775 2016-08-14T22:38:09 [trac] #2775/defect (new) updated by chrisj (I agree it is problematic for a running system. I had not seen the note in your git ...) http://devel.rtems.org/ticket/2775 2016-08-14T23:19:59 *** mumptai has quit IRC 2016-08-14T23:23:06 [trac] #2775/defect (closed) updated by Chris Johns (In [changeset:"01aa1ba34a1696736ed5200ffa1d4be9963d99b3/rtems"]: {{{ ...) http://devel.rtems.org/ticket/2775 2016-08-14T23:23:06 [git] 01aa1ba by Chris Johns: [libbsp/arm: Add the TTB table to the default MMU set up as read/write. This lets the table be changed at runtime for dynamic loading and debugger support. Closes #2775.] http://devel.rtems.org/changeset/01aa1ba/rtems