**** BEGIN LOGGING AT Thu Jul 31 03:00:01 2014 Jul 31 03:01:02 Abhishek_ : what was the kernel patch you pushed? Jul 31 05:40:08 panto : how are we going to go about refactoring the driver? Jul 31 05:51:47 we have to port to mainline first Jul 31 05:52:08 the it must turn into a subsystem Jul 31 05:52:11 that other drivers use Jul 31 05:54:08 panto : so, we will have rproc_lib.c that contain all the necessary functions Jul 31 05:54:23 then each driver will be independent of each other Jul 31 05:54:59 i.e. I don't have to patch rproc_lib.c for every driver I want to support? Jul 31 06:00:20 it's not going to be a library Jul 31 06:00:27 it's going to be a subsystem Jul 31 06:10:07 I wonder if you can expose the PRU as a MFD device Jul 31 06:15:08 what exactly do you mean by subsystem? (is it going to be something like a module?) Jul 31 06:15:16 panto ^ Jul 31 06:18:22 no Jul 31 06:18:28 ds2, MFD sucks Jul 31 06:18:38 it's not well thought out Jul 31 06:22:05 * karki breaks his head over not knowing what panto means by subsystem. he has no idea what to google for wither :p Jul 31 06:22:26 *either Jul 31 06:22:39 pinctrl is a subsystem Jul 31 06:22:44 reset is a subsystem Jul 31 06:23:29 the idea is that it's not a driver or library, it's something that provides services for other drivers Jul 31 06:24:01 ah! thanks panto :) Jul 31 11:42:08 karki: Look at https://github.com/RobertCNelson/bb-kernel/tree/am33x-v3.8/patches/BeagleLogic Jul 31 12:11:47 Abhishek_ : would you have to refactor it again after panto cleans up the rproc stuff ? Jul 31 12:17:04 what is an alternative to dd on windows ? Jul 31 12:17:16 install linux Jul 31 12:17:22 or install cygwin Jul 31 12:17:24 and your have dd Jul 31 12:17:38 av500: looking something that i can use from my qt app Jul 31 12:17:49 open/write/close? Jul 31 12:17:49 on linux i can just subprocess dd and it works Jul 31 12:18:02 but on windows have no clue how to dump the flash image onto emmc Jul 31 12:18:33 https://www.google.com/search?q=windows+equivalent+to+dd Jul 31 12:19:25 found some tools but still wanted a dd port if there is one :( Jul 31 12:20:35 i'm not really keen into making 2 verisons of my qt app Jul 31 13:02:57 Abhishek_ : and I needed some help with tracing flow of control in your modular driver. Jul 31 13:19:05 panto : once the rproc subsystem is ready, would we have to patch the subsystem for every new driver that comes in? (like abhishek modified rproc.c for beaglelogic) Jul 31 13:20:39 no Jul 31 13:57:32 panto : so now if I modularize my driver like Abhishek did, I'd have to redo stuff again when we clean up the rproc, right? Jul 31 13:58:04 your modularization is a good first step Jul 31 14:49:08 karki: pong Jul 31 14:49:32 Abhishek__ : :) Jul 31 14:49:38 I'm facing difficulties with connectivity over proxy, sockets do not seem to be working properly Jul 31 14:49:58 I have difficulties thanks to BSNL :( Jul 31 14:50:09 anyway, coming to your driver Jul 31 14:50:57 what is the flow of control from when you load the dtbo? Jul 31 14:51:40 Abhishek__ : I'm not able to cleanly figure out how the connection b/w the drivers work. Jul 31 14:51:52 karki: BeagleLogic is a misc device in a platform driver Jul 31 14:52:24 *platform device Jul 31 14:53:27 it's just a set of function pointers that are exchanged Jul 31 14:54:16 https://github.com/abhishek-kakkar/BeagleLogic/blob/master/beaglelogic-kernel-driver/BB-BEAGLELOGIC-00A0.dts#L209 triggers the probe of beaglelogic after the remoteproc module is probed. Jul 31 14:54:51 (which includes uploading the beaglelogic firmware in the PRUs, and the BeagleLogic driver verifies this before it loads, else it fails) Jul 31 14:58:35 ahhhr Jul 31 14:58:41 got disconnected Jul 31 14:58:51 but thanks for https://github.com/abhishek-kakkar/BeagleLogic/blob/master/beaglelogic-kernel-driver/BB-BEAGLELOGIC-00A0.dts#L209 Jul 31 14:58:57 it solves my doubt Jul 31 14:59:48 but I don't get how the pru_rproc is kicked off first..... Jul 31 14:59:59 why is that kicked off at all? Jul 31 15:00:53 Abhishek__ : because pru_rproc looks for compatible "ti,pru-rproc" Jul 31 15:02:41 karki: another thing is that the BeagleLogic module uses exported functions from the remoteproc module, so it is loaded first before beaglelogic can load Jul 31 15:03:04 ah, so it's because of the EXPORT Jul 31 15:03:05 ? Jul 31 15:04:21 Abhishek__ ^ Jul 31 15:04:33 yeah, and since the remoteproc module is compiled into the kernel, it is already present though Jul 31 15:07:25 Abhishek__ : so as far as I use EXPORT, the dependency causes the pru_proc probe to be kicked off before BeagleLogic probe. right? Jul 31 15:08:42 no, device tree fragments cause the probes to be triggered Jul 31 15:08:55 panto: Please correct me if I am wrong Jul 31 15:09:32 but no where in the dts is there a use of pru_rproc Jul 31 15:09:39 pru_rproc looks for compatible "ti,pru-rproc" Jul 31 15:11:37 there is Jul 31 15:19:35 Abhishek__ : where is the compatible "ti,pru-rproc" in your dts? Jul 31 15:21:48 84 Jul 31 15:21:50 k Jul 31 15:24:48 karki: then look at https://github.com/abhishek-kakkar/BeagleLogic/blob/master/beaglelogic-kernel-driver/beaglelogic_glue.h Jul 31 15:25:54 yep had a look at that :) Jul 31 15:26:50 I think that's pretty much all to the modularization on my part Jul 31 15:27:45 Thanks :) I'll ping if I'm stuck again! Jul 31 15:28:44 * Abhishek__ got a clear acrylic case for his BBB today :) Jul 31 15:34:35 you should have 3d printed it ;) Jul 31 15:35:32 I haven't really heard of a 3D printed acrylic case Jul 31 15:39:17 where did you order? Jul 31 15:44:44 I found it at eBay Jul 31 15:59:22 If I had a 3D printer I would surely have printed one for myself **** ENDING LOGGING AT Fri Aug 01 03:00:00 2014