**** BEGIN LOGGING AT Tue Oct 02 03:00:02 2018 Oct 02 03:42:47 So, P9.42 is connected to both C18 and B12. How can I enable pru_eCAP, without damaging B12? (e.g. B12 in mode pru0in4 ? ) Oct 02 06:03:14 why would it damage it? Oct 02 06:03:50 as long as you don't configure the two pins both into output mode and drive opposing values there's nothing to worry about Oct 02 06:57:09 I am trying to connect my BBBW with wireless router. I turned agent on, wifi enable and tether wifi disable. I scanned wifi and tried to connect with AP from services of connmanctl. But I am getting error Oct 02 06:57:47 connect-failed agent request canclled by connman Oct 02 06:57:53 operation aborted Oct 02 06:58:09 I am not sure what is wrong as this same BBBW works fine on other router Oct 02 08:41:55 clear Oct 02 08:42:59 I am getting connect-failed ; agent request cancelled by connman; operation aborted while trying to connect with wifi router using connman Oct 02 09:04:24 ? Oct 02 10:16:15 on other router....you have 2 routers of exact same model? Oct 02 11:44:51 what nodes do I need to overlay in order to enable the ecap module (in the PRU), and configure as pwm output? Oct 02 11:51:13 &am33xx_pinmux and.. ? Oct 02 12:18:36 you mean you want to use the pru ecap module with the linux driver rather than using it from pru? I suppose it could work Oct 02 12:44:51 I mean, &am33xx_pinmux{ pru_cap_pins: pru_cap {... then in which node does go the corresponding <&pru_cap_pins> ?? I'm using uio for the pru, and using pru_ecap from the pru Oct 02 12:45:17 then just attach the pinmux to the pru Oct 02 12:49:20 &pruss1 { ... pinctrl-0 = <&pru_cap_pins> ?? Oct 02 12:49:33 for example yes Oct 02 12:50:40 okay Oct 02 13:01:57 I think DT deserves an entire book... C pointers has one, grep has one, Regular Expressions has one.... Oct 02 13:03:33 the PRU deserves its own also Oct 02 13:21:08 I have no idea how you might be able to fill a book on DT Oct 02 13:21:14 nor C pointers for that matter Oct 02 13:32:00 there are at least four C pointers books out there Oct 02 13:36:47 I feel like, if someone feels the need for an entire book on C pointers, maybe they should reconsider their choice of programming language :P Oct 02 14:00:17 zmatt: or reconsider programming in general ;) Oct 02 14:07:54 or reconsider writing books and publish them Oct 02 14:46:08 can I ask questions right here? Oct 02 14:46:39 you just did Oct 02 14:49:07 thx. I want to know how to use bbb for setting mesh network. I'm setting mesh network using open80211s, Is it possible Oct 02 14:49:28 the bbbw you mean probably? Oct 02 14:50:00 yep Oct 02 14:50:14 I know the chipset supports 802.11s networks, but that's about all I know Oct 02 14:51:38 you'll probably want to disable connman and manually configure wpa_supplicant Oct 02 14:54:19 TI has some notes about mesh networks: http://www.ti.com/lit/an/swaa166/swaa166.pdf but they're referring to scripts that are part of TI's Processor SDK for Linux and not present on the beaglebone by default Oct 02 14:54:46 you could of course download the sdk and inspect those scripts Oct 02 14:54:57 oh thanks. Oct 02 15:13:15 why is the pin index multiplied by 4 as in pinctrl-single,pins = < ((i)*4) (m) ... > ?? Oct 02 15:13:54 because the pinctrl-single driver expects an offset rather than an index Oct 02 15:15:55 and that offset has anything to do with the ControlModule regs? Oct 02 15:16:03 yes Oct 02 15:18:00 it's the offset to the relevant padconf register within that register block Oct 02 15:21:13 ayjay_t: would you be interested in code to use the modern drm interface rather than the legacy fbdev interface? (that would also open the possibility of doing double-buffering to avoid tearing) Oct 02 15:23:02 it would also avoid messing with KD_GRAPHICS, it simply takes over the display and when the program exits it automatically switches back to whatever it was showing before Oct 02 15:23:48 * thinkfat raises head Oct 02 15:26:52 * zmatt wonders what thinkfat sees from this elevated position Oct 02 16:14:25 in an overlay, the name (e.g. spi0), after a label (e.g. spi0_pins:) and before pinctrl-single... how is choose? Oct 02 16:15:36 whatever you want, as long as it's different from other nodes and properties of the parent node (&am33xx_pinmux) Oct 02 16:15:52 in contrast, labels (spi0_pins:) must be globally unique Oct 02 16:42:13 are there any utzmatt: yeah sbolutely Oct 02 16:42:17 omg that came out wrong Oct 02 16:42:19 zmatt: absolutely Oct 02 16:42:38 is there still a bitmap like interface? Oct 02 16:43:07 and i'm really glad you took a look at our code- its my concept/architecture but it was implemented by my colleague who was an intern at that time Oct 02 16:45:17 i think it would probably make sense to include both implementations as either a run/compile-time switch Oct 02 16:45:43 i'm also sort of assuming there is a driver to run a display from the drm/fbdev interface over SPI Oct 02 16:55:21 ayjay_t: ignoring the setup/modesetting stuff, the drm lets you allocate one or more framebuffers (with pixelformat selected from a list of those supported by the hardware), mmap() them, and change which framebuffer should be displayed on screen (which takes effect on next vsync). you can also get an event to know when the vsync has happened Oct 02 16:57:52 i wonder if i can get it working with this project from amazon: http://a.co/d/hFzhMNN Oct 02 16:57:57 projecter* Oct 02 16:59:01 and dunno what sort of drivers exist for which SPI displays Oct 02 16:59:21 yeah, that's going to be an issue with small screens. but its such a common problem, i'd be its been solved Oct 02 16:59:24 bet* Oct 02 17:00:08 I know there's some stuff related to it, but no doubt there are an infinite number of different protocols used by spi displays so I doubt it's been "solved" Oct 02 17:00:45 in general I don't know how well the existing APIs are suited to "command mode" / "remote framebuffer" displays Oct 02 17:02:18 ah right there's an api to mark framebuffer regions as "dirty", I forgot about that Oct 02 17:03:34 i've implemented a couple screens and i can do without most of the advanced functionality, as long as i can set some commands to wrap a frame dump i'd be happy Oct 02 17:03:38 but its a different animal for sure Oct 02 17:03:58 so there's the tinydrm infrastructure in mainline for "small, simple displays" Oct 02 17:04:15 also a bunch of out-of-free tinydrm display drivers here; https://github.com/notro/tinydrm Oct 02 17:07:04 well, i already have a patch for the issue now which I'll implement this weekend if I get to it, and I'll try to stream a fix on twitch https://paste.ofcode.org/76bvEeKN2wgaE8yamnitsL Oct 02 17:07:26 i may start looking into your suggestions too Oct 02 17:09:00 does anyone have an hdmi mini display they use for the bbb that they could recommend? i'd rather do the fix on that first and then do it on a tft afterwards Oct 02 17:09:23 i dont want to buy one off amazon and then find it for whatever reason its not compatible, i don't know if any hdmi touch is good enough Oct 02 17:10:02 I'm writing a quick C++ wrapper for (the modern parts of) the drm api. it's been on my to-do list for far too long Oct 02 17:11:06 a lot of people obviously need a better solution for displays, i'm hoping uxmux takes off Oct 02 17:19:32 i also just order two of the new chinese $2 arm-wifi interfaces Oct 02 17:26:36 two pocket beagles too Oct 02 17:26:52 it might be some time for some kernel hacking! Oct 02 17:41:44 the 2/4 USB HUB cape looks good and tindie sells it but its e-post from japan and thats hard Oct 02 17:42:03 i can have ethernet by november lol Oct 02 17:51:21 https://github.com/beagleboard/am335x_pru_package/blob/master/pru_sw/example_apps/PRU_memAccess_DDR_PRUsharedRAM/PRU_memAcc_DDR_sharedRAM.c#L204-L218 Oct 02 17:51:53 how is possible to write to arbitrary offsets in DDR? Oct 02 17:58:55 ew who wrote this disgusting example Oct 02 17:59:12 you should never need to use /dev/mem when interacting with pru Oct 02 18:00:05 could they reserve that portion of DDR in the DT ? Oct 02 18:00:42 the amount of ddr memory allocated by the uio_pruss driver is specified in a module parameter Oct 02 18:02:05 its size can be queried using prussdrv_extmem_size() Oct 02 18:02:18 the pointer can be retrieved using prussdrv_map_extmem Oct 02 18:03:50 And it is contiguous? Oct 02 18:04:11 yes Oct 02 18:06:10 holy shit, what the fuck, the example you linked to directly meddles with DDR3 memory at some fixed address? o.O Oct 02 18:06:37 yes Oct 02 18:07:46 I think Oct 02 18:08:12 this example should probably be removed entirely Oct 02 18:09:39 or needs some serious modification at the very least Oct 02 18:09:48 PRU_PRUtoPRU_Interrupt ditto Oct 02 18:12:18 and there really needs to be a better C/C++ library for using uio-pruss... both libprussdrv itself and its examples are pretty gross Oct 02 18:13:45 So... can I follow this example instead https://github.com/google/prudaq/blob/master/src/examples/round_robin/round-robin.c ?? Oct 02 18:14:50 that seems fine Oct 02 18:14:58 (based on a very quick glance) Oct 02 18:16:22 hmm, maybe I should make an example of how to use py-uio to setup pru and then spawn a C/C++ process for high-performance interaction with pru via shared memory Oct 02 18:16:35 to get the best of both worlds Oct 02 18:26:51 i really want my office to use your work but they're not into it- they're using the tutorial chain of elinux i think? Oct 02 18:27:03 is your py-uio featured on elinux or bbb.io yet? Oct 02 18:27:09 (zmatt) Oct 02 18:27:20 I have no idea Oct 02 18:39:18 zmatt: Have you thinking rewrite a C/C++ version of py-uio? Oct 02 18:40:02 sure, just not the time and motivation yet Oct 02 18:40:15 it's just stuck somewhere on my to-do list Oct 02 19:40:33 zmatt do you have a spreadsheet about the am335 registers/mmap? like the one (sexy looking) for pins? Oct 02 19:42:13 I have one for the global memory map, i.e. all subsystems/peripherals Oct 02 19:42:27 :D Oct 02 19:43:07 obviously not for all registers of all peripherals since that would be completely insane, plus it would be a headache to try to encode all relevant type information in a spreadsheet Oct 02 19:44:09 https://goo.gl/UHF2Fy is my memory map spreadsheet for subarctic (am335x) and centaurus (dm814x) Oct 02 19:47:01 thanks :D Oct 02 19:47:29 and while I'm at it, the one for irqs/events => https://goo.gl/7YooOO Oct 02 19:50:58 beauty **** ENDING LOGGING AT Wed Oct 03 03:00:01 2018