**** BEGIN LOGGING AT Tue Mar 16 03:03:12 2021 Mar 16 03:28:12 just a thought, one of the largest internet spammers in the world, was found clubbed to death in their home in Russia. The person felt it was not only his right to spame email but that he was blessing the world by doing so. Result for 3 months email spam dropped 80% afterwards it resumed to "normal" albeit gradually. Most spammers are from India China and Russia. Mar 16 03:32:44 apparently his death was unrelated Mar 16 03:35:34 to be fair I nary heard what they determined happened. However it did result in a very drastic drop (ahem) just saying. Mar 16 03:36:00 quite a while ago though Mar 16 03:36:11 2005 Mar 16 03:37:38 that's when (at the time) email spam was consuming 30% of the internet total bandwidth and was costing 250million US a year in total, that includes payload lost service electricity. It does not include people being defrauded etc. Mar 16 03:40:22 Anyhow I was just thinking of the type of people who spam and their motivations. Mar 16 03:42:06 I wasn't thinking of eliminating them but demotivating instead. Mar 16 04:55:28 hello i need a help regarding with beaglebone black board Mar 16 06:02:19 Hey guys I have just bought a BBB wireless (Linux beaglebone 4.4.41-ti-r83 #1 SMP Tue Jan 17 00:01:19 UTC 2017 armv7l GNU/Linux). I connect to the board using USB cable connected to client USB fed from my Windows10 DELL 5550 Precision laptop and after some time (15-20min) the USB stops responding (USR0 LED stops blinking). I am wondering if it's Mar 16 06:02:20 due to the fact I am powering the board via USB ? Mar 16 06:03:18 sounds more like a kernel panic to me, unless all leds turn off including the power led Mar 16 06:03:31 also sounds like you're running quite an ancient system Mar 16 06:04:08 Leds continue turned on, only USR0 stops flashing Mar 16 06:04:49 You mean I should update Linux system ? Mar 16 06:05:15 I'd suggest reflashing to the latest iot image Mar 16 06:05:51 Okay. I'll try reflashing then. Thanks zmatt Mar 16 06:06:30 i.e. get the "AM3358 Debian 10.3 2020-04-06 4GB eMMC IoT Flasher" from https://beagleboard.org/latest-images, flash it to sd card (using Etcher is highly recommended), boot the beaglebone from that card and it should automatically reflash eMMC Mar 16 06:07:25 I see. Thanks zmatt very helpful ! Mar 16 06:08:04 preferably power on with the S2 button held down (the one closest to the card slot), you can let go once the power led turns on. this is usually not needed, but it may be if the bootloader on eMMC is ancient enough Mar 16 06:08:35 Got it ! Mar 16 06:08:40 (if the S2 button is held down at the moment of power-on, the bootloader on eMMC is bypassed and it'll use the one on SD card instead) Mar 16 06:09:27 Okay. Great, Thanks ! Mar 16 08:15:08 Hello everyone..I am trying to execute spidev_test on Beaglebone AI. However, I am receiving stream of 0's instead of actual data. How do I resolve this? Mar 16 08:24:07 zmatt: Did you try the spidev_test on Beaglebone AI? Mar 16 09:01:28 Hello, I have a problem with the configuration of the pwm with the Beaglebone black. I have installed the latest version of debian available (10.3) and looking on the internet it seems that the management has changed with the latest version. Mar 16 09:01:29 I disabled HDMI and was trying to configure PWM by anchoring on: Mar 16 09:01:30 and then Mar 16 09:01:30 echo 1> enable but it comes to me Mar 16 09:01:31 reported the following error: Mar 16 09:01:31 bash: echo: write error: Invalid argument Mar 16 09:02:43 how to use PWMs hasn't really changed Mar 16 09:02:57 have you configured period and duty_cycle first, before trying to enable? Mar 16 09:04:11 anpra: not yet Mar 16 09:04:37 I tried but also by entering: Mar 16 09:04:38 echo 25000> duty_cycle Mar 16 09:04:38 the same error is returned to me Mar 16 09:05:44 what period did you configure? Mar 16 09:08:10 I ran the period cat and it was set to 0! I tried to configure it to 50000 and by doing so I was able to also set duty cycle enable! Mar 16 09:08:58 But when I connect the oscilloscope I don't see anything. Mar 16 09:08:58 I am configuring the pwm 6: 0 and I am rightly in the p8_19 Mar 16 09:09:43 I wouldn't rely on pwm device numbering, there are convenient symlinks in /dev/pwm/ Mar 16 09:10:17 did you configure p8_18 to pwm using config-pin ? Mar 16 09:10:35 (config-pin p8_18 pwm) Mar 16 09:10:41 p8_19 sorry Mar 16 09:11:32 also that's not pwm-6:0, at least not on my beaglebone Mar 16 09:11:45 p8.19 is ehrpwm2a Mar 16 09:12:03 i.e. /dev/pwm/ehrpwm2a/ Mar 16 09:12:45 when i write "config-pin p8_19 pwm" Mar 16 09:12:51 ERROR: open() for /sys/devices/platform/ocp/ocp:P8_19_pinmux/state failed, No such file or directory Mar 16 09:13:09 uhh that's strange, it works fine here Mar 16 09:13:27 are you booting from eMMC or from sd card? Mar 16 09:15:34 Actually it is the first time that I have been given this error, I had already entered this command and it had not given any errors. I have done various tests and most likely I have made some big mistakes myself. I'm using the SD, I try to do the flash again which I probably solve !! Thank you very much you have been very useful Mar 16 09:16:42 flashing to eMMC is recommended yes Mar 16 09:17:12 (an old bootloader on eMMC can cause weird problems when booting from sd card) Mar 16 09:22:12 Perfect! thanks a lot again! Mar 16 11:51:35 m Mar 16 12:43:55 How to wake up from suspend using a GPIO input on Bank0? Mar 16 16:46:09 Hi,  Anyone has any suggestions on how to wake up from suspend using a GPIO input on Bank0? I found there is no wakeup state for any gpio pins under /sys/class/gpio Mar 16 17:02:03 I'm curious: If the PocketBeagle was designed today, do folks think that the OSD3558 would still be the best SoC, or has something better suited come out in the few years since? Mar 16 17:04:41 I think the pocketbeagle was designed because the "chip" was available Mar 16 17:08:33 And that having a chip available that fit that particular niche isn't common? Mar 16 17:09:10 I haven't searched so I couldn't say Mar 16 17:09:33 are imx or whatever socs available in bare die form? Mar 16 17:12:19 you have to buy wafers directly from the supplier. almost no one sells them through distribution. Mar 16 17:12:48 right, and it's not a given that they can be obtained at all Mar 16 17:15:35 Ok, that makes sense. But of course there are plenty of other ARM SoCs that might work too. Mar 16 17:16:05 yes, if you can buy the unpackaged silicon Mar 16 17:17:21 right. That seems to be the issue. Most of those SoCs seem to be impossible to actually order. :D Mar 16 17:20:33 normally packaged imx chips are readily available Mar 16 17:33:52 Hello, beautiful people! Do you have an approximate price for the BeagleV board? Will it ship to Mexico? Mar 16 17:34:36 It would be nice to play around with RISC-V. Mar 16 17:37:54 Oh, it seems this room is quite lonely. Mar 16 17:38:18 or nobody knows the answer Mar 16 17:38:53 I would assume they're aiming for a price below $100 Mar 16 17:39:00 or at least not a lot more Mar 16 17:39:09 I say it because I see only two people in the participants list. Mar 16 17:39:22 really? I see 181 Mar 16 17:39:33 It's probably my client. Mar 16 17:39:58 Not sure. Mar 16 17:40:34 OK, now I see everyone. :P Mar 16 17:41:04 Sorry. Mar 16 17:41:34 most are just lurking, obviously Mar 16 17:42:32 The ones using Matrix will probably receive the message until minutes or hours later, given the latency of the IRC bridge. Mar 16 17:43:09 At least that's what I've experienced. Mar 16 17:53:31 It seems to be working OK today. I'm on matrix and was having a normal conversation with mru a bet ago. Mar 16 18:03:09 rschulman: it's a bit of a strange question, since I think the goal of the pocketbeagle was to design a tiny board around the osd335x, hence necessarily no other SoC would fit that purpose :P Mar 16 18:04:15 to assess whether "better" SoCs are available for some particular purpose, you'd first need to define that purpose Mar 16 18:06:07 for applications that benefit from PRU, the am335x may very well still be the best option (if the higher-end TI SoCs are overkill for the application) Mar 16 18:08:05 I personally don't know of a SoC I'd rather be using, but I also haven't really looked (and of course I'd rather stick with a SoC I know deeply than switch to one that might on paper be "better" but whose problems I don't know) Mar 16 18:39:51 zmatt: That's all really useful context, thank you! Mar 16 18:41:30 I'm thinking of a SBC that can run linux, have wifi and BT, and operate off battery at least for a few hours, all while staying within roughly 40x40 mm. Mar 16 18:41:59 And that would have fairly little else in the way of interfaces. Mar 16 18:44:31 looking at the rpi zero w, it could probably be condensed to that size if you omit some interfaces and maybe put components on both sides Mar 16 18:44:42 of course that's of no use to you since you can't buy the SoC Mar 16 18:47:02 Hi @zmatt  , Is it possible to wake up from suspend using a GPIO input on Bank0? Mar 16 18:47:53 I'm pretty sure I've mentioned before that I don't know how to configure wakeup sources Mar 16 18:48:02 like, yes the hardware can do so just fine Mar 16 18:49:06 and actually I think when I experimented with suspend (which was a _long_ time ago) I don't think I had to do anything special to get it to wake up from gpio0 interrupts, but it's quite possible I just misremember Mar 16 18:50:17 have you tried configuring interrupts on the gpio? set the "edge" attribute of the gpio in sysfs to "rising", "falling", or "both" Mar 16 18:51:28 there are many wakeup sources preconfigured. It can be PWR_BUT, UART0_RXT, RTC, etc. But I found that there is no wakeup (item) in the exported gpio pins under /sys/class/gpio folder. Mar 16 18:52:55 rschulman: the AM335x is also available in a 13x13mm package (ZCE) Mar 16 18:53:34 zmatt: Thanks, yeah, that makes sense. Mar 16 18:54:26 rschulman: of course you'd still need space for ram and eMMC in that case, and power circuitry (though iirc it is possible to use a simpler power scheme than the usual PMIC if you use the ZCE and don't need cpu freq above 500 MHz, but I don't know the details) Mar 16 18:54:41 there are a lot of kernel wakeups: for example:  ./kernel/irq/19/wakeup, But I don't know how to relate the number, i.e 19 to a GPIO pin Mar 16 18:56:45 Or in reverse order, for GPIO0_26, what is the number mapped to kernel: /sys/kernel/nnn ? Mar 16 18:57:31 why would you need to know irq numbers? those are dynamically allocated at runtime, there's no fixed mapping Mar 16 18:59:00 /sys/kernel/irq/ is debug information Mar 16 18:59:04 There is a wakeup under every sys/kernel/nnn/wakeup Mar 16 18:59:40 Just some thoughts for testing, to see it can work Mar 16 18:59:47 ? Mar 16 19:00:01 you can't enable/disable wakeup there, it's a readonly attribute Mar 16 19:01:27 I see. Ok. I guess it cannot be done for a GPIO pin to trigger a wakeup Mar 16 19:01:37 if a gpio has interrupts enabled you can find it in /proc/interrupts Mar 16 19:02:23 then you guess wrong, gpio interrupts are supported. just because nobody here right now happens to know the answer to your question doesn't mean something "cannot be done" Mar 16 19:02:41 very few people use suspend Mar 16 19:03:15 also, I gave a suggestion on something to try. have you tried it? Mar 16 19:03:32 19:50 <@zmatt> have you tried configuring interrupts on the gpio? set the "edge" attribute of the gpio in sysfs to "rising", "falling", or "both" Mar 16 19:05:11 I haven't tried interrupts, yet Mar 16 19:05:37 wait I just realized I do know of a way to explicitly request wakeups on a gpio: you can use an overlay to create a gpio-keys device and set the "wakeup-source" property: https://github.com/mvduin/overlay-utils/blob/master/gpio-wakeup-key.dtsi Mar 16 19:06:10 it's not very convenient, but I don't know what other options linux offers to configure wakeup sources Mar 16 19:07:40 Do I need to do anything on GPIO0_26? Is it pre-configured as a wakeup source? Mar 16 19:08:02 I can try Mar 16 19:10:34 GPIO0_26 (P8_14) is used for something else , in our board Mar 16 19:10:47 so use a different one, it's just an example Mar 16 19:17:52 the hardware supports wakeup from "standby" using any gpio (or indeed any interrupt), and wakeup from "deepsleep" using any gpio0-pin (as well as usb, uart0, i2c0, rtc alarm, and tsc) Mar 16 19:20:01 I tried on GPIO0_20, it didn't work Mar 16 19:20:39 Maybe I need to configure the interrupt for gpio0_20 Mar 16 19:23:49 tiger: you can try, and I suggested that _half an hour ago_, I don't know if it suffices. it is possible linux currently lacks any convenient way of enabling wakeup sources from userspace, it's possible the device-tree method I suggested 20 minutes ago is the only way Mar 16 19:25:01 and this concludes the help you're going to get from me, I'm kinda done with trying to hold your hand while you're mostly ignoring my suggestions Mar 16 19:25:09 good luck. Mar 16 19:29:19 Really appreciate your help. Mar 16 23:07:27 i am finding my pocketbeagle very buggy. no amount of trying got usb internet going. wifi did work but the device needs to be replugged after boot Mar 17 02:32:25 How would I enable SPI on the AI? I am currently trying to configure bbai-bone-buses.dtsi w/ a dts file. I have nothing so far. Mar 17 02:33:53 Currently, I have reviewed the am5729-beagleboneai.dts and am5729-bone-common-univ.dts Mar 17 02:34:43 I found that currently the SPIDEV 2.0 pins are already at 'default' on the AI w/ cs, clock, COPI, and CIPO. Mar 17 02:35:47 They are P9.17a/18a/22b/21b. Mar 17 02:37:35 W/ relation to a .dts file, which I will keep learning about for the sake of learning, what exact config. would the AI understand in syntax of a .dts file? I mean...do I need any additional info. to give the .dts file to describe the SPI bus hardware outside of what is listed currenctly? Mar 17 02:43:19 I know it is late and I am going to try a loop back test real quickly... Mar 17 02:43:39 If you understand my issue, please go back in time to help this little man win a prize! **** ENDING LOGGING AT Wed Mar 17 02:59:56 2021