**** BEGIN LOGGING AT Wed Feb 24 03:02:58 2021 Feb 24 03:13:57 kveremitz: btw, why the "k" ? Feb 24 03:15:34 this is my new [to me] kubuntu laptop.. I have a new znc account to separate from the old veremitz with the baggage that went with it Feb 24 03:15:43 kinda didn't re-fudge it til sorta recentlyish Feb 24 14:30:03 Has anyone tried Android automotive OS on any beaglebone platform boards? Feb 24 14:33:34 Hello, I'm trying to get interrupt through gpio, exactly gpio 60 (P9.12). Feb 24 14:34:00 cat /proc/interrupts Feb 24 14:34:27 show me the interrupt has been registered Feb 24 14:34:32 64: 0 4804c000.gpio 28 Edge irqtest Feb 24 14:35:05 but a got no interruption Feb 24 14:35:36 can someone guide me on this Feb 24 14:36:06 http://paste.debian.net/1186745/ this is my dts part for device Feb 24 14:36:27 and driver code http://paste.debian.net/1186744/ I'm trying Feb 24 17:37:20 lucascastro: why are you using a gpio declaration instead of an irq declaration? the gpio controler is also an irq controller Feb 24 17:37:38 your driver does not need to know or care that the irq is delivered via a gpio Feb 24 17:39:44 lucascastro: https://github.com/mvduin/py-uio/blob/master/dts/gpio-irq.dtsi#L27-L28 Feb 24 17:45:41 and then use platform_get_irq() Feb 24 17:47:07 (if you have multiple irqs it's probably best to give them names using interrupt-names property and use platform_get_irq_byname() .. both functions also have an _optional() variant which doesn't log an error in the kernel log if the irq can't be obtained) Feb 24 17:49:48 and then devm_request_irq() on the irq number it returns Feb 24 17:50:33 (which is generally preferred over plain request_irq() since it makes sure the irq is released if your device is removed, including if the probe fails after requesting the irq) Feb 24 18:04:05 zmatt, I'll take a look at that you have pointed. Feb 24 18:04:45 My code is an example from Linux Device Driver Cookbook from packt. So I'm a little lost at that. Feb 24 18:04:56 packtpub Feb 24 18:05:20 if this driver code is an example, then the source of that code is probably a bad source of examples Feb 24 18:05:39 I found a code that really works from derek molloy, but not using device tree. Feb 24 18:06:03 use DT, use the two calls I just mentioned Feb 24 18:06:30 So, I'll try to get it to use device tree. But before I'm going to take a look what you pointed. Feb 24 18:07:29 device tree is what's responsible for your driver getting probed at all Feb 24 18:08:18 kernel sees the declaration in DT, creates a platform device for it, searches for a driver based on the declared "compatible" property, and calls its probe Feb 24 18:10:40 (or if it can't find the driver it calls out to userspace which will search the module index to find the kernel module containing the appropriate driver and automatically load it) Feb 24 18:31:00 lucascastro: btw your DT fragment doesn't show *where* you are declaring your device... since it's evidently a virtual platform device (not corresponding to any peripheral on the SoC) probably the best place to declare it is just in / (if this is an overlay then use &{/} instead of /) Feb 24 18:31:55 also the status = "okay"; is pointless nonsense... that's something you do to enable a device that has declared as disabled, i.e. to override a previous status = "disabled"; property in that node Feb 24 18:32:16 it serves no purpose when creating a new node Feb 24 18:35:01 lucascastro: you'll sometimes see people declaring virtual devices in &ocp ... that's simply wrong, &ocp is a flattened representation of the on-chip interconnects where all of the SoC's peripherals are located, it should not be used for devices that are not actually located on these interconnects Feb 24 18:35:09 anyway, afk Feb 24 21:42:04 bah, what is catching the pins? i keep getting pin PIN2 already requested by 44e10800.pinmux; cannot claim for ocp:P8_05_pinmux Feb 24 21:43:34 ooops wait, i think i forgot to disable universal cape Feb 24 21:47:25 hi all. im not using nelsons debian image, instead i am using an arch linux image. to make an backup of the emmc i am booting from an sd card and using dd to copy the hole content of the emmc to an external media. to flash a new beaglebone i simply booting up from sd card and using dd command to flash the emmc. after booting from the emmc Feb 24 21:47:26 everything seems okay. i can connect over rndis. but the wifi is disabled by rfkill. i have manualy to unblock the devie, then it works. in the source image (the image that i was backuping using dd) the wifi is not blocked. can anybody gibe me a hint what i am doing wrong ? Feb 24 23:56:16 Is anyone using machinekit or -hal emcapplication? Feb 24 23:56:52 I made a compilation example for making things available if you are using it... Feb 24 23:57:13 natively and not a cross-comp. but it works... Feb 24 23:57:56 I am asking to see if anyone is making a Cape for such a compilation. Oh! Feb 24 23:58:15 Does the Replicape work w/ the machinekit source and linuxcnc? Feb 25 00:13:03 ... Feb 25 00:13:18 GenTooMan: How yee? Feb 25 00:56:58 forget about my comp. My comp. is busted now b/c of emcapplication being something or other. I have not figured it out yet. Blah. **** ENDING LOGGING AT Thu Feb 25 02:59:57 2021