**** BEGIN LOGGING AT Sat May 08 02:59:57 2021 May 08 03:15:28 zeddii: thank you for that. May 08 06:15:53 Hi, how should I go about debugging why a Kconfig option isn't ending up enabled in my kernel? May 08 06:16:21 Ones adjacent to it in the config are being enabled just fine May 08 06:27:45 Hmm. Seems like sccs aren't updated, since trying to disable the option that was working leaves it still enabled on a new kernel May 08 06:30:37 Aha. .kernel-meta/configs hangs around and is a cached version from the devtool modify May 08 06:52:27 oe-local-files seems to be the culprit. Mirrored my modifications there and it compiles in. May 08 08:12:20 @dp: I assume you use yocto kernel tooling (scc/cfg and friends) May 08 08:13:29 @dp: try this here (for now): https://yoctoproject.blogspot.com/2020/08/yocto-kernel-tooling.html May 08 08:14:33 RobertBerger, thank you - it looks like my knowledge of non-kernel packages doesn't carry over so well to the kernel since it is a special case May 08 08:14:36 I'll have a read May 08 08:36:32 The `merge_config_build.log` shows that my cfg option is not being set May 08 08:37:12 Could this be because I am patching this Kconfig value in? I've backported CONFIG_GPIO_REGMAP to an older kernel May 08 08:41:33 It is ending up in mismatch.cfg, but not invalid.cfg, not sure what that indicates yet though May 08 08:49:10 @dp: did you create config fragments? May 08 08:49:39 I did yeah. There is a config fragment living in my layer, included in SRC_URI for my linux-yocto bbappend which contains CONFIG_GPIO_REGMAP=y among some other misc tweaks May 08 08:49:49 @dp: I suggest you do something like bitbake virtual/kernel -c menuconfig May 08 08:50:12 @dp: and create a config fragment: bitbake virtual/kernel -c diffconfig May 08 08:50:21 Hmm I'll give that a go May 08 08:50:43 If you manage to set the kernel config in menuconfig your chances might be better May 08 08:50:58 it could be just some dependency in the kernel you are missing May 08 08:51:05 Yeah. There don't seem to be any off prerequisites outside of GPIOLIB, but could be I am missing it still May 08 08:52:40 Just to understand - you backported CONFIG_GPIO_REGMAP to an older kernel and it does not show up in menuconfig. Is this the problem? May 08 08:53:03 Yes - just realised it doesn't appear in mainline kernel either. Presumably only selected by a specific driver May 08 08:54:02 a quick googling says it's not configurable: https://www.kernelconfig.io/config_gpio_regmap May 08 08:54:30 Yeah, it's selected by another driver (the only in-tree to use it) it would seem May 08 08:54:50 GPIO_SL28CPLD selects it May 08 08:55:24 Yep and CONFIG_GPIOLIB May 08 08:55:33 and CONFIG_REGMAP May 08 08:55:50 which depend again on other things May 08 08:56:04 I have regmap and gpiolib passing configuration in Yocto May 08 08:56:39 But should CONFIG_GPIO_REGMAP not be configurable for use in out-of-tree modules? Just not from menuconfig perhaps. May 08 09:23:39 Temporarily hacked it to be so, until I get my driver building in-tree. Seems to work May 08 09:33:49 @dp: wait a bit May 08 09:34:00 if it's out of tree it's a different story May 08 09:34:36 The upstream version is definitly not out of tree May 08 09:34:53 But May 08 09:35:21 If it's out of tree it can not be "y" but must be "m" May 08 09:36:10 I mean that my module using the gpio_regmap library is out of tree May 08 09:36:39 so try "m" May 08 09:36:48 or just load it ;) May 08 09:37:10 if it's out of tree it will not show up in menuconfig May 08 09:37:34 since you need in tree, kconfig and kbuild for it to show up in menuconfig May 08 09:38:08 Oh my module is building as a separate yocto recipe, I was only having trouble getting CONFIG_GPIO_REGMAP enabled for it May 08 09:38:17 Ah OK May 08 09:38:36 So independent of your module you want to turn on CONFIG_GPIO_REGMAP May 08 09:38:52 Yeah May 08 09:38:58 than it's as I said further above - try to turn it on with menuconfig May 08 09:39:08 create a cfg fragment May 08 09:39:21 add the fragment to the kernel recipe May 08 09:39:29 Yes, it was not visible in menuconfig because it's non-configurable/hidden, so I just made it configurable, and now my old config segment works :) May 08 09:39:31 and try menuconfig again - and it should be enabled ;) May 08 09:39:54 OK cool! May 08 09:40:01 A bit hacky, but this is just to get a demo kernel module going, and out-of-tree was the quicker way to iterate rather than rebuilding whole kernel May 08 09:40:10 sure May 08 09:40:13 Thanks for your help, I didn't know Kconfig entries could be invisible! May 08 09:40:31 I typically do all this kernel hacking just with the SDK May 08 09:40:43 once I get it to work I add it to the recipes May 08 09:40:47 You're welcome May 08 09:41:08 BTW you also might want to have a look at yocto dev day lectures and videos May 08 09:41:18 There are workflows with kernel hacking and devtool May 08 10:43:05 kanavin: I'm adding a patch to make smp arm/x86 to -next for testing May 08 10:52:28 hmm, gitlab server database upgrades has taken out the autobuilder for a repo I don't care about for this build :( May 08 10:52:37 * RP looks at rburton and jonmason May 08 10:53:04 (meta-kernel which we presumably still need for older releases of meta-arm) May 08 11:05:09 * RP realises there is a workaround May 08 11:21:06 RP: would it still be possible to override is 4 is not enough for some? May 08 11:21:40 moto-timo, I'll be working for linutronix from August May 08 11:27:32 RP: e.g. we use 8 :) May 08 11:30:44 kanavin: you'd just set QB_SMP? May 08 11:31:36 RP: yeah, but wouldn't QB_SMP=4 from qemuboot-x86.inc take precendence? Admittedly I am not sure what would be the final value then. May 08 11:32:47 kanavin: I think it would depend how and where but the values are all intended to be overridable May 08 11:33:43 rburton: another edk2 reproducibility issue: http://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20210508-wdqu25lh/packages/diff-html/:( May 08 11:34:19 RP: distro config I guess, if you think that'll work, then I'm fine :) May 08 11:36:25 hey i am using busybox. on insertion of eth table i am not able to get ip automatically . i used udhcpc -i eth0 command to get ip. how to make my yocto device to get auto ip? May 08 11:37:38 table=cable* May 08 11:38:42 kanavin: I'm fairly sure it should. If not we can revise based on a problem May 08 11:39:34 2nd is when i am using pip3 to install any package i am getting ssl error. i am on dunfell build of yocto. pip is 20.x , even in python reqquests i have to put verify=False. how to solve the certs issue? May 08 11:52:44 drofty: I'd imagine you're missing udev or equivalent to handle the device config May 08 11:53:30 drofty: other issue could be missing ssl lib or ca-certs ? May 08 11:58:00 after installing udev i have to edit config too for eth0, right? May 08 12:05:21 drofty: I'd have thought it would be setup for autoconfigure by default but I don't remember May 08 12:05:39 there are no ca-certs in layers.openembedded. is there is any way to add auto certs? May 08 12:10:14 drofty: sorry, ca-certificates May 08 12:11:11 thank you i will include this in my builds May 08 15:20:07 Hi May 08 15:20:10 Team May 08 15:20:17 I need to add a library May 08 15:20:23 in yacto May 08 15:20:28 please help May 08 15:38:49 kanavin: https://autobuilder.yoctoproject.org/typhoon/#/builders/44/builds/3435/steps/11/logs/stdio - small tweak needed :) May 08 17:09:23 i am not able to understand that how to install udev. i want to get automatically ip when eth cable is connected. on google it showing use systemd. currently my build is using udhcpc of busybox i have to manually type commaand to get auto ip. can anyone suggest easy way to do it and i want it be small. my device is weak May 08 17:10:12 or i have to swtich to systemd only? busybox kinda looks good for me. all things are tiny in size May 08 17:49:40 kanavin: testing one in -next May 08 18:04:39 I saw that libstatgrab is there in meta-perl but in my project iam not able to see libstatgrab in meta-perl can I know how to add May 08 18:45:03 kanavin: congratulations! May 08 18:47:26  Could not inherit file classes/ptest-perl.bbclass May 08 18:47:37 Iam getting this erro **** ENDING LOGGING AT Sun May 09 03:00:50 2021