**** BEGIN LOGGING AT Fri May 26 03:00:04 2017 May 26 05:40:22 hi, anyone knows what's the point of specifying a patch file and not apply it? e.g. file://foo.patch;apply=no May 26 07:43:17 After building an image, is there a package manifest located somewhere? May 26 07:53:29 nevermind. my eyes are not what the once were :p May 26 14:13:32 is there a way to generate a kernel with embedded initramfs (ie. fitImage with three entries, kernel initramfs dtb) and have that on UBIFS volume in UBI , without much hackery , ideally using bitbake foo-image ? May 26 14:14:09 I know I can have initramfs bundled with the kernel fitImage , but if I set FS_IMAGETYPES , it produces a ubi with kernel without initramfs and the UBI also contains the content of the initramfs May 26 14:14:13 that's not what I want May 26 14:39:18 Marex: i think you need to change kernel-fitimage to use the .ubi image instead of the initramfs *.cpio* targets May 26 14:40:39 nrossi: then the ubi would be embedded into the fitImage , no ? May 26 14:40:50 nrossi: I want the fitImage embedded into the ubi May 26 14:42:20 Marex: oh right sorry, thats a bit different. Any reason you can't put the kernel-image in your initramfs? May 26 14:42:31 Marex: sorry your UBI* May 26 14:43:00 (kernel-image being the package) May 26 14:46:47 My god, if Xilinx and Altera fan boys can have a civilized conversation, there is hope for world peace! May 26 14:46:50 :) May 26 14:48:02 Crofton|work: who said i was a Xilinx fanboy? :P May 26 14:48:10 well May 26 14:48:17 I can make stuff up :) May 26 14:49:01 hmm, 2 Jefros are better than one May 26 14:49:11 Crofton|work: since your here, did you get a chance to look at the mailing list user? May 26 14:49:29 no, been hoping to get some indication in my inbox May 26 14:49:36 happens when you send an email? May 26 14:49:46 hm i wish i could remember my windows bitlocker password May 26 14:49:58 I wish I could just remember May 26 14:50:07 that might be my problem May 26 14:50:42 Crofton|work: Every time :|, which is quite a few in the past day or two. It seems the email server ignores the Errors-To and any list info and just returns the mail delivery error to the original sender May 26 14:51:33 have you tried sending a direct email? May 26 14:51:52 Crofton|work: nope will give it a try now May 26 14:52:07 if that fails, send it to me and I can remove the guy May 26 14:59:03 Crofton|work: Unzustellbar! Failed with the direct email, address for the user is mir@lynx-technik.com May 26 14:59:22 OK May 26 15:01:37 nrossi: you mean an image which just contains the fitImage ? May 26 15:01:43 nrossi: I mean, the kernel image ? May 26 15:02:00 nrossi: how could that work if the image is what's bundled with the kernel ? there's a bit of a chicken-and-egg problem here May 26 15:02:53 nrossi, done May 26 15:03:00 Crofton|work: thanks May 26 15:03:01 thanks for help checking May 26 15:03:21 Marex: two images no, just like a normal 'initramfs' setup May 26 15:04:37 Marex: and one image literally just contains the kernel-image-fitimage package. May 26 15:04:56 Marex: for the full, 'image in and image' experience :) May 26 15:16:37 nrossi: what's "kernel-image-fitimage" ? I can't find that package :) May 26 15:17:35 Marex: its dynamic, generated by kernel.bbclass from current image types, see here -> http://git.openembedded.org/openembedded-core/tree/meta/classes/kernel.bbclass#n54 May 26 15:20:50 nrossi: ah, I'll try that May 26 15:58:49 nrossi: there seem to be no actual package with the kernel containing the initramfs, so an image cannot bundle it into itself May 26 16:06:43 Marex: ah yes... it appears no "install" of the .initramfs (or FILES values for that matter). :| you will need to modify kernel.bbclass or make your own image type May 26 16:09:02 damn ! May 26 16:09:27 nrossi: I hoped to avoid touching this awful dependency mess in kernel.bbclass May 26 16:11:18 Marex: now that i think about it, someone submitted some patches for this initramfs bundling... https://patchwork.openembedded.org/patch/138964/ May 26 16:12:05 Marex: just looking at the recipe, you could just copy the .initramfs from deploydir and populate it as a "kernel-initramfs" recipe into your UBI image May 26 16:14:38 oh interestting May 26 17:00:47 nrossi: creating empty image is not that easy, huh May 26 17:00:52 nrossi: I mean, completely empty May 26 17:04:00 Marex: can be a pain, post process commands can make it really easy if you don't mind wasteful dependences May 26 17:04:46 nrossi: postprocess with "zap everything and copy over the correct fitImage" ? May 26 17:05:00 Marex: yep :) May 26 17:06:20 nrossi: that's executed before ubinize ? May 26 17:07:43 Marex: not sure, but if ubinize is part of the image fstype then should be May 26 17:09:26 but it's also run before generation of the CPIO image ? May 26 17:09:33 (for the initramfs) May 26 17:09:40 I'd need to somehow determine when is the right time to nuke everything May 26 17:25:20 Marex: the image commands should run independently for each image. As long as you set them in the image recipe or equivalent **** ENDING LOGGING AT Sat May 27 03:00:01 2017