[02:24] Hi, am trying to build a custom kernel (4.4.20).. for kdump, ie, a crash kernel [02:25] Trouble is, am facing all kinds of issues when the kexec'd kernel boots - mainly to do with config and kernel cmdline params to pass [02:26] Is there a specific kernel config file available for the Ubuntu crash kernel image? [02:28] specifically, for Ubuntu 16.04 (xenial) [02:56] hey all, can someone refer me to a guide on the current, supported way of customizing a kernel on ubuntu? (xenial, but looking to get 4.10 kernel) preferably from a kernel-source package versus a git checkout? [03:24] well it looks like actually there isn't a source package beyond 4.4 so... git it is [09:31] ppisati: Hi [09:31] ppisati: ogra_ asked me to ping you [09:31] ppisati: Could you tell me all the kernel config options required for Snappy please? [09:31] ppisati: Is there a Wiki? [09:32] i think there was a script actually [09:33] lag, it could even be that these scripts are integrated into the snapcraft plugin ... https://github.com/snapcore/snapcraft/blob/master/demos/96boards-kernel/snapcraft.yaml has a demo snapcraft.yaml [09:34] (but i'm, not actually sure that checks for all config options, so better wait for ppisati's answer :) ) [09:34] ogra_: Thanks [09:46] lag: or here - https://github.com/snapcore/sample-kernels [09:46] lag: look under kernel/config/snaps [09:47] *kernel/config/snappy [09:52] ppisati, heh, you should really change CONFIG_SQUASHFS=m there [09:52] so people are not forced to put that into the initrd [10:03] ogra_: yeah, i need to update the configs there [10:03] ogra_: probably next week i'll push an update to the AA stuff, and i'll bundle the config chanfges [10:03] +1 [10:08] ppisati: Thanks [12:52] Hello, I need to do a bissect on the 4.8 mainline kernel [12:52] what is the quickest way to only rebuild the vmlinuz bisected file w/o all the modules ? [12:53] mainline or our own kernel for that matter so either one is fine [12:54] caribou, CONFIG_MODULES=n ? [12:55] there is no split in the build between built-in and modules. its just the one call to the kernel makefile [12:55] rtg: cool, thanks [12:56] or, you could do a native build after generating the config, then 'make vmlinuz' [12:56] caribou, I often do something like this: [12:56] fdr clean prepare-generic [12:56] cp debian/build/build-generic/.config . [12:56] make oldconfig scripts prepare [12:57] make vmlinuz [12:58] rtg: great; will try that [12:58] rtg: something has changed in the /proc/vmcore elf representation which breaks makedumpfile (or at least make it spit tons of errors) [14:24] Hi, linux-image-generic-lts-yakkety fails ot install on xenial, which I believe is due to its absence from linux-lts-yakkety - 4.8.0-25.27~16.04.1 in https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/ppa/+packages [14:24] ^ apw / rtg I've been told that you were the right guys to ping about it [14:26] sajoupa, how does it fail to install ? missing dependencies ? [14:26] rtg: yes [14:26] sajoupa, apw pointed out that I might have uploaded the meta package to the wrong pocket. [14:26] lemme check [14:26] https://pastebin.canonical.com/168001/ [14:27] indeed I did [14:31] sajoupa, I've uploaded a new version. Give it an hour or so to publish. [14:31] rtg: cool, thanks ! [15:07] rtg: installing fine now [15:08] sajoupa, ack [15:33] Hello there. Sorry if this is the wrong channel for "user questions", but I installed a 32-bit 4.8.0 daily kernel (tested with the 11 October and the 13 October one actually) and it hangs after grub. [15:33] Not a single line from the kernel printed, which makes me think that it is not properly loaded. The 4.8.0 non-git one boots fine. Does that bring anything to mind? Including problems between keyboard and chair... [15:34] Just to be clear, 4.8.0 works, 4.8.0-999 does not. [15:40] michael-vb, we tend to have a more issues with the debug kernels for sure they are built in an automated way [15:42] Right. I was asking here because it looks like the kernel possibly isn't loaded at all, in which case there might be some easy way to solve it (for me). [15:43] though the -999s are dalies if i remember the monenclature correctly, those have whatever random crack is in the aossicated tree that day [15:43] Of course, I am certainly not shocked, shocked to find one not working. [15:44] Or even two. [15:44] depending on the day that could even be the middle of the merge window [15:46] But the description of a hang after grub but before the first kernel message is printed does not bring anything to mind? [15:48] Actually it does a mode set after selecting the kernel in grub and it hangs on an underscore cursor in the top-left corner. Does grub trigger that, or would it be the kernel at least starting to boot? [15:51] michael-vb, when there is a failed boot grub the next boot hands off with the framebuffer blank with just a cursor [15:52] so that could easily be cratering in the first bits of the kernel [15:53] Thanks. [16:30] query on the overlay stuff in y -- given that overlayfs is no longer recognized, should there be something that goes into /etc/schroot/chroot.d/ and post-install of an appropriate kernel ensures that it says overlay for each schroot? I guess that could be a sbuild change. Wasn't sure if there was already a bug for this === mamarley_ is now known as mamarley === nacc_ is now known as nacc === teward- is now known as teward === xnox_ is now known as xnox === mterry_ is now known as mterry === om26er_ is now known as om26er === Trevinho_ is now known as Trevinho === kees_ is now known as kees === _fortis_ is now known as _fortis === cjwatson_ is now known as cjwatson