[09:40] i find these instructions a bit ambiguous https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1453391/comments/14 [09:40] Ubuntu bug 1453391 in linux (Ubuntu) "Very slow Qualcomm Atheros AR8121/AR8113/AR8114 Gigabit or Fast Ethernet (rev b0)" [Medium,Confirmed] [10:38] Hello [10:39] I filed a kernel bug and was asked to do a full commit bisect [10:39] but I can't follow the guide on the wiki [10:39] as I can't find the debian.master file where to make the changes [10:40] should I just adjust the version number in makefile instead? [12:05] ePirat: try ./debian.master/changelog [12:26] Taloth, there is no debian.master folder [12:26] http://kernel.ubuntu.com/git/ubuntu/ubuntu-trusty.git/tree/debian.master [12:27] which version were you supposed to bisect? [12:27] I am using the vivid repo [12:28] 3.18.6 to 3.19.x (most recent one I guess) [12:28] you already bisected the released versions (package)? [12:30] yes, those are the working/non-working versions I identified, I can't narrow down more as other 3.18 versions higher mess up everything making reproducing my bug impossible there [12:30] the issue is that wifi stops working after some time/data downloaded [12:31] even though it still shows up as connected it behaves like it is not [12:31] for testing I need to be able to run the system for at least an hour while downloading stuff [12:31] In the bugreport you mentioned 3.18.6 working while 3.18.11 doesn't. [12:32] so I assume you'll be commit bisecting from there on. [12:33] 3.18.11 shows different non working behavior than I have actually observed in 3.19 (i get popups that errors occured in the system but those seems to not really be related to wifi) [12:33] what's the next lower version you checked? [12:34] 3.18.0-10 or something else? [12:36] let me check (need to boot the affected machine) [12:38] 3.18.6-031806 was the last working version for me, 3.18.11-031811 was the one I could not test properly with [12:38] As for your original question, if you check out the repo from http://kernel.ubuntu.com/git/ubuntu/ubuntu-vivid.git the ./debian.master dir should be there. [12:39] Maybe you're bisecting the upstream kernel tree? which wouldn't have debian.master [12:40] I cloned from git://kernel.ubuntu.com/ubuntu/ubuntu-vivid.git [12:42] I do have a debian.master folder in master, but not in my branch which I am using for bisect (which is currently at 0ce3c06) [12:47] which branch are you bisecting on? seems to me like you should be using master. [12:48] I followed the guide linked in the bug ticket, so I branched from master in a branch called mybisect and started the bisect [12:54] which commit did you mark as good? [12:56] bad: 83f803a good: f8e80fa [12:58] but yeah, you're in the commit history of the upstream kernel, hence the missing debian.master. tbh, not sure how to normally deal with that. [13:00] oh [13:00] ok [13:01] but did you install 3.18.0-14.15 from package? [13:04] not sure… could you explain me how this versioning stuff works? I am confused by all the -x.x suffixes [13:06] that's because ubuntu is based on the mainline kernel, which uses different version numbers. [13:07] Your problem boils down to https://wiki.ubuntu.com/Kernel/KernelBisection#How_do_I_bisect_the_upstream_kernel.3F [13:08] ah [13:08] welcome down the rabbit hole btw. [13:08] :D [13:09] :P [13:09] I wish my bug would at least be easier to reproduce [13:11] i had a bug that made mono crash on users systems at unpredictable times. sometimes minutes sometimes days. Took quite some time till I isolated the kernel commit breaking it. [13:11] fortunately wasn't hardware related like your wifi issue [13:13] usually I would just have used an older versions but there my trackpoint click buttons do not work properly… [13:14] On the other hand, now you'll helping to make the world a better place :P [13:15] it seems switching kernels so often messed something up as I now get the system error popup even on previously fine kernel versions… great… [13:15] and I can't even report those as report fails with some error… [13:18] I wonder why no other thinkpad x250 user has had this wifi problem yet [13:18] they quite possibly did, and are angry atm but don't know how to fix it. [13:20] tbh, reporting the ubuntu kernel bug was one of the hardest bug reporting experiences I ever had [13:26] I wonder if it wouldn't be possible to narrow it down to a few commits, since it is some wifi related thing? [13:30] that's where a bisect comes in. [13:33] from what I could see the bisect is 13 steps, over 7000 commits. It's going to take a while. [13:35] indeed, specially since reproducing the bug takes hours sometimes [13:38] but don't give up, take your time, you'll be helping out the devs immensely. It will be near impossible for them to reproduce, much less fix. [13:40] ok :) [13:41] I think I will reinstall ubuntu later (specially since 15.04 is officially released now) and try to narrow it a bit more down with packaged kernel versions