IRC log of #cubox of Tue 13 Mar 2018. All times are in CET < Back to index

10:43 Ke> let's see if mcbin survives 8TiB btrfs raid5 scrub without crashing
10:44 Ke> if, then it might be just that my pcie sata controller does not like my SSD
14:07 jnettlet[m] 14:07 * jnettlet[m] is waiting patiently to see if it survives
14:10 Ke> 43 GiB later crashed
14:10 Ke> WARNING: CPU: 1 PID: 0 at /build/linux-KWzmYg/linux-4.15.4/kernel/rcu/tree.c:2792 rcu_process_callbacks+0x530/0x560
14:11 Ke> there is also another rcu crash here
14:11 Ke> but definitely problems are with rcu
14:14 jnettlet[m]> that could be memory related...or it could be thermal related
14:14 jnettlet[m]> you are running at 2Ghz ?
14:14 Ke> yes
14:15 Ke> not sure why people seem to think rcu problems are hw related
14:17 Ke> but setting 1.3GHz should be easy enough
14:18 Ke> I'll try it today
14:20 suihkulokki> Ke: does the error look like bugs.debian.org/891467
14:21 Ke> wow, how did you do it
14:21 Ke> you god
14:21 Ke> I actually like did search engine
14:21 Ke> suihkulokki: why you no fix already
14:22 Ke> suihkulokki: anyway, since debian maintainers hate me, and kernel-package no longer works, what god should I believe in, last time I saw suggestions for replacements, they were not too constructive
14:23 suihkulokki> ke :) all I did was put tree.c:2792 rcu_process_callbacks into google
14:23 Ke> I guess I could just get the src pkg and do patch apply and build
14:24 suihkulokki> I guess google thinks I visit debian.org more often than you, and gave the link to me
14:24 Ke> well I don't hate rms and use duckduckgo, I guess that's a fatal error
14:25 Ke> (to be honest, I kind of dislike rms for spreading notion that embedded peripheral fw is better than host loadable)
14:25 Ke> suihkulokki: for the kernel-package problem?
14:26 suihkulokki> Ke: upstream kernel sources have "make bindeb-pkg" target
14:27 Ke> thanks, I'll try that out
14:28 suihkulokki> so git clone kernel.org .. zcat /boot/config... > .config; make oldconfig; make -jX bindeb-pkg
14:28 Ke> yes yes
14:29 suihkulokki> for debian packages apt-get source linux; cd linux-*/; sudo apt-get build-dep -y .; dpkg-buildpackage -rfakeroot -B -jX
14:30 suihkulokki> end result should be pretty much the same - a deb you can install and which you can then select from grub
14:31 Ke> yup
14:31 Ke> suihkulokki: is there any hope of that patch making it to the debian kernel?
14:32 Ke> I can test it soon, to verify that it also fixes my problems
14:32 suihkulokki> Ke: It's CC:d stable so yes
14:37 suihkulokki> Ke: you'll probably want to build linux-stable 4.15.y .. the patch might be there already
15:17 Ke> hmm, perhaps, I'll check
15:17 Ke> if so, I might just pick up kernel from experimental or so
16:41 Ke> now I have a build-vm
16:41 Ke> let's see who wins, the bug or the compile
16:53 Ke> suihkulokki: is in linus's tree but not in stables
16:56 Ke> was added yesterday
16:56 Ke> but not released
16:58 Ke> 4.15.10 patches are being listed right this minute
17:07 Ke> is in next 4.14 kernel
17:07 Ke> I guess it's in 4.15.10 also then
18:17 Ke> yeah, it's in 4.15.10
18:18 Ke> I guess I will wait until thursday
19:16 Ke> ok, now finally compiling
19:17 Ke> huge effort ot find a solution for a problem that would have been fixed by just waiting
19:17 Ke> why am I only not lazy when it does not count
19:20 Ke> suihkulokki: you omitted the part, where you disable the system certificates for signarute checking or install appropriate debian certs, I did the former, which was to disable REGDB signature checking and system wide certs
22:24 Ke> this is getting a bit xkcd success, the wifi broke down trying to boot the newer kernel, which did not even print anything to serial
22:24 Ke> I think I'll just wait for the debian packaged kernel
22:28 Ke> not my best day
22:28 Ke> now the wifi works again after me fixing nothing
22:29 Ke> because, why wouldn't it get fixed the same way it broke
22:31 Ke> kernel not working at all despite being very closely related to the debian kernel sounds a bit like a toolchain problem