=== shengyao_afk is now known as shengyao | ||
=== yofel_ is now known as yofel | ||
=== Guest61012 is now known as fmasi_afk | ||
=== cmagina_ is now known as cmagina | ||
=== kentb-out is now known as kentb | ||
=== gfrog_ is now known as gfrog | ||
=== fmasi_afk is now known as fmasi | ||
=== fmasi is now known as Guest1765 | ||
=== Guest1765 is now known as fmasi_wfh | ||
Sargun | Why is CONFIG_FUNCTION_TRACER set to y in the standard kernel config? | 15:28 |
---|---|---|
Sargun | (on AMD64 Raring's kernel config) | 15:28 |
smb | Sargun, Because it otherwise would not be possible to enable function tracing on a stock kernel (it is a bool and should be y for i386, too). If that is the answer you are looking for | 15:35 |
bjf | Sargun, i believe it is enabled so that ftrace can be used to debug kernel problems | 15:35 |
bjf | Sargun, i believe we benchmarked it and found it added little/no overhead if you were not in fact using ftrace | 15:36 |
apw | yeah i believe the same | 15:37 |
smb | right, according to help 5 NOPs at the entry of each function which should have not much impact | 15:38 |
Sargun | Hm | 15:41 |
Sargun | This is unfortunate. I really want to enable KMEMCHECK on my kernels, which conflicts with CONFIG_FUNCTION_TRACER (it requires it to be =n). | 15:42 |
Sargun | I guess is CONFIG_FUNCTION_TRACER doesn't get used during normal operation, this is fine | 15:45 |
apw | Sarvatt, if you are changing your config you can change both anyhow right ? | 15:46 |
bjf | Sargun, ^ | 15:50 |
apw | bah i hate auto complete ... it is meant to pick the person who spoke last, that is working, NOT | 15:52 |
Sargun | apw: Yeah, I was just wondering if CONFIG_FUNCTION_TRACER was used for anything other than debugging. | 15:56 |
apw | no indeed, just for debug, but ... you dont want to change your system to start using debugging tools when you need them | 15:57 |
cking | brendand, i've built a new version of fwts for testing on that ARM box - it's in ppa:colin-king/white | 16:25 |
brendand | cking, cool - i'll give it a try | 16:26 |
cking | great! | 16:26 |
* cking waits with baited breath | 16:53 | |
cking | brendand_, any results? | 17:22 |
brendand_ | cking, yeah, passed | 17:24 |
brendand_ | cking, how did you change the test? | 17:24 |
cking | brendand_, cool, I'll send the patches out for review then \o/ | 17:24 |
cking | brendand_, i re-wrote the support libs last week but forgot to tweak the test according to the new api, so it failed last week, today I fixed the test and it works as expected | 17:25 |
=== fmasi_wfh is now known as fmasi_afk | ||
hallyn | rtg: i'm sitting here on my saucy laptop, getting the same kvm modprobe error you were on trusty. has that been figured out yet? | 20:25 |
apw | hallyn, hey which kerle is that | 21:21 |
apw | hallyn, is that a home brew | 21:25 |
hallyn | no | 21:28 |
hallyn | 3.11.0-12-generic #19-Ubuntu | 21:28 |
hallyn | (cut-paste is being a b*, refusing to cut at words) | 21:29 |
apw | hallyn, i don't think i see that on my machine when it is on 3.11 kernels | 21:30 |
rtg | hallyn, hmm, kvm_amd is working for me. You have Intel, right ? | 21:31 |
apw | hallyn, can you try the kernle in -proposed -13 ? | 21:31 |
apw | hallyn, i have a kvm_intel (the one showing the issue originally) but i believe it works fine on the 3.11 kernles | 21:31 |
hallyn | just dget it from the build logs? | 21:31 |
hallyn | not the logs of course | 21:32 |
hallyn | about time to call eod judging by reading my own comments | 21:32 |
* hallyn tries https://launchpad.net/~canonical-kernel-team/+archive/ppa/+build/5146181/+files/linux-image-3.11.0-13-generic_3.11.0-13.20_amd64.deb | 21:34 | |
hallyn | will be awhile | 21:34 |
apw | hallyn, i have just booted 3.11-12.19 and 3.12-13.20 and both work fine, 3.12-0 does not | 21:43 |
hallyn | apw: (i've installed but not yet rebooted) have you figured out wha'ts up with 3.12-0? | 21:48 |
apw | hallyn, nope rtg is on the case, its not h/w specific, seems to be a kvm core issue | 21:49 |
hallyn | wowzer | 21:50 |
rtg | hallyn, its a debugfs issues, so nothing too critical. I should be able to at least pin it down soon. | 21:51 |
hallyn | rtg: oh right, i forgot that showed up in syslog | 22:02 |
hallyn | but holy cheeseballs! somehow kvm got disabled in bios. | 22:03 |
hallyn | <think><think><think> <lightbulb> that effing windows 7 install i did to try and update bios. I bet it disabled it! | 22:04 |
rtg | hallyn, bummer dude | 22:07 |
apw | hallyn, man ... m$ should really be taken out and ... | 22:07 |
hallyn | you must not assume malice on th... oh wait this isn't debian-devel | 22:18 |
hallyn | i dunno, do they want to then charge me to reenable it for me? | 22:18 |
rtg | hallyn, the 3.12 kvm issues appears to be genuine. I've instrumented the create debugfs dir path and should have some notions soon. | 22:25 |
Malediction | Could anyone tell me why, since the latest release upgrade, I have 64 rcuos/<number> and rcuob/<number> processes on a single core VM? (or better yet, tell me what docs/man I can read to better understand them) | 23:29 |
bjf | Malediction, http://lwn.net/Articles/518953/ | 23:51 |
Malediction | Ah, thank you. :) | 23:54 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!