/srv/irclogs.ubuntu.com/2013/11/06/#ubuntu-kernel.txt

keesany reason we build the ARM kernels with CONFIG_OABI_COMPAT?00:13
bjfkees, can't think of any reason. we don't notmally turn on anything with EXPERIMENTAL in the description. rtg would be the one to ask however00:16
bjfkees, file a bug?00:17
bjfkees, or mail to the mailing list00:17
keesbjf: okay, cool00:18
lebcan anyone help with this00:58
lebhttps://gist.github.com/tildeleb/732900500:58
infinityleb: Not really a kernel question, but you want sys/inotify.h, not linux/inotify.h... You should never include linux/ headers from userspace.01:06
lebI know it's no a kernel question01:07
lebsorry about that01:07
lebbut other channels didn't have much of a clue01:07
netrunner__hello everybody01:07
lebtrying to compile an older inotifyd so this is not my code01:08
netrunner__i need support due to a fail at the new kernels01:08
netrunner__http://imagebin.org/27589701:09
netrunner__http://imagebin.org/27590001:09
netrunner__output of dmesg:01:09
netrunner__http://paste.ubuntu.com/6367846/01:09
netrunner__does anyone have a idea?01:10
netrunner__some more information:01:10
netrunner__the 3.2.1-52-generic-pae kernel is the last one who works01:10
lebinfinity thanks01:10
netrunner__the newer one dont work01:11
netrunner__cant boot01:11
lebchanging from linix/ to sys/ solved the issue01:11
netrunner__how i do that??01:11
netrunner__i am not an expert01:11
lebmuchas gracias01:11
netrunner__jep, thanks in advance for help :-)01:12
netrunner__or is this message for another guy?01:12
netrunner__anybody here who can help me?01:13
bjfnetrunner__, what i suggest is you boot the working kernel and file a bug01:31
bjfnetrunner__, what timezone are you in?01:31
bjfnetrunner__, once you file the bug and we take a look at it you are going to get asked to test some kernels01:33
netrunner__MEZ01:33
netrunner__ok, i have already boot the working kernel01:34
netrunner__how can i file a bug01:34
bjfkinda late for you :-) (or early depending on how you look at it)01:34
netrunner__?01:34
bjfnetrunner__, from a terminal type: "ubuntu-bug linux"01:35
netrunner__it is not possible because i have a kubuntu ditro01:38
netrunner__distro01:38
bjfnetrunner__, https://bugs.launchpad.net/ubuntu/+source/linux/+filebug01:39
netrunner__http://paste.ubuntu.com/6368039/01:50
netrunner__that is the output of "ubuntu-bug linux01:51
netrunner__i have netrunner - a kubuntu distro01:51
netrunner__should i really register at: https://bugs.launchpad.net/ubuntu/+source/linux/+filebug?01:51
bjfnetrunner__, yes, we need the bug filed01:53
netrunner__ok i will sign up01:53
netrunner__ok, how can i file this bug now?02:10
bjfnetrunner__, got to the url i posted ^02:10
netrunner__jep, there is a wizard, but i dont know the cause of the problem02:13
netrunner__and i dont know the package the wizard is asking for02:13
bjfnetrunner__, the package is already set to "linux" just fill out the summary and further information (what the probelm is and how to reproduce) then "submit bug report" at the bottom02:14
=== bjf is now known as bjf[afk]
netrunner__Can i also add pictures to the bug report?02:20
=== henrix_ is now known as henrix
=== lifeless_ is now known as lifeless
=== fmasi_afk is now known as fmasi
=== fmasi is now known as fmasi_afk
=== fmasi_afk is now known as fmasi
* apw yawns09:28
=== fmasi is now known as fmasi_afk
=== fmasi_afk is now known as fmasi
reiseihi, all! I'm compiling kernel for omap4, but got the error: objdump: /lib/libc.so.6: File format not recognized10:47
reiseidpkg-shlibdeps: error: objdump gave error exit status 1 How can I fix it?10:47
reiseiAny advice will be appreciated11:34
ppisatirobher: didn't you leave some debug printk around in your mac address learning patch?11:54
ppisatirobher: printk("unlearned addr %pM\n", addr);11:54
ppisatirobher: printk("learned addr %pM\n", src_addr);11:54
apwreisei, compiling in what environment, that sounds like you are running a forign binary12:04
reiseiapw: already found that was some kind of ASCII file. Just copy the library itself and it work.12:05
reiseiapw: arm-linux-gnueabi- lost some of the files...12:12
ppisatirobher: and one last thing that it's a bit counterintuitive, why do you learn the src_addr on the xmit path? since we are on the way out, shouldn't we learn the dst_addr or maybe the src_addr on receive?12:19
* ppisati should put some printk and check by himself...12:19
apwhttps://wiki.ubuntu.com/ARM64/FoundationModel12:30
athirai am new to contributing to ubuntu launchpad.i reported a bug in secure-delete package and i have attached and submitted the patch.i didnt set a reviewer for my patch.can anyone tell me how to set a reviewer?12:55
robherppisati: yes, those printks should be removed. We need to learn the src addr so that received pkts with that dst addr are not dropped. We'll never get the pkt on the receive side if the addr is not known.12:56
apwathira, you would normally subscribe the ubuntu-sru team if it is for a stable release, or for devel just ask the patch-pilot on #ubuntu-devel13:37
=== shadeslayer_ is now known as shadeslayer
apwhallyn_, hey .. we have reports of networking issues inside lxc containers on 3.12 kernels15:24
apwhallyn_, mean anything to you ?15:24
hallyn_apw: well there's the known offloading of checksums for veth as of 3.11 or 3.10...15:24
hallyn_ubuntu software as of saucy should be updatd to handle it15:25
hallyn_but what sort of issues are yous eeing?15:25
smbapw, hallyn_ maybe we should find out whether the setup has openvswitch for the container involved15:27
ogasawaraah, wasn't that one of the busted dkms packages for 3.12?15:29
smbone of the complicated cases where features moved to the in-kernel module and caused a lot of head scrating as of how to cope with the dkms module (especially in P userspace)15:30
smb*scratching15:30
* hallyn_ is surprised at the # issues in 3.1215:31
ppisatiback in 2015:45
smbapw, Hm, just my few cents here... maybe you are already further in looking but from the console logs its hard to say whether network went away completely or "just" some fail in the resolver15:50
apwsmb, i think the implecation of the further testing is that it works outside the lxc and not inside15:53
apwhallyn_, so in this case we have a report of an lxc container where the 'host' can talk to archive.u.c, and the lxc container processes cannot15:55
apwhallyn_, this is from a test case which works with a downgraded kernel15:55
apwhallyn_, do we know of any configuration changes between 3.11 and 3.12 in that area, i assume more is available there namespace wise15:56
hallyn_apw: i can't think of anything...15:57
hallyn_apw: is there an lp bug for this?15:57
apwhallyn_, not as yet16:00
apwhallyn_, is this something i can easily (or you) test in isolation 16:01
apwthat network isolation is working on 3.12, do we have any tests ?16:01
hallyn_apw: well lp:~serge-hallyn/+junk/lxc-tests has some.  but i've been experimeinting with containers on 3.12 for the last few days with no issue16:03
hallyn_haven't upgraded kernel in 2 days probably16:03
apwhallyn_, ok that is a good data point, what kernel would you have tested ours or something of your own16:03
apwbug #124859016:03
ubot2Launchpad bug 1248590 in linux (Ubuntu) "on qa-radeon-7750, containers lose network connectivity" [Critical,Confirmed] https://launchpad.net/bugs/124859016:03
apwhallyn_, thats the one16:03
hallyn_apw: I'm on 3.12.0-1-generic #3-Ubuntu SMP Tue Oct 29 18:41:32 right now16:04
apwhallyn_, ok thats the archive one ...16:05
hallyn_apw: alas the dmesg in that bug is from a run where they haven't tried a container16:06
apwhallyn_, they say it is an autopilot test, and it goes wrong in the 'update yourself' phase thereof16:08
hallyn_apw: ok, but - the containers, are they newly created each time, or are they using stale ones?16:10
hallyn_if they are using older contaienrs (or debian) then they can get the old dnsmasq which won't work with newer kernel16:10
apwhallyn_, do you know the otto tool at all, its that one16:10
hallyn_i don't16:11
hallyn_also, 3.11 is not old enough to make a difference with that16:11
hallyn_i'm readig the linked incident log right now16:12
=== fmasi is now known as Guest16174
=== Guest16174 is now known as fmasi_wfh
=== Mikee_C is now known as Mikee_C_afk
keeswhee, it only took me 5 months to get this backported and tested, but I've sent a Precise ARM seccomp-bpf patch to the list now.20:57
hallyn_apw: good news, reversing the CLONE_PARENT restriction which broke lxc-attach seems to be acceptable upstraem.23:11
* hallyn_ out for the day23:16

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!