/srv/irclogs.ubuntu.com/2013/05/09/#ubuntu-kernel.txt

* rtg_ -> EOD00:04
=== optimusprimem is now known as optimus-afk
infinityapw: Bah.  Headers for manta and mako didn't get flavourized, so they're sharing package namespace.05:58
infinityapw: Care to fix, lowlatency/ppc style?05:59
infinityapw: (Also true for grouper, it's just lucky that it has no other 3.1.10 kernels to compete with for namespace)06:00
* apw moans about spinning globes08:45
* RAOF pours apw a steaming hot cup of bees to cheer him up.09:11
apwRAOF, very thoughtful, interestingly they have been in my house the last couple of days, i should have thought of drinking them09:12
RAOFThey are gloriously spicy!09:12
apwdid you avoid the ubu-flu this time?  seems the .eu contingent has been struck down09:14
zequenceapw: All lowlatencies prepared, including raring (first time doing that one, but the only difference I saw was that the changelog said raring instead of raring-proposed after doing the scripted update, which I suppose is meant to happen).10:39
apwyep, there is no need for -proposed in any release now, the archive rewrites it for us10:40
* ppisati -> (long) lunch break10:53
=== akashj87_ is now known as akashj87
ignatenkobrainhi all !11:31
* henrix -> late lunch13:29
=== kentb-out is now known as kentb
sconklinapw: status on the cve tracker?14:28
apwsconklin, should be all live and well now14:30
apwsconklin, it now has the new HWE bits in it etc, and seems to be updating right now14:30
apws/right/correctly14:30
sconklinapw: ack, I'll run the merge in a bit14:31
apwrtg_, arn't we meant to select the AP with the best signal if there are more than one with the same SSID ?14:40
rtg_apw, seems like that was the algo that was described to me14:41
rtg_modulo supported bit rate14:41
rtg_max bit rate14:41
apwrtg_, i have sent you my channel dump, it is selecting cell 01 over cell 0614:44
apwwh14:44
apwwhich as far as i can see if about as wrong as it could be14:45
rtg_apw, max bit rate is 9 v.s. 614:45
rtg_doh! nm14:45
rtg_apw, 01 has a lot higher signal strength14:46
apwrtg_, how so14:46
apwthe signal levels are in negative units14:46
apwthe signal levels are in negative units14:47
rtg_-75 is higher then -53. get cking to explain it.14:47
rtg_or sconklin14:47
apwit is?14:47
apware you sure14:47
rtg_('cause I can't remember)14:47
apwrtg_, this here wifi analyser has bars which are better the higher they are, and the scale is 0dBm at the top and -100dBm at the bottom14:48
sconklingenerally, signal strength is not inverted from the measurements, although it is common for the measurements to be negative if they are less than some standard. Lower is weaker.14:48
apwrtg_, plus the cell 01 AP is really far away, and 06 is really near14:48
apwsconklin, so -100 is worse than -50 yes ?14:49
rtg_ok, perhaps I am misremembering14:49
apw(these are wifi cell stengths)14:49
* apw contends it is doing something dumb14:50
rtg_apw, mathieu trudeau was messing with the selection algo. perhaps you should consult him.14:50
apwrtg_, is that a userspace thing, or a kernel thing14:51
rtg_apw, supplicant IIRC14:51
apwok14:51
apwi hate networking, it is soooo difficult to get right14:52
apwhmm it just changed over, now that is odd14:57
henrixsmb: hi! re your "ACK w/whine" :)15:01
smbhenrix, yeees? :)15:01
henrixsmb: i send a single patch when git am succeeds without any output15:02
henrixsmb: whenever there's some warn/... i send different patches for each serie15:02
henrixsmb: even if git am do the right thing15:02
smbah ok. probably makes sense then15:03
henrixsmb: i do a cherry-pick, a format-patch and check to wich series it applies cleanly15:03
smbhenrix, I guess I should be looking a bit more carefully at the surroundings if there is multiple "cherry-picks". did not spot anything on those two15:06
henrixsmb: there should be... let me check again15:07
smbhenrix, not sure it matters. Looks like they get applied as we speak15:08
henrixsmb: yeah, they are slightly different. maybe i could just request a cherry-pick for all the relevant series instead, when they are clean cherry-picks15:09
henrixsmb: instead of sending 3 different patches. at least, there will be less noise in the ml15:09
henrix(says the guy that keeps spamming the list with 3.5.y mails :) )15:09
smbheh :)15:09
smbhenrix, Maybe it is only me but for now when I see cherry-pick in the patch I assume it is the same "quality" (iow format-patch taken as git am)15:12
henrixsmb: so, what would be your suggestion?  should i replace the 'cherry picked from...' by 'back ported from...' when the patch doesn't match the original commit?15:14
henrixi'm ok with that15:15
smbI probably would do that, yeah. And I would use "cherry-picked" and "backported" because I am anal... :-P15:15
henrixheh, ok. i'll do that then. thanks15:17
rtg_ogasawara, do you have the HW to repro the bug in bug #1178191 ?15:52
ubot2`Launchpad bug 1178191 in intel "[Ubuntu13.04 final]S3/S4 automatic resume when connect the USB3.0 hub and storage on SharkBay platforms" [Undecided,New] https://launchpad.net/bugs/117819115:52
zedtuxHello everyone. I already spoke in that channel with jpds about my issue https://bugs.launchpad.net/ubuntu/+source/linux/+bug/929715 where I have posted a git repo with an example of code that reproduce the issue. Could someone tell me what is the next step for my issue ?15:54
ubot2`Launchpad bug 929715 in linux (Ubuntu) "netfilter_ipv4.h:13:47: fatal error: limits.h: No such file or directory" [Medium,Confirmed]15:54
ogasawarartg_: I've got the shark bay desktop, but I don't have a usb 3.0 device.  lemme check if I can reproduce without it.15:55
rtg_zedtux, your include issue is prolly fixed in 3.8 or 3.9 with the transition to uapi. I see that limits.h now lives at include/uapi/linux/limits.h16:06
zedtuxsorry rtg_ what means "prolly" ?16:08
rtg_zedtux, probably16:08
zedtuxHa! :D16:08
zedtuxBut in my latest comment I say that I tried with kernel version 3.8.0-19 and still have the issue. So let's hope with the 3.9 version.16:09
henrixzedtux: why are you undef'ing __KERNEL__ in a kernel module? haven't took a look at the details, but i'm pretty sure that's not a good idea16:14
zedtuxhenrix, Otherwise I have a compilation issue. Let me find the error message16:14
zedtuxhenrix, see my comment: // This is needed in order to have access to NF_IP_PRE_ROUTING and NF_IP_POST_ROUTING16:15
zedtuxAlso often I saw this (like in this link: http://stackoverflow.com/questions/12073963/how-to-access-data-payload-from-tcphdr-sk-buff-struct-on-debian-64-bits)16:17
zedtuxhenrix, (http://www.lainoox.com/tag/netfilter-firewall/)16:17
henrixzedtux: right, i'm not familiar with that code. it doesn't seem to make sense to me, but you may be right16:21
henrixzedtux: however, if i grep the kernel code for NF_IP_PRE_ROUTING i find exactly one hit -- and its meant to be used in user space only16:22
zedtuxhenrix, I understand no worry. You're right it's strange...16:22
henrixzedtux: so, you may be using the wrong define. but again, i may be wrong16:22
zedtuxhenrix, ha.. that could explain the strange undef and compilation issue..16:23
zedtuxSee u and thank you rtg_ and henrix16:27
ogasawarartg_: I'm unable to reproduce with a non usb 3.0 device.  I'll see about picking one up later and re-testing.16:39
rtg_ogasawara, ack16:40
sconklinapw: here?18:38
apwsconklin, yo19:45
sconklinheya19:46
apws'up19:46
sconklinquestion: does the cve tracking magic update by using upstream sha1s in commits to our trees, even in the absence of a CVE field in the commit?19:47
apwyes it only uses the sha1s19:47
apwthe cve information is not looked at at all19:47
sconklinok. does anything in our process look at the CVE field?19:47
apwthis means that stable updates triggre cve closure correctly19:47
apwi think launchpad does a little to tie an upload to a CVE19:48
sconklinyes, I agree that this is the way it should work (not that it matters) :-)19:48
sconklinok, that was all19:48
sconklinthanks19:48
apwnp night19:48
=== smb` is now known as smb
infinityrtg_: https://launchpad.net/ubuntu/+source/linux-meta-lts-raring looks more sensible now.21:57
rtg_infinity, ack21:57
infinity(This is what happens when SRUs don't go through the usual tracking process, I let it slip through)21:58
infinityLuckily, recovering dead packages is totally doable now.21:58
=== kentb is now known as kentb-out
rtg_infinity, I don't think this was supposed to be an SRU, was it ? It was an initial upload.22:02
infinityrtg_: Well, sure, but an initial upload to proposed, not release, so still SRUish in the process.22:02
infinityrtg_: (Could have skipped a bunch of verification steps and such, sure)22:02
rtg_hmm22:03
infinityMostly, it's just that without the tracking bug to remind me to copy everything that related to linux-foo, I forgot, and then linux-meta got overridden with the new one from the PPA being copied.  Not world-ending, and mostly my fault.22:03
infinityJust saying a tracking bug would have held my hand. :)22:03
rtg_infinity, ok, I'll make sure to get it right when  upload saucy LTS et al22:04

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