[01:30] I am new to kernel programming and I am working on a test project (I understand there are things that already do this)that takes receives a packet via NETFILTER down in the kernel and just drops it. What i want to do is check if the packet against a list of approved ip's and ports. From what I have been reading passing the packet up into User Mode is my only option, especial if the list is sotred in a DB. I am curious wha [01:30] t is the best way for a pacet to be pased to a user-mode program verified then have its findings sent back to the kernel module with a yes allow the packet or no drop the packet? [01:31] I am new to kernel programming and I am working on a test project (I understand there are things that already do this)that takes receives a packet via NETFILTER down in the kernel and just drops it. [01:31] What i want to do is check if the packet against a list of approved ip's and ports. From what I have been reading passing the packet up into User Mode is my only option, especial if the list is sotred in a DB. [01:31] I am curious what is the best way for a pacet to be pased to a user-mode program verified then have its findings sent back to the kernel module with a yes allow the packet or no drop the packet? [01:33] honcho: openflow controller does that [01:34] looking that up now [01:34] thanks [01:36] is there a way to drop packets from user mode, or does that have to happen in a kernel module? [07:38] moin [08:54] smb: *nudge* [08:54] infinity, hmmmm? [08:54] smb: Do I get a linux-ec2 upload today? [08:55] infinity, Sure, if I can get out of Brad or henrix_ how I am supposed to make SB behave [08:55] smb: https://bugs.launchpad.net/ubuntu/+source/linux-ec2/+bug/1137197 is waiting patiently on you. :) [08:55] Launchpad bug 1137197 in kernel-sru-workflow "linux-ec2: 2.6.32-351.62 -proposed tracker" [Medium,In progress] [08:56] infinity, The git repo is pushed since Friday at least [08:56] smb: How is shankbot misbehaving? [08:56] infinity, By completely ignoring me... well or I don't know how to tell him [08:57] Ignoring what? Until it's uploaded, there's not much for it to do. [08:57] infinity, Thought the magic now is to set upload-to-ppa to confirmed [08:58] It doesn't upload for you... [08:58] * infinity is potentially confused by what you're driving at. [08:59] infinity, Could well be that *I* am quite confused by the last email telling me not to set the prepare-task to fixed anymore [08:59] but instead change upload-to-ppa but then it may still be someone else that would do the upload (as they did before) [08:59] smb: You're meant to just upload stuff to the PPA, and let the bot set prepare-* to Fix Released when it sees it all done, yes. [09:00] infinity, I guess the detail here is that before henrix_ was doing the uploads, I only did the git prepare [09:00] smb: Ahh. Well, you could ask him nicely to do so again. :P [09:01] smb: But I'd personally prefer it if the person prepping the git tree also checked his work by making a source package. [09:01] infinity, Have to talk to him [09:01] smb: And once you've done that, you may as well upload it. [09:02] infinity, I would not mind either if its clear that thats is the process [09:02] smb: Well, I'd ask bjf, but he seems to have run away from the channel. [09:03] infinity, Oh, I think he does this thing, this sleep thing you know [09:03] smb: I don't understand. [09:04] Hm, I am awake and its morning and he lives on the other side of the continent across the ocean... [09:05] * smb thinks there might be a lot of smileys missing [09:05] it really is awkward having to sleep, we'd be so much more productive if we weren't mortal [09:11] ++ === henrix_ is now known as henrix [09:18] cking: I've slept about three hours since Friday. Are you implying that I'm immortal? [09:18] infinity, henrix, ok, I uploaded something (hopefully the correct lucid ec2) ;) [09:18] henrix: I think smb either needs a crash course in process, or you need to upload linux-ec2 for him. :) [09:18] smb: Hah, jinx. [09:19] infinity, a implies b does not mean b implies a ;-) [09:21] * smb wonders whether he should have taken the opportunity to silently swap the ec2 kernel by the version without the big xen-classic patch... [09:22] smb: ? [09:22] infinity: smb: ok, i'm looking at the logs here... is there something you would like me to do? :) [09:22] infinity, you don't wanna know. these are not the robots you are looking for... [09:22] or that's all sorted out? [09:22] smb: Oh, as a Xen fan, you may be happy to know that Xen on ARM is already functional for both armv7 and armv8. [09:22] smb: And they're upstreaming it all like good little boys. [09:23] smb: So, it might beat KVM to full paravirt on ARM. Neener, neener to the KVM people. :P [09:23] henrix, no, think we are good. we just need to clarify wheter I should do the uploads myself in future as well [09:24] infinity, As a Xen "fan" yeah, glad they do upstream. Though as not owning more arm than I really have to I would not have any hw to run it on [09:24] :) [09:24] henrix: ack [09:24] grr.... [09:24] smb: ack :) [09:24] heh :D [09:24] * henrix needs to get out of weekend timezone [09:25] henrix: You talk to yourself on the weekends? [09:25] infinity: i do! [09:25] henrix, Probably that was part of the last email about using /changing the status of upload-to-ppa to confirmed. But I missed the part defining who would need to do the upload. I mean I am fine with doing it. Just need to add that to my brain. :) [09:26] infinity, me to... ok, mostly because I am the only one listening [09:27] smb: I'm not sure if anything has formally changed WRT who uploads, but I'll reiterate that I'd rather the developer who prepared the tree actually create a source packages (and, I dunno, test build it, is that asking too much? :P), at which point, why would you then have someone else do exactly the same thing again, when you could just upload it yourself? === zequence_ is now known as zequence === tyhicks` is now known as tyhicks [09:30] infinity, Right, well that actually was overdone in the past as I did a test-build which henrix repeated. So it has really some advanage to do it myself. It just was different process before. Not that it cannot be changed, but we just need to agree to do it differently (and have all the same recollection of how) [09:31] henrix, uh oh... 95% of ppa space reached [09:32] smb: infinity: i'm ok with that as well. we just need to sync with bjf i guess [09:32] smb: that's not good. there's probably lots of stuff to be deleted. i'll take a look at that [09:33] There doesn't seem to be much cruft in there. [09:33] henrix, yeah, unfortunately junk is not collected automatically... [09:33] lts-backport-natty is about the only cruft... [09:34] infinity, hm, maybe depedns on the view... though I guess *you* would see it all [09:34] Used to be all the superseded packages lingering [09:34] superseded ones should get deleted... [09:34] In theory. [09:35] (But no, I can't see any special view, I can't even see the views you see, since I'm not in the team) [09:35] Someone should add me. [09:36] infinity, practically I see 177 packages in the delete view... We probably should get you there... think Tim was admin... and Leann [09:37] Yeah. I'd argue that adding me would be incorrect from an org chart POV, but that's BS, cause half the people in that LP team aren't Leann's reports. [09:38] apw: Want to add me to ~canonical-kernel-team? Kthx. ;) [09:38] infinity, agreed... oh apw could do too [09:38] infinity, he has not yet shown up... [09:39] smb: He's in the channel, that's shown up enough for me. [09:39] Ok, its Monday, a day before the vUDS that still does not seem to have all sessions... [09:39] infinity, Thats bip to you [09:42] Doh... having the delete PPA button right next to the view details one isn't very helpful on Monday mornings... [09:43] henrix, just a note (though maybe you know already) when you delete superseded packages, only select about 10 or so at a time [09:43] smb: yeah, i guess i've been there already :) [09:44] henrix, ok, yeah... guess we all have. :-P [09:44] smb: anyway, i can't see the ppa size... "Failed to fetch repository size" [09:44] henrix, yeah, just saw the 95% mentioned in the upload email I got [09:44] times out for me too [09:45] ah, ok [09:47] smb: lp is unusable atm, i'm getting timeouts everytime i try to do something [09:47] infinity, i don't have any objection, i can see a couple of positive benefits given the specific relationship you have, that whole tema is a bunch of mess [09:48] infinity, so i'll discuss with leann et al about a general overhall [09:48] henrix, you have to delete a very few at a time else it barfs [09:48] henrix, i managed to delete all of natty (4) at once [09:48] apw: Yeah, the team doesn't relate to organisational units at all, or I wouldn't have asked. [09:48] apw: 4 at once? wow! :p [09:49] henrix, Hm, I could at least do 3... [09:49] yeah, but i'm unable to get a list of the packages. i keep getting a timeout [09:49] infinity, 'before' it was used for status and it wasn't great to have others on there, then we added -distro and it is now that so i think it should be ok [09:50] infinity, really we should have a separate -upload or -ppa group which has the ppa on it, but we [09:50] should have done that a long time ago ... sigh [09:50] henrix, Hm, that seems to work for me... so maybe I should go deleting [09:50] apw: Yeah, silly status tracking is why slangasek kicked me out of canonical-foundations. :P [09:50] smb: i'm getting the timeout when i try to apply filters [09:50] (Which is irksome, cause that punts me from PPAs and recipes that I used to have access to) [09:51] henrix, Ok, let me go ahead [09:51] smb: ack, thanks [09:51] Re-using OU/status teams for anything functional is probably a mistake. [09:51] he should jsut make a -status or -i-care-about and use that [09:51] or even -reports [09:51] yeah ... concur on organisational again functional [09:56] infinity, i see the l-lowlatency ones we did via PPA made it to -proposed, seemed ok i assume [09:57] I've yet to test boot the lowlatencies. I'll do that now [10:07] henrix, Ok, I deleted all superseded packages except one per package. Still cannot calculate the current size but deletions do take some time. [10:08] smb: cool, thanks. for some reason, i'm still getting timeouts when i try to filter to see superseded packages only... [10:09] henrix, Oh maybe that is the reason. I did not attempt to. But in the overview the superseded are a different shade of gray [10:11] smb: ah, right. so you just deleted them from the original listing. ok, i'll try to remember that next time :p [10:16] henrix, Hm, since that ec2 upload was a ABI change, I also prepared the meta package... Now how do I tell that in the bug? there is only a prepare-meta task, but since I should not change the other prepare task either, I am not sure? [10:17] smb: afaik you just need to upload the -meta package. i believe the bot will figure out that there was an ABI bump and will look for it [10:17] ok,... lets see [10:22] apw, don't suppose you're about? [11:28] lantizia, hi [11:31] apw, hey... sorry to bother you - do you recall https://bugs.launchpad.net/ubuntu/precise/+source/linux/+bug/1105230 ? [11:31] Launchpad bug 1105230 in linux "CONFIG_SOUND_OSS_CORE{,_PRECLAIM} seem to have become reenabled in 12.04 and later" [Medium,Confirmed] [11:31] lantizia, some indeed ? [11:32] well I'd love to test the changes but it seems every kernel i've got from kernel-ppa-pre-proposed (precise) doesn't have the fix [11:33] i'm guessing for precise it'll say "fix released" when it's in the main repo... but right now it says "confirmed" so does that mean someone has made the changes somewhere in a pre-release kernel for precise? [11:34] nope, that means it is believed to be an issue there [11:34] the changes were stalled because turning off CONFIG_SOUND_OSS was difficult in the odler releases [11:34] i think we were intending to just turn off _PRECLAIM and see if that was enough [11:34] but it's only PRECLAIM that was causing the issue [11:34] and ... that does not seem to have been done [11:35] some of that would be because we would want to see the changes bake in raring for a bit without anyone noticing, for example the ubuntu-studio folk [11:35] who care about sound a lot more than we do generally [11:36] lantizia, you are using which kernel on precise there ? [11:36] well preclaim doesn't add or take away any function or feature - it just reserves the space for a feature that isn't even enabled anyway [11:37] indeed, but thats the process. i think we can call raring baked without issues by now though [11:37] i suppose I'm saying - there is zero chance of anyone even noticing... if they do notice, what they'll be noticing is they now *can* use oss reserved addresses where as before they couldnt - but that's surely not going to break something [11:37] currently i'm on 3.5.0-18-generic [11:38] but i'm not sure if that's from the ppa's or not now ... i'd have to check [11:38] might have rolled back since it didn't do anything [11:38] ok so a quantal kernel, so i'll prepare some test kernels against P and Q and get you to test them [11:38] and if that works i can get it out for review and application [11:38] yay :D thank you very muchly [12:10] apw, i'm installing 13.04 in a VM to test the lack of PRECLAIM there for now (i.e. my mission to get osspd working which needs the OSS device name space available for it to fill) [12:15] bah, where did the morning go? [12:19] henrix: Hey, did you need more testing, and if so what kind, for 1111416 (nfs4.1)? [12:20] maswan: no, i guess we're good with that. i'll submit a patch to fix that [12:20] maswan: sorry, i should have done that for this cycle already but completely missed it [12:20] henrix: excellent! [12:42] lantizia, pointers to the test kernels in your bug [12:43] apw, ooh thanks - I can test the precise one on my physical machine, is it ok if I test quantal on a vm? [12:43] yes [12:43] assuming it can show the issue [12:43] and I'm testing them for... ? just to see if it fixes my issue (which is using OSS name space) [12:45] yes, and to see they still boot of course, but the first tells you that [12:45] ok :D will let you know as a reply in the bug? [12:45] lantizia, and poke me here when you've done it, else i'll forget [12:50] * ppisati is baffled [13:00] * ppisati disappears for ~10min [13:29] henrix, This mornings operation cleansweep finally shows results. LP will let you know the repo size again. Its down to about 33%. :) [13:30] smb: sweet! that was a nice cleanup :) [13:30] henrix, Yep. :) We could be even more aggressive by removing _all_ superseded, but hey. ;) [13:31] smb: heh. i guess this cleanup will allow us to survive for a while now [13:31] @ubuntu-devel: too long, couldn't read [13:31] lol [14:16] apw, So far so good... http://www.lantizia.me.uk/so-far-so-good.jpeg === kentb-out is now known as kentb === lynxman_ is now known as lynxman [15:38] * ogasawara back in 20 [15:46] * rtg uploads raring -10.19 rebased on v3.8.2 [15:50] rtg, sweet [15:51] apw, drat, forgot the orig tarball [15:51] rtg, meh, we'll survive [15:51] I'll try next time... [15:52] good to point it out tho. === rsalveti_ is now known as rsalveti [16:10] * bjf back in 20 [16:30] back [16:31] bjf: thanks for your link yesterday. that was exactly what I was looking for and was able to build the proper ubuntu kernel packages (with our needed patches) [16:31] dirtyfreebooter, np [16:43] * ppisati -> gym [16:43] * smb curses about the day being gone that far already [17:13] smb, yep, it's been one of those kind of days [17:14] must be the weather [17:18] lantizia, did those kernels solve your issue [17:20] apw, well they boot - not got around to testing osspd yet as having diffculties with an app that needs to run through it (old 32-bit thing) will keep you updated though [17:20] gotta run right now though to bbl [17:38] for uds, I take it the kernel/hardware track is merged with foundations now? [17:39] arges, yes [17:39] bjf: thanks just making sure i'm not missing something === henrix is now known as henrix_ === henrix_ is now known as henrix [18:27] * rtg -> lunch === shadeslayer is now known as evilshadeslayer === jsalisbury changed the topic of #ubuntu-kernel to: Home: https://wiki.ubuntu.com/Kernel/ || Ubuntu Kernel Team Meeting - Tues March 12th, 2013 - 17:00 UTC || If you have a question just ask, and do wait around for an answer! === rsalveti_ is now known as rsalveti [20:53] I found a bug with a particular integrated sound card and have a fix; where do you guys prefer I report the bug? [20:58] * rtg -> EOD [21:04] Hyperiant, It's best to open a Launchpad bug report and attach the patch there. === henrix is now known as henrix_ === kentb is now known as kentb-out === evilshadeslayer is now known as shadeslayer