[00:45] mrec__: might be a good idea for you do to a DKMS package [00:45] mkrufky: and make your package remove any pre-existing em28xx [00:45] oops, i meant mrec__ ^ [00:46] i *highly* suggest using the in-kernel dvb-core rather that installing your own... otherwise you'll break somebody else's driver [00:47] i did this by using the headers in /lib/modules/`uname -r`/build/drivers/media/dvb/dvb-core/*.h [00:47] ...and you can depend on their presence if you are doing a DKMS build [02:20] Hi [10:36] fair warning...rebasing on -rc6 now, so I would hold off pushing anything to the repo [10:42] rtg: -rc6 included a fix to xen/time.c...different than mine, but fixed the same problem [10:42] so it wasn't just us [10:43] BenC: cool. how did they fix it? [10:46] rtg: They used iter_div_u64_rem() [10:50] BenC: gotta go. home gain, home again, jiggety jig. === dashua__ is now known as dashua [12:41] what's the irc nick of tim gardner? [12:43] maks_: it is 'rtg' [12:43] ok thanks [12:43] rtg f448b9a70cda0ed9878db485de30363f9c9db6e1 is a dup [12:43] you'll find the same entry below in blacklist with proper desc [12:49] speaking of ubuntu hardy tree, as you might have guessed ;) [12:49] hmm rtg is not here [12:49] will repost [13:41] maks_: rtg is traveling today, he will be back online prob Monday [14:51] mjg59: Do you recall this patch: hostap: send events on data interface as well as master interface [14:52] mjg59: I'm wondering if it's still needed [14:52] I expect so, but haven't tested [15:02] mjg59: is it something that should go upstream? [15:02] Conceivably [15:02] Probably best to check it with Dan Williams [15:03] Ok [15:04] -rw-r--r-- 1 bcollins bcollins 48234 2008-06-13 09:45 ubuntu-stuff.diff [15:04] That our current delta against 2.6.26-rc2 [15:05] good chunk of that is vesafb modular [15:05] err, 2.6.26-rc6 [15:30] Can someone help me out a bit with bug 222703? [15:30] Launchpad bug 222703 in linux "Suspend fails after first time" [Undecided,Incomplete] https://launchpad.net/bugs/222703 [15:31] I can't determine the cause [15:31] should I just mark it as confirmed and leave it to you, or do you want some more informaton? === qense is now known as qense|dinner === pgraner_ is now known as pgraner === qense|dinner is now known as qense [18:02] BenC: ping [18:03] pgraner: Yes? [18:03] BenC: hey, I want to work on a adding a patch and building. Specifically this one http://marc.info/?l=linux-usb&m=121080766320582&w=4 [18:04] BenC: its upstream and I actually need it I'd like to build a custom pkg just for me. Where do we start, I can't seem to find the one page to get me started. [18:06] pgraner: https://wiki.ubuntu.com/KernelTeam/KnowledgeBase [18:06] KernelMaintenance page has a lot of good info [18:07] or [18:07] https://help.ubuntu.com/community/Kernel/Compile [18:07] pgraner: If you want a patch on top of the current git in a ppa, I could push you a script as well === emgent_ is now known as emgent [21:18] kirkland: (switching here) yeah, I think the problem is discussed in the bug you found [21:18] kees: the problem that pitti mentioned? [21:19] (95089) yeah [21:19] okay, i've asked Serge Hallyn to add some comments to the bug [21:20] kees: in case there's something we're missing [21:21] kirkland: yeah, it wasn't entirely obvious to me, but after looking at CAP_SETPCAP, I tended to agree with pitti. [21:21] kirkland: if CONFIG_SECURITY_FILE_CAPABILITIES could operate without it, that'd be nice [21:21] kees: right, i think hallyn will be able to clarify what is/isn't possible [21:24] if you look at the code for where CAP_SETPCAP is defined, it is clearly ifdef'd with CONFIG_SEC.._FILE_CAP.. but I don't know why [21:26] kees: hallyn added some info to https://bugs.edge.launchpad.net/ubuntu/+source/linux/+bug/95089 [21:27] heya hallyn, thanks for the details -- this is an area I'm much less familiar with. :) [21:28] np [21:28] i'm not quite clear on the history [21:29] at the start of the thread, were file capabilities enabled and cap-setpcap turned off, or was it just a regular kernel without file capabilities? [21:30] hallyn: I'm a bit unclear myself. I think the issue is with the #ifdefs that seem to allow CAP_SETPCAP when CONFIG_SEC.._FILE_CAP..=yes [21:30] and without additional context, it seems dangerous [21:30] is there a gitweb site or something where i can see the current code and .configs for default builds? I assume not, just would be kick-ass if so... [21:30] there is, yes, kernel.ubuntu.com [21:30] kees: Should get to your libx86 stuff next week, I've just got nv40 suspend/resume working [21:30] (without any libx86 stuff. Hurrah!) [21:31] kees: cool! thanks. [21:31] mjg59: nice! how does it work without need to talk to the bios?? [21:31] hallyn: current hardy: http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=summary current intrepid: http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-intrepid.git;a=summary [21:31] kees: Uses nouveau's bios table interpreter [21:32] kees: yes, seeing cap-setpcap enabled would set off warning bells :) though we did try to comment it in include/linux/capability.h [21:32] mjg59: niiiice. I'm so glad that project is coming along. [21:32] kees: nvidia and AMD both have tables that are used by the drivers, and just have a small x86 interpreter for it that sits in the bios [21:33] Execute the init table, restore a couple of magic values that would be programmed by the platform bios, run the lvds init code, restore the registers and wham! [21:33] kees: excellent, so I can take a look at the apparmor/caps integration. What about .configs? available? [21:33] An entirely broken console, but working X [21:33] hallyn: the configs are in the debian/configs directory [21:33] excellent [21:34] hallyn: though depending on build options, they are merge together. other folks here might be more helpful in describing that, though. [21:34] hallyn: so, in comment 4, pitti is correct? [21:36] hallyn: this page may be useful for you as well, if you're going to build modified ubuntu kernels: https://help.ubuntu.com/community/Kernel/Compile [21:37] kees: was hoping not to build, just figure out the state of the current kernel [21:38] i.e. i'm afraid one or two apparmor lines may need tweaking [21:39] hallyn: that's fine -- intrepid doesn't have apparmor in it yet, I'm waiting for upstream to finish their port to the current linus tree. [21:39] hallyn: but they're quite responsive about patches [21:39] kees: pitti == martin pitt? If so, he's only right if CONFIG_SECURITY_FILE_CAPABILITIES=n [21:39] hallyn: check in on oftc in #apparmor [21:39] hallyn: okay. so if I boot a CONFIG_SECURITY_FILE_CAPABILITIES=y kernel, I should still get the same output he's showing? [21:40] kees: no, bc cap_setpcap should then be in root's sets [21:40] kees: "debian/configs directory" -> can' tfind it, what is the url? [21:41] hallyn: is there a potentially security vulnerability if CONFIG_SECURITY_FILE_CAPABILITIES=n ? [21:41] hallyn: http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-hardy.git;a=tree;f=debian/config;hb=HEAD [21:41] hallyn: pitti seems to be saying that have cap_setpcap in root's sets is, in itself, dangerous. [21:44] kees: cap_setpcap means something different if file caps are on [21:45] kirkland: do you mean if cap_setpcap is allowed? yes. [21:45] ah-ha -- that's the piece that wasn't obvious to me. [21:45] kees: so yes it looks like apparmor would need some changes - at least for setxattr - to support file caps. [21:45] I'll look at it some more next week. [21:45] hallyn: I'm sure they'd be open to it. [21:45] hallyn: okay, cool. [21:46] kees: where would I send a patch? [21:46] i guess i can just join #apparmor next week [21:47] hallyn: yeah, that's probably the best place to start [21:47] hallyn: they have some public mailing lists too, but IRC is good. If you don't get any response, email me and I'll go ping them directly via email. [21:49] ok, thanks. [21:50] cool, thanks for digging into this -- I've been curious about what was needed for sane filecap support but hadn't had time to look into it. [21:52] kees: hallyn: hey, thanks for taking the time to connect [22:11] 'Collin King' happen to be here ? :) [22:30] somebody here that is able to include a patch into Hardy kernel ? :( [22:34] dupondje: best to open a bug report with the details and the patch [22:35] https://bugs.launchpad.net/ubuntu/+source/linux/+bug/235889 [22:35] enjoy [22:35] :) [22:36] dupondje: looks like it's already being discussed there. :) [22:36] its taking such a huge time to get solved :( [22:36] the patch is there [22:36] I tested it .. it works :( [22:37] hallyn: if you're still around, jjohansen is one of the AppArmor upstreams. [22:37] thanks kees [22:38] jjohansen: i haven't look at it enough, but i think in order for apparmor (in ubuntu) to work with filecaps, apparmor will need a few tweaks, such as calling the cap setxattr hook [22:40] dupondje: it can take time -- but it looks like it's progressing. [22:40] yes, the apparmor hook needs to call cap_inode_setxattr its a fix that is pending [22:41] dupondje: The hardy kernel is currently frozen for the 8.04.1 point release, but I suspect your patch and bug will be addressed shortly after that. [22:41] jjohansen: oh ok, saves me some time :) thanks [22:41] dupondje: I also have a patch queued for after the point release. I realise it's frustrating, but QA processes just can't allow for last-minute changes. :/ [22:41] hallyn: np. [22:42] its a fucking deathlock :( kinda important [22:42] that controller is used in ALOT of servers [22:42] dupondje: Swearing won't help much, nor will it make us delay a point release for a bug (no matter how important that bug may be to some people) [22:45] I know, but I just think its kinda an important bug, as its a verry common used controller in servers [22:46] if your server in a datacenter crashes because of it, its not that funny :) [22:50] I never claimed it was funny. Just that, unfortunately, we have release priorities that mean that you might have to carry a local patch/fork for a bit until after 8.04.1 is out (much like I also have to). [22:50] when is 8.04.1 comming out tho ? [22:52] It's scheduled for end-of-month, barring any massive setbacks. [22:52] Because the devel team is split between intrepid and hardy right now, the hardy point release test cycles are long and careful. [22:53] what they test btw ? :p [22:53] I just test its not working good ;)