Ubuntu Wiki

Ubuntu Open Week - Ubuntu Kernel Q+A - Leann Ogasawara - Fri, May 1st, 2009

(05:02:10 PM) charlie-tca: Next, we have the always interesting "Ubuntu Kernel Questions and Answers" starring the fantastic ogasawara
(05:02:22 PM) ogasawara: charlie-tca: thanks!
(05:02:28 PM) ogasawara: Hi Everyone!  Welcome to the Kernel Q+A session.
(05:02:42 PM) ogasawara: My name is Leann Ogasawara and I'm the Ubuntu kernel team's dedicated Kernel QA Engineer.
(05:02:56 PM) ogasawara: I'm actually filling in for Pete Graner, the Ubuntu kernel team manager.  He was originally going lead this session.
(05:03:10 PM) ogasawara: I work closely with the Ubuntu kernel team and try to keep an eye on the current Ubuntu kernel landscape as well as anything approaching on the horizon.
(05:03:27 PM) ogasawara: That being said, for this Open Week session I'll try to touch on some of the highlights of the Jaunty Jackalope 9.04 kernel development cycle as well as topics to be discussed and lined up for the upcoming Karmic Koala 9.10 release.
(05:03:51 PM) ogasawara: Here's a quick summary of what I'll get to today:
(05:04:00 PM) ogasawara: 1) Mainline Kernel Builds
(05:04:00 PM) ogasawara: 2) Better upstream kernel bugzilla to Launchpad bug linking
(05:04:00 PM) ogasawara: 3) Kernel Config Sanity Review
(05:04:00 PM) ogasawara: 4) Ubuntu Kernel Version
(05:04:00 PM) ogasawara: 5) Newer Kernel's on latest LTS release
(05:04:00 PM) ogasawara: 6) Removal of LRM
(05:04:02 PM) ogasawara: 7) Suspend Resume
(05:04:04 PM) ogasawara: 8) HWDB Workshop
(05:04:08 PM) ogasawara: 9) Sponsoring /staging drivers
(05:04:10 PM) ogasawara: 10) Remove AUFS
(05:04:12 PM) ogasawara: 11) Kernel Mode Setting
(05:04:14 PM) ogasawara: 12) EXT4 by default
(05:04:16 PM) ogasawara: 13) Improving Wifi
(05:04:18 PM) ogasawara: 14) Remove AppArmor
(05:04:45 PM) ogasawara: As you can see, we have quite a few items to cover.
(05:04:58 PM) ogasawara: If you have any questions, please post them to #ubuntu-classroom-chat.  I'll try to field and answer them as best I can.
(05:05:53 PM) ogasawara: Ok, lets get started!
(05:06:00 PM) ogasawara: 1) Mainline Kernel Builds
(05:06:06 PM) ogasawara: https://wiki.ubuntu.com/KernelMainlineBuilds
(05:06:13 PM) ogasawara: Ahhh, this was one of my favorite highlights of the Jaunty development cycle.
(05:06:23 PM) ogasawara: At the Jaunty UDS (Ubuntu Developer Summit) the kernel team decided it would be beneficial to begin building mainline vanilla kernels.
(05:06:38 PM) ogasawara: These mainline kernels are built from the unmodified upstream vanilla kernel source but use the Ubuntu kernel configuration files.
(05:06:52 PM) ogasawara: These are then packaged as Ubuntu .deb files for easier installation.
(05:07:06 PM) ogasawara: The advantage of providing these kernels for public consumption is they can often help isolate issues.
(05:07:20 PM) ogasawara: Users can identify if a bug is either :  1) a result of a set of Ubuntu specific kernel patches, 2) fixed upstream, or 3) Exists upstream and the upstream kernel developers should be notified.
(05:07:39 PM) ogasawara: The Ubuntu kernel team also generates an Ubuntu kernel version to Upstream kernel version mapping.
(05:07:54 PM) ogasawara: It is a one to one mapping of which upstream kernel version the corresponding Ubuntu kernel version was based on.
(05:08:10 PM) ogasawara: See http://kernel.ubuntu.com/~kernel-ppa/info/kernel-version-map.html
(05:08:22 PM) ogasawara: It's useful for helping determine which upstream kernel version(s) may be of interest to test.
(05:08:34 PM) ogasawara: Going forward into the Karmic development cycle, the Ubuntu kernel team will continue to generate the upstream mainline kernel builds and encourage any testing.
(05:08:48 PM) ogasawara: The 2.6.30-rc4 build is already available for those interested.
(05:10:05 PM) ogasawara: QUESTION: In UDS Jaunty (or was it UDS Intrepid) somebody spoke about dropping i386 support and preserving i586 and i686. Is that still planned?
(05:10:53 PM) ogasawara: artir: for karmic I believe there is plans for a generic 32 and 64 bit support
(05:11:12 PM) ogasawara: artir: no i586 and i686 though
(05:11:22 PM) ogasawara: moving on ...
(05:11:29 PM) ogasawara: 2) Better upstream kernel bugzilla to Launchpad bug linking
(05:11:42 PM) ogasawara: Obviously being able to test the latest upstream kernel is helpful when reporting bugs upstream.
(05:11:56 PM) ogasawara: When a bug has been reported to the upstream bugzilla (kernel.bugzilla.org), an upstream bug watch can then be set from the Launchpad bug report to the upstream bug report.
(05:12:27 PM) ogasawara: See https://wiki.ubuntu.com/Bugs/Watches
(05:12:34 PM) ogasawara: The upstream bug watch will allow bug subscribers to track the Launchpad bug report and also receive updates regarding any status changes in the upstream bug.
(05:12:51 PM) ogasawara: At the Jaunty UDS, communication also began between the upstream kernel bugzilla maintainers and the Launchpad team.
(05:13:04 PM) ogasawara: The Launchpad team have developed a Launchpad plugin to notify bug subscribers of any comments posted in the upstream bug report and vice versa.
(05:13:19 PM) ogasawara: The only issue is that the upstream kernel bugzilla needed to be updated to version 3.0 or newer in order to use the plugin.
(05:13:33 PM) ogasawara: The upstream kernel bugzilla has since been updated to version 3.2.2.
(05:13:44 PM) ogasawara: However, I'm unsure of the current status of the Launchpad plugin being installed.  I'm sure this will be followed up on during the Karmic development cycle.
(05:13:59 PM) ogasawara: Also, for a broader view of the Ubuntu to upstream bug linkage, see https://launchpad.net/ubuntu/+upstreamreport and https://wiki.ubuntu.com/Bugs/Upstream/UpstreamReport
(05:14:14 PM) ogasawara: next ...
(05:14:23 PM) ogasawara: 3) Kernel Config Sanity Review
(05:14:32 PM) ogasawara: At the Jaunty UDS the Ubuntu kernel team sat down and went through the entire Ubuntu kernel config.
(05:14:43 PM) ogasawara: See https://wiki.ubuntu.com/KernelTeam/2.6.28-2-generic-config
(05:14:54 PM) ogasawara: The purpose was to determine which features/drivers should be built in, enabled as a module, or disabled by default.
(05:15:09 PM) ogasawara: Improving boot performance by having modules built in by default was also kept under consideration while reviewing these options.
(05:15:21 PM) ogasawara: See https://wiki.ubuntu.com/KernelTeam/Specs/BootPerformance
(05:15:30 PM) ogasawara: The resulting config then served as the generic config file that was used for the Jaunty kernels.
(05:15:42 PM) ogasawara: Modifications were then made on a per request basis which usually came in the form of a bug report filed in Launchpad.
(05:15:57 PM) ogasawara: The kernel config options for the upcoming Karmic release will likely be based on the final Jaunty kernel configs.
(05:16:10 PM) ogasawara: As usual, kernel config options will continue to be re-examined on a per request basis.
(05:16:25 PM) ogasawara: I'd also point you to https://wiki.ubuntu.com/KernelTeam/Specs/KarmicKernelFlavours for additional information.
(05:16:34 PM) ogasawara: artir: ^^
(05:16:40 PM) ogasawara: Ok, next item ...
(05:16:48 PM) ogasawara: 4) Ubuntu Kernel Version
(05:17:21 PM) ogasawara: just a sec, am going to field a question really quick
(05:17:23 PM) ogasawara: <KhaaL> QUESTION: Can we expect tailor-made config for UNR or MID editions, for example with a I/O scheduler that's more SSD friendly than cfq?
(05:19:03 PM) ogasawara: KhaaL: unfortunately, I don't have an answer for that.  However, the kernel team hangs out in #ubuntu-kernel and could likely give you better info than I.
(05:19:10 PM) sconklin: I can address that
(05:19:30 PM) sconklin: I help maintain the kernels used for netbook devices.
(05:19:39 PM) ogasawara: sconklin: sweet, take it away
(05:20:34 PM) sconklin: We'll accept any requests for improvements. These usually come in the form of a bug report, and are addressed by the kernel team, with input from people who are ectually testing on the devices.
(05:21:09 PM) sconklin: For SSD in particular, we (the kernel team) are beginning more testing of more devices, so I expect to see more attention paid to those.
(05:21:20 PM) sconklin: that's all I have
(05:21:35 PM) ogasawara: sconklin: thanks
(05:21:38 PM) ogasawara: <kees> QUESTION: how does Ubuntu keep CONFIG options from getting lost between stable releases every 6 months?
(05:22:01 PM) ogasawara: kees: good question.  typically the config for the next release is based on the previous
(05:22:15 PM) ogasawara: kees: if however, something gets missed or dropped, please let us know!
(05:22:37 PM) ogasawara: ok 4) Ubuntu Kernel Version
(05:22:50 PM) pgraner: ogasawara: we do a review at UDS as well to make sure we don't drop things
(05:23:15 PM) ogasawara: The upcoming Ubuntu Kernel Version is always a topic of discussion for every UDS.
(05:23:26 PM) ogasawara: The 6 month Ubuntu release cycle is just long enough that it's entirely possible that one, if not two, upstream kernel releases will happen during a single Ubuntu release cycle.
(05:23:40 PM) ogasawara: When deciding which upstream kernel version to converge on, the kernel team will take into account a few factors . . .
(05:23:51 PM) ogasawara:  * When will the upstream release likely occur in relation to the Ubuntu development cycle.  Typically, it's better if the Ubuntu kernel is able to stabilize earlier in the cycle to have more time to test and fix bugs.
(05:24:07 PM) ogasawara:  * The kernel team will also take into account the value of work in a newer upstream version.  They'll also consider the regression potential.
(05:24:21 PM) ogasawara:  * They also have to look at the impact on the rest of the distribution.  If they decide to go with a newer kernel not only do they have to consider kernel specific regressions but also possible userspace regressions.
(05:24:37 PM) ogasawara:  * They also try to determine a what point within the development cycle will they refuse to adopt a newer kernel version.  e.g. Alpha 3 milestone?
(05:24:50 PM) ogasawara: However, with all the above taken into consideration, as you saw with the Intrepid release they bumped to the newer 2.6.27 kernel quite late in the cycle.
(05:25:15 PM) ogasawara: With that said, at the upcoming Karmic UDS the kernel team will review if Karmic will converge on 2.6.30 or maybe 2.6.31
(05:25:35 PM) ogasawara: This also leads into the following . . .
(05:25:40 PM) ogasawara: 5) Newer Kernel's on latest LTS (Long Term Support) release
(05:25:54 PM) ogasawara: Backporting a newer kernel to a LTS release was a topic of discussion at the previous UDS but nothing concrete was decided.
(05:26:07 PM) ogasawara: As mentioned above, many of the factors taken into consideration for picking the actively developed kernel version apply to backporting a newer kernel to a previous release.
(05:26:23 PM) ogasawara: Some additional items that also factor into the debate are:
(05:26:31 PM) ogasawara: * Kernel team resources available to backport a newer kernel
(05:26:38 PM) ogasawara:  * QA resources and hardware for testing
(05:26:48 PM) ogasawara: * Incompatibilities between the newer kernel and older userspace
(05:26:55 PM) ogasawara: * And as always, the fear of introducing new regressions
(05:27:03 PM) ogasawara: Currently the Ubuntu kernel team has settled on pulling in the upstream stable patch sets (for ex. 2.6.28.x) as SRU's (Stable Release Updates).
(05:27:18 PM) ogasawara: See https://wiki.ubuntu.com/StableReleaseUpdates
(05:27:26 PM) ogasawara: For a Non-LTS release, the Ubuntu kernel team will continue to apply the upstream stable patch sets as SRU's for approximately 3-4 months following a release.
(05:27:40 PM) ogasawara: After that initial 3-4 month period, they will cease merging the upstream stable patch sets and strictly focus on security fixes and critical bug fixes.
(05:27:58 PM) ogasawara: For an LTS release, the upstream stable patch sets will be applied as SRU's for the duration of it's life.
(05:28:09 PM) ogasawara: Security and critical bug fixes will carry on for the same duration as well.
(05:28:19 PM) ogasawara: For more information, refer to https://wiki.ubuntu.com/KernelTeam/KernelUpdates
(05:28:50 PM) ogasawara: Next . . .
(05:28:57 PM) ogasawara: 6) Removal of LRM (linux-restricted-modules)
(05:29:06 PM) ogasawara: As some of you may recall, some drivers that had previously existed in the linux-restricted-modules package, eg. nvidia and fglrx, were removed and split out into their own respective DKMS style packages.
(05:29:24 PM) ogasawara: For those that don't know, DKMS (Dynamic Kernel Module Support) is a framework that allows you to manage sources for modules.
(05:29:36 PM) ogasawara: It allows modules to be automatically built against the kernel being installed.
(05:29:47 PM) ogasawara: Generally, it provides the capability for users to run third party modules and have those modules stay in sync with a distro provided kernel.
(05:30:02 PM) ogasawara: Ubuntu also uses DKMS for modules provided by vendors, especially those that contain a binary blob or other closed source portion (e.g. broadcom, fglrx, nvidia).
(05:30:29 PM) ogasawara: The Ubuntu kernel team continues to encourage more use of DKMS, especially with regards to restricted modules.
(05:30:43 PM) ogasawara: As such, for the upcoming Karmic release they are debating on removing the linux-restricted-modules package entirely.
(05:30:57 PM) ogasawara: I believe the linux-restricted-modules package currently contains the madwifi and broadcom drivers.
(05:31:09 PM) ogasawara: It's already been mentioned that the Ubuntu kernel team will not pull in any further updates for madwifi.
(05:31:24 PM) ogasawara: Users are encouraged to instead use the free ath5k/ath9k drivers vs. the proprietary ath_pci driver.
(05:31:54 PM) ogasawara: I've also heard broadcom will be removed from lrm as well.
(05:32:19 PM) ogasawara: Final discussions will happen at the Karmic UDS, stay tuned . . .
(05:32:41 PM) ogasawara: <phitoo> QUESTION: I like using OPENVZ and it's very disappointing that it's only provided in the 8.04 release. I am now using intrepid with the 8.04 kernel but my first attempt with jaunty has been a failure. Isn't there some way to provide an openvz kernel with each release?
(05:33:09 PM) ogasawara: phitoo: not without someone from the community jumping in and helping out.
(05:33:27 PM) ogasawara: <alanbell> QUESTION: how many watts does the LPIA kernel save? Or what other benefits are there for LPIA over the generic x86 build?
(05:34:09 PM) ogasawara: alanbell: none, LPIA is identical to i386
(05:35:15 PM) ogasawara: <kees> QUESTION: are there plans to have DKMS run during package install rather than at boot-time?  it would improve the first-boot-on-a-new-kernel experience (e.g. distro upgrades)
(05:35:49 PM) ogasawara: kees:  not sure if this has been talked about yet.  I'll suggest they add it to their UDS topic list.
(05:36:17 PM) ogasawara: ok next, else we'll run out of time . . .7) Suspend Resume
(05:36:26 PM) ogasawara: Suspend and Resume (typically resume) continues to be problematic so the Ubuntu kernel team decided to make it a focus for the Jaunty development cycle.
(05:36:42 PM) ogasawara: The kernel team had put together a suspend/resume test script which users could run to determine the stability of suspend/resume on their respective system(s).
(05:36:58 PM) ogasawara: They also added hooks to apport to automatically detect and report suspend/resume failures to Launchpad.
(05:37:17 PM) ogasawara: With Jaunty's Beta release a call for testing announcement was posted.
(05:37:26 PM) ogasawara: See https://lists.ubuntu.com/archives/ubuntu-devel-announce/2009-March/000556.html
(05:37:32 PM) ogasawara: General debugging pages were also written.
(05:37:38 PM) ogasawara: See https://wiki.ubuntu.com/DebuggingKernelSuspendHibernateResume
(05:38:03 PM) ogasawara: An overwhelming number of bugs came flooding in, https://bugs.launchpad.net/ubuntu/+source/linux/+bugs?&field.tag=apport-kerneloops+suspend&field.tags_combinator=ALL
(05:38:18 PM) ogasawara: The team is still trying to process the bug reports and find similarities between hardware and/or drivers being used.
(05:38:31 PM) ogasawara: Unfortunately with the large volume of bugs coming in it was decided to temporarily disable the apport automated suspend/resume bug reporting mechanism for Jaunty's final release.
(05:38:49 PM) ogasawara: This will be re-visited for the Karmic development cycle.
(05:39:08 PM) ogasawara: <bryce> QUESTION:  Does suspend/resume and hibernate/resume work properly on your laptop ogasawara?
(05:39:38 PM) ogasawara: bryce: indeed suspend/resume works wonderfully on my laptop (Dell inspiron 1420, all intel hw)
(05:39:59 PM) ogasawara: bryce: although I recently did a hibernate but using a 2.6.30 kernel and I did not resume
(05:40:20 PM) ogasawara: bryce: I haven't had time to investigate the root cause
(05:40:33 PM) ogasawara: next, 8) HWDB Workshop
(05:40:43 PM) ogasawara: Leading up to the Jaunty UDS the Ubuntu kernel team had been holding discussions with the Launchpad team to investigate building a hardware database in which end users could submit their hardware profiles to Launchpad.
(05:41:02 PM) ogasawara: In theory, those profiles could then be linked to bug reports or have their data mined in general to determine for example how widespread a hardware specific issue may affect the overall Ubuntu user base.
(05:41:19 PM) ogasawara: Taking the above mentioned suspend/resume bugs as an example, having a hwdb would be ideal.
(05:41:30 PM) ogasawara: It would allow the Ubuntu kernel team to quickly determine which bugs relate to a specific piece of hardware or driver and then group them to one master bug.
(05:41:43 PM) ogasawara: Subsequently the team could then examine how many users this issue may potentially affect to help prioritize work to be done and resources to be allocated.
(05:42:02 PM) ogasawara: Some development for the hwdb has already began and work in this area will continue to be carried through the Karmic development cycle.
(05:42:15 PM) ogasawara: next . . .
(05:42:23 PM) ogasawara: 9) Sponsoring /staging drivers
(05:42:35 PM) ogasawara: For those not familiar, the staging directory "was created to hold drivers and filesystems and other semi-major additions to the Linux kernel that are not ready to be merged at this point in time.  It is here for companies and authors to get a wider range of testing, and to allow for other members of the community to help with the development of these features for the eventual inclusion into the main kernel tree."
(05:43:04 PM) ogasawara: The staging directory is maintained upstream and is a part of the latest Jaunty kernel.
(05:43:15 PM) ogasawara: The Ubuntu kernel team promotes the testing of drivers etc. which exist in the staging directory.
(05:43:28 PM) ogasawara: If there is a module of interest which is currently not enabled within the staging directory for the Ubuntu kernels, feel free to notify the kernel team of your interest in testing the module and having it enabled.
(05:43:45 PM) ogasawara: It's best to file a bug report in Launchpad with this request.  The Ubuntu kernel team will likely review the request and hopefully enable the module for further testing.
(05:43:59 PM) ogasawara: If there is code you know which exists outside of the mainline kernel tree that is a good candidate to submit for inclusion in drivers/staging, please let the corresponding authors know.
(05:44:19 PM) ogasawara: This ties into the next item. . .
(05:44:26 PM) ogasawara: 10) Remove AUFS
(05:44:38 PM) ogasawara: The Ubuntu kernel team met with AUFS maintainer Junjiro Okajima at the Jaunty UDS.
(05:45:00 PM) ogasawara: At that time Okajima-san was encouraged to get his code merged upstream into the drivers/staging directory.
(05:45:15 PM) ogasawara: The patches were submitted upstream but unfortunately, it seems they were rejected.
(05:45:26 PM) ogasawara: The reasons are captured at http://marc.info/?l=linux-kernel&m=123938533724484&w=2
(05:45:44 PM) ogasawara: Going forward (likely at the upcoming Karmic UDS) this topic will be re-examined and debated as to whether AUFS will be removed entirely.
(05:45:58 PM) ogasawara: Some alternatives would be to possibly go back to unionfs although the move to aufs was prompted by unionfs issues.
(05:46:12 PM) ogasawara: Another option might be device-mapper snapshots.
(05:46:21 PM) ogasawara: Regardless, any alternative will have to be sufficiently investigated and verified to provide support for the Ubuntu LiveCD.
(05:46:42 PM) ogasawara: I unfortunately don't know much more regarding this issue and I imagine it will warrant a new spec/blueprint in Launchpad and another UDS session for Karmic.
(05:46:51 PM) Jack_ is now known as JackWat
(05:47:30 PM) ogasawara: <zhurai> Question: what's AUFS?
(05:47:41 PM) zhurai: (it was already answered)
(05:47:41 PM) ogasawara: http://en.wikipedia.org/wiki/AuFS
(05:47:49 PM) ogasawara: <sbeattie> QUESTION: would it make sense to have the staging drivers enabled in a separate linux-staging-modules package?
(05:48:23 PM) ogasawara: sbeattie: actually that's sorta what the linux-ubuntu-modules package used to provide
(05:48:46 PM) ogasawara: sbeattie: but seeing as they've since removed lum, I doubt we'll see a separate linux-staging-modules package
(05:48:58 PM) ogasawara: 11) Kernel Mode Setting
(05:49:06 PM) ogasawara: This is a big topic I know a lot of people are interested in.
(05:49:17 PM) ogasawara: This will be on the agenda as well for Karmic UDS kernel discussions.
(05:49:27 PM) ogasawara: Given the fact that KMS requires some synchronization with userspace, ie. Xorg, KMS will likely be disabled for the first few Karmic Alpha releases until userspace catches up.
(05:49:39 PM) ogasawara: However, I believe the intention is to have this enabled for Karmic's final release.
(05:49:48 PM) ogasawara: For those not entirely familiar, Kernel Mode Setting will provide things such as:
(05:49:57 PM) ogasawara: * Improved suspend/resume support since the kernel will no longer have to rely on external resources to restore devices to a proper state after resume.
(05:50:11 PM) ogasawara: * Also, with this functionality moving to the kernel, the mode should also be restored automatically and more quickly.
(05:50:22 PM) ogasawara: * Similarly, virtual terminal switching should also see an improvement and likely be more reliable.
(05:50:34 PM) ogasawara: * You'll also see improved debugging as KMS should allow the kernel to display panic/oops messages rather than a hard system hang with no helpful indication of what went wrong.
(05:50:44 PM) ogasawara: For those interested in testing, I'd recommend following the Karmic Alpha release notes for when this will be available.
(05:50:54 PM) ogasawara: Ok, moving on . . .
(05:51:01 PM) ogasawara: 12) EXT4 by default
(05:51:07 PM) ogasawara: The Jaunty 9.04 release supported the ability to install the new ext4 file system, but ext3 remained the default file system for Jaunty.
(05:51:17 PM) ogasawara: Going forward, ext4 will be considered as the default file system based on the initial testing and feedback received during the Jaunty cycle.
(05:51:30 PM) ogasawara: One of the higher profile bugs regarding ext4 in Jaunty has since been resolved:
(05:51:38 PM) ogasawara: See https://bugs.launchpad.net/ubuntu/+bug/317781?comments=all
(05:51:45 PM) ogasawara: Please let the Ubuntu kernel team know of any issues you experience with ext4 to help them make an educated decision for recommending ext4 to be the default file system for the upcoming Karmic release.
(05:52:00 PM) ogasawara: For tracking purposes, the Ubuntu kernel team would prefer you file a bug report in Launchpad for any Ubuntu specific ext4 issues.
(05:52:14 PM) ogasawara: If you are able to confirm the issue with the upstream kernel as well, please also notify the upstream ext4 kernel developers.
(05:52:25 PM) ogasawara: Next item . . .
(05:52:43 PM) ogasawara: 13) Improving Wifi
(05:52:50 PM) ogasawara: The Jaunty UDS had some upstream linux-wireless representation in attendance.
(05:53:00 PM) ogasawara: This was a great time for the Ubuntu kernel team to interact with the upstream developers to get their feedback on what the Ubuntu kernel team could do to help support the upstream linux-wireless community.
(05:53:15 PM) ogasawara: It's also an opportunity to discover how Ubuntu as a distro can help end users engage in testing the latest compat-wireless stack and provide feedback directly to the upstream developers.
(05:53:28 PM) ogasawara: The following wiki was drafted as a result:
(05:53:34 PM) ogasawara: https://wiki.ubuntu.com/KernelTeam/LinuxWireless
(05:53:43 PM) ogasawara: A reference to this doc was also made from the upstream wireless.kernel.org wiki
(05:53:52 PM) ogasawara: http://wireless.kernel.org/en/users/Documentation/Reporting_bugs#Distributionspecificnotes
(05:54:04 PM) ogasawara: Another related item I'd like to point out is that the linux-backports-modules package typically contains an updated compat-wireless stack.
(05:54:23 PM) ogasawara: It's a great way to quickly test if a current wifi issue may be resolved with this updated stack.
(05:54:34 PM) ogasawara: The version of the compat-wireless stack in linux-backports-modules is updated fairly regularly but it is by no means updated daily to contain the bleeding edge bits.
(05:54:46 PM) ogasawara: If you want the bleeding edge, refer to http://wireless.kernel.org/en/users/Download
(05:54:55 PM) ogasawara: The practice of providing an updated compat-wireless stack via linux-backports-modules will continue with the Karmic release.
(05:55:08 PM) ogasawara: I'd also like to take a moment to mention CRDA (Central Regulatory Domain Agent) at this time.
(05:55:16 PM) ogasawara: http://wireless.kernel.org/en/developers/Regulatory/CRDA
(05:55:21 PM) ogasawara: "CRDA acts as the udev helper for communication between the kernel and userspace for regulatory compliance."
(05:55:31 PM) ogasawara: A spec was written for the Jaunty cycle to provide CRDA support.
(05:55:38 PM) ogasawara: See https://wiki.ubuntu.com/KernelTeam/Specs/JauntyCRDA
(05:55:45 PM) ogasawara: CRDA has since been packaged for use within Jaunty.
(05:55:50 PM) ogasawara: See https://launchpad.net/ubuntu/+source/wireless-crda
(05:56:06 PM) ogasawara: <yusuf> QUESTION: I Have a problem where the free ath5k driver drops 90% of pacakges in ad-hoc mode. If madwifi is to be discontinued will ath5k be useble in ad-hoc mode?
(05:56:25 PM) ogasawara: yusuf: I'd encourage you to first test with linux-backports-modules if you haven't already
(05:56:36 PM) ogasawara: real quick last item
(05:56:38 PM) ogasawara: 14) Remove AppArmor
(05:56:46 PM) charlie-tca: 3 minute warning
(05:56:49 PM) ogasawara: Discussions need to happen between the Ubuntu kernel team and the security team as to whether AppArmor will still be supported or if Ubuntu will move to a different security model, eg. SELinux.
(05:57:05 PM) ogasawara: Unfortunately that's about all I know about the current status of this topic at this point in time.
(05:57:25 PM) ogasawara: That's really all I had on my list to cover.  I would however like to point to a few general reference docs as well.
(05:57:33 PM) ogasawara: For general information regarding the Ubuntu kernel team, refer to https://wiki.ubuntu.com/KernelTeam
(05:57:41 PM) ogasawara: For those wanting to get involved, take a look at https://wiki.ubuntu.com/KernelTeam/GettingInvolved
(05:57:49 PM) ogasawara: And lastly, I'd like to point out https://wiki.ubuntu.com/KernelTeam/KnowledgeBase which contains a wealth of information.
(05:57:58 PM) ogasawara: Ok everyone, thanks for listening.  Enjoy the weekend!
(05:58:05 PM) jcastro: thanks ogasawara!