20070529

Logs

TZ UTC+1

05:01   BenC    Ok, everyone here?
05:01   pkl_    yes
05:01   zul     yep
05:01   BenC    Anyone who wants to follow along, agenda is at https://wiki.ubuntu.com/KernelTeam/Meeting although we only loosely follow it
05:02   BenC    This meeting, I wanted to start off with the team catching each other up on what we're doing
05:03   BenC    Kyle just reminded me that rtg (Tim) probably is on vacation
05:03   pkl_    on a raft in Montana somewhere...
05:03   BenC    I'll start. but I'm pretty basic. Been merging patches from feisty into gutsy kernel, and sending those patches upstream. Working on a -6.12 upload right now
05:04   BenC    Today I'll start merging some more patches, some of the bigger ones like toshiba_acpi
05:04   BenC    I may end up doing that in linux-ubuntu-modules instead, and disable the one in the kernel
05:05   BenC    kylem: how you been doing?
05:06   kylem   i haven't really done anything kernel related in the las tweek.
05:06   kylem   getting intel X support up to snuff for gutsy and then back for feisyt.
05:07   BenC    mobile stuff?
05:08   kylem   BenC_, no.
05:08   BenC_   Ok, we can catch up on the other stuff later then
05:08   BenC_   pkl_: how's things with you?
05:10   pkl_    I've been following post-feisty kernel upload bugs and emails.  Going to start a new release, and integrate the current crop of patches to be put in.
05:10   BenC    pkl_: did you discover what caused the storage bug?
05:10   cr3     kylem: I noticed important changes to the intel X driver in gutsy such as support for rotation
05:10   pkl_    In addition, done a bit of work on the mobile stuff, mainly Unionfs 2.0 status, and ODF (On disk format) investigation
05:11   pkl_    Will be falling back to looking at Squashfs instrumentation to investigate the liveCD slowdown soon.
05:12   pkl_    What storage bug BTW?  There's lots of the upload broke booting, fstab, resume etc.  All down to the reversion of libata support for PATA.
05:12   BenC    I don't think it's really a bug, but some people noticed hdX changing to sdX
05:13   BenC    I think they had hardcoded stuff in fstab, not using UUID's
05:13   BenC    or maybe it was sdX to hdX
05:13   pkl_    Most seem to have noticed sdX changing to hdX...  All the breakage seems to have been caused by hardcoding devices in various files.
05:14   BenC    one other common non-bug I noticed was people complaining we "broke" their locally compiled modules, so I'm suggesting to Kees to update the security notices on ABI bump kernels to include a note saying "ABI changes, recompile your locally compiled modules"
05:15   BenC    pkl_: any real regressions come out of that?
05:16   pkl_    Couple of people complaining that the ps/2 fix doesn't fix VIA chipsets.  Some VIA chipsets seem to be exhibiting the same problems as SiS chipsets, but the fix doesn't work and so the underlying cause is obviously different.  What the cause is however is not apparent.
05:16   BenC    pkl_: Are the VIA chipsets a regression from edgy still?
05:18   pkl_    I've not any regressions so far...  There's been some complaints that some -16.28 packages have been missing, or dependencies still based on -15.27...
05:18   mdz     pkl_: what's the status of the mobile kernel flavour?
05:18   mdz     pkl_: did you receive the config from intel?
05:18   BenC    pkl_: We can discuss this more on kernel-team@ if you need some help with these bugs...sounds like overall there weren't any major issues with -16 vs. -15
05:19   pkl_    The VIA chipsets are a regression from edgy, in addition mice on VirtualPC are still broken.
05:19   pkl_    mdz:  I've received nothing from Intel.
05:19   BenC    Ok, I remember the VirtualPC one being the initial test case...I couldn't get VirtualPC working though
05:19   BenC    mdz: Re: Intel-Mobile stuff, we haven't gotten a single bit from them, that I know of
05:20   BenC    mdz: Should Phillip post to ubuntu-mobile requesting updates on kernel side?
05:21   mdz     BenC: he should contact rob rhoads directly
05:21   BenC    Ok
05:21   mdz     and then mail ubuntu-mobile to let folks know what's happening, once there's something to tell
05:23   BenC    Noted, thanks
05:23   BenC    I want to skip ahead to KVM/cr3 topic...
05:24   cr3     ahoy, the KVM has been connected to each desktop in the hardware certification lab
05:24   cr3     to access the hardware certification network, I have decided to configure a VPN with dynamic addressing so that multiple can connect at the same time
05:25   BenC    What software do we use to access it?
05:25   cr3     to access the machines, there is a Java client which enables the user to select the machine
05:26   cr3     to poweron the machines, the machines are connected to PDUs which can manage the power to the machines
05:26   cr3     the BIOS of the machines are all configured to poweron when power is detected on the power supply
05:26   cr3     this is necessary because the machines are still located in an office which becomes prohibitively hot when the machines are all turned on
05:27   BenC    So we should power them down when we aren't using them?
05:27   cr3     to address this problem, I will be sending quotes for colocation centers this afternoon which should be approved in a matter of a week or two
05:27   cr3     BenC: ideally, yes
05:28   cr3     the process of configuring the VPN and using the client software should be straightforward, the problem is mostly having to deal with the PDUs which I would like to spare for you guys
05:28   BenC    Is there a howto page where we can go on how to configure the vpn and client?
05:28   cr3     so, I see a couple options: 1. until we get a colocation center, the distro team could ping me to poweron machines; 2. I could provide detailed instructions on managing the PDUs sometime this week.
05:28   BenC    I have to be honest, it's been years since I used a VPN on Linux
05:29   cr3     BenC: not yet, but it's on my TODO list. when would you like the instructions, just for the VPN and client to start with?
05:30   BenC    yeah, just being able to test this setup would be a good first step
05:31   cr3     BenC: sure, I could aim to provide these instructions sometime tomorrow and how should I inform the distro team about it?
05:31   BenC    Any chance the machines could have a CDRW media inserted in them constantly so we can do "burn ISO, reboot to livecd, install, reboot to new system" tests?
05:32   BenC    cr3: I can give you an email address to send it to
05:32   cr3     BenC: about half the machines don't even have a CDROM device because I often receive bare motherboards from Intel
05:32   BenC    cr3: Ah, right...is there a PXE boot server?
05:33   cr3     BenC: yes, and I would like to spend the next couple weeks improving the usability of the PXE boot server which I call a certification appliance
05:33   cr3     the reason for this naming is that I would eventually like it to become a reusable component, for people like Intel for example
05:34   BenC    cool
05:34   cr3     mvo has also expressed interest in running a certification appliance, just for the automatic building process
05:35   cr3     so, we could schedule discussing the certification appliance at the next kernel meeting in a couple weeks
05:35   BenC    Sounds good...thanks for your time cr3
05:35   cr3     BenC: does that seem like a reasonable timeline for you?
05:35   BenC    yes, right after Tribe 1 is when we'll probably want to start using it
05:36   BenC    that's when our bug level increases from more exposure
05:36   cr3     IOU: 1. VPN and KVM client instructions tomorrow; 2. certification appliance on June 12th.
05:36   cr3     we should also consider discussing the status of colocation at the next meeting
05:36   BenC    Ok, noted for next agenda
05:36   cr3     thanks!
05:37   BenC    have a good week cr3, and thanks again
05:37   BenC    Alright, moving on to Tribe 1 preparation
05:38   BenC    2.6.22-rc3, aka Ubuntu-2.6.22-6.12, was uploaded this morning, but I have a ppc build failure to contend with
05:39   BenC    I'll have that done by today, and do lum/lrm/meta behind it, so hopefully by the end of tomorrow, things will be stable
05:40   BenC    that will give us only 8 days till tribe 1, so this is most likely the kernel that will be in that milestone
05:40   BenC    fyi, Tribe is scheduled for June 7th
05:40   BenC    *1
05:41   BenC    I'm not aware of any showstoppers for this, but anyone have any known issues we should be worrying about?
05:43   BenC    Eerie silence :)
05:43   pkl_    well  assumed silence meant no from me :)
05:44   BenC    You guys running gutsy kernel yet?
05:44   kylem   yeah.
05:44   kylem   on all my machines.
05:44   BenC    nice
05:45   pkl_    no, I've been running the test Feisty kernel.  Now it's released, I could move to gutsy...
05:46   BenC    pkl_: you gotten a xeon mobo yet?
05:46   BenC    probably good if you could keep feisty on at least one machine
05:46   BenC    even if it's dual-boot
05:47   pkl_    BenC: not yet (xeon mobo).
05:47   BenC    well, I think that's it for us...any topics anyone wants to bring up before we cut out?
05:49   BenC    Excellent, thanks for coming everyone, and have a good week...thanks to pkl and rtg (absent) for their first security uploads this past week

MeetingLogs/Kernel/20070529 (last edited 2008-08-06 16:16:34 by localhost)