[01:35] <BenC> "...this document refers to domains as abstract isolated environments in the platform to which a subset of host physical memory is allocated."
[01:35] <BenC> I love reading tech speak to my friends, just to see the look on their face
[01:36] <BenC> Keeps them from asking me the same question all the time.."What are you reading?"
[01:42] <Nafallo> haha
[01:51] <zul> BenC: virt?
[01:53] <BenC> VT-d
[01:53] <BenC> IO virtualization from Intel
[01:54] <BenC> pretty cool stuff, let's you put a devices IO into a domain so that only the driver, or the guest system can access it...sort of like jail for devices
[01:55] <BenC> that way the device cannot scribble DMA all over memory, and guest systems can have direct access to devices on the host without worrying about compromising the host
[01:57] <BenC> So you can insert a network card, and SATA card, and have it assigned directly to the virtual machine
[01:57] <zul> neet
[02:01] <Nafallo> sounds kewl :-)
[02:02] <lifeless> BenC: ping
[02:03] <BenC> lifeless: yo
[02:03] <lifeless> think I'm running into the multiply-mounted squashfs bug in 2.6.10-17 here with ubuntu-on-tap
[02:03] <lifeless> wondering what the workaround on the livecd was so I can reproduce it
[02:04] <BenC> The workaround was to not do it
[02:04] <BenC> the fix is in edgy-security
[02:04] <lifeless> ok,so I'm using the livecd squashfs as-is, I presume that means I have the fix...
[02:05] <BenC> are you getting a zlib oops from the kernel?
[02:05] <lifeless> yes
[02:05] <BenC> you have the workaround on the current edgy live-cd
[02:05] <lifeless> the stock 6.10 iso ?
[02:05] <BenC> should be, yes
[02:06] <BenC> the workaround was to copy directly from the squasfs root mount, as opposed to mounting temp and copying from there
[02:06] <lifeless> righto, so that should still be in place
[02:06] <lifeless> let me see
[02:07] <BenC> best person to ask is cjwatson, he does ubiquity and did the workaround
[02:07] <lifeless> now you say that there is a fix - you mean kernel fix yes? - in edgy-security ?
[02:07] <BenC> right, the kernel in edgy-security has a fixed squashfs
[02:07] <BenC> there was talk of an edgy point-release to pull it and other things in, but I've no idea if that will happen
[02:08] <lifeless> ah-ha, I dont have security on my master machine
[02:08] <kinema> Is there a list of patches that have been applied to the various "official" Ubuntu kernels?  I looked on the wiki but wasn't able to find anything
[02:08] <lifeless> I'll install that kernel and generate a initrd and see.
[02:08] <lifeless> thanks!
[02:09] <BenC> kinema: It's all in the git repo via revision history
[02:10] <kinema> hmm... that's a pretty obious one is it?  now i feel kind of stupid.
[02:11] <kinema> BenC: Have you had any luck getting the liveCD to boot in KVM?  I saw you making inquires on the KVM list recently.
[02:12] <BenC> kinema: I'm not sure I'm qualified to do the real-mode 32-bit changes needed for it
[02:12] <lifeless> BenC: whats the fixed kernel-version ?
[02:12] <kinema> Understandable... that's some pretty gritty work.
[02:13] <BenC> lifeless: Can't recall
[02:20] <kinema> Does anyone know if it's possible to patch a kernel with the d80211/devicescape code using just the headers or if a from the ground up custom kernel is required.
[02:23] <mjg59> There's a one-line patch needed to the main kernel
[02:24] <kinema> mjg59: for d80211?  I thought all the code was in the wireless-dev tree still.
[02:33] <kinema> mjg59: what one-line patch?
[02:41] <kinema> what kernel is feisty herd 2 running?
[02:43] <crimsun> kinema: 2.6.20-rc3
[02:43] <crimsun> or rc4, don't remember immediately, but it has many, many patches
[02:43] <Nafallo> rc4 I think
[02:44] <kinema> if it's heavily patched I think I'll just download linus' rc5 and patch it myself.
[02:57] <kinema> ugh.... gitweb is slow!
[02:57] <kinema> painfully slow
[03:01] <lifeless> BenC: sweet, thank you, its good
[03:02] <BenC> lifeless: good deal
[07:44] <lifeless> BenC: around ?
[07:53] <fabbione> lifeless: he just went to sleep
[11:29] <\sh> hmmm...why is UTS_RELEASE not set anymore in /usr/include/linux/version.h ?
[01:41] <BenC> \sh: It's in linux/utsrelease.h
[01:43] <BenC> \sh: If you are needing it for userspace, use the libc call to get it, if you need it for a kernel module, use utsname()->release
[02:40] <jes-o-mat> hi
[02:48] <jes-o-mat> kylem, BenC: I'd like to retest #37452 but till now it's quite unclear to which git you have uploaded your patch. http://kernel.org/git/ states that both of your dapper-updates repos haven't been updated since weeks?
[02:48] <BenC> jes-o-mat: Use dapper-proposed repo to get the built kernel
[02:49] <kylem> er, oops, did i forget to push? :\
[02:51] <kylem> i did, terribly sorry.
[02:51] <jes-o-mat> BenC: is there also a proposed update for the u-i somewhere?
[02:51] <BenC> u-i?
[02:52] <jes-o-mat> ubuntu-installer =)
[02:52] <jes-o-mat> . o O (just named it like that because of d-i ;)
[02:52] <BenC> the installer doesn't get rebuilt in -proposed, but you can surely rebuild it using the debs/udebs in -proposed
[02:53] <jes-o-mat> but will there be some officially supported install medias available?
[02:54] <jes-o-mat> so to say 6.06.2?
[02:56] <BenC> https://wiki.ubuntu.com/KernelUpdates
[02:57] <BenC> that explains how fixes in -proposed get to stable
[02:57] <BenC> basically, it needs testing before it gets into a stable kernel, and then it's left in the hands of the release team as to whether or not a point release occurs for new CD's
[02:58] <Mithrandir> BenC: s/release team/stable release team/ :-)
[02:58] <kylem> BenC, http://www.reghardware.co.uk/2007/01/15/intel_robson_pcie_approval/
[02:59] <jes-o-mat> BenC: where shall test-reports being send to? reply to #37452
[03:00] <BenC> kylem: I wonder if the linux community has even pondered how to make use of that
[03:00] <BenC> jes-o-mat: To the bug report is the best place
[03:00] <kylem> BenC, i am... :)
[03:00] <BenC> kylem: Storing your pr0n passwords is not a legitimate use :P
[03:00] <kylem> jes-o-mat, if you could cc me privately as well, it would be good. my bugmail folder is pretty large.
[03:00] <kylem> BenC, no... suspend to disk.
[03:01] <BenC> Using it for /boot would be nice too
[03:01] <BenC> wonder if the BIOS provides compatibility for accessing it like a normal drive
[03:01] <kylem> jes-o-mat, i picked the patches to backport out of one report there, if they didn't help, i can pretty quickly have a new kernel out for you to test, but if it requires me finishing going through bugs then it will take longer. :)
[03:02] <jes-o-mat> the x4100 are currently blocked by some collegues, but I expect that I can reinstall them tomorrow
[03:03] <kylem> ok.
[03:03] <jes-o-mat> but it's quite unhandy, because I need to install to a single disk from CD, install kernel from proposed and then assemble the RAID to verify
[03:07] <jes-o-mat> kylem: that is just waiting for your changes: http://boschman.de/gallery/v/bjoern/tec/sun-x4100/dsc00077.jpg.html :-)
[03:26] <zul> argh...i guess i know what im going to be doing this weekend...thanks pitti
[03:42] <zul> Mithrandir: ping
[03:42] <Mithrandir> zul: You sent me a contentless ping.  This is a contentless pong.  Please provide a bit of information about what you want and I will respond when I am around.
[03:53] <jes-o-mat> Mithrandir: The Answer to The Ultimate Question Of Life, the Universe and Everything
[03:54] <jes-o-mat> hm it does not seem to be an infobot :)
[03:57] <zul> umm...no
[04:08] <Mithrandir> jes-o-mat: no, I'm not an infobot.
[04:10] <jes-o-mat> Mithrandir: I just thought that because of your generic answer style :)
[04:10] <Mithrandir> oh, it's a script, sure.
[04:10] <jes-o-mat> Mithrandir: ping
[04:10] <Mithrandir> jes-o-mat: You sent me a contentless ping.  This is a contentless pong.  Please provide a bit of information about what you want and I will respond when I am around.
[04:10] <jes-o-mat> I see
[04:11] <jes-o-mat> hehe
[04:12] <zul> ah man im not going to get any work done tonight american idol is on tonight
[05:22] <jes-o-mat> BenC: Unfortunatelly I do not find any linux-image on dapper-proposed?
[05:23] <BenC> either it failed to build, or an archive person hasn't processed it yet
[05:23] <jes-o-mat> where to check for that?
[05:28] <tepsipakki> it isn't in the queue
[05:29] <tepsipakki> 15:49 < kylem> er, oops, did i forget to push? :\
[05:29] <tepsipakki> 15:51 < kylem> i did, terribly sorry.
[05:33] <kylem> i uploaded
[05:33] <kylem> i didn't push the tree to korg.
[05:35] <tepsipakki> oh, sorry
[05:36] <kylem> lol, np
[05:37] <tepsipakki> -proposed ones don't show up in the build queue :)
[05:38] <tepsipakki> or -security, for that matter
[05:39] <tepsipakki> anyway, the images are not in archive.u.c?
[05:55] <jes-o-mat> kylem: where did you upload your changes?
[06:58] <zul> the kernel.org hg tree sucks..
[09:01] <BenC> That's just dandy
[09:01] <BenC> vbox requires you to disable NMI watchdog on the host
[10:33] <kylem> eep.
[12:08] <Keybuk> random thought ... is it possible to re-exec a crashed process based purely on the memory map of the process?
[12:15] <kylem> Keybuk, how do you mean?
[12:16] <kylem> Keybuk, like, program fully terminates, dumps core. then some time x later you want to rerun it? or immediately re-execute it on crash?