[12:35] <BenC> fabbione: Not sure if the ones that have built-in firewire can boot from sbp2 or not
[12:36] <BenC> fabbione: Actually, I think it is possible...I remember my Blade 100 had sbp2 support in firmware
[12:36] <fabbione> BenC: hmmm
[12:36] <fabbione> BenC: ok.. people can file bugs.. for now i am happy to support scsi/ide/sbus
[02:13] <zul> good luck
[02:22] <bronson> BenC: what are your -updates trees? Are they a place to stage patches before merging into the dist tree?
[02:23] <bronson> i.e. patches are baked in ubuntu-edgy-updates.git, then pulled into ubuntu-edgy.git when it's time to release a new kernel?
[02:48] <BenC> bronson: They are in -updates, and supposed to be put into -proposed where the get tested before inclusion in the main tree
[03:09] <bronson> BenC: so, if I want to keep my linux-vserver kernel on top of Edgy development, I should periodically rebase it on top of -updates.  That way, when you do build a new kernel, my derivative kernel will be pretty much ready to go?
[03:09] <bronson> or..  maybe I should track -proposed?
[03:10] <BenC> keep it on edgy proper
[03:10] <BenC> -updates can be broken at any time
[03:10] <bronson> Ah good.  OK, will do.
[12:50] <gnomefreak> BenC: are you here? initramfs-tools was unable to overwrite
[12:50] <gnomefreak> is it safe to force the overwrite like i would with any other package?
[12:55] <gnomefreak> btw no its not its missing depends (seems to be a bunch of crap that is stopping it
[02:45] <fabbione> BenC: davem has a firewire device here.. we will get the OBP mapping today
[02:45] <fabbione> he also thinks we can boot from it
[02:45] <fabbione> but only if the device is plugged in at poweron
[02:45] <BenC> fabbione: Yeah, I think I tested that a loooong time ago
[02:46] <BenC> I'm pretty sure you can, if you have the latest OBP for the box
[02:46] <BenC> fabbione: Write my name on that firewire drive too, so dave remembers where it came from :)
[02:46] <fabbione> ahaha
[02:47] <fabbione> it's just a shell script to parse the obppath sysfs attribute really
[02:47] <fabbione> did you try booting your e3k with the new patched?
[02:47] <fabbione> patches
[02:47] <fabbione> find /sys -name "obppath*"
[02:50] <BenC> not yet, I'll give it a try later
[02:50] <fabbione> ok
[03:33] <maks_> BenC thanks for the merge
[03:34] <maks_> i seem to have posted my review of your changes to the wrong channel, will recap
[03:35] <maks_> * check_minkver() lives in hook-functions you should have it twice now
[03:35] <maks_> * +rm -f /etc/initramfs-tools/modules is doubled in postrm
[03:36] <maks_> * ubuntu used to use busybox mount for nfs
[04:49] <BenC> maks_: Ah, thanks
[05:08] <kylem> BenC, i merged my -updates and the dapper security stuff, what do you think of a -proposed upload?
[05:10] <BenC> kylem: I'll look at it in just a couple of minutes
[05:15] <BenC> maks_: Can you merge the s/Build-Depends/Build-Depends-Indep/ change?
[05:16] <BenC> maks_: And if you could add " | busybox-initramfs" to the depends, that would fix all the debian/control merges
[05:18] <maks_> BenC remind me why is the s/Build-Depends/Build-Depends-Indep/ needed?
[05:19] <maks_> yes i'll add that to the depends list
[05:19] <BenC> maks_: Because initramfs-tools is binary-indep, so the build-depends are actuallt for the binary-indep portion of the build
[05:19] <BenC> On the Ubuntu build, we only upload source
[05:19] <BenC> so the binary-indep portion gets built on i386
[05:20] <BenC> in order for the buildd to process that correctly, the build-deps need to reflect reality
[05:20] <maks_> ok, thanks for the info
[05:21] <BenC> maks_: One other question I had was the {-q,-Qb} modprobe inconsistencies we have
[05:21] <BenC> I think the -b makes it respect modprobe.d blacklists
[05:21] <maks_> debian modprobe didn't have that arg irc
[05:21] <BenC> and the -Q makes it quiet about ignoring a blacklisted module (which -q wont do)
[05:22] <BenC> the man page doesn't show it, but "modprobe -h" will show it
[05:22] <maks_> no it's not there
[05:22] <BenC> it has to be, unless you guys have a super old modprobe
[05:22] <BenC> what version of module-init-tools?
[05:23] <maks_> ii  module-init-to 3.3-pre3-1
[05:23] <BenC> is it -Q or -b that it doesn't have?
[05:23] <maks_> -Q
[05:24] <BenC> ah, right, it's an ubuntu patch it seems
[05:24] <BenC> maybe -qb then
[05:24] <maks_> ok
[05:24] <BenC> the -b is the important part, because users need to be able to blacklist modules and have initramfs scripts respect that (e.g. for the case of buggy force loaded modules)
[05:25] <maks_> fully agreed
[05:25] <BenC> maybe you could just export one MODPROBE_OPTIONS in /init
[05:26] <BenC> modprobe will respect it, and that'll localize any changes we have to make for -Q
[05:30] <BenC> maks_: In scripts/init-top/framebuffer, the VESA and SPLASH vars aren't used anywhere...I thought it was a leftover Ubuntu patch, but it's in your source
[05:31] <maks_> # Set modprobe env
[05:31] <maks_> MODPROBE_OPTIONS="-qb"
[05:31] <maks_> export ${MODPROBE_OPTIONS}
[05:32] <maks_> BenC: indeed, kick those
[05:33] <BenC> maks_: Just uploaded ubuntu4 with the changes you mentioned, thanks.
[05:34] <maks_> cool i get the diff, i'm monitoring :)
[05:43] <maks_> BenC: the discussed changes on my side are in 0.85f, need to push 0.85e first into testing
[05:46] <maks_>   * debian/control: Change Build-depends-indep to Build-depends as we need
[05:46] <maks_>     debhelper and cdbs for the clean target, fulfills policy 7.6.
[05:48] <maks_> so
[05:48] <maks_> not sure if i want to pick that up
[05:53] <BenC> hmm
[05:53] <fabbione> BenC: get ready for ubuntu5 ? :)
[05:54] <BenC> fabbione: Might just be 0.85fubuntu1 :)
[05:55] <fabbione> ehe
[05:56] <BenC> ubuntu5 uploaded...glad this isn't a heave package to build
[05:56] <BenC> heavy
[05:57] <fabbione> BenC: are we sure it doesn't have lvm or mdadm stuff?
[05:58] <BenC> it has an mdrun that only gets used if mdadm isn't also there
[05:58] <Mithrandir> BenC: fubuntu < ubuntu. :-P
[05:59] <fabbione> hmmm
[06:00] <BenC> fabbione: you asked me to remove it :P
[06:00] <BenC> I don't have any machines with a lvm/md root to test
[06:00] <fabbione> lvm is ok.. must go
[06:00] <fabbione> i am not sure what that mdrun can do without mdadm installed
[06:01] <fabbione> i guess i will have to clean up that clusterfuck when i am back in january
[06:01] <fabbione> and the first guy that's gonna complain about the abuse of the word clusterfuck will get to solve all the mess
[06:02] <BenC> fabbione: it runs if there is no local-top/mdadm, but it exects mdadm binary to be present if raid is required
[06:02] <fabbione> ok, then it's totally useless :P
[06:02] <fabbione> once you install mdadm you get local-top/mdadm :)
[06:03] <BenC> probably debian/ubuntu friendly
[06:04] <fabbione> probably
[06:04] <fabbione> or probably due to lack of communication between the 3 developers that can't agree on where these scripts should exists
[06:04] <BenC> maks_: Am I correct that the resume stuff in initramfs supports suspend2?
[06:04] <fabbione> because our mdadm is the same as debian
[06:49] <zul> yeah so one of the managers here at work changed the locks fo the server room without telling us
[06:49] <kylem> i think that means you're fired
[06:49] <kylem> or evicted
[06:50] <zul> yay... just after i got my stock options
[07:50] <maks_> fabbione, BenC: the mdrun hook is for partial upgrades from sarge -> etch, where you have initramfs-tools but not yet the new mdadm
[07:50] <fabbione> maks_: kthx.. BenC we can kill it.. we don't support partial upgrades anyway
[07:51] <maks_> yes there should be some code in mkinitramfs that can go as well
[07:51] <maks_> maraked as to be removed after the etch release
[07:52] <maks_> BenC: afaik suspend2 has it's own i-t boot hook, but i might be wrong never used that crap