[12:16] Micksa: Rock [12:16] Micksa: vbetool is probably what you need === doko [~doko___@dsl-082-082-201-086.arcor-ip.net] has joined #ubuntu-kernel [05:33] mjg59: the PCI config for the card gets trashed [05:40] I believe it's PCI express [05:43] vebtool doesn't appear to help [05:52] okay, so maybe I'll have to read the PCI spec 8) [05:57] g'night all [06:27] morning === dilinger waves [06:31] hey dilinger [06:47] lamont: i guess you did commit to your local tree, didn't you? [07:12] dilinger: btw did you ever tested clvm ? [07:13] (since you are one of the lvm2 folks ;)) [07:14] nope [07:14] i haven't played w/ GFS stuff at all [07:15] you don't need GFS to play with clvm :) [07:16] i guess i am going to play with it for you soon [07:17] oh. well, no, i've only dealt w/ basic lvm [07:17] clvm is cluster lvm, no? [07:17] since redhat ate sistina, and joe thornber stopped working there, i haven't really kept up. they're no fun anymore :) [07:18] eheh ok [07:18] yeah it's cluster lvm [07:25] i never really had the opportunity to play w/ it [07:25] oky doky :) [07:25] sistina freed GFS around the time i quit voxel, i believe [07:25] s/sistina/redhat/ [07:25] yeah [07:25] and i don't have the hardware at home [07:25] and my current job is big on openafs [07:25] you need 3 pc to do that [07:25] s/pc/boxes/ [07:25] i didn't have 3 working pcs :) [07:26] i ended up buying a whole bunch of parts, i had a bunch of dead machines [07:26] and sparc boxes [07:26] well you can mix hardware... [07:26] i wouldn't quite trust the sparcs w/ stuff like that [07:27] why not? [07:28] 'cause i'm happy if i can get the things to boot; clustering seems like an extravagancy [07:28] well but testing doesn't mean production :) [07:29] btw.. did you finally got your appartment? [07:29] so that we can turn one of your sparcs into a buildd? [07:29] yep [07:29] cool :) [07:29] i moved in last weekend [07:29] still making it habitable [07:29] congratulation :) [07:29] yeah i can imagine [07:29] thanks :) [07:30] the 280's sitting on the floor, i need to get all my machines up and running still [07:30] no rush [07:30] i'm not sure what i'm going to do about the internet connection [07:30] why? [07:30] i tried contacting roadrunner, but they haven't gotten back to me [07:30] in the meantime, someone in the building has an open AP that i've been using [07:30] are you still without net at home? [07:30] ahhh [07:30] ehehhe [07:32] i'll try to get it online sometime this week, though [07:32] well without bw it might be an issue [07:33] maybe i should see if i can host it at work [07:33] although getting it there will be.. interesting [07:33] and i should probably move wanna-build to a public host [07:34] dilinger: for the size? :) [07:34] they're also firewalled like crazy; does anything need to connect to the buildd, or does the buildd do all the connecting out [07:34] ? [07:34] yea, the size and weight. i certainly couldn't move it myself [07:34] buildd connects to outside [07:34] nothing connect to buildd [07:34] carrying from my car to my room, when i first got it, nearly killed me [07:35] ehe i can imagine [07:36] buildd needs ssh access to wanna-build [07:36] and clearly access to a public sparc archive [07:36] ok [07:36] like http://ports. [07:36] so it ould be fine to host it at work [07:37] i guess there is no way to ssh in, is it? [07:37] i'd have to set up a tunnel [07:37] which wouldn't be permanent [07:38] hmmmm [07:38] of course, if you can find me hosting in nyc somewhere, i can put it there instead.. [07:39] it was more to be able to do admin tasks, like updating chroot, unfucking chroots [07:39] yeah but i know shit about hosting in NYC [07:39] it's like 4000 Miles from here if not more :) [07:41] heh [07:42] building iseries kernel is a royal pain in the butt :) [07:42] hm, my laundry's probably done.. [07:43] i need to wake up my wife and listen to her for the next hour... [07:44] given that we have been having a fight all night [07:44] i would rather do your laundry :) [07:44] ttyl === Seveas [~seveas@seveas.demon.nl] has joined #ubuntu-kernel [01:40] fabbione: no, I simply failed to actually do the commit. :-( [02:43] fabbione: Would it be possible to set CONFIG_ACPI_DEBUG to y for a while? [02:50] lamont: ah ok :) [02:50] mjg59: there will be 2 -dbg images with the next upload [02:50] 686-dbg and 386-dbg with all the DEBUG=y [02:50] i guess that includes ACPI too [02:51] fabbione: Oh, rock [02:52] fabbione: Did you get my mail? [02:52] mjg59: for the libata suspend/resume patch is not really go.. we will drah 2.6.12 into main and as default kernel within the next 2 weeks :/ [02:52] fabbione: This code needs testing. It's not going to get tested unless we put it somewhere [02:52] And it's vital for ACPI on new laptops [02:52] mjg59: i need to be at least sure it's not going to eat people disks [02:53] It /looks/ fine [02:53] make it twiddlable by a boot parameter? [02:53] The only time this code patch is followed is during suspend or resume [02:53] Uh, path [02:53] It won't affect any other use [02:53] ah it's from Jeff [02:54] i will ask him [02:58] hmm he is not online [02:58] fabbione: ok, I have an oops now. What do you want me to do with it? [02:58] Mithrandir: stick it somewhere i can look at it? [02:59] or the best option would be to print it in 4 layers toilet paper for firther usage :) [02:59] fabbione: I mildly hate you then, since it means I have to copy half a screenfull off by hand. Or do you do digicam shots? :) [02:59] lamont: can you commit asap? i might upload the kernel either today or tomorrow my morning (your late evening) [02:59] fabbione: http://lkml.org/lkml/2005/5/23/116 is the disclaimer [02:59] Mithrandir: it needs to be the full OOPS [02:59] But various people are using it without complaint so far [03:00] fabbione: sure, but is a picture of it ok or do you really want the text? [03:00] digikam is ok if fully readable [03:00] ok, fine. [03:01] fuck ppc and its random process killer [03:01] it should have finished the ppc64 build while i was asleep [03:09] mjg59: Patch 2/2 for review by the linux-scsi crowd. [03:09] where is 1/2 ? [03:10] nver mind [03:11] 509700 build-power4-smp [03:11] 561520 build-powerpc64-smp [03:11] YEAH [03:11] elmo and the mirrors will love me :) [03:13] fabbione: 1/2 is unrelated [03:13] yup [03:18] fabbione: http://err.no/tmp/DSC00879.JPG [03:20] Mithrandir: ok... [03:21] it's probably the biggest backtrace you've seen, at 2.7MB. :P [03:21] ahhaha [03:26] Mithrandir: is that card usb or pci? [03:26] pci [03:26] ok [03:28] * acx_ether_to_txdesc [03:28] * [03:28] * Uses the contents of the ether frame to build the elements of [03:28] * the 802.11 frame. [03:28] * [03:28] * We don't actually set up the frame header here. That's the [03:28] * MAC's job. We're only handling conversion of DIXII or 802.3+LLC [03:28] * frames to something that works with 802.11. [03:30] Mithrandir: given that the driver is compiled with DEBUG [03:30] can you show me the logs from kern.log or syslog? [03:30] they should be there somewhere [03:30] fabbione: the box is hung when that oops happens. [03:31] Mithrandir: yes, but the logs should be written before that :) [03:31] they aren't. Or wasn't last time. [03:32] WTF!!!!1 [03:32] I HATE EXTERNAL DRIVERS [03:32] or, which part of the logs? [03:32] THEY SHOULD DIE OF SLOW PAINFUL DEATH [03:32] acxlog(L_DEBUG, "tx: 802.3 len: %d\n", skb->len); [03:32] there should be entries like that around [03:33] acxlog(L_DEBUG, "tx: DIXII len: %d\n", skb->len); [03:33] no tx: from today. [03:34] do you actually get any data out of the device before the crash? [03:34] data like IP connections [03:35] sure, I can ping my router, for instance. [03:36] ssh-ing made it die. [03:37] Mithrandir: was that piece of code working in 2.6.10? [03:37] (can't remember if you told me) [03:37] fabbione: 2.6.10 works just fine on the box, yes. [03:37] ok [03:37] the driver is in the kernel and I haven't reviewed any changes. [03:38] yeah the acx100 is anyway an external one [03:40] UHAAAA [03:41] the code has been changed hell of a lot [03:41] - e_snap = (wlan_snap_t*)((UINT8*)e_llc + sizeof(wlan_llc_t)); [03:41] + e_snap = (wlan_snap_t*)((u8*)e_llc + sizeof(wlan_llc_t)); [03:41] stuff like this might make it amd64 unfriendly :) [03:41] Mithrandir: do you have any option to test it on i386? [03:42] I can install i386 on the box and try, sure. [03:42] it'll take me an hour or so, though. [03:42] but, food now [03:43] sure no rush [03:43] but if it works on i386.. well there is only one solution [03:43] bomb upstream :) [03:47] blah [03:52] hey zul [04:03] dpkg-deb: building package `linux-image-2.6.12-1-powerpc64-smp' in `../linux-image-2.6.12-1-powerpc64-smp_2.6.11.93-1.2_powerpc.deb'. [04:05] dpkg-deb: building package `linux-image-2.6.12-1-iseries-smp' in `../linux-image-2.6.12-1-iseries-smp_2.6.11.93-1.2_powerpc.deb'. [04:05] there we go [04:07] /usr/share/kernel-wedge/commands/copy-modules: line 13: 14622 Segmentation fault mv $tmpdir/work.new $tmpdir/work [04:07] YEAHH!!!!!! [04:07] GO KERNEL-WEDGE IT'S YOUR BDAY! [04:08] i am afraid kernel wedge needs some ppc64 love too [04:09] man [04:09] I wanna hack kernels all day [04:09] I can do that [04:10] you don't believe me huh :) [04:21] Micksa: there are tons of bugs that need love :) [04:21] welcome to provide patches to fix them [04:21] there are always bugs [04:22] it's mostly a time issue [04:22] well, uh, also [04:22] do you work for canonical? [04:28] yes [04:35] fabbione: i wont be around much tomorrow since i have an interview === Micksa [~mslade@203-217-18-166.perm.iinet.net.au] has joined #ubuntu-kernel [04:36] crap [04:37] lemme try that again [04:37] fabbione: do you work for canonical? [04:38] Micksa: yes [04:38] zul: no problem.. i am fighting with ppc64 [04:38] that must be fun [04:38] see, you're getting paid to do this stuff :) [04:38] I could SO do that [04:38] but man I'm crap at making good use of spare time [04:39] I mean, shit, I spent all day yesterday trying to get my laptop working 8) [04:39] oh and linux-non-supported-modules is coming along nicely [04:39] zul: it was fun until i hitted the random PPC segfaults [04:39] he [04:39] hehe [04:39] zul: ah nice [04:39] i guess i am going to try to hack on 64 images only [04:39] building 8 images takes too long [04:40] it does...usually yes :) [04:40] Micksa: well up to you to get known.. a good point to start is MOTU :) [04:40] *sigh* i have to wear a tie tomorrow [04:40] amen [04:40] i hate ties [04:40] MOTU? [04:41] Micksa: #ubuntu-motu [04:41] Master Of The Universe [04:41] it's a good point to start [04:42] mmmkay === Micksa commences lurking [04:44] do we have a list of which external-drivers run on which arch [04:44] yet [04:48] zul: no, you will need to grab the images from archive and ports [04:48] and check them there [04:48] thats what i thought [04:53] i have the feeling that ppc buildd will never make a kernel build [04:54] ppc has this nice issue of doing random segfaults [04:54] it did it 4 times in one build [04:54] meh...drop ppc :) [04:54] hehehe [04:55] with Apple switching to intel, it might be an option :) [04:56] fabbione: I'll believe that when I see it. [04:56] Mithrandir: well it was on /. yesterday [04:56] fabbione: "IT WAS ON SLASHDOT. OH GOD IT MUST BE TEH TRUE!" [04:57] :-) [04:57] ahhaha [04:57] AH SHIT..ok i am too tired [04:57] time to stop for today [04:57] the OCFS2 update is FTBFS [04:57] and of course i can't build ppc64 anymore [04:57] what slashdot always reports the truth? [04:57] that makes all my d-i changes pointless [04:58] fs/configfs/inode.c:51: error: unknown field `memory_backed' specified in initializer [04:58] BAH === Mithrandir installs i386 breezy kernels [05:11] bah fixed [05:23] fabbione: well, if I switch to i386, it doesn't associate at all, even. :P [05:24] Mithrandir: fun :) bug upstream :) [05:24] or check if there is a newer version [05:24] fabbione: it's the stock kernel driver. [05:25] 2.6.12 doesn't have lrm [05:25] :P [05:26] um...is the hotplug stuff the same from the old version and the new version of acx100 === Mithrandir tries to parse zul's sentence a few times an fails. [05:27] s/an/and/ [05:28] bah...the acx100 driver we are using is from a different source has the *.bin files in the hotplug directory the same as the 2.6.10 version? [05:29] http://www.tsn.ca/columnists/bob_mckenzie.asp [05:29] oops....wrong one [05:29] http://rhlx01.fht-esslingen.de/~andi/acx100 [05:30] I had to symlink in the firmware from 2.6.10 since the kernel didn't find it otherwise. [05:30] wierd [05:31] hmmmmmmmmmmm [05:31] HMMMMMMM [05:31] i think the acx100 firmwares are in l-r-m [05:31] they probably need to be updated [05:31] yes they are [05:31] the acx100 hotplug code from 2.6.10 to 2.6.12 is different [05:32] bbl need to eat [05:33] bah ocfs2 update is fucked === fabbione reverts [05:33] hm, ok. [05:35] except there's no newer firmware upstream? [05:39] hm, I'll see if compiling the newest and shiniest external driver works. [05:41] is there a new version? [05:42] acx100 | 0.2.0pre8+fixes52 | ok | 18/04/2005 | http://acx100.sourceforge.net/ http://rhlx01.fht-esslingen.de/~andi/acx100/ [05:42] this is the one in the kernel right now [05:42] anyway i need to go and cook dinner [05:42] ppc64 is pissing me off too much [05:43] yes, there's a newer upstream [05:44] fixes56 [05:44] Mithrandir: ok.. test that one and let us know :) [05:44] we will update the driver anyway, but if it fixes the problem i will find the time to include it in the next release [05:44] i am off :) === Micksa [~mslade@203-217-18-166.perm.iinet.net.au] has joined #ubuntu-kernel [09:29] * committed kernel-team@ubuntu.com--2005/kernel-debian--pre1,2--2.6.11.93--patch-25 [09:29] that was either last night or first thing this morning... [09:29] almost 8 hours ago === Seveaz [~seveas@seveas.demon.nl] has joined #ubuntu-kernel