NCommander | I did all that | 00:01 |
---|---|---|
TheMuso | NCommander: Ok. | 00:01 |
NCommander | Very strange | 00:01 |
TheMuso | Let me know when its ready to pull. | 00:01 |
NCommander | http://kernel.ubuntu.com/git?p=mcasadevall/ubuntu-jaunty-ports.git;a=summary - I pushed it, but I don't see the tag .... | 00:04 |
NCommander | TheMuso, oh, it seems tags aren't automatically pushed | 00:10 |
=== asac_ is now known as asac | ||
TheMuso | NCommander: aaaaahhhhh! | 00:12 |
NCommander | TheMuso, lets try that again | 00:13 |
TheMuso | This is a nice learning experience. | 00:13 |
NCommander | I dunno fi I call the lack of automatic pushing of tags a feature or a bug | 00:13 |
NCommander | Its uploading now | 00:13 |
TheMuso | Ok. | 00:13 |
TheMuso | I call it a bug personally, but thats just me. | 00:14 |
NCommander | SOmeone should note that you need to git push --tags :-) | 00:14 |
TheMuso | Yeah. | 00:14 |
NCommander | Its designed so if you pull someone elses branch, you don't get private tags from during release | 00:14 |
NCommander | wow, 10MB worth of tags | 00:14 |
NCommander | Linus has been busy ;-) | 00:14 |
TheMuso | Ouch. | 00:15 |
NCommander | (ubuntu-ports tags are obviously enough Ubuntu-ports vs Ubuntu so we can't accidently get the same tag) | 00:15 |
TheMuso | NCommander: Well you probably need to change the debian/rules files to look for those tags then. | 00:15 |
TheMuso | Otherwise the insertchanges et al commands will fail. | 00:15 |
NCommander | Probably | 00:15 |
NCommander | TheMuso, that change is in my branch, once 1.1 is released, it will be available for 1.2 (or 2.1) | 00:17 |
TheMuso | NCommander: ok. | 00:17 |
NCommander | TheMuso, TADA http://kernel.ubuntu.com/git?p=mcasadevall/ubuntu-jaunty-ports.git;a=summary | 00:24 |
NCommander | The changelog is amusingly long | 00:25 |
NCommander | and | 00:25 |
NCommander | I think incorrect | 00:25 |
NCommander | Damn it | 00:25 |
NCommander | Freaking insert changes pulled in stuff it shouldn't | 00:25 |
NCommander | actually, wait, it only grabbed 2.6.27->2.6.27.4 | 00:26 |
NCommander | Which is correct | 00:26 |
=== CarlFK1 is now known as CarlF1 | ||
TheMuso | NCommander: ok. | 00:27 |
NCommander | Ok, so just grab my branch, build a source package (with a orig.tar.gz), sign and upload | 00:27 |
TheMuso | Will do. | 00:28 |
TheMuso | NCommander: Do you want me to give it a test build on PowerPC hardware? | 00:53 |
NCommander | TheMuso, sure | 01:42 |
NCommander | (I tested it on my box, but I haven't tested it since fixing the other archs, but there were no code changes) | 01:42 |
TheMuso | Right. | 01:53 |
NCommander | TheMuso, if you do, save the abi folder | 01:53 |
NCommander | We'll need to generate those on all architectures from this release forward, since we'll have lbm and lrm soon | 01:53 |
TheMuso | NCommander: I'm building in an LVM snapshot with sbuild. | 01:53 |
NCommander | oh | 01:54 |
NCommander | Ok | 01:54 |
TheMuso | I can grab it during the build however if you like. | 01:54 |
NCommander | Its not a priority until we start getting lrm and module packages | 01:54 |
TheMuso | Is it created early on? | 01:54 |
NCommander | No | 01:54 |
NCommander | Towards the very end of the build | 01:54 |
TheMuso | Right. | 01:55 |
TheMuso | NCommander: actually I will build it manually, so I can get the ABI folder. | 02:00 |
* NCommander nods | 02:00 | |
TheMuso | NCommander: I thought there was a bug opened for ports kernel in intrepid to not use gcc 4.1 for powerpc, yet I see it in the build-deps. | 02:01 |
NCommander | I haven't cleared it yet | 02:01 |
NCommander | 1.2 build task | 02:01 |
TheMuso | ok | 02:02 |
NCommander | (doko wants to zap 4.1, once 1.1 is uploaded, that will be the first thing I do, since I need to make sure we don't have any nasty ices on any platform and still have working kernels) | 02:02 |
NCommander | i.e., I want a baseline to work against | 02:02 |
TheMuso | Right. | 02:03 |
TheMuso | NCommander: Kernels are cooking. | 02:34 |
NCommander | I thought I smelled something burning ;-) | 02:34 |
* NCommander runs from the bad pun | 02:34 | |
TheMuso | It still made me laugh however. | 02:34 |
NCommander | Yay, 7 packages pending uploading (with an eight pending successful pbuildering) | 02:35 |
* TheMuso slaps his head. I should have told it to use both CPUs to build faster... | 02:35 | |
NCommander | -j2 is broken | 02:35 |
TheMuso | Oh ok. | 02:36 |
TheMuso | That solves that then. | 02:36 |
NCommander | It kinda works | 02:36 |
NCommander | But you get a weird race condition which causes dpkg to explode | 02:36 |
TheMuso | Right. | 02:36 |
NCommander | I need to sit down and debug it at some point | 02:36 |
TheMuso | I wonder if that can be sorted out. | 02:36 |
NCommander | Enviornmental variable which just passed -j to the kernel build | 02:36 |
NCommander | WHich is what is already implemented | 02:36 |
TheMuso | Yeah I saw that. | 02:36 |
NCommander | (KERNEL_CONCURENCY_LEVEL or something) | 02:36 |
TheMuso | Yeah. | 02:37 |
TheMuso | NCommander: powerpc kernel cooked, powerpc-smp kernel now cooking. | 03:19 |
* NCommander holds the extinisher ready | 03:21 | |
TheMuso | lol | 03:22 |
TheMuso | NCommander: powerpc64-smp now cooking. | 03:37 |
NCommander | TheMuso, can you test run them? | 03:38 |
TheMuso | NCommander: As in try to boot from them? Sure. | 03:38 |
NCommander | TheMuso, since we're on the topic of things to be done | 03:50 |
NCommander | TheMuso, can I get a few sponsoring on other packages? | 03:50 |
TheMuso | NCommander: Sure. Shall we take it to a more appropriate channel? | 03:56 |
NCommander | motu or devel, your pick | 03:56 |
TheMuso | NCommander: doesn't bother me, whatever suits you. | 03:57 |
TheMuso | :p | 03:57 |
TheMuso | \/c | 03:57 |
TheMuso | Wow documentation takes a while. | 04:47 |
NCommander | TheMuso, yup | 04:50 |
* NCommander should have told you to debuild -B | 04:50 | |
TheMuso | Oh well. | 04:51 |
NCommander | TheMuso, did you get your zinc acocunt fixed? | 04:51 |
TheMuso | NCommander: Its in progress I believe. | 04:53 |
TheMuso | NCommander: kernels done. | 05:05 |
TheMuso | I can smoke test powerpc, and powerpc64-smp. I don't have a generic powerpc-smp box though. | 05:05 |
NCommander | now do they boot? | 05:05 |
TheMuso | Bout to try powerpc64-smp. | 05:05 |
NCommander | powerpc-smp will work fine on powerpc, or on a powerpc64-smp box | 05:06 |
TheMuso | Ok. | 05:06 |
TheMuso | NCommander: booting into powerpc64-smp. | 05:08 |
* NCommander crosses fingers | 05:08 | |
TheMuso | It boots. | 05:09 |
TheMuso | Ok sound works, all disks present and correct, hrm not sure what else to check. | 05:11 |
NCommander | Good enough for me | 05:16 |
NCommander | The powerpc kernel been tested | 05:16 |
TheMuso | testing powerpc-smp now. | 05:16 |
NCommander | TheMuso, just as a reminder, remember to leave my name on the changelog (on the bottom) so I get the build failure notices if any occur :-) | 05:16 |
TheMuso | NCommander: since you fixed the changelog, that is not a problem. | 05:17 |
TheMuso | hrm powerpc-smp doesn't want to boot. | 05:20 |
TheMuso | on my G5./ | 05:20 |
NCommander | coooooool | 05:20 |
TheMuso | yeah | 05:20 |
TheMuso | it brings up a screen with a lot of OpenFirmware stuff, or thats what it looks like. | 05:20 |
TheMuso | ok will try the powerpc and powerpc-smp kernels on my mini. | 05:22 |
NCommander | Hrm | 05:23 |
NCommander | the smp kernel might be hosed | 05:23 |
TheMuso | haha and the old boot option is broken. Time to dig out a desktop CD from hardy... | 05:23 |
NCommander | I'll have to smash it with a stick but unless I can find someone with a 32 bit SMP box | 05:23 |
NCommander | TheMuso, maybe the -smp kernel for some ungodly reason tries to put the G5 into little endian mode which would crash and burn | 05:26 |
TheMuso | possibly. | 05:28 |
TheMuso | NCommander: what makes you think that? | 05:29 |
NCommander | Its the only thing that could possibly cause the kernel to fault all the way back into open firmware | 05:29 |
TheMuso | NCommander: right | 05:29 |
NCommander | At least thats what I'm thinking | 05:29 |
NCommander | If it works on your mini, yay, if not | 05:29 |
NCommander | 1.2 :-) | 05:29 |
* NCommander hears a gunshot | 05:30 | |
* TheMuso slaps his head. | 05:30 | |
TheMuso | I was trying to boot old on my mini, not the G5. | 05:30 |
NCommander | oh | 05:31 |
NCommander | Your mini crashed and burned | 05:31 |
NCommander | Nice :-/ | 05:31 |
TheMuso | no | 05:31 |
NCommander | Wait | 05:31 |
NCommander | huh? | 05:31 |
NCommander | What machine failed to boot the smp kernel? | 05:31 |
TheMuso | When I tried to boot into another kernel on what I thought was my G5, it didn't work, but it was on my mini. | 05:32 |
NCommander | Oh I see | 05:32 |
TheMuso | the G5 failed to boot the powerpc-smp kernel. | 05:32 |
* NCommander is resmoketesting the -smp kernel | 05:32 | |
NCommander | I dunno why we have an non-smp/smp varient | 05:32 |
NCommander | Every other architecture is compiled to smp | 05:32 |
TheMuso | Yeah I know what you mean. | 05:33 |
NCommander | THere may be issues with the smp kernel on non-smp PPC hardware | 05:34 |
NCommander | THats the only reason I can think of | 05:34 |
NCommander | The G4 made it through a full startup with the SMP kernel | 05:34 |
NCommander | (its a single chip box) | 05:34 |
TheMuso | Right. | 05:34 |
NCommander | I think there must be hardware the SMP kernel crashes and burns with | 05:34 |
NCommander | TheMuso, I think we can write the issue as "Needs Investigation" and upload it | 05:35 |
TheMuso | Ok. | 05:35 |
NCommander | TheMuso, I just had a brainwave | 05:40 |
NCommander | TheMuso, your mini might have faulted because the kernel uses OpenFirmware to determine the presence of a second processor | 05:40 |
NCommander | Want to be on your mini crashed and burned due bad OF information? | 05:40 |
NCommander | *bet | 05:40 |
TheMuso | NCommander: it was the G5. | 05:41 |
TheMuso | NCommander: not the mini. | 05:41 |
NCommander | Oh! | 05:42 |
NCommander | same point applies ;-) | 05:42 |
TheMuso | Right. | 05:42 |
TheMuso | But powerpc64-smp worked. | 05:42 |
NCommander | and powerpc-smp worked on non-smp hardware | 05:42 |
NCommander | We need someone with an SMP G4 | 05:43 |
TheMuso | Yep. | 05:43 |
TheMuso | ok powerpc works on the mini | 05:47 |
NCommander | so either smp is broken on smp hardware :-/, or smp is broken on smp64 | 05:49 |
TheMuso | NCommander: NCommander Well 64-smp is fine, as I am now back on that with my G5, and both CPUs are showing. | 05:52 |
NCommander | I think infinity or lamont have to have a multicore non-SMP G5 | 05:52 |
NCommander | can you boot a powerpc kernel on your G5? | 05:53 |
NCommander | (maybe its a general 64-bit issue vs. smp) | 05:53 |
TheMuso | NCommander: the powerpc kernel non-smp boots fine. | 05:53 |
NCommander | ok | 05:53 |
NCommander | We'll have to mark that on the Needs Debugging list | 05:53 |
TheMuso | NCommander: powerpc-smp boots on mini. | 05:59 |
NCommander | woooo | 06:00 |
NCommander | so its just a 64-bit issue | 06:00 |
TheMuso | sounds like it. | 06:01 |
NCommander | do we care enough to fix it at this very moment? | 06:02 |
TheMuso | I don't. | 06:03 |
* NCommander doesn't | 06:06 | |
TheMuso | NCommander: i have some things to do so I'll look further into this later. | 06:13 |
NCommander | TheMuso, the smp crash, or uploading ;-)? | 06:13 |
TheMuso | NCommander: uploading. | 06:13 |
NCommander | ok, no | 06:13 |
NCommander | *np | 06:13 |
NCommander | ah, abogani | 07:32 |
abogani | NCommander: Hi | 07:32 |
NCommander | abogani, I've been working on linux-rt-powerpc, and I found a rather disturbing problem | 07:33 |
NCommander | The current rt kernel in Intrepid isn't actually RT :-/ | 07:33 |
abogani | :-? | 07:33 |
NCommander | (CONFIG_PREEMPT_RT is not defined) | 07:33 |
NCommander | and I can be blamed for it | 07:34 |
* NCommander whacks head on the desk | 07:34 | |
abogani | No way: it is defined. | 07:34 |
NCommander | Nope | 07:35 |
NCommander | abogani, here's the debdiff | 07:35 |
NCommander | oh wait | 07:36 |
NCommander | bah | 07:37 |
NCommander | Screw me, I can't spell "PREEMPT" | 07:37 |
NCommander | morning smb_tp | 08:05 |
smb_tp | Good morning NCommander | 08:06 |
NCommander | smb_tp, how goes your morning | 08:06 |
smb_tp | NCommander, Good so far. :) Just got the first coffee. Saw your mail | 08:07 |
NCommander | smb_tp, TheMuso will be uploading the first ports kernel soon, and I'm bending linux-ports-lrm into shape :-) | 08:08 |
smb_tp | NCommander, Sounds great. :) I guess I will give the mail some time today to be seen by others and then pull it to the repo. | 08:10 |
NCommander | \o/ | 08:11 |
NCommander | smb_tp, if you have sometime at some point to sit down with the ia64 config files, and figure out what things should be built as modules, it would be apperiated because then I can fix d-i for ia64 | 08:11 |
smb_tp | NCommander, I only had a config that compiled (but had to discard serial-modules) and produced some udebs but whether this really is useful for the installer, I don't know. | 08:16 |
NCommander | smb_tp, someone who knows the hardware needs to sitdown and smash the kernel into shape :-/. Even HPPA has working d-i/udeb support and thats the architecture thats usually broken | 08:16 |
smb_tp | NCommander, True. I will try to ask BenC whether he got a little time to help. At least he has the HW | 08:18 |
NCommander | Oh, I thought you had ia64 | 08:18 |
NCommander | d'oh | 08:18 |
smb_tp | NCommander, No, I only have access to one (but only as user, so I can't install new kernels) | 08:20 |
NCommander | oh right | 08:20 |
=== lamont` is now known as lamont | ||
nodeboy_999 | hello | 14:12 |
nodeboy_999 | does apt-get <linux-kernel> provide a static or dynamic kernel image for any given kernel? | 14:12 |
=== smurfix is now known as smurf | ||
drunkenkilla | hello | 15:04 |
drunkenkilla | i got a problem | 15:04 |
drunkenkilla | i can't change the brightness...the fn-keys and the applet in the panel doesn't work | 15:05 |
drunkenkilla | i tested even now pre-release of fedora 10 and i could change the brigthness with the applet in the panel | 15:05 |
mjg59 | F10 implements opregion support which is required for backlight control on a bunch of Intel graphics laptops | 15:15 |
mjg59 | It'll be in 2.6.28 of the upstream kernel, but it's not in Intrepid | 15:15 |
psusi | isn't there a libata option somewhere for verbose debugging output? | 17:07 |
rtg | lamont: https://edge.launchpad.net/ubuntu/+source/linux-backports-modules-2.6.27 has finished building. this time it should _really_ fix your b44 issue. | 17:11 |
lamont | rtg: and which suite does it show up in, I wonder? | 17:12 |
lamont | intrepid-proposed? | 17:12 |
rtg | lamont: yes. | 17:13 |
lamont | kewl | 17:13 |
rtg | lamont: adventurous souls such as yourself should always have -proposed enabled :) | 17:14 |
lamont | heh. not. | 17:14 |
lamont | well. maybe pinned at 25 :-) | 17:14 |
BenC1 | rtg: I'm worried about the patch that fixes backlight key...it's not in upstream yet | 18:05 |
drunkenkilla | BenC1: which patch? i have problems too witch backlight | 18:06 |
BenC1 | drunkenkilla: what's the problem you are having? | 18:07 |
=== BenC1 is now known as BenC | ||
drunkenkilla | i can't change the brightness | 18:07 |
drunkenkilla | i'm using ubuntu 8.10 and i can't change the brightness with the fn-keys and the applet in the panel | 18:08 |
BenC | Could be fixed by this patch | 18:08 |
drunkenkilla | but today i tested pre-release of fedora 10 and i could change the brightness with the applet, but not with the fn-keys | 18:08 |
BenC | The case I'm fixing is where the brightness key causes two key press events | 18:09 |
drunkenkilla | i will show you dmesg... | 18:09 |
drunkenkilla | BenC: http://paste.ubuntu.com/67995/ | 18:10 |
BenC | drunkenkilla: full dmesg would be better | 18:13 |
drunkenkilla | BenC: http://paste.ubuntu.com/67998/ | 18:15 |
drunkenkilla | re | 18:29 |
drunkenkilla | BenC: when will come this patch? | 18:50 |
BenC | drunkenkilla: not sure | 18:51 |
mjg59 | BenC: You get an ACPI key event and an event from the keyboard? | 18:52 |
psusi | what did you have to do to see KERN_DEBUG printks again? | 19:46 |
Pici | Howdy kernel people. It looks like the current kernel compiling help on help.u.com assumes that the user is running Intrepid. I don't really know enough about kernel compiling to split out the directions, but this is the wiki diff that appears to be the culprit: https://help.ubuntu.com/community/Kernel/Compile?action=diff&rev1=69&rev2=61 | 20:05 |
Pici | linux-kernel-devel is not on Intrepid and makedumpfile isnt on anything pre intrepid | 20:06 |
BenC | mjg59: yeah | 21:02 |
NCommander | hey BenC | 21:11 |
BenC | NCommander: hey | 21:45 |
NCommander | BenC, did you see my email to the list? | 21:46 |
mjg59 | BenC: I suspect gnome-power-manager is the right place to handle that | 22:00 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!