| berto- | following the instructions on the following page i was able to build my own kernel. after rebooting i'm able to SSH into the box, but there is no console (TTY1). any ideas? https://help.ubuntu.com/12.04/installation-guide/i386/kernel-baking.html | 02:19 |
|---|---|---|
| * apw whines about the early hour | 09:18 | |
| stgraber | have to get used to it again :) | 09:19 |
| * apw reboots into a 2013 kernel ... | 09:26 | |
| === yofel_ is now known as yofel | ||
| === amitk is now known as amitk-afk | ||
| === chiluk_away is now known as chiluk | ||
| === amitk-afk is now known as amitk | ||
| * ogasawara stretches and waves | 13:38 | |
| apw | hny ogasawara :) | 13:55 |
| * rtg has accumulated a simply staggering amount of email in 2 weeks | 13:56 | |
| ogasawara | my inbox was actually manageable, although I was staying on top of it over the break | 13:57 |
| rtg | sforshee, do I want BRCM_TRACING=y for 3.8 ? | 15:02 |
| sforshee | rtg, yes, let's have BRCM_TRACING=y and BRCMDBG=n | 15:05 |
| rtg | sforshee, ack | 15:05 |
| sforshee | rtg, I think there were some mac80211 trace options I wanted to enable too but I can send a patch | 15:05 |
| rtg | sforshee, wait until I've got the rebase complete. I'll send out a note | 15:06 |
| sforshee | rtg, ack. I doubt I'll get to it before then end of the day anyway, most of today will be playing catch-up. | 15:07 |
| rtg | bjf, 3 of us responded.... | 15:35 |
| bjf | rtg, i heard a "yup" | 15:35 |
| rtg | bjf, ok, its hard to tell when _you_ don't respond | 15:35 |
| * ogasawara back in 20 | 15:53 | |
| === JanC_ is now known as JanC | ||
| slangasek | cking: ping - did you see my request in bug #1040557 to push your samsung-laptop changes to precise-proposed so that we can build live images from there and test the fix? | 16:40 |
| ubot2 | Launchpad bug 1040557 in Ubuntu CD Images "UEFI boot live-usb bricks SAMSUNG 530U3C,np700z5c laptop" [Critical,Confirmed] https://launchpad.net/bugs/1040557 | 16:40 |
| bjf | slangasek, i'm working on the next round of kernels as we "speak" | 16:48 |
| slangasek | bjf: is there a commit from cking somewhere that you can perhaps pull in for this? | 16:49 |
| bjf | slangasek, it's already in the tree and will be part of this update | 16:49 |
| slangasek | oh, awesome - thanks! | 16:49 |
| cking | ..the crank is turning so to speak | 16:51 |
| bjf | indeed | 16:51 |
| rtg | ogasawara, pushed raring master-next. build tested on amd64/i386. still working on armhf. updated dropped.txt with the usual suspects. | 16:52 |
| ogasawara | rtg: ack, thanks. I'll give it a spin on my kit. | 16:52 |
| * rtg -> lunch | 18:32 | |
| rtg | ogasawara, repushed raring with some version fixes. its now 3.8.0-0.1 | 19:06 |
| ogasawara | rtg: ack | 19:06 |
| infinity | Ooo, 3.8? | 19:14 |
| infinity | That fixes my tmpfs leak. Want. | 19:14 |
| rtg | it even seems to work so far | 19:14 |
| === chiluk is now known as chiluk_away | ||
| === chiluk_away is now known as chiluk | ||
| === chiluk is now known as chiluk_away | ||
| === chiluk_away is now known as chiluk | ||
| dobey | if i clone the git repo for the kernel, and do debuild -S, where does it get the orig source from? | 20:16 |
| * ogasawara lunch | 20:21 | |
| * rtg finally catches up on LKML | 20:25 | |
| * rtg -> EOD | 20:31 | |
| kamal | dobey: it doesn't get the orig source from anywhere ... you need to manually download it and place it in "../" (right above the git repo) before you do your debuild -S | 21:02 |
| dobey | it appears a bit more complex than that | 21:02 |
| kamal | dobey: if you don't do that, then it will build a *full* source package | 21:02 |
| dobey | debuild -S just errors | 21:02 |
| dobey | because there's no debian/changelog (it's under debian.master/) | 21:02 |
| kamal | dobey: you must do "fakeroot debian/rules clean" first ... that'll setup debian/ | 21:03 |
| dobey | ah | 21:03 |
| kamal | (i.e. it'll copy bits from debian.master) | 21:03 |
| kamal | dobey: fakeroot debain/rules clean is poorly named ... it would be better called "clean_and_prepare" | 21:03 |
| kamal | dobey: another tip: for the kernel tree in particular, you'll want to use: debuild -S -I -i (to make it not try to package up the .git/ dir and other junk bits) | 21:04 |
| dobey | kamal: and what would the orig source be, given there's no debian/patches directory, and the kernel has patches which aren't in the upstream orig source? | 21:07 |
| dobey | i'm trying to understand how i can add some patches to it, and build a package, so i can test them | 21:08 |
| kamal | dobey: the orig source is in fact, the upstream kernel release tarball, e.g. linux_3.5.0.tar.gz but renamed to linux_3.5.0.orig.tar.gz as is normal for debian packages | 21:11 |
| kamal | dobey: you can download that .orig.tar.gz file from Launchpad ... | 21:12 |
| kamal | https://launchpad.net/ubuntu/+archive/primary/+files/linux_3.5.0.orig.tar.gz | 21:13 |
| kamal | (modify as needed for your version) | 21:13 |
| kamal | dobey: there's no debian/patches directory because this particular debian package just doesn't use debian/patches -- instead, all the changes (the delta from that .orig) are just glommed into another big tarball at the time you run debuild -S | 21:14 |
| kamal | dobey: for your case, you should just *apply* the patches directly to your git repo, then add a new "customdobeyversion" entry to debian.master/changelog, then fakeroot debian/rules clean (to copy that into debian/), then debuild -S . . . | 21:16 |
| kamal | you'll end up with a new customdobeyversion source package in ../ | 21:16 |
| === chrisccoulson_ is now known as chrisccoulson | ||
| === henrix_ is now known as henrix | ||
| === henrix is now known as henrix_ | ||
| dobey | well i managed to get a source package with at least one of the patches i want to test. i hope it doesn't blow up | 21:54 |
| kamal | dobey: excellent. good luck! | 21:58 |
| dobey | well, this is unfortunate, and not particularly helpful either: http://pastebin.ubuntu.com/1490051/ | 22:16 |
| kamal | dobey: the paste doesn't show the "real" error, which surely happened many many lines above that ... the kernel's parallel make gets run in "-k" (try to keep going after an error) mode, so the trick there is to capture the whole build output log, then search backwards from the bottom for "Error" | 22:18 |
| dobey | make[3]: openssl: Command not found | 22:28 |
| dobey | make[3]: *** [signing_key.x509] Error 127 | 22:28 |
| dobey | that i guess? | 22:28 |
| kamal | dobey: that sure looks like a problem to me! but I don't know how to fix it | 22:29 |
| dobey | but why would the kernel package not have a build-depends on openssl if it requires it? this is git://kernel.ubuntu.com/ubuntu/ubuntu-raring.git save for the i915 drm patch i pulled in from dan's intel-fixes branch, and my changelog entry to tweak the version | 22:31 |
| kamal | dobey: I think the openssl "signed kernel" stuff is new for raring, for whatever that's worth | 22:32 |
| * kamal tries a test raring build | 22:32 | |
| dobey | but the latest changelog entry is for the same version i'm currently running, so it's surely been built by the builders, which is weird | 22:34 |
| dobey | i'm happy to build without that for now if i can. i just want to see if this patch fixes my video issue | 22:34 |
| kamal | dobey: yes, the buildlog here does refer to openssl: https://launchpad.net/ubuntu/+source/linux/3.7.0-7.15/+build/4065696 | 22:36 |
| dobey | indeed. and pbuilder very helpfully just deletes the tree when it's done, regardless of failure | 22:36 |
| dobey | interestingly though, "openssl" is not installed via build-depends according to that build log | 22:42 |
| kamal | dobey: I was just going to point that out! | 22:42 |
| === chiluk is now known as chiluk_away | ||
| dobey | ok, found where i need to add it to build-depends, i think | 22:44 |
| kamal | dobey: but that won't explain why it builds okay on launchpad ;-) | 22:45 |
| kamal | dobey: is your pbuilder up to date? | 22:45 |
| dobey | yes, i did update right before the build | 22:46 |
| dobey | and i'm running on a raring host as well | 22:46 |
| kamal | dobey: my raring kernel test build (using the schroot builder on tangerine) did run openssl, just like the launchpad build. I still don't know what's pulling it in though. | 22:59 |
| dobey | kamal: maybe sbuild has it installed in the chroot by default, while pbuilder doesn't? | 23:00 |
| kamal | seems unlikely to me | 23:00 |
| kamal | dobey: I take that back ... openssl does indeed appear to be installed in the schroot on my build machine -- I don't know if that's the standard default or not. | 23:10 |
| kamal | .... and it's not installed in my raring pbuilder. | 23:10 |
| kamal | so, yes, I'm inclined to think that this is a bug in raring's Build-depends... it should list openssl explicitly. | 23:12 |
| dobey | now i'm getting this: http://pastebin.ubuntu.com/1490157/ | 23:14 |
| kamal | dobey: progress! | 23:14 |
| dobey | i think i'll have to deal with it later though. | 23:15 |
| kamal | dobey: the easiest solution to that is to just disable the ABI check ... I can help you with that when you're ready to proceed. holler for me. | 23:15 |
| dobey | thanks! | 23:15 |
| dobey | later | 23:16 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!