apw | crshman, right, they are raw builds, which arn't building, because upstream has yet to apply the fix, sigh | 10:12 |
---|---|---|
* xnox doesn't like our kernels | 11:10 | |
* xnox goes to fix | 11:10 | |
apw | xnox, ? | 11:11 |
xnox | apw, is there a brand new way to checkout ubuntu linux trees? do i need to use launchpad, or can i use kernel.ubuntu.com? | 11:11 |
* xnox did not have the memo | 11:11 | |
xnox | apw, just packaging and user-space bits, nothing drastic. | 11:11 |
apw | they are all in the original kernel.ubuntu.com place, though some of the newer ones are master in LP, and they are all available there too | 11:12 |
xnox | apw, oh, and i've built the s390x kernel package... it didn't build me the kernel image / udebs. used 4.2.0-19.23 | 11:12 |
apw | xnox, how did you build it | 11:13 |
xnox | apw, dpkg-buildpackage -b | 11:13 |
xnox | it gave me source, doc, headers, tools-common, cloud-tools-common, and linux-libc-dev and that's it =( | 11:13 |
apw | which branch of what repo ? | 11:14 |
xnox | pull-lp-source linux =) | 11:14 |
xnox | hence looking for repositories now ;-) | 11:14 |
xnox | http://kernel.ubuntu.com/git/ubuntu/ubuntu-xenial.git/ is out of date, so i need launchpad i guess. | 11:14 |
apw | yep, then you will only get linux-libc-dev, that is a stage1 only build in the archive | 11:15 |
apw | you need the ubuntu-xenial master-next to get anything mroe | 11:15 |
xnox | https://code.launchpad.net/~ubuntu-kernel/ubuntu/+source/linux/+git/xenial -> looks fresh | 11:15 |
xnox | ack | 11:15 |
apw | xnox, yes that is the official master for xenial | 11:15 |
xnox | and i don't like the droppage of libunwind8-dev =) | 11:15 |
apw | heh whyso ? | 11:16 |
xnox | we should totally use elfutils which has unwind support on s390x | 11:16 |
xnox | and like switch to it from libunwind throughout, because hipsters =) | 11:16 |
apw | if you have the energy :) and being you you have a lot of that | 11:17 |
xnox | well, my hands are tied, as llvm is building.... | 11:17 |
xnox | and there is one more llvm to build. | 11:18 |
* xnox should have done a shallow clone =( | 11:23 | |
apw | xnox, or a --reference linux-linus clone | 11:25 |
xnox | that would have worked, if i were a kernel developer =) | 11:25 |
apw | all serious developers have linux clone :) | 11:25 |
apw | plus kernel.org's pipe is really fat | 11:26 |
apw | xnox, you will need to touch some abi-ignore files into s390x as there are no s390x builds yet, chicken-egg-chicken | 11:29 |
apw | xnox, echo 1 >debian.master/abi/4.3.0-0.7/ignore; echo 1 >debian.master/abi/4.3.0-0.7/ignore.modules | 11:29 |
apw | else it will barf on you | 11:29 |
apw | ass not that | 11:30 |
apw | xnox, mkdir debian.master/abi/4.3.0-0.7/s390x; echo 1 >debian.master/abi/4.3.0-0.7/s390x/ignore; echo 1 >debian.master/abi/4.3.0-0.7/s390x/ignore.modules | 11:30 |
apw | xnox, that | 11:30 |
xnox | apw, you have access to a devirt ppa to test build things on all arches, right? | 11:31 |
apw | xnox, yes | 11:31 |
xnox | apw, https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1519820 | 13:23 |
ubot5 | Ubuntu bug 1519820 in linux (Ubuntu) "xenial linux master-next FTBFS on s390x" [Undecided,New] | 13:23 |
apw | xnox, you can turn that off in debian.master/rules.d/s390x.mk do_zfs=false sort of thing | 13:25 |
xnox | apw, tah, let me test that. | 13:25 |
* xnox will submit patches with all changes in the end. | 13:25 | |
apw | xnox, the build i did for infinity was against the 4.2 base, so i am not supprised you are seeing issues | 13:30 |
apw | xnox, though it has been test built (other than zfs because that won't cross compile) | 13:30 |
apw | xnox, and thanks | 13:41 |
xnox | apw, in linux-image-extra there are: | 13:41 |
xnox | -rw-r--r-- root/root 24590 2015-11-25 14:26 ./lib/modules/4.3.0-0-generic/kernel/drivers/s390/block/dasd_fba_mod.ko | 13:41 |
xnox | -rw-r--r-- root/root 178446 2015-11-25 14:26 ./lib/modules/4.3.0-0-generic/kernel/drivers/s390/block/dasd_eckd_mod.ko | 13:41 |
xnox | -rw-r--r-- root/root 190646 2015-11-25 14:26 ./lib/modules/4.3.0-0-generic/kernel/drivers/s390/block/dasd_mod.ko | 13:41 |
xnox | -rw-r--r-- root/root 21502 2015-11-25 14:26 ./lib/modules/4.3.0-0-generic/kernel/drivers/s390/block/dasd_diag_mod.ko | 13:41 |
xnox | .. | 13:41 |
xnox | i think these are ought to move to linux-image, rather than stay in extra. | 13:41 |
apw | that looks like it should be in linux-image as that is likley the disks :) | 13:42 |
apw | xnox, concur | 13:42 |
xnox | and i also want dasd-modules and dasd-modules-extra udebs | 13:42 |
xnox | like in debian. | 13:42 |
xnox | in debian they put: | 13:42 |
xnox | -rw-r--r-- root/root 163968 2015-11-10 20:42 ./lib/modules/4.2.0-1-s390x/kernel/drivers/s390/block/dasd_mod.ko | 13:42 |
xnox | -rw-r--r-- root/root 24896 2015-11-10 20:42 ./lib/modules/4.2.0-1-s390x/kernel/drivers/s390/block/dasd_fba_mod.ko | 13:42 |
xnox | -rw-r--r-- root/root 179880 2015-11-10 20:42 ./lib/modules/4.2.0-1-s390x/kernel/drivers/s390/block/dasd_eckd_mod.ko | 13:42 |
xnox | into dasd-modules udeb, | 13:42 |
xnox | and the dasd_diag_mod.ko into dasd-modules-extra udeb. | 13:42 |
xnox | not sure if that's worth the split or not. | 13:43 |
xnox | apw, do you want a bug report for that, or shall I try to implement this udeb bit? | 13:43 |
apw | i am happy to do it, but a bug would be nice | 13:43 |
apw | the bit where an arch deviates from the norm is a bit messy to say the least | 13:43 |
apw | xnox, for the instant i'd suggest adding them to the scsi-modules list and making that Provide: dasd-modules | 13:44 |
apw | and then i can fix it properly slower | 13:44 |
xnox | apw, cool. i kind of pre-emptively uploaded d-i that expects to find dasd-modules =) | 13:45 |
xnox | due to monkey-see debian, monkey-do the same | 13:45 |
xnox | apw, "patch" for disable of zfs -> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1519820/comments/2 | 13:49 |
ubot5 | Ubuntu bug 1519820 in linux (Ubuntu) "xenial linux master-next FTBFS on s390x" [Undecided,Confirmed] | 13:49 |
xnox | submitted trivial bug fix to upstream to fix spl, but not sure if want to carry that without zfs. | 13:49 |
xnox | linked to from https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1519814 | 13:49 |
ubot5 | Ubuntu bug 1519814 in linux (Ubuntu) "xenial, linux-master FTBFS on s390x" [Undecided,Incomplete] | 13:49 |
apw | if that makes it build, we should likely get it tested by someone | 13:51 |
apw | i am testing a fix for the udebs now, if you get my bug filed :) | 13:52 |
xnox | apw, well that patch makes the "spl" bits build, but zfs bails out. | 13:52 |
xnox | apw, yes sir. | 13:52 |
apw | xnox, i guess we'll turn it off, and file a bug for someone to turn it back on once we get a machine we can boot | 13:53 |
xnox | apw, https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1519833 | 13:55 |
ubot5 | Ubuntu bug 1519833 in linux (Ubuntu) "xenial master-next needs a different udeb split for x390x" [Undecided,New] | 13:55 |
xnox | apw, let me file a bug with s390x tag. As it will most likely be me again. | 13:55 |
* xnox will check priority on zfs work too. | 13:56 | |
xnox | ok, so this is looking lovely. If you give me a patch for udebs to have dasd modules, i should be in a position to spin up a test d-i build. | 13:57 |
apw | xnox, building it again now, looking like it will work ... | 13:58 |
xnox | \o/ =_ | 13:59 |
xnox | \o/ =) | 13:59 |
apw | xnox, i'll push fixes for all of the above to the tree shortly | 13:59 |
xnox | cool. | 13:59 |
* xnox goes to make an egg soup, whilst attempting to poach an egg. | 14:09 | |
apw | whoops | 14:16 |
apw | xnox, ok ... i've pushed a quartet of changes to the tip of master-next for your issues | 14:47 |
xnox | tah | 14:51 |
* xnox will finish figuring out how to unbreak xorg-gtk3-mir loop and will get back to that | 14:51 | |
xnox | apw, yo. very good it builds in sbuild without a hitch. Some typos though: | 15:41 |
xnox | apw, s/dasd-modules-extra/dasd-extra-modules/ possibly my mistake, sorry about miss-leading you. | 15:41 |
apw | bah | 15:41 |
xnox | apw, and `dasd_mod.ko` ended up in both dasd-modules and the extra variant, that's a conflict. | 15:41 |
apw | no that is normal | 15:42 |
xnox | dasd-extra-modules should have just the dasd_diag_mod, and like depend on dasd-modules i guess. | 15:42 |
xnox | oh, is it? | 15:42 |
apw | hmmm, it may go away if it depends, i'll try that | 15:43 |
apw | it is normal for all of the deps to end up in a module, but it may cope with interdeps | 15:43 |
xnox | yeah it has: Depends: kernel-image-4.3.0-0-generic-di, storage-core-modules-4.3.0-0-generic-di | 15:43 |
xnox | but needs the dasd-modules dependency too | 15:43 |
apw | the extra is so small i think it might make sense to be merged | 15:43 |
xnox | me too | 15:43 |
xnox | kill the extra, and just stick it all into dasd-modules. | 15:43 |
xnox | and i'm fine to keep this "divergence" from debian. | 15:44 |
apw | i'll confirm the dep helps or not, if not, then i recon i will just wack it | 15:44 |
xnox | cool. | 15:45 |
xnox | it does build natively in a clean sbuild without any problems now. Thanks a lot =) | 15:45 |
* xnox compiles mir | 15:46 | |
apw | if i add a Provides: you would be covered anyhow i think | 15:46 |
apw | xnox, ok ... i've forced master-next with the updated udeb names, getting the Depend: right has cleared out the duplicate | 15:51 |
xnox | \o/ | 15:52 |
xnox | i will rebuild it again, after getting some chocolate. | 15:52 |
apw | i've also fixed builds for !s390x too :) | 15:52 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!