[10:33] @LocutusOfBorg, hello! the virtualbox upload to fix building for 5.8 kernels in focal is still on the upload queue? [10:55] klebers: that appears to be the case [10:56] klebers: https://launchpad.net/ubuntu/focal/+queue?queue_state=1&queue_text= [10:56] @enyc, thanks, I'll contact someone from the sru team [10:56] klebers: groovy20.10 virtualbox6.1.16 went through ok [10:57] needs to go into focal-proposed [10:57] yeah, I noticed it seems to be missing the update only for focal-proposed [11:01] morning guys [11:01] * enyc meows [11:01] is there a main bug about that 5.8 hwe thing on focal? [11:01] ... Not sure about 5.8 kernel schedule (if there even is one) [11:02] lotuspsychje: https://bugs.launchpad.net/kernel-sru-workflow/+bug/1907407 ?? [11:02] Ubuntu bug 1907407 in linux-hwe-5.8 (Ubuntu Focal) "focal/linux-hwe-5.8: 5.8.0-33.36~20.04.1 -proposed tracker" [Medium,Fix released] [11:02] lemme check tnx [11:03] klebers, yes [11:04] 5.8.0-33 tho, needs newer [11:04] enyc: some users claiming that their 20.04 box received the 5.8 kernel in their regular updates [11:04] and now its raining bugs on wifi/5.8 in #ubuntu-bugs-announce [11:05] @lotuspsychje, this is the latest tracking bug for hwe-5.8: https://bugs.launchpad.net/kernel-sru-workflow/+bug/1910008 [11:05] Ubuntu bug 1910008 in Kernel SRU Workflow prepare-package-signed "focal/linux-hwe-5.8: 5.8.0-37.42~20.04.1 -proposed tracker" [Medium,In progress] [11:05] lotuspsychje: aah ;-( [11:06] lotuspsychje: that should only be the case if they installed -hwe explicitly ? [11:06] @lotuspsychje, we are aware of the bcmwl issue. Unfortunately the dkms was not tested as we expected before the hwe update to 5.8 [11:06] enyc: no, i think they got it on clean installs or something, can that be? [11:06] klebers: clem of linuxmint, did a test of bcmwl-source and did a backport [11:07] for 20.1-edge linuxmint image based upon ubuntu 20.04 .... [11:07] lots of people saying they NEED newer kernel for ryzen and so-on! [11:07] fun =) [11:08] lotuspsychje: I can only imagine 3rd-party / respin images .... many of which exist especially to provide newer kernel [11:10] enyc: ill see if i can ask their dpkg logs, if more claims arise [11:56] klebers: wich bug ID can we use to forward the dupes of that broadcom bug? [12:06] eh-oh! https://bugs.launchpad.net/bugs/1910709 nvidia-340 broken as well as bcmwl-source [12:06] Ubuntu bug 1910709 in linux (Ubuntu) "DKMS linux-headers-5.8.0-34-generic update brokes nvidia-340 " [Undecided,Incomplete] [12:23] [ok] [12:31] 340 is EOL [12:34] But why is 5.8 suddenly getting installed? [12:34] if the hwe metapackage is installed [12:36] It happened to me when installing 20.04 the other night with an ISO from April [12:46] lotuspsychje, I think this bug report is a good candidate to mark all the ones as its duped: https://bugs.launchpad.net/ubuntu/+source/bcmwl/+bug/1910555 [12:46] Ubuntu bug 1910555 in bcmwl (Ubuntu) "Driver build error on kernel 5.8, but works on 5.4" [Undecided,Confirmed] [12:48] lets see [12:48] By default the 20.04 media installs the hwe kernel for desktop images. The hwe meta was still pointing to the 5.4 generic kernel, until this week when we released an update to pull the hwe-5.8 backport [12:48] ok tnx klebers ill spread the word [12:49] 20.04.2 will install 5.8 as default [12:49] lotuspsychje, thanks [14:49] lotuspsychje: https://bugs.launchpad.net/bugs/1901904 vbox package ready to test, soon in focal-proposed, etc. [14:49] Ubuntu bug 1901904 in virtualbox-guest-additions-iso (Ubuntu Groovy) "[SRU] virtualbox 6.1.10-dfsg-1~ubuntu1.20.04.1 ADT test failure with linux-hwe-5.8 5.8.0-25.26~20.04.1" [Undecided,New] [19:41] hrrm ... I'm not a ubuntu kernel git expert yet but ...... [19:41] https://lore.kernel.org/stable/20210107143052.592801948@linuxfoundation.org/T/#u [19:42] ^^ This amdgpu fix be good to have in ubuntu kernels, will apparently be in 5.10.6 upstream kernel... I would guess it will 'filter' into all kernel.org LTS kernels [4.19 and 5.4 particularly], but not necessarially 5.8 as upstream kernel.org don't support that anymore [19:49] tjaalton: if nvidia-340 = EOL, why is it still present? some still need it for their gpus? ubuntu ought to do something to warn about the incompatibility if going to kernel >5.4? etc [19:52] enyc: build fixes are trivial aiui, it's just that there are no cve fixes to it anymore [19:54] also, we maintain our own stable trees, so fixes sent to stable@ get backported to the distro kernels we have which are EOL upstream [19:54] like 5.8, until groovy itself is EOL [19:54] tjaalton: OK #1872950 exists I see... hope that gets sorted by 20.04.2 . [20:06] no way [20:06] ah [20:06] uh [20:06] needs a new bug [20:07] that one is closed [20:57] enyc: bug 1910709 is more fresh [20:57] bug 1910709 in nvidia-graphics-drivers-340 (Ubuntu Focal) "DKMS linux-headers-5.8.0-34-generic update brokes nvidia-340 " [Undecided,Confirmed] https://launchpad.net/bugs/1910709 [21:15] please retry https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/bootstrap/+build/20785961 [21:39] tjaalton: aha, thankyou =) [21:40] for some reason the autopkgtest shows green... [21:40] it should catch build errors I think [21:41] tseliot isn't here [21:41] tjaalton: the package / source package as built just puts files into the dkms dir, its' a "runtime" (user) build to build against particular kernel headers ....? [21:42] that's what I mean [21:43] tjaalton: ok so where's the spec of what autopkgtest does, etc.? [21:43] The only dkms package that I know is broken is rtl8812au-dkms [21:44] jeremy31: https://bugs.launchpad.net/ubuntu/+source/rtl8812au/+bug/1884646 ? [21:44] Ubuntu bug 1884646 in rtl8812au (Ubuntu) "rtl8812au 4.3.8.12175.20140902+dfsg-0ubuntu12 ADT test failure with linux-5.8 5.8.0-1.2" [Undecided,Confirmed] [21:44] jeremy31: can you comemnt in the buf that it is still valid bug etc.? [21:44] err still fails an 5.8.0-36 and affects focal and groovy, or so? as appropriate [21:44] enyc: It has been broken since 16.04 [21:45] jeremy31: o yes lol lots of similar build breakage [21:46] The Makefile and/or dkms.conf needs to be changed or it builds against the wrong headers when a new kernel is installed and people get exec format errors trying to load the module as the vermagic isn't correct [21:48] comment #3 on https://bugs.launchpad.net/ubuntu/+source/rtl8812au/+bug/1733379 shows this [21:48] Ubuntu bug 1733379 in rtl8812au (Ubuntu) "RTL8812AU driver working no more after kernel 4.4.0-101-generic update" [Undecided,Confirmed] [21:48] enyc: spec? just look at it? [21:51] in fact there is no autopkgtest [21:57] so I don't know what it's actually testing [21:58] 'dkms-autopkgtest' [21:58] tjaalton, hi, could you retry https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/bootstrap/+build/20785961 [21:59] no [21:59] don't know how [21:59] oh [21:59] thanks [21:59] retried