[09:17] hello folks, can you please sponsor the fix for bionic/cosmic? https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1798921 ? anything we should do to have it? [09:17] Ubuntu bug 1798921 in linux (Ubuntu Cosmic) "sky2 ethernet card don't work after returning from suspension" [Undecided,New] === cpaelzer__ is now known as cpaelzer [09:53] Hello everyone, slight problem with my Mint 19 (based on 18.04) and its sleep-wake-NoNetwork problem. Two Marvell 88E8053 with the "sky2" driver, both lose their network with a "Network cable disconnected". [09:53] I found this workaround script that works great (https://askubuntu.com/questions/1029250/ubuntu-18-04-ethernet-disconnected-after-suspend) but it seems to me that it's a bug. [09:53] And then I found the exact bug for the exact "sky2" driver: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1798921 [09:54] Ubuntu bug 1798921 in linux (Ubuntu Cosmic) "sky2 ethernet card don't work after returning from suspension" [Undecided,New] [09:55] Now, since the Linux world is something far in the past for me, that I took a renewed interest upon, and since I'm on Mint but all the fingers point to Ubuntu, how does one handle this? What am I supposed to do? Create a new Mint ticket? Tag along the existing Ubuntu one? [09:57] Ticket 1798921 talks about the issue being fixed a week ago, on 2019-04-02, but it's only available on kernel 4.4.0-145.171. What's the norm in these situations? Do the changes/fixes propagate? Should I just stick with the workaround script and forget about the whole thing? [09:57] TIA for any thoughts/comments on this... [09:58] hello folks, can you please sponsor the fix for bionic/cosmic? https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1798921 ? anything we should do to have it? [09:58] Ubuntu bug 1798921 in linux (Ubuntu Cosmic) "sky2 ethernet card don't work after returning from suspension" [Undecided,New] [09:59] socratis, I asked this before your join [09:59] Ah, thanks LocutusOfBorg!!! [09:59] It needs a ... sponsor? Like the Olympics? MasterCard and Heineken? :D [10:00] Generally speaking, if I find an issue with the kernel, should I be talking to the Mint people or the Ubuntu people? Do the Ubuntu people "hate" the Mint people/distro? ;) [10:01] arighi, ^^ maybe you are the best person to answer this... can you consider adding that one-line patch on next bionic/cosmic update? [10:01] socratis, the kernel should be the same, so "we" are the best people to ask [10:02] Excellent! Nice to know... [10:05] smb, since you fixed it on trusty... may I ask why you didn't forward the patch to bionic, cosmic and disco? [10:06] Funny you should mention that "nickname", I'm trying to solve a read-only SaMBa share issue! :D [10:07] Oh, and since I have your attention, and since this is the Ubuntu channel, and since we've known each other for a long time from the #vbox channel... [10:07] LocutusOfBorg: Did you get any reports from people that couldn't install VirtualBox 6.0.x on Ubuntu 16.04 LTS? [10:08] Just since yesterday, I've had 3 reports already! [10:08] (methinks I should double-post in #vbox or #vbox-dev just to be safe...) [10:11] socratis, lets move on vbox [10:11] Okie dokie... [10:13] LocutusOfBorg, if you talk about the sky2 bug, that was xenial. And as to why not port to newer releases, it seems nobody reminded us (in the sense of having the issue with those kernels) and there was just too many other things going on [10:14] I do! Put me on the list smb... ;) [10:15] smb, I see disco is already ok, so it should be a matter of bionic and cosmic, if you can add it to the patch queue... [10:15] I'm simply a n00b in this and I didn't even know whether I should first talk to the Mint people (Mint19 here), the Ubuntu people or the kernel people... [10:18] LocutusOfBorg, I had already added nominations for those to the bug report. We will see whether this would be an easy pick. socratis you probably can help by adding a comment about having this on bionic/cosmic [10:18] Okie dokie... On it. I have to register, right? (of course...) [10:20] socratis, is how we control spam indeed [10:24] I know... ;) [10:51] socratis, ok, looks like simple pick and I sent it to our mailing list (the sky2 patch). So it should make it at some point [10:51] Excellent! Thanks smb... [12:11] Is there a known issue with linux-headers-hwe-lowlatency-18.04-edge from bionic-proposed, via linux-headers-5.0.0-8-lowlatency, which is preventing DKMS module builds due to missing symlink targets. E.g. "lrwxrwxrwx 1 root root 39 Mar 19 21:22 /usr/src/linux-headers-5.0.0-8-lowlatency/block -> ../linux-hwe-edge-headers-5.0.0-8/block" ? apt-file search does not find the target path and I cannot find it [12:11] in a .postinst script either === chrisccoulson_ is now known as chrisccoulson [13:44] since edge is now coming from disco... sforshee, cascardo ^? === ddstreet_away is now known as ddstreet [13:55] I will look at that soonish [13:57] TJ-, cascardo: I just downloaded and extracted linux-headers-5.0.0-8-lowlatency_5.0.0-8.9~18.04.1_amd64.deb and the symlink is there [13:58] sforshee: hmmm! let me check what's installed again [14:05] sforshee: whih package do you see the target path (/usr/src/linux-hwe-edge-headers-5.0.0-8/block) being installed/created from ? I can't see it here with the same package installed (Installed: 5.0.0-8.9~18.04.1) [14:05] TJ-: https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/ppa/+build/16519051/+files/linux-headers-5.0.0-8-lowlatency_5.0.0-8.9~18.04.1_amd64.deb [14:06] I just downloaded and extracted it, I did not install it [14:07] sforshee: I don't see the path in the data.tar.gz of that package [14:11] $ wget https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/ppa/+build/16519051/+files/linux-headers-5.0.0-8-lowlatency_5.0.0-8.9~18.04.1_amd64.deb [14:11] $ dpkg-db -R linux-headers-5.0.0-8-lowlatency_5.0.0-8.9~18.04.1_amd64.deb ext [14:11] $ ls -l ext/usr/src/linux-headers-5.0.0-8-lowlatency/block [14:11] lrwxrwxrwx 1 sforshee sforshee 39 Mar 19 16:22 ext/usr/src/linux-headers-5.0.0-8-lowlatency/block -> ../linux-hwe-edge-headers-5.0.0-8/block [14:11] TJ-: ^ [14:11] typo, s/dpkg-db/dpkg-deb/ [14:12] I got that one :D [14:12] as in, the typo [14:18] funny thing is that it was a copy/paste, so I have no idea where it went [14:21] sforshee: Sometihng weird going on, I replicated your commands and there is no ./usr/src/linux-hwe-edge-headers-5.0.0-8/ created [14:21] sforshee: in case there is some confusion here, the sym-links are there, the *target* is missing [14:22] TJ-: ok, I misunderstood I thought you said the symlink was missing [14:22] sforshee: no, target of the symlinks are all 'dead', and I can't find which package is supposed to install to /usr/src/linux-hwe-edge-headers-5.0.0-8/ [14:28] TJ-: yeah that does look like a problem, I'd expect those headers to come from linux-headers-5.0.0-8 but that also produces symlinks [14:28] cascardo: ^ [14:28] phew! not me going lala! [14:29] It knocked out wireguard else I wouldn't have noticed, which was unexpected since there is no problem with the mainline 5.0 builds and the OS had switched on last reboot [14:32] it should have been caught in testing, but it looks like adt has never run for that kernel for some reason [14:32] yeah, so I will look at the problem there, and probably have it fixed on a 5.0.0-9 upload [14:33] cascardo: or -10, I built a -9 in bootstrap but there's an apparmor regression that needs fixing so I don't plan to put that into -proposed [14:35] cascardo: for the main kernel that is, I didn't build an hwe-edge kernel [14:35] sforshee: ack, will work on the fixes, and wait for a -10 before rebasing [14:36] sforshee: I will drop snapdragon altogether from main and meta package [14:37] cascardo: hmm now that brings up a problem, I guess we've had hwe/hwe-edge packages for snapdragon but will not have them moving forward [14:37] apw: ^ any suggestions on how to handle that? [14:40] sforshee, keep building them in the -hwe ? [14:40] for the main package, I don't think that's an easy thing to do [14:40] apw: even if we don't expect they will work? [14:41] during rebase for 5.1, I had to drop them all, lots of snapdragon patches causing build failures [14:41] hmmm, are you intending to break them [14:41] oh [14:41] we dropped support for snapdragon from the main kernel branch and are using a topic branch now [14:41] we might need to ask pp for a 5.1 version [14:41] or whatever hwe is [14:41] and we will have double the work? [14:41] whatever we had for 5.0 was never confirmed to work, afaik [14:42] maybe we could have the hwe meta package point to the generic arm64 one [14:42] cascardo, that is at an older version no ? [14:42] though that would likely not work, or have lots of "bugs" [14:42] I think he means the generic arm64 5.0 kernel [14:42] yep [14:43] we signed up for double the work when we split it out ... [14:43] otherwise, just keep them at an older version, ie, do not provide a new meta package [14:43] that effectivly abandons them without cves [14:44] we will have a 5.1 and so on linux-snapdragon, right ? [14:44] I'm assuming we will [14:44] I don't disagree. but if we love those snapdragon hwe users so much, we will take not only double the work, but risk regressions for other users of the master kernel [14:45] apw: would you suggest we provide yet another repo for a snapdragon-hwe? [14:45] it's sounding to me like we'll have to start producing linux-snapdragon-hwe packages [14:45] based on disco/linux-snapdragon, for example? [14:45] cascardo, i am struggling to see how we can avoid doig something, if we supported snapdragon in an hwe kernel [14:45] too bad we just didn't support it [14:46] we may be able to leave them on this version and not roll them, but it is this version we are breaking [14:47] its a problem indeed, if we had never had hwe version of it, we would be better off [14:47] sforshee, you could look at the anonuncement to see if we ever said we supported it on snapdragon [14:47] sforshee, whetner we might expect anyone to be using it [14:48] apw: which announcement would that be? I do wonder whether anyone uses it at all [14:48] sforshee, that is the worry, that you have to do it, and it is never used [14:49] apw: I'm just wondering where we would have announced it. In the release notes for one of the point releases maybe? [14:50] sforshee, yeah it would be the point release announcement i guess [14:50] sforshee, that is the only place i can imagine it could have been listed [14:51] apw: it's all your fault - bug 1798352 ;-) [14:51] bug 1798352 in linux-meta-snapdragon (Ubuntu Cosmic) "linux-snapdragon: missing meta packages for this flavour" [Critical,Fix released] https://launchpad.net/bugs/1798352 [14:51] sforshee, i am sure i didn't do that off my own bat; i imagine i was doing it because someone noticed [14:52] apw: yeah and it wasn't just for -hwe. That bug was mentioned here - https://wiki.ubuntu.com/BionicBeaver/ReleaseNotes/ChangeSummary/18.04.2 [14:52] that's the closes I've found to us announcing anything [14:52] *closest [14:53] sforshee, yeah that was an adam request iirc [14:59] right now, 5.0 will work just fine with snapdragon, we haven't dropped patches yet [15:00] so, we could just leave them as it is, until we get to later kernels, like 5.1 [15:12] cascardo: I did drop the patches. We could reapply them, but I had already had to drop some because of conflicts. Afaik the support we had in the 5.0 master kernel was never tested at all [15:14] even if we dodged the issue that way for 5.0 we have to solve it later [15:18] yeah, so we'd better just have that linux-snapdragon-hwe instead of reapplying them. if there are conflicts, it's not worth fixing them === Orphis_ is now known as Orphis === SimonNL is now known as SimonNL_Afk [17:32] TJ-: found the cause for the missing files on the headers package, will fix it on the next upload [17:33] apw: so, I will follow up with dropping the snapdragon flavour from the bionic/hwe-edge kernel, we'll have to sort it out some other way [17:33] cascardo: great news :) === tdaitx_ is now known as tdaitx