=== gerald is now known as Guest85444 [04:10] yo [06:46] ppisati, no! (too early) [07:49] yeah that is far toooo early [07:50] But not too early for the BEES! [07:55] apw: any chance of this patch getting into trusty? https://bugs.freedesktop.org/show_bug.cgi?id=58378 [07:55] Freedesktop bug 58378 in Driver/nouveau "[NV86] Distorted graphics on NVIDIA GeForce 8400M G after upgrade the kernel to 3.5.x (and RHEL 6.5) or later" [Normal,Resolved: fixed] [07:56] apw: I've not been able to use the ubuntu dashboard (switched to i3) due to corruption issues since 12.04 [07:56] the patch made it into 3.14-rc1 [07:59] related ubuntu bug I filed (and ignored eventually): https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1158689 [07:59] Launchpad bug 1158689 in linux (Ubuntu) "10de:0422 bringing up dash causes screen corruption on nouveau" [Medium,Incomplete] [08:01] or perhaps you look into graphics RAOF? [08:02] If you need a kernel patch, you need apw! [08:02] well there is little change for the release kernel, but ... for an sru i can have a look [08:02] (Or someone else on the kernel team ☺) [08:02] amitk, could you add the details to the bug as well [08:02] * smb wonders whether amitk remembers that deadline thing in general [08:03] apw: what details do you need? I'll link to the freedesktop bug [08:03] amitk, that'll do i recon [08:03] smb: its done when its done ;) I can't promise deadlines for scheduler changes ;) [08:04] amitk, Just wondering about a patch that you say made it into 3.14-rc1. Thought that was a bit ago. [08:06] smb: I haven't tracked this bug for a while since I don't use the ubuntu dashboard anymore. This came up again when I was testing out Trusty images over the weekend and I decided to look if it was fixed anywhere [08:06] smb: and I knew you were getting close to a release, but no I didn't know when exactly [08:07] :) [08:07] amitk, So quick to forget. :) [08:09] smb: hardly forgotten, more like I don't upgrade my devbox on a 6-monthly cadence anymore ;) [08:10] apw: done linking against the upstream bug [08:15] amitk, this is going to be a hard ask, it is rather large: [08:15] 33 files changed, 500 insertions(+), 308 deletions(-) [08:18] apw: did you look at commit 4019aaa2b314a5be9886ae1db64ff8c6d3c060ed? [08:19] apw: I only see 17 files changed, 287 insertions(+), 35 deletions(-) [08:19] apw: or is there a dependency patch that I missed? [08:20] amitk, yes, there are two, one of which 'rewrites most of the quirk init code' [08:20] apw: aah [08:20] though even the 17 is pretty invasive [08:20] apw: true [08:22] apw: the patch itself is important for nouveau users that are currently seeing errors in their logs because the driver tries to use HW blocks that aren't even present [08:23] apw: I'll just switch to mainline kernels [08:27] amitk, i'll spin a test kernel and you can test it, at least we will know for sure it is the fix [08:27] i can then propose it and see what happens [08:32] apw: thanks [09:15] Can someone take a look at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1303419 ? I found a workaround but it'll be nice to have it as default. [09:15] Launchpad bug 1303419 in linux (Ubuntu) "Brightness keys don't work on HP Elitebook Folio 1040 G1" [Undecided,Confirmed] [09:18] jpds, will have a look [09:35] apw: Thanks. === zequence_ is now known as zequence === cjwatson_ is now known as cjwatson [09:56] amitk, test kernels linked in the bug [10:11] apw: will test soon and get back, thanks again [10:35] jpds, possible fix applied to test kernels, urls in the bug [10:37] apw: Testing... [10:40] jpds, it is not 100% clear from the description which ids are covered, so it may not, we shall see [10:41] apw: It worked. [10:45] \o/ [10:45] jpds, ok ta, put that in the bug for me [10:46] apw: Already done. [10:50] jpds, submitted to our list for consideration, shame you didn't get this in before freeze, but hey [10:51] apw: I got the laptop yesterday! [10:51] ;-) === inaddy is now known as tinoco [11:21] jpds, heh === elmo_ is now known as elmo [13:14] rtg, i just slammed head to fix a bug number in one of my patches [13:16] apw, ack [13:16] apw, did I mark the wrong bug fix committed then ? [13:17] rtg, you did on my error, i fixed that up as well [13:17] cool [13:17] rtg, you seem to have skipped over "rds: prevent dereference of a NULL device in rds_iw_laddr_check", i was going to apply that if you arn't doing it ... its a big heap of applying i don't want to duplicate :) [13:18] apw, I just haven't gotten to it yet, but go ahead. [13:18] rtg, ack [13:18] I was busy updating kernels and such. [13:18] in fact, I'm about to reboot now. === tinoco is now known as inaddy === inaddy is now known as tinoco === brendand_ is now known as brendand [14:54] is there a difference between checking for 'mem' in /sys/power/state vs using pm-is-supported --suspend? [14:54] i'm sure there is a difference, but i'd like to know what that is [14:55] if mem exists but pm-is-supported fails, does that mean suspend *can't* work on that system, or might it mean there's a bug? [14:56] similarly, if mem doesn't exist, might it be possible for pm-is-supported to succeed? === shuduo is now known as shuduo_afk [15:10] brendand, it checks mem in /sys/power/state as one of the options, there are two others, so to answer your questions, yes and yes [15:11] it is not clear how pm-is-supports --suspend could fail if mem is in /sys/power/state however [15:14] apw, when installing from the server ISO having a driver referenced in d-i/modules/nic-modules should be sufficient for network access, right ? For example, 'debian.master/d-i/modules/nic-modules:mlx4_en ?'. [15:14] rtg, i believe so indeed [15:15] apw, and it _should_ get bundled into the initrd when the kernel is installed on the target disk. [15:15] the criteria there are unrelated, networking is not normally seen as boot essential is it ? [15:16] apw, Infiniband might be. Guess I'd better find out. [15:16] nor would it normally be, given you open the real disk, and find the drivers there [15:16] yeah ... that [15:16] apw, netboot? [15:16] oh dear, if pm-suspend can't do real hybrid sleep then it cheats :/ [15:17] smb, that has a different initrd [15:17] brendand, "oh dear" ? [15:17] apw, hopefully one that considers networking as boot essential [15:17] hopefully nideed [15:18] apw, from the point of view of someone testing that hybrid sleep works, yes [15:19] smb, i just updated "clark" and rebooted, on startup i can now seem my old VMs, good, but they do not start [15:19] Error starting domain: internal error: libxenlight failed to create new domain 'grovel-r64' [15:20] apw, Would be nice to have some more useful messages. Unfortunately those hide usually in /var/log/libvirt/libxl/... on the host [15:21] tseliot, have you heard any howling about 4K monitor support getting broken with this last nVidia-331 update ? [15:21] smb, yeah that is crap, it may be an old VM with "pre VG rename" paths in it [15:22] rtg, i don't think it's 4k support. i think nvidia is borked [15:22] rtg, i'm guessing you are in low-res mode [15:22] apw, True. Well it could be anything. Sadly the error reporting through virt-manager is... suboptimal [15:22] rtg, and your jumbo-tron doesn't have a low-res support [15:22] rtg: the driver we have in Ubuntu is not the latest from Nvidia. I haven't tested it with 4K but 2560x1440 works fine here [15:22] bjf, oh, you mean altogether borked ? I'm in "no mode" black screen [15:23] rtg, is your monitor "black" or "off", mine seems to be off [15:23] smb, ok this was "my fault" [15:23] apw, black [15:23] they seem converted just fine [15:23] rtg, i'm in low-res on a "normal" monitor which might be different on the jumbo-tron [15:23] rtg: in the latest update I only added support for Linux 3.14 [15:23] apw, Uh, what kept them from starting then? [15:24] somewhen back in the mists of time (read more than 2 weeks back) i renamed my datavg to clarkvg [15:24] and this had the old, you cannot have inserted the wrongo path [15:25] apw, Ah! Ok, makes some sense then. Just not the nicest thing on earth when it comes to errors. As much as a python stackrace can ever be called nice [15:26] ok this was a raring machine, that makes sense being as it isn't supported any mroe, /me tries an s instance [15:26] apw, No the convertion should not change anything there [15:26] smb, ok my S which i had fixed previously appears correctly and starts correctly. i'd say that is success, got a bug # i can report that testing in ? [15:29] apw, No, I probably should open one... and maybe ... hm... just occured to me that re-creating those VMs that are in the old path but not the new may be annoyingly bringing back things on every upgrade [15:30] smb, you have to mark that you have done it and do it just once indeed [15:30] apw, except maybe when adding a '--force' [15:31] smb, or rename the previous ones to .old sort of thing [15:31] so you can rename them back if you need to but only process the ones not .old [15:32] apw, I would be careful there. As I cannot be 100% sure what the old xend does with those. [15:34] smb, move them from like /foo/bar/machines/* to /foo/bar/machines.old/* [15:36] apw, If I do that, and someone switches back to the old toolstack, the domains are gone there. [15:38] smb, yes but i may have changed them anyhow, so you need to convert them back [15:38] smb, which is why i think we should convert them, pop somethign up to tell you that, so you arn't supprised they are gone when you downgrade [15:39] smb, but overall they work well so that at least is good [15:39] apw, it already tells you about the conversion in the log. And I rather won't go on a reverse. Thats deprecated anyhow [15:40] And adds too much complexity [15:40] smb, right that is why i don't think we care that they are gone if you go back, that can be documented on something the log points to perhaps [15:40] like a wiki for downgraders [15:40] or something [15:41] smb, though your bug should have a release notes task so we can warn people this is going to happen to them in advance as well [15:42] apw, Hm, I am rather adding something preventing a second convertion. Feels more straight forward [15:44] we should still release note the conversion so people can think about the risks, as they will be higher than normal for xen use, which is almost noone of course [15:46] apw, Agreed. I think that bug I am reporting right now can also get a release-notes task [15:51] smb, yeah agree [15:52] apw, bug 1303886 created [15:52] Launchpad bug 1303886 in xen (Ubuntu) "[Trusty] Xend managed domains disappear after upgrade" [Undecided,New] https://launchpad.net/bugs/1303886 [15:54] apw, Oh one thing that prevents the renaming. Unfortunately I have to take care about xen only and xen with libvirt and I would not know in the postinst which case [15:55] smb, i've added the release notes task [15:55] apw, ok [15:55] smb, if you can only "do it" once ever that seems fine, and safer [16:00] * cking drops in to listen to the canonical "plan" via icecast.. [16:00] oops [16:01] cking, New level of annoying elevator musak [16:01] arggh === gerald is now known as Guest12248 [16:34] apw: bugfix confirmed and commented on bug [16:34] apw: thanks [18:45] zequence, BenC new SRU cycle just started [18:58] bjf: Yep. I'm all done except for testing. === DavidDuffey is now known as dduffey === mjeanson_ is now known as mjeanson === ChickenCutlass_ is now known as ChickenCutlass === Prf_Jako1 is now known as Prf_Jakob