| === himcesjf_ is now known as him-cesjf | ||
| f_g | is the rationale for the ("tentative") choice of 4.15 for Bionic public somewhere? I already asked on-list some time ago (in reply to an even earlier question regarding the same), but there hasn't been a reply | 09:44 |
|---|---|---|
| apw | f_g, mostly it is about the tension between new h/w support which is useful for upcoming machines and stability from a longer baked kernel; that take with a view on when we will get the kernle relative to our cycle | 09:48 |
| f_g | apw: but why specifically 4.15 instead of 4.14, which is an upstream LTS release? is there some feature that landed in 4.15 that is more important than upstream LTS support? just "it's one cycle newer" seems kinda.. thin? :P | 09:52 |
| f_g | I know the general balance between newer kernel - better HW support - shiny new features / older kernel - less breakage - more stability | 09:54 |
| apw | f_g, we get input from vendors on where they are landing feature support, and that tends to be late in the game | 09:54 |
| apw | the upstream stable kernles are generally chosen to be non-alighned with our cycle | 09:54 |
| apw | we have to balance the work we will have to do porting stable fixes, against porting features | 09:55 |
| apw | though iirc there was a change to how long LTS stables will be supported which may well make them more attractive | 09:56 |
| f_g | apw: I am not sure whether the latter applies to upstream stable in general, or just certain ones? 4.4 got 6 years IIRC | 09:57 |
| apw | thye all started claiming to be 2, but i do think they may well be doing longer now | 09:57 |
| apw | 4.15 is still a tentative, and we will have a hard think before we move 4.14 -> 4.15 as we are not committed till that point | 09:58 |
| f_g | apw: understood. thanks for taking the time to answer - it looks surprising from an outsider / downstream perspective, especially since 16.04 had 4.4 | 09:59 |
| apw | f_g, that was somewhat lucky for sure | 09:59 |
| f_g | we moved from 4.4 to 4.10 and now 4.13 (downstream, based on Ubuntu kernel packages), and especially the move to 4.13 has been quite painful so we are a bit wary of 4.15 | 10:00 |
| apw | it is a difficult balancing act, newer is almost always better in terms of h/w support a big driver for desktop users who want the new shiney laptop to just work | 10:02 |
| f_g | I know - which is why we use Ubuntu's as a base - there is no other distro kernel with the same mix of uptodateness and HW testing/support ;) | 10:02 |
| apw | so we must be doing something right | 10:03 |
| f_g | :) | 10:03 |
| apw | we will be reviewing the plan as we always do as we get each kernel in and "stable", 4.14 is still bouncing around in -proposed | 10:04 |
| f_g | most of the pain with 4.13 was not Ubuntu's fault at all btw - just the usual upstream breakage. but much of it only got fixed in 4.15 and 4.14 stable updates, not in the short period of upstream 4.13 support | 10:05 |
| f_g | great, I'll keep an eye out for future updates / plans then! | 10:05 |
| apw | f_g, it is easier when there is an upstream LTS nearby to keep those updates flowing that is for sure | 10:06 |
| apw | 4.13 is as far from an upstream stable as you can be | 10:06 |
| f_g | indeed. there's only LP#1726519 left atm thankfully, but that seems to have stalled upstream | 10:09 |
| dupondje | pcieport 0000:00:1d.0: AER: Corrected error received: id=00e8 => any idea on such error? tested daily kernel, but still seems to happen :( | 12:52 |
| dupondje | 00:1d.0 PCI bridge: Intel Corporation Sunrise Point-H PCI Express Root Port #9 (rev f1) (prog-if 00 [Normal decode]) | 12:55 |
| dupondje | hmm :) | 12:55 |
| dupondje | oh my nvme disk is on that :( | 13:00 |
| dupondje | Is there some daily linux-firmware package available? | 18:29 |
| === ghostcube is now known as badcube | ||
| === badcube is now known as ghostcube | ||
| === hggdh is now known as desole | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!