=== amitk_ is now known as amitk | ||
apw | leitao (N,BFTL), no, ppc64el is not compatible with 7.2 without patching right now | 11:35 |
---|---|---|
apw | gcc 7.2 | 11:35 |
tseliot | apw: hi, can you approve nvidia-graphics-drivers-384 in artful NEW, please? | 13:25 |
apw | tseliot, will look shortly | 13:26 |
tseliot | apw: thanks! | 13:26 |
apw | tseliot, does this only upgrade -375, does it not also upgrade -367? are you relying on the chain there ? | 13:40 |
tseliot | apw: yes, it only upgrades 375. 375 had transitional packages for 367 | 13:41 |
tseliot | which should still be there in the archive | 13:41 |
apw | tseliot, and done | 13:52 |
tseliot | apw: thanks a lot! | 13:52 |
apw | tseliot, and it has cratered on armhf | 13:55 |
apw | sed: can't read NVIDIA-Linux-armv7l-gnueabihf-384.90/nvidia_icd.json.template: No such file or directory | 13:55 |
tseliot | apw: let me check | 13:55 |
tseliot | oh | 13:56 |
tseliot | so they only introduced the change on x86 and x86_64 | 13:56 |
tseliot | I'll fix it | 13:56 |
tseliot | apw: 384.90-0ubuntu2 fixes that on armhf | 14:19 |
tjaalton | apw: I should probably upload the oem kernel to xenial NEW.. | 14:39 |
tjaalton | soon | 14:39 |
apw | tjaalton, you would need to upload it to the CKT PPA | 14:39 |
apw | so it gets built right | 14:39 |
tjaalton | ah, ok | 14:45 |
leitao_ | It seems that mainline ppc64el kernels are failing to build | 15:04 |
leitao_ | http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.14-rc3/ | 15:04 |
leitao_ | I am wondering if this is a toolchain being more strict issue. | 15:05 |
apw | leitao_, yes, it is the compiler being more strict, i think sforshee sent a fix for that somewhere | 15:06 |
leitao_ | nice. | 15:06 |
leitao_ | sforshee, is it a kernel issue? If so, I can pester the powerpc maintainer to get it accepted. | 15:08 |
apw | leitao_, we definatly applied something to the kernel source to avoid/fix the issue | 15:10 |
leitao_ | apw, but it does not fix the mainline builds, right? | 15:11 |
apw | nope, they will need the fix to land upstream, or some manual action | 15:12 |
apw | them failing is an indicator that the mainline kernel is broke | 15:12 |
leitao_ | apw, by the way, did you see that there is a regression on artful/master-next on P9? | 15:38 |
apw | leitao_, got a pointer to the regresion ? | 15:40 |
leitao_ | apw, yes | 15:41 |
leitao_ | sent to the kernel-team mailing list. | 15:41 |
leitao_ | apw, "[Artful][PATCH 03/16] powerpc/perf: Add nest IMC PMU support" | 15:41 |
leitao_ | A patchset that we sent and it seems broken | 15:42 |
leitao_ | a full revert would be ideal | 15:42 |
apw | sforshee, ^ | 15:42 |
=== leitao_ is now known as leitao | ||
apw | leitao, well damn that is poor timing | 15:45 |
leitao | apw, :-| | 15:46 |
leitao | apw, and it breaks it completely, not even booting a p9. | 15:46 |
apw | leitao, how did that get past testing, given it is a patchset for P9 | 15:47 |
leitao | apw, that is a question we are trying to understand internally. | 15:48 |
leitao | Gustavo said he tested and it worked. | 15:48 |
leitao | But I was not sure if he enabled the CONFIG_VAS option | 15:48 |
leitao | which really enables the feature. | 15:48 |
apw | leitao, anyhow, i'll poke people to sort that out, as right now the kernel planned for the final kernel for artful has that junk applied | 15:48 |
leitao | apw, ok! | 15:49 |
leitao | apw, sorry about it again. We are addressing it internally. This must never happen and it will certainly bring a lot of burden to you. | 15:52 |
apw | leitao, it is particularly painful as this is the last upload for artful, for kernel freeze | 15:52 |
leitao | apw, yup | 15:53 |
=== ghostcube_ is now known as ghostcube | ||
sforshee | leitao: http://lkml.kernel.org/r/20170928133340.19312-1-seth.forshee@canonical.com is the patch I sent for the boot regression | 18:26 |
sforshee | leitao: I'll revert that series and respin artful | 18:27 |
sforshee | leitao: sorry, for the build regression | 18:28 |
leitao | sforshee, sorry for the regression. :-) | 18:28 |
leitao | sforshee, I am not sure if powerpc maintainer will find you patch at LKML. They usually look at the powerpc subsystem. | 18:29 |
leitao | let me pester him with it. | 18:29 |
sforshee | leitao: I sent it to the linuxppc-dev list, maintainers also included too | 18:29 |
sforshee | but also to lkml | 18:29 |
leitao | ah, ok! | 18:29 |
sforshee | that's just the easiest way to give a link, I only need the message id | 18:30 |
sforshee | leitao: also http://lkml.kernel.org/r/20170928133426.19454-1-seth.forshee@canonical.com for a selftest build failure | 18:31 |
leitao | sforshee, just pestered the powerpc maintainer. I hope to have a review by tomorrow | 18:31 |
sforshee | both of those are because of a new compiler | 18:31 |
leitao | yea. these compiler guys.... :-P | 18:32 |
sforshee | leitao_: you still around? | 20:47 |
leitao_ | sforshee, I am | 20:47 |
sforshee | leitao_: okay, let me double check with you which commits I should revert | 20:47 |
leitao_ | sforshee, I would revert the whole series. It seems to be completely broken. | 20:48 |
sforshee | what I did was the ones for bug 1718293 | 20:48 |
ubot5 | bug 1718293 in The Ubuntu-power-systems project "[Ubuntu 17.10] POWER9 - Base - Integrate P9 VAS (Virtual Accelerator Switchboard) support in kernel" [High,Fix committed] https://launchpad.net/bugs/1718293 | 20:48 |
leitao_ | I understand that disabling VAS=y would also work | 20:48 |
leitao_ | sforshee, that is the bug. | 20:48 |
sforshee | but I wonder if the ones for bug 1718292 should also be reverted or not | 20:48 |
ubot5 | bug 1718292 in The Ubuntu-power-systems project "POWER9: NX842 module changes" [High,Incomplete] https://launchpad.net/bugs/1718292 | 20:48 |
leitao_ | sforshee, were patches from 1718292 commited? | 20:49 |
leitao_ | nm, it seems so. | 20:50 |
sforshee | yeah | 20:50 |
leitao_ | sforshee, I do not know, let me try to discover. | 20:51 |
sforshee | thanks | 20:51 |
leitao_ | sforshee, you will need to revert 1718292 also. | 21:01 |
sforshee | leitao_: ack, thanks | 21:06 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!