dileks_webchat | hi | 07:08 |
---|---|---|
dileks_webchat | I am not sure if this is the correct channel for asking | 07:08 |
dileks_webchat | but according to release-schedule xenial-b1 shall be available for download | 07:08 |
dileks_webchat | today | 07:08 |
dileks_webchat | did that happen? if yes can you point me to the download-url? | 07:09 |
dileks_webchat | OK, I found https://wiki.ubuntu.com/XenialXerus/Beta1 which offers some iso-images | 07:16 |
dileks_webchat | away again, hehe | 07:20 |
snonez | hi, I have a question about kdump, http://stackoverflow.com/questions/29448158/enable-kdump-on-a-compiled-linux-kernel, | 09:09 |
snonez | the answer here doesn't solve the problem, | 09:10 |
apw | snonez, alwys helpfil to know what the problem is that you are seeing | 09:14 |
apw | and also if you could give us some background as to what you are trying to do | 09:14 |
Madkiss | We have just confirmed that https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1505948 also affects the Ubuntu standard kernel for Xenial :( | 09:18 |
ubot5` | Launchpad bug 1505948 in linux (Ubuntu Wily) "Memory arena corruption with FUSE (was Memory allocation failure crashes kernel hard, presumably related to FUSE)" [High,Confirmed] | 09:18 |
snonez | apw, the problem is here, http://stackoverflow.com/questions/29448158/enable-kdump-on-a-compiled-linux-kernel | 09:21 |
apw | Madkiss, this is the one (iirc) where if you use the "next" version of fuse which uses async IO things explode right ? | 09:22 |
Madkiss | apw: correct | 09:23 |
apw | Madkiss, phew, ok, i think we knew it would affect the main kernle, that this was likely a gneeric issue, i'll read the current status shortly | 09:24 |
apw | snonez, so your load fails with "no hole found" yes ? | 09:24 |
apw | snonez, what crashkernel= setting are you using on the kernel command line ? | 09:25 |
snonez | apw: yes, "Could not find a free area of memory of *** bytes..." | 09:27 |
snonez | "locate_hole failed" | 09:27 |
snonez | * failed to load kdump kernel | 09:27 |
snonez | apw: crashkernel=384M-:128M the default value | 09:27 |
apw | snonez, and this is a self compiled kernel i think, is the config somewhere ? | 09:27 |
apw | snonez, and is this your actual value: 0x9521000 bytes | 09:28 |
snonez | no, it comes from the orig config | 09:29 |
snonez | apw: 0x9ba5000 | 09:30 |
apw | snonez, so ... that is 163205120 bytes, or 155.6M which is more than 128M so ... it won't fit | 09:31 |
snonez | apw: make it bigger? | 09:32 |
apw | worth a try | 09:32 |
apw | snonez, as an experiment i'd make it -:256M and see if that makes the issue go away | 09:33 |
Madkiss | apw: I think the reason we're not seeing more reports of this until now is that the libfuse in 14.04 deoes not support AIO. | 09:34 |
snonez | apw, it works,, | 09:34 |
apw | snonez, did you say that you built your own kernel here, how big is your vmlinux that it wouldn't fix there | 09:34 |
snonez | I make it 256M and the error msg gone | 09:34 |
Madkiss | apw: once we have a newer version of libfuse, we're likely to see more users running into this. | 09:34 |
apw | Madkiss, yep this needs to be fixed, with luck we won't get hat version in xenial, so we have a little time to find and fix it in YY | 09:35 |
Madkiss | apw: I'm trying to figure out what exact libfuse version contains the problematic code. | 09:35 |
apw | Madkiss, thanks, if you could put that in the bug tooo so the info is all in one place | 09:36 |
apw | i really need a reproduce by, but i've not had time to sit with th eproblem and think about it since last time we talked | 09:36 |
snonez | apw: vmlinuz or vmlinux? | 09:37 |
apw | both | 09:37 |
snonez | apw, vmlinuz 6446944, vmlinux 356899822 | 09:38 |
apw | that doesn't seem so big, if you are loading vmlinuz, hrm | 09:38 |
* cking wonders if it's been allocated, one can check with cat /proc/iomem | grep Crash | 09:42 | |
snonez | apw, the orig kernel size(vmlinuz) is bigger than the one i build | 09:43 |
apw | snonez, yeah, all a bit odd | 09:45 |
dileks_webchat | hi apw long time no see/read | 09:53 |
dileks_webchat | bye | 09:58 |
snonez | apw, thx | 10:01 |
Madkiss | apw: my understanding is that the libfuse version in our local setup is libfuse3, which is currently not included in Ubuntu 14.04. | 10:09 |
Madkiss | I would, however, like to point out that under no circumstances should a kernel crash, and I am rather sure this is not even related at all to libfuse directly. | 10:09 |
Madkiss | I assume that libfuse3 triggers a function that makes the kernel run into a panic. However, any user being allowed to use fuse will likely be able to do the same, and hence crash the system. | 10:10 |
Madkiss | apw: Would you be willing to look into the problem if we deliver some piece of code that reliably crashes a Xenial system where fuse is in the kernel? | 10:14 |
Madkiss | when executed as a user? | 10:14 |
apw | Madkiss, concur you should never be able to crash a kernel, the primary concern close to release is that this is not osmething that users will hit easily | 11:12 |
apw | Madkiss, if you have a reproduce by that will trigger it then yes i would like to have me or someone here look at it | 11:13 |
luc4 | Hello! My ubuntu laptop seems to have a couple of issues when resuming from stand by. Both keyboard and trackpad frequently wake up dead. I'd file a bug report for this, but what kernel should I use for the test? Latest in repo? I'm typically asked to install some other kernels from some other source. | 11:25 |
apw | luc4, you should update to the archive latest and confirm the bug there, then report it against that | 11:38 |
apw | luc4, the next steps are often to test other "nearby" kernels to determine where if any a fix exists | 11:39 |
luc4 | apw: sorry, pc went to standby :-D so here http://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D? v4.5-rc5-wily? | 11:46 |
apw | luc4, no as in letting update manager install the latest (or apt-get dist-upgrade) | 11:47 |
apw | we care that that one is broken | 11:47 |
apw | (or not | 11:47 |
apw | and then you will likely get asked to test some of those mainline kernels to confirm if like v4.5 is broken as part of diagnosis | 11:47 |
luc4 | apw: sorry, not sure I understand. I'm now using the latest from apt. Is there anything else I should test? | 11:48 |
apw | but your bug should be filed from an ubuntu kernel which has the issue if at all possible (some bugs not so much) | 11:48 |
apw | you asked wehre you should file a bug, file it against archive latest | 11:48 |
apw | yes you can then test the highest v4.5-rc that you found | 11:49 |
luc4 | apw: ah ok, now I'm using 4.2.0-30-generic. I should file against that. | 11:49 |
apw | yes because that is what you want fixed, the archive kernel for your series | 11:50 |
luc4 | apw: thanks! | 11:50 |
snonez | apw, sry to bother you again, how many minutes does kdump takes, which level do you set? | 12:17 |
snonez | I set level 17, x86_64, 18 minutes, still running | 12:18 |
Madkiss | apw: alright, I will see what I can do. | 13:07 |
apw | 18m sounds like a long time to dump to me, but i can't say i dump large amchines very often | 13:17 |
tjaalton | apw: hey, I'm not seeing my stuff on xenial master-next? tseliot would rebase his on top because some core drm commits are shared | 14:16 |
apw | tjaalton, UBUNTU: SAUCE: i915_bpo: Revert "drm/i915: Switch DDC when reading the EDID" looks to be on there to me | 14:20 |
tjaalton | apw: hmm where is that? I have lp as the main source | 14:21 |
tjaalton | something happening now :) | 14:22 |
tjaalton | or just my local clone not set up properly | 14:24 |
tjaalton | git fetch does nothing | 14:24 |
tjaalton | git fetch origin works | 14:24 |
tjaalton | yep, it's all good | 14:25 |
lamont | jsalisbury: I stand corrected - the 300 retry kernel resulted in one display, not two | 17:57 |
lamont | bug updated | 17:58 |
jsalisbury | lamont, ack, thanks for testing. I'll take a look | 17:58 |
jsalisbury | lamont, I'm talking with upstream about the bug. I posted a link to the upstream thread in the bug. Can you post the monitor type and the type of cable used? | 18:30 |
lamont | sure | 18:30 |
jsalisbury | lamont, they state that a single link dvi/hdmi cable didn’t work and dual link worked | 18:30 |
lamont | single link dvi-d cable it is | 18:32 |
jsalisbury | lamont, great, thanks. I'll update upstream | 18:36 |
lamont | not sure whether or not a dual-link is possible for the other end | 18:36 |
lamont | nah... it looks like it could take dual-link | 18:37 |
* lamont jsut deosn't have one | 18:37 | |
jsalisbury | lamont, ack, it looks like they are coming up with patches, so I can build a new kernel for you once one is available. | 18:39 |
lamont | nice | 18:40 |
sorinello | hello. can someone giver me an estimate until https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1548587 hits the official channels ? | 20:33 |
ubot5` | Launchpad bug 1548587 in linux (Ubuntu Wily) "Ubuntu 15.10 VMWare guest won't show UI after upgrading to 4.2.0-30" [High,In progress] | 20:33 |
apw | kamal, ^ | 20:33 |
kamal | sorinello, apw: the team is discussing that now. I'll let you know when we know. | 20:35 |
sorinello | thanks a lot kamal , apw | 20:37 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!