| === henrix_ is now known as henrix | ||
| * ppisati -> reboot | 07:16 | |
| ppisati | smb: guess what? it doesn't work... | 07:19 |
|---|---|---|
| ppisati | smb: i can hear you, but can't answer... | 07:19 |
| smb | I realized soon agter I tried to talk to you... | 07:20 |
| smb | after | 07:20 |
| === fmasi_afk is now known as fmasi | ||
| apw | smb, so it is possible. boot to grub, edit the kernel command line, fix root= to the new dm mapper name for root and add break=mount | 09:28 |
| apw | smb, at that point lvm vgrename the group and let the boot continue, update-grub, grub-install and reboot again | 09:28 |
| smb | apw, Ok, at least one method working. Not sure a rename reboot and change root= and then update-grub again would not work as well. And not sure either the install-grub thing is needed there at all... | 09:31 |
| apw | smb, if you rename then reboot you have to incant at grub-rescue to get to to the grub prompt | 09:33 |
| apw | it is easier to incant at the normal menu thing :) | 09:33 |
| smb | Yeah I think | 09:33 |
| * smb reboots | 11:30 | |
| ppisati | brb | 11:46 |
| * henrix -> lunch | 11:59 | |
| === fmasi is now known as fmasi_lunch | ||
| === fmasi_lunch is now known as fmasi | ||
| amitk | cking: you used to have a script to exercise suspend/resume a bazillion times, did you have something for hibernate? | 12:53 |
| cking | amitk, it's in fwts, e.g. use fwts s4 --s4-multiple=10000 | 12:54 |
| amitk | cking: urghhh... tied to x86? | 12:54 |
| cking | more like pm-utils | 12:54 |
| amitk | cking: the s4 worries me, I wanted to point to it for ARM platforms | 12:55 |
| apw | amitk, how does one ask for hib on arm | 12:55 |
| apw | echo disk >/proc/thingy ? | 12:55 |
| amitk | apw: we're fixing that :) | 12:55 |
| amitk | apw: by adding the required hooks | 12:55 |
| apw | what i mean is pmutils uses ech | 12:56 |
| apw | echo mem or echo disk into /proc, if you respond to those then i suspect things will work | 12:56 |
| amitk | apw: right, that is the only interface i want to depend on, I'll have a look at what fwts expects. I'm hoping it isn't too tied to x86'isms | 12:57 |
| amitk | cking: fwts git tree? | 12:57 |
| cking | amitk, git://kernel.ubuntu.com/hwe/fwts | 12:57 |
| amitk | apw: echo mem > /sys/power/state interface, that is | 12:57 |
| cking | amitk, you may get away with faking a pm-hibernate script, that's all it uses | 12:58 |
| amitk | cking: we pushed some patches to migrate workqueues and other irqs away from idle cores in 3.10 and 3.11, you might find them useful | 12:59 |
| cking | amitk, for ARM specifically, or generic arch changes? | 13:00 |
| amitk | cking: most of our work is core frameworks, the only ARM-specific stuff is driver consolidation | 13:00 |
| cking | amitk, so I could theoretically be able to measure this on x86 if I enable CONFIG_WQ_POWER_EFFICIENT_DEFAULT=y | 13:04 |
| * cking pops it on the TODO list | 13:05 | |
| amitk | cking: depends on your idle profile (drivers using wqs), we only converted 3-4 subsytems to use POWER_EFFICIENT WQs by default. There might be more subystem in use on x86 | 13:05 |
| amitk | cking: driver writers who don't have to worry about performance (cache) are starting to convert their drivers to use this now | 13:06 |
| amitk | cking: so yes, we see improvements | 13:07 |
| amitk | idle states are deeper and longer | 13:07 |
| cking | ack | 13:08 |
| === cmagina_ is now known as cmagina | ||
| amitk | cking: will try to dump info and exchange notes with you guys as time permits, gtg. Bye | 13:26 |
| cking | amitk, ok, I'm always around to help | 13:26 |
| amitk | cking: thanks | 13:26 |
| tyrant | i have a kernel 3.2.0-49 installed and i want to rebuild a module for that kernel, where do i get the source code for that particular version (same vermagic)? the sources apt gives me has this vermagic | 13:59 |
| tyrant | 3.2.46 SMP mod_unload 686 | 13:59 |
| lfaraone | If you're running the raring kernel on 12.04 LTS and upgrade to 12.10, what happens to your kernel version? | 14:15 |
| === kentb-out is now known as kentb- | ||
| === kentb- is now known as kentb | ||
| === fmasi is now known as fmasi_afk | ||
| apw | lfaraone, from P+R kernel to real R ? the kernel would be near identicle, so you would upgrade to the replacement kernel | 18:41 |
| apw | which would be basically the same bits | 18:41 |
| lfaraone | apw: P+R kernel to Q. | 18:42 |
| apw | oh, you will end up with both kernel still i expect | 18:42 |
| apw | so you would probabally continue to use the R kernel, which isn't going to update | 18:42 |
| lfaraone | apw: Is the R kernel offered as a package in Q? | 18:42 |
| apw | hmmm not idea | 18:42 |
| apw | nope | 18:43 |
| apw | though that seems like an odd upgrade | 18:43 |
| apw | cirtainly odd to move from P+R to Q and not then on to R | 18:44 |
| lfaraone | So if you install 12.04.3 or whatever with the HWE raring kernel and then want to upgrade to the (still supported) 12.10 but not 13.04, because lets say there's some incompatability with your frobnicator. | 18:45 |
| lfaraone | You end up in a situation without updates :/ | 18:45 |
| lfaraone | I'm trying to figure out if I need to backport the 3.8-enablement patches for OpenAFS to PQ or just P. | 18:48 |
| apw | lfaraone, that does sound like an accurate statement, not idea | 18:49 |
| apw | ideal indeed | 18:49 |
| apw | i wonder if we can force the lts backports off on upgrade | 18:50 |
| lfaraone | apw: that would be unfortunate if I actually need the R kernel for my hardware, though, right? | 18:50 |
| apw | if you arn't supported on Q in theory you cannnot upgrade from P+R though Q anyho | 18:50 |
| apw | as you should have never been able to install Q in the first place, but would have been forced to R | 18:51 |
| apw | but this is all far from ideal in either direction | 18:51 |
| * apw makes sure it is on our next team agenda | 18:57 | |
| * apw calls it a day | 18:57 | |
| kentb | do we typically update pciutils with a new set of pci.ids in a stable release? for example, 12.04.3? There are some new Ivy Bridge servers on the horizon that'll need the id bump so they're recognized properly in 'lspci'. Or is the answer to run 'update-pciids' (which seems to work)? | 19:56 |
| kentb | 'they're' meaning memory controllers and the like | 19:57 |
| === shadeslayer_ is now known as shadeslayer | ||
| mojo706 | hello wonderful people on the Kernel Team | 22:19 |
| mojo706 | please assist this user http://ubuntuforums.org/showthread.php?t=2164127 | 22:19 |
| mjg59 | mojo706: It's not a kernel bug - the panic is because it can't start init | 22:32 |
| mjg59 | And it can't start init because your hard drive is dead | 22:32 |
| mojo706 | mjg59, but I have read somewhere else that it got fixed after a reinstall of Ubuntu but in this case there are important files. | 22:34 |
| ShapeShifter499 | hi | 23:49 |
| ShapeShifter499 | I have a nook hd tablet that as far as I know is OMAP4. Do I have to do any edits to ubuntu arm to get it to boot and run off my system? | 23:49 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!