=== BenC- is now known as BenC [10:56] erle- (N,BFTL), they are vanilla almost completely === vHanda_ is now known as vHanda [11:51] henrix, will you likley pick up smb's patches for v3.16 stable or should i review them for application to utopic direct [11:56] apw: i haven't look yet but i guess i'll pick them for stable [11:56] apw: actually, i believe one of the commits was already tagged for stable [11:56] apw: i'll have a look at them this afternoon [11:58] henrix, ahh yes so it is, and the other foundational, ok good. [12:51] henrix, i've reviewed them best i can anyhow, whether they come in via stable or direct makes no odds [12:53] apw: ack, thanks. i'll review them as well in a bit (and ACK them) [13:57] jibel, yo ... seems we still have a slew of failures on 3.2.0-73-generic dkms tests, do we have any idea if this is testbed or kernel yet ? [13:59] apw, hey, no sorry, I'm swamped in phone testing and didn't have time to look at it more than yesterday. I'll try to capture the output of the console. [13:59] henrix, have we had any complaints booting 3.2.0-73-generic kernels at all in VMs ? we see a lot of the DKMS testing fail during boot, but that might not be anything to do with the kernel [14:00] apw: not that i'm aware, no [14:01] henrix, me either thanks [14:02] apw: usually joe ping us when this sort of regressions occur. and he didn't :) [14:02] heheh yeah i was thinking the same [16:01] In kernel 3.18RC1 this kernel config perameter is: # CONFIG_SCSI_MQ_DEFAULT is not set [16:01] In kernel 3.18RC2 and beyond, the kernel config permeter is: CONFIG_SCSI_MQ_DEFAULT=y [16:01] Resulting in loss of the ability to set the IO scheduler via /sys/block/sda/queue/scheduler [16:01] From the add a CONFIG_SCSI_MQ_DEFAULT option commit message: [16:02] > Add a Kconfig option to enable the blk-mq path for SCSI by default [16:02] > to ease testing and deployment in setups that know they benefit [16:02] > from blk-mq. [16:02] How do we know that all systems benifit from blk-mq? [16:02] It seems complicated to have to re-compile the kernel to get the other IO scheduler options back. [16:02] By the way, my system does seem to benifit from blk-mq, I just didn't understand why I couldn't observe and change the IO scheduler anymore, and so isolated the change. [16:21] dsmythies, hmmm [16:23] dsmythies, could you file a bug against linux for me with details of that, as though it affects no kernels in the wild, it will soon [16:23] dsmythies, and shove the bug # in here [16:28] apw: O.K., but give me some time to do so. [16:42] dsmythies, great, as soon as i have that i can write up my findings and get it changed [17:14] I cann't seem to set "linux" anymore. I think I have in the past. Anyway, bug 1397061 against "Ubuntu" [17:14] bug 1397061 in Ubuntu "Kernel Config setting: CONFIG_SCSI_MQ_DEFAULT" [Undecided,New] https://launchpad.net/bugs/1397061 [17:15] dsmythies, ack i'll sort it out [17:26] apw: what is the kernel command line option to enable blk-mq? [17:39] dsmythies, something like scsi.use_blk_mq=yes/no [17:40] dsmythies, and it may well switchable at runtime [17:40] /sys/module/scsi_mod/parameters/use_blk_mq [17:45] scsi_mod.use_blk_mq=yes/no perhaps [18:33] apw: thanks [19:21] zequence: https://bugs.launchpad.net/ubuntu/+source/linux-lowlatency/+bug/1396193 is all you. [19:21] Launchpad bug 1396193 in Kernel SRU Workflow "linux-lowlatency: -proposed tracker" [Medium,In progress]