[02:57] <sasha_> Does anyone here have instructions to create a kernel patch from a github driver file? Thanks.
[19:59] <awreece> I'm noticing a behavior I'm having difficulty explaing -- my application is doing exceptional amounts of `mmap`s from many threads and is contending in the kernel on `mm->mmap_sem`. My expectation is that CPU profiling would reveal this contention; however, I appear to be missing it. 
[19:59] <awreece> If I use the sched:sched_switch tracepoints with perf, I can very easily see that I'm getting dequeued
[20:02] <awreece> I'm surprised that I don't see more CPU usage here: I guess `rwsem_down_read/write_failed` immediately deschedule if it fails?
[20:05] <awreece> I must be blind, I don't see anything in the tree documenting rwsem
[20:05] <awreece> `git grep rwsem -- Documentation`
[21:44] <Noskcaj> Is there any recent change that would cause 15.10 to fail to upgrade to 16.04 with linux-image-4.3.0-2-generic install failure? I think it's lp:1527854