19:02 karolherbst: Lyude: do you want to fix the SCHED_ERROR thing as well?
19:03 Lyude: karolherbst: yeah - I can send out a patch for it in a bit
19:03 karolherbst: wait a sec
19:03 karolherbst: Lyude: https://gist.githubusercontent.com/karolherbst/1b80ca96f334ebe4347b9dc1f8d8e562/raw/5d380824dfe19fbdec7fe814461f016643897040/gistfile1.txt :D
19:03 karolherbst: but I hoped you could just add it to yours
19:04 karolherbst: hitting those is usually still super fatal, but at least _some_ things work a little better
19:04 Lyude: karolherbst: ahhh lol, yeah I just wanted to push it since I have a habit of sometimes forgetting :)
19:04 Lyude: and yeah. imho considering this makes it more likely for us to get more useful debugging info from customers by default if their kernel logs would have been underrun otherwise, seems worth it to me
19:04 Lyude: *overru
19:04 Lyude: **overrun
19:04 karolherbst: yep
19:05 karolherbst: that SCHED stuff really hits the kernel though... ssh and stuff still runs, but kernel compilation becomes impossible because the entire IO system is overflooded
19:05 karolherbst: 99% of the time you wait on flushes
19:05 Lyude: karolherbst: either way though that patch is: Reviewed-by: Lyude Paul <lyude@redhat.com>. Feel free to just send it onto the ml for the record and push it after
19:05 Lyude: karolherbst: yeah, dmesg is surprisingly slow
19:06 karolherbst: yeah.. probably
19:06 Lyude: I suppose that's why we have ftrace though
19:07 karolherbst: yeah.. ftrace is awesome, just a little annoying to use