=== llstarks_ is now known as llstarks
=== llstarks_ is now known as llstarks
[06:49] <mlankhorst> Hello, world!\n
[06:50] <llstarks> hey mlankhorst 
[06:50] <mlankhorst> morning
[06:50] <llstarks> did your fencing work make it to sumit and airlied in time for 3.11?
[06:50] <mlankhorst> not the fencing itself, it wasn't planned for it
[06:50] <mlankhorst> but the wound/wait mutexes did
[06:51] <llstarks> does this allow in-order frames?
[06:52] <mlankhorst> not yet
[06:52] <mlankhorst> intel needs to be modifiied to support fencing internally first
[06:55] <mlankhorst> and what I did was a hack to only support it on the command submission buffers
[06:56] <llstarks> cool
[06:56] <mlankhorst> but it should be possible, it just means disintegrating the dev_struct lock in intel that currently protects everything
[06:57] <llstarks> you lost me heh. drm and 3am don't mix
[06:58] <llstarks> going through your just-now commits and the 3.11 merges. looks nice
[07:01] <mlankhorst> everything in for-airlied and for-airlied-next is queued
[07:01] <mlankhorst> rework-v11 is going through core/locking
[07:03] <llstarks> i'll go through the mutex stuff in the morning when i'm rested enough to trace it. always enjoyed mutexes and spinlocks when learning c
[07:04] <mlankhorst> llstarks: https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git/log/?h=core/mutexes
[07:06] <mlankhorst> Those lockdep tests saved my life once already, when one of the core maintainers had a locking selftest failure when he merged another mutex patch that wasn't adjusted for the w/w mutex changes. :P
[07:36] <RAOF> mlankhorst: Oh, hai! You wouldn't happen to know off the top of your head how I stop the outputless x-x-v-ati from binding on my hybrid system?
[07:51] <mlankhorst> RAOF: erm not from head, you could disable autoAddGPU
[07:54] <mlankhorst> that disables the initial scan at least :P
[08:31] <tjaalton> hmm the soft cursor with mir is kinda distracting :P
[08:32] <tjaalton> also doesn't seem to use mouse acceleration
[08:45] <RAOF> tjaalton: It should use acceleration; it's using the full X input stack.
[08:45] <tjaalton> RAOF: the x cursor does, but the big soft one doesn't
[08:46] <RAOF> You mean the big partially-transparent one?
[08:46] <tjaalton> yes
[08:46] <RAOF> That's the hardware cursor. Yes, indeed, it doesn't have mouse acceleration.
[08:46] <tjaalton> ah, i heard it was software.. anyway, ok
[08:46] <RAOF> The X cursor is software.
[08:47] <tjaalton> so I got it backwards then
[08:48] <RAOF> Right.
[08:48] <RAOF> Also, does that cursor move for you? It doesn't with synaptics. At least once X starts.
[08:48] <tjaalton> it does with the thinkpad nipple
[08:48] <RAOF> Ah, evdev's not exclusive then.
[09:23] <tjaalton> suspend seems to fail with xmir
[09:34] <RAOF> As in: fails to suspend, or fails to resume?
[09:34] <tjaalton> fails to suspend
[09:34] <RAOF> The latter might be failure on our part to do the intricate VT dance
[09:35] <tjaalton> though this is a quirky old thinkpad so it might be something else
[09:35] <RAOF> Does it try to suspend?
[09:35] <tjaalton> yeah, the crescent moon is blinking
[09:36] <RAOF> <shrug>
[09:36] <RAOF> Interesting.
[09:38] <tjaalton> hmm, hitting the hotkey-combo "freezes" the machine, meaning it changes vt
[09:38] <tjaalton> oh, now it suspended fine
[09:39] <tjaalton> resume worked
[09:46] <tjaalton> so.. quirky old laptop :)
[10:14] <ricotz> mlankhorst, hi :)
[10:15] <ricotz> mlankhorst, i am hoping you can get pixman updated to 0.30
[10:19] <mlankhorst> ricotz: ideally through debian first
[10:20] <ricotz> mlankhorst, right, altough you updated 0.28 in debian git for ubuntu only too
[10:20] <mlankhorst> yeah but that was due to freeze
[10:20] <ricotz> ok
[11:04] <RAOF> Hah. xf86-video-ati doesn't work particularly well when you hand the fd of an i915 device.
[11:04] <RAOF> Who'd've thought?
[11:05] <mlankhorst> RAOF: I caught a kernel panic in nouveau that way
[11:05] <RAOF> mlankhorst: What, handing a nouveau fd off to i915 or something?
[11:06] <mlankhorst> yeah
[13:09] <mlankhorst> I'm running a torture test on nouveau for some fun
[13:11] <mlankhorst> running the max-texture-size piglit test until things break
=== llstarks_ is now known as llstarks
[15:29] <smartboyhw> bryce, private message?
=== llstarks_ is now known as llstarks
[17:12] <Sarvatt> ricotz: go figure, right after you updated https://devtalk.nvidia.com/default/topic/549155
[17:20] <llstarks> dammit nvidia, still no egl?
[17:21] <llstarks> they put out wddm1.3 drivers for windows 8.1 but oh no, experimental stuff for linux is offlimits
=== llstarks_ is now known as llstarks
=== llstarks_ is now known as llstarks
[20:26] <ricotz> Sarvatt, hehe, i knew it