<alyssa---->
robher: tomeu: By the way, are we still aiming for 5.2?
<anarsoul>
isn't it a bit late assuming 5.1 window is closed?
<Lyude>
yeah
<Lyude>
better to wait for 5.2 at this point
<alyssa---->
Lyude: That's what I said...?
<Lyude>
Oh
<Lyude>
I think we both misread your post at the same time
<Lyude>
what a coincidence
<alyssa---->
tomeu: robher: Nothing on the TODO list looks like a blocker, I think
<robher>
alyssa----: there is the question of doing GPU VA space management in userspace instead.
<alyssa---->
robher: :/
vstehle has quit [Ping timeout: 245 seconds]
NeuroScr has quit [Quit: NeuroScr]
<alyssa---->
Time to look into cubemaps, I think
<alyssa---->
At least enough so we go from "crash" to "fail" :P
jolan has quit [Quit: leaving]
jolan has joined #panfrost
NeuroScr has joined #panfrost
jolan has quit [Read error: Connection timed out]
jolan has joined #panfrost
<alyssa---->
Oh, this is some interesting lowering
<alyssa---->
Well, an opt I guess
<alyssa---->
TL;DR "hard to compute constants" can be moved out of a fragment shader and into an intermediate shader stage and fed in as a uniform
_whitelogger has joined #panfrost
NeuroScr has quit [Quit: NeuroScr]
vstehle has joined #panfrost
<tomeu>
cubemaps!
<tomeu>
robher, alyssa----: will look at that in a few days
<tomeu>
at the very least, to have a better idea of how the UAPI would look like
_whitelogger has joined #panfrost
afaerber has quit [Quit: Leaving]
afaerber has joined #panfrost
chewitt has joined #panfrost
afaerber has quit [Quit: Leaving]
chewitt has quit [Quit: Zzz..]
chewitt has joined #panfrost
chewitt has quit [Quit: Zzz..]
chewitt has joined #panfrost
<tomeu>
alyssa----: so, the problem with -brefract on weston (and possibly any other scene) is that glmark2 tries to run there at a much higher rate than 60fps
<tomeu>
and the DRM backend isn't blocking after every job as the non-drm does
<tomeu>
instead, it returns a fence that gallium can choose to wait on or not
<tomeu>
what gallium doesn't know is that jobs share scrap buffers
<tomeu>
so it doesn't wait before a new render starts, and panfrost overwrites data that current jobs depended on
<tomeu>
if I actually wait for rendering to finish in force_flush_fragment, stuff looks fine
MoeIcenowy has quit [Quit: ZNC 1.6.5+deb1+deb9u1 - http://znc.in]
MoeIcenowy has joined #panfrost
chewitt has quit [Read error: Connection reset by peer]
chewitt has joined #panfrost
<chewitt>
@alyssa---- you posted this patch to the list but I didn't see it get merged
<chewitt>
after mesa changes merged last night things are quite fluid in the Kodi GUI now
<chewitt>
still some mem leakage over time
<chewitt>
no change to the dim/black screensavers refusing to run
<chewitt>
i've switched my test codebase to the drm driver now :)
Elpaulo has quit [Read error: Connection reset by peer]
Elpaulo has joined #panfrost
afaerber has joined #panfrost
chewitt has quit [Remote host closed the connection]
<alyssa---->
tomeu: You say it like "if I actually wait for rendering to finish in force_flush_fragment" is a hack, but what do you think "Force flush" implied? ;P
<alyssa---->
chewitt: Yeah, that patch is a major hack. The real solution will take a lot more thinking (to avoid the performance disaster associated with that patch) but other items are higher-prio... so this way you can merge the patch for Kodi and have those bugs fixed, and I can continue to get reliable perf numbers? :P
<alyssa---->
Might have to revisit it soon for dEQP, though
<alyssa---->
Happy to hear the rendering glitches fixed themselves. Not really sure what the culprit was, since I was fixing a lot of rather obscure stuff, but hey :P
<tomeu>
alyssa----: yeah, that was just to explain what was going on
<alyssa---->
:+1:
<Lyude>
tomeu: btw, curious: will panfrost's kernel driver have SAME_VA and such like kbase?
<tomeu>
Lyude: I think it's not clear atm, robher is taking care of all the MM stuff
<alyssa---->
there's discussion on the list rn about this
afaerber has quit [Quit: Leaving]
pH5 has joined #panfrost
NeuroScr has joined #panfrost
<robher>
Lyude: For my own education, why do you care or want that?
<Lyude>
robher: I'm honestly just curious
<Lyude>
since you guys had been talking about doing VAs in userspace earlier
<robher>
Lyude: I'm just trying to gauge whether we really need to support it from the start.