alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - https://gitlab.freedesktop.org/panfrost - Logs https://freenode.irclog.whitequark.org/panfrost - <daniels> avoiding X is a huge feature
stikonas has quit [Remote host closed the connection]
<alyssa> tomeu: Also, a bunch of fbo.render.* are flip-flops ...
vstehle has quit [Ping timeout: 250 seconds]
robclark has quit [Read error: Connection reset by peer]
robclark has joined #panfrost
ric96 has quit [Ping timeout: 258 seconds]
ric96 has joined #panfrost
<alyssa> You know starting to think all of fbo.* is a flip-flop. Bleh.
* hanetzer holds his neovim close :3
<hanetzer> so, finals time for me, just took one today. the immediatly gradable multiple choice portion of the exam netted me a 95% average on it
<alyssa> Congrats :)
<alyssa> hanetzer: If we're judging by conformance, I'm not doing so well on tests myself ...
anarsoul has quit [Read error: Connection reset by peer]
anarsoul has joined #panfrost
<alyssa> I think we just hit B- for GLES2, tho
<alyssa> Implementing programmable blending would get us up to a B+, though
<anarsoul> alyssa: nice
<HdkR> GL_MESA_programmable_blending
_whitelogger has joined #panfrost
<alyssa> Did some compiler work tonight, down to 2524 failures. That's a win!
<alyssa> tomeu: (See the for-alyssa branch)
<alyssa> tomeu: Also, um, if a test that was expected to fail now passes, why does that cause CI to fail? This is bound to be a productivity nosediver, if I have to wait 40 minutes (!) from a successful commit fixing tests to, you know, moving on..
vstehle has joined #panfrost
<tomeu> alyssa: because if we don't remove the expected failure when a test starts to pass, we won't realize when a regression comes in and starts to fail again
BenG83 has joined #panfrost
<tomeu> alyssa: you can still send the patches for review and move on while the CI does its thing
<tomeu> and consider the CI results as if it was one more review comment
BenG83 has quit [Quit: Leaving]
_whitelogger has joined #panfrost
pH5 has joined #panfrost
cwabbott_ has joined #panfrost
stikonas has joined #panfrost
cwabbott has quit [Ping timeout: 245 seconds]
cwabbott_ is now known as cwabbott
stikonas has quit [Remote host closed the connection]
raster has joined #panfrost
raster has quit [Ping timeout: 252 seconds]
raster has joined #panfrost
cwabbott_ has joined #panfrost
cwabbott has quit [Ping timeout: 258 seconds]
cwabbott_ is now known as cwabbott
_whitelogger has joined #panfrost
rhyskidd has joined #panfrost
Elpaulo has joined #panfrost
afaerber has joined #panfrost
rhyskidd has quit [Ping timeout: 258 seconds]
rhyskidd has joined #panfrost
rhyskidd has quit [Ping timeout: 245 seconds]
rhyskidd has joined #panfrost
<alyssa> tomeu: Hmm, this is true, yes..
yann has joined #panfrost
<bbrezillon> alyssa: did you have that wallpaper logic working?
<bbrezillon> looks like the code was tested in a specific situation according to one of this comment https://gitlab.freedesktop.org/mesa/mesa/blob/master/src/gallium/drivers/panfrost/pan_wallpaper.c#L207
<bbrezillon> so I guess it worked at some point
raster has quit [Remote host closed the connection]
raster has joined #panfrost
pH5 has quit [Quit: bye]
pH5 has joined #panfrost
<tomeu> alyssa: got these differences in rk3228 vs rk3399: https://gitlab.freedesktop.org/tomeu/mesa/-/jobs/280251
<tomeu> is that expected?
<urjaman> 3228? *blinks*
<tomeu> sorry, I meant rk3288
<alyssa> bbrezillon: I don't think so, I got half through and grumbled my way to oblivion, hence the giant #if 0
<bbrezillon> alyssa: oh, BTW, job->clear is always zero in my case
<alyssa> bbrezillon: Good to know. Unrelated issue, then!
<bbrezillon> so, even if there are missing job->clear clears, I'm not falling in that case
<bbrezillon> oh, and I had cases ->display_target != ->pipe_framebuffer, so I'm not sure we're reloading the right texture
<bbrezillon> but using the one pointed by ->pipe_framebuffer does not solve my problem :-/
<alyssa> tomeu: The log from that job looks right?
<alyssa> Actually, it looks very similar to the RK3399 log, so maybe I'm reading the wrong line
<alyssa> tomeu: The diff produced is pretty small and looks like broken flip-floppish tests anyway
<alyssa> Admittedly I have no idea what's up with 'dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_bvec4_qs_fragment'
yann has quit [Remote host closed the connection]
raster has quit [Remote host closed the connection]
pH5 has quit [Quit: bye]
stikonas has joined #panfrost
<ezequielg> not sure if anybody is around.
<ezequielg> i am running linux-next, with the panfrost driver, and i have latest mesa installed.
<ezequielg> on a rk3399 rockpi.
<ezequielg> i was testing wayland, and noticed that the hdmi goes black and only changed areas are painted.
<ezequielg> the rest goes black.
<ezequielg> as if it's not rendering areas that hasn't changed.
<ezequielg> is this a known issue?
<TheCycoONE> yes
<ezequielg> :-(
<ezequielg> so close
<bbrezillon> ezequielg: that's what I'm trying to fix :)
<bbrezillon> ezequielg: easy way to have that fixed => https://gitlab.freedesktop.org/mesa/mesa/merge_requests/658
<bbrezillon> you'd need to patch panforst 0 for the BUFFER_AGE cap as done here https://gitlab.freedesktop.org/yuq825/mesa/commit/fb48f39be29d8dda1936e3253966ad6a13b2cb8e
<bbrezillon> *to return 0
<ezequielg> ah, I recall daniels mentioned this.
<ezequielg> I'm doing --use-pixman for now @bbrezillon as a quick test
cwabbott has quit [Quit: cwabbott]
cwabbott has joined #panfrost
<anarsoul> bbrezillon: I think MR with this change was rejected
<anarsoul> consensus was that we still want to export buffer age extension and let application decide whether it's beneficial or not
cwabbott has quit [Client Quit]
cwabbott has joined #panfrost
TheKit has quit [Remote host closed the connection]
<TheCycoONE> what about a driconf option ala vmware?
<HdkR> driconf should really just be used as a last resort, implementing what an app wants or fixing driver bugs is the better outcome :D
<anarsoul> HdkR: app doesn't always want the best thing for driver
<HdkR> but the app is always right :D
<HdkR> Let's fix subroutines </s>
afaerber has quit [Quit: Leaving]
TheKit has joined #panfrost
<alyssa> HdkR is correct.
<HdkR> alyssa: Are you volunteering for subroutines? :P
<alyssa> Please no
<alyssa> :P
<HdkR> :D
mearon has quit [*.net *.split]
<HdkR> Today and Tomorrow is panfrost day
<HdkR> ...After I eat a late lunch
<alyssa> HdkR: Are we dubbing May 3 panfrost today?
<alyssa> I thought robclark already made July 4 "GPU driver independence [from the blob] day"
<HdkR> I only watch subs, no dubs
Lyude has quit [*.net *.split]
robertfoss has quit [*.net *.split]
<robclark> every day is "GPU driver independence day" \o/
<alyssa> robclark: Hooray!
<alyssa> HdkR: Luchame.
robher has quit []
<HdkR> luchame?
robher has joined #panfrost
Elpaulo has quit [Ping timeout: 252 seconds]
robertfoss has joined #panfrost
Lyude has joined #panfrost
mearon has joined #panfrost
tgall_foo has quit [Read error: Connection reset by peer]
NeuroScr has quit [Quit: NeuroScr]
stikonas has quit [Ping timeout: 250 seconds]
<HdkR> Alright. I want to get a newer kernel on one of my ODROID-N2 devices
<HdkR> narmstrong: Which kernel should I use if I want to build a new kernel for my ODROID-N2 with panfrost DRM?