ChanServ changed the topic of #lima to: Development channel for open source lima driver for ARM Mali4** GPUs - Kernel driver has landed in mainline, userspace driver is part of mesa - Logs at https://oftc.irclog.whitequark.org/lima/
adjtm has quit [Remote host closed the connection]
adjtm has joined #lima
adjtm has quit [Remote host closed the connection]
adjtm has joined #lima
<anarsoul> rellla: enunes: I've got 3D textures somewhat working :)
<anarsoul> it passes basic piglit tests, tex3d-depth1 fails because we don't have clamp_to_border support
<anarsoul> it turns out utgard texture descriptor has something in common with midgard :)
<anarsoul> I also have strong suspicion that wrap modes are also here, we currently treat them as individual bits, but it's too wasteful
<anarsoul> it's likely enum
<anarsoul> similar to midgard
<anarsoul> just need some experiments to figure out what each value means
<rellla> anarsoul: hey, nice work, will look into it ;)
Danct12 has quit [Quit: Quitting]
Danct12 has joined #lima
Danct12 has quit []
adjtm is now known as Guest1995
adjtm has joined #lima
Guest1995 has quit [Read error: No route to host]
Danct12 has joined #lima
drod has joined #lima
<anarsoul> rellla: the same MR now has more wrap modes :)
<anarsoul> also I found border color
<anarsoul> more piglit fixes :P
<anarsoul> tex3d-depth1 now passes
<anarsoul> oh, and new failures
<rellla> anarsoul: you are too fast for me to review :)
<anarsoul> these are trivial changes
<anarsoul> is there a way to get reference image and image diff in piglit?
<rellla> yes, the last one. a few hours ago i thought, if i should tell you, where the border color is. though i suspected it 3 bits later ;)
<rellla> not sure about piglit, i guess #dri-devel knows if possible
<anarsoul> rellla: I tried it first, it didn't work
<anarsoul> apparently ARM loves tight bit packing
<rellla> regarding shaders unknown_1 i guess you are right. at least some of the 5 bits seem to be a value, likely absolute lod.
Danct12 has quit [Quit: Quitting]
<anarsoul> well, WRAP_CLAMP doesn't seem to work
<anarsoul> rellla: it could be absolute lod bias
<anarsoul> well, constant lod bias
<anarsoul> and you need all 5 bits if you want it to be negative
<anarsoul> OK, fixed GL_CLAMP failures with alyssa help :)
<anarsoul> and added 1D textures support
<anarsoul> once it merges we officially we'll support more features than blob :)
<adjtm> anarsoul, which opengl standard is this texture work for?
<anarsoul> adjtm: desktop gl? 2.1 since it's what we expose
<anarsoul> I guess mesa should also expose ES2 extension for 3D textures
<adjtm> so it could affect es2 in android, right?
<adjtm> by the way, thanks for your work
<anarsoul> adjtm: I may affect es2 on android if mesa exposes OES_texture_3D and GL_EXT_texture_border_clamp
<anarsoul> adjtm: np, are you using lima in replicant? :)
<anarsoul> (sorry, I don't remember from which project you are coming)
<adjtm> yes, replicant11, but it's in very early stage of development
<adjtm> all devices that are going to have initial support have mali 400
<adjtm> basically galaxy s3, note2 and pinephone
<adjtm> lima is far from being the most lacking part of replicant 11 :)
<anarsoul> I skimmed through the code and looks like mesa doesn't expose GL_EXT_texture_border_clamp
drod has quit [Remote host closed the connection]
drod has joined #lima
<anarsoul> rellla: can you share your piglit cmdline?
drod has quit [Remote host closed the connection]
adjtm has quit [Remote host closed the connection]
adjtm has joined #lima
adjtm has quit [Remote host closed the connection]
adjtm has joined #lima