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/
<anarsoul> OK, so liveness of output regs is weird :)
camus has joined #lima
camus1 has quit [Ping timeout: 480 seconds]
<anarsoul> almost got it working, just need to fix loops :)
camus1 has joined #lima
camus has quit [Ping timeout: 480 seconds]
<anarsoul> fragdepth_gles2 passes
<anarsoul> yay!
camus has joined #lima
camus1 has quit [Read error: Connection reset by peer]
camus1 has joined #lima
camus has quit [Ping timeout: 480 seconds]
camus has joined #lima
camus1 has quit [Read error: Connection reset by peer]
<anarsoul> I pushed it to gitlab, see https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/13830 if you're curious
<anarsoul> there's no dEQP regressions, I'll run piglit to see if there's any changes there
camus1 has joined #lima
camus has quit [Ping timeout: 480 seconds]
camus has joined #lima
camus1 has quit [Read error: Connection reset by peer]
camus1 has joined #lima
camus has quit [Ping timeout: 480 seconds]
<anarsoul> well, 12 piglit fixes, 1 regression
<anarsoul> it regresses in spec@ext_framebuffer_object@fbo-depth-sample-compare which relies on working gl_FragDepth, so I really have no idea how it used to pass :)
<anarsoul> fwiw it also fails on i915 and vc4
camus1 has quit [Remote host closed the connection]
camus has joined #lima
gcl_ has joined #lima
camus1 has joined #lima
gcl has quit [Ping timeout: 480 seconds]
camus has quit [Ping timeout: 480 seconds]
camus has joined #lima
camus1 has quit [Ping timeout: 480 seconds]
<rellla> anarsoul: thanks, i will have a look into your code
<rellla> btw, is the description for depth_test & 0x00001c00 right? at least 0x400 seems misleading to me.
gcl_ has quit [Ping timeout: 480 seconds]
camus1 has joined #lima
camus has quit [Ping timeout: 480 seconds]
camus has joined #lima
camus1 has quit [Ping timeout: 480 seconds]
camus1 has joined #lima
camus has quit [Ping timeout: 480 seconds]
enunes has quit [Quit: ZNC - https://znc.in]
<anarsoul> rellla: I'll double check later
<rellla> anarsoul: it's not related to your MR, it's just for my understanding ;)
<anarsoul> rellla: so 0x800 is set when shader writes depth
<anarsoul> 0x400 is set in depth reload when shader writes depth
<anarsoul> and 0x400 is set only for Z24S8 format when we use this weird format 0x3-something for sampling
<anarsoul> 0x1000 is set for writing stencil
<anarsoul> so I'm not really sure on 0x400 either
drod has joined #lima
worldlier has joined #lima
worldlier has quit [autokilled: #nottor is permanently shut down, do not join it. - Contact support@oftc.net for help. (2021-11-17 17:00:33)]
<rellla> yeah, 0x800 and 0x1000 is clear, but 0x400 is a bit weird ;)
camus has joined #lima
camus1 has quit [Read error: Connection reset by peer]
<anarsoul> rellla: it looks like it has to be used with this weird reload format
<anarsoul> so I'd call it something like "write z24"?
<anarsoul> rellla: so are you working on genxml?
<rellla> anarsoul: no.
<rellla> not yet maybe, i just had a look into panfrost and others
<anarsoul> btw, I think that we can't implement gl_FragStencilRefARB with lima
<anarsoul> gl_FragStencilRefARB is int while it looks like all the outputs in Mali4x0 are floats
<anarsoul> so I'm not sure if we even need to normalize it
<anarsoul> so no GL_ARB_shader_stencil_export for lima :)
camus1 has joined #lima
camus has quit [Read error: Connection reset by peer]
camus has joined #lima
camus1 has quit [Ping timeout: 480 seconds]
drod has quit [Remote host closed the connection]