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/
hexdump0815 has joined #lima
hexdump01 has quit [Ping timeout: 480 seconds]
dsimic is now known as Guest17030
dsimic has joined #lima
Guest17030 has quit [Ping timeout: 480 seconds]
<uis>
sarbes: If a fragment program samples from a texture target on a texture
<uis>
defined in section 3.8.9, the result will be the vector
<uis>
image unit where the bound texture object is not complete, as
<uis>
(R, G, B, A) = (0, 0, 0, 1).
<anarsoul>
there is no such thing as "texture object is not complete" in hardware. If anything is wrong in texture descriptor GPU will crash in one way or another
<sarbes>
There might be a texture type for incomplete textures. There is an ARM errata talking about a "NO_TEXTURE" type, which I suspect is 0x3f.
<sarbes>
uis @_oftc_uis:matrix.org: if you are refering to the depth mode of my recent MR, I'd have to test it again if I really wanted to know what the value "0" means. But as it seems that Mesa bakes the mode into the shader, I don't see too much value in that.
<sarbes>
This mode might just be invalid, or it might require some other fields to be set up differently.
<sarbes>
This being said, the bit "Unknown 5" (6:15) has some effect on shadows, but it might be layout related.