ChanServ changed the topic of #lima to: Development channel for open source lima driver for ARM Mali4** GPUs - Kernel has landed in mainline, userspace driver is part of mesa - Logs at https://people.freedesktop.org/~cbrill/dri-log/index.php?channel=lima and https://freenode.irclog.whitequark.org/lima - Contact ARM for binary driver support!
warpme_ has quit [Quit: Connection closed for inactivity]
<anarsoul> enunes: rellla: any objections on https://gitlab.freedesktop.org/mesa/mesa/merge_requests/3335? it's a one-liner that should fix chromium-based browsers
<anarsoul> rellla: ref_value should be set if cso->alpha.enabled is set
<anarsoul> see ./freedreno/a2xx/fd2_zsa.c:fd2_zsa_state_create() for reference
<anarsoul> rellla: oh, I misinterpreted number from https://pastebin.com/raw/2qcH6c8E
<anarsoul> so you have 0xf4 in upper 16 bits
megi has quit [Ping timeout: 260 seconds]
yuq825 has joined #lima
dddddd has quit [Remote host closed the connection]
Barada has joined #lima
yuq825 has quit [Remote host closed the connection]
yuq825 has joined #lima
mripard has joined #lima
megi has joined #lima
warpme_ has joined #lima
dddddd has joined #lima
yann has quit [Ping timeout: 240 seconds]
yann has joined #lima
megi has quit [Quit: WeeChat 2.7]
Barada has quit [Quit: Barada]
<rellla> anarsoul: exactly. i have 0xf412 in the upper 16bits. the values vary in the different clears and do only appear, if STENCIL_TEST is enabled.
<rellla> and the only value i could see, that also varies, is glClearDepth ... :/ though i can't find any algorithm to convert the hex to a corresponding float.
<rellla> btw, i don't think it's alpha->ref_value, because i could not see anything in deqp code, that sets or even changes it...
<rellla> i'm dumping mali deqp again and will grep "STENCIL_TEST" to see, if any other test touches it
megi has joined #lima
yuq825 has quit [Quit: Leaving.]
<rellla> ... i think these upper bits aren't related at all ...
BenG83 has joined #lima
<anarsoul> rellla: it could be left over of alpha-ref value if alpha test it enabled
<anarsoul> rellla: there should be alpha function somewhere as well
<anarsoul> I have strong suspicion that it's in 0x20 [8] multisample
<anarsoul> lower 3 bits that have comment "always set? could be another CompareFunc"
<anarsoul> see "CompareFunc" below on this page, 111 GL_ALWAYS
<anarsoul> so alpha test always passes if it's 7 (all 3 bits are set), regardles of what ref value is :)
jernej has quit [Ping timeout: 260 seconds]
warpme_ has quit [Quit: Connection closed for inactivity]
jernej has joined #lima
chewitt has quit [Quit: Zzz..]
BenG83 has quit [Ping timeout: 268 seconds]
<anarsoul> rellla: so looks like bit 0x2000 in PLBU_CMD_PRIMITIVE_SETUP() has to do something with clipping
<anarsoul> blob has it set even for points, so it's not primitive type
<anarsoul> also when I set it it fixes some of the tests (missing points are now here) but breaks other tests (since it doesn't clip some of the points)
<anarsoul> but then /* 0x100947a0 (0x00000020) */0x0000f007/* MULTI_SAMPLE: points */