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!
megi has quit [Ping timeout: 240 seconds]
yuq825 has joined #lima
chewitt has joined #lima
dddddd has quit [Ping timeout: 260 seconds]
chewitt has quit [Quit: Zzz..]
chewitt has joined #lima
chewitt has quit [Quit: Zzz..]
Net147 has quit [Quit: Quit]
Net147 has joined #lima
chewitt has joined #lima
megi has joined #lima
BenG83 has joined #lima
yuq825 has quit [Ping timeout: 260 seconds]
_whitelogger has joined #lima
dddddd has joined #lima
_whitelogger has joined #lima
abordado has joined #lima
yuq825 has joined #lima
megi has quit [Ping timeout: 240 seconds]
buzzmarshall has joined #lima
megi has joined #lima
yuq825 has quit [Quit: Leaving.]
abordado has quit [Read error: Connection reset by peer]
abordado has joined #lima
chewitt has quit [Quit: Zzz..]
buzzmarshall has quit [Quit: Leaving]
buzzmarshall has joined #lima
drod has joined #lima
buzzmarshall has quit [Remote host closed the connection]
egbert has joined #lima
drod has quit [Quit: Ухожу я от вас (xchat 2.4.5 или старше)]
<anarsoul> rellla: any progress with functional.buffer.write.use.index_array.array issue?
BenG83 has joined #lima
<rellla> anarsoul: no. i guess tomorrow i can do some work again. i think we should check dEQP-GLES2.functional.draw.draw_arrays.* and dEQP-GLES2.functional.draw.draw_elements.* test first
<rellla> they should be smaller and easier to "reproduce"
<rellla> in the meantime i fixed mali_syscall_tracker and wrote a little bash script...
<rellla> which is running all deqp tests one after another, producing dumps and separate them in different directories.
<rellla> once finished, i will upload them. they probably can be used for stream analysis and as a reference :)
<anarsoul> yeah, definitely
<anarsoul> I'm looking into deqp code for this test
<anarsoul> I wonder what if we specify index size incorrectly
<anarsoul> I mean to the hardware
<rellla> yeah. i guess, the complete dumps will reveal that
<anarsoul> we need to parse index size as well
<rellla> i will wait until the run completes and then will add some more parses. what should we prettify, too?
<rellla> attributes, varyings?
<anarsoul> yeah, parsing attributes and varyings descriptors would be nice
<rellla> where do you want to catch index_size?
<anarsoul> looks like it's already parsed, so nevermind
<anarsoul> I wonder why start is 1 in first DRAW_ELEMENTS in lima, but 0 in blob
<rellla> if /* 0x10092450 (0x00000050) */0x00003200 0x1000010b/* PRIMITIVE_SETUP: prim: LINES, cull: 0 (0x0), index_size: 1 */ is correct, then yes
<rellla> i have a strong guess, that solving that issue will solve dEQP-GLES2.functional.buffer.write.use.*, dEQP-GLES2.functional.draw.* and dEQP-GLES2.functional.vertex_arrays.* ...
<anarsoul> maybe
<anarsoul> it's likely another one-liner
<anarsoul> "start: 1" doesn't make sense to me
<anarsoul> :\
<rellla> what dumps do you refer to?
<anarsoul> and http://imkreisrum.de/deqp/index_array.array_mali/mali.dump.0002 where start is 0 in DRAW_ELEMENTS
<anarsoul> oh
<anarsoul> btw, note SHADER_INFO: prefetch: disabled in mali
abordado has quit [Read error: Connection reset by peer]
<rellla> but bit 21 is set instead
<anarsoul> "prefetch" seems to be a number, not flag
<anarsoul> see ./ir/gp/codegen.c
<anarsoul> it's last instruction that touches attribute
<anarsoul> *attribute(s)
<rellla> ok, but then just the parser is wrong?!
<anarsoul> yeah
<anarsoul> I wonder why blob splits vertex shader invocations
<anarsoul> note that it invokes vertex shader 24 + 183 + 27 = 234 times
<anarsoul> lima invokes it 254 times
<rellla> thats what i meant two days ago
<anarsoul> that's interesting :)
<anarsoul> in theory it's totally legal to have several commands to invoke vertex shader as long as it's synchronized with plbu
<anarsoul> and it is in this case
<anarsoul> well, it looks like all the lines are here, but color is not valid
<anarsoul> it means that vs ran successfully and at least wrote gl_Position correctly
BenG83 has quit [Quit: Leaving]
<anarsoul> I'm not so sure about varyings