00:06
NeuroScr has quit [Quit: NeuroScr]
00:53
NeuroScr has joined #panfrost
01:14
chewitt has joined #panfrost
01:23
NeuroScr has quit [Quit: NeuroScr]
01:40
stikonas has quit [Read error: Connection reset by peer]
02:11
<
alyssa >
I mean, literally, like,
02:12
<
alyssa >
varying vec4 whatever;
02:12
<
alyssa >
whatever = vec4(1.0, 0.0, 0.0, 1.0);
02:12
<
alyssa >
Super dumb but it's legal :p
02:13
<
alyssa >
Not, like, a special "constant varying" or anything
02:13
<
alyssa >
Just.... a constant, written to a varying
02:13
<
alyssa >
Which is a special case for us
02:16
<
alyssa >
Psst constant varying is an oxymoron :p
02:20
<
HdkR >
alyssa: a constant varying in newer GLSL terms effectively means the "flat" qualifier. So it appears as if it is constant from the fragment side
02:20
<
alyssa >
Not what I meant in the MR but ok
02:20
<
HdkR >
Yea, bit different
02:21
<
alyssa >
Regardless uh
02:21
<
alyssa >
The MR still fixes an actual bug :p
04:30
<
alyssa >
^ Got the other one for ya
04:32
<
alyssa >
Oh hey, we don't fail all of dEQP anymore
04:37
<
alyssa >
...only most of it
05:13
chewitt has quit [Max SendQ exceeded]
05:15
chewitt has joined #panfrost
06:26
anarsoul has quit [Remote host closed the connection]
06:27
anarsoul has joined #panfrost
06:30
anarsoul_ has joined #panfrost
06:30
anarsoul has quit [Client Quit]
06:31
<
alyssa >
What do you know? I tried a bunch of random stuff and now all of the Panfrost bugs went away!
06:32
<
alyssa >
So either I actually fixed something among the random changes, or I somehow installed llvmpipe :P
06:33
<
anarsoul_ >
can you run q3arena now?
06:33
<
alyssa >
By "all", well, it was all the same bug, but
06:33
<
alyssa >
- Missing triangles in -bshadow
06:33
<
alyssa >
- Missing parts in various renders
06:34
<
alyssa >
- Black rectangles in Weston
06:34
<
alyssa >
- Flashing in Kodi
06:34
<
alyssa >
- Missing draws in single-frame apps
06:34
<
alyssa >
- Failing all of dEQP
06:34
<
alyssa >
Speaking of, hey, let's retry dEQP
06:35
<
HdkR >
But does the Quake 2 ray tracing game work now? :P
06:38
anarsoul_ has quit [Read error: Connection reset by peer]
06:59
<
alyssa >
^ I'm satisfied
07:00
<
alyssa >
---Why does [REDACTED] get stuck on the loading screen now?
07:00
<
chewitt >
oooh.. shiny
07:01
<
alyssa >
chewitt: Screen saver still broken, and video playback is wacked but you said that doesn't matter on Amlogic... besides that we seem to be working well :)
07:01
* alyssa
tries to figure out when [REDACTED] regressed
07:02
<
chewitt >
lrusak asked me to test a patch earlier that results in text but no background GUI
07:03
<
alyssa >
chewitt: Blugh, I forgot Kodi was a moving target :p
07:03
<
chewitt >
slowly moving target for most GBM things
07:03
<
chewitt >
that's the first change in a few months
07:04
<
HdkR >
alyssa: I'm 100% fine with this
07:05
<
alyssa >
Update; [REDACTED] didn't regress, I'm just bad at testing :p
07:09
<
chewitt >
@alyssa what board are you testing Kodi on now? .. playback should be okay on most things
07:09
<
alyssa >
chewitt: Board?
07:09
<
alyssa >
It's just my Panfrost dev laptop (RK3399)
07:09
<
alyssa >
No VPU acceleration setup, since mainline kernel
07:10
<
alyssa >
[Okay, apparently there's work to get mainline VPU accel. But I don't have that yet :P]
07:10
<
chewitt >
ahh, ok, yeah, that's still wip
07:11
<
chewitt >
rebuilt/updated and I still have the black screen with white text (due to the lrusak patch) but flickering is definitely gone :)
07:11
<
chewitt >
I'll drop the kodi patch and have a play..
07:13
<
alyssa >
Is that patch pushed to kodi master (so if I pull/rebuild things will replicate nicely?)
07:17
<
chewitt >
no that's not PR'd/merged yet
07:23
<
alyssa >
We're passing actual dEQP tests!!
07:23
<
alyssa >
We pass 24/24 of dEQP-GLES2.functional.draw.draw_arrays.*
07:24
<
alyssa >
Cherry picked I will admit
07:24
<
alyssa >
But still! We were at 0/24 yesterday!
07:27
<
HdkR >
Time to upstream
07:30
<
alyssa >
HdkR: Pending on kernel, sure :P
07:31
<
chewitt >
with latest changes (and without lrusak's patch) Kodi GUI is rather nice
07:31
<
chewitt >
it still memleaks badly.. but all the glitches in the GUI are gone
07:33
<
alyssa >
I can work on memleaks next burst of energy I get, bugsquashing is only fun for so long ;)
07:34
<
alyssa >
Oh, dear, it appears I've substantially regressed [REDACTED]
07:34
<
alyssa >
HdkR: Don't merge the thing I just opened
07:36
<
alyssa >
The greatest thing about the phrase "[REDACTED]" is that it can expand to two totally separate things
07:39
<
alyssa >
HdkR: OK now it's good
07:48
anarsoul has joined #panfrost
08:37
chewitt has quit [Quit: Adios!]
08:59
NeuroScr has joined #panfrost
09:29
NeuroScr has quit [Quit: NeuroScr]
10:00
sphalerite has quit [Quit: rebooting!]
10:05
sphalerite has joined #panfrost
10:38
NeuroScr has joined #panfrost
10:59
stikonas has joined #panfrost
11:07
NeuroScr has quit [Quit: NeuroScr]
11:53
_whitelogger has joined #panfrost
11:58
stikonas has quit [Remote host closed the connection]
12:47
_whitelogger has joined #panfrost
15:17
Elpaulo has quit [Quit: Elpaulo]
15:39
pH5 has quit [Ping timeout: 245 seconds]
15:42
rhyskidd has quit [Quit: rhyskidd]
15:42
rhyskidd has joined #panfrost
16:07
rhyskidd has quit [Quit: rhyskidd]
16:08
rhyskidd has joined #panfrost
16:08
BenG83 has quit [Read error: Connection reset by peer]
16:10
Elpaulo has joined #panfrost
16:27
stikonas has joined #panfrost
16:41
stikonas has quit [Remote host closed the connection]
16:48
<
alyssa >
Oh hey, [REDACTED} gets a teensy bit further now
16:56
<
sphalerite >
supertuxkart? :D
16:56
<
alyssa >
sphalerite: Other [REDACTED]
16:57
<
sphalerite >
and why is it redacted?
17:05
<
Lyude >
Can we just say the name of the thing?
17:05
<
Lyude >
it feels really weird and unusual we can't talk about hardware being worked on in here
17:25
<
alyssa >
Well, I guess I can let the cat out since there's no bag :(
17:25
<
alyssa >
[REDACTED} alternately refers to SuperTuxKart or GNOME Shell. Both are severely broken so there's that :p
17:29
<
sphalerite >
OpenClonk when?
17:29
<
alyssa >
OpenClonk?
17:39
<
alyssa >
If I cherrypick aggressively, I can find apps that work right with 'frost ;P
17:46
<
sphalerite >
openclonk.org
17:47
<
sphalerite >
hedgewars!
17:47
<
alyssa >
I was actually going to say Neverball, but sure :p
17:49
<
alyssa >
Actually, re Neverball... it has one critical bug, and it's.... exactly the same as STK's
17:54
<
alyssa >
(Background/grass/etc is TOTALLY corrupted and has random massive tris in random places and it's whaaa?)
17:57
<
alyssa >
Guessing something isn't happy about these big tris
18:00
<
alyssa >
I wonder if it's depth near/far
18:02
<
alyssa >
Eh nope, it's still being set to (0, 1) like we hardcode as is. Can't be that
18:12
rhyskidd has quit [Quit: rhyskidd]
18:13
rhyskidd has joined #panfrost
18:42
<
rtp >
alyssa: oh, I'm surprised you didnt notice the fnctl(F_DUPFD) thing. it was quite obvious to me when looking at the strace output :P
18:44
BenG83 has joined #panfrost
19:33
<
alyssa >
rtp: I can't be amazing at
_everything_ ;)
19:33
<
alyssa >
Collaboration! Teamwork! Friendship!
19:43
<
anarsoul >
long time no see :)
19:44
<
alyssa >
You guys know each other?
19:45
<
anarsoul >
well, it's a small world
19:46
<
alyssa >
Fair enough
19:59
stikonas has joined #panfrost
20:15
<
alyssa >
Looking into gl_PointSize
20:17
<
alyssa >
Okay, reading panwrap, when gl_PointSize is written in the vertex shader, we have:
20:17
<
alyssa >
- varying_count incresed by two
20:17
<
alyssa >
- uniform_count increased by one but I think that's spurious
20:18
<
alyssa >
- Some new stuff in the shader I'll read in a moment
20:18
<
alyssa >
- Two new varyings as expected
20:18
<
alyssa >
- New uniform, again as expected
20:18
<
alyssa >
- line_width field is set to -0.00f (negative zero?!)
20:19
<
alyssa >
- unknown_draw ORed with 0x100 (so, 0x3100 instead of 0x3000)
20:19
<
alyssa >
Let's look at the shader
20:20
<
alyssa >
So, we have a st_vary_32 to address 1
20:21
<
alyssa >
Er, no, I think that's gl_Position actually
20:24
<
alyssa >
min(max(a, b), c)
20:24
<
alyssa >
That's, clamp, uh
20:26
<
alyssa >
Point is, it's writing the transformed point size to hr53.z and then st_vary_16'ing it
20:26
<
alyssa >
Let's look at the varying descriptor next, I suppose
20:27
<
alyssa >
Of course I don't have the new formats in this panwrap version, annoying
20:29
<
alyssa >
First varying is MALI_R16F, which is what you would expect
20:30
<
alyssa >
xyyy swizzle, for some reason (if I decoded that right -- I have to do this stuff manually since old panwrap :p)
20:31
<
alyssa >
index set to 2, which is bizarre since there's only varying buffers decoded but we might have an off-by-one
20:33
<
alyssa >
Might be easier to just start trying things in a branch, heh "P
21:32
<
alyssa >
HdkR: Actually, don't merge the MR I sent, I'm going to try to do proper PointSize support to get it over with :p
21:33
<
ezequielg >
alyssa: we were right
21:33
<
ezequielg >
nv12 is just a special format
21:33
<
ezequielg >
it's 0xc
21:34
<
ezequielg >
:triumph_emoji:
21:34
<
ezequielg >
you know what's funny
21:34
<
ezequielg >
i was about to give u
21:34
<
ezequielg >
then i said, what if it's the next one.
21:34
<
ezequielg >
etc etc
21:34
<
ezequielg >
rocky song climing the stairs, etc
21:34
<
alyssa >
Gambler's fallacy pays off?!
21:36
<
ezequielg >
lol, i didn't know there was a hole fallacy
21:37
<
alyssa >
A whole fallacy, yes :)
21:37
<
ezequielg >
right hole, whole
21:38
<
ezequielg >
if it sounds right, it's right - and that's how i learned to speak anglais
21:38
<
ezequielg >
ok, merge request
21:38
<
alyssa >
Anglais? I thought the trick was if it sounds right, it's wrong :P
21:39
<
alyssa >
Source: many years of speaking English
21:40
<
alyssa >
So kmscube -Bnv12 works now?
21:42
<
alyssa >
Awesome :)
21:46
<
alyssa >
The harder problem is getting Kodi/mpv/etc to use native NV12... =P
21:53
<
ezequielg >
...a problem for another day
21:53
<
ezequielg >
\o ttys
21:53
<
alyssa >
ezequielg: (Did you send the MR?)
21:53
<
ezequielg >
i'm on it
21:56
<
ezequielg >
alyssa: anything special?
21:56
<
ezequielg >
do you want to squash?
21:56
<
ezequielg >
pff, it's just one commit.
21:56
<
ezequielg >
got confused with rebase vs. merge
21:57
<
ezequielg >
done, it's #29
21:58
<
alyssa >
Thank you!
22:17
<
alyssa >
Oh, this is interesting
22:18
<
alyssa >
line_width is interpreted as point_size, if point_size is constant (i.e. not written in the shader)
22:18
<
alyssa >
(I'm not sure if that's a case you can get to from GLES, but that's a different issue, haha)
22:22
<
alyssa >
Unpopular opinion: GL_POINTS should've been tossed along with QUADS
22:24
<
alyssa >
What the huh?
22:27
<
alyssa >
...Okay, so it's shoving a pointer into line_width
22:28
stikonas has quit [Remote host closed the connection]
22:28
<
alyssa >
Midgard never ceases to amaze me
22:29
stikonas has joined #panfrost
22:36
NeuroScr has joined #panfrost
22:41
<
alyssa >
Ooookay, that's kind of terrifying(ly awesome) but I digress
22:41
<
alyssa >
The 0x100 in the unknown_draw means "interpret line_width as a pointer"
22:41
<
alyssa >
And I'm assuming that's how we link the varying in
22:48
<
stikonas >
KDE Plasma bootsplash actually renders nicely now :). Although plasmashell itself is still not working properly
22:50
<
alyssa >
stikonas: Progress is progress! :)
22:51
<
stikonas >
yeah, panfrost team is doing a great job
22:51
<
alyssa >
Thank you :)
22:54
<
Lyude >
robclark: didn't you do a bunch of stuff with 96boards?
22:54
<
Lyude >
robclark: was wondering if you remember what the easiest way to get Fedora onto one would be
23:49
<
alyssa >
...I did not realise gl_PointSize would become an all-day project
23:49
<
alyssa >
(Still not there yet!)
23:51
<
alyssa >
Whatever, I'm having fun and that's what matters :)
23:56
<
robclark >
Lyude, at least a couple 96boards are supported out of the box in fedora.. the ones that have upstream u-boot and kernel support..
23:56
<
robclark >
for the others, dd fedora img to sdcard or usb disk, and make your own kernel arrangements..
23:57
<
alyssa >
robclark: The Embedded Linux Experience (TM)
23:57
<
robclark >
more or less, yup