2021-07-26 22:56
ChanServ changed the topic of #dri-devel to: <ajax> nothing involved with X should ever be unable to find a bar
00:02
dllud_ has quit [Server closed connection]
00:02
dllud has joined #dri-devel
00:03
pcercuei has quit [Quit: dodo]
00:08
Wally has quit [Remote host closed the connection]
00:22
nchery has quit [Remote host closed the connection]
00:22
nchery has joined #dri-devel
00:24
loki_val has joined #dri-devel
00:27
crabbedhaloablut has quit [Ping timeout: 480 seconds]
00:27
thellstrom has quit [Remote host closed the connection]
00:28
thellstrom has joined #dri-devel
00:34
Wally has joined #dri-devel
00:35
mhenning has quit []
00:37
gruetzkopf has quit [Server closed connection]
00:37
gruetzkopf has joined #dri-devel
00:42
Wally has quit [Remote host closed the connection]
00:48
rasterman has joined #dri-devel
01:02
rkanwal has quit []
01:02
rkanwal has joined #dri-devel
01:15
tursulin has quit [Read error: Connection reset by peer]
01:16
rkanwal has quit []
01:25
rasterman has quit [Quit: Gettin' stinky!]
01:47
jewins1 has joined #dri-devel
01:51
columbarius has joined #dri-devel
01:52
jewins has quit [Ping timeout: 480 seconds]
01:53
co1umbarius has quit [Ping timeout: 480 seconds]
02:06
ngcortes has quit [Remote host closed the connection]
02:12
shashanks has joined #dri-devel
02:20
Emantor has joined #dri-devel
02:21
shashanks has quit [Ping timeout: 480 seconds]
02:30
tarceri_ has quit [Remote host closed the connection]
02:31
soreau has quit [Read error: Connection reset by peer]
02:31
soreau has joined #dri-devel
02:53
pnowack has quit [Quit: pnowack]
03:04
tarceri has joined #dri-devel
03:20
jewins1 has quit [Ping timeout: 480 seconds]
03:44
shashanks has joined #dri-devel
03:52
shashanks has quit [Ping timeout: 480 seconds]
03:52
devilhorns has joined #dri-devel
03:52
devilhorns has quit [Remote host closed the connection]
03:59
cheako has joined #dri-devel
04:03
shashanks has joined #dri-devel
04:05
Guest1230 is now known as DrNick
04:11
shashanks has quit [Ping timeout: 480 seconds]
04:12
tarceri has quit [Remote host closed the connection]
04:12
tarceri has joined #dri-devel
04:15
OftenTimeConsuming has quit [Remote host closed the connection]
04:15
OftenTimeConsuming has joined #dri-devel
04:27
tarceri has quit [Remote host closed the connection]
04:27
tarceri has joined #dri-devel
04:39
mclasen has joined #dri-devel
04:40
maxzor has joined #dri-devel
04:55
ppascher has joined #dri-devel
05:06
_alice is now known as arisu
05:20
Duke`` has joined #dri-devel
05:26
thellstrom has quit [Ping timeout: 480 seconds]
05:32
tarceri has quit [Ping timeout: 480 seconds]
05:34
JohnnyonFlame has quit [Ping timeout: 480 seconds]
05:44
mattrope has quit [Read error: Connection reset by peer]
05:55
Hi-Angel has joined #dri-devel
05:58
Wally has joined #dri-devel
06:01
tarceri has joined #dri-devel
06:06
<
Wally >
[re-asking] How does dri share the bus of (character) devices that it accesses with other kernel modules, and how can I access it in a kernel module?
06:07
sarnex has quit [Read error: Connection reset by peer]
06:07
sarnex has joined #dri-devel
06:07
<
Wally >
...to access lower level functions(read, write, seek) or it
06:14
Wally has quit [Remote host closed the connection]
06:15
camus has joined #dri-devel
06:16
Wally has joined #dri-devel
06:17
camus1 has quit [Remote host closed the connection]
06:18
Wally has quit [Remote host closed the connection]
06:20
Wally has joined #dri-devel
06:21
Wally_ has joined #dri-devel
06:22
Wally is now known as Guest1340
06:22
Wally_ is now known as Wally
06:28
Wally has quit [Remote host closed the connection]
06:28
Guest1340 has quit [Remote host closed the connection]
06:29
siqueira_ has joined #dri-devel
06:31
siqueira has quit [Ping timeout: 480 seconds]
06:33
pjakobsson has quit [Ping timeout: 480 seconds]
06:34
siqueira_ has quit []
06:36
siqueira has joined #dri-devel
06:48
danvet has joined #dri-devel
06:50
pjakobsson has joined #dri-devel
06:50
nchery has quit [Read error: Connection reset by peer]
06:56
mclasen has quit [Ping timeout: 480 seconds]
07:04
shashanks has joined #dri-devel
07:08
Wally has joined #dri-devel
07:30
alanc has quit [Remote host closed the connection]
07:31
alanc has joined #dri-devel
07:44
idr has quit [Ping timeout: 480 seconds]
07:47
kts has joined #dri-devel
07:50
lemonzest has joined #dri-devel
07:53
maxzor has quit [Remote host closed the connection]
07:58
<
airlied >
Wally: it doesn't it binds directly to the pci device and uses the mmio resources
08:05
Duke`` has quit [Ping timeout: 480 seconds]
08:06
neonking has joined #dri-devel
08:07
Duke`` has joined #dri-devel
08:14
YuGiOhJCJ has joined #dri-devel
08:35
Duke`` has quit [Ping timeout: 480 seconds]
08:43
Duke`` has joined #dri-devel
08:55
<
Wally >
airlied: so instead of accessing chardev, etc. it interacts with mmio?
08:56
Wally has quit [Remote host closed the connection]
09:08
<
cheako >
ne1 use zink? I'm trying to use my vulkan patch for No Man's Sky on GL applications like Oxygen Not Included.
09:09
<
HdkR >
vulkan patch?
09:09
tarceri has quit [Remote host closed the connection]
09:10
tarceri has joined #dri-devel
09:12
<
HdkR >
I guess I've never noticed weird frame timing problems but I also don't play NMS or use a composited desktop environment :)
09:24
Akari has quit [Quit: segmentation fault (core dumped)]
09:27
Akari has joined #dri-devel
09:31
gouchi has joined #dri-devel
09:46
thellstrom has joined #dri-devel
09:55
gouchi has quit [Remote host closed the connection]
09:56
gouchi has joined #dri-devel
10:05
Surkow|laptop has quit [Remote host closed the connection]
10:14
kts has quit [Quit: Konversation terminated!]
10:15
pcercuei has joined #dri-devel
10:36
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
11:02
OftenTimeConsuming has quit [Remote host closed the connection]
11:03
OftenTimeConsuming has joined #dri-devel
11:08
flacks has quit [Quit: Quitter]
11:12
flacks has joined #dri-devel
11:19
shashanks has quit [Remote host closed the connection]
11:31
rasterman has joined #dri-devel
11:37
Surkow|laptop has joined #dri-devel
11:44
Hi-Angel has quit [Ping timeout: 480 seconds]
12:19
shashanks has joined #dri-devel
12:22
dv_ has quit [Server closed connection]
12:23
dv_ has joined #dri-devel
12:24
Ziemas has quit [Server closed connection]
12:24
Ziemas has joined #dri-devel
12:26
ella-0_ has joined #dri-devel
12:28
ella-0 has quit [Remote host closed the connection]
12:28
shashanks has quit [Ping timeout: 480 seconds]
12:29
shashanks has joined #dri-devel
12:44
gouchi has quit [Remote host closed the connection]
12:57
FireBurn has joined #dri-devel
13:05
JohnnyonFlame has joined #dri-devel
13:10
JoshuaAshton has quit [Server closed connection]
13:10
JoshuaAshton has joined #dri-devel
13:12
tpalli has quit [Server closed connection]
13:12
tpalli has joined #dri-devel
13:14
rellla has quit [Server closed connection]
13:20
rellla has joined #dri-devel
13:20
pcercuei has quit [Ping timeout: 480 seconds]
13:20
calebccff has quit [Server closed connection]
13:20
calebccff has joined #dri-devel
13:27
pcercuei has joined #dri-devel
13:37
shashanks has quit [Remote host closed the connection]
13:38
camus1 has joined #dri-devel
13:40
FireBurnUK has joined #dri-devel
13:40
FireBurn has quit [Read error: Connection reset by peer]
13:41
opotin61 has quit [Server closed connection]
13:41
opotin61 has joined #dri-devel
13:43
camus has quit [Ping timeout: 480 seconds]
13:43
FireBurnUK has quit [Read error: Connection reset by peer]
13:43
FireBurnUK has joined #dri-devel
13:44
rkanwal has joined #dri-devel
13:48
FireBurnUK has quit [Quit: Konversation terminated!]
13:48
FireBurnUK has joined #dri-devel
13:57
marex has quit [Server closed connection]
13:57
marex has joined #dri-devel
14:01
xantoz has quit [Server closed connection]
14:01
xantoz has joined #dri-devel
14:07
zamundaaa has quit [Server closed connection]
14:17
zamundaaa has joined #dri-devel
14:36
mclasen has joined #dri-devel
14:37
shashanks has joined #dri-devel
14:41
tomeu has quit [Server closed connection]
14:41
tomeu has joined #dri-devel
14:46
shashanks has quit [Ping timeout: 480 seconds]
14:53
rkanwal has quit [Ping timeout: 480 seconds]
14:55
dos1 has quit [Server closed connection]
14:55
dos1 has joined #dri-devel
15:02
mstoeckl has quit [Server closed connection]
15:02
mstoeckl has joined #dri-devel
15:02
pinchartl has quit [Server closed connection]
15:03
pinchartl has joined #dri-devel
15:04
thellstrom has quit [Ping timeout: 480 seconds]
15:09
iokill has quit [Server closed connection]
15:09
iokill has joined #dri-devel
15:12
Haaninjo has joined #dri-devel
15:14
FireBurnUK is now known as FireBurn
15:23
Arsen has quit [Server closed connection]
15:23
Arsen has joined #dri-devel
15:24
Arsen is now known as Guest1373
15:26
pcercuei has quit [Quit: Lost terminal]
15:29
pcercuei has joined #dri-devel
15:32
shashanks has joined #dri-devel
15:33
rkanwal has joined #dri-devel
15:37
shashanks has quit [Remote host closed the connection]
15:53
rkanwal has quit [Ping timeout: 480 seconds]
16:01
libv has quit [Server closed connection]
16:01
libv has joined #dri-devel
16:05
shashanks has joined #dri-devel
16:06
camus has joined #dri-devel
16:07
yoslin has quit [Server closed connection]
16:07
yoslin has joined #dri-devel
16:13
shashanks has quit [Ping timeout: 480 seconds]
16:18
mclasen has quit [Ping timeout: 480 seconds]
16:32
camus has joined #dri-devel
16:51
APic has quit [Server closed connection]
16:51
APic has joined #dri-devel
16:52
gawin has joined #dri-devel
16:53
degasus has quit [Server closed connection]
16:53
degasus has joined #dri-devel
16:53
degasus is now known as Guest1380
16:57
<
gawin >
iirc Pavel made some changes for broken liveness when there's break in code
17:00
mclasen has joined #dri-devel
17:03
mhenning has joined #dri-devel
17:06
ced117 has quit [Server closed connection]
17:06
ced117 has joined #dri-devel
17:08
mclasen has quit [Ping timeout: 480 seconds]
17:09
vup has quit [Server closed connection]
17:09
vup has joined #dri-devel
17:10
dwlsalmeida has quit [Server closed connection]
17:11
dwlsalmeida has joined #dri-devel
17:11
pq has quit [Server closed connection]
17:11
pq has joined #dri-devel
17:12
qyliss has quit [Server closed connection]
17:12
qyliss has joined #dri-devel
17:14
CATS has quit [Server closed connection]
17:14
CATS has joined #dri-devel
17:16
Guest1380 is now known as degasus
17:20
camus has quit [Remote host closed the connection]
17:34
vup has joined #dri-devel
17:35
cheako has joined #dri-devel
17:37
shashanks has joined #dri-devel
17:39
tanty has quit [Server closed connection]
17:40
mszyprow has joined #dri-devel
17:40
tanty has joined #dri-devel
17:43
gouchi has joined #dri-devel
17:43
nsneck has quit [Server closed connection]
17:44
nsneck has joined #dri-devel
17:45
shashanks has quit [Ping timeout: 480 seconds]
17:46
ivyl has quit [Server closed connection]
17:46
ivyl has joined #dri-devel
17:46
thellstrom has joined #dri-devel
17:51
Guest838 has quit [Server closed connection]
17:51
dreda has joined #dri-devel
17:51
mszyprow has quit [Ping timeout: 480 seconds]
17:52
dreda is now known as Guest1387
18:03
AndroidC512L has joined #dri-devel
18:04
LaserEyess has quit [Server closed connection]
18:04
LaserEyess has joined #dri-devel
18:29
hakzsam has quit [Server closed connection]
18:30
hakzsam has joined #dri-devel
18:38
mripard has quit [Server closed connection]
18:38
mripard has joined #dri-devel
18:51
Wally has joined #dri-devel
18:51
tango_ has quit [Server closed connection]
18:51
tango_ has joined #dri-devel
18:51
sravn has quit [Server closed connection]
18:51
sravn has joined #dri-devel
18:55
LexSfX has joined #dri-devel
18:59
gawin has quit [Quit: Konversation terminated!]
19:09
xerpi[m] has quit [Server closed connection]
19:09
xerpi[m] has joined #dri-devel
19:11
DrNick has quit [Server closed connection]
19:11
DrNick has joined #dri-devel
19:11
DrNick is now known as Guest1401
19:12
heftig has quit [Server closed connection]
19:12
heftig has joined #dri-devel
19:13
danylo has quit [Server closed connection]
19:13
danylo has joined #dri-devel
19:18
ella-0[m] has quit [Server closed connection]
19:18
ella-0[m] has joined #dri-devel
19:19
Hi-Angel has joined #dri-devel
19:21
tzimmermann has joined #dri-devel
19:23
neobrain[m] has quit [Server closed connection]
19:23
neobrain[m] has joined #dri-devel
19:28
pcercuei has quit [Quit: brb]
19:31
pcercuei has joined #dri-devel
19:35
Wally has quit [Remote host closed the connection]
19:43
HayashiEsme[m] has quit [Server closed connection]
19:43
HayashiEsme[m] has joined #dri-devel
19:44
lemonzest has quit [Quit: WeeChat 3.4]
19:45
Newbyte has quit [Server closed connection]
19:45
Newbyte has joined #dri-devel
19:45
<
marex >
The powervr stuff looks great so far ... did they forget to include firmware ?
19:45
<
marex >
[ 1.833512] powervr fd000000.gpu: [drm]
*ERROR* failed to load firmware powervr/rogue_4.45.2.58_v1.14.fw (err=-2)
19:48
<
alyssa >
there's a linux-firmware fork too
19:49
<
marex >
aha, thank you
19:49
<
alyssa >
seems to be a different version than you want though
19:50
<
marex >
I wonder if this slightly different version would actually run ...
19:51
<
imirkin >
just a rename away
19:52
mszyprow has joined #dri-devel
19:52
<
marex >
imirkin: half-way there
19:54
<
marex >
[ 1.837114] powervr fd000000.gpu: [drm] loaded firmware powervr/rogue_4.45.2.58_v1.14.fw
19:54
<
marex >
[ 1.845235] powervr fd000000.gpu: [drm]
*ERROR* Unsupported fw version 4.40.2.51
19:54
<
imirkin >
they thought of everything
19:54
<
marex >
all right ... time for hex editor
19:54
dafna2[m] has quit [Server closed connection]
19:54
dafna2[m] has joined #dri-devel
19:54
<
imirkin >
well, probably it wants something in the new firmware
19:55
<
marex >
I suspect that 4.40.2.48 is some core version
19:55
<
imirkin >
(all those years of math finally paying off...)
19:56
<
imirkin >
ah yeah. looks convincing.
19:56
<
marex >
if you look around these 4.x headers, they all define what looks like different erratas to different hardware mutations
19:56
<
imirkin >
right. so presumably you need the one that matches your core.
19:57
<
marex >
either that, or it can work with some rendering corruption
19:57
<
marex >
so much for a quick test :)
19:58
<
marex >
hm, lemme comment out the firmware version checks first
20:02
mszyprow has quit [Ping timeout: 480 seconds]
20:30
<
simon-perretta-img >
marex: Just out of curiosity, what hardware are you testing that on?
20:31
<
marex >
simon-perretta-img: renesas r8a779{51,60,65,90}
20:31
<
marex >
right now, the r8a77960, because that's pvr 6250 too
20:31
<
marex >
i.e. closest to what should be supported
20:31
<
marex >
the 77951 is Unsupported BVNC: 4.46.6.62
20:34
<
alyssa >
marex: (also, we have #powervr now)
20:55
<
imirkin >
anholt: i think there's some deeper issue with deqp-runner missing results ... i'm getting a bunch of them with nfs, i've bumped the timeout to 5 mins and it didn't help
20:55
<
imirkin >
it's not just "nfs is slow"
20:55
<
imirkin >
still getting "Failure getting run results: parsing results: Reading from dEQP: timed out waiting for fd to be ready"
20:57
<
imirkin >
anholt: e.g. my c210.r1.qpa is just plain truncated towards the end
20:57
<
imirkin >
it's like there's a flush missing, or you don't wait for the glcts process to exit, or ...
21:02
gawin has joined #dri-devel
21:25
gawin has quit [Ping timeout: 480 seconds]
21:28
mszyprow has joined #dri-devel
21:37
Hi-Angel has quit [Remote host closed the connection]
21:50
nsneck has quit [Quit: bye]
21:51
nsneck has joined #dri-devel
21:51
danvet has quit [Ping timeout: 480 seconds]
21:57
rkanwal has joined #dri-devel
21:57
thellstrom has quit [Ping timeout: 480 seconds]
22:02
Duke`` has quit [Ping timeout: 480 seconds]
22:02
<
alyssa >
spec@arb_texture_float@fbo-blending-formats@GL_LUMINANCE32F_ARB
22:02
<
alyssa >
...does this pass for anyone?
22:02
<
alyssa >
the test seems bogus, I don't think that format has to be colour renderable?
22:03
<
imirkin >
but if it is, it probably has to work
22:04
<
imirkin >
Testing GL_LUMINANCE32F_ARB
22:04
<
imirkin >
PIGLIT: {"subtest": {"GL_LUMINANCE32F_ARB" : "pass"}}
22:04
<
imirkin >
(on NVIDIA GF108)
22:05
<
imirkin >
there's a slight annoyance in that e.g. D3D10 only requires certain formats to be blendable
22:05
<
imirkin >
but GL makes no such distinction
22:05
<
alyssa >
_mesa_validate_framebuffer refuses any L/A/I/R formats
22:05
neonking has quit []
22:05
<
alyssa >
which begs the question, how is piglit rendering anyway
22:05
<
alyssa >
luminance, intensity, etc formats are marked as GL_FRAMEBUFFER_UNSUPPORTED
22:06
<
imirkin >
i don't think that function is called
22:07
<
imirkin >
you should be able to GC it
22:07
<
imirkin >
probably stopped getting called in the dri driver removal
22:07
<
imirkin >
or a later cleanup
22:08
<
imirkin >
not sure that any of those are strictly required
22:08
<
alyssa >
rendering luminance/intensity formats is possible for panfrost, but I see no spec requirement so I went to disable the formats, but bam
22:09
<
imirkin >
disabling them should work
22:09
<
imirkin >
although ... maybe not? dunno
22:09
<
alyssa >
"should" ;)
22:09
<
imirkin >
i haven't looked at the compat spec in quite a while
22:09
<
imirkin >
some might be "req rend"
22:10
<
alyssa >
oh right I forgot opengl has a spec ;-p
22:10
<
alyssa >
that format does not exist in the 4.6 compat spec, it's ARB only
22:11
<
imirkin >
hrmph. very odd. they seem to be missing from the table in the GL 4.6 compat spec?
22:11
<
alyssa >
LUMINANCE32F_ARB is arb-only
22:11
<
alyssa >
and it's meant to be a texture format only, I suspect
22:11
<
alyssa >
and this is failing on a bunch of drivers so I assume either mesa/st bug or piglit bug
22:11
<
imirkin >
more likely "no one cared to make it pass"
22:12
<
alyssa >
(crocus, zink/lvp, zink/radv, v3d, freedreno)
22:12
<
imirkin >
crocus ... on gen4/5?
22:12
<
imirkin >
or later?
22:12
<
imirkin >
yeah, iirc that has issues with ... float
22:13
<
alyssa >
st_format has RGBA32F as a fallback for LUMNINANCE32F
22:13
<
alyssa >
RGBA32F is col renderable
22:13
Company has joined #dri-devel
22:13
<
imirkin >
you are mistaken.
22:13
<
alyssa >
(at least for us?)
22:13
<
imirkin >
then you aren't going to be desktop GL-compatible
22:14
<
imirkin >
RGBA32F is not only color-renderable, but it's "req rend" and "req tex"
22:14
<
imirkin >
oh. which is what you said.
22:14
<
imirkin >
i read that as "not color renderable"
22:14
<
alyssa >
I suspect mesa/st is deciding L32F is renderable, because R32F is renderable and it falls back L32F -> R32F
22:14
<
alyssa >
which, I mean, "technically accurate" but
22:15
<
imirkin >
should be fine, no?
22:15
<
imirkin >
when does that fail?
22:15
<
imirkin >
assuming you have texture swizzles in place
22:15
<
alyssa >
not texture swizzles, render target swizzles...
22:15
<
alyssa >
oh, wait. right.
22:15
<
imirkin >
luminance rendering is a pretty dubious thing in the first place, tbh
22:15
<
imirkin >
like let's say r != g, what to do
22:16
<
alyssa >
that piglit tests interactions that don't make any sense
22:16
<
imirkin >
it's reasonable to say "undefined" or "we'll just take r"
22:16
<
alyssa >
I want to just disable the luminance rendering for panfrost
22:16
<
alyssa >
the r!=g behaviour
22:16
<
imirkin >
i think you'd be within your rights to do so
22:16
<
imirkin >
in which case you should get skips
22:16
<
imirkin >
for that test
22:18
mszyprow has quit [Ping timeout: 480 seconds]
22:21
<
alyssa >
i don't think it happens though, due to the fallback in mesa/st
22:27
gouchi has quit [Remote host closed the connection]
22:27
LexSfX has joined #dri-devel
22:27
<
imirkin >
alyssa: but as far as the gallium driver is concerned, it should be fine i think?
22:28
<
imirkin >
oh, no, it's not
22:28
<
imirkin >
coz R32F = (r,0,0,1) while L32F is (l,l,l,0)
22:28
<
imirkin >
which makes the alpha blend stuff not work
22:28
<
alyssa >
mesa/st bug then?
22:28
<
imirkin >
i mean ...... yes and no
22:29
<
imirkin >
it's reasonable in general, and better than nothing
22:29
<
imirkin >
but it fails for this one case
22:29
<
imirkin >
this is true of lots of fallbacks in there
22:29
<
alyssa >
...do apps use L32F rendering?
22:29
Haaninjo has quit [Quit: Ex-Chat]
22:29
<
alyssa >
(I32F has a similar issue, I think?)
22:30
<
imirkin >
my solution was to just stick a L32F mapping in nouveau and a bunch of other drivers
22:30
<
imirkin >
and that tends to Just Work (tm)
22:30
<
imirkin >
but if you have super-custom blend you might need an extra line of code here and there
22:30
<
imirkin >
I is not renderable, i think
22:30
<
imirkin >
so less of an issue
22:30
<
alyssa >
piglit tries anyway
22:31
<
imirkin >
oh. hm. so it does.
22:31
<
imirkin >
and it passes on nouveau. heh.
22:31
<
imirkin >
i guess it's not as non-renderable as i remmebered ;)
22:32
<
imirkin >
let me glance at GL 3.0
22:33
<
imirkin >
that should be a lot more represntative. this stuff got moved around a lot since then
22:39
<
alyssa >
adding an L32F mapping that's marked as "texturable but not renderable" makes the problem go away
22:39
<
alyssa >
pretty stupid.
22:39
<
imirkin >
then st/mesa doesn't make the fallback.
22:40
<
imirkin >
if you have it explicitly in the driver
22:41
<
imirkin >
then st/mesa doesn't make the fallback to the other format
22:41
<
alyssa >
it's a stupid fix
22:41
<
alyssa >
mesa/st shouldn't be falling back to known-broken behaviour, for an optional feature
22:41
<
imirkin >
rendering L is a silly sub-feature
22:42
<
imirkin >
texturing it is a lot more plausible
22:42
<
imirkin >
and works totally fine.
22:42
<
imirkin >
with the fallback that is there
22:42
<
imirkin >
don't want to throw out the baby with the bathwater
22:42
<
alyssa >
mesa/st should be falling back for texturing but not for rendering, is my point
22:42
<
imirkin >
it's too late by then
22:42
<
imirkin >
step 1: make texture
22:42
<
imirkin >
step 2: bind texture to fbo
22:43
<
imirkin >
hard to make step 1 do something else if you don't know that step 2 will occur
22:43
<
imirkin >
for some super-basic formats, we require render support when creating a regular texture
22:43
<
alyssa >
mesa/st should not be advertising L32_FLOAT as a renderable format if it falls back to R32_FLOAT
22:43
<
imirkin >
(like RGBA8)
22:44
<
alyssa >
doing so is within the spec, that's what you get when you add a mapping for L32 -> R32 texture-only in the backend
22:44
<
imirkin >
that would require a
_lot_ more tracking
22:44
<
imirkin >
right now it says "oh yeah, i'll use PIPE_FORMAT_R32_FLOAT"
22:44
<
imirkin >
and then the other thing says "oh yeah, PIPE_FORMAT_R32_FLOAT is totally fine for rendering"
22:44
<
alyssa >
yes. that's the bug, right theere.
22:45
<
imirkin >
i would not oppose you doing the work to do all that extra tracking
22:45
<
imirkin >
but this is pretty much one of the very few cases where it matters
22:45
<
imirkin >
and/or is desirable
22:46
agx_ has quit [Server closed connection]
22:46
agx has joined #dri-devel
22:59
gawin has joined #dri-devel
23:07
JohnnyonFlame has quit [Ping timeout: 480 seconds]