ella-0_ has quit [Read error: Connection reset by peer]
saurabhg has joined #dri-devel
chslt^ has quit [Ping timeout: 480 seconds]
bmodem has quit []
mclasen_ has joined #dri-devel
chslt^ has joined #dri-devel
mclasen has quit [Ping timeout: 480 seconds]
Duke`` has joined #dri-devel
mclasen_ has quit [Ping timeout: 480 seconds]
Daanct12 has quit [Remote host closed the connection]
itoral has joined #dri-devel
Daanct12 has joined #dri-devel
aravind has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
chslt^ has quit [singleton.oftc.net synthon.oftc.net]
dviola has quit [singleton.oftc.net synthon.oftc.net]
swivel has quit [singleton.oftc.net synthon.oftc.net]
emersion has quit [singleton.oftc.net synthon.oftc.net]
sarnex has quit [singleton.oftc.net synthon.oftc.net]
everfree has quit [singleton.oftc.net synthon.oftc.net]
unerlige has quit [singleton.oftc.net synthon.oftc.net]
rsripada has quit [singleton.oftc.net synthon.oftc.net]
sumits has quit [singleton.oftc.net synthon.oftc.net]
mdnavare has quit [singleton.oftc.net synthon.oftc.net]
pzanoni has quit [singleton.oftc.net synthon.oftc.net]
mattrope has quit [singleton.oftc.net synthon.oftc.net]
sagar_ has quit [singleton.oftc.net synthon.oftc.net]
LexSfX has quit [singleton.oftc.net synthon.oftc.net]
JohnnyonFlame has quit [singleton.oftc.net synthon.oftc.net]
alanc has quit [singleton.oftc.net synthon.oftc.net]
krushia has quit [singleton.oftc.net synthon.oftc.net]
cheako has quit [singleton.oftc.net synthon.oftc.net]
flto has quit [singleton.oftc.net synthon.oftc.net]
rsripada_ has quit [singleton.oftc.net synthon.oftc.net]
anujp_ has quit [singleton.oftc.net synthon.oftc.net]
gpiccoli has quit [singleton.oftc.net synthon.oftc.net]
arnd_ has quit [singleton.oftc.net synthon.oftc.net]
zf has quit [singleton.oftc.net synthon.oftc.net]
Lightsword_ has quit [singleton.oftc.net synthon.oftc.net]
jhli_ has quit [singleton.oftc.net synthon.oftc.net]
sumoon_ has quit [singleton.oftc.net synthon.oftc.net]
codingkoopa9 has quit [singleton.oftc.net synthon.oftc.net]
Kayden has quit [singleton.oftc.net synthon.oftc.net]
xperia64 has quit [singleton.oftc.net synthon.oftc.net]
Ryback_ has quit [singleton.oftc.net synthon.oftc.net]
mattst88 has quit [singleton.oftc.net synthon.oftc.net]
naseer has quit [singleton.oftc.net synthon.oftc.net]
orbea has quit [singleton.oftc.net synthon.oftc.net]
rg3igalia has quit [singleton.oftc.net synthon.oftc.net]
andrey-konovalov has quit [singleton.oftc.net synthon.oftc.net]
lstrano_ has quit [singleton.oftc.net synthon.oftc.net]
Lyude has quit [singleton.oftc.net synthon.oftc.net]
soreau has quit [singleton.oftc.net synthon.oftc.net]
Peuc has quit [singleton.oftc.net synthon.oftc.net]
kem has quit [singleton.oftc.net synthon.oftc.net]
oneforall2 has quit [singleton.oftc.net synthon.oftc.net]
ZeZu has quit [singleton.oftc.net synthon.oftc.net]
aswar002 has quit [singleton.oftc.net synthon.oftc.net]
eukara has quit [singleton.oftc.net synthon.oftc.net]
jstultz has quit [singleton.oftc.net synthon.oftc.net]
bluebugs has quit [singleton.oftc.net synthon.oftc.net]
rpigott has quit [singleton.oftc.net synthon.oftc.net]
agd5f has quit [singleton.oftc.net synthon.oftc.net]
jljusten has quit [singleton.oftc.net synthon.oftc.net]
rcf has quit [singleton.oftc.net synthon.oftc.net]
cphealy has quit [singleton.oftc.net synthon.oftc.net]
hfink has quit [singleton.oftc.net synthon.oftc.net]
Karyon has quit [singleton.oftc.net synthon.oftc.net]
rohiiyer0 has quit [singleton.oftc.net synthon.oftc.net]
tlwoerner has quit [singleton.oftc.net synthon.oftc.net]
bl4ckb0ne has quit [singleton.oftc.net synthon.oftc.net]
vyivel has quit [singleton.oftc.net synthon.oftc.net]
JTL has quit [singleton.oftc.net synthon.oftc.net]
radii_ has quit [singleton.oftc.net synthon.oftc.net]
narmstrong_ has quit [singleton.oftc.net synthon.oftc.net]
Rayyan has quit [singleton.oftc.net synthon.oftc.net]
ifreund has quit [singleton.oftc.net synthon.oftc.net]
rib has quit [singleton.oftc.net synthon.oftc.net]
tfiga has quit [singleton.oftc.net synthon.oftc.net]
Prf_Jakob has quit [singleton.oftc.net synthon.oftc.net]
kennylevinsen has quit [singleton.oftc.net synthon.oftc.net]
lumag_ has quit [singleton.oftc.net synthon.oftc.net]
Akari has quit [singleton.oftc.net synthon.oftc.net]
pixelcluster has quit [singleton.oftc.net synthon.oftc.net]
mwalle has quit [singleton.oftc.net synthon.oftc.net]
BobBeck has quit [singleton.oftc.net synthon.oftc.net]
gallo2 has quit [singleton.oftc.net synthon.oftc.net]
dwlsalmeida has quit [singleton.oftc.net synthon.oftc.net]
pochu has quit [singleton.oftc.net synthon.oftc.net]
yoslin has quit [singleton.oftc.net synthon.oftc.net]
illwieckz has quit [singleton.oftc.net synthon.oftc.net]
MrCooper has quit [singleton.oftc.net synthon.oftc.net]
dos1 has quit [singleton.oftc.net synthon.oftc.net]
FLHerne has quit [singleton.oftc.net synthon.oftc.net]
lplc has quit [singleton.oftc.net synthon.oftc.net]
dliviu has quit [singleton.oftc.net synthon.oftc.net]
dv_ has quit [singleton.oftc.net synthon.oftc.net]
columbarius has quit [singleton.oftc.net synthon.oftc.net]
rsalvaterra has quit [singleton.oftc.net synthon.oftc.net]
minecrell has quit [singleton.oftc.net synthon.oftc.net]
angerctl has quit [singleton.oftc.net synthon.oftc.net]
PiGLDN[m] has quit [singleton.oftc.net synthon.oftc.net]
Anson[m] has quit [singleton.oftc.net synthon.oftc.net]
onox[m] has quit [singleton.oftc.net synthon.oftc.net]
naheemsays[m] has quit [singleton.oftc.net synthon.oftc.net]
undvasistas[m] has quit [singleton.oftc.net synthon.oftc.net]
reactormonk[m] has quit [singleton.oftc.net synthon.oftc.net]
sigmoidfunc[m] has quit [singleton.oftc.net synthon.oftc.net]
chivay has quit [singleton.oftc.net synthon.oftc.net]
pushqrdx[m] has quit [singleton.oftc.net synthon.oftc.net]
yshui` has quit [singleton.oftc.net synthon.oftc.net]
Vin[m] has quit [singleton.oftc.net synthon.oftc.net]
sigmaris has quit [singleton.oftc.net synthon.oftc.net]
enunes has quit [singleton.oftc.net synthon.oftc.net]
DavidHeidelberg[m] has quit [singleton.oftc.net synthon.oftc.net]
linkmauve has quit [singleton.oftc.net synthon.oftc.net]
shadeslayer has quit [singleton.oftc.net synthon.oftc.net]
adavy has quit [singleton.oftc.net synthon.oftc.net]
danylo has quit [singleton.oftc.net synthon.oftc.net]
bnieuwenhuizen has quit [singleton.oftc.net synthon.oftc.net]
glennk has quit [singleton.oftc.net synthon.oftc.net]
jeeeun841 has quit [singleton.oftc.net synthon.oftc.net]
go4godvin has quit [singleton.oftc.net synthon.oftc.net]
Arsen has quit [singleton.oftc.net synthon.oftc.net]
tomba has quit [singleton.oftc.net synthon.oftc.net]
bylaws has quit [singleton.oftc.net synthon.oftc.net]
karolherbst has quit [singleton.oftc.net synthon.oftc.net]
agx has quit [singleton.oftc.net synthon.oftc.net]
padovan has quit [singleton.oftc.net synthon.oftc.net]
tomeu has quit [singleton.oftc.net synthon.oftc.net]
sravn has quit [singleton.oftc.net synthon.oftc.net]
Lynne has quit [singleton.oftc.net synthon.oftc.net]
libv has quit [singleton.oftc.net synthon.oftc.net]
akselmo has quit [singleton.oftc.net synthon.oftc.net]
Newbyte has quit [singleton.oftc.net synthon.oftc.net]
Emantor has quit [singleton.oftc.net synthon.oftc.net]
mriesch has quit [singleton.oftc.net synthon.oftc.net]
pjakobsson has quit [singleton.oftc.net synthon.oftc.net]
jernej has quit [singleton.oftc.net synthon.oftc.net]
mceier has quit [singleton.oftc.net synthon.oftc.net]
Lucretia has quit [singleton.oftc.net synthon.oftc.net]
loki_val_ has quit [singleton.oftc.net synthon.oftc.net]
i-garrison has quit [singleton.oftc.net synthon.oftc.net]
haagch has quit [singleton.oftc.net synthon.oftc.net]
Thymo has quit [singleton.oftc.net synthon.oftc.net]
knr has quit [singleton.oftc.net synthon.oftc.net]
haasn has quit [singleton.oftc.net synthon.oftc.net]
DragoonAethis has quit [singleton.oftc.net synthon.oftc.net]
DPA- has quit [singleton.oftc.net synthon.oftc.net]
pendingchaos has quit [singleton.oftc.net synthon.oftc.net]
Ella[m] has quit [singleton.oftc.net synthon.oftc.net]
jkhsjdhjs has quit [singleton.oftc.net synthon.oftc.net]
aissen has quit [singleton.oftc.net synthon.oftc.net]
ds` has quit [singleton.oftc.net synthon.oftc.net]
jkqxz has quit [singleton.oftc.net synthon.oftc.net]
unevenrhombus[m] has quit [singleton.oftc.net synthon.oftc.net]
viciouss[m] has quit [singleton.oftc.net synthon.oftc.net]
bluepenquin has quit [singleton.oftc.net synthon.oftc.net]
urja has quit [singleton.oftc.net synthon.oftc.net]
exit70[m] has quit [singleton.oftc.net synthon.oftc.net]
cyrozap has quit [singleton.oftc.net synthon.oftc.net]
martijnbraam has quit [singleton.oftc.net synthon.oftc.net]
Strit[m] has quit [singleton.oftc.net synthon.oftc.net]
tintou has quit [singleton.oftc.net synthon.oftc.net]
KunalAgarwal[m]1 has quit [singleton.oftc.net synthon.oftc.net]
marex has quit [singleton.oftc.net synthon.oftc.net]
simon-perretta-img has quit [singleton.oftc.net synthon.oftc.net]
nanonyme has quit [singleton.oftc.net synthon.oftc.net]
kchibisov has quit [singleton.oftc.net synthon.oftc.net]
glehmann has quit [singleton.oftc.net synthon.oftc.net]
dj-death has quit [singleton.oftc.net synthon.oftc.net]
ickle_ has quit [singleton.oftc.net synthon.oftc.net]
mupuf has quit [singleton.oftc.net synthon.oftc.net]
egbert has quit [singleton.oftc.net synthon.oftc.net]
Guest2694 has quit [singleton.oftc.net synthon.oftc.net]
vsyrjala has quit [singleton.oftc.net synthon.oftc.net]
Thaodan has quit [singleton.oftc.net synthon.oftc.net]
ced117 has quit [singleton.oftc.net synthon.oftc.net]
kgz has quit [singleton.oftc.net synthon.oftc.net]
OftenTimeConsuming has quit [singleton.oftc.net synthon.oftc.net]
ella-0 has quit [singleton.oftc.net synthon.oftc.net]
srslypascal has quit [singleton.oftc.net synthon.oftc.net]
jfalempe has quit [singleton.oftc.net synthon.oftc.net]
italove has quit [singleton.oftc.net synthon.oftc.net]
leandrohrb has quit [singleton.oftc.net synthon.oftc.net]
milek7 has quit [singleton.oftc.net synthon.oftc.net]
cmeissl[m] has quit [singleton.oftc.net synthon.oftc.net]
RAOF has quit [singleton.oftc.net synthon.oftc.net]
g0b has quit [singleton.oftc.net synthon.oftc.net]
alatiera has quit [singleton.oftc.net synthon.oftc.net]
kj has quit [singleton.oftc.net synthon.oftc.net]
Adrinael has quit [singleton.oftc.net synthon.oftc.net]
vup2 has quit [singleton.oftc.net synthon.oftc.net]
gruetzkopf has quit [singleton.oftc.net synthon.oftc.net]
tanty has quit [singleton.oftc.net synthon.oftc.net]
moony has quit [singleton.oftc.net synthon.oftc.net]
llyyr has quit [singleton.oftc.net synthon.oftc.net]
kallisti5[m] has quit [singleton.oftc.net synthon.oftc.net]
CME has quit [singleton.oftc.net synthon.oftc.net]
Company has quit [singleton.oftc.net synthon.oftc.net]
alarumbe has quit [singleton.oftc.net synthon.oftc.net]
ppascher has quit [singleton.oftc.net synthon.oftc.net]
tango_ has quit [singleton.oftc.net synthon.oftc.net]
bbrezillon has quit [singleton.oftc.net synthon.oftc.net]
Surkow|laptop has quit [singleton.oftc.net synthon.oftc.net]
TMM has quit [singleton.oftc.net synthon.oftc.net]
digetx has quit [singleton.oftc.net synthon.oftc.net]
RSpliet has quit [singleton.oftc.net synthon.oftc.net]
jadahl has quit [singleton.oftc.net synthon.oftc.net]
iokill has quit [singleton.oftc.net synthon.oftc.net]
javierm has quit [singleton.oftc.net synthon.oftc.net]
lanodan has quit [singleton.oftc.net synthon.oftc.net]
itoral has quit [singleton.oftc.net synthon.oftc.net]
zamundaaa[m] has quit [singleton.oftc.net synthon.oftc.net]
Andy[m] has quit [singleton.oftc.net synthon.oftc.net]
neobrain has quit [singleton.oftc.net synthon.oftc.net]
shoragan has quit [singleton.oftc.net synthon.oftc.net]
dafna33[m] has quit [singleton.oftc.net synthon.oftc.net]
KunalAgarwal[m] has quit [singleton.oftc.net synthon.oftc.net]
srslypascal has joined #dri-devel
i-garrison has joined #dri-devel
Thaodan has joined #dri-devel
Sachiel has joined #dri-devel
melissawen has joined #dri-devel
robher has joined #dri-devel
kts has quit [Quit: Konversation terminated!]
sergi has joined #dri-devel
Major_Biscuit has joined #dri-devel
sergi has quit []
sergi has joined #dri-devel
sergi has quit []
tzimmermann has joined #dri-devel
alanc has joined #dri-devel
frieder has joined #dri-devel
sergi has joined #dri-devel
kts has joined #dri-devel
gouchi has joined #dri-devel
gouchi has quit [Remote host closed the connection]
thellstrom has joined #dri-devel
Daaanct12 has joined #dri-devel
crabbedhaloablut has joined #dri-devel
Daaanct12 has quit [Remote host closed the connection]
Daanct12 has quit [Ping timeout: 480 seconds]
thellstrom has quit [Ping timeout: 480 seconds]
aravind has quit [Read error: Connection reset by peer]
rasterman has joined #dri-devel
thellstrom has joined #dri-devel
Major_Biscuit has quit [Ping timeout: 480 seconds]
mvlad has joined #dri-devel
wvanhauwaert has joined #dri-devel
wv has joined #dri-devel
Terman has joined #dri-devel
wvanhauwaert has quit [Ping timeout: 480 seconds]
Major_Biscuit has joined #dri-devel
chaim has joined #dri-devel
wv has quit []
wvanhauwaert has joined #dri-devel
Daanct12 has joined #dri-devel
lynxeye has joined #dri-devel
tursulin has joined #dri-devel
Major_Biscuit has quit [Ping timeout: 480 seconds]
camus has quit [Read error: Connection reset by peer]
Major_Biscuit has joined #dri-devel
camus has joined #dri-devel
Daanct12 has quit [Ping timeout: 480 seconds]
tursulin has quit [Ping timeout: 480 seconds]
rpigott has joined #dri-devel
vliaskov has joined #dri-devel
pcercuei has joined #dri-devel
saurabh_1 has joined #dri-devel
Major_Biscuit has quit [Ping timeout: 480 seconds]
apinheiro has joined #dri-devel
camus has quit [Remote host closed the connection]
camus has joined #dri-devel
bmodem has joined #dri-devel
whald has joined #dri-devel
rkanwal has joined #dri-devel
<whald>
hi! i'm failing to use gbm_bo_create any BO with GBM_FORMAT_NV12 and could use some pointers as to when this should succeed. with all my attempts I only get "invalid argument" errors, on Intel and amdgpu. the RGB formats work without problems, though.
<whald>
should I use the gbm_bo_create_with_modifiers(2) function? and if, which modifiers have a chance to succeed?
<lynxeye>
whald: Most (all?) drivers don't support rendering into NV12, so they won't allow to allocate GBM buffers in this format. If you only need to construct some buffers for sampling, you can allocate the individual planes with R and RG format and then import them.
mclasen has joined #dri-devel
<whald>
lynxeye, thanks for the info. I somehow suspected I have to got that route but had hope I can skip the extra effort. :-)
<pq>
lynxeye, even when omitting GBM_BO_USE_RENDERING ? Or did whald use that?
<whald>
pq, lynxeye, I tried several GBM_BO_USE flags like rendering, linear, 0, it's been always the same for me.
<pq>
aha
<whald>
what I really want to do is to have those buffers, gbm_bo_map them, dump some YUV data there and then use them as texture in GL by importing using that EGL dmabuf import extension (forgot the name). that all works when going the RGB route, but I'd like to do the RGB -> YUV conversion in a shader and I'd also like not to blow up the YUV data to RGB (for YUV420 it's 1.5 bytes per pixel, for RGB it's 4 bytes and that hurts)
<lynxeye>
pq: Most drivers don't even know what NV12 is. They can deal with it on import, but even then most drivers rely on the frontend to lower the import into R and RG buffers and rewrite the shader for them.
<lynxeye>
There is no such lowering for the allocation path.
<lynxeye>
whald: Yep, although a bit clumsy, allocating the NV12 planes as R and RG buffers of the right dimensions will probably work. Then you just dump your Y and UV data into those buffers and import them as a EGL image and treat them as NV12 at that point.
devilhorns has joined #dri-devel
<pq>
alright
KunalAgarwal[m] has joined #dri-devel
KunalAgarwal[m] has quit []
KunalAgarwal[m] has joined #dri-devel
icecream95 has quit [Ping timeout: 480 seconds]
mi6x3m has joined #dri-devel
<mi6x3m>
Hey Mesa, why isn't there a vulkan mega driver like there is with OpenGL?
<mi6x3m>
so the whole thing can work nicely with the Vulkan loader from Khronos?
epoll has joined #dri-devel
<pixelcluster>
I doubt the loader has anything to do with it
<pixelcluster>
In fact the code that sets up entrypoints for each driver is shared among all Vulkan drivers
Namarrgon has joined #dri-devel
<mi6x3m>
i mean yeah but they are still separate .so files
<mi6x3m>
in contrast to gallium where all drivers are 1 .so file
heat_ has joined #dri-devel
srslypascal is now known as Guest4253
srslypascal has joined #dri-devel
kts has quit [Quit: Konversation terminated!]
Guest4253 has quit [Ping timeout: 480 seconds]
<javierm>
daniels: hi, it seems dri-devel is again having issues again today?
<daniels>
javierm: just flushed a load of stuff
bmodem has quit [Ping timeout: 480 seconds]
kts has joined #dri-devel
mclasen has quit [Ping timeout: 480 seconds]
<javierm>
daniels: Ok, thanks! It worked
<daniels>
np
wv has joined #dri-devel
kts has quit [Quit: Konversation terminated!]
wvanhauwaert has quit [Remote host closed the connection]
wv__ has joined #dri-devel
Peste_Bubonica has joined #dri-devel
wv has quit [Ping timeout: 480 seconds]
<pinchartl>
lynxeye: do you know what hardware could support YUV "rendering" ? we're looking at implementing a camera ISP in software with the GPU
<lynxeye>
pinchartl: I don't think any HW can do it natively. You could probably emulate YUV444 via MRT output from the fragment shader and I think new Vivante GPUs can do a blit from RGB to subsampled YUV, but not sure how useful that would be for your use-case.
<pinchartl>
packed YUV 4:4:4 should be fine, as that's 3x8b per pixel, but it would be nice to support NV12 as that's what many encoders expect
kts has joined #dri-devel
<pinchartl>
well, worst case an RGB output is still better than nothing
<pq>
maybe you can first render a temporary YUV444 and then implement a pass to convert that to NV12 or whatever. Or is two-pass prohibitive?
<pq>
1st pass: render Y into NV12 Y channel and UV into full resolution temp; 2nd pass: take UV temp and downsample into NV12 UV
mclasen has joined #dri-devel
<pq>
or maybe multi-render-target + multi-sampling rendering could offer tricks?
<pinchartl>
that's a good idea
<pinchartl>
there's lots to experiment with :-)
<pq>
multi-render into full-resolution Y and sub-sized UV; Y converts each sub-sample into separate pixel while UV collapses the 4 sub-samples into one? I've no idea if GPUs bend into that.
<pinchartl>
there's plenty of work to be done. at the moment we're trying to figure out what is the right device to pass to gbm :-) /dev/dri/card0 will work in many cases, but for systems that have standalone GPUs that may not be right
* pinchartl
is a bit lost in the gbm-related EGL extensions
<pq>
Do you need GBM for allocating buffers or only for rendering?
<pq>
OpenGL or Vulkan?
<pinchartl>
to allocate buffers
<pq>
I guess allocating buffers is the primary problem... but which device to use depends on where those buffers are intended to be used: in KMS, or only fed back to the GPU again? or to some other device?
<pinchartl>
opengl for now, vulkan support in lower-end ARM platforms is not great yet
<pinchartl>
they're primarily meant for rendering by the GPU-ISP. they may be shared with other devices, but that's currently a separate issue
<pinchartl>
we're still missing a good API for cross-device buffer allocation
<pq>
exactly :-/
* pinchartl
wonders what happened to the unified memory allocator
agx has joined #dri-devel
<pinchartl>
while on the EGL topic
<pq>
with or without the unified memory allocator, I believe you first need to know which devices are going to be acessing the buffer before you can allocate
<pinchartl>
but I've seen sources that told it can also be used for off-screen rendering
<pinchartl>
when given a GBM buffer
<pinchartl>
is there any good overview/tutorial document on the topic of off-screen rendering and how it ties to EGL and GBM ?
<pq>
pinchartl, what "onscreen" means is up to the EGL platform to define.
<pq>
in EGL GBM platform case, "onscreen" only means that the buffers are compatible with the given KMS device. They don't magically get actually displayed, you still need to drive KMS manually.
<pinchartl>
aahhh
<pinchartl>
of course if "onscreen" doesn't mean "on screen", then, ... :-)
<pinchartl>
the name "window" in the function is also misleading then. I suppose it comes from historical reasons
<pq>
indeed
<pq>
in EGL terminology, GBM is a "window system"
<pinchartl>
so that's the right API to create a surface for off-screen rendering ?
<pq>
GBM has the API so that after you eglSwapBuffers, you get the gbm_bo you just rendered into, and then export dmabuf and/or import to KMS, etc.
<pq>
pinchartl, there are a few :-)
<pq>
EGL Surfaceless platform is the easiest to initialize, but you cannot tell it which GPU to use, and it won't allocate interoperable buffers for you.
<pinchartl>
and I suppose there's no single API that I can/should use on all platforms, because that would be too easy ? :-)
<pq>
EGL Device platform allows you to choose your GPU, but it's not an interoperable allocator either, AFAIU.
<pq>
EGL GBM platform, using gbm_surface, can give you buffers that are compatible with your chosen KMS device.
<pq>
all EGL platforms also support off-screen rendering a few ways (some, like Surfaceless, support nothing else)
<pinchartl>
thanks for the information
<pinchartl>
we'll keep investigating
<pq>
there are EGL pbuffer, which is perhaps not useful here as you cannot export them
<pinchartl>
if anyone is aware of a good tutorial, that would be really helpful too
saurabhg has quit [Ping timeout: 480 seconds]
<pinchartl>
yes, I need dmabuf import and export here. that's why I was looking at GBM
saurabh_1 has quit [Ping timeout: 480 seconds]
<pq>
you can also allocate with whatever, and then import that as dmabuf through EGL into a GL texture, and then use GL FBO to render into that texture
<pq>
allocating is again the problem
<pinchartl>
yes, sometimes I don't have a "whatever" to allocate from :-)
<pq>
then you cannot allocate
<pq>
stop.
<pinchartl>
well, the GPU can be one of those "whatever". I meant I don't have a separate standard API to retrieve buffer from
<pq>
yeah, that's the elephant in everyone's house
<pq>
for the rendering side, I would perhaps suggest EGL Device platform and importing elsewhere allocated dmabuf for rendering
<pq>
but if you want to go with EGL GBM platform + dmabuf import, that should work equally well although it is much more Linux specific
<pq>
if you go with EGL GBM platform + gbm_surface, then you are supposed to initialize GBM with a KMS device, and the buffers you get from it will be suitable for the KMS device, but who knows if they are good for anything else.
<pinchartl>
so GBM can't allocate buffers from a standalone GPU, it needs a KMS device ?
<pq>
it can, but :-)
<pq>
that's "allocate somewhere else" case: initialize gbm_device with whatever device you want to allocate from, and use gbm_bo_create_with_modifiers2() or so
<pq>
export a dmabuf, import through EGL if you want to render to it with OpenGL, and import to anywhere else you want - hoping it works
<pq>
pixel format modifiers and GBM usage flags only go so far, but it's probably the best bet right now, I guess
* pinchartl
will sleep over it
<pinchartl>
thanks again for all the pointers, that was very helpful
<pq>
there is still the open question of which DRM device to initialize gbm_device with, and is that any good to begin with for whatever encoder device you want to feed those dmabuf to
<pq>
if the intention is *not* to display those buffers with KMS, then use a render node for GBM, I think
<pinchartl>
in practice I suppose I'll have a single GPU and a single display device on the platforms that this code will run on (at least in production, for development we can run on x86, so there are more options)
<pq>
yeah, hardware knowledge will help you make the right choices with these questions that do not have a generic answer
<pq>
alternatively, maybe you need to allocate through the encoder device? but can the GPU write to the buffer then?
<pq>
presumably the hardware platform vendor has thought about how this should work? ;-)
pcercuei has quit [Read error: Connection reset by peer]
pcercuei has joined #dri-devel
<pq>
pinchartl, so, yeah. I guess the conclusion is: gbm_device from the render node, gbm_bo_create_with_modifiers2(), export dmabuf, and go on from there, for the allocation.
<pq>
in that case for rendering, EGL GBM platform vs. EGL Device platform probably doesn't make a difference.
<pq>
so might as well go with GBM platform, since you already have the gbm_device
<pq>
ok, I guess that's the end of the lecture :-D
apinheiro has quit [Ping timeout: 480 seconds]
chslt^ has joined #dri-devel
rgallaispou has joined #dri-devel
zehortigoza has joined #dri-devel
saurabhg has joined #dri-devel
fxkamd has joined #dri-devel
agd5f has joined #dri-devel
loki_val has joined #dri-devel
srslypascal is now known as Guest4264
srslypascal has joined #dri-devel
camus has quit []
saurabh_1 has joined #dri-devel
crabbedhaloablut has quit [Ping timeout: 480 seconds]
Guest4264 has quit [Ping timeout: 480 seconds]
saurabhg has quit [Ping timeout: 480 seconds]
sergi has quit [Ping timeout: 480 seconds]
mclasen_ has joined #dri-devel
zehortigoza has quit [Remote host closed the connection]
<alyssa>
nir_opt_remove_exit_barriers, delete the barrier in things like
<alyssa>
kernel main() { lots of work; barrier(); }
<alyssa>
Oddly expressing this in NIR is a lot trickier than in my backend IR
fxkamd has quit []
<alyssa>
One reason NOT to do this in NIR is that better results might be possible after scheduling (if your backend scheduler can sink barriers)
<alyssa>
(I think ours can)
<alyssa>
I suppose that pass could be beefed up to ignore instructions that don't interact with the barrier, but that's starting to sound really complicated and maybe wontfix
technopoirot has joined #dri-devel
chslt^ has quit [Ping timeout: 480 seconds]
<alyssa>
---
<alyssa>
I see llvmpipe timeout on dEQP-GLES31.functional.ssbo.layout.random.all_shared_buffer.36
<alyssa>
This is a notoriously hard test for RA, maybe should be skipped for llvmpipe?
<zmike>
really feels like that's not going to be happy with non-spirv shaders
<jekstrand>
zmike: That comment is a bit of a lie. It has nothing to do with GLSL vs. SPIR-V. It has everything to do with whether or not you have explicitly sized types.
<zmike>
hm
<zmike>
I guess I'll fire off some runs locally and see what happens
<alyssa>
IDK what election it is but with a campaign slogan about unbreaking CI, you have my vote
chslt^ has quit [Ping timeout: 480 seconds]
thellstrom has quit [Ping timeout: 480 seconds]
sergi has quit []
pcercuei has quit [Quit: brb]
<anholt_>
wait, today is a us holiday for google, too? I should probably go play in the dirt or something instead of playing in ci.
pcercuei has joined #dri-devel
pcercuei has quit []
sergi has joined #dri-devel
pcercuei has joined #dri-devel
kts has quit [Quit: Konversation terminated!]
sergi has quit [Ping timeout: 480 seconds]
ngcortes has joined #dri-devel
sergi has joined #dri-devel
pcercuei has quit [Quit: brb]
pcercuei has joined #dri-devel
saurabhg has quit [Ping timeout: 480 seconds]
camus has quit [Ping timeout: 480 seconds]
pcercuei has quit [Quit: brb]
pcercuei has joined #dri-devel
sul has joined #dri-devel
gouchi has joined #dri-devel
gouchi has quit []
chslt^ has joined #dri-devel
CME_ has joined #dri-devel
sergi has quit [Ping timeout: 480 seconds]
<zmike>
jekstrand: good news, your proposed changes didn't make cts explode
<jekstrand>
zmike: :D
<jekstrand>
zmike: Should I be holding my breath for bad news?
<zmike>
no
<zmike>
there is no bad news this time
<jekstrand>
:D
apinheiro has joined #dri-devel
maxzor__ has joined #dri-devel
TMM has joined #dri-devel
mvlad has quit [Remote host closed the connection]
<mmind00>
bbrezillon: re dw-hdmi-qp ... not that I know of ... I just received a rk3588 radxa board yesterday, so haven't even looked to much at it yet
Haaninjo has quit [Quit: Ex-Chat]
ngcortes has quit [Ping timeout: 480 seconds]
pixelclu- has joined #dri-devel
pixelclu- has quit [Ping timeout: 480 seconds]
pixelcluster has joined #dri-devel
zehortigoza has quit [Remote host closed the connection]
bnieuwenhuizen has joined #dri-devel
bnieuwenhuizen has quit []
bnieuwenhuizen has joined #dri-devel
bnieuwenhuizen is now known as Guest4288
Guest4288 is now known as bnieuwenhuizen
maxzor_ has joined #dri-devel
maxzor__ has quit [Ping timeout: 480 seconds]
wvanhauwaert has joined #dri-devel
wv__ has quit [Read error: Connection reset by peer]
vsyrjala_ has joined #dri-devel
<vliaskov>
Hi, I see an Xorg crash on TigerLake-H GT1 [UHD Graphics] [8086:9a60] in iris https://pastebin.com/5T98zfKx on older Mesa-20.2.4. it looks like iris_bo_map, called from iris_texture_subdata, fails, returning NULL dst and causing the crash. I saw upstream "0022a11f iris: Disable tiled memcpy for Tile4", part of a larger XeHP+ patchset https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/12132 . But I am
<vliaskov>
assuming the UHD graphics I mentioned is not XeHP+ nchery, is it? thanks for any pointers regarding this.
<nchery>
vliaskov: No, it's not XeHP+
<vliaskov>
thanks, good to know this isn't relevant
wv has joined #dri-devel
wvanhauwaert has quit [Read error: Connection reset by peer]
wv__ has joined #dri-devel
<nchery>
vliaskov: running with the environment variable INTEL_DEBUG=buf might give some more clues as to why dst is NULL
wv has quit [Ping timeout: 480 seconds]
<vliaskov>
thanks, I don't have the system atm, but I 'll retry soon with that, and repost here if I have more clues/questions
alyssa has left #dri-devel [#dri-devel]
krushia has joined #dri-devel
sul has quit [Ping timeout: 480 seconds]
sul has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
apinheiro has quit [Quit: Leaving]
Major_Biscuit has joined #dri-devel
rkanwal has quit [Ping timeout: 480 seconds]
wv has joined #dri-devel
Major_Biscuit has quit [Ping timeout: 480 seconds]
ngcortes has joined #dri-devel
wv__ has quit [Ping timeout: 480 seconds]
pcercuei has quit [Quit: dodo]
FireBurn has joined #dri-devel
<FireBurn>
Hi, is there a way to bisect easily when there are multiple patches that either need to be applied or reverted from the tree?
<Sachiel>
you can have a script do it and use git bisect run
rgallaispou1 has joined #dri-devel
<FireBurn>
And how would it know what to revert / apply?
<Sachiel>
if you have to figure that out each time, you are out of luck