<bkrause>
hi, anyone here with access to the dri-devel mailing list administration? looks like one of my mails (sent while I was not yet subscribed) got stuck in moderation
<danvet>
tzimmermann, fd.o m-l are a bit wobbly, emersion is working on it
<danvet>
should recover now
<tzimmermann>
danvet, thanks
srslypascal has quit [Ping timeout: 480 seconds]
frankbinns has joined #dri-devel
Danct12 has quit [Ping timeout: 480 seconds]
Danct12 has joined #dri-devel
Haaninjo has joined #dri-devel
srslypascal has joined #dri-devel
Daaanct12 has quit [Ping timeout: 480 seconds]
rkanwal has joined #dri-devel
srslypascal is now known as Guest2365
srslypascal has joined #dri-devel
Guest2365 has quit [Ping timeout: 480 seconds]
mclasen has joined #dri-devel
srslypascal has quit [Ping timeout: 480 seconds]
gawin has quit [Ping timeout: 480 seconds]
<javierm>
tzimmermann: did you see Zack report on a NULL pointer reference on remove conflicting framebuffers ?
<javierm>
I'm not able to reproduce but looking at the call stack, it seems to me that's a double free of the platform device
nchery has quit [Read error: Connection reset by peer]
YuGiOhJCJ has quit [Remote host closed the connection]
YuGiOhJCJ has joined #dri-devel
<javierm>
tzimmermann: and I think that can happen with simpledrm if both sysfb and the aperture core try to unregister the pdev
<javierm>
tzimmermann: in drm_aperture_detach_firmware()
<javierm>
unsure how that could be prevented, because sysfb has no way to know if something else unregistered a device that was registered by it
<javierm>
we really need that unified aperture register and unregistration, I may make some time to work on that on top of your patch that moved out of drm
frankbinns1 has joined #dri-devel
flacks has quit [Quit: Quitter]
frankbinns has quit [Remote host closed the connection]
flacks has joined #dri-devel
<javierm>
tzimmermann: that way sysfb could call to the aperture helpers to acquire and release the pdev and we could have a central place to do it
JohnnyonFlame has joined #dri-devel
<tzimmermann>
javierm, ok
<tzimmermann>
make sense
<tzimmermann>
javierm, you should reach out to alex williamson. i told him to add my patch to the vfio tree. you'd now need a topic branch, i guess. or you revert the current syfb patch and wait for 5.20
<tzimmermann>
javierm, an update for vfio has been posted yesterday. maybe you can review
sagar_ has quit [Read error: Connection reset by peer]
sagar_ has joined #dri-devel
anarsoul|2 has joined #dri-devel
anarsoul has quit [Read error: No route to host]
saurabhg has joined #dri-devel
neoXite___ has quit []
Company has joined #dri-devel
MajorBiscuit has joined #dri-devel
<tzimmermann>
javierm, i think i just saw that bug :/
<javierm>
tzimmermann: I think that the we should just revert the part that unregister the device in sysfb and keep the sysfb disable
<javierm>
since that part would be handled by the unified aperture core
<tzimmermann>
javier, can you please send the revert?
<tzimmermann>
reverting patch 3 should do the job
<javierm>
tzimmermann: yes, I will but first wanted to understand the problem better and get more info from zack
<javierm>
tzimmermann: I don't think we need a revert of patch #3 but just a partial revert of patch #2 that drops the sysfb_unregister() bit
<javierm>
since as mentioned we still want to disable sysfb pdev registration, regardless of what unregister the pdev
<tzimmermann>
javierm, simpledrm got released, but the global variable is still there
<tzimmermann>
and as sysfb_disable is called unconditionally, it'll free the device again
saurabhg has quit [Ping timeout: 480 seconds]
<tzimmermann>
javierm, ^ that 'unconditionally' is part of the problem here , I guess
bmodem1 has quit []
bmodem has joined #dri-devel
<javierm>
tzimmermann: yes, that's why I said that is a double free but what confuses me is that a) I can't reproduce it and b) drm_aperture_detach_firmware() is called after sysfb_disable() AFAICT
<javierm>
tzimmermann: but yeah, I guess that dropping sysfb_unregister() without fully understanding why doesn't happen in all cases makes sense anyways
<javierm>
since we should make the aperture core to handle everything
<tzimmermann>
i couldn't reproduce it on ast. but it happens on mgag200 :)
toolchains has quit [Remote host closed the connection]
<javierm>
tzimmermann: oh, interesting
toolchains has joined #dri-devel
<javierm>
tzimmermann: Ok, I'll post the patch then after lunch. Thanks for your help
<javierm>
tzimmermann: about vfio, yes I've in my queue to review that patch-set including your patch
<tzimmermann>
javierm, ast first calls sysfb_disable, which removes simpledrm. the later call to drm_aperture_detach_firmware does nothing
<tzimmermann>
mgag200 first runs drm_aperture_detach_firmware(), which removes simpledrm. it then calls sysfb_disable, which segfaults
bmodem has quit [Ping timeout: 480 seconds]
toolchains has quit [Ping timeout: 480 seconds]
vliaskov has joined #dri-devel
bmodem1 has quit [Ping timeout: 480 seconds]
toolchains has joined #dri-devel
ElementW has joined #dri-devel
toolchains has quit [Ping timeout: 480 seconds]
shankaru has quit [Quit: Leaving.]
<tzimmermann>
javierm, i sent a pathc to dri-devel
<tzimmermann>
works on my mgag200
Danct12 has quit [Remote host closed the connection]
<javierm>
tzimmermann: ah! that explains why I was confused because I was looking at the call stack that first called sysfb and then the drm_aperture_detach_firmware()
<javierm>
and assumed that order would be the same for pci and non-pci cases
toolchains has joined #dri-devel
pixelcluster has quit [Read error: Connection reset by peer]
gawin has joined #dri-devel
toolchains has quit [Ping timeout: 480 seconds]
<gawin>
anholt_: I'm bumpting up question: nir_cf_node_loop should only happen with dynamic loop or any not unrolled loop?
<gawin>
*bumping
pixelcluster has joined #dri-devel
frankbinns2 has joined #dri-devel
frankbinns1 has quit [Remote host closed the connection]
<lynxeye>
narmstrong: What is the rationale for preferring YUV422 over RGB or even YUV444 in the DW hdmi bus format negotiation?
g0b has joined #dri-devel
frankbinns1 has joined #dri-devel
frankbinns2 has quit [Ping timeout: 480 seconds]
nchery has joined #dri-devel
camus has quit [Ping timeout: 480 seconds]
alyssa has quit [coherence.oftc.net synthon.oftc.net]
bl4ckb0ne has quit [coherence.oftc.net synthon.oftc.net]
vyivel has quit [coherence.oftc.net synthon.oftc.net]
emersion has quit [coherence.oftc.net synthon.oftc.net]
anujp_ has quit [coherence.oftc.net synthon.oftc.net]
idr has quit [coherence.oftc.net synthon.oftc.net]
ifreund has quit [coherence.oftc.net synthon.oftc.net]
rcf has quit [coherence.oftc.net synthon.oftc.net]
rohiiyer has quit [coherence.oftc.net synthon.oftc.net]
soreau has quit [coherence.oftc.net synthon.oftc.net]
kem has quit [coherence.oftc.net synthon.oftc.net]
kchibisov has quit [coherence.oftc.net synthon.oftc.net]
aswar002 has quit [coherence.oftc.net synthon.oftc.net]
pzanoni has quit [coherence.oftc.net synthon.oftc.net]
lstrano has quit [coherence.oftc.net synthon.oftc.net]
Ryback_ has quit [coherence.oftc.net synthon.oftc.net]
jhli has quit [coherence.oftc.net synthon.oftc.net]
narmstrong has quit [coherence.oftc.net synthon.oftc.net]
jstultz has quit [coherence.oftc.net synthon.oftc.net]
Lightsword has quit [coherence.oftc.net synthon.oftc.net]
hfink has quit [coherence.oftc.net synthon.oftc.net]
radii has quit [coherence.oftc.net synthon.oftc.net]
nchery has quit [coherence.oftc.net synthon.oftc.net]
alanc has quit [coherence.oftc.net synthon.oftc.net]
sarnex has quit [coherence.oftc.net synthon.oftc.net]
oneforall2 has quit [coherence.oftc.net synthon.oftc.net]
anholt_ has quit [coherence.oftc.net synthon.oftc.net]
cheako has quit [coherence.oftc.net synthon.oftc.net]
agd5f has quit [coherence.oftc.net synthon.oftc.net]
Peuc_ has quit [coherence.oftc.net synthon.oftc.net]
kisak has quit [coherence.oftc.net synthon.oftc.net]
samueldr has quit [coherence.oftc.net synthon.oftc.net]
eukara has quit [coherence.oftc.net synthon.oftc.net]
flto has quit [coherence.oftc.net synthon.oftc.net]
jbarnes has quit [coherence.oftc.net synthon.oftc.net]
Lyude has quit [coherence.oftc.net synthon.oftc.net]
smaeul has quit [coherence.oftc.net synthon.oftc.net]
mattrope has quit [coherence.oftc.net synthon.oftc.net]
YuGiOhJCJ has quit [coherence.oftc.net synthon.oftc.net]
JohnnyonFlame has quit [coherence.oftc.net synthon.oftc.net]
sagar_ has quit [coherence.oftc.net synthon.oftc.net]
jekstrand has quit [coherence.oftc.net synthon.oftc.net]
dolphin has quit [coherence.oftc.net synthon.oftc.net]
ManMower has quit [coherence.oftc.net synthon.oftc.net]
jljusten has quit [coherence.oftc.net synthon.oftc.net]
cengiz_io has quit [coherence.oftc.net synthon.oftc.net]
zzag has quit [coherence.oftc.net synthon.oftc.net]
ZeZu has quit [coherence.oftc.net synthon.oftc.net]
everfree has quit [coherence.oftc.net synthon.oftc.net]
jolan has quit [coherence.oftc.net synthon.oftc.net]
clever has quit [coherence.oftc.net synthon.oftc.net]
flacks has quit [coherence.oftc.net synthon.oftc.net]
jrayhawk has quit [coherence.oftc.net synthon.oftc.net]
graphitemaster has quit [coherence.oftc.net synthon.oftc.net]
austriancoder has quit [coherence.oftc.net synthon.oftc.net]
isinyaaa has quit [coherence.oftc.net synthon.oftc.net]
andrey-konovalov has quit [coherence.oftc.net synthon.oftc.net]
Sachiel has quit [coherence.oftc.net synthon.oftc.net]
airlied has quit [coherence.oftc.net synthon.oftc.net]
Guest358 has quit [coherence.oftc.net synthon.oftc.net]
ajax has quit [coherence.oftc.net synthon.oftc.net]
sh-zam has quit [coherence.oftc.net synthon.oftc.net]
dschuermann_ has quit [coherence.oftc.net synthon.oftc.net]
robink_ has quit [coherence.oftc.net synthon.oftc.net]
mmenzyns_ has quit [coherence.oftc.net synthon.oftc.net]
jessica_24 has quit [coherence.oftc.net synthon.oftc.net]
zmike has quit [coherence.oftc.net synthon.oftc.net]
seanpaul__ has quit [coherence.oftc.net synthon.oftc.net]
CosmicPenguin has quit [coherence.oftc.net synthon.oftc.net]
MTCoster has quit [coherence.oftc.net synthon.oftc.net]
leo60228 has quit [coherence.oftc.net synthon.oftc.net]
jhugo_ has quit [coherence.oftc.net synthon.oftc.net]
Karyon has quit [coherence.oftc.net synthon.oftc.net]
abhinav__ has quit [coherence.oftc.net synthon.oftc.net]
tlwoerner has quit [coherence.oftc.net synthon.oftc.net]
siqueira has quit [coherence.oftc.net synthon.oftc.net]
mslusarz has quit [coherence.oftc.net synthon.oftc.net]
dri-logger has quit [coherence.oftc.net synthon.oftc.net]
exit70 has quit [coherence.oftc.net synthon.oftc.net]
OftenTimeConsuming has quit [coherence.oftc.net synthon.oftc.net]
anarsoul|2 has quit [coherence.oftc.net synthon.oftc.net]
Kayden has quit [coherence.oftc.net synthon.oftc.net]
rpigott has quit [coherence.oftc.net synthon.oftc.net]
demarchi has quit [coherence.oftc.net synthon.oftc.net]
enilflah has quit [coherence.oftc.net synthon.oftc.net]
LexSfX has quit [coherence.oftc.net synthon.oftc.net]
sumoon_ has quit [coherence.oftc.net synthon.oftc.net]
zf has quit [coherence.oftc.net synthon.oftc.net]
mattst88 has quit [coherence.oftc.net synthon.oftc.net]
krushia has quit [coherence.oftc.net synthon.oftc.net]
xyene has quit [coherence.oftc.net synthon.oftc.net]
mclasen has quit [coherence.oftc.net synthon.oftc.net]
orbea has quit [coherence.oftc.net synthon.oftc.net]
ogabbay has quit [coherence.oftc.net synthon.oftc.net]
unerlige4 has quit [coherence.oftc.net synthon.oftc.net]
rsripada has quit [coherence.oftc.net synthon.oftc.net]
mdnavare has quit [coherence.oftc.net synthon.oftc.net]
tchar_ has quit [coherence.oftc.net synthon.oftc.net]
reduz_ has quit [coherence.oftc.net synthon.oftc.net]
norris has quit [coherence.oftc.net synthon.oftc.net]
kathleen___ has quit [coherence.oftc.net synthon.oftc.net]
bcheng has quit [coherence.oftc.net synthon.oftc.net]
remexre has quit [coherence.oftc.net synthon.oftc.net]
sumits has quit [coherence.oftc.net synthon.oftc.net]
eletrotupi has quit [coherence.oftc.net synthon.oftc.net]
SolarAquarion has quit [coherence.oftc.net synthon.oftc.net]
rossy has quit [coherence.oftc.net synthon.oftc.net]
reductum_ has quit [coherence.oftc.net synthon.oftc.net]
steev has quit [coherence.oftc.net synthon.oftc.net]
arnd_ has quit [coherence.oftc.net synthon.oftc.net]
quantum5_ has quit [coherence.oftc.net synthon.oftc.net]
JTL has quit [coherence.oftc.net synthon.oftc.net]
cphealy has quit [coherence.oftc.net synthon.oftc.net]
mareko has quit [coherence.oftc.net synthon.oftc.net]
glisse has quit [coherence.oftc.net synthon.oftc.net]
lileo has quit [coherence.oftc.net synthon.oftc.net]
kurufu has quit [coherence.oftc.net synthon.oftc.net]
melissawen has quit [coherence.oftc.net synthon.oftc.net]
lemes has quit [coherence.oftc.net synthon.oftc.net]
xperia64 has quit [coherence.oftc.net synthon.oftc.net]
ernstp has quit [coherence.oftc.net synthon.oftc.net]
cmarcelo has quit [coherence.oftc.net synthon.oftc.net]
rodrigovivi has quit [coherence.oftc.net synthon.oftc.net]
krh has quit [coherence.oftc.net synthon.oftc.net]
angular_mike_____ has quit [coherence.oftc.net synthon.oftc.net]
rg3igalia has quit [coherence.oftc.net synthon.oftc.net]
tfiga has quit [coherence.oftc.net synthon.oftc.net]
rib___ has quit [coherence.oftc.net synthon.oftc.net]
benettig has quit [coherence.oftc.net synthon.oftc.net]
ezequielg has quit [coherence.oftc.net synthon.oftc.net]
hwentlan____ has quit [coherence.oftc.net synthon.oftc.net]
robher has quit [coherence.oftc.net synthon.oftc.net]
SanchayanMaity has quit [coherence.oftc.net synthon.oftc.net]
mmx_in_orbit has quit [coherence.oftc.net synthon.oftc.net]
Frogging101 has quit [coherence.oftc.net synthon.oftc.net]
sauce has quit [coherence.oftc.net synthon.oftc.net]
swivel has quit [coherence.oftc.net synthon.oftc.net]
Simonx22 has quit [coherence.oftc.net synthon.oftc.net]
dianders has quit [coherence.oftc.net synthon.oftc.net]
daniels has quit [coherence.oftc.net synthon.oftc.net]
gawin has quit [coherence.oftc.net synthon.oftc.net]
Company has quit [coherence.oftc.net synthon.oftc.net]
Haaninjo has quit [coherence.oftc.net synthon.oftc.net]
ella-0 has quit [coherence.oftc.net synthon.oftc.net]
rcn-ee_ has quit [coherence.oftc.net synthon.oftc.net]
eric_engestrom has quit [coherence.oftc.net synthon.oftc.net]
cwabbott has quit [coherence.oftc.net synthon.oftc.net]
robclark has quit [coherence.oftc.net synthon.oftc.net]
lemonzest has quit [coherence.oftc.net synthon.oftc.net]
famfo has quit [coherence.oftc.net synthon.oftc.net]
ced117 has quit [coherence.oftc.net synthon.oftc.net]
Akari has quit [coherence.oftc.net synthon.oftc.net]
egbert has quit [coherence.oftc.net synthon.oftc.net]
viciouss[m] has quit [coherence.oftc.net synthon.oftc.net]
mupuf has quit [coherence.oftc.net synthon.oftc.net]
lplc has quit [coherence.oftc.net synthon.oftc.net]
Ella[m] has quit [coherence.oftc.net synthon.oftc.net]
bbrezillon has quit [coherence.oftc.net synthon.oftc.net]
Thaodan has quit [coherence.oftc.net synthon.oftc.net]
RAOF has quit [coherence.oftc.net synthon.oftc.net]
cmeissl[m] has quit [coherence.oftc.net synthon.oftc.net]
aissen has quit [coherence.oftc.net synthon.oftc.net]
urja has quit [coherence.oftc.net synthon.oftc.net]
unevenrhombus[m] has quit [coherence.oftc.net synthon.oftc.net]
martijnbraam has quit [coherence.oftc.net synthon.oftc.net]
bluepenquin has quit [coherence.oftc.net synthon.oftc.net]
exit70[m] has quit [coherence.oftc.net synthon.oftc.net]
dv_ has quit [coherence.oftc.net synthon.oftc.net]
cyrozap has quit [coherence.oftc.net synthon.oftc.net]
Strit[m] has quit [coherence.oftc.net synthon.oftc.net]
tintou has quit [coherence.oftc.net synthon.oftc.net]
Terman_ has quit [coherence.oftc.net synthon.oftc.net]
pcercuei has quit [coherence.oftc.net synthon.oftc.net]
lynxeye has quit [coherence.oftc.net synthon.oftc.net]
danvet has quit [coherence.oftc.net synthon.oftc.net]
mvlad has quit [coherence.oftc.net synthon.oftc.net]
haasn has quit [coherence.oftc.net synthon.oftc.net]
rsalvaterra has quit [coherence.oftc.net synthon.oftc.net]
TMM has quit [coherence.oftc.net synthon.oftc.net]
minecrell has quit [coherence.oftc.net synthon.oftc.net]
jeeeun841 has quit [coherence.oftc.net synthon.oftc.net]
digetx has quit [coherence.oftc.net synthon.oftc.net]
jani has quit [coherence.oftc.net synthon.oftc.net]
alarumbe has quit [coherence.oftc.net synthon.oftc.net]
pendingchaos has quit [coherence.oftc.net synthon.oftc.net]
simon-perretta-img_ has quit [coherence.oftc.net synthon.oftc.net]
pochu has quit [coherence.oftc.net synthon.oftc.net]
vjaquez has quit [coherence.oftc.net synthon.oftc.net]
hakzsam has quit [coherence.oftc.net synthon.oftc.net]
Koniiiik has quit [coherence.oftc.net synthon.oftc.net]
Surkow|laptop has quit [coherence.oftc.net synthon.oftc.net]
jfalempe has quit [coherence.oftc.net synthon.oftc.net]
gpiccoli has quit [coherence.oftc.net synthon.oftc.net]
gio has quit [coherence.oftc.net synthon.oftc.net]
ElementW has quit [coherence.oftc.net synthon.oftc.net]
DragoonAethis has quit [coherence.oftc.net synthon.oftc.net]
imre has quit [coherence.oftc.net synthon.oftc.net]
pepp has quit [coherence.oftc.net synthon.oftc.net]
adavy has quit [coherence.oftc.net synthon.oftc.net]
kusma has quit [coherence.oftc.net synthon.oftc.net]
mighty17 has quit [coherence.oftc.net synthon.oftc.net]
nielsdg has quit [coherence.oftc.net synthon.oftc.net]
JosExpsito[m] has quit [coherence.oftc.net synthon.oftc.net]
jannau has quit [coherence.oftc.net synthon.oftc.net]
milek7_ has quit [coherence.oftc.net synthon.oftc.net]
alatiera has quit [coherence.oftc.net synthon.oftc.net]
DPA- has quit [coherence.oftc.net synthon.oftc.net]
hikiko has quit [coherence.oftc.net synthon.oftc.net]
tango_ has quit [coherence.oftc.net synthon.oftc.net]
libv has quit [coherence.oftc.net synthon.oftc.net]
qyliss has quit [coherence.oftc.net synthon.oftc.net]
CounterPillow has quit [coherence.oftc.net synthon.oftc.net]
CATS has quit [coherence.oftc.net synthon.oftc.net]
q66 has quit [coherence.oftc.net synthon.oftc.net]
CME has quit [coherence.oftc.net synthon.oftc.net]
bgs has quit [coherence.oftc.net synthon.oftc.net]
nanonyme has quit [coherence.oftc.net synthon.oftc.net]
sven has quit [coherence.oftc.net synthon.oftc.net]
APic has quit [coherence.oftc.net synthon.oftc.net]
marcan has quit [coherence.oftc.net synthon.oftc.net]
gallo2 has quit [coherence.oftc.net synthon.oftc.net]
enunes has quit [coherence.oftc.net synthon.oftc.net]
calebccff has quit [coherence.oftc.net synthon.oftc.net]
V has quit [coherence.oftc.net synthon.oftc.net]
jasuarez has quit [coherence.oftc.net synthon.oftc.net]
egalli has quit [coherence.oftc.net synthon.oftc.net]
kallisti5[m] has quit [coherence.oftc.net synthon.oftc.net]
cwfitzgerald[m] has quit [coherence.oftc.net synthon.oftc.net]
macc24 has quit [coherence.oftc.net synthon.oftc.net]
doras has quit [coherence.oftc.net synthon.oftc.net]
mairacanal[m] has quit [coherence.oftc.net synthon.oftc.net]
feaneron has quit [coherence.oftc.net synthon.oftc.net]
r[m] has quit [coherence.oftc.net synthon.oftc.net]
gagallo7[m] has quit [coherence.oftc.net synthon.oftc.net]
ambasta[m] has quit [coherence.oftc.net synthon.oftc.net]
ella-0[m] has quit [coherence.oftc.net synthon.oftc.net]
Mershl[m] has quit [coherence.oftc.net synthon.oftc.net]
enick_254 has quit [coherence.oftc.net synthon.oftc.net]
gdevi has quit [coherence.oftc.net synthon.oftc.net]
cleverca22[m] has quit [coherence.oftc.net synthon.oftc.net]
jenatali has quit [coherence.oftc.net synthon.oftc.net]
dcbaker has quit [coherence.oftc.net synthon.oftc.net]
Tooniis[m] has quit [coherence.oftc.net synthon.oftc.net]
turol has quit [coherence.oftc.net synthon.oftc.net]
tonyk has quit [coherence.oftc.net synthon.oftc.net]
tales-aparecida has quit [coherence.oftc.net synthon.oftc.net]
swick has quit [coherence.oftc.net synthon.oftc.net]
leandrohrb has quit [coherence.oftc.net synthon.oftc.net]
evadot has quit [coherence.oftc.net synthon.oftc.net]
padovan has quit [coherence.oftc.net synthon.oftc.net]
Stary has quit [coherence.oftc.net synthon.oftc.net]
rawoul has quit [coherence.oftc.net synthon.oftc.net]
Venemo has quit [coherence.oftc.net synthon.oftc.net]
ickle has quit [coherence.oftc.net synthon.oftc.net]
HdkR has quit [coherence.oftc.net synthon.oftc.net]
tagr has quit [coherence.oftc.net synthon.oftc.net]
pH5 has quit [coherence.oftc.net synthon.oftc.net]
BobBeck has quit [coherence.oftc.net synthon.oftc.net]
mlankhorst has quit [coherence.oftc.net synthon.oftc.net]
jcristau has quit [coherence.oftc.net synthon.oftc.net]
mal has quit [coherence.oftc.net synthon.oftc.net]
tomeu has quit [coherence.oftc.net synthon.oftc.net]
Plagman has quit [coherence.oftc.net synthon.oftc.net]
frankbinns1 has quit [coherence.oftc.net synthon.oftc.net]
MajorBiscuit has quit [coherence.oftc.net synthon.oftc.net]
vliaskov has quit [coherence.oftc.net synthon.oftc.net]
rkanwal has quit [coherence.oftc.net synthon.oftc.net]
ppascher has quit [coherence.oftc.net synthon.oftc.net]
columbarius has quit [coherence.oftc.net synthon.oftc.net]
Lucretia-backup has quit [coherence.oftc.net synthon.oftc.net]
tanty has quit [coherence.oftc.net synthon.oftc.net]
MrCooper has quit [coherence.oftc.net synthon.oftc.net]
pinchartl has quit [coherence.oftc.net synthon.oftc.net]
zamundaaa has quit [coherence.oftc.net synthon.oftc.net]
romangg has quit [coherence.oftc.net synthon.oftc.net]
akselmo has quit [coherence.oftc.net synthon.oftc.net]
ivyl has quit [coherence.oftc.net synthon.oftc.net]
ana has quit [coherence.oftc.net synthon.oftc.net]
kbingham has quit [coherence.oftc.net synthon.oftc.net]
haagch has quit [coherence.oftc.net synthon.oftc.net]
illwieckz has quit [coherence.oftc.net synthon.oftc.net]
colemickens has quit [coherence.oftc.net synthon.oftc.net]
unrelentingtech has quit [coherence.oftc.net synthon.oftc.net]
mmind00 has quit [coherence.oftc.net synthon.oftc.net]
pjakobsson has quit [coherence.oftc.net synthon.oftc.net]
Emantor has quit [coherence.oftc.net synthon.oftc.net]
Sumera[m] has quit [coherence.oftc.net synthon.oftc.net]
znullptr[m] has quit [coherence.oftc.net synthon.oftc.net]
gnustomp[m] has quit [coherence.oftc.net synthon.oftc.net]
loki_val has quit [coherence.oftc.net synthon.oftc.net]
xantoz has quit [coherence.oftc.net synthon.oftc.net]
mwk has quit [coherence.oftc.net synthon.oftc.net]
jernej has quit [coherence.oftc.net synthon.oftc.net]
LaserEyess has quit [coherence.oftc.net synthon.oftc.net]
dliviu has quit [coherence.oftc.net synthon.oftc.net]
dwlsalmeida has quit [coherence.oftc.net synthon.oftc.net]
hch12907 has quit [coherence.oftc.net synthon.oftc.net]
robertfoss[m] has quit [coherence.oftc.net synthon.oftc.net]
shoragan has quit [coherence.oftc.net synthon.oftc.net]
neobrain[m] has quit [coherence.oftc.net synthon.oftc.net]
jekstrand[m] has quit [coherence.oftc.net synthon.oftc.net]
Andy[m] has quit [coherence.oftc.net synthon.oftc.net]
zamundaaa[m] has quit [coherence.oftc.net synthon.oftc.net]
T_UNIX has quit [coherence.oftc.net synthon.oftc.net]
chema has quit [coherence.oftc.net synthon.oftc.net]
masush5[m] has quit [coherence.oftc.net synthon.oftc.net]
dhanuka[m] has quit [coherence.oftc.net synthon.oftc.net]
ramacassis[m] has quit [coherence.oftc.net synthon.oftc.net]
YaLTeR[m] has quit [coherence.oftc.net synthon.oftc.net]
Dylanger has quit [coherence.oftc.net synthon.oftc.net]
ralf1307[theythem][m] has quit [coherence.oftc.net synthon.oftc.net]
LaughingMan[m] has quit [coherence.oftc.net synthon.oftc.net]
MatrixTravelerbot[m] has quit [coherence.oftc.net synthon.oftc.net]
heftig has quit [coherence.oftc.net synthon.oftc.net]
Mis012[m] has quit [coherence.oftc.net synthon.oftc.net]
hasebastian[m] has quit [coherence.oftc.net synthon.oftc.net]
halfline[m] has quit [coherence.oftc.net synthon.oftc.net]
arisu has quit [coherence.oftc.net synthon.oftc.net]
moben[m] has quit [coherence.oftc.net synthon.oftc.net]
x512[m] has quit [coherence.oftc.net synthon.oftc.net]
DrNick has quit [coherence.oftc.net synthon.oftc.net]
mripard has quit [coherence.oftc.net synthon.oftc.net]
probablymoony has quit [coherence.oftc.net synthon.oftc.net]
Ristovski has quit [coherence.oftc.net synthon.oftc.net]
bkrause has quit [coherence.oftc.net synthon.oftc.net]
mstoeckl has quit [coherence.oftc.net synthon.oftc.net]
ccr has quit [coherence.oftc.net synthon.oftc.net]
pq has quit [coherence.oftc.net synthon.oftc.net]
lcn has quit [coherence.oftc.net synthon.oftc.net]
neobrain has quit [coherence.oftc.net synthon.oftc.net]
degasus has quit [coherence.oftc.net synthon.oftc.net]
rellla has quit [coherence.oftc.net synthon.oftc.net]
robertfoss has quit [coherence.oftc.net synthon.oftc.net]
italove has quit [coherence.oftc.net synthon.oftc.net]
Ziemas has quit [coherence.oftc.net synthon.oftc.net]
tjaalton has quit [coherence.oftc.net synthon.oftc.net]
JoshuaAshton has quit [coherence.oftc.net synthon.oftc.net]
pixelcluster has quit [coherence.oftc.net synthon.oftc.net]
tursulin has quit [coherence.oftc.net synthon.oftc.net]
rasterman has quit [coherence.oftc.net synthon.oftc.net]
tzimmermann has quit [coherence.oftc.net synthon.oftc.net]
mszyprow has quit [coherence.oftc.net synthon.oftc.net]
g0b has quit [coherence.oftc.net synthon.oftc.net]
ds` has quit [coherence.oftc.net synthon.oftc.net]
RSpliet has quit [coherence.oftc.net synthon.oftc.net]
jkhsjdhjs has quit [coherence.oftc.net synthon.oftc.net]
angerctl has quit [coherence.oftc.net synthon.oftc.net]
i-garrison has quit [coherence.oftc.net synthon.oftc.net]
neonking has quit [coherence.oftc.net synthon.oftc.net]
jkqxz has quit [coherence.oftc.net synthon.oftc.net]
mceier has quit [coherence.oftc.net synthon.oftc.net]
mriesch has quit [coherence.oftc.net synthon.oftc.net]
sravn has quit [coherence.oftc.net synthon.oftc.net]
sul has quit [coherence.oftc.net synthon.oftc.net]
shadeslayer has quit [coherence.oftc.net synthon.oftc.net]
vup has quit [coherence.oftc.net synthon.oftc.net]
karolherbst has quit [coherence.oftc.net synthon.oftc.net]
gruetzkopf has quit [coherence.oftc.net synthon.oftc.net]
dos1 has quit [coherence.oftc.net synthon.oftc.net]
Arsen has quit [coherence.oftc.net synthon.oftc.net]
FLHerne has quit [coherence.oftc.net synthon.oftc.net]
marex has quit [coherence.oftc.net synthon.oftc.net]
Thymo has quit [coherence.oftc.net synthon.oftc.net]
Newbyte has quit [coherence.oftc.net synthon.oftc.net]
chivay has quit [coherence.oftc.net synthon.oftc.net]
Lynne has quit [coherence.oftc.net synthon.oftc.net]
dreda_ has quit [coherence.oftc.net synthon.oftc.net]
sigmaris has quit [coherence.oftc.net synthon.oftc.net]
yoslin has quit [coherence.oftc.net synthon.oftc.net]
lanodan has quit [coherence.oftc.net synthon.oftc.net]
bnieuwenhuizen has quit [coherence.oftc.net synthon.oftc.net]
iokill has quit [coherence.oftc.net synthon.oftc.net]
glennk has quit [coherence.oftc.net synthon.oftc.net]
dj-death has quit [coherence.oftc.net synthon.oftc.net]
sigmoidfunc[m] has quit [coherence.oftc.net synthon.oftc.net]
reactormonk[m] has quit [coherence.oftc.net synthon.oftc.net]
tomba has quit [coherence.oftc.net synthon.oftc.net]
undvasistas[m] has quit [coherence.oftc.net synthon.oftc.net]
pushqrdx[m] has quit [coherence.oftc.net synthon.oftc.net]
naheemsays[m] has quit [coherence.oftc.net synthon.oftc.net]
go4godvin has quit [coherence.oftc.net synthon.oftc.net]
DavidHeidelberg[m] has quit [coherence.oftc.net synthon.oftc.net]
onox[m] has quit [coherence.oftc.net synthon.oftc.net]
Anson[m] has quit [coherence.oftc.net synthon.oftc.net]
Vin[m] has quit [coherence.oftc.net synthon.oftc.net]
PiGLDN[m] has quit [coherence.oftc.net synthon.oftc.net]
yshui` has quit [coherence.oftc.net synthon.oftc.net]
bylaws has quit [coherence.oftc.net synthon.oftc.net]
danylo has quit [coherence.oftc.net synthon.oftc.net]
vsyrjala has quit [coherence.oftc.net synthon.oftc.net]
glehmann has quit [coherence.oftc.net synthon.oftc.net]
mntmn has quit [coherence.oftc.net synthon.oftc.net]
linkmauve has quit [coherence.oftc.net synthon.oftc.net]
javierm has quit [coherence.oftc.net synthon.oftc.net]
Adrinael has quit [coherence.oftc.net synthon.oftc.net]
Prf_Jakob has quit [coherence.oftc.net synthon.oftc.net]
kgz has quit [coherence.oftc.net synthon.oftc.net]
jadahl has quit [coherence.oftc.net synthon.oftc.net]
sumits has joined #dri-devel
shoragan has joined #dri-devel
Strit[m] has joined #dri-devel
kallisti5[m] has joined #dri-devel
Andy[m] has joined #dri-devel
zamundaaa[m] has joined #dri-devel
BobBeck has joined #dri-devel
neobrain has joined #dri-devel
<g0b>
v3dv/broadcom people, why is only the VK_SUBGROUP_FEATURE_BASIC_BIT supported ? HW limitation (i don't get how that would work though?), or just a feature that's not been done yet ?
saurabhg has joined #dri-devel
<narmstrong>
lynxeye: HDMI spec
<narmstrong>
content type should be implemented and selection based on that, YUV is best for video content
pcercuei has joined #dri-devel
toolchains has joined #dri-devel
<lynxeye>
narmstrong: Yea, without a indication that the content going over the link is really video, it seems a bit strange to prefer a subsampled format.
toolchains has quit [Ping timeout: 480 seconds]
frankbinns has joined #dri-devel
iive has joined #dri-devel
MajorBiscuit has joined #dri-devel
toolchains has joined #dri-devel
<jekstrand>
jenatali, kusma: RE the driverID for dozen. Do we want it to be MESA or MICROSOFT? I don't personally care, but it seems like something someone should think about and make an intentional decision.
<jenatali>
jekstrand: Good question. Lemme check if my management cares
<jekstrand>
Question is which they value more: credit or plausible deniability. :D
<kusma>
jekstrand: Good point. I don't have a strong opinion here... My feeling is that this is an "upstream project", so MESA is fine, but 🤷
<jenatali>
Yeah, my default is that what's proposed is the right thing here, it is a Mesa driver, just a layered one that happens to be funded by one company rather than a community effort
<jekstrand>
Makes sense. ANV is a weird one because it's a Mesa driver but is definitely VK_DRIVER_ID_INTEL_*
toolchains has quit [Ping timeout: 480 seconds]
<jekstrand>
And that was an intentional decision.
<jekstrand>
We didn't want developers seeing VK_MESA and thinking it wasn't a production-grade driver.
<jenatali>
Yeah makes sense
<jekstrand>
And it was (and still is) the only HW-vendor-funded Mesa Vulkan driver.
<jekstrand>
So it's a bit weird like that.
saurabhg has quit [Ping timeout: 480 seconds]
<jenatali>
Though that's an unfortunate reason to make that decision, I wouldn't expect that connotation
<jekstrand>
You're also deep in the Mesa code and you know it's solid.
nchery has joined #dri-devel
<jekstrand>
There's a lot of Mesa FUD kicking around.
<jekstrand>
I think RADV has gone a very long way towards cleaning it up, which is really good.
<jenatali>
Interesting. Guess I just don't run in those circles
<jekstrand>
But, at the time, it wasn't a crazy line of thinking.
<jenatali>
When embarking on our GL/CL effort ~3 years ago, in my mind it was the only viable approach to have anything remotely quality in any reasonable timeframe
<jekstrand>
Also, all that FUD is really amplified at Intel internally so I may have a skewed view.
<jenatali>
Anyway, if management comes back and says we want MICROSOFT I'll probably create a counter-proposal PR so it's coming from a Microsoft person rather than a Mesa person, but I expect they either won't care or will want it this way
toolchains has joined #dri-devel
toolchains has quit [Read error: No route to host]
toolchains has joined #dri-devel
<jekstrand>
jenatali: Sounds good
toolchains has quit [Ping timeout: 480 seconds]
<alyssa>
15:28 < jekstrand> And it was (and still is) the only HW-vendor-funded Mesa Vulkan driver.
srslypascal has joined #dri-devel
<alyssa>
jekstrand: IMG, soon, hopefully?
<jekstrand>
alyssa: Hopefully!
<zmike>
yeah otherwise what did I even do all this tiler optimizing for
<alyssa>
jekstrand: :D
<jekstrand>
zmike: funzies
<zmike>
I'm not familiar with the concept
Duke`` has joined #dri-devel
<alyssa>
jekstrand: this might be a dumb question but is it possible to generate pack_uvec4_to_uint ops
<jekstrand>
alyssa: What do you mean by "generate"?
<alyssa>
oh, only from the pack_*norm_4x8 lowering. i see
<alyssa>
well, I don't see glsl-to-nir or vtn generating it
<alyssa>
glsl lowers it unconditionally to bitshifts
<jekstrand>
alyssa: That's believable.
<jenatali>
jekstrand: Looks like the answer is "don't care" so MESA it is
<jekstrand>
jenatali: :+1:
<alyssa>
jekstrand: on one hand, bifrost can implement pack_uvec4_to_uint in one instruction, so it's a shame to have such a massive pile of bitshifts
<alyssa>
(two instructions on valhall)
<alyssa>
on the other hand, I don't have a single shader in my shader-db that uses any pack_u* or pack_s* function, so I really don't care :p
<jekstrand>
Oh the other hand, bifrost?
<jekstrand>
alyssa: Yeah, I think that's why no one's bothered to do the plumbing. If we suddenly find something that's pack_*norm bound, we can fix it then.
<alyssa>
yeah
<alyssa>
that instruction is more useful for OpenCL
<alyssa>
but there NIR would emit an 8-bit vec4 for that instruction
<alyssa>
(8-bit nir_op_vec4, I mean)
<jekstrand>
We'd probably do something rubbish for the pack, though.
<jekstrand>
One day, I want to make pack/unpack suck less.
devilhorns has joined #dri-devel
<alyssa>
ok, I think this is all the fallout from enabling arb_gpu_shader5 (+ other shader caps) on valhall
gouchi has joined #dri-devel
mbrost has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
fcarrijo has joined #dri-devel
abws has joined #dri-devel
Duke`` has joined #dri-devel
fcarrijo has quit []
<zmike>
dcbaker: you may want to pick !16977 to stable just to cut down on ci noise/runtimes
mbrost has quit [Read error: Connection reset by peer]
<zmike>
if yes then I'll take care of getting it submitted+merged
frankbinns has quit [Remote host closed the connection]
alyssa has left #dri-devel [#dri-devel]
MajorBiscuit has quit [Ping timeout: 480 seconds]
<karolherbst>
zmike: would have to think about it more... is it just the 1.0 / 1.625 case?
<zmike>
unsure, danylo was the one who was debugging it
<karolherbst>
well.. normally you'd have to check if the result fits within glsl requierements
<karolherbst>
so the issue that imirkin and I tried to fix was that if the hw doesn't behave the same on each channel
<karolherbst>
but I don't think we worked on relaxing the precision reqs at all
<karolherbst>
if the result is simply not precise enough that's a different problem than if the test is wrong
<danylo>
karolherbst: fmod(14.625, 1.625) - that's what failed in the cts test
<karolherbst>
ahh
<danylo>
On Turnip the result is 1.625 instead of expected 0.0
<karolherbst>
mhhh
<karolherbst>
how is / implemented on turnip?
<karolherbst>
because that's pretty much 0.0 no matter how you look at it
<karolherbst>
14.625 / 1.625 is 9.0
<karolherbst>
and floor 9.0 is 9
<danylo>
For turnip zink lowers fp64
<karolherbst>
sure
<karolherbst>
but if you'd do that in fp32 space, you'd get 0.0
<cwabbott>
no, not necessarily
<karolherbst>
so I'd say the fp64 lowering is a bit inprecise?
<cwabbott>
the spec allows some wiggle room for / precision
<karolherbst>
yeah.... I guess that's true
<cwabbott>
but I'm surprised we're hitting this when other soft-fp64 users aren't
<cwabbott>
I think that's just intel?
<karolherbst>
but still
<karolherbst>
if it's a good enough soft fp64, then it should be 0.0
<karolherbst>
those aren't odd or edge case numbers
<danylo>
During the fmod calculations we do "double(1.0) / double(1.625)" which on Turnip gives 0.6153846153846153 (3fe3b13b13b13b13) and on RADV gives 0.6153846153846154 (3fe3b13b13b13b14)
<danylo>
That tiny difference is enough to tip the scales
<karolherbst>
yeah
<karolherbst>
so if the floor becomes < 9.0 it returns 8
<danylo>
E.g if we do one more Newton-Raphson step - we get expected result
<karolherbst>
yeah... so this all depends on the division to return a precise value for numbers you can expect it to return such
<danylo>
And the core difference is in frcp which in Turnip translates into HW instruction frcp
<danylo>
s/frcp/rcp/
<karolherbst>
but I'd agree, that returning 1.625 would be valid
<danylo>
but on RADV it gives a bit imprecise result
nsneck has joined #dri-devel
<karolherbst>
yeah soo.. we use RCP as well
<cwabbott>
danylo: really, I thought that double rcp was translated to the softfp64 version
<cwabbott>
which I believe uses just integers
<karolherbst>
but not sure if the patch to CTS is correct
<lina>
That's a program written in D, with GL bindings I had to hack EGL support into, running on EGL-surfaceless with a pbuffer, on Mesa, with alyssa's AGX driver, patched to add a DRM backend with a UABI I just made up for a nonexistent kernel driver, on top of drm-shim, with a backend that embeds a Python interpreter, which imports a prototype "kernel-side" driver Python script (that runs in userspace), which
<lina>
reads BOs out of the drm-shim memfd and uploads them to the GPU along with sending commands to the firmware and managing memory, all driving a binary proxy protocol over a USB serial TTY, which is a USB device exposed by a proxy shim running on bare metal on an M1 Mac Mini, which the Python side then instructs to blit the right render target to the HDMI framebuffer when rendering is complete.
* ccr
feels scared :P
<lina>
Hey, glmark2 even runs at 15FPS or so on this!
<lina>
Oh yeah, the USB device is USB2.0 (which is why it's so hilariously slow)
<karolherbst>
lina: hi there (didn't know you were here with us all along!)
ngcortes has joined #dri-devel
<lina>
karolherbst: I idle around ^^
<karolherbst>
cool
nchery has quit [Ping timeout: 480 seconds]
vyivel has joined #dri-devel
ella-0_ has joined #dri-devel
gawin has joined #dri-devel
mbrost_ has joined #dri-devel
mbrost has quit [Remote host closed the connection]
mclasen_ has joined #dri-devel
ngcortes has quit [Ping timeout: 480 seconds]
Duke`` has quit [Ping timeout: 480 seconds]
heat has joined #dri-devel
mclasen_ has quit [Ping timeout: 480 seconds]
lemonzest1 has joined #dri-devel
Duke`` has joined #dri-devel
ngcortes has joined #dri-devel
nchery has joined #dri-devel
DottorLeo has joined #dri-devel
<DottorLeo>
hi!
<DottorLeo>
What is the time zone for Gert Wollny? :)
<anholt_>
europe, I'm pretty sure.
eukara has joined #dri-devel
srslypascal has quit [Quit: Leaving]
sul has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
alyssa has joined #dri-devel
<alyssa>
dcbaker: eric_engestrom: For a pair of patches of the form "small refactor to make the next patch possible / patch that needs backporting", would you prefer those be squashed into a single commit to backport (as opposed to Cc stable on the firts and Fixes on the second)?
<alyssa>
For stuff that doesn't backport, I try to change only 1 component per commit (where reasonable), so those would be split out
<alyssa>
But I don't know if backport work scales with # of patches (independent of # of lines)
<dcbaker>
alyssa: I really don't care either way, I have a script that does the work, so the CC: is easy with me, and is probably less work for you
DottorLeo has quit [Quit: Leaving]
<alyssa>
ack
<alyssa>
no less work either way -- squashing is trivial
lemonzest1 has quit []
<alyssa>
anholt_: v3d has the comment:
<alyssa>
/* The blitter util dereferences the color regardless, even though the
<alyssa>
* gallium clear API may not pass one in when only Z/S are cleared.
<alyssa>
Any idea if this is still the case? I don't see u_blitter doing this but I don't have any v3d hardware to test if dropping the workaround would break the world.
<alyssa>
I guess I can use CI driven development ....
nchery has quit [Ping timeout: 480 seconds]
<anholt_>
yeah, dunno any more
<alyssa>
will give it a try I guess
<alyssa>
dates back to 2018, wild times.. :p
<alyssa>
wait I was already writing Mesa code then. hmph.
mszyprow has joined #dri-devel
<ccr>
time flies by, eh
mszyprow has quit [Ping timeout: 480 seconds]
<anarsoul>
alyssa: pre-panfrost times? :)
<alyssa>
ccr: not really, it feels like a lifetime ago
<alyssa>
anarsoul: I started work on the midgard compiler in Jan 2018 \s/
<anarsoul>
OK, early panfrost times
<alyssa>
yeah..
<ccr>
ice age!
<ccr>
pan .. frost .. see .. /me takes his coat
mszyprow has joined #dri-devel
srslypascal has joined #dri-devel
gouchi has quit [Remote host closed the connection]
mclasen has joined #dri-devel
heat has quit [Remote host closed the connection]
heat has joined #dri-devel
abws has quit [Ping timeout: 480 seconds]
<alyssa>
why the heck is ERROR - Test dEQP-GLES31.functional.compute.shared_var.work_group_size.float_64_1_1: Fail: See "output/c61.r1.caselist.txt"
<alyssa>
er
<alyssa>
why the heck is ERROR - Test dEQP-GLES31.functional.compute.shared_var.work_group_size.float_64_1_1: Fail: See "output/c61.r1.caselist.txt"
<alyssa>
ok I just can't use copypaste apparently
<alyssa>
why the heck is that test flaking when run in deqp-runner for the whole gles31, but not for a subset >_>
<alyssa>
ok. managed to get it to fail on the own with deqp-runner in the background. I guess that supports the race condition hypothesis, where we "almost" always win the race