ChanServ changed the topic of #dri-devel to: <ajax> nothing involved with X should ever be unable to find a bar
<alyssa>
nothing quite like a kernel hang so bad the UART stops before you get any debug info
<danvet>
alyssa, yeah I wonder whether we should guarantee that the free job callback runs in process ctx
<danvet>
in drm/sched I mean
<danvet>
also the patch that added the irq_work shouldn't have regressed, since this could be called from fence put which is in an irqsave spinlock or something like that
<danvet>
well when you kill a process that is, maybe just not something that gets much testing
<danvet>
plus also needs lockdep enabled or it wont splat
<danvet>
but yeah that all means you'll now always run in interrupt-like context
<alyssa>
danvet: easily willing to believe panfrost was broken before too :v
<danvet>
alyssa, maybe just reply to that patch on dri-devel asking what should be done
<alyssa>
and yes this bug is "difficult" to trigger
<danvet>
since it's way too late to trust my brain over here to come up with good ideas :-P
<alyssa>
(I think it requires multiple processes crashing and faulting at the same time plus some bad luck with a race condition or two...)
<danvet>
yeah refcounting is fun that way
<alyssa>
which, I mean, that's perfect for running 8 instances of the conformance tests in parallel on a completely broken mesa
<danvet>
I think really good practice is that you put a might_lock() into any _put() with the worst lock this can take
<danvet>
and then enable lockdep for testing
<alyssa>
hmm?
<alyssa>
(update: reverted lockdep splat, now I get a different panic, yee)
<danvet>
well usually the final put is called from somewhere benign
<alyssa>
I will say the deqp run lasted much longer so that's probably a new, even rarer issue...?
<danvet>
so to make sure you're always unlucky, add might_lock() for the nasty locks, and then lockdep is much better informed about your dependency graph and can catch issues
<alyssa>
okay..
<danvet>
that ooks like a fun backtrace
<danvet>
"fun"
ngcortes has joined #dri-devel
ngcortes has quit [Remote host closed the connection]
<danvet>
sometimes what's needed is an additional lockdep key to connect the pieces
ngcortes has joined #dri-devel
<danvet>
like anytime we call job_put add a might_lock(job_put_lockdep_key)
<danvet>
and then also grab that around the actual free_job callback
<danvet>
that way anyway the driver might ever grab gets pulled into any potential final free callsite
<danvet>
repeat with any other _put() you have
<alyssa>
that one is a null deref
<danvet>
it's a really nasty exercise to catch locking design pain
<danvet>
hm I guess not enough arm asm knowledge here
<danvet>
meaning, zero :-)
<alyssa>
hmm?
<alyssa>
mmu_as_get(..) is getting called with a NULL argument
<alyssa>
losing some race
<alyssa>
i don't think i have the brain cycles to parse that tonight
pcercuei has quit [Quit: dodo]
<alyssa>
It's admittedly unclear to me how this pointer can ever end up null
iive has quit []
<alyssa>
brb rewriting panfrost.ko in rust
tursulin has quit [Remote host closed the connection]
<airlied>
alyssa: racey mcracey
<robher>
alyssa: spinlocks can be used in irq, but the irqsave version has to be used in thread ctx.
mbrost has joined #dri-devel
danvet has quit [Ping timeout: 480 seconds]
<alyssa>
i know wht some of those words are
<alyssa>
um
<alyssa>
that's my cue to step away from the computer and call it a night
<ajax>
alyssa: "panfrust" was right there, come on
macromorgan has quit [Read error: Connection reset by peer]
macromorgan has joined #dri-devel
<imirkin>
jenatali: fyi, ARB_sample_shading is in no way required for ES 3.1. ARB_gpu_shader5 requires more than what's needed for ES 3.1, so we "work around" that by checking that iamges/etc are supported but not necessarily that ext.
<imirkin>
[just looking at your comment in the MR to add those exts for d3d12]
<jenatali>
Yeah I saw. You can claim ESSL3.1 instead of GLSL4.0 but oh well
macromorgan is now known as Guest1217
Guest1217 has quit [Read error: Connection reset by peer]
<imirkin>
jenatali: you should still claim ESSL 310 - iirc it enables some stuff, i forget exactly what
<imirkin>
or maybe it was 320 which enables stuff? i forget
macromorgan has joined #dri-devel
<jenatali>
Looked like GLSL400 was a superset when I looked
<jenatali>
I'm seeing GLES3.1 getting reported with that series
<imirkin>
they're separate pipe_cap's
<imirkin>
best to report it "correctly"
<imirkin>
even if it "works" the other way
cphealy_ has joined #dri-devel
bl4ckb0ne has quit [Remote host closed the connection]
ifreund has quit [Remote host closed the connection]
ddevault has quit [Remote host closed the connection]
dnkl has quit [Remote host closed the connection]
bl4ckb0ne has joined #dri-devel
<jenatali>
Yeah fair, I'll add it to that patch
dnkl has joined #dri-devel
ifreund has joined #dri-devel
ddevault has joined #dri-devel
<jenatali>
Thanks for looking
pnowack has quit [Quit: pnowack]
<alyssa>
jenatali: ARB_sample_shading is the one I keep forgetting is ES3.2 and not ES3.1
bbrezillon has quit [Ping timeout: 480 seconds]
<alyssa>
and accidentally nerdsniped myself into implementing on panfrost
mripard has quit [Ping timeout: 480 seconds]
<jenatali>
Yeah when I implemented it for GL4 I had a test failures from not supporting GPU_shader5 and tests that missed checking for it
<jenatali>
So I went ahead and implemented both lol
cphealy has quit [Ping timeout: 480 seconds]
<imirkin>
you'll need it eventually anyways
<jenatali>
Yep
ngcortes_ has quit []
<airlied>
ajax: not sure, just another area of 8k to check into it
oneforall2 has quit [Remote host closed the connection]
<ishitatsuyuki>
any guides on using asan with radv? what LD_PRELOADs do I need to set when running apps?
<airlied>
just the asan one?
tarceri has quit [Server closed connection]
tarceri has joined #dri-devel
<ishitatsuyuki>
ok, thanks
oneforall2 has joined #dri-devel
cef has quit [Server closed connection]
cef has joined #dri-devel
hch12907 has quit [Server closed connection]
hch12907 has joined #dri-devel
mwalle has quit [Quit: WeeChat 3.0]
dviola has joined #dri-devel
ramaling_ has quit [nucleus.oftc.net coherence.oftc.net]
radii has quit [nucleus.oftc.net coherence.oftc.net]
sdutt has quit [nucleus.oftc.net coherence.oftc.net]
Kayden has quit [nucleus.oftc.net coherence.oftc.net]
flacks has quit [nucleus.oftc.net coherence.oftc.net]
tlwoerner has quit [nucleus.oftc.net coherence.oftc.net]
ezequielg has quit [nucleus.oftc.net coherence.oftc.net]
sagar__ has quit [nucleus.oftc.net coherence.oftc.net]
ishitatsuyuki has quit [nucleus.oftc.net coherence.oftc.net]
xyene has quit [nucleus.oftc.net coherence.oftc.net]
unerlige has quit [nucleus.oftc.net coherence.oftc.net]
Ryback_ has quit [nucleus.oftc.net coherence.oftc.net]
samueldr has quit [nucleus.oftc.net coherence.oftc.net]
bluebugs has quit [nucleus.oftc.net coherence.oftc.net]
sneil has quit [nucleus.oftc.net coherence.oftc.net]
ivyl has quit [nucleus.oftc.net coherence.oftc.net]
rcf has quit [nucleus.oftc.net coherence.oftc.net]
dviola has quit [nucleus.oftc.net coherence.oftc.net]
LexSfX has quit [nucleus.oftc.net coherence.oftc.net]
alanc has quit [nucleus.oftc.net coherence.oftc.net]
linearcannon has quit [nucleus.oftc.net coherence.oftc.net]
shoragan has quit [nucleus.oftc.net coherence.oftc.net]
punit has quit [nucleus.oftc.net coherence.oftc.net]
alyssa has quit [nucleus.oftc.net coherence.oftc.net]
Lightsword has quit [nucleus.oftc.net coherence.oftc.net]
xperia64 has quit [nucleus.oftc.net coherence.oftc.net]
flto has quit [nucleus.oftc.net coherence.oftc.net]
robink has quit [nucleus.oftc.net coherence.oftc.net]
anholt has quit [nucleus.oftc.net coherence.oftc.net]
siqueira has quit [nucleus.oftc.net coherence.oftc.net]
dri-logger has quit [nucleus.oftc.net coherence.oftc.net]
mattst88 has quit [nucleus.oftc.net coherence.oftc.net]
lstrano has quit [nucleus.oftc.net coherence.oftc.net]
dnkl has quit [nucleus.oftc.net coherence.oftc.net]
mattrope has quit [nucleus.oftc.net coherence.oftc.net]
ifreund has quit [nucleus.oftc.net coherence.oftc.net]
mclasen has quit [nucleus.oftc.net coherence.oftc.net]
join_subline has quit [nucleus.oftc.net coherence.oftc.net]
sarnex has quit [nucleus.oftc.net coherence.oftc.net]
eukara has quit [nucleus.oftc.net coherence.oftc.net]
ddevault has quit [nucleus.oftc.net coherence.oftc.net]
enilflah has quit [nucleus.oftc.net coherence.oftc.net]
sadlerap has quit [nucleus.oftc.net coherence.oftc.net]
nikitalita48 has quit [nucleus.oftc.net coherence.oftc.net]
ajax has quit [nucleus.oftc.net coherence.oftc.net]
remexre has quit [nucleus.oftc.net coherence.oftc.net]
rpigott has quit [nucleus.oftc.net coherence.oftc.net]
fluix has quit [nucleus.oftc.net coherence.oftc.net]
mmenzyns has quit [nucleus.oftc.net coherence.oftc.net]
zf has quit [nucleus.oftc.net coherence.oftc.net]
reductum has quit [nucleus.oftc.net coherence.oftc.net]
exit70_ has quit [nucleus.oftc.net coherence.oftc.net]
sh-zam has quit [nucleus.oftc.net coherence.oftc.net]
anujp has quit [nucleus.oftc.net coherence.oftc.net]
jhli has quit [nucleus.oftc.net coherence.oftc.net]
quantum5 has quit [nucleus.oftc.net coherence.oftc.net]
ogabbay_ has quit [nucleus.oftc.net coherence.oftc.net]
mareko has quit [nucleus.oftc.net coherence.oftc.net]
CosmicPenguin has quit [nucleus.oftc.net coherence.oftc.net]
zzag has quit [nucleus.oftc.net coherence.oftc.net]
dolphin has quit [nucleus.oftc.net coherence.oftc.net]
cyrozap has quit [nucleus.oftc.net coherence.oftc.net]
jolan_ has quit [nucleus.oftc.net coherence.oftc.net]
orbea has quit [nucleus.oftc.net coherence.oftc.net]
imirkin has quit [nucleus.oftc.net coherence.oftc.net]
airlied has quit [nucleus.oftc.net coherence.oftc.net]
kem has quit [nucleus.oftc.net coherence.oftc.net]
kurufu has quit [nucleus.oftc.net coherence.oftc.net]
craftyguy has quit [nucleus.oftc.net coherence.oftc.net]
emersion has quit [nucleus.oftc.net coherence.oftc.net]
sauce has quit [nucleus.oftc.net coherence.oftc.net]
TD-Linux has quit [nucleus.oftc.net coherence.oftc.net]
jbarnes has quit [nucleus.oftc.net coherence.oftc.net]
SanchayanMaity has quit [nucleus.oftc.net coherence.oftc.net]
ManMower has quit [nucleus.oftc.net coherence.oftc.net]
hwentlan___ has quit [nucleus.oftc.net coherence.oftc.net]
krh has quit [nucleus.oftc.net coherence.oftc.net]
chadv1 has quit [nucleus.oftc.net coherence.oftc.net]
robclark has quit [nucleus.oftc.net coherence.oftc.net]
eric_engestrom has quit [nucleus.oftc.net coherence.oftc.net]
clever has quit [nucleus.oftc.net coherence.oftc.net]
daniels has quit [nucleus.oftc.net coherence.oftc.net]
Sachiel has quit [nucleus.oftc.net coherence.oftc.net]
cmarcelo has quit [nucleus.oftc.net coherence.oftc.net]
zmike has quit [nucleus.oftc.net coherence.oftc.net]
melissawen has quit [nucleus.oftc.net coherence.oftc.net]
gruetze_ has quit [nucleus.oftc.net coherence.oftc.net]
Thymo has quit [nucleus.oftc.net coherence.oftc.net]
gpiccoli has quit [nucleus.oftc.net coherence.oftc.net]
thellstrom has quit [nucleus.oftc.net coherence.oftc.net]
alatiera has quit [nucleus.oftc.net coherence.oftc.net]
jfalempe has quit [nucleus.oftc.net coherence.oftc.net]
agx has quit [nucleus.oftc.net coherence.oftc.net]
kmn has quit [nucleus.oftc.net coherence.oftc.net]
OftenTimeConsuming has quit [nucleus.oftc.net coherence.oftc.net]
MrCooper has quit [nucleus.oftc.net coherence.oftc.net]
undvasistas[m] has quit [nucleus.oftc.net coherence.oftc.net]
zamundaaa has quit [nucleus.oftc.net coherence.oftc.net]
PiGLDN[m] has quit [nucleus.oftc.net coherence.oftc.net]
kalli0815[m] has quit [nucleus.oftc.net coherence.oftc.net]
jenatali has quit [nucleus.oftc.net coherence.oftc.net]
mauld has quit [nucleus.oftc.net coherence.oftc.net]
dv_ has quit [nucleus.oftc.net coherence.oftc.net]
pendingchaos has quit [nucleus.oftc.net coherence.oftc.net]
dafna2[m] has quit [nucleus.oftc.net coherence.oftc.net]
Ziemas has quit [nucleus.oftc.net coherence.oftc.net]
Venemo has quit [nucleus.oftc.net coherence.oftc.net]
ds` has quit [nucleus.oftc.net coherence.oftc.net]
jernej has quit [nucleus.oftc.net coherence.oftc.net]
Guest890 has quit [nucleus.oftc.net coherence.oftc.net]
CME has quit [nucleus.oftc.net coherence.oftc.net]
yshui` has quit [nucleus.oftc.net coherence.oftc.net]
q66 has quit [nucleus.oftc.net coherence.oftc.net]
V has quit [nucleus.oftc.net coherence.oftc.net]
CounterPillow has quit [nucleus.oftc.net coherence.oftc.net]
Ristovski has quit [nucleus.oftc.net coherence.oftc.net]
jcristau has quit [nucleus.oftc.net coherence.oftc.net]
Surkow|laptop has quit [nucleus.oftc.net coherence.oftc.net]
hakzsam has quit [nucleus.oftc.net coherence.oftc.net]
bl4ckb0ne has quit [nucleus.oftc.net coherence.oftc.net]
Strit[m] has quit [nucleus.oftc.net coherence.oftc.net]
mupuf has quit [nucleus.oftc.net coherence.oftc.net]
Lucretia has quit [nucleus.oftc.net coherence.oftc.net]
pjakobsson has quit [nucleus.oftc.net coherence.oftc.net]
libv has quit [nucleus.oftc.net coherence.oftc.net]
moben[m] has quit [nucleus.oftc.net coherence.oftc.net]
mstoeckl has quit [nucleus.oftc.net coherence.oftc.net]
robertmader[m] has quit [nucleus.oftc.net coherence.oftc.net]
zzoon[m] has quit [nucleus.oftc.net coherence.oftc.net]
YaLTeR[m] has quit [nucleus.oftc.net coherence.oftc.net]
doras has quit [nucleus.oftc.net coherence.oftc.net]
marex has quit [nucleus.oftc.net coherence.oftc.net]
Vin[m] has quit [nucleus.oftc.net coherence.oftc.net]
swick has quit [nucleus.oftc.net coherence.oftc.net]
pepp has quit [nucleus.oftc.net coherence.oftc.net]
neobrain[m] has quit [nucleus.oftc.net coherence.oftc.net]
ubitux has quit [nucleus.oftc.net coherence.oftc.net]
Tooniis[m] has quit [nucleus.oftc.net coherence.oftc.net]
sigmaris has quit [nucleus.oftc.net coherence.oftc.net]
rellla has quit [nucleus.oftc.net coherence.oftc.net]
mort_6 has quit [nucleus.oftc.net coherence.oftc.net]
opotin has quit [nucleus.oftc.net coherence.oftc.net]
pH5 has quit [nucleus.oftc.net coherence.oftc.net]
LaserEyess has quit [nucleus.oftc.net coherence.oftc.net]
unrelentingtech has quit [nucleus.oftc.net coherence.oftc.net]
dottedmag has quit [nucleus.oftc.net coherence.oftc.net]
vup has quit [nucleus.oftc.net coherence.oftc.net]
Wallbraker[m] has quit [nucleus.oftc.net coherence.oftc.net]
mmind00 has quit [nucleus.oftc.net coherence.oftc.net]
shadeslayer has quit [nucleus.oftc.net coherence.oftc.net]
tpalli has quit [nucleus.oftc.net coherence.oftc.net]
apinheiro[m] has quit [nucleus.oftc.net coherence.oftc.net]
samuelig has quit [nucleus.oftc.net coherence.oftc.net]
halfline[m] has quit [nucleus.oftc.net coherence.oftc.net]
dcbaker has quit [nucleus.oftc.net coherence.oftc.net]
i-garrison has quit [nucleus.oftc.net coherence.oftc.net]
karolherbst has quit [nucleus.oftc.net coherence.oftc.net]
tagr has quit [nucleus.oftc.net coherence.oftc.net]
pq has quit [nucleus.oftc.net coherence.oftc.net]
el0y has quit [nucleus.oftc.net coherence.oftc.net]
dllud has quit [nucleus.oftc.net coherence.oftc.net]
iokill has quit [nucleus.oftc.net coherence.oftc.net]
nielsdg has quit [nucleus.oftc.net coherence.oftc.net]
danylo has quit [nucleus.oftc.net coherence.oftc.net]
Emantor has quit [nucleus.oftc.net coherence.oftc.net]
tomeu82 has quit [nucleus.oftc.net coherence.oftc.net]
bnieuwenhuizen has quit [nucleus.oftc.net coherence.oftc.net]
pinchartl has quit [nucleus.oftc.net coherence.oftc.net]
FLHerne has quit [nucleus.oftc.net coherence.oftc.net]
JoshuaAshton has quit [nucleus.oftc.net coherence.oftc.net]
haagch_ has quit [nucleus.oftc.net coherence.oftc.net]
GyrosGeier has quit [nucleus.oftc.net coherence.oftc.net]
dos1 has quit [nucleus.oftc.net coherence.oftc.net]
ced117 has quit [nucleus.oftc.net coherence.oftc.net]
gnustomp[m] has quit [nucleus.oftc.net coherence.oftc.net]
sravn has quit [nucleus.oftc.net coherence.oftc.net]
tonyk has quit [nucleus.oftc.net coherence.oftc.net]
Mershl[m] has quit [nucleus.oftc.net coherence.oftc.net]
egalli has quit [nucleus.oftc.net coherence.oftc.net]
RAOFhehis[m] has quit [nucleus.oftc.net coherence.oftc.net]
Koniiiik has quit [nucleus.oftc.net coherence.oftc.net]
hatrix has quit [nucleus.oftc.net coherence.oftc.net]
unidan has quit [nucleus.oftc.net coherence.oftc.net]
gagallo7[m] has quit [nucleus.oftc.net coherence.oftc.net]
yoslin has quit [nucleus.oftc.net coherence.oftc.net]
ccr has quit [nucleus.oftc.net coherence.oftc.net]
Sumera[m] has quit [nucleus.oftc.net coherence.oftc.net]
baryluk has quit [nucleus.oftc.net coherence.oftc.net]
enunes has quit [nucleus.oftc.net coherence.oftc.net]
x512[m] has quit [nucleus.oftc.net coherence.oftc.net]
HayashiEsme[m] has quit [nucleus.oftc.net coherence.oftc.net]
degasus has quit [nucleus.oftc.net coherence.oftc.net]
go4godvin has quit [nucleus.oftc.net coherence.oftc.net]
javierm has quit [nucleus.oftc.net coherence.oftc.net]
loki_val has quit [nucleus.oftc.net coherence.oftc.net]
italove3 has quit [nucleus.oftc.net coherence.oftc.net]
xerpi[m] has quit [nucleus.oftc.net coherence.oftc.net]
Prf_Jakob has quit [nucleus.oftc.net coherence.oftc.net]
LaughingMan[m] has quit [nucleus.oftc.net coherence.oftc.net]
adavy has quit [nucleus.oftc.net coherence.oftc.net]
rawoul has quit [nucleus.oftc.net coherence.oftc.net]
jasuarez has quit [nucleus.oftc.net coherence.oftc.net]
evadot has quit [nucleus.oftc.net coherence.oftc.net]
_alice has quit [nucleus.oftc.net coherence.oftc.net]
Eighth_Doctor has quit [nucleus.oftc.net coherence.oftc.net]
dj-death has quit [nucleus.oftc.net coherence.oftc.net]
exit70[m] has quit [nucleus.oftc.net coherence.oftc.net]
tintou has quit [nucleus.oftc.net coherence.oftc.net]
Terman_ has quit [nucleus.oftc.net coherence.oftc.net]
kgz has quit [nucleus.oftc.net coherence.oftc.net]
haasn has quit [nucleus.oftc.net coherence.oftc.net]
boistordu_ex has quit [nucleus.oftc.net coherence.oftc.net]
co1umbarius has quit [nucleus.oftc.net coherence.oftc.net]
ella-0 has quit [nucleus.oftc.net coherence.oftc.net]
glennk has quit [nucleus.oftc.net coherence.oftc.net]
Akari` has quit [nucleus.oftc.net coherence.oftc.net]
lcn has quit [nucleus.oftc.net coherence.oftc.net]
Namarrgon has quit [nucleus.oftc.net coherence.oftc.net]
tango_ has quit [nucleus.oftc.net coherence.oftc.net]
Erandir has quit [nucleus.oftc.net coherence.oftc.net]
Lynne has quit [nucleus.oftc.net coherence.oftc.net]
aura[m] has quit [nucleus.oftc.net coherence.oftc.net]
chivay has quit [nucleus.oftc.net coherence.oftc.net]
cwfitzgerald has quit [nucleus.oftc.net coherence.oftc.net]
Anson[m] has quit [nucleus.oftc.net coherence.oftc.net]
gdevi has quit [nucleus.oftc.net coherence.oftc.net]
jekstrand[m] has quit [nucleus.oftc.net coherence.oftc.net]
naheemsays[m] has quit [nucleus.oftc.net coherence.oftc.net]
reactormonk[m] has quit [nucleus.oftc.net coherence.oftc.net]
sigmoidfunc[m] has quit [nucleus.oftc.net coherence.oftc.net]
zamundaaa[m] has quit [nucleus.oftc.net coherence.oftc.net]
JosExpsito[m] has quit [nucleus.oftc.net coherence.oftc.net]
mntmn has quit [nucleus.oftc.net coherence.oftc.net]
leandrohrb has quit [nucleus.oftc.net coherence.oftc.net]
bgs has quit [nucleus.oftc.net coherence.oftc.net]
xantoz has quit [nucleus.oftc.net coherence.oftc.net]
lplc has quit [nucleus.oftc.net coherence.oftc.net]
DPA has quit [nucleus.oftc.net coherence.oftc.net]
pmoreau has quit [nucleus.oftc.net coherence.oftc.net]
anparra[m] has quit [nucleus.oftc.net coherence.oftc.net]
mwk has quit [nucleus.oftc.net coherence.oftc.net]
pushqrdx[m] has quit [nucleus.oftc.net coherence.oftc.net]
MrR[m] has quit [nucleus.oftc.net coherence.oftc.net]
Dylanger has quit [nucleus.oftc.net coherence.oftc.net]
T_UNIX has quit [nucleus.oftc.net coherence.oftc.net]
alarumbe has quit [nucleus.oftc.net coherence.oftc.net]
kusma has quit [nucleus.oftc.net coherence.oftc.net]
Viciouss has quit [nucleus.oftc.net coherence.oftc.net]
hasebastian[m] has quit [nucleus.oftc.net coherence.oftc.net]
ppascher has quit [nucleus.oftc.net coherence.oftc.net]
urja has quit [nucleus.oftc.net coherence.oftc.net]
neobrain has quit [nucleus.oftc.net coherence.oftc.net]
xxmitsu has quit [nucleus.oftc.net coherence.oftc.net]
Mis012[m] has quit [nucleus.oftc.net coherence.oftc.net]
Newbyte has quit [nucleus.oftc.net coherence.oftc.net]
dliviu has quit [nucleus.oftc.net coherence.oftc.net]
BobBeck has quit [nucleus.oftc.net coherence.oftc.net]
jani has quit [nucleus.oftc.net coherence.oftc.net]
jkqxz has quit [nucleus.oftc.net coherence.oftc.net]
cleverca22[m] has quit [nucleus.oftc.net coherence.oftc.net]
calebccff has quit [nucleus.oftc.net coherence.oftc.net]
HdkR has quit [nucleus.oftc.net coherence.oftc.net]
mal has quit [nucleus.oftc.net coherence.oftc.net]
sven has quit [nucleus.oftc.net coherence.oftc.net]
ella-0[m] has quit [nucleus.oftc.net coherence.oftc.net]
jannau has quit [nucleus.oftc.net coherence.oftc.net]
Stary has quit [nucleus.oftc.net coherence.oftc.net]
marcan has quit [nucleus.oftc.net coherence.oftc.net]
fahien has quit [nucleus.oftc.net coherence.oftc.net]
kallisti5[m] has quit [nucleus.oftc.net coherence.oftc.net]
robertfoss has quit [nucleus.oftc.net coherence.oftc.net]
lanodan_ has quit [nucleus.oftc.net coherence.oftc.net]
mriesch has quit [nucleus.oftc.net coherence.oftc.net]
milek7 has quit [nucleus.oftc.net coherence.oftc.net]
Adrinael has quit [nucleus.oftc.net coherence.oftc.net]
dt9 has quit [nucleus.oftc.net coherence.oftc.net]
ickle_ has quit [nucleus.oftc.net coherence.oftc.net]
imre has quit [nucleus.oftc.net coherence.oftc.net]
turol has quit [nucleus.oftc.net coherence.oftc.net]
ceyusa has quit [nucleus.oftc.net coherence.oftc.net]
romangg has quit [nucleus.oftc.net coherence.oftc.net]
mceier has quit [nucleus.oftc.net coherence.oftc.net]
vsyrjala has quit [nucleus.oftc.net coherence.oftc.net]
chema has quit [nucleus.oftc.net coherence.oftc.net]
DrNick has quit [nucleus.oftc.net coherence.oftc.net]
egbert has quit [nucleus.oftc.net coherence.oftc.net]
heftig has quit [nucleus.oftc.net coherence.oftc.net]
APic has quit [nucleus.oftc.net coherence.oftc.net]
hikiko_ has quit [nucleus.oftc.net coherence.oftc.net]
Plagman has quit [nucleus.oftc.net coherence.oftc.net]
jadahl has quit [nucleus.oftc.net coherence.oftc.net]
tjaalton has quit [nucleus.oftc.net coherence.oftc.net]
tomba has quit [nucleus.oftc.net coherence.oftc.net]
MatrixTravelerbot[m] has quit [nucleus.oftc.net coherence.oftc.net]
bylaws has quit [nucleus.oftc.net coherence.oftc.net]
moony has quit [nucleus.oftc.net coherence.oftc.net]
dwlsalmeida has quit [nucleus.oftc.net coherence.oftc.net]
illwieckz has quit [nucleus.oftc.net coherence.oftc.net]
minecrell has quit [nucleus.oftc.net coherence.oftc.net]
qyliss has quit [nucleus.oftc.net coherence.oftc.net]
oneforall2 has quit [nucleus.oftc.net coherence.oftc.net]
mbrost has quit [nucleus.oftc.net coherence.oftc.net]
ngcortes has quit [nucleus.oftc.net coherence.oftc.net]
cphealy_ has quit [nucleus.oftc.net coherence.oftc.net]
macromorgan has quit [nucleus.oftc.net coherence.oftc.net]
jewins1 has quit [nucleus.oftc.net coherence.oftc.net]
pzanoni` has quit [nucleus.oftc.net coherence.oftc.net]
anarsoul has quit [nucleus.oftc.net coherence.oftc.net]
demarchi has quit [nucleus.oftc.net coherence.oftc.net]
sumits has quit [nucleus.oftc.net coherence.oftc.net]
soreau has quit [nucleus.oftc.net coherence.oftc.net]
holmanb has quit [nucleus.oftc.net coherence.oftc.net]
jekstrand has quit [nucleus.oftc.net coherence.oftc.net]
JohnnyonFlame has quit [nucleus.oftc.net coherence.oftc.net]
rossy has quit [nucleus.oftc.net coherence.oftc.net]
MTCoster__ has quit [nucleus.oftc.net coherence.oftc.net]
rib___ has quit [nucleus.oftc.net coherence.oftc.net]
cengiz_io has quit [nucleus.oftc.net coherence.oftc.net]
SolarAquarion has quit [nucleus.oftc.net coherence.oftc.net]
lileo__ has quit [nucleus.oftc.net coherence.oftc.net]
jjardon___ has quit [nucleus.oftc.net coherence.oftc.net]
mdnavare has quit [nucleus.oftc.net coherence.oftc.net]
rsripada has quit [nucleus.oftc.net coherence.oftc.net]
kisak has quit [nucleus.oftc.net coherence.oftc.net]
dschuermann has quit [nucleus.oftc.net coherence.oftc.net]
mmx_in_orbit has quit [nucleus.oftc.net coherence.oftc.net]
markyacoub has quit [nucleus.oftc.net coherence.oftc.net]
jessica_24 has quit [nucleus.oftc.net coherence.oftc.net]
abhinav___ has quit [nucleus.oftc.net coherence.oftc.net]
hfink has quit [nucleus.oftc.net coherence.oftc.net]
andrey-konovalov has quit [nucleus.oftc.net coherence.oftc.net]
curro has quit [nucleus.oftc.net coherence.oftc.net]
aswar002 has quit [nucleus.oftc.net coherence.oftc.net]
mslusarz has quit [nucleus.oftc.net coherence.oftc.net]
rando25892 has quit [nucleus.oftc.net coherence.oftc.net]
neoXite_ has quit [nucleus.oftc.net coherence.oftc.net]
marcheu has quit [nucleus.oftc.net coherence.oftc.net]
seanpaul__ has quit [nucleus.oftc.net coherence.oftc.net]
rodrigovivi has quit [nucleus.oftc.net coherence.oftc.net]
agd5f has quit [nucleus.oftc.net coherence.oftc.net]
tzanger has quit [nucleus.oftc.net coherence.oftc.net]
tfiga has quit [nucleus.oftc.net coherence.oftc.net]
angular_mike_____ has quit [nucleus.oftc.net coherence.oftc.net]
graphitemaster has quit [nucleus.oftc.net coherence.oftc.net]
robher has quit [nucleus.oftc.net coherence.oftc.net]
trn has quit [nucleus.oftc.net coherence.oftc.net]
arnd_ has quit [nucleus.oftc.net coherence.oftc.net]
eletrotupi has quit [nucleus.oftc.net coherence.oftc.net]
tchar has quit [nucleus.oftc.net coherence.oftc.net]
jrayhawk has quit [nucleus.oftc.net coherence.oftc.net]
Lyude has quit [nucleus.oftc.net coherence.oftc.net]
JTL has quit [nucleus.oftc.net coherence.oftc.net]
cwabbott has quit [nucleus.oftc.net coherence.oftc.net]
rg3igalia__ has quit [nucleus.oftc.net coherence.oftc.net]
swivel has quit [nucleus.oftc.net coherence.oftc.net]
dianders has quit [nucleus.oftc.net coherence.oftc.net]
austriancoder has quit [nucleus.oftc.net coherence.oftc.net]
glisse has quit [nucleus.oftc.net coherence.oftc.net]
narmstrong has quit [nucleus.oftc.net coherence.oftc.net]
jstultz has quit [nucleus.oftc.net coherence.oftc.net]
jljusten has quit [nucleus.oftc.net coherence.oftc.net]
steev has quit [nucleus.oftc.net coherence.oftc.net]
FluffyFoxeh has quit [nucleus.oftc.net coherence.oftc.net]
eric_engestrom has joined #dri-devel
xxmitsu has joined #dri-devel
zf has joined #dri-devel
alanc has joined #dri-devel
sarnex has joined #dri-devel
exit70 has joined #dri-devel
nikitalita48 has joined #dri-devel
mclasen has joined #dri-devel
jhli has joined #dri-devel
Frogging101 has joined #dri-devel
melissawen has joined #dri-devel
Frogging101 is now known as FluffyFoxeh
Duke`` has joined #dri-devel
ifreund_ has joined #dri-devel
ddevault_ has joined #dri-devel
dnkl_ has joined #dri-devel
hwentlan____ has joined #dri-devel
fluix has joined #dri-devel
Frogging101 has joined #dri-devel
SanchayanMaity_ has joined #dri-devel
daniels has joined #dri-devel
ogabbay__ has joined #dri-devel
eric_engestrom has quit [Ping timeout: 480 seconds]
FluffyFoxeh has quit [Read error: Connection reset by peer]
Frogging101 is now known as FluffyFoxeh
zzag has joined #dri-devel
eric_engestrom has joined #dri-devel
rpigott has joined #dri-devel
camus1 has quit [Ping timeout: 480 seconds]
punit has joined #dri-devel
camus has joined #dri-devel
radii has joined #dri-devel
tzimmermann has joined #dri-devel
mclasen has quit [Ping timeout: 480 seconds]
itoral has joined #dri-devel
thellstrom1 has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
mszyprow has joined #dri-devel
camus1 has joined #dri-devel
camus has quit [Ping timeout: 480 seconds]
ishitatsuyuki has joined #dri-devel
jbarnes has joined #dri-devel
zf has quit [Remote host closed the connection]
mmenzyns has joined #dri-devel
camus1 has quit [Ping timeout: 480 seconds]
zf has joined #dri-devel
gpiccoli has joined #dri-devel
ddevault_ has left #dri-devel [#dri-devel]
itoral has quit [Remote host closed the connection]
itoral has joined #dri-devel
camus has joined #dri-devel
ppascher has joined #dri-devel
alanc has quit [Remote host closed the connection]
alanc has joined #dri-devel
ramaling has joined #dri-devel
Company has joined #dri-devel
ahajda has joined #dri-devel
MajorBiscuit has joined #dri-devel
danvet has joined #dri-devel
shoragan has joined #dri-devel
rgallaispou has joined #dri-devel
<pq>
danvet, FWIW, I think black&white only is a good idea, because who knows what the color management properties have been set to. And maybe you're on a grayscale or 2-color display.
<danvet>
pq, yeah for the emergency log max contrast is best
<danvet>
also easiest
mripard has joined #dri-devel
<danvet>
but I think drmcon actually provided a full console, just bypassig all of fbcon
bbrezillon has joined #dri-devel
<danvet>
and there I guess people would like some colors for highlighting
<danvet>
javierm, that syzbot bisect is funny
camus1 has joined #dri-devel
<danvet>
probably pre-existing issue, but with a reproducer should be doable to track down
<emersion>
what's the difference between drmcon and kmscon? drmcon is still in-kernel?
<danvet>
yeah
<danvet>
I think ideally we only need the drm emergency log + kmscon in userspace
<danvet>
but for smoother transition on all distros maybe a real drm console is better
<danvet>
or maybe both
<danvet>
since the real console sucks more as emergency logger
camus has quit [Read error: Connection reset by peer]
<danvet>
airlied, tzimmermann I'll leave the late drm-misc-next-fixes pull for -rc2 and don't spin another one
<danvet>
it doesn't look like it's super important to land these for -rc1
<tzimmermann>
danvet, ok
<tzimmermann>
i was late this week. sorry
<danvet>
yeah no problem
<danvet>
also if I'd still done it it would have been in time :-)
<MrCooper>
ajax airlied: FWIW, DSC is working fine here (with Navi 21, it's required for 240 Hz with the Samsung G9 :)
tursulin has joined #dri-devel
<tzimmermann>
danvet, i usually try to send PRs on thursday. but yesterday i got perma-interrupted
<danvet>
yeah it happens and really no big deal
<danvet>
for something serious I'd have spun up a pr to linus, but doesn't look like
<tzimmermann>
and i tested simpledrm on opensuse to work with the proprietary graphics driver of a vendor that shall not be named here. that was frustrating
<tzimmermann>
at least it worked
<javierm>
tzimmermann: great. People reported issue with simplefb but your upcoming patches to unregister the pdev when removing conflicting fb will fix that
eukara_ has joined #dri-devel
quantum5_ has joined #dri-devel
enilflah_ has joined #dri-devel
lstrano_ has joined #dri-devel
ManMower_ has joined #dri-devel
sadlerap1 has joined #dri-devel
mareko_ has joined #dri-devel
jolan has joined #dri-devel
sh_zam has joined #dri-devel
<tzimmermann>
javierm, that bug report you linked. do you remember what driver version that-other-driver had? i expected failure, but when i used 470.* it worked
airlied_ has joined #dri-devel
Sachiel_ has joined #dri-devel
mmenzyns has quit [synthon.oftc.net weber.oftc.net]
FluffyFoxeh has quit [synthon.oftc.net weber.oftc.net]
fluix has quit [synthon.oftc.net weber.oftc.net]
radii has quit [synthon.oftc.net weber.oftc.net]
rpigott has quit [synthon.oftc.net weber.oftc.net]
orbea has joined #dri-devel
rgallaispou has quit []
mmenzyns has joined #dri-devel
<javierm>
tzimmermann: one sec, let me dig that from the fedora channel irc logs
krh has joined #dri-devel
<javierm>
< kwizart>| this is with 510.39.01 beta
dolphin has joined #dri-devel
<javierm>
tzimmermann: ^
<tzimmermann>
thanks
cmarcelo has joined #dri-devel
clever has joined #dri-devel
join_subline has joined #dri-devel
ajax has joined #dri-devel
kurufu has joined #dri-devel
radii has joined #dri-devel
TD-Linux has joined #dri-devel
imirkin has joined #dri-devel
kem has joined #dri-devel
emersion has joined #dri-devel
Ultrasauce has joined #dri-devel
FluffyFoxeh has joined #dri-devel
cyrozap has joined #dri-devel
robclark has joined #dri-devel
chadv1 has joined #dri-devel
fluix has joined #dri-devel
reductum has joined #dri-devel
rpigott has joined #dri-devel
remexre has joined #dri-devel
zmike has joined #dri-devel
anujp has joined #dri-devel
craftyguy has joined #dri-devel
pcercuei has joined #dri-devel
CosmicPenguin has joined #dri-devel
itoral has quit [Remote host closed the connection]
itoral has joined #dri-devel
camus has joined #dri-devel
Lightsword_ has joined #dri-devel
camus1 has quit [Ping timeout: 480 seconds]
Lightsword_ has quit []
Lightsword has joined #dri-devel
tarceri has quit [Read error: Connection reset by peer]
Lightsword_ has joined #dri-devel
Lightsword has quit [Read error: Connection reset by peer]
rgallaispou has joined #dri-devel
Peroverik has joined #dri-devel
mlankhorst has joined #dri-devel
Lightsword_ has quit []
Peroverik has quit []
Peroverik has joined #dri-devel
TD-Linux has quit [Ping timeout: 480 seconds]
TD-Linux has joined #dri-devel
camus has quit []
Peroverik has quit []
TD-Linux has quit [Ping timeout: 480 seconds]
Peroverik has joined #dri-devel
rasterman has joined #dri-devel
Peroverik has quit []
TD-Linux has joined #dri-devel
demarchi has joined #dri-devel
pnowack has joined #dri-devel
robert_mader has joined #dri-devel
Lightsword has joined #dri-devel
<robert_mader>
pq, swick, Company: I think one important thing no note on gw point is that youtube and all those streaming platforms today often have a small, semitransparent logo layered on top of their videos. The logo is important for brand recognition of all those streamers and expected to look "right" - i.e. have non-linear blending (if we agree to choose that term). So if we want Wayland to allow browsers to be energy efficient with streaming
<robert_mader>
platforms how they are today (and at least for the next few years), some kind of guarantee for non-linear blending would be needed. IIUC jadahl s proposal to have that for pure SDR subsurface trees would fit the bill.
<Company>
so that's basically the subtitle thing I mentioned
<jadahl>
heh, a third channel to discuss the same thing :P
<Company>
just that it's a dumb but permanent subtitle
<robert_mader>
And one that really cares about colors, unfortunately :(
<jadahl>
robert_mader: yea i know about that scanout-killing logo. just as frustrating as the rounded corners next to the panel in gnome-shell
<Company>
robert_mader: how does that work on Windows/Android/OSX/other desktops?
<Company>
do you know?
Haaninjo has joined #dri-devel
<robert_mader>
Company: AFAIK they just give the guarantee for non-linear blender. However, Windows/DirectComposition does not enable HDR by default and MacOS/CoreAnimation I don't understand yet how they sneak HDR elements in.
<swick>
robert_mader: import the video frame to GL/Vk blend the logo in non-linear space, attach to wl_surface. worse than 2 overlays, better than compositing the video with the whole scene…
<swick>
robert_mader: what exactly do they guarantee? non-linear blending isn't *a* thing.
<Company>
robert_mader: I was wondering about Windows with enabled HDR
<Company>
(and a MacOS display-p3)
<robert_mader>
Company: I think Windows hasn't figured out how to have good performance and HDR at the same time - which is why they don't enable it.
<robert_mader>
swick: I mostly care about display compositor offloading.
<Company>
now I need a HDR monitor so I can enable it and watch amouranth for science?
<robert_mader>
*display controller offloading
<swick>
robert_mader: that doesn't answer any questions I have
<robert_mader>
swick: lets call it CSS compatible blending?
<swick>
and what exactly is that?
<Company>
CSS is undefined ;)
<robert_mader>
well, that simple blending function that's currently used everywhere? How should I call it?
<swick>
the blending function is the same for linear and non-linear blending
<Company>
robert_mader: what's the state with CSS blending and colorspaces, do you know?
<swick>
so, you mean sRGB inverse EOTF non-linearity on source and destination?
<robert_mader>
swick: yes, I think so. Can we find a shorter catchy name for it?
<swick>
sure, sRGB blending
<Company>
the dark in the middle blending
JohnnyonFlame has joined #dri-devel
<Company>
twitch doesn't have any overlays btw
<robert_mader>
Company: IIUC there are just proposals so far but a severe lack of actual implementations. But gw would be a better person ask
<Company>
at least not on the desktop
<Company>
well, it fades out all the overlays
<Company>
it has tons of overlays the moment I accidentally move the mouse
<robert_mader>
Well, asked another Firefox graphics person that's deeper into the topic to come here for some more direct conversations.
mareko_ is now known as mareko
<pq>
robert_mader, wrong channel?
<robert_mader>
pq:
<robert_mader>
pq: uh, indeed, should have been in #wayland!
<robert_mader>
sorry
flacks has joined #dri-devel
rgallaispou has left #dri-devel [#dri-devel]
YuGiOhJCJ has joined #dri-devel
tobiasjakobi has joined #dri-devel
tobiasjakobi has quit [Remote host closed the connection]
mclasen has joined #dri-devel
itoral has quit [Remote host closed the connection]
Peste_Bubonica has joined #dri-devel
itoral has joined #dri-devel
itoral has quit [Remote host closed the connection]
itoral has joined #dri-devel
aravind has quit [Ping timeout: 480 seconds]
mlankhorst has quit [Ping timeout: 480 seconds]
JohnnyonFlame has quit [Ping timeout: 480 seconds]
mlankhorst has joined #dri-devel
JohnnyonFlame has joined #dri-devel
devilhorns has joined #dri-devel
soreau has joined #dri-devel
Karyon has joined #dri-devel
ManMower_ is now known as ManMower
gruetzkopf has joined #dri-devel
gruetzkopf has quit []
iive has joined #dri-devel
JohnnyonFlame has quit [Ping timeout: 480 seconds]
mszyprow has quit [Ping timeout: 480 seconds]
JohnnyonFlame has joined #dri-devel
<tzimmermann>
danvet, do you remember why exactly fbdev deferred io is incompatible with shmem helpers?
nchery has joined #dri-devel
jewins has joined #dri-devel
ppascher has quit [Ping timeout: 480 seconds]
heckt has joined #dri-devel
verifying has joined #dri-devel
heckt has quit [Remote host closed the connection]
itoral has quit [Remote host closed the connection]
sdutt has joined #dri-devel
oneforall2 has joined #dri-devel
robert_mader has left #dri-devel [#dri-devel]
verifying has quit [Remote host closed the connection]
verifying has joined #dri-devel
gruetzkopf has joined #dri-devel
sdutt has quit []
sdutt has joined #dri-devel
tursulin has quit [Quit: Konversation terminated!]
flto has joined #dri-devel
tursulin has joined #dri-devel
verifying has quit [Remote host closed the connection]
verifying has joined #dri-devel
verifying has quit [Remote host closed the connection]
verifying has joined #dri-devel
verifying has quit [Remote host closed the connection]
<danvet>
tzimmermann, thrashes struct page fields and upsets shmem
<danvet>
iirc the solution would be to intercept page faults and keep track of dirty pages in a separately allocated bitmask
<danvet>
the tricky bit would be to figure out how to intercept these vma_ops, while still passing everything else through to the underlying gem vma ops
<danvet>
so that it all neatly works with shmem, ttm and maybe even cma backend
ella-0_ has joined #dri-devel
ahajda has quit []
mlankhorst has joined #dri-devel
<tzimmermann>
danvet, the thing is, it actually works :)
<tzimmermann>
probably by chance
<tzimmermann>
it mmaped a shmem BO via fbdev with deferred io
<tzimmermann>
the fbdev tests run quite a bit of mmap'ed io and it all works
<tzimmermann>
shmem's vma open/close callbacks get and put the pages. but fbdev already did that when it vmap'ed the BO memory
<tzimmermann>
and the vma fault callback is probably compatible by chance
mattrope has joined #dri-devel
<danvet>
tzimmermann, hm
<danvet>
tzimmermann, that really should blow up eventually
<danvet>
fb_deferred_io_mkwrite takes ownership of the page->lru
<tzimmermann>
both, shmem and fbdevio, use vma->vm_private_data. but that can be avoided easily and doesn't seem to be a problem. i'm kni of lost now, what else might be broken
<danvet>
and for a page in pagecache that's owned by the pagecache
<tzimmermann>
danvet, i see. that's a good point
<danvet>
but also I have at most cargo-culted understanding of the vm
<danvet>
so who knows
<tzimmermann>
:D
<danvet>
but yeah when I looked without typing my idea was to create a fbdev_vma ops
<danvet>
which forward everything to the next level (tricky, since we'd need to store the real vma_ops somewhere)
<danvet>
and intercept mkwrite to set a bitfield of dirty pages
<danvet>
and for ->fault it would make sure the page is only ever read-faulted
<danvet>
which could be tricky if the fault handler inserts the pte, but iirc you can supply the desired pte mode somewhere
<danvet>
it all looked very messy and I just didn't care enough to figure out the details and type it up
<tzimmermann>
danvet, i had similar thoughts
<danvet>
tzimmermann, I think we could just keep our vma pointer and anything else we need in a separate lookup tree or so
<danvet>
to avoid the problem of us having to overwrite ->vma_ops and not really being able to overwrite vm_private (or the real ops would just break)
<danvet>
I wanted to discuss this also with willy at some conference, whether it's reasonable to intercept faults like this in a generic way
<danvet>
but well this pandemic thing happened
<tzimmermann>
in the fbdevio callbacks, i saved and restored the driver's original values for vma->vm_private_data and vma->vm_ops. it looks a bit ugly, but has little overhead
<danvet>
ah yeah I guess that works too
<danvet>
a bit more code in wrappers but all doable
<tzimmermann>
for the lru problem, i guess i'll try your idea of a bitarray
<danvet>
hm so one thing I'm not sure about is locking
<danvet>
i.e. can you replace vm_ops in callbacks
<danvet>
I think for most only mmap_sem is held in read mode
<danvet>
which would indicate the answer is "no" :-(
<tzimmermann>
danvet, handling that is part of the save/restore code
<danvet>
save/restore code?
<tzimmermann>
save/restore for the driver's original vma fields. around each wrapped vma op
<tzimmermann>
never mind
<tzimmermann>
i'm also considering to remove the fbdefio entirely from fbdev emulation. instead drm_fbdev_fb_mmap() would tell the gem callback that the mmap comes from fbdev and let's the gem memory manager handle the details. hopefully the resulting gem code would be sharable in a helper lib
<tzimmermann>
just an idea so far
<tzimmermann>
for gem shmem, the mmap would work similar to regular mmap, but gem shmem would collect pagefaults for the BO and tell fbdev to to damage ahndign every few millisecs. i wonder if that would work
<danvet>
yeah, but that means more driver code
<danvet>
which is annoying
<danvet>
I think we should be able to get away with overwriting vma_ops
<danvet>
and then we could do the replacement for vma_private under a lock of our own in defio code
<danvet>
the problem I think is really only mm calling into vma_ops
<danvet>
and I don't think any of our gem vma ops care about the ops pointer being right
<danvet>
so something like struct defio_vm_private { void * real_vm_private; long *dirty_pages_bitfield; struct mutex lock; };
<tzimmermann>
maybe i try both ideas.
Peste_Bubonica has quit [Quit: Leaving]
nchery has quit [Ping timeout: 480 seconds]
nchery has joined #dri-devel
Duke`` has joined #dri-devel
<danvet>
javierm, tzimmermann btw since at least fedora still has VGA16FB enabled, do you plan to replace that?
<danvet>
or can simpledrmfb already drive that in the reasonable modes (without modesetting)?
<danvet>
asking since Gerd on dri-devel said that the plane is to only disable vesafb and efifb
<cwabbott>
long story short, I now have patches to mesa and VK-GL-CTS for all the remaining issues, and that's the last mesa thing
<tzimmermann>
danvet, we'll also replace all fbdev drivers on suse with simpledrm. we have some obscure drivers enable and no one knows whether they are needed. if anyone asks for them, i'll simply wite drm replacements
<alyssa>
cwabbott: any shader calculating determinants in a critical path is doomed
<alyssa>
then again, ray tracing
<cwabbott>
alyssa: well, I was more motivated by the test fails
Peroverik has joined #dri-devel
<alyssa>
and deleting code, it seems :D
<cwabbott>
at least the new thing is less code + faster too
<alyssa>
woo
Duke`` has quit []
<pendingchaos>
graphitemaster: with gl_SubgroupInvocationID, I think so, but probably not with acceptable performance
<pendingchaos>
subgroupBroadcast() for booleans can be vote(val && gl_SubgroupInvocationID == index), 32 of those for a integer broadcast, a bunch to read all values across invocations and implement a shuffle, reductions can be implemented on shuffles
Duke`` has joined #dri-devel
JohnnyonF has joined #dri-devel
JohnnyonFlame has quit [Ping timeout: 480 seconds]
ahajda has quit []
<alyssa>
cwabbott: I do wonder if it makes sense to put that in nir_builtin_builder so it can be shared with glsl->nir
Sachiel_ is now known as Sachiel
<alyssa>
I guess GLSL IR doesn't model determinants
<alyssa>
never mind
<alyssa>
i regret looking at the implementation of builtin_builder::_determinant_mat4
<alyssa>
jekstrand: ^ a use case for compiling glsl to spirv and then use vtn
JohnnyonFlame has joined #dri-devel
JohnnyonF has quit [Ping timeout: 480 seconds]
<alyssa>
zmike: can u_vbuf (or something else) unroll indexing entirely?
<alyssa>
like, rewrite indexed draws into direct draws
JohnnyonF has joined #dri-devel
<alyssa>
for debug
<zmike>
I think that'd be a primconvert thing?
<imirkin>
alyssa: we do that in nv50/nvc0 for small draws. makes sense if all the bits are client-side
<alyssa>
wait i think i figured out what i screwed up
JohnnyonFlame has quit [Ping timeout: 480 seconds]
mszyprow has joined #dri-devel
Peroverik has quit [Ping timeout: 480 seconds]
Peroverik has joined #dri-devel
pnowack has quit [Quit: pnowack]
devilhorns has quit []
mszyprow has quit [Ping timeout: 480 seconds]
Peroverik has quit []
hch12907 has joined #dri-devel
<alyssa>
ah, it wants the index buffer to be aligned in memory
Peroverik has joined #dri-devel
<imirkin>
so picky
pnowack has joined #dri-devel
Peroverik has quit []
Peroverik has joined #dri-devel
Peroverik has quit []
Peroverik has joined #dri-devel
mszyprow has joined #dri-devel
thellstrom has joined #dri-devel
thellstrom1 has quit [Read error: Connection reset by peer]
<ccr>
at least it's not D&D style alignment. imagine - instead of 2^n alignment, you'd have good/evil/neutral etc.
<emersion>
i didn't ask you whether you would use it or not
<emersion>
i asked you whether you'd object to merging
<emersion>
introducing the concept of encoders to newcomers is pretty difficult
<emersion>
"this one thing you need to use but really shouldn't care about"
<emersion>
not having this function is the only reason why i need to teach about encoders at all
<emersion>
i don't especially care whether this is useful for existing users
<emersion>
i just want to make it easy for new users
<ccr>
>_> that feeling when you realize that you just shot yourself in the foot with git while splitting a commit by doing "reset --hard" at wrong place/time, and losing 2 days worth of work. and sweating profusely try extundelete to see if something could be rescued. but no. and then realizing that git rebase --show-current-patch has the previous state still, and the feeling of relief is undescribable.
<anholt>
ccr: have you heard the good word of "git reflog"?
<ccr>
anholt, not really, I've mostly been a mercurial user.
<ccr>
ah
<anholt>
it's one of the coolest tools in the swiss army chainsaw
<anholt>
basically, if you've ever[*] git committed it, it's not really lost. [*] not strictly true, but pretty close
<ccr>
hmmm, cool
<anholt>
austriancoder, jasuarez: were one of you using a cisco switch for power control of ci farm?
<Venemo>
ccr: if the file is still open in a text editor, it is likely you can restore it as easily as Ctrl+Z
<jasuarez>
Not a Cisco but a swich with PoE to turn on off devices
<anholt>
I seemed to remember someone with a bm power script that looked like cisco config, but now I can't find it.
<jasuarez>
IIRC its a Netgear and I think it had an interface Cisco compatible
<austriancoder>
anholt: no .. I am using an IP based relay board
<jasuarez>
Based on Telnet
<jasuarez>
We had a script to control it
<jasuarez>
But then switched to SNMP
ngcortes has joined #dri-devel
<ccr>
Venemo, too late for that, but reflog helped (well, it saved me from re-writing commit message for one of the commits)
<ccr>
anholt, thanks for the reflog tip!
Viciouss has joined #dri-devel
gouchi has quit [Remote host closed the connection]
MajorBiscuit has quit [Ping timeout: 480 seconds]
danvet has quit [Ping timeout: 480 seconds]
<mdnavare>
vsyrjala: For the affected CRTC DRM patch that you reviewed, should I add a Fixes tag?