alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - Logs https://oftc.irclog.whitequark.org/panfrost - <macc24> i have been here before it was popular
CrazyEddy has joined #panfrost
CrazyEddy has quit [Remote host closed the connection]
atler is now known as Guest2169
atler has joined #panfrost
Schoentoon22 has joined #panfrost
Schoentoon22 has quit [Remote host closed the connection]
Guest2169 has quit [Ping timeout: 480 seconds]
radix42__ has joined #panfrost
radix42__ has quit [Read error: Connection reset by peer]
FelixErnst-M has joined #panfrost
FelixErnst-M has quit [Remote host closed the connection]
camus has joined #panfrost
phalanx has joined #panfrost
phalanx has quit [Remote host closed the connection]
pastly-antispam has quit [autokilled: Suspected spammer. Mail support@oftc.net with questions (2021-06-15 02:03:48)]
pastly-antispam has joined #panfrost
camus1 has joined #panfrost
camus has quit [Ping timeout: 480 seconds]
ryjm has joined #panfrost
ryjm has quit [Remote host closed the connection]
camus1 has quit [Remote host closed the connection]
camus has joined #panfrost
alad has joined #panfrost
alad has quit [Remote host closed the connection]
zhongfu_ has joined #panfrost
zhongfu_ has quit [Remote host closed the connection]
camus has quit [Remote host closed the connection]
camus has joined #panfrost
jambalaya has quit [Ping timeout: 480 seconds]
tlwoerner__ has joined #panfrost
jambalaya has joined #panfrost
tlwoerner_ has quit [Ping timeout: 480 seconds]
Ivan5-t has joined #panfrost
Ivan5-t has quit [Remote host closed the connection]
hanetzer has quit [Quit: WeeChat 3.0.1]
hanetzer has joined #panfrost
heyj has joined #panfrost
heyj has quit [Remote host closed the connection]
floof58 has quit [Ping timeout: 480 seconds]
floof58 has joined #panfrost
kwmiebach has joined #panfrost
kwmiebach has quit [Remote host closed the connection]
Ottomite_Desk has joined #panfrost
Ottomite_Desk has quit [Remote host closed the connection]
camus has quit []
camus has joined #panfrost
camus has quit []
camus has joined #panfrost
whynot has joined #panfrost
whynot has quit [Remote host closed the connection]
tlwoerner__ has quit []
tlwoerner has joined #panfrost
floof58 has quit [Remote host closed the connection]
moniker-- has joined #panfrost
moniker-- has quit [Remote host closed the connection]
andrey-konovalov has joined #panfrost
floof58 has joined #panfrost
DustyDingo has joined #panfrost
DustyDingo has quit [Remote host closed the connection]
zezinho-M has joined #panfrost
zezinho-M has quit [Remote host closed the connection]
leah has joined #panfrost
alpernebbi has joined #panfrost
camus has quit [Remote host closed the connection]
camus has joined #panfrost
dtrainor has joined #panfrost
dtrainor has quit [Remote host closed the connection]
Sammitch has joined #panfrost
Sammitch has quit [Remote host closed the connection]
floof58 has quit [Read error: Connection reset by peer]
floof58 has joined #panfrost
Ekkie has joined #panfrost
Ekkie has quit [Remote host closed the connection]
alpernebbi has quit [Quit: alpernebbi]
LordLamer has joined #panfrost
LordLamer has quit [Remote host closed the connection]
floof58 has quit [Ping timeout: 480 seconds]
floof58 has joined #panfrost
rasterman has joined #panfrost
floof58 has quit [Read error: No route to host]
floof58 has joined #panfrost
floof58 has quit [Remote host closed the connection]
floof58 has joined #panfrost
jernej has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net]
floof58 has quit [Remote host closed the connection]
floof58 has joined #panfrost
thecycoone has quit []
thecycoone has joined #panfrost
camus1 has joined #panfrost
camus has quit [Remote host closed the connection]
floof58 has quit [Remote host closed the connection]
floof58 has joined #panfrost
jernej has joined #panfrost
jernej has quit []
jernej has joined #panfrost
floof58 has quit [Read error: No route to host]
floof58 has joined #panfrost
floof58_ has joined #panfrost
floof58 has quit []
tonyt has joined #panfrost
tonyt has quit [Remote host closed the connection]
jernej has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net]
jernej has joined #panfrost
jernej has quit [Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net]
Belus-M has joined #panfrost
Belus-M has quit [Remote host closed the connection]
jernej has joined #panfrost
Danct12 has quit [Quit: Quitting]
Danct12 has joined #panfrost
Danct12 has quit []
Danct12 has joined #panfrost
mofm has joined #panfrost
mofm is now known as Guest2260
Guest2260 has quit [Remote host closed the connection]
<bbrezillon> stepri01: I'm resurrecting https://patchwork.kernel.org/project/dri-devel/patch/20190816093107.30518-2-steven.price@arm.com/, and it looks like timeout handling gets trickier with 2 jobs per slot :-(
<bbrezillon> the job passed to `->timedout_job()` is not necessarily the one that timedout
Danct12 has quit [Quit: Quitting]
<bbrezillon> hm, might be simpler than I thought, I can just signal the fence on the job that finished in the timeout handler, and let the core restart the timer and timeout a second time on the faulty job
<rasterman> bah
<rasterman> better
<rasterman> this is new...
<rasterman> glmark.... segv
<alyssa> uh oh
<alyssa> hw?
<rasterman> midgard
<alyssa> flavour? scene?
<rasterman> instantly on start
<alyssa> eyes
<rasterman> doesnt even get a window up
<alyssa> wat..
<alyssa> k, need to stuff my face with food then will look
<rasterman> yeah... wat... :)
<rasterman> i have not debugged yet beyond that
<rasterman> e works. efl stuff works.. glmark.. nope
<rasterman> https://termbin.com/r682\there better
<rasterman> full bt no wraps :)
<rasterman> fr 0
<rasterman> oops
<rasterman> instr is garbage
<rasterman> 1038 switch (instr->type) {
<rasterman> and
<rasterman> (gdb) p instr
<rasterman> $1 = (nir_instr *) 0x8e
<rasterman> waaaait... how does this not barf in frame 3 when it does
<rasterman> 277 nir_builder_instr_insert(build, &load_const->instr);
camus has joined #panfrost
<rasterman> don't tell me the stack is junk too... :|
camus1 has quit [Ping timeout: 480 seconds]
<alyssa> erm
<alyssa> valgrind?
<rasterman> doing that right now
<alyssa> thx
<rasterman> same story actually
<rasterman> fr 3...
<rasterman> 277 nir_builder_instr_insert(build, &load_const->instr);
<rasterman> (gdb) p load_const
<rasterman> $2 = (nir_load_const_instr *) 0x8e
<rasterman> wtf?
<rasterman> sorry brain only 10% on this - on zoom call
ChristinaAguiler has joined #panfrost
ChristinaAguiler has quit [Remote host closed the connection]
WoC has joined #panfrost
<rasterman> oh wait never mind.. union. thus why ptr is the same
<rasterman> and this means rzalloc is returning junk.. which leads me right back to wtf?
<rasterman> well rzalloc_size ...
Manderson has joined #panfrost
Manderson is now known as Guest2272
Guest2272 has quit [Remote host closed the connection]
<rasterman> sorry ... added stuff on line
Lightsword23 has joined #panfrost
Lightsword23 has quit [Remote host closed the connection]
WoC has quit [Remote host closed the connection]
<rasterman> and no with -O0 ... bug is somewhere else ... pan_unpack returns junk
<rasterman> format is PIPE_FORMAT_B10G10R10A2_UNORM
<rasterman> 1x1, depth 1, bits 32....
<rasterman> {format = PIPE_FORMAT_B10G10R10A2_UNORM, name = 0xfffff71fff98 "PIPE_FORMAT_B10G10R10A2_UNORM", short_name = 0xfffff71fffb8 "b10g10r10a2_unorm", block = {width = 1, height = 1, depth = 1, bits = 32}, layout = UTIL_FORMAT_LAYOUT_PLAIN, nr_channels = 4, is_array = 0, is_bitmask = 1, is_mixed = 0,
<rasterman> shift = 30}}, swizzle = "\002\001\000\003", colorspace = UTIL_FORMAT_COLORSPACE_RGB}
<rasterman> is_unorm = 1, is_snorm = 0, channel = {{type = 1, normalized = 1, pure_integer = 0, size = 10, shift = 0}, {type = 1, normalized = 1, pure_integer = 0, size = 10, shift = 10}, {type = 1, normalized = 1, pure_integer = 0, size = 10, shift = 20}, {type = 1, normalized = 1, pure_integer = 0, size = 2,
<rasterman> hmmm
<rasterman> wasaaait... the format handling doesnt handle this!
<rasterman> it handles PIPE_FORMAT_R10G10B10A2_UNORM but not BGR... :)
<rasterman> but wait. unreachable() should have probably aborted?
<alyssa> rasterman: unreachable is compiled out in release builds iirc
<rasterman> oh ha!
<rasterman> the fprintf could have been better than just printing the name as i SEE
<rasterman> [build] use-vbo=false:PIPE_FORMAT_B10G10R10A2_UNORM
<rasterman> :)
<rasterman> why do i see this problem and not you?
<alyssa> e maybe?
<rasterman> well this is the client segfaulting...
<rasterman> glmark
<alyssa> right but you're running it under e
<alyssa> BGR10A2 is the window format the compositor negotiated
<alyssa> my compositor probably does BGRA8
<rasterman> might change the eglconfig backbuffer fmt chosen by mesa
<rasterman> but i didnt think e would try use this...
<alyssa> anyway been working on that file today
<rasterman> maybe because it doesnt use an alpha channel and thus is promoting to highest bit depth possible?
<alyssa> this is a good retroactive excuse
<rasterman> everythgin else i have uses a full alpha (eg for shadow area) so thus probably gets BGRA8888
<rasterman> really this just needs handlers for pack/unpack unorm_bgr_1010102
<alyssa> yeah sec
<alyssa> i've been fixing that whole file to not suck
<alyssa> applying Delete The Code methodology(TM)
<alyssa> 15 files changed, 545 insertions(+), 911 deletions(-)
<rasterman> hehehe well may as well throw this in too :)
<rasterman> oh wait...
<rasterman> i know... e prefers BGRA not RGBA texture formats
<rasterman> always.... and i note that bgr is not handled there - just the rgb ones
<rasterman> so you probably do get 10bit ... but rgb instead of bgr.
<rasterman> (e/efl prefer bgr because traditionally outside of big endian sgi land .. bgr was by far the more common colorspace supported by pc hardware)
<alyssa> rasterman: alyssa/mesa:better-fb is totally untested
<rasterman> an mr?
<alyssa> will come once it's tested
<alyssa> kinda nervous it'll break skadi
<rasterman> so where is this magic branch?
<alyssa> alyssa/mesa:better-fb
camus1 has joined #panfrost
<rasterman> but where? it's not a mr - i can't find it. it's not a branch in mesa... ?
camus has quit [Ping timeout: 480 seconds]
<robmur01> break Skadi> isn't she already broken? (#4795)
<robmur01> (I massively corrupted my Juno's filesystem so haven't had a chance to test anything since then)
<rasterman> hahaha ... a broken juno. never seen that before...
<rasterman> robmur01: why cant gitlab find that?
<rasterman> i was searching for it... :|
<alyssa> robmur01: are you giving me areason to break her more
<rasterman> alyssa: cang thats a lot of commits on top.
<alyssa> robmur01: I thought I fixed that bug
<robmur01> quite possibly - I'd like to play with the conformance fixes but I basically need to rescue some stuff and reinstall from scratch
* robmur01 might cave in and build a Debian FS in the hope of getting deqp to actually work
<rasterman> alyssa: yeah.. just some more case:'s to handle BGR there works.
<rasterman> just threw those in. as you've already got a fat lot of changes there... just throw those extra cases in for bgr
<alyssa> robmur01: conformance fixes were for es3.1 on freya
<alyssa> don't get your hopes up :-p
<alyssa> (and, er, bifrost)
<rasterman> gah
<rasterman> glmark terrain is broken again... its' all flat
<robmur01> yeah, but, trickle down... you never intended to make T620 work better, but you *have* :P
<rasterman> this was broken, then fixed.. and now broken again :)
<macc24> o/ cphealy
<alyssa> robmur01: Darn! :-p
<alyssa> terrain works fine for me
<robmur01> sanity check: definitely not an old glmark without the mediump/FP16 bugs fixed?
<rasterman> alyssa: https://termbin.com/cimu <- that is all that's needed to fix .. well the segv. but other bugs like the terrain stuff are some nir/compiler bug methinks
<alyssa> the bgr10a2_uint case is wrong
<alyssa> but the unorm one is right
<alyssa> i think
<alyssa> old glmark
<rasterman> hmmm i compiled this one not that long ago...
<rasterman> i think
<rasterman> unless that was another one of my boards/systems... i lose track
<rasterman> oh wait
<rasterman> it IS old...
<rasterman> hmmm then another onwe of my systems i built a newer one ...
<rasterman> but i saw this problem before and fixed it... :/
<rasterman> dang... i have not run glmark outside of a mali in a long time...
<rasterman> alyssa: no- latest git master glmark - same terrain problem
<rasterman> i suspect u have a fix in your tree that i dont have
PocketKnife has joined #panfrost
PocketKnife has quit [autokilled: Suspected spammer. Mail support@oftc.net with questions (2021-06-15 18:12:28)]
alteredbeast has joined #panfrost
alteredbeast has quit [Remote host closed the connection]
<rasterman> wait a xsec.. wth.. it didnt rebuild glmark?
<rasterman> too many ssh sessions.... i got confused :)
<rasterman> gagh .. out of disk space
WoC has joined #panfrost
rasterman has quit [Quit: Gettin' stinky!]
keith_miller has joined #panfrost
keith_miller has quit [Remote host closed the connection]
rasterman has joined #panfrost
WoC has quit [Remote host closed the connection]
warpme_ has quit [Quit: Connection closed for inactivity]
MinimalSpace has joined #panfrost
camus has joined #panfrost
camus1 has quit [Ping timeout: 480 seconds]
DmitriyMarkov-t has joined #panfrost
DmitriyMarkov-t has quit [Remote host closed the connection]
thecycoone has quit []
thecycoone has joined #panfrost
camus1 has joined #panfrost
camus has quit [Read error: Connection reset by peer]
thecycoone has quit []
thecycoone has joined #panfrost
MinimalSpace has quit [Ping timeout: 480 seconds]
Jeff-M has joined #panfrost
Jeff-M has quit [Remote host closed the connection]
MinimalSpace has joined #panfrost
andrey-konovalov has quit [Ping timeout: 480 seconds]
camus has joined #panfrost
camus1 has quit [Ping timeout: 480 seconds]
Majiir has joined #panfrost
Majiir has quit [Remote host closed the connection]
_ has joined #panfrost
_ has quit [Remote host closed the connection]
MinimalSpace has quit [Ping timeout: 480 seconds]
MinimalSpace has joined #panfrost
martineau has joined #panfrost
martineau has quit [Remote host closed the connection]
pastly-antispam has quit [Quit: WeeChat 2.3]
ntninja-M has joined #panfrost
ntninja-M has quit [autokilled: Possible spambot. Mail support@oftc.net if you think this is in error. (2021-06-15 23:42:11)]
rasterman has quit [Quit: Gettin' stinky!]