alyssa changed the topic of #panfrost to: Panfrost - FLOSS Mali Midgard & Bifrost - https://gitlab.freedesktop.org/panfrost - Logs https://freenode.irclog.whitequark.org/panfrost - <daniels> avoiding X is a huge feature
davidlt has quit [Remote host closed the connection]
davidlt has joined #panfrost
davidlt has quit [Read error: Connection reset by peer]
davidlt has joined #panfrost
vstehle has quit [Ping timeout: 245 seconds]
davidlt has quit [Remote host closed the connection]
davidlt has joined #panfrost
davidlt has quit [Read error: Connection reset by peer]
davidlt has joined #panfrost
davidlt has quit [Remote host closed the connection]
davidlt has joined #panfrost
forkbomb has quit [Remote host closed the connection]
forkbomb has joined #panfrost
vstehle has joined #panfrost
Elpaulo has quit [Read error: Connection reset by peer]
Elpaulo has joined #panfrost
davidlt has quit [Ping timeout: 245 seconds]
davidlt has joined #panfrost
davidlt_ has joined #panfrost
davidlt has quit [Read error: Connection reset by peer]
pH5 has joined #panfrost
raster has joined #panfrost
sphalerite_ has joined #panfrost
davidlt has joined #panfrost
davidlt_ has quit [Read error: Connection reset by peer]
<tomeu> Prf_Jakob: awesome, thanks
davidlt has quit [Ping timeout: 268 seconds]
<rcf> EmilKarlson: maybe now is the time to port alacritty to gles....
<HdkR> <3 alacritty
<EmilKarlson> well I don't need super speed, just to be able to read normal dmesg output without pager
<rcf> I have had good luck with the vte terminals (of which gnome-terminal is one).
<EmilKarlson> I would rather have sandboxed terminal that can't access anything
<EmilKarlson> yes
<EmilKarlson> lxterminal seems to be almost exact copy of gnome-terminal for an external observer
<tomeu> Prf_Jakob: regarding --print-failing, it seems to default to true
sphalerite has quit [Quit: WeeChat 2.4]
sphalerite_ is now known as sphalerite
davidlt has joined #panfrost
davidlt has quit [Ping timeout: 248 seconds]
davidlt has joined #panfrost
davidlt has quit [Ping timeout: 245 seconds]
_whitelogger has joined #panfrost
ckeepax has quit [Quit: WeeChat 2.2]
<tomeu> cross compiling to aarch64-linux
<tomeu> error: No valid LLVM Toolchains found!
<tomeu> Prf_Jakob: could you check what I could be missing in https://gitlab.freedesktop.org/tomeu/mesa/-/jobs/513800 ?
ckeepax has joined #panfrost
<tomeu> Prf_Jakob: this run should match closer what virgl does, but still no luck: https://gitlab.freedesktop.org/tomeu/mesa/-/jobs/513851
cwabbott has quit [Quit: cwabbott]
cwabbott has joined #panfrost
<Prf_Jakob> tomeu: I'll take a look
<Prf_Jakob> tomeu: Default seems to be false
<Prf_Jakob> Ooh, it's because of the regression file it prints it out.
<Prf_Jakob> That's a easy fix
<tomeu> Prf_Jakob: nice
<tomeu> made some progress with building volt stuff: https://gitlab.freedesktop.org/tomeu/mesa/-/jobs/514049
<tomeu> guess gdc is linking to the wrong llvm version
<Prf_Jakob> It's not linking it at all? That's odd
<tomeu> Prf_Jakob: had to play with PATH for battery to be find llvm-config
adjtm_ has quit [Ping timeout: 248 seconds]
<tomeu> do I need to do something similar to LD_LIBRARY_PATH?
<Prf_Jakob> Why isn't llvm just installed as normally?
<tomeu> Prf_Jakob: how ould be normally?
<Prf_Jakob> The debian package?
<tomeu> Prf_Jakob: this are all debian packages
<Prf_Jakob> Then why are you setting path to '/usr/lib/llvm-8/bin'?
<tomeu> Prf_Jakob: because battery seems to search for llvm-config in the default PATH, which is only installed by llvm-7
<Prf_Jakob> Aaah don't install the 'llvm' package, only install llvm-7 and llvm-8
<Prf_Jakob> Or well only the llvm-X packages you need
<Prf_Jakob> But that shouldn't matter tho
<tomeu> Prf_Jakob: afaics, the llvm package isn't being installed : https://gitlab.freedesktop.org/tomeu/mesa/-/jobs/514049
megi has joined #panfrost
<Prf_Jakob> Huh... that's interesting.
<Prf_Jakob> tomeu: "llvm-8-dev" is not installed
<tomeu> ok, trying with it
<tomeu> I'm not setting PATH, and battery is still able to find llvm stuff
<tomeu> but I'm still getting a error: No valid LLVM Toolchains found!
<tomeu> without further info
<Prf_Jakob> There is a .battery/config-log.txt file
<Prf_Jakob> tomeu: New release of battery 0.1.22, minor fixes. Including writing out where the log file is. :)
unoccupied has quit [Ping timeout: 246 seconds]
davidlt has joined #panfrost
<tomeu> WARNING: possible circular locking dependency detected
<tomeu> robher: not sure what happened later: https://lava.collabora.co.uk/scheduler/job/1769192
forkbomb has quit [Quit: In the beginning the Universe was created. This has made a lot of people very angry and been widely regarded as a bad move.]
forkbomb has joined #panfrost
jolan has quit [Quit: leaving]
jolan has joined #panfrost
<robher> tomeu: looks like a regulator issue, not something in panfrost.
adjtm has joined #panfrost
davidlt has quit [Ping timeout: 272 seconds]
gcl has joined #panfrost
unoccupied has joined #panfrost
tgall_fo_ has joined #panfrost
tgall_foo has quit [Ping timeout: 276 seconds]
<tomeu> robher: got the same lockdep warning on 5.2.0
<tomeu> so I guess the hang is unrelated
<robher> tomeu: I get a long pause booting with lockdep enabled, but it happens much earlier for me.
<tomeu> it's still hanging on 5.2 with lockdep: https://lava.collabora.co.uk/scheduler/job/1769301
<tomeu> that has CONFIG_DETECT_HUNG_TASK, so I hope I get some details shortly
<tomeu> haven't seen that one ye
<tomeu> *yet
davidlt has joined #panfrost
raster has quit [Remote host closed the connection]
pH5 has quit [Quit: bye]
pH5 has joined #panfrost
tgall_fo_ has quit [Read error: Connection reset by peer]
<shadeslayer> frick yeah, progress
jcureton has joined #panfrost
<alyssa> I like horses.
<alyssa> Neigh me.
<shadeslayer> ok, so at least with both the BO Cache patches AND the MIR for_next patches, plasma shows up and then leads to this https://paste.ubuntu.com/p/2PXmTpYfhK/
<shadeslayer> so, at least, some progress from last week :)
stikonas has joined #panfrost
<daniels> shadeslayer: valgrind again? :P
stikonas has quit [Read error: Connection reset by peer]
stikonas_ has joined #panfrost
stikonas_ has quit [Remote host closed the connection]
stikonas has joined #panfrost
<shadeslayer> daniels: that's tricky since this crash happens in Plasma Shell
<shadeslayer> so running with valgrind just takes forever
<daniels> heh
<shadeslayer> daniels: do you reckon something like building with address sanitizer would help here
<shadeslayer> meh, nvm, it would if it was supported on ARM64 :/
<daniels> hmm, the problem with ASan is that it also doesn't know about ralloc
<shadeslayer> oh :(
<HdkR> Would be nice if asan had a way to inject some handlers for custom userspace memory allocators
* alyssa is down the tiler rabbit hole too deep
<alyssa> Abort
<alyssa> But I think I got something correct
<alyssa> ....Just can't figure out for the life of me why it works
* alyssa spoke too soon
<alyssa> Okay what in the name of equestria is going on here.
<alyssa> This hardware is playing games with me
<alyssa> OKAY this is NOT okay
<alyssa> There's a 4096x4096 level, isn't there, Midgard?
<alyssa> yOu are cRazy
jcureton has quit [Ping timeout: 260 seconds]
<alyssa> Tada!
<alyssa> Got the computation correct!!!
<alyssa> Specifically, for every scene in glmark --validate,
<alyssa> Given a (width, height, mask) tuple, we compute bit-identical sizes compared to the blob
<anarsoul> alyssa: congrats :)
<alyssa> anarsoul: heh, thanks :)
<alyssa> anarsoul: I'm on a mission to eliminate command stream bugs for good.
<alyssa> Mostly because the hardware fault reporting doesn't give enough detail for effective debugging, so I'm writing my own tooling
<alyssa> I wouldn't be surprised if the vendors have tools like this internally
<anarsoul> alyssa: nice
<anarsoul> I've discovered nasty bug on lima recently (after I finally got glamor working)
<alyssa> Ah?
<anarsoul> see how glxgears freezes when xterm window is moved and then tries to catch up
<alyssa> Super glitchy but I think that's a panfrost bug on my end :blush:
<alyssa> Hrmm
stikonas has quit [Ping timeout: 252 seconds]