ChanServ changed the topic of #etnaviv to: #etnaviv - the home of the reverse-engineered Vivante GPU driver - Logs https://oftc.irclog.whitequark.org/etnaviv
otavio_ has quit [Remote host closed the connection]
otavio_ has joined #etnaviv
karolherbst_ has joined #etnaviv
karolherbst has quit [Ping timeout: 480 seconds]
Leopold_ has joined #etnaviv
chewitt has joined #etnaviv
mvlad has joined #etnaviv
otavio_ has quit [Ping timeout: 480 seconds]
pcercuei has joined #etnaviv
lynxeye has joined #etnaviv
<tomeu> this may be in the top 10 most beautiful changelogs I have ever crafted:
<tomeu> src/etnaviv/ci/etnaviv-rusticl-fails.txt | 303 ----------------------------------------------------------------------------------------------
<tomeu> src/etnaviv/ci/etnaviv-rusticl-flakes.txt | 52 ----------------
<tomeu> src/gallium/drivers/etnaviv/etnaviv_compiler_nir_emit.c | 14 +++++
<lynxeye> oh wow
<tomeu> actually, that is really too true to be believed
<tomeu> I'm undoing some changes in piglit and retesting right now :(
<tomeu> was really happy for a couple of minutes :)
lynxeye has quit [Ping timeout: 480 seconds]
Leopold_ has quit [Remote host closed the connection]
Leopold_ has joined #etnaviv
lynxeye has joined #etnaviv
pcercuei has quit [Quit: Lost terminal]
pcercuei has joined #etnaviv
karolherbst_ is now known as karolherbst
otavio_ has joined #etnaviv
<tomeu> well, it fixes 100 tests, not 300
<tomeu> but I still need to investigate further integer conversions
<tomeu> i2i seems to, at least on this NPU, have some kind of mask in src1, as an immediate
<tomeu> the 100 fixed tests are after handling unsigned integers, tomorrow will see how many more are fixed after handing signed ones
<tomeu> what I find best though is that only 4 flaky tests remain after this change
otavio_ has quit [Remote host closed the connection]
pcercuei has quit [Quit: brb]
pcercuei has joined #etnaviv
karolherbst_ has joined #etnaviv
karolherbst has quit [Ping timeout: 480 seconds]
<rah> I managed to trigger the crash again, some photos:
<cphealy> rah: Can you try taking screenshots to see if the corruption is present there?
<cphealy> Once it behaves this way, are you back to everything working if you close the app?
<rah> cphealy: it's locked hard
<rah> sorry I didn't mention that here
<rah> I'm getting a hard lock with display corruption
<rah> those are photos of the machine hard locked
<cphealy> Ahh, got it. When you say locked hard, do you mean the only solution is a hard reboot? What if you enable watchdog?
<rah> yes, the only solution is a hard reboot
<mntirc> rah: i haven't seen this type of corruption yet. do you already have a ssh server set up on the machine? perhaps you can watch the kernel log etc from another machine
<mntirc> and maybe only the desktop is locking up
<mntirc> another thing to try to isolate whether it's gpu related is to run the app with LIBGL_ALWAYS_SOFTWARE=1
<rah> mntirc: there is an SSH server and it becomes non-responsive; there's also nothing in the logs after restarting the machine
<cphealy> Sounds like possibly a hard lock of the SoC.
<mntirc> rah: good to know. i though more like running `dmesg -w`
<mntirc> maybe you would still see something like "mmu fault" from the driver
<rah> mntirc: I understood what you meant
lynxeye has quit [Quit: Leaving.]
karolherbst_ is now known as karolherbst
mvlad has quit [Remote host closed the connection]
JohnnyonF has joined #etnaviv
JohnnyonF has quit [Read error: Connection reset by peer]
JohnnyonFlame has quit [Ping timeout: 480 seconds]
Leopold_ has quit [Remote host closed the connection]
Leopold has joined #etnaviv
pcercuei has quit [Quit: dod]