JohnnyonFlame has quit [Ping timeout: 480 seconds]
Surkow|laptop has quit [Ping timeout: 480 seconds]
Surkow|laptop has joined #etnaviv
frieder has joined #etnaviv
pjakobsson has quit []
pjakobsson has joined #etnaviv
lynxeye has joined #etnaviv
mvlad has joined #etnaviv
pjakobsson has quit [Remote host closed the connection]
pjakobsson has joined #etnaviv
pcercuei has joined #etnaviv
otavio_ has quit [Remote host closed the connection]
otavio_ has joined #etnaviv
frieder has quit [Ping timeout: 480 seconds]
frieder has joined #etnaviv
frieder has quit [Ping timeout: 480 seconds]
frieder has joined #etnaviv
<mntirc>
i'm trying to track down a fiddly bug where opening a file (gtk3 chooser window) in chromium causes it to crash the whole xwayland+sway stack, and i've now caught it in gdb attached to xwayland
<mntirc>
segmentation fault in 0x0000ffffa82cdd08 in etna_flush_resource () from /usr/local/lib/aarch64-linux-gnu/dri/mxsfb-drm_dri.so
frieder has quit [Remote host closed the connection]
<mntirc>
well, this is 22.2.0, will update to head first
<lynxeye>
mntirc: I think I already see the issue. With the multithreading fixes in 22.2 we got rid of some extra reference to resources, but we also don't take a proper reference to the resources tracked in flush_resources, so the framebuffer resource may already be destroyed when you get to this context flush point.
<mntirc>
lynxeye: nice, i was going to log all the pointers there but they probably still would look valid (i.e. not 0)?
<mntirc>
what i can say is that before the crash, rsc->render is not nil (but it usually always is), instead it's something like 0x800000008
<lynxeye>
mntirc: I just CCed you on a MR. Can you check if this fixes your issue?
<mntirc>
lynxeye: great, will do
<mntirc>
lynxeye: it seems to work, testing a bit more
<mntirc>
lynxeye: yes, it works, fantastic
<mntirc>
this was a big issue also for my colleague whose chromium was constantly crashing
<lynxeye>
Yay! I would feel really good about this turnaround time, if it wasn't me who introduced the bug in the first place. :/
<lynxeye>
mntirc: Thanks for the report with useful callstack!