ChanServ changed the topic of #zink to: official development channel for the mesa3d zink driver || https://docs.mesa3d.org/drivers/zink.html
hikiko has joined #zink
<anholt> zmike: asan was fine with the local-size test, so I don't actually think it's heap corruption
<zmike> anholt: yep, like I said, I've never been able to catch it at a point where it can be debugged
<anholt> there are leaks at the end of that program, which makes me wonder if we have any threads still running as other stuff is being torn down.
<zmike> the backtraces I've seen are from asserts in vk object
<zmike> one of the more frustrating problems I've seen...it's been around for at least a year I think