Monday after spring break right back getting dragged into (gpu) memory management fun. I guess time to re-drop my tongue-in-cheek take on this:
memory management is trivial, right until you add reclaim
then it becomes impossible
Remote
🐧sima🐧
@sima@chaos.social
mastodon 4.6.9relapsing linux kernel maintainer or something
formerly @danvet on birdsite, she/they mostly, queer af & adhd neurospicy 🐧 🦈 🦀
toots are a mix of shitposting, despair about tech I work on my job under #kernel, mostly good self-made #food, vaguely deniable shark posting and the oddball very personal thing that I need to get out
0 Followers
0 Following
8 Posts
Joined November 01, 2022
Blog:
Replying to
@airlied dreading that thread, but I think I need to jump in to write Miguel a thank you note
@drakulix I'm afraid, the answer to that is "no", at least for the kernel at large
I'm more hopeful than before all this that maybe it could become a "yes" for at least some more subsystems. which would at least help us to keep dri-devel a safer place. still not safe enough I'm afraid, but one small step at a time and all that
Replying to
@drakulix I learned a few hard lessons about myself, one is that I cannot protect or even provide much help to people who look at the kernel and go "I can fix this"
to avoid unecessary pain I've decided to be very blunt and open about how I see things, to avoid any disappointment due to misplaced hopes
Replying to
@alwayscurious@infosec.exchange @julienbarnoin@mastodon.gamedev.place this was a much more general observation I've made all across various issues. the upstream dma_fence rules are just one example. and it's not even the only issue that plagues gpu memory management reclaim
Replying to
@brauner@mastodon.social spring break, so not
but link for my entertainment and suffering?