#asahi · 2023-04-08 18:11
respIs there any further reading for KVM about this?#asahi · 2023-04-08 18:09
respWhat confuses me is the KVM part, don't guest kernels have to share a page size with the host?#asahi · 2023-04-08 18:08
respThe idea is just a stop gap until we can do it with the Linux userland, not a permanent solution #asahi · 2023-04-08 18:07
respOh no I understand on macOS you can do it, It's just under Linux we can't mix and match page sizes#asahi · 2023-04-08 17:48
respThe qemu idea isn't a solution longterm, but it would be a stop gap if it could be simplified into a script to wrap a program that expects 4k pages.#asahi · 2023-04-08 17:48
respI desperately want to play around with qemu userland emulation for AARC64 4k pages on AARCH64 16k page size hosts#asahi · 2023-04-08 17:47
respI can't afford to mess up my macOS install.#asahi · 2023-04-08 17:47
respI can't wait for my college to end for the year so I can safely play around with Asahi#asahi · 2023-04-08 17:45
respI can wrap my mind around QEMU emulating a different ARM CPU, but just hosting a guest kernel with different page sizes?#asahi · 2023-04-08 17:44
respI'm curious, how is that possible?#asahi · 2023-04-08 17:43
respIt really can't be worse than running FEX with x86->ARM translation #asahi · 2023-04-08 17:42
respHorribly inefficient. But its a workaround.#asahi · 2023-04-08 17:42
respemulate a 4k page ARM system on Asahi and pass a x11 socket through to the host over a virtual bridge.#asahi · 2023-04-08 17:41
respFor the people who desperately need 4k page expectign software on M1 why not just run it in qemu?#asahi · 2023-04-08 17:39
respBut I had a probably *very stupid idea* #asahi · 2023-04-08 17:38
respI'm not sure if I should talk in here or #asahi-offtopic as it relates to Asahi