Berrysoft opened issue #4992:
After #4990 merged, we can build wasmtime for Windows ARM64, and it should work without trap handling.
Berrysoft edited issue #4992:
After #4990 merged, we can build wasmtime for Windows ARM64, and it should work without trap handling.
We now need to add trap handling to make it totally work.
Berrysoft commented on issue #4992:
https://github.com/bytecodealliance/wasmtime/blob/main/cranelift/codegen/src/isa/aarch64/mod.rs#L152
Berrysoft edited a comment on issue #4992:
Berrysoft edited a comment on issue #4992:
alexcrichton labeled issue #4992:
After #4990 merged, we can build wasmtime for Windows ARM64, and it should work without trap handling.
We now need to add trap handling to make it totally work.
hwoodiwiss commented on issue #4992:
Is this something that's being looked at actively at the moment, and would a community contributions be welcome for this?
hwoodiwiss edited a comment on issue #4992:
Is this something that's being looked at actively at the moment, and would a community contribution be welcome?
cfallin commented on issue #4992:
No one has indicated they're working on it, so if you or anyone who is familiar with Windows on Arm is willing to build it, we'd very much welcome it!
dpaoliello commented on issue #4992:
It's hackathon week at Microsoft, so I'm going to have a go at this as my project...
alexcrichton commented on issue #4992:
Oh nice, that'd be great! If you have any questions feel free to reach out here or on Zulip
dpaoliello commented on issue #4992:
I have a draft PR: #9266
Currently most tests are passing, but there is a crash in the JITted code. Unfortunately, there doesn't seem to be debug info (or stack is corrupted just for the JIT frames) and WinDbg is crashing when I try to capture a Time Travel Trace.
dpaoliello edited a comment on issue #4992:
I have a draft PR: #9266
Update: I misread the stack, looks like it's a crash due to a "malformed function table during unwind" - investigating further...
dpaoliello commented on issue #4992:
Update: I got everything working (required a more careful reading of the spec and how wasmtime emits instructions).
PR has been published.
cfallin commented on issue #4992:
Very exciting, thanks @dpaoliello! It looks like Alex has looked at the PR at a high level; I'll do a detailed review.
cfallin closed issue #4992:
After #4990 merged, we can build wasmtime for Windows ARM64, and it should work without trap handling.
We now need to add trap handling to make it totally work.
kpreisser commented on issue #4992:
Thank you very much for implementing unwind support for Windows Arm64, and providing release artifacts!
When running the unit tests ofwasmtime-dotnet
on Windows 11 Version 23H2 Arm64 (using thewasmtime-dev-aarch64-windows-c-api.zip
artifact of the latest dev release), all tests now succeed.
Last updated: Dec 23 2024 at 13:07 UTC