Stream: git-wasmtime

Topic: wasmtime / PR #2157 Deduplicate for_each_libcall, assert,...

Wasmtime GitHub notifications bot (Aug 21 2020 at 21:30):

yurydelendik opened PR #2157 from libcalls-wasmtime to main:

We don't have listed all Libcalls in the for_each_libcall at the moment. (And I noticed Libcall::Probestack reloc record in one of the large WASI programs)

This patch:

• [x] removes for_each_libcall macro from the obj crate
• [x] adds __assert_all_for_each_libcalls to catch missing Libcall (and warns for duplicate)
• [x] Implement wasmtime_memXXX and stub rest

Not sure what are the plans. Can we use stuff from the compiler-builtins crate here?

Wasmtime GitHub notifications bot (Aug 21 2020 at 21:36):

bjorn3 submitted PR Review.

Wasmtime GitHub notifications bot (Aug 21 2020 at 21:36):

bjorn3 submitted PR Review.

Wasmtime GitHub notifications bot (Aug 21 2020 at 21:36):

bjorn3 created PR Review Comment:

This can be unreachable!(). Wasmtime doesn't use TLS in the jitted code.

Wasmtime GitHub notifications bot (Aug 21 2020 at 21:36):

bjorn3 created PR Review Comment:

You would have to use something like extern "C" { #[linkage_name = "__rust_probestack"] fn wasmtime_probestack(); }. The abi of probestack is incompatible with C, so using a wrapper function with abi C won't work.

Wasmtime GitHub notifications bot (Aug 24 2020 at 16:06):

alexcrichton submitted PR Review.

Wasmtime GitHub notifications bot (Aug 24 2020 at 16:06):

alexcrichton submitted PR Review.

Wasmtime GitHub notifications bot (Aug 24 2020 at 16:06):

alexcrichton created PR Review Comment:

Could this actually get removed since I don't think it's used anywhere any more?

Wasmtime GitHub notifications bot (Aug 24 2020 at 16:06):

alexcrichton created PR Review Comment:

Out of curiosity, where do these get generated in cranelift?

Wasmtime GitHub notifications bot (Aug 24 2020 at 16:13):

bjorn3 submitted PR Review.

Wasmtime GitHub notifications bot (Aug 24 2020 at 16:13):

bjorn3 created PR Review Comment:

I assume they will be generated when implementing the bulk memory wasm proposal.

Wasmtime GitHub notifications bot (Aug 24 2020 at 17:04):

yurydelendik submitted PR Review.

Wasmtime GitHub notifications bot (Aug 24 2020 at 17:04):

yurydelendik created PR Review Comment:

FWIW I'm getting Libcall::Probestack relocation entries in large code base. (I'll find out where exactly it is coming from.)

Wasmtime GitHub notifications bot (Mar 13 2024 at 02:18):

alexcrichton closed without merge PR #2157.

Wasmtime GitHub notifications bot (Mar 13 2024 at 02:18):

alexcrichton commented on PR #2157:

This PR is pretty old in Wasmtime now and would need a good bit of rebasing to get it applying again. Enough time has passed as well that enough has changed under the hood, so I'm going to close this

Last updated: Nov 22 2024 at 16:03 UTC