diff options
author | Ayke van Laethem <[email protected]> | 2022-10-16 00:44:38 +0200 |
---|---|---|
committer | Ron Evans <[email protected]> | 2022-10-19 18:36:53 +0200 |
commit | 65d65c131345e47581835e76df5fd22fced15158 (patch) | |
tree | f3d70027c1c993959ba6657143d0ebdf3a123cea /targets | |
parent | 6b46ae261a7709d893ced7c831c4ef24a4c16d4f (diff) | |
download | tinygo-65d65c131345e47581835e76df5fd22fced15158.tar.gz tinygo-65d65c131345e47581835e76df5fd22fced15158.zip |
wasm: fix GC scanning of allocas
Scanning of allocas was entirely broken on WebAssembly. The code
intended to do this was never run. There were also no tests.
Looking into this further, I found that it is actually not really
necessary to do that: the C stack can be scanned conservatively and in
fact this was already done for goroutine stacks (because they live on
the heap and are always referenced). It wasn't done for the system stack
however.
With these fixes, I believe code should be both faster *and* more
correct.
I found this in my work to get opaque pointers supported in LLVM 15,
because the code that was never reached now finally got run and was
actually quite buggy.
Diffstat (limited to 'targets')
-rw-r--r-- | targets/wasi.json | 3 | ||||
-rw-r--r-- | targets/wasm.json | 3 |
2 files changed, 6 insertions, 0 deletions
diff --git a/targets/wasi.json b/targets/wasi.json index e710b4bb1..6cec6be47 100644 --- a/targets/wasi.json +++ b/targets/wasi.json @@ -18,6 +18,9 @@ "--stack-first", "--no-demangle" ], + "extra-files": [ + "src/runtime/asm_tinygowasm.S" + ], "emulator": "wasmtime {}", "wasm-abi": "generic" } diff --git a/targets/wasm.json b/targets/wasm.json index 2bcada5fe..26494cc45 100644 --- a/targets/wasm.json +++ b/targets/wasm.json @@ -19,6 +19,9 @@ "--stack-first", "--no-demangle" ], + "extra-files": [ + "src/runtime/asm_tinygowasm.S" + ], "emulator": "node {root}/targets/wasm_exec.js {}", "wasm-abi": "js" } |