alexcrichton requested dicej for a review on PR #10011.
alexcrichton opened PR #10011 from alexcrichton:emit-optimized-clif
to bytecodealliance:main
:
This commit updates the implementation of the
wasmtime
CLI flag--emit-clif
to emit the IR post-optimization rather than pre-optimization. This is now emitting the IR that's directly fed to the backend for lowering which reflects any mid-level transformation that's performed. It should still be possible to recover pre-optimized CLIF with-O opt-level=0
.This then additionally refactors functions to emit CLIF in a more "core" location so all trampolines get their CLIF emitted as well in case those are needed for debugging.
<!--
Please make sure you include the following information:
If this work has been discussed elsewhere, please include a link to that
conversation. If it was discussed in an issue, just mention "issue #...".Explain why this change is needed. If the details are in an issue already,
this can be brief.Our development process is documented in the Wasmtime book:
https://docs.wasmtime.dev/contributing-development-process.htmlPlease ensure all communication follows the code of conduct:
https://github.com/bytecodealliance/wasmtime/blob/main/CODE_OF_CONDUCT.md
-->
alexcrichton requested wasmtime-core-reviewers for a review on PR #10011.
dicej submitted PR review.
alexcrichton updated PR #10011.
alexcrichton commented on PR #10011:
Looks like all
disas
tests are changing notably for a few minor differences:
- Running through the compilation pipeline removes all
global_value
instructions (replaces them with their equivalents)- Unnecessary block params are removed
alexcrichton has enabled auto merge for PR #10011.
alexcrichton merged PR #10011.
bjorn3 commented on PR #10011:
It should still be possible to recover pre-optimized CLIF with -O opt-level=0.
Doesn't that still perform legalizations?
alexcrichton commented on PR #10011:
That's true yeah,
--emit-clif
won't work for that any more
Last updated: Jan 24 2025 at 00:11 UTC