[RyuJIT Wasm] Cast Operations Follow Up #122862
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Co-authored-by: Copilot <[email protected]>
github-actions
bot
added
the
area-CodeGen-coreclr
Contributor
|
Tagging subscribers to 'arch-wasm': @lewing, @pavelsavara |
This was referenced Jan 7, 2026
Open
…some more casting cases; Add genFloatToIntCast
adamperlin
commented
Jan 14, 2026
adamperlin
commented
Jan 14, 2026
adamperlin
commented
Jan 14, 2026
adamperlin
force-pushed
the
adamperlin/ryujit-wasm-cast
branch
from
521f043 to
c88d66b
Compare
…nt saturating cast opcodes
adamperlin
commented
Jan 14, 2026
adamperlin
commented
Jan 14, 2026
Contributor
There was a problem hiding this comment.
Pull request overview
This PR extends WebAssembly JIT cast operation support by implementing saturating float-to-int conversions and small integer casts with proper sign extension.
Changes:
- Added support for casts from small int types (byte, short) with proper sign/zero extension
- Implemented float-to-int casts using saturating non-trapping conversion instructions (requiring 2-byte opcodes)
- Extended opcode type from uint8_t to uint16_t to support 2-byte instruction encoding
- Moved GenIntCastDesc constructor out of TARGET_64BIT ifdefs to support WASM
Reviewed changes
Copilot reviewed 9 out of 9 changed files in this pull request and generated 5 comments.
Show a summary per file
| File | Description |
|---|---|
| src/coreclr/jit/morph.cpp | Skip helper call morphing for float-to-long casts on WASM (direct saturating instruction support) |
| src/coreclr/jit/instrswasm.h | Add load-with-extension, conversion, and saturating truncation instructions; align formatting |
| src/coreclr/jit/gentree.cpp | Update cost estimation for WASM cast operations |
| src/coreclr/jit/emitwasm.cpp | Change opcode type to uint16_t and add 2-byte opcode emission logic |
| src/coreclr/jit/cpp.hint | Add multiple assert macro definitions for IntelliSense |
| src/coreclr/jit/codegenwasm.cpp | Implement genIntToIntCast, genFloatToIntCast, genFloatToFloatCast; add PackTypes helper; handle TYP_INT from TYP_LONG in genCodeForLclVar |
| src/coreclr/jit/codegenlinear.cpp | Move genCodeForCast and GenIntCastDesc to support WASM; add WASM to TARGET_64BIT conditionals |
| src/coreclr/jit/codegeninterface.h | Change instInfo type to uint16_t for WASM |
| src/coreclr/jit/codegen.h | Add WASM to TARGET_64BIT conditionals in GenIntCastDesc |
AndyAyersMS
reviewed
Jan 14, 2026
adamperlin
force-pushed
the
adamperlin/ryujit-wasm-cast
branch
from
d86e911 to
cc76584
Compare
Contributor
Author
|
@dotnet/jit-contrib Initial feedback should be implemented now so this is ready for another review pass! |
kg
reviewed
Jan 16, 2026
| // | ||
| static constexpr uint32_t PackOperAndType(genTreeOps oper, var_types toType, var_types fromType) | ||
| { | ||
| if (fromType == TYP_BYREF) |
kg
reviewed
Jan 16, 2026
| // | ||
| static constexpr uint32_t PackTypes(var_types toType, var_types fromType) | ||
| { | ||
| if (toType == TYP_BYREF) |
| m_extendSrcSize = srcSize; | ||
| } | ||
|
|
||
| #ifndef TARGET_WASM |
Contributor
There was a problem hiding this comment.
Do you need this ifdef? I think things will work correctly without it (since we don't actually do any containment for casts yet).
Comment on lines
+554
to
+555
| GetEmitter()->emitIns(INS_i32_and); | ||
| ins = (toType == TYP_LONG) ? INS_i64_extend_u_i32 : INS_none; |
Contributor
There was a problem hiding this comment.
Suggested change
| GetEmitter()->emitIns(INS_i32_and); | |
| ins = (toType == TYP_LONG) ? INS_i64_extend_u_i32 : INS_none; | |
| ins = INS_i32_and; |
This type of cast always produces a TYP_INT value.
| // Return Value: | ||
| // oper and the types packed into an integer that can be used as a switch value/case | ||
| // | ||
| static constexpr uint32_t PackOperAndType(genTreeOps oper, var_types toType, var_types fromType) |
Contributor
There was a problem hiding this comment.
This variant looks unused.
| instruction ins = INS_none; | ||
| assert(fromType == TYP_INT || fromType == TYP_LONG); | ||
|
|
||
| genConsumeOperands(cast); |
Contributor
There was a problem hiding this comment.
Need an NYI for checked (gtOverflow) casts. Also applies to genFloatToIntCast.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR addresses some review feedback in #122301 and fills in more codegen for casts on Wasm.
Notably:
GenIntCastDescwas moved out of ifdefs and adapted for Wasm to support this. We don't yet support pushing down casts into loads (i.e., theLOAD_*integer cast types).