Stream: git-wasmtime

Topic: wasmtime / issue #13229 ISLE iconst_u doesn't support flo...

Wasmtime GitHub notifications bot (Apr 29 2026 at 13:15):

alexcrichton opened issue #13229:

This input:

test interpret
test run
set opt_level=speed
target x86_64

function %a(f64) {
block0(v0: f64):
    v11 = band_not v0, v0
    return
}

crashes with:

$ cargo run test ../out.clif
    Finished `dev` profile [unoptimized + debuginfo] target(s) in 0.12s
     Running `/home/alex/code/wasmtime/target/debug/clif-util test ../out.clif`

thread 'worker #0' (1457745) panicked at /home/alex/code/wasmtime/target/debug/build/cranelift-codegen-e399dfecead1442e/out/isle_opt.rs:25792:5:
internal error: entered unreachable code: no rule matched for term iconst_u at src/prelude_opt.isle line 165; should it be partial?
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
[2026-04-29T13:14:38Z ERROR cranelift_filetests::concurrent] FAIL: panicked in worker #0: internal error: entered unreachable code: no rule matched for term iconst_u at src/prelude_opt.isle line 165; should it be partial?
FAIL ../out.clif: panicked in worker #0: internal error: entered unreachable code: no rule matched for term iconst_u at src/prelude_opt.isle line 165; should it be partial?
1 tests
Error: 1 failure

I believe this is a regression from https://github.com/bytecodealliance/wasmtime/pull/13199 (cc @bongjunj)

Wasmtime GitHub notifications bot (Apr 29 2026 at 13:15):

alexcrichton added the fuzz-bug label to Issue #13229.

Wasmtime GitHub notifications bot (Apr 29 2026 at 13:16):

alexcrichton added the cranelift label to Issue #13229.

Wasmtime GitHub notifications bot (Apr 29 2026 at 14:38):

bongjunj commented on issue #13229:

Didn't expect float types can be matched to band op... Will look into this shortly

Wasmtime GitHub notifications bot (Apr 29 2026 at 14:52):

bongjunj commented on issue #13229:

https://github.com/bytecodealliance/wasmtime/pull/13063/changes#diff-2041f67049d5ac3d8f62ea91d3cb45cdb8608d5f5cdab988731ae2addf90ef01L50

https://github.com/bytecodealliance/wasmtime/pull/13063/changes#diff-2041f67049d5ac3d8f62ea91d3cb45cdb8608d5f5cdab988731ae2addf90ef01R386-R389

I was inspired to write the PR because, in the change above, eq, iadd, isub operations are not ty_int-constrained.
Why specifically band is causing a bug, or just we didn't observe one in eq/iadd/isub cases?

Last updated: May 03 2026 at 22:13 UTC