Stream: git-wasmtime

Topic: wasmtime / PR #13525 wasmtime: Use fine-grained alias reg...

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 18:04):

fitzgen opened PR #13525 from fitzgen:fine-grained-memory-global-table-alias-regions to bytecodealliance:main:

Defines an AliasRegionKey that gets turned into a Cranelift ir::AliasRegionData and yields a different region per defined memory/global/table vs GC heap vs vmctx structures, etc...

Relatively small diff in reality, only a few hundred lines, but the disas test expectations all got updated.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 18:04):

fitzgen requested cfallin for a review on PR #13525.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 18:04):

fitzgen requested wasmtime-compiler-reviewers for a review on PR #13525.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 18:04):

fitzgen requested wasmtime-core-reviewers for a review on PR #13525.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 19:32):

:thumbs_up: cfallin submitted PR review:

Looks great! One question about new tests but otherwise LGTM.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 19:32):

:speech_balloon: cfallin created PR review comment:

Since we now have separate regions per memory, do we have a test that shows e.g. store-to-load forwarding working with interleaved stores to distinct memories in a multi-memory guest? (I skimmd over the list of tests touched here but it's huge; curious if you explicitly probed this)

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 19:32):

:speech_balloon: cfallin edited PR review comment.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 19:52):

:memo: fitzgen submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 19:52):

:speech_balloon: fitzgen created PR review comment:

We have basically this at the Cranelift filetest level, but not at the Wasm level. I can add a new disas test.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 20:17):

:memo: fitzgen submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 20:17):

:speech_balloon: fitzgen created PR review comment:

Ah wait, yes, I already added such a test: tests/disas/alias-region-memories.wat

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 20:57):

fitzgen updated PR #13525.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 20:57):

fitzgen has enabled auto merge for PR #13525.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 21:59):

fitzgen updated PR #13525.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 22:11):

fitzgen updated PR #13525.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 22:21):

fitzgen added PR #13525 wasmtime: Use fine-grained alias regions for memory, globals, tables, vmctx, and GC heap to the merge queue.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 22:45):

:check: fitzgen merged PR #13525.

view this post on Zulip Wasmtime GitHub notifications bot (Jun 01 2026 at 22:45):

fitzgen removed PR #13525 wasmtime: Use fine-grained alias regions for memory, globals, tables, vmctx, and GC heap from the merge queue.

Last updated: Jul 29 2026 at 05:03 UTC