Add support for map type by yordis · Pull Request #12216 · bytecodealliance/wasmtime

yordis · 2025-12-23T21:42:43Z

Context

This is adding support for map based on WebAssembly/component-model#554

References

github-actions · 2025-12-23T23:46:00Z

Subscribe to Label Action

cc @fitzgen

Details

This issue or pull request has been labeled: "wasmtime:api", "wasmtime:c-api", "wasmtime:config", "wizer"

Thus the following users have been cc'd because of the following labels:

fitzgen: wizer

To subscribe or unsubscribe from this label, edit the .github/subscribe-to-label.json configuration file.

Learn more.

github-actions · 2025-12-24T01:06:25Z

Label Messager: wasmtime:config

It looks like you are changing Wasmtime's configuration options. Make sure to
complete this check list:

If you added a new Config method, you wrote extensive documentation for
it.

Details

Our documentation should be of the following form:

Short, simple summary sentence.

More details. These details can be multiple paragraphs. There should be
information about not just the method, but its parameters and results as
well.

Is this method fallible? If so, when can it return an error?

Can this method panic? If so, when does it panic?

# Example

Optional example here.

If you added a new Config method, or modified an existing one, you
ensured that this configuration is exercised by the fuzz targets.

Details

For example, if you expose a new strategy for allocating the next instance
slot inside the pooling allocator, you should ensure that at least one of our
fuzz targets exercises that new strategy.

Often, all that is required of you is to ensure that there is a knob for this
configuration option in wasmtime_fuzzing::Config (or one
of its nested structs).

Rarely, this may require authoring a new fuzz target to specifically test this
configuration. See our docs on fuzzing for more details.
If you are enabling a configuration option by default, make sure that it
has been fuzzed for at least two weeks before turning it on by default.

Details

To modify this label's message, edit the .github/label-messager/wasmtime-config.md file.

To add new label messages or remove existing label messages, edit the
.github/label-messager.json configuration file.

Learn more.

alexcrichton · 2026-01-07T18:09:25Z

On a skim this looks like it's all in the right direction, thanks! As as a heads up the wasm-tools deps will be updated in #12254 which'll avoid the need for git deps. I'll take a closer look once this is further along in CI passing tests

github-actions · 2026-01-08T11:46:55Z

Subscribe to Label Action

cc @fitzgen

Details

This issue or pull request has been labeled: "fuzzing"

Thus the following users have been cc'd because of the following labels:

fitzgen: fuzzing

To subscribe or unsubscribe from this label, edit the .github/subscribe-to-label.json configuration file.

Learn more.

.gitignore

alexcrichton

This is looking quite good to me, thanks for the thorough tests!

I haven't scrutinized the trampoline generation nor the lifting/lowering yet, but I can do that once the tests added here are passing (the #[ignore] ones at least).

If you can one thing I'd also recommend is modeling as many tests as possible as a *.wast test since that's generally the easiest to run and share (albeit difficult to write and debug)

crates/test-util/src/component.rs

crates/test-util/src/component_fuzz.rs

tests/misc_testsuite/component-model/types.wast

.gitignore

alexcrichton

Mostly some thoughts about deduplication/sharing of the gnarliest bits below --

crates/environ/src/fact/trampoline.rs

crates/wasmtime/src/runtime/component/func/typed.rs

crates/wasmtime/src/runtime/component/values.rs

tests/all/component_model/func.rs

Signed-off-by: Yordis Prieto <yordis.prieto@gmail.com>

…model This commit introduces the Map and MapEntry classes, enabling the representation of map values in the component model. The Map class allows for the creation and iteration of key/value pairs, enhancing the functionality of the wasmtime component API. Additionally, the .gitignore file is updated to exclude build artifacts from the crates/c-api directory.

This commit introduces support for HashMap<K, V> in the component model, allowing maps to be represented as list<tuple<K, V>> in the canonical ABI. It includes implementations for the ComponentType, Lower, and Lift traits for HashMap, enabling type checking, lowering to flat representations, and lifting from memory. Additionally, the maximum depth for type generation in the fuzzing utility is updated to accommodate the new map type.

This commit removes the previous wasm features configuration and adds new functions for creating a map-configured engine. The `map_config` and `map_engine` functions are introduced to facilitate the use of the component model with maps in tests, ensuring that the engine is properly configured for map types in the component model.

…existing tests This commit introduces a new WAST test file specifically for testing various map types in the component model. Additionally, it removes the redundant map type definitions from the existing types.wast file to streamline the test suite.

…eturn with call

…n typed.rs

…ved iteration and memory management. Introduce new implementations for ComponentType, Lower, and Lift traits for std::collections::HashMap, enhancing support for map types in the component model.

The translate_map function had two categories of bugs preventing map adapter trampolines from working: 1. Wasm stack discipline: local_set_new_tmp emits LocalSet which pops from the stack, but was called when the stack was empty (to "pre-allocate" locals). Fixed by computing values first, then calling local_set_new_tmp to consume them—matching translate_list's pattern. Also removed an erroneous LocalTee that left an orphan value on the stack. Affected: src_byte_len, dst_byte_len, cur_src_ptr, cur_dst_ptr. 2. Pointer advancement: after value translation, the pointer still points at the value start. The code only advanced by trailing padding instead of value_size + trailing_padding, causing every loop iteration to re-read the same memory. Also fixes entry layout to use proper record alignment rules (entry align = max(key_align, value_align), value at aligned offset).

Val::Map already holds Vec<(Val, Val)> which derefs to &[(Val, Val)], matching lower_map's signature directly. The intermediate Vec allocation and deep clone of every key/value pair was redundant.

yordis · 2026-02-27T06:04:56Z

@alexcrichton addressed the comments, about the sharing stuff, feedback welcome there, I am not sure if it is the best way or not; not sure what you had in mind

alexcrichton · 2026-02-27T18:10:08Z

crates/wasmtime/src/runtime/component/values.rs

+    let tuple_alignment = key_abi.align32.max(value_abi.align32);
+    let tuple_size = usize::try_from(offset).unwrap();
+    let tuple_size = (tuple_size + usize::try_from(tuple_alignment)? - 1)
+        & !(usize::try_from(tuple_alignment)? - 1);


Reading over crates/environ/src/component/types.rs again, I think it would be best to store the "tuple ABI" inside of the TypeMap structure. That way it wouldn't need to be caluclated at runtime either here or in the typed.rs case and it would only be calculated in one location.

That might also be a good location to store a value_offset: u32 value which is the offset from the start of the tuple to the value, so it also wouldn't need to be recomputed here for example.

alexcrichton · 2026-02-27T18:11:53Z

crates/wasmtime/src/runtime/component/func/typed.rs

+#[cfg(feature = "std")]
+unsafe impl<K, V> Lift for std::collections::HashMap<K, V>
+where
+    K: Lift + Eq + Hash,
+    V: Lift,
+{
+    fn linear_lift_from_flat(
+        cx: &mut LiftContext<'_>,
+        ty: InterfaceType,
+        src: &Self::Lower,
+    ) -> Result<Self> {
+        let (key_ty, value_ty) = match ty {
+            InterfaceType::Map(i) => {
+                let m = &cx.types[i];
+                (m.key, m.value)
+            }
+            _ => bad_type_info(),
+        };
+        // FIXME(#4311): needs memory64 treatment
+        let ptr = src[0].get_u32();
+        let len = src[1].get_u32();
+        let (ptr, len) = (usize::try_from(ptr)?, usize::try_from(len)?);
+        lift_std_map(cx, key_ty, value_ty, ptr, len)
+    }
+
+    fn linear_lift_from_memory(
+        cx: &mut LiftContext<'_>,
+        ty: InterfaceType,
+        bytes: &[u8],
+    ) -> Result<Self> {
+        let (key_ty, value_ty) = match ty {
+            InterfaceType::Map(i) => {
+                let m = &cx.types[i];
+                (m.key, m.value)
+            }
+            _ => bad_type_info(),
+        };
+        debug_assert!((bytes.as_ptr() as usize) % (Self::ALIGN32 as usize) == 0);
+        // FIXME(#4311): needs memory64 treatment
+        let ptr = u32::from_le_bytes(bytes[..4].try_into().unwrap());
+        let len = u32::from_le_bytes(bytes[4..].try_into().unwrap());
+        let (ptr, len) = (usize::try_from(ptr)?, usize::try_from(len)?);
+        lift_std_map(cx, key_ty, value_ty, ptr, len)
+    }
+}


For lifting here what I'd recommend is that since this is implemented in terms of the fallible hash map this can delegate to the fallible version, then convert the fallible version to the std version.

Fore xample linear_lift_from_flat here would call HashMap::linear_lift_form_flat(...), and if successful, would convert that to std::collections::HashMap

alexcrichton · 2026-02-27T18:12:20Z

crates/wasmtime/src/runtime/component/func/typed.rs

+    fn typecheck(ty: &InterfaceType, types: &InstanceType<'_>) -> Result<()> {
+        match ty {
+            InterfaceType::Map(t) => {
+                let map_ty = &types.types[*t];
+                K::typecheck(&map_ty.key, types)?;
+                V::typecheck(&map_ty.value, types)?;
+                Ok(())
+            }
+            other => bail!("expected `map` found `{}`", desc(other)),
+        }
+    }
+}


This would be a good place to delegate to HashMap::typecheck(...) as well for example as opposed to duplicating the method body here.

alexcrichton · 2026-02-27T18:13:36Z

crates/wasmtime/src/runtime/component/func/typed.rs

+    fn linear_lower_to_flat<U>(
+        &self,
+        cx: &mut LowerContext<'_, U>,
+        ty: InterfaceType,
+        dst: &mut MaybeUninit<[ValRaw; 2]>,
+    ) -> Result<()> {
+        let (key_ty, value_ty) = match ty {
+            InterfaceType::Map(i) => {
+                let m = &cx.types[i];
+                (m.key, m.value)
+            }
+            _ => bad_type_info(),
+        };
+        let (ptr, len) = lower_map_iter(cx, key_ty, value_ty, self.len(), self.iter())?;
+        // See "WRITEPTR64" above for why this is always storing a 64-bit
+        // integer.
+        map_maybe_uninit!(dst[0]).write(ValRaw::i64(ptr as i64));
+        map_maybe_uninit!(dst[1]).write(ValRaw::i64(len as i64));
+        Ok(())
+    }
+
+    fn linear_lower_to_memory<U>(
+        &self,
+        cx: &mut LowerContext<'_, U>,
+        ty: InterfaceType,
+        offset: usize,
+    ) -> Result<()> {
+        let (key_ty, value_ty) = match ty {
+            InterfaceType::Map(i) => {
+                let m = &cx.types[i];
+                (m.key, m.value)
+            }
+            _ => bad_type_info(),
+        };
+        debug_assert!(offset % (Self::ALIGN32 as usize) == 0);
+        let (ptr, len) = lower_map_iter(cx, key_ty, value_ty, self.len(), self.iter())?;
+        *cx.get(offset + 0) = u32::try_from(ptr).unwrap().to_le_bytes();
+        *cx.get(offset + 4) = u32::try_from(len).unwrap().to_le_bytes();
+        Ok(())
+    }


What I might recommend for deduplicating this with the above HashMap implementation is to have two helpers on top of lower_map_iter, specifically linear_lower_map_to_memory and linear_lower_map_to_flat. That would have the boilerplate here (plus a generic iterator) and that way the hash map lower implementations would be one-liners to the shared functions.

yordis force-pushed the yordis/feat-map-support branch 3 times, most recently from 6b6f675 to cd2282b Compare December 23, 2025 21:53

yordis force-pushed the yordis/feat-map-support branch 5 times, most recently from db45a3a to 83dd7a7 Compare January 8, 2026 10:17

github-actions bot added the fuzzing Issues related to our fuzzing infrastructure label Jan 8, 2026

yordis force-pushed the yordis/feat-map-support branch from 83dd7a7 to b9ca740 Compare January 8, 2026 19:19

yordis commented Jan 16, 2026

View reviewed changes

.gitignore Show resolved Hide resolved

alexcrichton reviewed Jan 20, 2026

View reviewed changes

crates/test-util/src/component.rs Outdated Show resolved Hide resolved

crates/test-util/src/component_fuzz.rs Outdated Show resolved Hide resolved

tests/misc_testsuite/component-model/types.wast Outdated Show resolved Hide resolved

.gitignore Show resolved Hide resolved

yordis force-pushed the yordis/feat-map-support branch from 301e19b to d291adf Compare February 6, 2026 06:26

yordis requested a review from alexcrichton February 10, 2026 20:12

yordis force-pushed the yordis/feat-map-support branch from ddac63a to b6376cb Compare February 16, 2026 23:06

yordis marked this pull request as ready for review February 16, 2026 23:07

yordis requested review from a team as code owners February 16, 2026 23:07

alexcrichton reviewed Feb 20, 2026

View reviewed changes

yordis added 2 commits February 26, 2026 20:21

Add support for map type

b5a1d6d

Signed-off-by: Yordis Prieto <yordis.prieto@gmail.com>

yordis added 11 commits February 26, 2026 20:21

Add wasm_component_model_map configuration support

8735d2a

Format code

cf61139

Format C code

c2c44fd

Update component fuzzing and dynamic tests to replace call_and_post_r…

d9c039d

…eturn with call

Format code

a2d54a2

Refactor HashMap usage in typed.rs to use wasmtime_environ collections

bb6fb78

Fix HashMap initialization and insertion to handle potential errors i…

b0d318e

…n typed.rs

Refactor HashMap handling in typed.rs to use lower_map_iter for impro…

e644b8b

…ved iteration and memory management. Introduce new implementations for ComponentType, Lower, and Lift traits for std::collections::HashMap, enhancing support for map types in the component model.

yordis force-pushed the yordis/feat-map-support branch from b6376cb to e644b8b Compare February 27, 2026 01:21

yordis added 6 commits February 26, 2026 21:58

Refactor map entry layout calculations to use canonical ABI

9a9f37f

Remove unnecessary clone of map pairs during lowering

62c1ffb

Val::Map already holds Vec<(Val, Val)> which derefs to &[(Val, Val)], matching lower_map's signature directly. The intermediate Vec allocation and deep clone of every key/value pair was redundant.

Deduplicate map lift logic between HashMap implementations

b373bde

Deduplicate list and map sequence translation scaffolding

f54f788

Fix cargo fmt formatting issues

756a68e

yordis requested a review from alexcrichton February 27, 2026 06:05

alexcrichton reviewed Feb 27, 2026

View reviewed changes

Conversation

yordis commented Dec 23, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Context

References

Uh oh!

github-actions bot commented Dec 23, 2025

Subscribe to Label Action

Uh oh!

github-actions bot commented Dec 24, 2025

Label Messager: wasmtime:config

Uh oh!

alexcrichton commented Jan 7, 2026

Uh oh!

github-actions bot commented Jan 8, 2026

Subscribe to Label Action

Uh oh!

Uh oh!

alexcrichton left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

alexcrichton left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

yordis commented Feb 27, 2026

Uh oh!

alexcrichton Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

alexcrichton Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

alexcrichton Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

alexcrichton Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

alexcrichton Feb 27, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

yordis commented Dec 23, 2025 •

edited

Loading