[wasm][R2R] Request flat layout for composite image on WebAssembly by lewing · Pull Request #131354 · dotnet/runtime

lewing · 2026-07-24T23:03:36Z

On WebAssembly, PE images are flat-layout only (PEIMAGE_FLAT_LAYOUT_ONLY). A webcil composite R2R image has no OS-loadable/mapped layout, so requesting LAYOUT_LOADED yields no layout and trips the flat-suitability assert in GetOrCreateLayoutInternal. Request the flat layout explicitly under TARGET_WASM when obtaining the composite image layout in NativeImage::Open.

Guarded by #ifdef TARGET_WASM, so no behavior change on any other target.

Note

This pull request was authored with assistance from GitHub Copilot.

OpenR2RFromPE requested PEImageLayout::LAYOUT_LOADED for the composite R2R container. On WebAssembly (PEIMAGE_FLAT_LAYOUT_ONLY) there is no OS-loadable mapped layout, so the loaded-layout branch in GetOrCreateLayoutInternal is compiled out and the request degrades to nothing, tripping the _ASSERTE(bIsFlatLayoutSuitable) consistency check in Debug builds (and only silently degrading in Release). Every other wasm image open uses LAYOUT_ANY. Request LAYOUT_FLAT explicitly on TARGET_WASM so a webcil composite loads via the flat layout, matching the rest of the wasm image-load paths. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 50ddfb2c-b6f8-4847-81e7-44d37d44a175

azure-pipelines · 2026-07-24T23:04:16Z

Azure Pipelines: Successfully started running 3 pipeline(s). 13 pipeline(s) were filtered out due to trigger conditions. There may be pipelines that require an authorized user to comment /azp run to run.

Copilot

Pull request overview

Updates CoreCLR’s composite R2R loading path on WebAssembly to request a flat PE layout explicitly when opening the composite image, avoiding a debug assert caused by requesting a loaded/mapped layout on a flat-only platform.

Changes:

Under TARGET_WASM, call PEImage::GetOrCreateLayout(PEImageLayout::LAYOUT_FLAT) instead of LAYOUT_LOADED when opening the composite image.
Add in-code rationale explaining why the flat layout is required on WASM for webcil composites.

dotnet-policy-service · 2026-07-24T23:48:56Z

Tagging subscribers to 'arch-wasm': @lewing, @pavelsavara
See info in area-owners.md if you want to be subscribed.

jkoritzinsky

Is there a reason we aren't using the "native format composite image" path we added for Mach-O for WASM?

Use flat PE image layout on flat-layout-only targets and record the NativeImageEagerFixups to PregeneratedStringThunks lock order exposed by composite eager fixups. The flat-layout correction is based on dotnet#131354. Co-authored-by: Larry Ewing <lewing@microsoft.com> Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 44f4b4ea-409c-4a52-89a0-f488a89488b9

lewing · 2026-07-27T18:40:26Z

Is there a reason we aren't using the "native format composite image" path we added for Mach-O for WASM?

Right now wasm composite mode is being used as a minimum delta between the browser (webcil) target and wasi which effectively requires composite mode. Once wasm r2r is running tests in CI we can revisit the format.

davidwrighton · 2026-07-28T17:59:53Z

@lewing, have you validated that a composite image on the browser path works with this change as well as your wasi approach?

lewing · 2026-07-29T00:32:24Z

@lewing, have you validated that a composite image on the browser path works with this change as well as your wasi approach?

yes, it is both necessary and sufficient to wasm webcil r2r composite mode working on browser right now (just double checked)

flat layout:                      nested layout:
Ready to Run initialized          Ready to Run header not found: "System.Private.CoreLib"
  successfully: "System.Private   Ready to Run header not found: "System.Console"
  .CoreLib"                       Ready to Run header not found: "System.Runtime"
Ready to Run initialized          ...
  successfully: "System.Console"
Ready to Run initialized
  successfully: "System.Runtime"

Copilot

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

Suppressed comments (1)

src/coreclr/vm/nativeimage.cpp:145

The #endif trailing comment doesn’t match the corresponding #ifdef (PEIMAGE_FLAT_LAYOUT_ONLY). The repo’s native-code conventions call out matching #endif comments exactly to avoid confusion when refactoring preprocessor blocks.

#endif // TARGET_WASM

Co-authored-by: Jan Kotas <jkotas@microsoft.com>

Copilot

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

…-1 tests (#131883) ## Summary PR #131110 enabled CoreCLR browser-WASM out-of-process runtime tests. The newly enabled priority-1 coverage exposed eight unique failures. Four were composite WebCIL/ReadyToRun failures tracked by #131767 and are fixed on `main` by #131354. This change removes their temporary `ActiveIssue` annotations and Crossgen2 bypasses, restoring the intended R2R coverage. The unrelated field-layout suppression for #131421 remains. The other four tests are narrowly gated: - `GC/API/GC/GetTotalAllocatedBytesServerGC` is excluded on CoreCLR browser, which does not provide Server GC. - `GC/API/GC/GetTotalMemoryConcurrent` uses `PlatformDetection.IsMultithreadingSupported` through `ConditionalFact`. - `baseservices/threading/regressions/2164/foreground-shutdown` retains its explicit `Main` and returns success before creating a thread when multithreading is unsupported. - `JIT/jit64/regress/ddb/113574` is excluded only on CoreCLR browser because its multi-billion-iteration optimizing-JIT workload exceeds the browser interpreter's CI timeout. The runtime capability checks skip single-threaded browser and WASI while preserving desktop and threaded-browser coverage. The Server GC and interpreter-workload exclusions are tracked in #131321. ## Testing Built the Checked browser runtime from `main`, including #131354, then built and ran the complete priority-1 browser-WASM Node suite: | Scope | Total | Passed | Failed | Skipped | | --- | ---: | ---: | ---: | ---: | | Full suite | 14,745 | 13,823 | 0 | 922 | | Out-of-process subset | 672 | 560 | 0 | 112 | The four former WebCIL/R2R failures now pass: - `Regressions/coreclr/GitHub_49826/test49826` - `Regressions/coreclr/GitHub_49982/test49982` - `readytorun/tests/genericsload/callgenericctor` - `readytorun/tests/genericsload/usegenericfield` Additional focused validation for the runtime multithreading gates: - 8 MSBuild evaluations across desktop, threaded browser, single-threaded browser, and single-threaded WASI; neither project is permanently marked unsupported. - 2 targeted Checked browser builds, both with 0 warnings and 0 errors. - 2 single-threaded browser wrapper runs, both passing with expected/actual exit code 100. - Native `foreground-shutdown` execution exited with 100 after 2.28 seconds, confirming that its foreground thread still keeps the process alive after `Main` returns. > [!NOTE] > This pull request was created with assistance from GitHub Copilot. Copilot-Session: 19e61792-38f9-4f80-8d38-3e3e12d1275e

…otnet#131354) # Honour ErrorOnUnknownConfiguration on value conversion fail Fixes dotnet#98231 ## Problem `BinderOptions.ErrorOnUnknownConfiguration` is documented as controlling whether the binder throws "when converting a value", but it was only honoured on some of the paths that can fail: | Failure | Flag honoured? | |---|---| | Unknown configuration key | Yes | | Leaf value with no `TypeConverter` at all | Yes | | Collection item that fails to convert | Yes | | **Leaf value whose `TypeConverter` throws** | **No, always threw** | So a single malformed value in an otherwise valid configuration source would tear down binding even for callers that had deliberately left the flag at its default: ```csharp // appsettings.json: { "Timeout": "not-a-number" } var options = config.Get<MyOptions>(); // threw InvalidOperationException ``` This is inconsistent both with the documented contract on the property and with the sibling "no converter" case a few lines away in the same method, which already returned quietly. ## Fix Gate the rethrow on the flag, in both the reflection binder and the source generator. ### Reflection binder `BindInstance` no longer rethrows the conversion error unconditionally; it returns and leaves the binding point untouched, so the member keeps whatever value it already had. ### Source generator The emitted `ParseX(value, path)` helpers become `TryParseX(value, path, errorOnFailedBinding, out result)`: ```csharp public static bool TryParseInt(string value, string? path, bool errorOnFailedBinding, out int result) { try { result = int.Parse(value, NumberStyles.Integer, CultureInfo.InvariantCulture); return true; } catch when (!errorOnFailedBinding) { result = default; return false; } catch (Exception exception) { throw new InvalidOperationException($"Failed to convert configuration value '{value ?? "null"}' at '{path}' to type '{typeof(int)}'.", exception); } } ``` Call sites pass `binderOptions?.ErrorOnUnknownConfiguration is true`, except for two cases that must always report the failure: - `GetValueCore`, since `GetValue<T>` has no `BinderOptions` overload. - Constructor parameters with no declared default, which are unsatisfiable either way. These keep throwing the accurate `Failed to convert configuration value ...` rather than degrading to `... has no matching config`. A try/catch at each call site was considered instead, but it would also have swallowed nested binding errors in the dictionary case; a `Func<>`-based helper would have allocated. ## Behaviour change This is a breaking change and will need a breaking-change doc. | Scenario | Before | After | |---|---|---| | `Bind` / `Get<T>` with an unconvertible scalar, default options | throws | member keeps its existing value | | `GetSection("x").Get<int>()` with a bad value | throws | `default(int)` | | Same, with `ErrorOnUnknownConfiguration = true` | throws | throws (unchanged) | | `GetValue<T>` with a bad value | throws | throws (unchanged) | | Required constructor parameter with a bad value | throws | throws (unchanged) | | Constructor parameter with a declared default and a bad value | throws | falls back to the default | The `Get`/`Bind` versus `GetValue` asymmetry is deliberate and follows the design agreed in the issue, but it is user-visible. ## Also in this change - Fixes a latent generator bug: a primitive section with an empty value used to fall through `GetCore` to `throw new NotSupportedException("Unable to bind to type ...: generator did not detect the type as input.")`. It now returns `null`, matching the reflection binder. - Fixes a hard-coded path separator in `GeneratorTests.Helpers.cs` that stopped `/p:UpdateBaselines=true` working on non-Windows hosts. - Un-gates two `CollectionsBindingWithErrorOnUnknownConfiguration` tests that were restricted to the reflection binder; they now pass under source generation too. ## Testing - 4 existing tests that asserted the old unconditional throw now opt in with `ErrorOnUnknownConfiguration = true`. - New regression tests cover both flag states, the `GetValue` carve-out, and constructor parameters with and without declared defaults. All run in both binder modes. - 111 of the 120 generator baselines regenerated (`netcoreapp` and `net462`, `Version0` and `Version1`). Some `net462` `Version0` files additionally pick up unrelated drift corrections that had accumulated on `main`. ``` Microsoft.Extensions.Configuration.Binder.Tests 356 passed, 0 failed Microsoft.Extensions.Configuration.Binder.SourceGeneration.Tests 433 passed, 0 failed, 26 skipped ```

Copilot AI review requested due to automatic review settings July 24, 2026 23:03

github-actions Bot added the area-VM-coreclr label Jul 24, 2026

dotnet-policy-service Bot assigned lewing Jul 24, 2026

Copilot started reviewing on behalf of lewing July 24, 2026 23:04 View session

lewing requested review from AndyAyersMS and davidwrighton July 24, 2026 23:07

Copilot AI reviewed Jul 24, 2026

View reviewed changes

Comment thread src/coreclr/vm/nativeimage.cpp Outdated

lewing added the arch-wasm WebAssembly architecture label Jul 24, 2026

jkoritzinsky reviewed Jul 26, 2026

View reviewed changes

radekdoulik mentioned this pull request Jul 27, 2026

Enable out-of-process tests for CoreCLR browser WASM #131110

Merged

lewing requested a review from pavelsavara August 2, 2026 16:00

pavelsavara approved these changes Aug 3, 2026

View reviewed changes

jkotas reviewed Aug 3, 2026

View reviewed changes

Comment thread src/coreclr/vm/nativeimage.cpp Outdated

jkotas reviewed Aug 3, 2026

View reviewed changes

Comment thread src/coreclr/vm/nativeimage.cpp Outdated

feedback

003e3ff

Copilot AI review requested due to automatic review settings August 3, 2026 14:13

Copilot started reviewing on behalf of pavelsavara August 3, 2026 14:14 View session

Copilot AI reviewed Aug 3, 2026

View reviewed changes

jkotas approved these changes Aug 3, 2026

View reviewed changes

jkotas reviewed Aug 3, 2026

View reviewed changes

Comment thread src/coreclr/vm/nativeimage.cpp Outdated

Update src/coreclr/vm/nativeimage.cpp

7de70ad

Co-authored-by: Jan Kotas <jkotas@microsoft.com>

Copilot AI review requested due to automatic review settings August 3, 2026 15:36

Copilot started reviewing on behalf of pavelsavara August 3, 2026 15:37 View session

akoeplinger approved these changes Aug 3, 2026

View reviewed changes

Copilot AI reviewed Aug 3, 2026

View reviewed changes

radekdoulik mentioned this pull request Aug 3, 2026

[wasm][CoreCLR][R2R] Composite WebCIL images request an unsupported loaded PE layout #131767

Open

pavelsavara merged commit c8e00c9 into dotnet:main Aug 5, 2026
110 checks passed

radekdoulik mentioned this pull request Aug 5, 2026

[browser][CoreCLR] Restore R2R coverage and gate unsupported priority-1 tests #131883

Merged

dotnet-maestro Bot mentioned this pull request Aug 6, 2026

[main] Source code updates from dotnet/runtime dotnet/dotnet#8163

Merged

dotnet-milestone-bot Bot added this to the 11.0-rc1 milestone Aug 6, 2026

Navigation Menu

Uh oh!

[wasm][R2R] Request flat layout for composite image on WebAssembly - #131354

[wasm][R2R] Request flat layout for composite image on WebAssembly#131354
pavelsavara merged 3 commits into
dotnet:mainfrom
lewing:wasm-r2r-composite-flat-layout

lewing commented Jul 24, 2026

Uh oh!

azure-pipelines Bot commented Jul 24, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

dotnet-policy-service Bot commented Jul 24, 2026

Uh oh!

jkoritzinsky left a comment

Uh oh!

lewing commented Jul 27, 2026

Uh oh!

davidwrighton commented Jul 28, 2026

Uh oh!

lewing commented Jul 29, 2026 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

Uh oh!

Conversation

lewing commented Jul 24, 2026

Uh oh!

azure-pipelines Bot commented Jul 24, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

dotnet-policy-service Bot commented Jul 24, 2026

Uh oh!

jkoritzinsky left a comment

Choose a reason for hiding this comment

Uh oh!

lewing commented Jul 27, 2026

Uh oh!

davidwrighton commented Jul 28, 2026

Uh oh!

lewing commented Jul 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

7 participants

lewing commented Jul 29, 2026 •

edited

Loading