[cDAC] WebAssembly: real-data R2R stack-walk validation tests by lewing · Pull Request #131161 · dotnet/runtime

lewing · 2026-07-21T17:04:49Z

Summary

Follow-up to #130988 (cDAC WebAssembly support). That PR added the WASM cDAC stack-walk plug-in (WasmContext / WasmUnwinder / WasmR2RInfo / WasmFrameHandler) and the metadata-resolution contract fixes, with unit coverage built on synthetic mock values. This PR anchors the R2R forward path to real runtime behavior and corrects several inaccurate comments.

Tests

Empty range-list guard — s_pFunctionTableIndexRangeList is an empty (null) list head until the runtime populates it per module while applying eager fixups at module load. That empty state is legitimate — an interpreter-only process, or before any ReadyToRun module has loaded — so WasmR2RInfo.FindSection must treat it as a clean "no R2R sections" answer rather than dereferencing the null head. EmptyRangeList_ResolvesToNoSection locks that in.
Real-data R2R validation — two tests anchored to values captured from a live dispatching composite ReadyToRun WASM build (a merged CoreLib R2R image):
- The live FunctionTableIndexRangeSection node: MinFunctionTableIndex = 6259, NumRuntimeFunctions = 45283, single node (pNext = 0).
- ReadyToRunInfo.MinVirtualIP = 0 (single-image composite → base virtual IP reduces to BeginAddress).
- Per-frame values from disassembling that composite's R2R prolog for System.Exception..ctor: on entry it stores its function-table index (6259 + 1040) at frame+0 and its function-local virtual IP / 2 at frame+4.
RealDispatchingBuild_ResolvesFunctionTableIndexInRange validates FindSection / TryGetVirtualIPBase against the real range (including the exclusive [Min, Min+Num) top boundary). RealDispatchingBuild_UnwinderDecodesRealFrameVirtualIP drives WasmUnwinder.GetVirtualIP through the real WasmR2RInfo end-to-end — the first test exercising the unwinder against the real section reader rather than a fake — validating the × 2 encoding on the frame field with no shift on BeginAddress.

Comment corrections (no behavior change)

WasmR2RInfo: the class summary still listed ExecutionManager::IsFuncletFunctionIndex among the mirrored native APIs after that member was removed from IWasmR2RInfo in review. The class still reads the funclet flag internally, so only the API list changes.
WasmUnwinder: now also points at src/coreclr/vm/wasm/callhelpers.hpp, which defines the frame offsets and sentinels the unwinder mirrors, alongside helpers.cpp (the walk itself).
WasmR2RInfoTests: an earlier draft of the empty-list test claimed the runtime registers virtual-IP ranges "lazily ... until the first managed stack walk". That is incorrect — ReadyToRunInfo::RegisterVirtualIPRange runs per module during eager fixups at module load (Module::RunEagerFixupsUnlocked). The test stands; only its rationale was wrong.

Context

The captured values come from cross-session live debugging of a dispatching R2R-WASM build. A notable finding from that work: on WASM the R2R virtual-IP walk is DAC/reader-driven — managed StackTrace / EH / GC do not invoke it in-process; the WasmUnwinder itself is the walker, driven against a paused target. So these fixture-based tests (replaying captured structure values) are the right CI shape and require no WASM runtime in CI. A future PR can add a live reader-drives-debugger pass once broader R2R dispatch is available.

Testing

Full cDAC unit suite: 2766 passed / 0 failed / 16 skipped.
The integration test is mutation-verified: changing the unwinder's × 2 to × 3 fails it (9746 vs 9747), confirming it guards the real encoding rather than passing vacuously.

Note

This pull request was authored with the assistance of GitHub Copilot.

The runtime registers R2R virtual-IP ranges lazily, so s_pFunctionTableIndexRangeList is an empty (null) list head until the first managed stack walk — which is the actual runtime state today. Add a test asserting FindSection treats an empty list as a clean "no R2R sections" (TryGetVirtualIPBase / TryGetUnwindData return false) rather than dereferencing the null head. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: afa9efc6-f596-41bd-9070-dc651db8b7b4

Anchors the WasmR2RInfo / WasmUnwinder forward path to values captured from a live dispatching composite ReadyToRun WASM build (a merged CoreLib R2R image): the FunctionTableIndexRangeSection node { MinFunctionTableIndex = 6259, NumRuntimeFunctions = 45283 }, ReadyToRunInfo.MinVirtualIP = 0, and a dispatched System.Exception..ctor frame storing its function-table index (6259 + 1040) at frame+0 and its function-local virtual IP / 2 at frame+4. - RealDispatchingBuild_ResolvesFunctionTableIndexInRange: a real function-table index resolves through the single range node; base virtual IP = MinVirtualIP + BeginAddress; the [Min, Min + Num) boundaries are exclusive at the top. - RealDispatchingBuild_UnwinderDecodesRealFrameVirtualIP: end-to-end GetVirtualIP over the real frame through the real WasmR2RInfo, validating the * 2 encoding on the frame field with no shift on BeginAddress. Full cDAC suite: 2766 passed / 0 failed / 16 skipped. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: afa9efc6-f596-41bd-9070-dc651db8b7b4

azure-pipelines · 2026-07-21T17:05:40Z

Azure Pipelines: Successfully started running 4 pipeline(s). 12 pipeline(s) were filtered out due to trigger conditions. There may be pipelines that require an authorized user to comment /azp run to run.

dotnet-policy-service · 2026-07-21T17:05:56Z

Tagging subscribers to this area: @agocke
See info in area-owners.md if you want to be subscribed.

Copilot

Pull request overview

This PR strengthens cDAC’s WebAssembly ReadyToRun (R2R) stack-walk coverage by adding unit tests that validate WasmR2RInfo / WasmUnwinder behavior against an empty range-list state and against a “real-data” fixture modeled after a dispatching composite R2R WASM build.

Changes:

Extend the test target builder to optionally model an empty FunctionTableIndexRangeList (null list head).
Add EmptyRangeList_ResolvesToNoSection to lock in “empty list => clean miss” behavior for both TryGetVirtualIPBase and TryGetUnwindData.
Add two “real-data” fixture tests that validate range inclusion boundaries and end-to-end virtual IP reconstruction through WasmUnwinder.GetVirtualIP.

Comment-only; no behavior change. - WasmR2RInfoTests: the empty-range-list test claimed the runtime registers virtual-IP ranges "lazily ... until the first managed stack walk". That is wrong: ReadyToRunInfo::RegisterVirtualIPRange runs per module while applying eager fixups at module load (Module::RunEagerFixupsUnlocked). An empty list is still a legitimate state -- an interpreter-only process, or before any ReadyToRun module has loaded -- so the test stands; only its rationale was incorrect. - WasmR2RInfoTests: the captured-value comment implied the per-frame values were read from an observed live frame. They come from disassembling the composite's R2R prolog for System.Exception..ctor; say so. - WasmR2RInfo: the class summary still listed ExecutionManager::IsFuncletFunctionIndex among the mirrored native APIs after that member was removed from IWasmR2RInfo. The class still reads the funclet flag internally, so only the API list changes. - WasmUnwinder: point at src/coreclr/vm/wasm/callhelpers.hpp, which defines the frame offsets and sentinels the unwinder mirrors, alongside helpers.cpp (the walk). Full cDAC suite: 2766 passed / 0 failed / 16 skipped. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: afa9efc6-f596-41bd-9070-dc651db8b7b4

Copilot

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated no new comments.

…(wasi-sdk) (#131241) ## Summary On non-emscripten wasm (wasi-sdk), the cDAC contract descriptor getter `GetDotNetRuntimeContractDescriptor` is not exported, so `--gc-sections` strips the descriptor from the final module. This prevents a native host / the cDAC reader from discovering the runtime data-contract descriptor on WASI. The emscripten path already handles this via `EMSCRIPTEN_KEEPALIVE`, which both exports the getter and keeps the descriptor alive. There was no equivalent for the wasi-sdk (`__wasm__`, non-emscripten) toolchain. ## Fix Add an `#elif defined(EXPORT_CONTRACT) && defined(__wasm__)` branch that emits the getter with `__attribute__((export_name("Get" #CONTRACT_NAME)))`. The explicit wasm export both makes the symbol reachable by a native host and roots it so `--gc-sections` does not remove the descriptor — mirroring what `EMSCRIPTEN_KEEPALIVE` does on the emscripten path. The `GETTER_NAME` token-paste macros are hoisted above the `#if` (shared by both branches) and a `GETTER_NAME_STR` stringize pair is added to produce the literal export name `"GetDotNetRuntimeContractDescriptor"`. ## Validation - WASI CoreCLR build (`clr -os wasi -c Release`): compiles clean (0W/0E). - Confirmed the compiled object exports `GetDotNetRuntimeContractDescriptor` with the `no_strip` flag, so it survives `--gc-sections`. This unblocks live cDAC discovery on WASI (non-emscripten) — the runtime-side counterpart to the cDAC-wasm validation work (#131161). Fixture-based cDAC tests do not depend on it. > [!NOTE] > This PR was authored with the assistance of GitHub Copilot. --------- Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com> Copilot-Session: 2902374a-f352-4c46-bcec-8a35c97733ae Copilot-Session: 50ddfb2c-b6f8-4847-81e7-44d37d44a175

…validation

azure-pipelines · 2026-07-25T18:52:03Z

Azure Pipelines: Successfully started running 4 pipeline(s). 12 pipeline(s) were filtered out due to trigger conditions. There may be pipelines that require an authorized user to comment /azp run to run.

…dac-wasm-r2r-validation

Copilot

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

+        // A contiguous runtime function table covering local indices [0, RealLocalIndex]. Only the
+        // entry at RealLocalIndex carries a non-zero (non-funclet) BeginAddress; the rest read as zero.
+        uint runtimeFunctionStride = runtimeFunctionLayout.Stride;
+        uint tableEntries = RealLocalIndex + 1;
+        var runtimeFuncTableFrag = allocator.Allocate((ulong)(tableEntries * runtimeFunctionStride), "RuntimeFunctions");


lewing and others added 2 commits July 21, 2026 11:59

Copilot AI review requested due to automatic review settings July 21, 2026 17:04

github-actions Bot added the area-VM-coreclr label Jul 21, 2026

dotnet-policy-service Bot assigned lewing Jul 21, 2026

Copilot started reviewing on behalf of lewing July 21, 2026 17:05 View session

Copilot AI reviewed Jul 21, 2026

View reviewed changes

lewing mentioned this pull request Jul 22, 2026

[wasm] Export cDAC contract descriptor getter on non-emscripten wasm (wasi-sdk) #131241

Merged

Copilot AI review requested due to automatic review settings July 24, 2026 18:20

Copilot started reviewing on behalf of lewing July 24, 2026 18:21 View session

Copilot AI reviewed Jul 24, 2026

View reviewed changes

Merge remote-tracking branch 'origin/main' into lewing-cdac-wasm-r2r-…

8a8736b

…validation

lewing marked this pull request as ready for review July 25, 2026 18:51

Merge branch 'main' into lewing-cdac-wasm-r2r-validation

66f79a2

Copilot AI review requested due to automatic review settings July 25, 2026 18:51

Copilot started reviewing on behalf of lewing July 25, 2026 18:51 View session

Merge commit '66f79a20dcbe36b775f430e048e751bdb9a6b3c9' into lewing-c…

ccb0ca7

…dac-wasm-r2r-validation

Copilot AI reviewed Jul 25, 2026

View reviewed changes

Navigation Menu

Uh oh!

[cDAC] WebAssembly: real-data R2R stack-walk validation tests - #131161

[cDAC] WebAssembly: real-data R2R stack-walk validation tests#131161
lewing wants to merge 6 commits into
dotnet:mainfrom
lewing:lewing-cdac-wasm-r2r-validation

lewing commented Jul 21, 2026 •

edited

Loading

Uh oh!

azure-pipelines Bot commented Jul 21, 2026

Uh oh!

dotnet-policy-service Bot commented Jul 21, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

azure-pipelines Bot commented Jul 25, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

lewing commented Jul 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Tests

Comment corrections (no behavior change)

Context

Testing

Uh oh!

azure-pipelines Bot commented Jul 21, 2026

Uh oh!

dotnet-policy-service Bot commented Jul 21, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

azure-pipelines Bot commented Jul 25, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

lewing commented Jul 21, 2026 •

edited

Loading