Fix interpreter delegate calls through shuffle thunks

[BrzVlad]

StubLinker-generated wrapper and shuffle stubs are now tracked by the JIT code manager. NonVirtualEntry2MethodDesc can therefore find their code ranges even though they do not map directly to managed method bodies. The assertion path was hit from the interpreter delegate dispatch, passing an address to a shuffle thunk.

runtime/src/coreclr/vm/interpexec.cpp

Line 3503 in 643b896

if ((targetMethod = NonVirtualEntry2MethodDesc(targetAddress)) != NULL)

Recognize these stub block kinds and return null so callers use their existing non-method fallback path. Fixes assertions on the interpreter pipelines

ASSERT FAILED
Expression: !"NonVirtualEntry2MethodDesc failed"
Location: /__w/1/s/src/coreclr/vm/method.cpp:2428
Function: NonVirtualEntry2MethodDesc
Process: 90146

Regressed after #131903

[azure-pipelines]

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.

[BrzVlad]

/azp run runtime-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[BrzVlad]

/azp run runtime-libraries-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR adjusts NonVirtualEntry2MethodDesc to recognize additional non-method stub code block kinds (wrapper stubs and shuffle thunks) and return NULL so callers can take their existing non-method fallback path instead of hitting the debug assertion when given an entrypoint that doesn’t map to a managed method body.

Changes:

• Cache the StubCodeBlockKind for RangeList sections to avoid repeated GetCodeBlockKind() calls.
• After JitCodeToMethodInfo fails, query the stub code block kind and return NULL for STUB_CODE_BLOCK_WRAPPER_STUB and STUB_CODE_BLOCK_SHUFFLE_THUNK to avoid the assertion path.

[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/method.cpp:2427

• Dropping the debug _ASSERTE entirely removes a useful invariant check for genuinely unexpected fallthroughs. Given the current logic, fixup/stub precodes return early, so reaching the end of the function with a RANGE_SECTION_RANGELIST should now only be expected for wrapper stubs and shuffle thunks (which don’t map to a MethodDesc). Consider restoring a debug-only assert that limits the new “allowed” cases to those stub kinds, so other accidental regressions still trip early in debug builds.

return NULL;

[BrzVlad]

/ba-g failures unrelated

[BrzVlad]

StubLinker-generated wrapper and shuffle stubs are now tracked by the JIT code manager. NonVirtualEntry2MethodDesc can therefore find their code ranges even though they do not map directly to managed method bodies. The assertion path was hit from the interpreter delegate dispatch, passing an address to a shuffle thunk.

runtime/src/coreclr/vm/interpexec.cpp

Line 3503 in 643b896

if ((targetMethod = NonVirtualEntry2MethodDesc(targetAddress)) != NULL)

Recognize these stub block kinds and return null so callers use their existing non-method fallback path. Fixes assertions on the interpreter pipelines

ASSERT FAILED
Expression: !"NonVirtualEntry2MethodDesc failed"
Location: /__w/1/s/src/coreclr/vm/method.cpp:2428
Function: NonVirtualEntry2MethodDesc
Process: 90146

Regressed after #131903

[azure-pipelines]

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.

[BrzVlad]

/azp run runtime-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[BrzVlad]

/azp run runtime-libraries-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR adjusts NonVirtualEntry2MethodDesc to recognize additional non-method stub code block kinds (wrapper stubs and shuffle thunks) and return NULL so callers can take their existing non-method fallback path instead of hitting the debug assertion when given an entrypoint that doesn’t map to a managed method body.

Changes:

• Cache the StubCodeBlockKind for RangeList sections to avoid repeated GetCodeBlockKind() calls.
• After JitCodeToMethodInfo fails, query the stub code block kind and return NULL for STUB_CODE_BLOCK_WRAPPER_STUB and STUB_CODE_BLOCK_SHUFFLE_THUNK to avoid the assertion path.

[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/method.cpp:2427

• Dropping the debug _ASSERTE entirely removes a useful invariant check for genuinely unexpected fallthroughs. Given the current logic, fixup/stub precodes return early, so reaching the end of the function with a RANGE_SECTION_RANGELIST should now only be expected for wrapper stubs and shuffle thunks (which don’t map to a MethodDesc). Consider restoring a debug-only assert that limits the new “allowed” cases to those stub kinds, so other accidental regressions still trip early in debug builds.

return NULL;

[BrzVlad]

/ba-g failures unrelated

[BrzVlad]

StubLinker-generated wrapper and shuffle stubs are now tracked by the JIT code manager. NonVirtualEntry2MethodDesc can therefore find their code ranges even though they do not map directly to managed method bodies. The assertion path was hit from the interpreter delegate dispatch, passing an address to a shuffle thunk.

runtime/src/coreclr/vm/interpexec.cpp

Line 3503 in 643b896

if ((targetMethod = NonVirtualEntry2MethodDesc(targetAddress)) != NULL)

Recognize these stub block kinds and return null so callers use their existing non-method fallback path. Fixes assertions on the interpreter pipelines

ASSERT FAILED
Expression: !"NonVirtualEntry2MethodDesc failed"
Location: /__w/1/s/src/coreclr/vm/method.cpp:2428
Function: NonVirtualEntry2MethodDesc
Process: 90146

Regressed after #131903

[azure-pipelines]

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.

[BrzVlad]

/azp run runtime-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[BrzVlad]

/azp run runtime-libraries-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR adjusts NonVirtualEntry2MethodDesc to recognize additional non-method stub code block kinds (wrapper stubs and shuffle thunks) and return NULL so callers can take their existing non-method fallback path instead of hitting the debug assertion when given an entrypoint that doesn’t map to a managed method body.

Changes:

• Cache the StubCodeBlockKind for RangeList sections to avoid repeated GetCodeBlockKind() calls.
• After JitCodeToMethodInfo fails, query the stub code block kind and return NULL for STUB_CODE_BLOCK_WRAPPER_STUB and STUB_CODE_BLOCK_SHUFFLE_THUNK to avoid the assertion path.

[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/method.cpp:2427

• Dropping the debug _ASSERTE entirely removes a useful invariant check for genuinely unexpected fallthroughs. Given the current logic, fixup/stub precodes return early, so reaching the end of the function with a RANGE_SECTION_RANGELIST should now only be expected for wrapper stubs and shuffle thunks (which don’t map to a MethodDesc). Consider restoring a debug-only assert that limits the new “allowed” cases to those stub kinds, so other accidental regressions still trip early in debug builds.

return NULL;

[BrzVlad]

/ba-g failures unrelated

[BrzVlad]

StubLinker-generated wrapper and shuffle stubs are now tracked by the JIT code manager. NonVirtualEntry2MethodDesc can therefore find their code ranges even though they do not map directly to managed method bodies. The assertion path was hit from the interpreter delegate dispatch, passing an address to a shuffle thunk.

runtime/src/coreclr/vm/interpexec.cpp

Line 3503 in 643b896

if ((targetMethod = NonVirtualEntry2MethodDesc(targetAddress)) != NULL)

Recognize these stub block kinds and return null so callers use their existing non-method fallback path. Fixes assertions on the interpreter pipelines

ASSERT FAILED
Expression: !"NonVirtualEntry2MethodDesc failed"
Location: /__w/1/s/src/coreclr/vm/method.cpp:2428
Function: NonVirtualEntry2MethodDesc
Process: 90146

Regressed after #131903

[azure-pipelines]

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.

[BrzVlad]

/azp run runtime-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[BrzVlad]

/azp run runtime-libraries-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR adjusts NonVirtualEntry2MethodDesc to recognize additional non-method stub code block kinds (wrapper stubs and shuffle thunks) and return NULL so callers can take their existing non-method fallback path instead of hitting the debug assertion when given an entrypoint that doesn’t map to a managed method body.

Changes:

• Cache the StubCodeBlockKind for RangeList sections to avoid repeated GetCodeBlockKind() calls.
• After JitCodeToMethodInfo fails, query the stub code block kind and return NULL for STUB_CODE_BLOCK_WRAPPER_STUB and STUB_CODE_BLOCK_SHUFFLE_THUNK to avoid the assertion path.

[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/method.cpp:2427

• Dropping the debug _ASSERTE entirely removes a useful invariant check for genuinely unexpected fallthroughs. Given the current logic, fixup/stub precodes return early, so reaching the end of the function with a RANGE_SECTION_RANGELIST should now only be expected for wrapper stubs and shuffle thunks (which don’t map to a MethodDesc). Consider restoring a debug-only assert that limits the new “allowed” cases to those stub kinds, so other accidental regressions still trip early in debug builds.

return NULL;

[BrzVlad]

/ba-g failures unrelated

[BrzVlad]

StubLinker-generated wrapper and shuffle stubs are now tracked by the JIT code manager. NonVirtualEntry2MethodDesc can therefore find their code ranges even though they do not map directly to managed method bodies. The assertion path was hit from the interpreter delegate dispatch, passing an address to a shuffle thunk.

runtime/src/coreclr/vm/interpexec.cpp

Line 3503 in 643b896

if ((targetMethod = NonVirtualEntry2MethodDesc(targetAddress)) != NULL)

Recognize these stub block kinds and return null so callers use their existing non-method fallback path. Fixes assertions on the interpreter pipelines

ASSERT FAILED
Expression: !"NonVirtualEntry2MethodDesc failed"
Location: /__w/1/s/src/coreclr/vm/method.cpp:2428
Function: NonVirtualEntry2MethodDesc
Process: 90146

Regressed after #131903

[azure-pipelines]

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.

[BrzVlad]

/azp run runtime-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[BrzVlad]

/azp run runtime-libraries-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR adjusts NonVirtualEntry2MethodDesc to recognize additional non-method stub code block kinds (wrapper stubs and shuffle thunks) and return NULL so callers can take their existing non-method fallback path instead of hitting the debug assertion when given an entrypoint that doesn’t map to a managed method body.

Changes:

• Cache the StubCodeBlockKind for RangeList sections to avoid repeated GetCodeBlockKind() calls.
• After JitCodeToMethodInfo fails, query the stub code block kind and return NULL for STUB_CODE_BLOCK_WRAPPER_STUB and STUB_CODE_BLOCK_SHUFFLE_THUNK to avoid the assertion path.

[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/method.cpp:2427

• Dropping the debug _ASSERTE entirely removes a useful invariant check for genuinely unexpected fallthroughs. Given the current logic, fixup/stub precodes return early, so reaching the end of the function with a RANGE_SECTION_RANGELIST should now only be expected for wrapper stubs and shuffle thunks (which don’t map to a MethodDesc). Consider restoring a debug-only assert that limits the new “allowed” cases to those stub kinds, so other accidental regressions still trip early in debug builds.

return NULL;

[BrzVlad]

/ba-g failures unrelated

[BrzVlad]

StubLinker-generated wrapper and shuffle stubs are now tracked by the JIT code manager. NonVirtualEntry2MethodDesc can therefore find their code ranges even though they do not map directly to managed method bodies. The assertion path was hit from the interpreter delegate dispatch, passing an address to a shuffle thunk.

runtime/src/coreclr/vm/interpexec.cpp

Line 3503 in 643b896

if ((targetMethod = NonVirtualEntry2MethodDesc(targetAddress)) != NULL)

Recognize these stub block kinds and return null so callers use their existing non-method fallback path. Fixes assertions on the interpreter pipelines

ASSERT FAILED
Expression: !"NonVirtualEntry2MethodDesc failed"
Location: /__w/1/s/src/coreclr/vm/method.cpp:2428
Function: NonVirtualEntry2MethodDesc
Process: 90146

Regressed after #131903

[azure-pipelines]

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.

[BrzVlad]

/azp run runtime-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[BrzVlad]

/azp run runtime-libraries-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR adjusts NonVirtualEntry2MethodDesc to recognize additional non-method stub code block kinds (wrapper stubs and shuffle thunks) and return NULL so callers can take their existing non-method fallback path instead of hitting the debug assertion when given an entrypoint that doesn’t map to a managed method body.

Changes:

• Cache the StubCodeBlockKind for RangeList sections to avoid repeated GetCodeBlockKind() calls.
• After JitCodeToMethodInfo fails, query the stub code block kind and return NULL for STUB_CODE_BLOCK_WRAPPER_STUB and STUB_CODE_BLOCK_SHUFFLE_THUNK to avoid the assertion path.

[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/method.cpp:2427

• Dropping the debug _ASSERTE entirely removes a useful invariant check for genuinely unexpected fallthroughs. Given the current logic, fixup/stub precodes return early, so reaching the end of the function with a RANGE_SECTION_RANGELIST should now only be expected for wrapper stubs and shuffle thunks (which don’t map to a MethodDesc). Consider restoring a debug-only assert that limits the new “allowed” cases to those stub kinds, so other accidental regressions still trip early in debug builds.

return NULL;

[BrzVlad]

/ba-g failures unrelated

[BrzVlad]

StubLinker-generated wrapper and shuffle stubs are now tracked by the JIT code manager. NonVirtualEntry2MethodDesc can therefore find their code ranges even though they do not map directly to managed method bodies. The assertion path was hit from the interpreter delegate dispatch, passing an address to a shuffle thunk.

runtime/src/coreclr/vm/interpexec.cpp

Line 3503 in 643b896

if ((targetMethod = NonVirtualEntry2MethodDesc(targetAddress)) != NULL)

Recognize these stub block kinds and return null so callers use their existing non-method fallback path. Fixes assertions on the interpreter pipelines

ASSERT FAILED
Expression: !"NonVirtualEntry2MethodDesc failed"
Location: /__w/1/s/src/coreclr/vm/method.cpp:2428
Function: NonVirtualEntry2MethodDesc
Process: 90146

Regressed after #131903

[azure-pipelines]

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.

[BrzVlad]

/azp run runtime-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[BrzVlad]

/azp run runtime-libraries-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR adjusts NonVirtualEntry2MethodDesc to recognize additional non-method stub code block kinds (wrapper stubs and shuffle thunks) and return NULL so callers can take their existing non-method fallback path instead of hitting the debug assertion when given an entrypoint that doesn’t map to a managed method body.

Changes:

• Cache the StubCodeBlockKind for RangeList sections to avoid repeated GetCodeBlockKind() calls.
• After JitCodeToMethodInfo fails, query the stub code block kind and return NULL for STUB_CODE_BLOCK_WRAPPER_STUB and STUB_CODE_BLOCK_SHUFFLE_THUNK to avoid the assertion path.

[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/method.cpp:2427

• Dropping the debug _ASSERTE entirely removes a useful invariant check for genuinely unexpected fallthroughs. Given the current logic, fixup/stub precodes return early, so reaching the end of the function with a RANGE_SECTION_RANGELIST should now only be expected for wrapper stubs and shuffle thunks (which don’t map to a MethodDesc). Consider restoring a debug-only assert that limits the new “allowed” cases to those stub kinds, so other accidental regressions still trip early in debug builds.

return NULL;

[BrzVlad]

/ba-g failures unrelated

[BrzVlad]

StubLinker-generated wrapper and shuffle stubs are now tracked by the JIT code manager. NonVirtualEntry2MethodDesc can therefore find their code ranges even though they do not map directly to managed method bodies. The assertion path was hit from the interpreter delegate dispatch, passing an address to a shuffle thunk.

runtime/src/coreclr/vm/interpexec.cpp

Line 3503 in 643b896

if ((targetMethod = NonVirtualEntry2MethodDesc(targetAddress)) != NULL)

Recognize these stub block kinds and return null so callers use their existing non-method fallback path. Fixes assertions on the interpreter pipelines

ASSERT FAILED
Expression: !"NonVirtualEntry2MethodDesc failed"
Location: /__w/1/s/src/coreclr/vm/method.cpp:2428
Function: NonVirtualEntry2MethodDesc
Process: 90146

Regressed after #131903

[azure-pipelines]

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.

[BrzVlad]

/azp run runtime-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[BrzVlad]

/azp run runtime-libraries-interpreter

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[dotnet-policy-service]

Tagging subscribers to this area: @JulieLeeMSFT, @BrzVlad, @janvorli
See info in area-owners.md if you want to be subscribed.

[Copilot]

Pull request overview

This PR adjusts NonVirtualEntry2MethodDesc to recognize additional non-method stub code block kinds (wrapper stubs and shuffle thunks) and return NULL so callers can take their existing non-method fallback path instead of hitting the debug assertion when given an entrypoint that doesn’t map to a managed method body.

Changes:

• Cache the StubCodeBlockKind for RangeList sections to avoid repeated GetCodeBlockKind() calls.
• After JitCodeToMethodInfo fails, query the stub code block kind and return NULL for STUB_CODE_BLOCK_WRAPPER_STUB and STUB_CODE_BLOCK_SHUFFLE_THUNK to avoid the assertion path.

[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/method.cpp:2427

• Dropping the debug _ASSERTE entirely removes a useful invariant check for genuinely unexpected fallthroughs. Given the current logic, fixup/stub precodes return early, so reaching the end of the function with a RANGE_SECTION_RANGELIST should now only be expected for wrapper stubs and shuffle thunks (which don’t map to a MethodDesc). Consider restoring a debug-only assert that limits the new “allowed” cases to those stub kinds, so other accidental regressions still trip early in debug builds.

return NULL;

[BrzVlad]

/ba-g failures unrelated