[wasm][coreclr] Cache calli cookies by radekdoulik · Pull Request #127016 · dotnet/runtime

radekdoulik · 2026-04-16T17:14:06Z

Avoid repeated expensive calls to get calli cookie by caching it

Checked in HelloWorld

CalliCookie cache: 18 hits, 17 misses, 35 total (51.4% hit rate)

Avoid repeated expensive calls to get calli cookie by caching it

dotnet-policy-service · 2026-04-16T17:15:11Z

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

Copilot

Pull request overview

This PR reduces overhead in the WASM interpreter call path by caching the computed calli cookie on the MethodDesc, avoiding repeated (and potentially expensive) cookie computation for repeated invocations of the same managed method.

Changes:

Cache and reuse a per-MethodDesc calli cookie in InvokeManagedMethod on WASM.
Extend MethodDescCodeData (portable entrypoints + interpreter builds) to store a CalliCookie.
Add MethodDesc::{GetCalliCookie, SetCalliCookie} APIs to manage the cached value thread-safely.

Reviewed changes

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

File	Description
src/coreclr/vm/wasm/helpers.cpp	Uses `MethodDesc`-cached cookie to avoid repeated `GetCookieForCalliSig` calls.
src/coreclr/vm/method.hpp	Adds `CalliCookie` storage in `MethodDescCodeData` and declares accessor APIs under the relevant feature flags.
src/coreclr/vm/method.cpp	Implements thread-safe cookie get/set using `EnsureCodeDataExists` + interlocked/volatile operations.

…Cookie typedef - Define InterpreterCalliCookie typedef: function pointer on WASM, CallStubHeader* on non-WASM (Jan's feedback) - Unify SetCalliCookie/GetCalliCookie API, removing the separate SetCallStub/GetCallStub and FEATURE_PORTABLE_ENTRYPOINTS branching in MethodDesc - Cache calli cookie on targetMethod in the WASM calli path in interpexec.cpp (Jan's feedback) - Re-read cookie after SetCalliCookie in wasm/helpers.cpp to use the race-winning value (Aaron's feedback) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

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

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

radekdoulik · 2026-05-05T15:35:28Z

/ba-g infra issues

…127876) ## Problem PR #127016 updated `MethodDescCodeData::CallStub` added a new writer in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig` that caches a calli signature `CallStubHeader*` on the P/Invoke target `MethodDesc`. PR #127016 added a small "calli cookie" cache. Each managed method has one slot where this cookie is stored. The problem is that two different parts of the runtime now write to that one slot: - The interpreter-to-native call path stores a helper that's set up for the method's own signature - The new code in #127016 stores a helper that's set up for the signature of a `calli` instruction inside an IL stub. On iOS, when an app tries to execute address 0 the kernel does not give us a normal crash. It assumes the page is unsigned/tampered and kills the process with `SIGKILL` and a `CODESIGNING / Invalid Page` reason. ## Proposed fix Drop the new `pContextMD->{Set,Get}CalliCookie` cache reads/writes in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig`. The function now always computes the cookie via `GetCookieForCalliSig(sig, pContextMD)`, matching pre-#127016 behavior. Fixes#127869 #127863 #127867 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

radekdoulik · 2026-04-16T17:14:06Z

Avoid repeated expensive calls to get calli cookie by caching it

Checked in HelloWorld

CalliCookie cache: 18 hits, 17 misses, 35 total (51.4% hit rate)

Avoid repeated expensive calls to get calli cookie by caching it

dotnet-policy-service · 2026-04-16T17:15:11Z

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

Copilot

Pull request overview

This PR reduces overhead in the WASM interpreter call path by caching the computed calli cookie on the MethodDesc, avoiding repeated (and potentially expensive) cookie computation for repeated invocations of the same managed method.

Changes:

Cache and reuse a per-MethodDesc calli cookie in InvokeManagedMethod on WASM.
Extend MethodDescCodeData (portable entrypoints + interpreter builds) to store a CalliCookie.
Add MethodDesc::{GetCalliCookie, SetCalliCookie} APIs to manage the cached value thread-safely.

Reviewed changes

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

File	Description
src/coreclr/vm/wasm/helpers.cpp	Uses `MethodDesc`-cached cookie to avoid repeated `GetCookieForCalliSig` calls.
src/coreclr/vm/method.hpp	Adds `CalliCookie` storage in `MethodDescCodeData` and declares accessor APIs under the relevant feature flags.
src/coreclr/vm/method.cpp	Implements thread-safe cookie get/set using `EnsureCodeDataExists` + interlocked/volatile operations.

…Cookie typedef - Define InterpreterCalliCookie typedef: function pointer on WASM, CallStubHeader* on non-WASM (Jan's feedback) - Unify SetCalliCookie/GetCalliCookie API, removing the separate SetCallStub/GetCallStub and FEATURE_PORTABLE_ENTRYPOINTS branching in MethodDesc - Cache calli cookie on targetMethod in the WASM calli path in interpexec.cpp (Jan's feedback) - Re-read cookie after SetCalliCookie in wasm/helpers.cpp to use the race-winning value (Aaron's feedback) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

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

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

radekdoulik · 2026-05-05T15:35:28Z

/ba-g infra issues

…127876) ## Problem PR #127016 updated `MethodDescCodeData::CallStub` added a new writer in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig` that caches a calli signature `CallStubHeader*` on the P/Invoke target `MethodDesc`. PR #127016 added a small "calli cookie" cache. Each managed method has one slot where this cookie is stored. The problem is that two different parts of the runtime now write to that one slot: - The interpreter-to-native call path stores a helper that's set up for the method's own signature - The new code in #127016 stores a helper that's set up for the signature of a `calli` instruction inside an IL stub. On iOS, when an app tries to execute address 0 the kernel does not give us a normal crash. It assumes the page is unsigned/tampered and kills the process with `SIGKILL` and a `CODESIGNING / Invalid Page` reason. ## Proposed fix Drop the new `pContextMD->{Set,Get}CalliCookie` cache reads/writes in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig`. The function now always computes the cookie via `GetCookieForCalliSig(sig, pContextMD)`, matching pre-#127016 behavior. Fixes#127869 #127863 #127867 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

radekdoulik · 2026-04-16T17:14:06Z

Avoid repeated expensive calls to get calli cookie by caching it

Checked in HelloWorld

CalliCookie cache: 18 hits, 17 misses, 35 total (51.4% hit rate)

Avoid repeated expensive calls to get calli cookie by caching it

dotnet-policy-service · 2026-04-16T17:15:11Z

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

Copilot

Pull request overview

This PR reduces overhead in the WASM interpreter call path by caching the computed calli cookie on the MethodDesc, avoiding repeated (and potentially expensive) cookie computation for repeated invocations of the same managed method.

Changes:

Cache and reuse a per-MethodDesc calli cookie in InvokeManagedMethod on WASM.
Extend MethodDescCodeData (portable entrypoints + interpreter builds) to store a CalliCookie.
Add MethodDesc::{GetCalliCookie, SetCalliCookie} APIs to manage the cached value thread-safely.

Reviewed changes

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

File	Description
src/coreclr/vm/wasm/helpers.cpp	Uses `MethodDesc`-cached cookie to avoid repeated `GetCookieForCalliSig` calls.
src/coreclr/vm/method.hpp	Adds `CalliCookie` storage in `MethodDescCodeData` and declares accessor APIs under the relevant feature flags.
src/coreclr/vm/method.cpp	Implements thread-safe cookie get/set using `EnsureCodeDataExists` + interlocked/volatile operations.

…Cookie typedef - Define InterpreterCalliCookie typedef: function pointer on WASM, CallStubHeader* on non-WASM (Jan's feedback) - Unify SetCalliCookie/GetCalliCookie API, removing the separate SetCallStub/GetCallStub and FEATURE_PORTABLE_ENTRYPOINTS branching in MethodDesc - Cache calli cookie on targetMethod in the WASM calli path in interpexec.cpp (Jan's feedback) - Re-read cookie after SetCalliCookie in wasm/helpers.cpp to use the race-winning value (Aaron's feedback) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

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

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

radekdoulik · 2026-05-05T15:35:28Z

/ba-g infra issues

…127876) ## Problem PR #127016 updated `MethodDescCodeData::CallStub` added a new writer in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig` that caches a calli signature `CallStubHeader*` on the P/Invoke target `MethodDesc`. PR #127016 added a small "calli cookie" cache. Each managed method has one slot where this cookie is stored. The problem is that two different parts of the runtime now write to that one slot: - The interpreter-to-native call path stores a helper that's set up for the method's own signature - The new code in #127016 stores a helper that's set up for the signature of a `calli` instruction inside an IL stub. On iOS, when an app tries to execute address 0 the kernel does not give us a normal crash. It assumes the page is unsigned/tampered and kills the process with `SIGKILL` and a `CODESIGNING / Invalid Page` reason. ## Proposed fix Drop the new `pContextMD->{Set,Get}CalliCookie` cache reads/writes in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig`. The function now always computes the cookie via `GetCookieForCalliSig(sig, pContextMD)`, matching pre-#127016 behavior. Fixes#127869 #127863 #127867 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

radekdoulik · 2026-04-16T17:14:06Z

Avoid repeated expensive calls to get calli cookie by caching it

Checked in HelloWorld

CalliCookie cache: 18 hits, 17 misses, 35 total (51.4% hit rate)

Avoid repeated expensive calls to get calli cookie by caching it

dotnet-policy-service · 2026-04-16T17:15:11Z

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

Copilot

Pull request overview

This PR reduces overhead in the WASM interpreter call path by caching the computed calli cookie on the MethodDesc, avoiding repeated (and potentially expensive) cookie computation for repeated invocations of the same managed method.

Changes:

Cache and reuse a per-MethodDesc calli cookie in InvokeManagedMethod on WASM.
Extend MethodDescCodeData (portable entrypoints + interpreter builds) to store a CalliCookie.
Add MethodDesc::{GetCalliCookie, SetCalliCookie} APIs to manage the cached value thread-safely.

Reviewed changes

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

File	Description
src/coreclr/vm/wasm/helpers.cpp	Uses `MethodDesc`-cached cookie to avoid repeated `GetCookieForCalliSig` calls.
src/coreclr/vm/method.hpp	Adds `CalliCookie` storage in `MethodDescCodeData` and declares accessor APIs under the relevant feature flags.
src/coreclr/vm/method.cpp	Implements thread-safe cookie get/set using `EnsureCodeDataExists` + interlocked/volatile operations.

…Cookie typedef - Define InterpreterCalliCookie typedef: function pointer on WASM, CallStubHeader* on non-WASM (Jan's feedback) - Unify SetCalliCookie/GetCalliCookie API, removing the separate SetCallStub/GetCallStub and FEATURE_PORTABLE_ENTRYPOINTS branching in MethodDesc - Cache calli cookie on targetMethod in the WASM calli path in interpexec.cpp (Jan's feedback) - Re-read cookie after SetCalliCookie in wasm/helpers.cpp to use the race-winning value (Aaron's feedback) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

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

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

radekdoulik · 2026-05-05T15:35:28Z

/ba-g infra issues

…127876) ## Problem PR #127016 updated `MethodDescCodeData::CallStub` added a new writer in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig` that caches a calli signature `CallStubHeader*` on the P/Invoke target `MethodDesc`. PR #127016 added a small "calli cookie" cache. Each managed method has one slot where this cookie is stored. The problem is that two different parts of the runtime now write to that one slot: - The interpreter-to-native call path stores a helper that's set up for the method's own signature - The new code in #127016 stores a helper that's set up for the signature of a `calli` instruction inside an IL stub. On iOS, when an app tries to execute address 0 the kernel does not give us a normal crash. It assumes the page is unsigned/tampered and kills the process with `SIGKILL` and a `CODESIGNING / Invalid Page` reason. ## Proposed fix Drop the new `pContextMD->{Set,Get}CalliCookie` cache reads/writes in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig`. The function now always computes the cookie via `GetCookieForCalliSig(sig, pContextMD)`, matching pre-#127016 behavior. Fixes#127869 #127863 #127867 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

radekdoulik · 2026-04-16T17:14:06Z

Avoid repeated expensive calls to get calli cookie by caching it

Checked in HelloWorld

CalliCookie cache: 18 hits, 17 misses, 35 total (51.4% hit rate)

Avoid repeated expensive calls to get calli cookie by caching it

dotnet-policy-service · 2026-04-16T17:15:11Z

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

Copilot

Pull request overview

This PR reduces overhead in the WASM interpreter call path by caching the computed calli cookie on the MethodDesc, avoiding repeated (and potentially expensive) cookie computation for repeated invocations of the same managed method.

Changes:

Cache and reuse a per-MethodDesc calli cookie in InvokeManagedMethod on WASM.
Extend MethodDescCodeData (portable entrypoints + interpreter builds) to store a CalliCookie.
Add MethodDesc::{GetCalliCookie, SetCalliCookie} APIs to manage the cached value thread-safely.

Reviewed changes

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

File	Description
src/coreclr/vm/wasm/helpers.cpp	Uses `MethodDesc`-cached cookie to avoid repeated `GetCookieForCalliSig` calls.
src/coreclr/vm/method.hpp	Adds `CalliCookie` storage in `MethodDescCodeData` and declares accessor APIs under the relevant feature flags.
src/coreclr/vm/method.cpp	Implements thread-safe cookie get/set using `EnsureCodeDataExists` + interlocked/volatile operations.

…Cookie typedef - Define InterpreterCalliCookie typedef: function pointer on WASM, CallStubHeader* on non-WASM (Jan's feedback) - Unify SetCalliCookie/GetCalliCookie API, removing the separate SetCallStub/GetCallStub and FEATURE_PORTABLE_ENTRYPOINTS branching in MethodDesc - Cache calli cookie on targetMethod in the WASM calli path in interpexec.cpp (Jan's feedback) - Re-read cookie after SetCalliCookie in wasm/helpers.cpp to use the race-winning value (Aaron's feedback) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

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

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

radekdoulik · 2026-05-05T15:35:28Z

/ba-g infra issues

…127876) ## Problem PR #127016 updated `MethodDescCodeData::CallStub` added a new writer in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig` that caches a calli signature `CallStubHeader*` on the P/Invoke target `MethodDesc`. PR #127016 added a small "calli cookie" cache. Each managed method has one slot where this cookie is stored. The problem is that two different parts of the runtime now write to that one slot: - The interpreter-to-native call path stores a helper that's set up for the method's own signature - The new code in #127016 stores a helper that's set up for the signature of a `calli` instruction inside an IL stub. On iOS, when an app tries to execute address 0 the kernel does not give us a normal crash. It assumes the page is unsigned/tampered and kills the process with `SIGKILL` and a `CODESIGNING / Invalid Page` reason. ## Proposed fix Drop the new `pContextMD->{Set,Get}CalliCookie` cache reads/writes in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig`. The function now always computes the cookie via `GetCookieForCalliSig(sig, pContextMD)`, matching pre-#127016 behavior. Fixes#127869 #127863 #127867 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

radekdoulik · 2026-04-16T17:14:06Z

Avoid repeated expensive calls to get calli cookie by caching it

Checked in HelloWorld

CalliCookie cache: 18 hits, 17 misses, 35 total (51.4% hit rate)

Avoid repeated expensive calls to get calli cookie by caching it

dotnet-policy-service · 2026-04-16T17:15:11Z

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

Copilot

Pull request overview

This PR reduces overhead in the WASM interpreter call path by caching the computed calli cookie on the MethodDesc, avoiding repeated (and potentially expensive) cookie computation for repeated invocations of the same managed method.

Changes:

Cache and reuse a per-MethodDesc calli cookie in InvokeManagedMethod on WASM.
Extend MethodDescCodeData (portable entrypoints + interpreter builds) to store a CalliCookie.
Add MethodDesc::{GetCalliCookie, SetCalliCookie} APIs to manage the cached value thread-safely.

Reviewed changes

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

File	Description
src/coreclr/vm/wasm/helpers.cpp	Uses `MethodDesc`-cached cookie to avoid repeated `GetCookieForCalliSig` calls.
src/coreclr/vm/method.hpp	Adds `CalliCookie` storage in `MethodDescCodeData` and declares accessor APIs under the relevant feature flags.
src/coreclr/vm/method.cpp	Implements thread-safe cookie get/set using `EnsureCodeDataExists` + interlocked/volatile operations.

…Cookie typedef - Define InterpreterCalliCookie typedef: function pointer on WASM, CallStubHeader* on non-WASM (Jan's feedback) - Unify SetCalliCookie/GetCalliCookie API, removing the separate SetCallStub/GetCallStub and FEATURE_PORTABLE_ENTRYPOINTS branching in MethodDesc - Cache calli cookie on targetMethod in the WASM calli path in interpexec.cpp (Jan's feedback) - Re-read cookie after SetCalliCookie in wasm/helpers.cpp to use the race-winning value (Aaron's feedback) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

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

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

radekdoulik · 2026-05-05T15:35:28Z

/ba-g infra issues

…127876) ## Problem PR #127016 updated `MethodDescCodeData::CallStub` added a new writer in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig` that caches a calli signature `CallStubHeader*` on the P/Invoke target `MethodDesc`. PR #127016 added a small "calli cookie" cache. Each managed method has one slot where this cookie is stored. The problem is that two different parts of the runtime now write to that one slot: - The interpreter-to-native call path stores a helper that's set up for the method's own signature - The new code in #127016 stores a helper that's set up for the signature of a `calli` instruction inside an IL stub. On iOS, when an app tries to execute address 0 the kernel does not give us a normal crash. It assumes the page is unsigned/tampered and kills the process with `SIGKILL` and a `CODESIGNING / Invalid Page` reason. ## Proposed fix Drop the new `pContextMD->{Set,Get}CalliCookie` cache reads/writes in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig`. The function now always computes the cookie via `GetCookieForCalliSig(sig, pContextMD)`, matching pre-#127016 behavior. Fixes#127869 #127863 #127867 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

radekdoulik · 2026-04-16T17:14:06Z

Avoid repeated expensive calls to get calli cookie by caching it

Checked in HelloWorld

CalliCookie cache: 18 hits, 17 misses, 35 total (51.4% hit rate)

Avoid repeated expensive calls to get calli cookie by caching it

dotnet-policy-service · 2026-04-16T17:15:11Z

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

Copilot

Pull request overview

This PR reduces overhead in the WASM interpreter call path by caching the computed calli cookie on the MethodDesc, avoiding repeated (and potentially expensive) cookie computation for repeated invocations of the same managed method.

Changes:

Cache and reuse a per-MethodDesc calli cookie in InvokeManagedMethod on WASM.
Extend MethodDescCodeData (portable entrypoints + interpreter builds) to store a CalliCookie.
Add MethodDesc::{GetCalliCookie, SetCalliCookie} APIs to manage the cached value thread-safely.

Reviewed changes

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

File	Description
src/coreclr/vm/wasm/helpers.cpp	Uses `MethodDesc`-cached cookie to avoid repeated `GetCookieForCalliSig` calls.
src/coreclr/vm/method.hpp	Adds `CalliCookie` storage in `MethodDescCodeData` and declares accessor APIs under the relevant feature flags.
src/coreclr/vm/method.cpp	Implements thread-safe cookie get/set using `EnsureCodeDataExists` + interlocked/volatile operations.

…Cookie typedef - Define InterpreterCalliCookie typedef: function pointer on WASM, CallStubHeader* on non-WASM (Jan's feedback) - Unify SetCalliCookie/GetCalliCookie API, removing the separate SetCallStub/GetCallStub and FEATURE_PORTABLE_ENTRYPOINTS branching in MethodDesc - Cache calli cookie on targetMethod in the WASM calli path in interpexec.cpp (Jan's feedback) - Re-read cookie after SetCalliCookie in wasm/helpers.cpp to use the race-winning value (Aaron's feedback) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

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

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

radekdoulik · 2026-05-05T15:35:28Z

/ba-g infra issues

…127876) ## Problem PR #127016 updated `MethodDescCodeData::CallStub` added a new writer in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig` that caches a calli signature `CallStubHeader*` on the P/Invoke target `MethodDesc`. PR #127016 added a small "calli cookie" cache. Each managed method has one slot where this cookie is stored. The problem is that two different parts of the runtime now write to that one slot: - The interpreter-to-native call path stores a helper that's set up for the method's own signature - The new code in #127016 stores a helper that's set up for the signature of a `calli` instruction inside an IL stub. On iOS, when an app tries to execute address 0 the kernel does not give us a normal crash. It assumes the page is unsigned/tampered and kills the process with `SIGKILL` and a `CODESIGNING / Invalid Page` reason. ## Proposed fix Drop the new `pContextMD->{Set,Get}CalliCookie` cache reads/writes in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig`. The function now always computes the cookie via `GetCookieForCalliSig(sig, pContextMD)`, matching pre-#127016 behavior. Fixes#127869 #127863 #127867 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

radekdoulik · 2026-04-16T17:14:06Z

Avoid repeated expensive calls to get calli cookie by caching it

Checked in HelloWorld

CalliCookie cache: 18 hits, 17 misses, 35 total (51.4% hit rate)

Avoid repeated expensive calls to get calli cookie by caching it

dotnet-policy-service · 2026-04-16T17:15:11Z

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

Copilot

Pull request overview

This PR reduces overhead in the WASM interpreter call path by caching the computed calli cookie on the MethodDesc, avoiding repeated (and potentially expensive) cookie computation for repeated invocations of the same managed method.

Changes:

Cache and reuse a per-MethodDesc calli cookie in InvokeManagedMethod on WASM.
Extend MethodDescCodeData (portable entrypoints + interpreter builds) to store a CalliCookie.
Add MethodDesc::{GetCalliCookie, SetCalliCookie} APIs to manage the cached value thread-safely.

Reviewed changes

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

File	Description
src/coreclr/vm/wasm/helpers.cpp	Uses `MethodDesc`-cached cookie to avoid repeated `GetCookieForCalliSig` calls.
src/coreclr/vm/method.hpp	Adds `CalliCookie` storage in `MethodDescCodeData` and declares accessor APIs under the relevant feature flags.
src/coreclr/vm/method.cpp	Implements thread-safe cookie get/set using `EnsureCodeDataExists` + interlocked/volatile operations.

…Cookie typedef - Define InterpreterCalliCookie typedef: function pointer on WASM, CallStubHeader* on non-WASM (Jan's feedback) - Unify SetCalliCookie/GetCalliCookie API, removing the separate SetCallStub/GetCallStub and FEATURE_PORTABLE_ENTRYPOINTS branching in MethodDesc - Cache calli cookie on targetMethod in the WASM calli path in interpexec.cpp (Jan's feedback) - Re-read cookie after SetCalliCookie in wasm/helpers.cpp to use the race-winning value (Aaron's feedback) Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated 3 comments.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 2 comments.

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

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 1 comment.

Copilot

Pull request overview

Copilot reviewed 10 out of 10 changed files in this pull request and generated 3 comments.

radekdoulik · 2026-05-05T15:35:28Z

/ba-g infra issues

…127876) ## Problem PR #127016 updated `MethodDescCodeData::CallStub` added a new writer in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig` that caches a calli signature `CallStubHeader*` on the P/Invoke target `MethodDesc`. PR #127016 added a small "calli cookie" cache. Each managed method has one slot where this cookie is stored. The problem is that two different parts of the runtime now write to that one slot: - The interpreter-to-native call path stores a helper that's set up for the method's own signature - The new code in #127016 stores a helper that's set up for the signature of a `calli` instruction inside an IL stub. On iOS, when an app tries to execute address 0 the kernel does not give us a normal crash. It assumes the page is unsigned/tampered and kills the process with `SIGKILL` and a `CODESIGNING / Invalid Page` reason. ## Proposed fix Drop the new `pContextMD->{Set,Get}CalliCookie` cache reads/writes in `CInterpreterJitInfo::GetCookieForInterpreterCalliSig`. The function now always computes the cookie via `GetCookieForCalliSig(sig, pContextMD)`, matching pre-#127016 behavior. Fixes#127869 #127863 #127867 --------- Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

[wasm][coreclr] Cache calli cookies
644cd59
Avoid repeated expensive calls to get calli cookie by caching it

radekdoulik added this to the Future milestone Apr 16, 2026

radekdoulik added arch-wasm WebAssembly architecture area-VM-coreclr labels Apr 16, 2026

radekdoulik requested review from AaronRobinsonMSFT, Copilot and jkotas April 16, 2026 17:14

dotnet-policy-serviceBot assigned radekdoulik Apr 16, 2026

Copilotstarted reviewing on behalf of radekdoulik April 16, 2026 17:15 View session

CopilotAI reviewed Apr 16, 2026
View reviewed changes

AaronRobinsonMSFT reviewed Apr 16, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/wasm/helpers.cpp

build-analysisBot mentioned this pull request Apr 17, 2026
[browser] Could not resolve type with token 01000001 from typeref #126997
Open

jkotas reviewed Apr 17, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/wasm/helpers.cpp

jkotas reviewed Apr 17, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/method.cpp Outdated

radekdoulik requested review from BrzVlad, janvorli and kg as code owners April 26, 2026 20:01

Cache the cookie in CInterpreterJitInfo::GetCookieForInterpreterCalliSig
d7a3255

CopilotAI review requested due to automatic review settings April 26, 2026 20:29

Copilotstarted reviewing on behalf of radekdoulik April 26, 2026 20:30 View session

CopilotAI reviewed Apr 26, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/wasm/helpers.cpp
Comment threadsrc/coreclr/vm/jitinterface.cpp Outdated
Comment threadsrc/coreclr/vm/interpexec.cpp Outdated

This was referenced Apr 27, 2026
XHarness package install failure on iOS due to devicectl NSPOSIXErrorDomain error 49 #123796
Open
CONTRACT VIOLATION by PInvokeImportWorker #127231
Closed

jkotas approved these changes Apr 27, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/interpexec.cpp Outdated

radekdoulik added 2 commits April 27, 2026 12:58

Feedback
09c7676

Fix build
bbf1b3f

CopilotAI review requested due to automatic review settings April 27, 2026 13:09

Copilotstarted reviewing on behalf of radekdoulik April 27, 2026 13:10 View session

Copilotstarted reviewing on behalf of radekdoulik May 3, 2026 19:59 View session

CopilotAI reviewed May 3, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/method.hpp
Comment threadsrc/mono/browser/build/BrowserWasmApp.CoreCLR.targets Outdated

jkotas reviewed May 3, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/precode_portable.hpp Outdated

jkotas reviewed May 3, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/wasm/callhelpers-interp-to-managed.cpp Outdated

jkotas reviewed May 3, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/wasm/callhelpers-interp-to-managed.cpp Outdated

Feedback
c20fa67
Co-authored-by: Jan Kotas <jkotas@microsoft.com>

CopilotAI review requested due to automatic review settings May 4, 2026 10:54

Copilotstarted reviewing on behalf of radekdoulik May 4, 2026 10:54 View session

Feedback
a871a22

CopilotAI reviewed May 4, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/jitinterface.cpp

radekdoulik added 2 commits May 4, 2026 14:27

Don't forget the pointer dereference
022523e

Regenerate and fix non-pe case
bec30ff

CopilotAI review requested due to automatic review settings May 4, 2026 12:35

Copilotstarted reviewing on behalf of radekdoulik May 4, 2026 12:35 View session

CopilotAI reviewed May 4, 2026
View reviewed changes

Comment threadsrc/coreclr/vm/jitinterface.cpp
Comment threadsrc/tasks/WasmAppBuilder/coreclr/InterpToNativeGenerator.cs
Comment threadsrc/coreclr/vm/wasm/callhelpers-pinvoke.cpp

Merge branch 'main' into pe-cookie-cache
4027a23

radekdoulik merged commit 7018450 into dotnet:mainMay 5, 2026
122 of 126 checks passed

dotnet-maestroBot mentioned this pull request May 6, 2026
[main] Source code updates from dotnet/runtime dotnet/dotnet#6423
Merged

kotlarmilos mentioned this pull request May 6, 2026
Fix iOS arm64 CoreCLR crash from shared MethodDesc::CalliCookie slot #127876
Merged

kotlarmilos mentioned this pull request May 12, 2026
Interpreter SIGSEGV in latest maestro bump #128064
Closed

This was referenced May 13, 2026
[Perf] Linux/x64: 335 Improvements on 5/5/2026 5:18:31 PM +00:00 dotnet/perf-autofiling-issues#72828
Open
[Perf] Linux/x64: 253 Improvements on 5/5/2026 5:18:31 PM +00:00 dotnet/perf-autofiling-issues#72667
Open

kotlarmilos mentioned this pull request Jun 4, 2026
Stackoverflow at startup with CoreCLR on ios simulator x64 #128898
Closed

github-actionsBot locked and limited conversation to collaborators Jun 5, 2026