Enable GC stress for ReadyToRun code by Copilot · Pull Request #135794 · dotnet/runtime

Copilot · 2026-10-14T20:58:27Z

Instruction-level GC stress currently instruments JIT-compiled code but skips ReadyToRun methods.

Instrumentation: Call SetupGcCoverage before publishing R2R entry points, including async resumption stubs. Serialize setup and use canonical method identity/default code versions for shared generic code.
Patching: Support image-backed writes throughout instrumentation and instruction restoration, retaining executable-allocator mappings for JIT code.
Regression coverage: Verify collections and live-reference preservation in ordinary and shared-generic R2R methods.
Platform scope: Exclude Apple platforms and document the image-patching limitation.

Fixes Implement GC stress for R2R code. #108249

azure-pipelines · 2026-10-14T20:58:44Z

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 · 2026-10-14T21:01:04Z

Tagging subscribers to this area: @anicka-net, @dotnet/gc
See info in area-owners.md if you want to be subscribed.

VSadov · 2026-10-16T19:36:50Z

/azp run runtime-coreclr gcstress-extra, runtime-coreclr gcstress0x3-gcstress0xc

azure-pipelines · 2026-10-16T19:37:16Z

Azure Pipelines: Successfully started running 2 pipeline(s).

Co-authored-by: VSadov <8218165+VSadov@users.noreply.github.com>

VSadov · 2026-10-17T17:31:59Z

/azp run runtime-coreclr gcstress-extra, runtime-coreclr gcstress0x3-gcstress0xc

VSadov · 2026-11-04T21:35:49Z

@copilot - we do not want to add tests with this PR since. This is a test infrastructure change. There are plenty of tests that it runs already. Please remove all added tests.

Co-authored-by: VSadov <8218165+VSadov@users.noreply.github.com>

Copilot · 2026-11-04T21:39:49Z

@copilot - we do not want to add tests with this PR since. This is a test infrastructure change. There are plenty of tests that it runs already. Please remove all added tests.

Removed all four GC stress test source/project files added by this PR in commit 8b4e…. The runtime implementation and documentation changes remain.

VSadov · 2026-11-05T00:32:40Z

/azp run runtime-coreclr gcstress-extra, runtime-coreclr gcstress0x3-gcstress0xc

VSadov · 2026-11-05T00:32:53Z

/azp run runtime-coreclr r2r-extra

azure-pipelines · 2026-11-05T00:33:03Z

Azure Pipelines: Successfully started running 2 pipeline(s).

azure-pipelines · 2026-11-05T00:33:04Z

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

VSadov · 2026-11-05T20:54:07Z

The PR matches the baseline, with essentially identical failures. This time both PRs were run together, so the comparison is same-day. The PR head is 9dbdd81 and the baseline head is 447af12. Both pipelines started at about 00:33Z on Nov 5.

Scenarios: no difference. Both ran the same 51 jobs, and the same 6 top-level jobs failed (3 pipelines plus their Helix monitors).

	PR #135794	Baseline #136001
Build Analysis test failures	141	142
gcstress-extra, not displayed	68	68
gcstress0x3-0xc, not displayed	36	36
r2r-extra, not displayed	32	33

Displayed failures are the same on both sides:
- Callback* and MainProgramHandle (72 configs in gcstress-extra on each side)
- CollectibleTLSStaticCollection timeouts (Test failure: CollectibleTLSStaticCollection Timed Out #136891), 50 on each side
- Loader.0.3.WorkItemExecution, a BadExit crash on osx-arm64 gcstress0x3 with a similar failure rate on both sides (62.79% vs 63.64%). It is pre-existing and flaky, not caused by the PR.
Nothing PR-only is visible. The earlier JIT/opt/OSR/tailrecurse hit and the Windows x86 gcstress0xc_* crashes did not reappear. The earlier one-off AF: *(_UNCHECKED_OBJECTREF *)handle == NULL (HndCreateHandle called by getJitHandleForObject) #117138 hit is also gone.
x86 is clean: all Windows x86 build and test legs succeeded in the PR.

Caveats:

About 136 failures per run are hidden by the Build Analysis display limit. "Same failures" rests on matching totals and the visible entries.
AzDO and Helix were unreachable, so I couldn't get raw test-result counts or console logs.

I haven't checked which PR commit fixed the x86 crashes. They were already gone at the previous check.

jkotas

LGTM!

VSadov · 2026-11-10T19:10:20Z

/ba-g test failures unrelated/known. stress failures are the same as in a no-op run.

@jkotas

Thanks for asking me to work on this. I will get started on it and keep this PR's description up to date as I form a plan and make progress. > ---- > > *This section details on the original issue you should resolve* > > <issue_title>Assert failure: !th.IsNull()</issue_title> > <issue_description>## Build Information > Build: https://dev.azure.com/dnceng-public/cbb18261-c48f-4abb-8651-8cdcb5474649/_build/results?buildId=1654712 > Build error leg or test failing: OutOfProcessTest.JIT/Regression_3/DevDiv_461649/DevDiv_461649/DevDiv_461649.cmd >  > ## Error Message > > Fill the error message using [step by step known issues guidance](https://github.com/dotnet/arcade/blob/main/Documentation/Build%20Analysis/KnownIssueJsonStepByStep.md). > >  > > ```json > { > "ErrorMessage": "Failed test: OutOfProcessTest.JIT/Regression_3/DevDiv_461649/DevDiv_461649/DevDiv_461649.cmd", > "BuildRetry": false, > "ExcludeConsoleLog": false > } > ``` > >  > ### Known issue validation > **Build: 🔎** https://dev.azure.com/dnceng-public/public/_build/results?buildId=1654712 > **Error message validated:** `[Failed test: OutOfProcessTest.JIT/Regression_3/DevDiv_461649/DevDiv_461649/DevDiv_461649.cmd`] > **Result validation:** ❌ Known issue did not match with the provided build. > **Validation performed at:** 10/30/2026 1:07:04 AM UTC >  >  > > ### Report > #### Summary > |24-Hour Hit Count|7-Day Hit Count|1-Month Count| > |---|---|---| > |0|0|0| > </issue_description> > - Closes #136674 > > <agent_instructions>Re: also fails in #135794 > > **What fails** > > On x86 (non-Unix), `ResolveWorker` builds a resolve stub for a virtual or interface call. The stub needs the number of stack bytes the callee pops, so it calls `pMD->SizeOfArgStack()` on the method from the dispatch token. That call walks the method signature. For a value-type parameter or return it needs the type layout already loaded, because it runs under `ENABLE_FORBID_GC_LOADER_USE`. Otherwise `!th.IsNull()` asserts. > > **Why it started** > > Before #135805, the R2R caller's fixups loaded those types. Now `ExternalMethodFixupWorker` calls `PrepareForUseAsAVirtualCallTargetOfANativeImage(pGCRefMap)`. For a plain instance method with an import map, it only stores the map via `SetGCRefMap` and returns. So the method has a map but its value types may be unloaded. > > **The fix** > > The x86 GCRefMap already encodes the stack-pop size. The runtime already trusts it for the same purpose in `UpdateRegDisplay_Impl` (via `CbStackPopUsingGCRefMap`). For non-vararg methods, `CbStackPop()` equals `SizeOfArgStack()`. > > ```cpp > PTR_BYTE pGCRefMap = pMD->FindGCRefMap(objectType, true /* isDispatch */); > if (pGCRefMap != NULL) > { > GCRefMapDecoder decoder(pGCRefMap); > stackArgumentsSize = decoder.ReadStackPop() * sizeof(TADDR); > } > else > { > ENABLE_FORBID_GC_LOADER_USE_IN_THIS_SCOPE(); > stackArgumentsSize = pMD->SizeOfArgStack(); // unchanged fallback > } > ``` > > - It uses the same lookup `StubDispatchFrame` uses for GC scanning. > - Nothing is loaded and the scope isn't relaxed. It also adds `#include "gcrefmap.h"`, and it stays inside the existing x86 `#if`. > - The fallback should be safe. As far as I can tell from reading, preparation skips the load only when it stores a shareable map. Generic methods and methods that need a hidden generic argument are fully prepared. I haven't confirmed that. > > </agent_instructions> > > ## Comments on the Issue (you are @copilot in this section) > > <comments> > <comment_new><author>@jkotas</author><body> > Regression likely caused by #135806 cc @davidwrighton </body></comment_new> > </comments> - Fixes #136674 --------- Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com> Co-authored-by: VSadov <8218165+VSadov@users.noreply.github.com> Co-authored-by: David Wrighton <davidwr@microsoft.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> Copilot-Session: 6e1d3f7f-a06c-4561-8189-bdf699656b46

Copilot AI assigned Copilot and VSadov Oct 14, 2026

Copilot started work on behalf of VSadov October 14, 2026 20:58

View session

Copilot AI linked an issue Oct 14, 2026 that may be closed by this pull request

Implement GC stress for R2R code. #108249

Closed

github-actions Bot added the area-GC-coreclr label Oct 14, 2026

Copilot AI changed the title ~~[WIP] Add support for GC stress in R2R code~~ Enable GC stress for ReadyToRun code Oct 14, 2026

Copilot finished work on behalf of VSadov October 14, 2026 21:46

Copilot AI requested a review from VSadov October 14, 2026 21:46

janvorli reviewed Oct 15, 2026

View reviewed changes

Comment thread src/coreclr/vm/gccover.cpp Resolved

Copilot AI mentioned this pull request Oct 15, 2026

Fix helperildll build race in Crossgen2 determinism test #135878

Merged

Copilot started work on behalf of VSadov October 16, 2026 18:15

View session

Copilot finished work on behalf of VSadov October 16, 2026 18:32

Copilot AI and others added 2 commits October 17, 2026 10:31

Enable GC stress instrumentation for ReadyToRun code

afcafb0

Co-authored-by: VSadov <8218165+VSadov@users.noreply.github.com>

Fail fast when GC stress cannot change code protection

c9bf9db

Co-authored-by: VSadov <8218165+VSadov@users.noreply.github.com>

VSadov force-pushed the copilot/implement-gc-stress-for-r2r-code branch from 8909fd0 to c9bf9db Compare October 17, 2026 17:31

Copilot finished work on behalf of VSadov November 4, 2026 20:35

Copilot started work on behalf of VSadov November 4, 2026 21:36

View session

Remove GC stress tests from R2R change

9dbdd81

Co-authored-by: VSadov <8218165+VSadov@users.noreply.github.com>

Copilot finished work on behalf of VSadov November 4, 2026 21:40

VSadov requested a balanced review from Copilot November 4, 2026 22:13

Copilot started reviewing on behalf of VSadov November 4, 2026 22:13

View session

This comment was marked as outdated.

Sign in to view

build-analysis Bot mentioned this pull request Nov 5, 2026

TensorExtensionsSpanInSpanOut<Single> test failure #135718

Open

VSadov mentioned this pull request Nov 5, 2026

[release/11.0] Fix GC hole caused by adrp and add instructions on arm64 #136945

Open

2 of 4 tasks

VSadov requested review from jkotas and a balanced review from Copilot November 5, 2026 20:54

Copilot started reviewing on behalf of VSadov November 5, 2026 20:54

View session

This comment was marked as resolved.

Sign in to view

VSadov requested a balanced review from Copilot November 5, 2026 22:27

Copilot started reviewing on behalf of VSadov November 5, 2026 22:28

View session

This comment was marked as resolved.

Sign in to view

jkotas approved these changes Nov 10, 2026

View reviewed changes

VSadov merged commit 5c03d3b into main Nov 10, 2026
152 of 158 checks passed

VSadov deleted the copilot/implement-gc-stress-for-r2r-code branch November 10, 2026 19:11

dotnet-maestro Bot mentioned this pull request Nov 11, 2026

[main] Source code updates from dotnet/runtime dotnet/dotnet#10904

Merged

dotnet-milestone-bot Bot added this to the 12.0-preview1 milestone Nov 11, 2026

Conversation

Copilot AI commented Oct 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

azure-pipelines Bot commented Oct 14, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

dotnet-policy-service Bot commented Oct 14, 2026

Uh oh!

Uh oh!

VSadov commented Oct 16, 2026

Uh oh!

azure-pipelines Bot commented Oct 16, 2026

Uh oh!

VSadov commented Oct 17, 2026

Uh oh!

VSadov commented Nov 4, 2026

Uh oh!

Copilot AI commented Nov 4, 2026

Uh oh!

This comment was marked as outdated.

Uh oh!

VSadov commented Nov 5, 2026

Uh oh!

VSadov commented Nov 5, 2026

Uh oh!

azure-pipelines Bot commented Nov 5, 2026

Uh oh!

azure-pipelines Bot commented Nov 5, 2026

Uh oh!

VSadov commented Nov 5, 2026

Uh oh!

This comment was marked as resolved.

Uh oh!

This comment was marked as resolved.

Uh oh!

jkotas left a comment

Choose a reason for hiding this comment

Uh oh!

VSadov commented Nov 10, 2026

Uh oh!

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Copilot AI commented Oct 14, 2026 •

edited

Loading

azure-pipelines Bot commented Oct 14, 2026 •

edited

Loading