NET 7: Calling `GC.Collect()` causes Access Violation exception (cx0000005)

[nilekirk]

Description

Calling GC.Collect() when having sufficiently large objects in LOH and using the Server Garbage Collection causes Access Violation exception (cx0000005) being thrown from coreclr.dll

CoreCLR Version: 7.0.22.51805
.NET Version: 7.0.0
Description: The process was terminated due to an unhandled exception.
Exception Info: exception code c0000005

We have seen this on Windows 10 Enterprise and on Windows Server 2016

Reproduction Steps

Code

open System

[<Struct>]
type T = {
    A: DateTime
    B: int
    C: float
    D: byte }

let data: T array = Array.zeroCreate 2_000_000_000

GC.Collect()

printfn "All OK"

Project file

<Project Sdk="Microsoft.NET.Sdk">

  <PropertyGroup>
    <OutputType>Exe</OutputType>
    <TargetFramework>net7.0</TargetFramework>
    <ServerGarbageCollection>true</ServerGarbageCollection>
  </PropertyGroup>

  <ItemGroup>
    <Compile Include="Program.fs" />
  </ItemGroup>

</Project>

Run this with dotnet run -c release

Changes that leads to program running OK

• Allocating with size 1_000_000_000
• Changing the order of B and C in the struct

Expected behavior

The program writes All OK to the console.

Actual behavior

The program crashes and we see in the Event log that The process was terminated due to an unhandled exception with exception code c0000005

Regression?

We never saw this on NET 6, but it appeared immediately efter we upgraded to NET 7.

Known Workarounds

No response

Configuration

.NET version

CoreCLR Version: 7.0.22.51805
.NET Version: 7.0.0

OS

Edition Windows 10 Enterprise
Version 21H2
Installed on ‎03/‎10/‎2020
OS build 19044.2251
Experience Windows Feature Experience Pack 120.2212.4180.0

Architecture

x64

Other information

No response

I couldn't figure out the best area label to add to this issue. If you have write-permissions please help me learn by adding exactly one area label.

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

Issue Details

---

Description

Calling GC.Collect() when having sufficiently large objects in LOH and using the Server Garbage Collection causes Access Violation exception (cx0000005) being thrown from coreclr.dll

CoreCLR Version: 7.0.22.51805
.NET Version: 7.0.0
Description: The process was terminated due to an unhandled exception.
Exception Info: exception code c0000005

We have seen this on Windows 10 Enterprise and on Windows Server 2016

Reproduction Steps

Code

open System

[<Struct>]
type T = {
    A: DateTime
    B: int
    C: float
    D: byte }

let data: T array = Array.zeroCreate 2_000_000_000

GC.Collect()

printfn "All OK"

Project file

<Project Sdk="Microsoft.NET.Sdk">

  <PropertyGroup>
    <OutputType>Exe</OutputType>
    <TargetFramework>net7.0</TargetFramework>
    <ServerGarbageCollection>true</ServerGarbageCollection>
  </PropertyGroup>

  <ItemGroup>
    <Compile Include="Program.fs" />
  </ItemGroup>

</Project>

Run this with dotnet run -c release

Changes that leads to program running OK

• Allocating with size 1_000_000_000
• Changing the order of B and C in the struct

Expected behavior

The program writes All OK to the console.

Actual behavior

The program crashes and we see in the Event log that The process was terminated due to an unhandled exception with exception code c0000005

Regression?

We never saw this on NET 6, but it appeared immediately efter we upgraded to NET 7.

Known Workarounds

No response

Configuration

.NET version

CoreCLR Version: 7.0.22.51805
.NET Version: 7.0.0

OS

Edition Windows 10 Enterprise
Version 21H2
Installed on ‎03/‎10/‎2020
OS build 19044.2251
Experience Windows Feature Experience Pack 120.2212.4180.0

Architecture

x64

Other information

No response

Author:	nilekirk
Assignees:	-
Labels:	area-GC-coreclr, untriaged
Milestone:	-

[deeprobin]

Cannot reproduce with Workstation-GC.

[mrsharm]

Hi @nilekirk,

I am failing to repro this issue on my machine. I am getting Out of Memory Errors vs. an Access Violation when I set the heap hard limit to control the amount of memory allowed in an effort to replicate the conditions. Could you please provide me the amount of memory the machine you are using has? Also, could you provide a memory dump for this issue, as well?

Thanks in advance!

[samme78]

I can reproduce this, I get the same application error.

Faulting application name: ConsoleApp1.exe, version: 1.0.0.0, time stamp: 0x634eda42
Faulting module name: coreclr.dll, version: 7.0.22.51805, time stamp: 0x634ed04e
Exception code: 0xc0000005
Fault offset: 0x000000000013b93e
Faulting process id: 0x5698
Faulting application start time: 0x01d90ed91f3bd398
Faulting application path: C:\gitprojects\test\ConsoleApp1\ConsoleApp1\bin\release\net7.0\ConsoleApp1.exe
Faulting module path: C:\Program Files\dotnet\shared\Microsoft.NETCore.App\7.0.0\coreclr.dll
Report Id: ab6bce9e-0683-4f90-be8b-8aa580ae9dd3
Faulting package full name: 
Faulting package-relative application ID: 

OS
Edition Windows 10 Pro
Version 21H2
Installed on ‎2021-‎09-‎22
OS build 19044.2251
Experience Windows Feature Experience Pack 120.2212.4180.0

Computer
Processor 11th Gen Intel(R) Core(TM) i9-11950H @ 2.60GHz 2.61 GHz
Installed RAM 64,0 GB (63,2 GB usable)
System type 64-bit operating system, x64-based processor

Resolution
If I change the size to 1_000_000_000 or change the order of B and C in the struct the program works as expected and prints the "All OK".

[mrsharm]

Thanks, @samme78! We are able to consistently repro this issue and are working on diagnosing the cause to work on a fix. We appreciate you reporting this issue!