Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Test failure:Regressions\\coreclr\\GitHub_34094\\Test34094\\Test34094.cmd #57458

Closed
VincentBu opened this issue Aug 16, 2021 · 3 comments · Fixed by #57603
Closed

Test failure:Regressions\\coreclr\\GitHub_34094\\Test34094\\Test34094.cmd #57458

VincentBu opened this issue Aug 16, 2021 · 3 comments · Fixed by #57603
Assignees
Labels
arch-x64 arch-x86 area-CodeGen-coreclr CLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMI disabled-test The test is disabled in source code against the issue JitStress CLR JIT issues involving JIT internal stress modes needs-further-triage Issue has been initially triaged, but needs deeper consideration or reconsideration os-linux Linux OS (any supported distro) os-windows
Milestone

Comments

@VincentBu
Copy link
Contributor

Run: runtime-coreclr jitstress 20210814.1

Failed test:

CoreCLR windows x86 Checked jitstress2 @ Windows.10.Amd64.Open
- Regressions\\coreclr\\GitHub_34094\\Test34094\\Test34094.cmd

CoreCLR OSX x64 Checked jitstress2 @ OSX.1013.Amd64.Open
- Regressions/coreclr/GitHub_34094/Test34094/Test34094.sh

CoreCLR Linux x64 Checked jitstress2 @ Ubuntu.1804.Amd64.Open
- Regressions/coreclr/GitHub_34094/Test34094/Test34094.sh

CoreCLR OSX x64 Checked jitstress1 @ OSX.1013.Amd64.Open
- Regressions/coreclr/GitHub_34094/Test34094/Test34094.sh

One of error message:

Assert failure(PID 5988 [0x00001764], Thread: 3976 [0x0f88]): Assertion failed '(emitChkAlign == false) || (ins == INS_lea) || ((emitComp->compCodeOpt() == Compiler::SMALL_CODE) && (((size_t)addr & 3) == 0)) || (((size_t)addr & (byteSize - 1)) == 0)' in 'Test34094:TestAvxCompareGreaterThanOrEqualDouble():bool' during 'Emit code' (IL size 270)

File: D:\workspace\_work\1\s\src\coreclr\jit\emitxarch.cpp Line: 11534
Image: C:\h\w\9FF408CA\p\corerun.exe
Return code:      1
Raw output file:      C:\h\w\9FF408CA\w\A79708F5\uploads\Reports\Regressions.coreclr\GitHub_34094\Test34094\Test34094.output.txt
Raw output:
BEGIN EXECUTION
"C:\h\w\9FF408CA\p\corerun.exe" -p "System.Reflection.Metadata.MetadataUpdater.IsSupported=false" Test34094.dll
Expected: 100
Actual: -1073740286
END EXECUTION - FAILED
FAILED
Test Harness Exitcode is : 1
To run the test:

set CORE_ROOT=C:\h\w\9FF408CA\p
C:\h\w\9FF408CA\w\A79708F5\e\Regressions\coreclr\GitHub_34094\Test34094\Test34094.cmd
Expected: True
Actual:   False


Stack trace
   at Regressions_coreclr._GitHub_34094_Test34094_Test34094_._GitHub_34094_Test34094_Test34094_cmd()
@VincentBu VincentBu added arch-x86 os-linux Linux OS (any supported distro) os-windows JitStress CLR JIT issues involving JIT internal stress modes arch-x64 labels Aug 16, 2021
@dotnet-issue-labeler dotnet-issue-labeler bot added area-CodeGen-coreclr CLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMI untriaged New issue has not been triaged by the area owner labels Aug 16, 2021
@ghost
Copy link

ghost commented Aug 16, 2021

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

Issue Details

Run: runtime-coreclr jitstress 20210814.1

Failed test:

CoreCLR windows x86 Checked jitstress2 @ Windows.10.Amd64.Open
- Regressions\\coreclr\\GitHub_34094\\Test34094\\Test34094.cmd

CoreCLR OSX x64 Checked jitstress2 @ OSX.1013.Amd64.Open
- Regressions/coreclr/GitHub_34094/Test34094/Test34094.sh

CoreCLR Linux x64 Checked jitstress2 @ Ubuntu.1804.Amd64.Open
- Regressions/coreclr/GitHub_34094/Test34094/Test34094.sh

CoreCLR OSX x64 Checked jitstress1 @ OSX.1013.Amd64.Open
- Regressions/coreclr/GitHub_34094/Test34094/Test34094.sh

One of error message:

Assert failure(PID 5988 [0x00001764], Thread: 3976 [0x0f88]): Assertion failed '(emitChkAlign == false) || (ins == INS_lea) || ((emitComp->compCodeOpt() == Compiler::SMALL_CODE) && (((size_t)addr & 3) == 0)) || (((size_t)addr & (byteSize - 1)) == 0)' in 'Test34094:TestAvxCompareGreaterThanOrEqualDouble():bool' during 'Emit code' (IL size 270)

File: D:\workspace\_work\1\s\src\coreclr\jit\emitxarch.cpp Line: 11534
Image: C:\h\w\9FF408CA\p\corerun.exe
Return code:      1
Raw output file:      C:\h\w\9FF408CA\w\A79708F5\uploads\Reports\Regressions.coreclr\GitHub_34094\Test34094\Test34094.output.txt
Raw output:
BEGIN EXECUTION
"C:\h\w\9FF408CA\p\corerun.exe" -p "System.Reflection.Metadata.MetadataUpdater.IsSupported=false" Test34094.dll
Expected: 100
Actual: -1073740286
END EXECUTION - FAILED
FAILED
Test Harness Exitcode is : 1
To run the test:

set CORE_ROOT=C:\h\w\9FF408CA\p
C:\h\w\9FF408CA\w\A79708F5\e\Regressions\coreclr\GitHub_34094\Test34094\Test34094.cmd
Expected: True
Actual:   False


Stack trace
   at Regressions_coreclr._GitHub_34094_Test34094_Test34094_._GitHub_34094_Test34094_Test34094_cmd()
Author: VincentBu
Assignees: -
Labels:

arch-x86, os-linux, os-windows, JitStress, arch-x64, area-CodeGen-coreclr, untriaged

Milestone: -

@JulieLeeMSFT JulieLeeMSFT added this to the 6.0.0 milestone Aug 16, 2021
@JulieLeeMSFT JulieLeeMSFT added needs-further-triage Issue has been initially triaged, but needs deeper consideration or reconsideration and removed untriaged New issue has not been triaged by the area owner labels Aug 16, 2021
@JulieLeeMSFT
Copy link
Member

CC @dotnet/jit-contrib

@kunalspathak
Copy link
Member

kunalspathak commented Aug 17, 2021

I spent some time investigating. It seems that we do not properly propagate the sizeAttr of some HW instrinsic instructions because of which we think that we want a specific alignment which is different than the actual alignment. E.g. , in the test for INS_vinsertf128 which operates on 16 bytes data but we initialize id->_idOpSize == 32 and never adjust it like

switch (ins)
{
case INS_vextractf128:
case INS_vextracti128:
case INS_vinsertf128:
case INS_vinserti128:
{
sizeAttr = EA_16BYTE;
break;
}
case INS_pextrb:
case INS_pinsrb:
{
sizeAttr = EA_1BYTE;
break;
}
case INS_pextrw:
case INS_pextrw_sse41:
case INS_pinsrw:
{
sizeAttr = EA_2BYTE;
break;
}
case INS_extractps:
case INS_insertps:
case INS_pextrd:
case INS_pinsrd:
{
sizeAttr = EA_4BYTE;
break;
}
case INS_pextrq:
case INS_pinsrq:
{
sizeAttr = EA_8BYTE;
break;
}
default:
{
break;
}
}

It is surprising that we didn't change anything recently in this area and we are seeing this issue now although it might be always there. The fix might need inspecting the code paths that are impacted by this (it may be that only DEBUG / asserts are impacted) and then coming up with some common method that assigns right attrSize for these instructions.

I would like to mark this as .NET 7 (since the issue is there for a while) and continue investigating and submit a fix.

Need to reenable the test that was disabled in #57473

@kunalspathak kunalspathak modified the milestones: 6.0.0, 7.0.0 Aug 17, 2021
@ghost ghost added the in-pr There is an active PR which will close this issue when it is merged label Aug 17, 2021
@echesakov echesakov added the disabled-test The test is disabled in source code against the issue label Aug 21, 2021
@ghost ghost closed this as completed in #57603 Aug 30, 2021
@ghost ghost removed the in-pr There is an active PR which will close this issue when it is merged label Aug 30, 2021
@ghost ghost locked as resolved and limited conversation to collaborators Sep 30, 2021
This issue was closed.
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
arch-x64 arch-x86 area-CodeGen-coreclr CLR JIT compiler in src/coreclr/src/jit and related components such as SuperPMI disabled-test The test is disabled in source code against the issue JitStress CLR JIT issues involving JIT internal stress modes needs-further-triage Issue has been initially triaged, but needs deeper consideration or reconsideration os-linux Linux OS (any supported distro) os-windows
Projects
None yet
Development

Successfully merging a pull request may close this issue.

4 participants