Skip to content

[clr-ios] Fix Vector<T> width and R2R AVX baseline mismatch on x64 iOS/tvOS simulators #129012

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.

Sign up for GitHub

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

Open
kotlarmilos wants to merge 9 commits into
base: main
Choose a base branch
from
+24 −10
Open

[clr-ios] Fix Vector<T> width and R2R AVX baseline mismatch on x64 iOS/tvOS simulators #129012

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
bd8f26b
[clr-ios] Fix interpreter NullReferenceException with Vector<T> on no…
kotlarmilos Jun 4, 2026
33fb93c
Keep existing Vector<T> cap comments
kotlarmilos Jun 4, 2026
f75f6b7
Use 128-bit Vector<T> baseline for crossgen2 on Apple x64 targets
kotlarmilos Jun 5, 2026
f3ad776
Fail fast when stripped-IL R2R is disabled on a no-JIT runtime
kotlarmilos Jun 5, 2026
83b41ef
Drop iOS/tvOS device targets from x64 baseline exclusion
kotlarmilos Jun 5, 2026
6d35b32
Read stripped-IL flag from the component header
kotlarmilos Jun 5, 2026
0cb7e20
Fix stale Vector<T> comment to match the actual condition
kotlarmilos Jun 5, 2026
3912fb3
Remove redundant interpreter cap on preferred vector width
kotlarmilos Jun 5, 2026
41dd8e6
Avoid reading InterpMode config when overridden on no-JIT runtimes
kotlarmilos Jun 5, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 4 additions & 1 deletion src/coreclr/tools/Common/InstructionSetHelpers.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,7 +31,10 @@ public static InstructionSetSupport ConfigureInstructionSetSupport(string instru

if ((targetArchitecture == TargetArchitecture.X86) || (targetArchitecture == TargetArchitecture.X64))
{
if (isReadyToRun && targetOS != TargetOS.OSX && targetOS != TargetOS.MacCatalyst)
bool isAppleOS = targetOS is TargetOS.OSX or TargetOS.MacCatalyst
Comment thread
BrzVlad marked this conversation as resolved.
or TargetOS.iOSSimulator or TargetOS.tvOSSimulator;

if (isReadyToRun && !isAppleOS)
{
// ReadyToRun can presume AVX2, BMI1, BMI2, F16C, FMA, LZCNT, and MOVBE
instructionSetSupportBuilder.AddSupportedInstructionSet("x86-64-v3");
Expand Down
8 changes: 8 additions & 0 deletions src/coreclr/vm/ceeload.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3583,6 +3583,14 @@ void Module::RunEagerFixupsUnlocked()
{
_ASSERTE(IsReadyToRun());
GetReadyToRunInfo()->DisableAllR2RCode();

#ifndef FEATURE_DYNAMIC_CODE_COMPILED
if (GetReadyToRunInfo()->IsStrippedILBodies())
{
EEPOLICY_HANDLE_FATAL_ERROR_WITH_MESSAGE(COR_E_EXECUTIONENGINE,
W("ReadyToRun code was disabled by a failed eager fixup, but the image has stripped IL bodies and this runtime has no JIT fallback."));
}
#endif // !FEATURE_DYNAMIC_CODE_COMPILED
}
else
{
Expand Down
15 changes: 6 additions & 9 deletions src/coreclr/vm/codeman.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1331,7 +1331,12 @@ void EEJitManager::SetCpuInfo()
uint32_t maxVectorTBitWidth = (CLRConfig::GetConfigValue(CLRConfig::EXTERNAL_MaxVectorTBitWidth) / 128) * 128;

#if defined(FEATURE_INTERPRETER)
if (maxVectorTBitWidth != 128 && CLRConfig::GetConfigValue(CLRConfig::EXTERNAL_InterpMode) == 3)
#if defined(FEATURE_DYNAMIC_CODE_COMPILED)
bool interpreterOnly = CLRConfig::GetConfigValue(CLRConfig::EXTERNAL_InterpMode) == 3;
#else
bool interpreterOnly = true;
#endif
if (maxVectorTBitWidth != 128 && interpreterOnly)
{
// Disable larger Vector<T> sizes when interpreter is enabled
maxVectorTBitWidth = 128;
Expand Down Expand Up @@ -1701,14 +1706,6 @@ void EEJitManager::SetCpuInfo()

uint32_t preferredVectorBitWidth = (CLRConfig::GetConfigValue(CLRConfig::EXTERNAL_PreferredVectorBitWidth) / 128) * 128;

#ifdef FEATURE_INTERPRETER
if (CLRConfig::GetConfigValue(CLRConfig::EXTERNAL_InterpMode) == 3)
{
// Disable larger Vector<T> sizes when interpreter is enabled, and not compiling S.P.Corelib
preferredVectorBitWidth = 128;
}
#endif

if ((preferredVectorBitWidth == 0) && throttleVector512)
{
preferredVectorBitWidth = 256;
Expand Down
6 changes: 6 additions & 0 deletions src/coreclr/vm/readytoruninfo.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -249,6 +249,12 @@ class ReadyToRunInfo
return m_pHeader->CoreHeader.Flags & READYTORUN_FLAG_PARTIAL;
}

BOOL IsStrippedILBodies()
{
LIMITED_METHOD_CONTRACT;
return m_pHeader->CoreHeader.Flags & READYTORUN_FLAG_STRIPPED_IL_BODIES;
}
Comment thread
kotlarmilos marked this conversation as resolved.

void DisableAllR2RCode()
{
LIMITED_METHOD_CONTRACT;
Expand Down
Loading