Age | Commit message (Collapse) | Author | Files | Lines |
|
* Implement instantiating and unboxing through portable stublinker code
- Handle only the cases with register to register moves
- Shares abi processing logic with delegate shuffle thunk creation
- Architecture specific logic is relatively simple
- Do not permit use of HELPERREG in computed instantiating stubs
- Fix GetArgLoc such that it works on all architectures and OS combinations
Add a JIT stress test case for testing all of the various combinations
- Use the same calling convention test architecture that was used as part of tail call work
Rename secure delegates to wrapper delegates
- Secure delegates are no longer a feature of the runtime
- But the wrapper delegate lives on as a workaround for a weird detail of the ARM32 abi
|
|
* [release/3.1] Port fix for JIT silent bad code
Release/3.1 port of https://github.com/dotnet/runtime/pull/797.
Fixes https://github.com/dotnet/runtime/issues/764
The jit might incorrectly order a read from a struct field with an operation
that modifies the field, so that the read returns the wrong value.
Silent bad code; program behaves incorrectly.
Yes, introduced in the 3.0 cycle.
Verified the user's test case now passes; no diffs seen in any existing framework
or test code.
**Low**: the jit is now spilling the eval stack entries to temps in cases where it
did not before; this should be conservatively safe.
cc: @brucefo
____
If we're appending an assignment whose LHS is is a location within a local
struct, we need to spill all references to that struct from the eval stack.
Update the existing logic for this to handle the case where the LHS is a field
of a local struct, and the field is updated by unusual means (here, `initobj`).
Fixes dotnet/runtime#764.
* Fix test
|
|
If a call site fails tail call stress validation, don't consider it for
implicit tail calling either.
In normal jitting we defer this level of validation until impImportCall
to avoid duplicating work.
This avoids an assert when we have invalid IL.
Fixes #25027.
|
|
* Marking Vector128<T>.Count and Vector256<T>.Count as [Intrinsic]
* Fixing NI_Vector128_Count and NI_Vector256_Count to use clsHnd when getting the simdSize and baseType
* Applying the formatting patch.
* Changing some comments to just be "vector element count".
* Fixing impBaseIntrinsic to set the baseType so Vector128_Count and Vector256_Count don't return nullptr
|
|
Methods like System.Runtime.CompilerServices.Unsafe.As<TFrom, TTo> may
have struct reinterpretation when function signature specifies Struct1&
and the method returns Struct2& where Struct1 and Struct2 are different
structs. This may confuse jit optimizations (in particular, value
numbering) because fields of a struct of type Struct1 may be accessed
using handles of Struct2. This fix marks the source local involved in
such struct reinterpretation as having overlapping fields. That prevents
SSA builder from inserting the local into SSA.
Fixes #24159.
No diffs in framework assemblies and coreclr benchmarks.
|
|
hwintrinsic method. (#24327)
|
|
Fixes for tracking struct field sequences
|
|
|
|
- A GT_LCL_VAR may have a zeroOffset field
- Add an assert to prevent building field sequences with duplicates
- Fix fgMorphField when we have a zero offset field
Improve fgAddFieldSeqForZeroOffset
- Add JItDump info
- Handle GT_LCL_FLD
Changing the sign of an int constant also remove any field sequence information.
Added method header comment for fgAddFieldSeqForZeroOffset
Changed when we call fgAddFieldSeqForZeroOffset to be before the call to fgMorphSmpOp.
Prefer calling fgAddFieldSeqForZeroOffset() to GetZeroOffsetFieldMap()->Set()
|
|
We may see pin modifiers on locals that are not gc types.
Ignore these modifiers.
Closes #23950.
Also added a missing copyright header on an unrelated test.
|
|
Before implicit byrefs were tracked by setting lvIsParam and lvIsTemp.
This change explicitly adds a flag for implicitByRef instead of overloading.
In addition, it fixes the decision to copy an implicitByRef for arm64 varargs.
Temporarily bump weight on byref params to match old behavior and avoid codegen
diffs.
Re-enabled various tests and parts of tests.
Closes #20046
Closes #19860
|
|
* Support for Arm64 Vector ABI
Extend HFA support to support vectors as well as floating point types.
This requires that the JIT recognize vector types even during crossgen,
so that the ABI is supported consistently.
Also, fix and re-enable the disabled Arm64 Simd tests.
Fix #16022
|
|
(#23751)
* Moving some JITDUMP calls into Compiler::lookupNamedIntrinsic
* Marking the IsSupported methods as Intrinsic for all HWIntrinsic ISAs
* Updating the hwintrinsic importation to more generally handle IsSupported and ThrowPNSE
* Applying formatting patch.
* Adding using System.Runtime.CompilerServices to the various x86 PlatformNotSupported HWIntrinsic files
|
|
The assert introduced in #23570 was overly aggressive, and didn't account for the fact that pointer arithmetic can exist in the IL, not just introduced by morph.
Fix #23693
|
|
|
|
* Fix spill check for struct lclVars
With the 1st class struct changes, the `SPILL_APPEND` check for the case of an assignment to a lclVar needs to handle block ops as well as lclVar lhs.
Fix #23545
|
|
|
|
* Extract `impAppendStmt` and `impExtractLastStmt`.
* Delete `BEG_STMTS` fake stmt.
Use new functions to keep the list updated.
* Retype `impTreeList` and `impTreeLast` as statements.
Rename `impTreeList` and `impTreeLast` to show that they are statements.
* Fix fields that have to be stmt.
* Start using GenTreeStmt.
Change `optVNAssertionPropCurStmt` to use GenTreeStmt.
Replace `GenTree* stmt = block->bbTreeList` with `GenTreeStmt* stmt = block->firstStmt()`.
Save results of `FirstNonPhiDef` as `GenTreeStmt`.
* Replace do-while with for loop.
* Change type inside VNAssertionPropVisitorInfo.
* Delete unused args fron `optVNConstantPropOnTree`.
* Update fields to be stmt.
Update optVNConstantPropCurStmt to use Stmt.
Change `lvDefStmt` to stmt.
Update LoopCloning structs.
Update `optDebugLogLoopCloning`.
Make `compCurStmt` a statement.
Update declaration name in `BuildNode`.
* Clean simple cpp files.
Clean valuenum.
Clean ssabuilder.
Clean simd.
Clean optcse.
Clean loopcloning.
Clean copyprop.
Clean optimizer part1.
* Start cleaning importer, morph, flowgraph, gentree.
* Continue clean functons.
Clean assertionprop.
Clean morph.
Clean gentree.
Clean flowgraph.
Clean compiler.
Clean rangecheck.
Clean indirectcalltransofrmer.
Clean others.
* Create some temp stmt.
* Delete unnecessary noway_assert and casts.
* Init `impStmtList` and `impLastStmt` in release.
* Response review 1.
|
|
(#23028)
* Simplify and unify Vector64/128/256 platform-agnostic intrinsic handling
* Removed unsupported ISAs
|
|
* [WIP] Struct & SIMD improvements
- Enable CSE of struct values when handle is available (and add code to get the handle of HW SIMD types)
- Don't require block nodes for SIMD assignments
- Don't set `GTF_GLOB_REF` on `GT_OBJ` if it is local
- Set `lvRegStruct` on promoted SIMD fields
- Add tests for #19910 (fixed with this PR) and #3539 & #19438 (fixed with #21314)
- Additional cleanup
Fix #19910
|
|
* fix implicit constructor call
* extern c
format patch
* muti-line
* Remove direct constructor call
* Conversion
* Need paranthesis
* Return value on resize
* declspec(Thread)
* Ignore warnings for GCC
* Formatting issues
* Move cast to constant
|
|
.Net -> .NET
|
|
In #22791 I was creating struct assignments with COMMAs on the rhs, but that isn't handled downstream.
Fix #23059
|
|
`fgArgInfo::ArgsComplete()` checks for additional conditions requiring temps that are not checked for in the body of `fgMorphArgs()`. However, if there are no register args, it won't be called.
Fix #19256
|
|
|
|
implementations (#22705)
* Removing various S.R.I.Vector instance method APIs, since they should now be exposed as extension methods
* Updating the JIT to recognize the S.R.I.Vector extension methods.
* Updating various S.R.I.Vector test templates
* Regenerating the S.R.I tests that are created from a template.
* Fixing the numArgs for Base_Vector256_GetLower
* Fixing the handling for `Base_VectorXXX_As` to normalize the struct type.
* Adding the Base_Vector128_As intrinsics back for arm64
|
|
|
|
|
|
|
|
IsVarAddr was checking GTF_ADDR_ONSTACK to determine if
the GT_ADDR node is an address of a local. This change removes both
GTF_ADDR_ONSTACK and IsVarAddr and uses IsLocalAdrExpr instead.
IsLocalAddrExpr uses opcodes to determine if GT_ADDR node is
a local address.
GTF_ADDR_ONSTACK flag is ancient, added before 2002 so I couldn't find
the checkin that introduced it.
I changed the assert to a check and an assignment since simplifications
inside fgMorphArgs between
https://github.com/dotnet/coreclr/blob/1a1e4c4d5a8030cb8d82a2e5b06c2ab357b92534/src/jit/morph.cpp#L3709
(which causes https://github.com/dotnet/coreclr/blob/1a1e4c4d5a8030cb8d82a2e5b06c2ab357b92534/src/jit/morph.cpp#L3057)
and
https://github.com/dotnet/coreclr/blob/1a1e4c4d5a8030cb8d82a2e5b06c2ab357b92534/src/jit/morph.cpp#L3790
may result in more GT_ADDR nodes recognized by IsLocalAdrExpr.
x86 and x64 pmi frameworks had no code diffs and some gcinfo reductions
(15 methods with gcinfo diffs in x86).
Fixes #22190.
|
|
Atomic ops like GT_CMPXCHG and some HW intrinsic nodes act like assignements so impAppendStmt has to account for them. They can be top level nodes or they can appear in the RHS of a GT_ASG node that perhaps is not considered to have an assignment side effect itself.
|
|
Change test priority to 0
|
|
|
|
|
|
and a tail call optimization
Explicit tail calls are now checked for and blocked from performing an unboxing operation in impDevirtualizeCall
If late devirtualization calls impDevirtualizeCall with an IMPLICIT_TAILCALL we will clear this flag if we
decide to perform the unboxing operation.
|
|
Some IR cleanup
|
|
* Copy address-taken SIMD intrinsic
This occurs, for example, when boxing the result of a SIMD intrinsic. This was being handled for the HW intrinsic case, but not the SIMD Vector intrinsics. Also, eliminate `OperIsSimdHWIntrisic` since it redundantly checks for the case of a SIMD result, even though it was always called in a context where the result is known to be a struct type.
Fix #21854
|
|
Add methods that answer the general question of whether or not
the jit is optimizing the code it produces.
Use this to replace composite checks for minopts and debug
codegen (the two modes where the jit is not optimizing).
|
|
|
|
attribute is respected by the runtime (#21601)
* Updating the SIMD intrinsics to check for the `[Intrinsic]` attribute
* Ensure that the existing Vector/Vector<T> intrinsic methods are marked as such.
* Fixing Vector<T>.DotProduct and Vector.AndNot to be treated as intrinsic
* Fixing `SIMDIntrinsicBitwiseAndNot` to swap the operands for XARCH
|
|
intrinsics when available. (#21602)
* Removing the explicit value assignment to NamedIntrinsic enum values
* Renaming NI_Math_Round/NI_MathF_Round to NI_System_Math_Round/NI_System_MathF_Round
* Updating `FusedMultiplyAdd` for `System.Math` and `System.MathF` to use intrinsics when available
|
|
This change modified the importer to create GenTreeAllocObj node for
box and newobj instead of a helper call in R2R mode. ObjectAllocator phase
decides whether the object can be allocated on the stack or has to be created
on the heap via a helper call.
To trigger object stack allocation COMPlus_JitObjectStackAllocation has
to be set (it's not set by default).
|
|
Don't require BLK nodes for SIMD
|
|
Optimize away the unbox type test when the jit knows the type of object
being unboxed and can resolve the type comparison at jit time.
Closes #14473.
|
|
Eliminate most cases where an OBJ or BLK node is required for SIMD
values. The exception is the case where a value produced by an intrinsic
(SIMD or HWIntrinsic) is used as an argument but the argument is of a
different SIMD type (e.g. a different baseType),
|
|
intrinsics (#21432)
* Updating Vector128.CreateScalar and Vector256.CreateScalar to be implemented using other intrinsics
* Updating Vector128.Equals and Vector256.Equals to be implemented using other intrinsics
* Updating Vector256.WithLower, Vector256.GetUpper, and Vector256.WithUpper to be implemented using other intrinsics
* Updating Vector128.Create(T) and Vector256.Create(T) to be implemented using other intrinsics
* Fixing the `NI_Base_Vector256_As` intrinsics to only fold the cast if AVX is supported and add a clarifying comment to the Vector128/256.Equals code
* Changing the various `*Software()` local functions in Vector128/256 to be `SoftwareFallback()`
|
|
Vector256.GetLower to be intrinsics (#21351)
* Moving CreateScalarUnsafe, ToScalar, Vector128.ToVector256Unsafe, and Vector256.GetLower to be intrinsics
* Adding containment support to the helper intrinsics
|
|
Lay the groundwork for guarded devirtualization of virtual and interface
calls in the jit.
Introduce the notion of a guarded devirtualization candidate and identify
these if regular devirtualization fails. Use simple heuristics to produce
a class to guess for. Require that the method that would be invoked if the class
guess is correct be a plausible inline candidate.
Generalize the calli transformer to become an indirect call transformer.
This runs after importation because it needs to introduce control flow and
runs before inlining so that the new direct calls it introduces can be inlined.
Implement the transformation to duplicate the call site, devirtualize on the side
where the class is now known exactly, and turn the resulting direct call into an
inline candidate.
Add a motivation and design document.
|
|
|
|
featureSIMD is disabled (#21274)
|