ラベル: area-CodeGen-coreclr

dotnet/runtimeにマージされたPull RequestをAIで日本語要約

ラベル一覧へ戻る

ラベル: area-CodeGen-coreclr の Pull Request の一覧を表示します。上から新しい順に表示されています。

  1. #136260 JIT: Clear GC poll flags when converting unreachable blocks to throws
  2. #136297 [RISC-V] Fix 64-bit constant truncation on LLP64 hosts
  3. #136203 Fix virtual vtable call lowering without fixed registers
  4. #135970 JIT: treat throwing trees as GC safe points
  5. #135638 JIT: make sure the block a catch returns to generates an instruction
  6. #135296 [wasm] [spmi] Collect SPMI contexts from libraries via crossgen2
  7. #136100 JIT: Devirtualize calls before probing them in optimized instrumented code
  8. #130717 JIT: Materialize partially-constant Vector.Create as CNS_VEC + inserts
  9. #135543 [wasm] Regalloc: Allow Non-exposed GC Ref Locals to Live in Wasm Locals
  10. #135853 JIT: spill local refs for COMMA-wrapped local stores in impAppendStmt
  11. #131428 Ensure that x64 disassembly names operands at the size actually encoded
  12. #136120 Correct x86 instruction metadata
  13. #135991 [wasm] Fix single-register struct call result stores (narrow stores, GC write barrier)
  14. #134910 [JIT] Rationalize: Re-run pre-order processing on identity shuffle rewrite in PreOrderVisit
  15. #136149 JIT: Preserve volatile static reads when folding readonly fields
  16. #136021 JIT: keep lea esp, [ebp] in x86 epilogs
  17. #135823 JIT: preserve null checks when folding ThrowIfNull in Tier0
  18. #135893 JIT: Gate single-edge root profile trust on hotness or speed preference
  19. #135837 JIT: ignore finally-entry backedges when fixing continuation weights
  20. #136040 JIT: inline box Nullable<T> in ReadyToRun
  21. #135146 [wasm] [JIT] Contain Array Data Offsets
  22. #136043 JIT: clear BBF_DONT_REMOVE on the new preheader in loop inversion
  23. #136085 JIT: Exclude byref-like structs from async continuation capture
  24. #136083 [release/11.0] Preserve ARM64 flags dependencies during indirection reordering
  25. #135999 fold more redundant nullchecks
  26. #136080 [release/11.0] JIT: Make GenTree::Compare more conservative around several node types
  27. #136082 [release/11.0] Report byref-like struct arguments to the GC in helper-based tail calls
  28. #136079 [release/11.0] JIT: Use local's GC type when unspilling a register candidate
  29. #135768 Absolute/global addressing changes for native AOT/WASM
  30. #135724 JIT: Keep volatile loads when folding full-range relational comparisons
  31. #135850 JIT: Treat stores to EH-live locals as visible when hoisting
  32. #135946 JIT: Don't assert when printing open generic method names in disasm
  33. #135851 JIT: Propagate rejectNegativeConst through lcl + cns in monotonicity check
  34. #135835 JIT: preserve GTF_IND_INITCLASS when inlining BOX of Nullable<T>
  35. #135293 [wasm][JIT] Remove assertions assuming the SP arg of a fast tail call is a GT_PHYSREG
  36. #135833 JIT: sign-extend folded 32-bit RotateLeft/RotateRight constants
  37. #135834 JIT: spill pending side effects before signed Log2 expansion
  38. #135739 JIT: Remove redundant Span.Slice/Vector128.Create checks in vectorized loops
  39. #135577 [wasm] Avoid Spilling Many GC-ref Typed Tree Temps
  40. #135602 JIT: reject method entry inside handlers or filters
  41. #135827 JIT: check for a conditional branch that falls off the end of the method before linking its false edge
  42. #135793 Emit base-only LEAs as register moves
  43. #135761 Do not emit try table on native AOT
  44. #135723 JIT: Do not propagate non-nullness through isinst helpers
  45. #135722 JIT: Respect ordering side effects on the second operand when swapping
  46. #135725 JIT: Reject native-int bounds checks in range check cloning
  47. #135721 JIT: Model xadd as writing its register operand
  48. #135720 JIT: Do not contain local addresses of volatile indirections on xarch
  49. #135719 JIT: Stop the redundant zero-init scan at blocks with other predecessors
  50. #135745 Re enable vector3Interop for gcstress
  51. #135760 Do not use CORINFO_HELP_INITCLASS on native AOT
  52. #135599 JIT: use the carry flag to detect SP wraparound in genLclHeap on ARM
  53. #135596 JIT: support register-condition SELECT_INC/NEG/INV on arm64
  54. #135575 Preserve ARM64 flags dependencies during indirection reordering
  55. #135600 JIT: use the second field's GC type for the ldp hi register in genPut…
  56. #135686 Emit a zero-offset local address as a mov
  57. #135692 JIT: Preserve Half estimate ISA dependencies in ReadyToRun
  58. #135685 Remove the redundant jump in the HW intrinsic jump table fallback
  59. #131826 [APX] Add Jmpabs support.
  60. #135458 JIT: preserve value evaluation when expanding reference array stores
  61. #135594 JIT: mark pseudo return buffer temp as multi-reg return
  62. #135595 JIT: don't contain BITCAST under small local stores
  63. #135474 Fix fat calls in native AOT/WASM
  64. #134420 [release/11.0] Fix VNNI blend merge operand preservation
  65. #134421 [release/11.0] Fix unsigned vector division correction lane width
  66. #134430 [release/11.0] Preserve contained-load exceptions during embedded masking
  67. #135629 JIT: Simplify ReadUtf8 immutable data extraction
  68. #135382 JIT: Restore insertps source-lane fold for floating-point inserts
  69. #135605 JIT: exclude handles from numeric range reasoning
  70. #135484 JIT: preserve relocatable handles when simplifying ordered comparisons
  71. #135485 JIT: model the signed Log2 fallback as throwing
  72. #135459 JIT: preserve negation of division by the signed minimum
  73. #135486 JIT: copy large local structs through their address when decomposing
  74. #135461 JIT: retain complete metadata of reassociated folded constants
  75. #135457 JIT: require nonnegative bounds for symbolic range limits
  76. #135380 Optimize Wasm comparisons against zero
  77. #135421 JIT: Fold unsigned constant range tests
  78. #135364 JIT: Preserve null checks in SequenceEqual VN folding
  79. #135415 Fix x86 GC holes in xchg/rorx and GC suspension of spinning threads
  80. #135363 JIT: Make GenTree::Compare more conservative around several node types
  81. #134970 JIT: Account for partial SIMD callee saves in hoisting
  82. #135431 JIT: Invalidate DFS annotations before recomputing the DFS tree
  83. #135327 Fix assertion (node->GetRegNum() != REG_NA) during LSRA
  84. #135381 JIT: Use local's GC type when unspilling a register candidate
  85. #135373 JIT: Preserve inlining of profiled memcmp fallback
  86. #135307 JIT: Fold SequenceEqual self-comparisons with constant length
  87. #134710 [release/10.0] JIT: handle redefined enumerators in conditional escape analysis
  88. #135135 Skip badinlinee under GC stress
  89. #135134 Fix JIT over reporting RCPC2
  90. #134549 [wasm] Remove Wasm NYI crossgen fallback
  91. #134709 [release/11.0] JIT: handle redefined enumerators in conditional escape analysis
  92. #135069 [release/11.0] JIT - Fix 'OperIs...' assert during forward substitution
  93. #134969 JIT: Skip unrolling tests stored in locals
  94. #134185 JIT: preserve local access type in relop assertion propagation
  95. #135072 JIT: Remove dead/redundant GTF_* flags
  96. #134962 JIT - Fix 'OperIs...' assert during forward substitution
  97. #134665 [release/11.0] Fix null vnStore dereference in optAssertionProp_HWIntrinsic
  98. #134940 JIT: don't skip covariant store check for shared sealed element types
  99. #134981 [wasm][R2R] Fix null ReturnTypeDesc dereference in SpillStructCallResult
  100. #134898 Fix Enum.Equals folding for shared-generic nested enums
  101. #134971 JIT: Preserve finalizable allocations during hoisting
  102. #135026 JIT: remove unused helpers
  103. #135027 [wasm][R2R] Re-enable b51875 and b77707 after #134979
  104. #134984 [wasm][JIT] Use TYP_SIMD16 register type for 16-byte struct layouts on Wasm
  105. #134668 Extend extend tail dup to get more ccmp
  106. #134973 JIT: mark GC safe point after stelem.ref call transformation
  107. #134927 JIT: Use a visitor for local occurrences through LocalsTreeList
  108. #134974 JIT: fix profile update when folding a switch with shared case targets
  109. #134979 JIT: make Wasm frames that call a finally unwindable
  110. #134880 JIT fix for m_regAlloc->isRegCandidate assert under optrepeat
  111. #134466 JIT: Fix several range check correctness bugs
  112. #132752 Arm64-SVE: Support accessing unknown-sized Tier-0 locals from Tier-1 (OSR)
  113. #134802 JIT: Use local definition visitors in async mutation analysis
  114. #134758 JIT: don't fall through from a Wasm callfinally across a try_table end
  115. #134666 [JIT] [wasm]: Compare TYP_REF references directly with null using i32.eqz
  116. #134703 Fix missing env.vars in JIT tests
  117. #134861 Remove allocation assertion from Runtime_134605 regression test
  118. #134718 JIT: Remove GTF_DEBUG_NODE_MORPHED tracking
  119. #134048 JIT: Defer physical promotion readbacks across basic blocks
  120. #134782 JIT: initialize tracked-local counts before liveness
  121. #134502 [WIP] Fix integer-divide overflow in JIT strength reduction
  122. #134744 JIT: Extend non-null assertion propagation to pointer-sized comparisons
  123. #134680 JIT: Fix normalization check for parameter register fields
  124. #134765 [wasm][R2R] Re-enable open virtual delegate runtime tests
  125. #134708 JIT: expand Object.GetType at Tier0
  126. #134658 JIT: handle redefined enumerators in conditional escape analysis
  127. #134644 JIT: Type implicit byref args as TYP_I_IMPL
  128. #134686 [wasm][RyuJIT] Restore __stack_pointer in reverse P/Invoke epilogs
  129. #134624 JIT: Make multiple try entries in an SCC an implLimitation in fgwasm
  130. #134490 JIT: Improve Wasm PEP calls
  131. #134634 JIT: Preserve SIMD operand order before morph
  132. #133304 Implicit byref arguments are never on the GC heap
  133. #134590 SPMI: Add -f as an alias for -filter in superpmi.py
  134. #134560 JIT: Track AVX state from register-producing widths
  135. #134530 Avoid repeated SIMD mask proof traversal in value numbering
  136. #134387 JIT: Guard comma throw propagation types
  137. #133892 JIT: handle local field stores in object stack allocation
  138. #133776 JIT: Preserve SIMD operand evaluation and ordering
  139. #133642 JIT: Preserve exception order in loop hoisting
  140. #134522 Preserve scalar negation types when folding FMA operands
  141. #134386 JIT: don't speculate ordering-side-effect upper bounds in FoldNeverNegativeRangeTest
  142. #134375 JIT: handle relational operators in range overflow analysis
  143. #133867 JIT: reject variable-count XOR rotation recognition
  144. #133826 Fix Enum.Equals folding for nullable boxed arguments
  145. #133600 JIT: keep the null check when a method table load folds to a constant
  146. #132260 [LoongArch64] Fix some new failures in Release libs.tests after PR#128604.
  147. #134511 JIT: Do not reuse local addresses across suspension points in async methods
  148. #134075 JIT: Use integral ranges to narrow comparisons with memory operands
  149. #132534 Use explicit secret arguments in IL stubs
  150. #134518 JIT: Include auxiliary types in hardware intrinsic value numbering
  151. #134521 Preserve element types when folding SIMD Dot broadcasts
  152. #134520 Preserve vector results when folding AVX-512 mask complements
  153. #134437 JIT: Fix profiled memcmp specialization for inline candidates
  154. #134441 Fix null vnStore dereference in optAssertionProp_HWIntrinsic
  155. #134411 JIT: Simplify and disable overlapping struct copy regression test
  156. #134366 Preserve contained-load exceptions during embedded masking
  157. #134358 [JIT] [Wasm]: Check IsInvariantInRange for LEA base addr in GetFoldableAddrMode
  158. #133416 JIT: Model promoted struct definitions through visitors
  159. #134410 JIT: Fix AVX state tracking for unrolled block operations
  160. #134362 Fix unsigned vector division correction lane width
  161. #134364 Fix VNNI blend merge operand preservation
  162. #133905 [WIP] Fix GC hole in JIT due to if-conversion
  163. #134345 JIT: Bypass bulk write barrier wrapper for small struct copies
  164. #134351 [release/11.0] JIT: Prevent loop hoisting across volatile reads
  165. #134350 [release/11.0] JIT: guard overflow in loop cloning offset limits
  166. #134349 [release/11.0] JIT: Fix promoted liveness in forward substitution
  167. #134348 [release/11.0] JIT: Validate jump threading phi inputs
  168. #134347 [release/11.0] JIT: validate RBO relop rewrite operands
  169. #134344 [release/11.0] Validate store-coalescing constants before modifying IR
  170. #134325 JIT: Set profile flag on async OSR jump blocks
  171. #134076 [wasm][coreclr] Re-enable Regression3_Regressions for interpreter
  172. #133801 JIT: Fix zero-flag reuse for BSF/BSR
  173. #133976 Validate store-coalescing constants before modifying IR
  174. #134184 JIT: Avoid null predecessor dereference after tail duplication
  175. #134279 JIT: Preserve exceptions during jump threading
  176. #133690 JIT: Generalize entry-block parameter register rewriting to all blocks
  177. #134214 [release/11.0] Fix not-greater-than predicate when lowering SIMD selects
  178. #134215 [release/11.0] Fix scalar-typed SIMD operands in Dot and mask inversion lowering
  179. #134213 [release/11.0] Preserve mask element granularity in SIMD transformations
  180. #134288 Re-enable badcodeinsidefinally for Wasm R2R
  181. #134205 JIT: Validate jump threading phi inputs
  182. #134198 JIT: Make LACKS_RETURN callsite-specific
  183. #133803 JIT: Fix block-store address containment boundary
  184. #134160 JIT: Fix value profiling in optimized instrumented tiers
  185. #134203 JIT: Fix vector IsSupported folding for constrained interface calls
  186. #133703 JIT: Fix promoted liveness in forward substitution
  187. #134098 JIT: don't elide the MD-array covariance check for covariant element types
  188. #134100 JIT: use a call GC poll for blocks that end an EH region
  189. #134140 JIT: Ensure the importer does not reorder GTF_ORDER_SIDEEFF with GTF_GLOB_REF
  190. #134151 Fix XARCH importation of isPositive and isNegative
  191. #133802 JIT: Preserve overflow checks when folding scaled indexes
  192. #134097 JIT: stop tail duplication from looping forever around a conditional …
  193. #133827 JIT: Fix null-check ordering in EH regions
  194. #134099 JIT: preserve side effect ordering when unrolling UTF16 comparisons
  195. #132517 JIT: Enable async inlining stress under JitStress
  196. #133834 JIT: guard overflow in loop cloning offset limits
  197. #133994 Fix for antigen build
  198. #133990 JIT: Fix truncated histogram likelihoods
  199. #133215 Fix Vector.Dot on arm64
  200. #133993 JIT: Attribute GDV inline failures to candidate
  201. #132981 Fix for out of range mask pattern assert
  202. #134033 [release/11.0] JIT: Refine promotion costing for cold small parameter fields
  203. #133808 JIT: Invalidate zero-trip guards across iterator or limit redefinitions
  204. #133766 JIT: Properly handle unknown starting values for overflow checks in SCEV analysis
  205. #133603 JIT: Refine promotion costing for cold small parameter fields
  206. #133869 JIT: Preserve checked additions in rotation recognition
  207. #133825 JIT: Prevent loop hoisting across volatile reads
  208. #133866 Fix double monitor release in synchronized runtime-async methods
  209. #133708 Preserve mask element granularity in SIMD transformations
  210. #132221 JIT: contain shifts into byref ADD on arm64
  211. #133597 JIT: fix BITCAST removal widening a small-typed load
  212. #133855 JIT: Fix GC reporting for xarch interlocked reference results
  213. #133735 JIT: Preserve constrained boxing order
  214. #133685 JIT: Respect offsets in non-null assertion propagation
  215. #133740 JIT: Fix MD array store evaluation order
  216. #133805 JIT: Preserve evaluation order in GC bulk copies
  217. #133806 JIT: Preserve receiver side effects in constrained GetType calls
  218. #133686 JIT: Fix hang when a switch targets an empty-block cycle
  219. #133649 JIT: Initialize Wasm resume IP to a rethrowing value
  220. #133749 Preserve profiling for managed intrinsic implementations
  221. #133737 Fix JIT this pointer null check elision
  222. #133790 JIT: Fix loop cloning with nested EH regions
  223. #133727 JIT: Preserve 64-bit bit-count value numbers
  224. #133646 JIT: fix checked subtraction loop unrolling
  225. #133709 Fix scalar-typed SIMD operands in Dot and mask inversion lowering
  226. #133590 [release/11.0] JIT: Dispatch OSR resumptions in tail-await-only methods
  227. #133700 JIT: stop inferring through bitwise complement
  228. #133621 JIT: validate RBO relop rewrite operands
  229. #133510 JIT: Fix wasm ABI register type mismatch
  230. #133589 Fix not-greater-than predicate when lowering SIMD selects
  231. #133539 Fix vector complement-add folding for floating-point types
  232. #133235 JIT: avoid GS cookie conflict with secret stub argument in non-standard register
  233. #133591 Group JitBlue regression tests by runner
  234. #133161 [release/10.0] JIT: Add visit budget checks to IsMonotonicallyIncreasing and ComputeDoesOverflow
  235. #133528 JIT: fix wasm conditional fallthrough
  236. #133542 Fix SIMD complement and scalar FMA folding
  237. #133541 Fix optimizations of explicitly rounded hardware intrinsics
  238. #133511 JIT: Normalize small unmanaged args on wasm
  239. #132423 JIT: Improve modelling of AsyncResumedDef store in value numbering
  240. #133257 [release/10.0] Fix SIMD MinMax constant special cases
  241. #133485 Guard stale VN store in unused GetType extraction
  242. #133531 JIT: expand all stack array allocations in a statement
  243. #133477 [release/11.0] Fix discarding of NRE's when unboxing nullables in Mono and JIT
  244. #133482 [release/11.0] JIT: guard negative-offset bounds inference against underflow
  245. #133481 [release/11.0] JIT: fix range assertions for bypassed phi blocks
  246. #132822 JIT: add JitEnablePhaseChecks, and disable it for the base SuperPMI asm diffs JIT
  247. #133486 Fix missing Helix queue variables in SuperPMI replay pipelines
  248. #133358 JIT: Dispatch OSR resumptions in tail-await-only methods
  249. #133093 JIT: fix TrailingZeroCount constant folding on 32-bit targets
  250. #133434 JIT: preserve recursive tail call flag during inlining
  251. #133324 JIT: Improve cast helper fallbacks
  252. #133138 JIT: Tail-merge fix self-loop handling
  253. #133281 [Wasm RyuJIT] Fix: Pass Call's struct type to Wasm ABI Classifier Instead of Dest Store Type
  254. #133256 [release/11.0] Fix SIMD MinMax constant special cases
  255. #133351 Fix discarding of NRE's when unboxing nullables in Mono and JIT
  256. #133272 JIT: guard negative-offset bounds inference against underflow
  257. #130588 JIT: Elide transitive array bounds checks
  258. #133268 JIT: fix range assertions for bypassed phi blocks
  259. #133228 JIT: Sharpen condition to avoid preserving call definition targets in continuations
  260. #133079 JIT: Expand REFANYTYPE QMARK before physical promotion
  261. #133310 Narrow browser WASM R2R test exclusions
  262. #133227 Use rolling internal Windows queue for helix_windows_x64_latest_internal
  263. #133206 JIT: Synthesize profile data when invalid counts are discarded
  264. #133039 Restore chaos65204782cs as a standalone JIT test
  265. #133173 Fix SIMD MinMax constant special cases
  266. #133210 [wasm] Narrow multi-use operands read out of long locals
  267. #132816 JIT: extend relaxed checks through morph
  268. #133231 Update jit-format workflow to install .NET 10 SDK
  269. #133136 Fix Wasm SCC dispatcher predecessor rewrites
  270. #133168 [release/11.0] Fix SIMD primitive zero initialization
  271. #133170 [release/10.0] Fix SIMD primitive zero initialization
  272. #133134 Run browser wasm runtime tests with ReadyToRun
  273. #132850 JIT: stop EnableExtraSuperPmiQueries from changing what gets compiled
  274. #133095 JIT: Include method name in Release disassembly footer
  275. #133132 Fix Wasm SCC profile weights for #133120
  276. #133100 Fix SIMD primitive zero initialization
  277. #133051 [release/11.0] JIT: update impFixupStructReturnType to handle GT_RET_EXPR
  278. #132986 JIT: update impFixupStructReturnType to handle GT_RET_EXPR
  279. #133016 [release/11.0] Preserve side effects in SIMD classification helpers
  280. #132631 [release/10.0] JIT: fix switch peeling stealing the wrong operand of its compare
  281. #132629 [release/10.0] JIT: don't use a jump-threaded block's sharpened predicate for dominator-based inference
  282. #132628 [release/10.0] JIT: only imply non-null assertions from exact-type assertions
  283. #133014 JIT: Fix Crossgen2 assertion for non-inlineable tail-call GDV
  284. #132375 Don't require GDV targets to be inlineable
  285. #132849 JIT: check operand def chains for overflow in RangeCheck::ComputeDoesOverflow
  286. #132539 JIT/Interp: Remove MAYBENULL TypeHandle -> RuntimeType Helper uses
  287. #132921 Preserve side effects in SIMD classification helpers
  288. #132963 [LoongArch64] Fix the GCReg update for the relocation case (pcalau12i + addi.d/ld.d) of NativeAOT, the GCReg should update after calculation completed.
  289. #132896 Avoid unnecessary JIT data section constants
  290. #132527 Add a SuperPMI collection over the Aspire CLI (NativeAOT)
  291. #130870 Fix LargestVectorSize and buffer size in ARM HW intrinsic tests
  292. #132795 JIT: Fix empty-block cycle hang
  293. #131746 Arm64-SVE: Fixes for returning Vector<T> through the return buffer
  294. #132487 JIT: relaxed early ir checks
  295. #132793 JIT: don't truncate 64-bit constants in IsConstantInt32Assertion
  296. #132236 [Wasm RyuJIT] Mark SP/FP delta computation helpers as unused
  297. #132741 Fix RISC-V JIT opcode format mismatch
  298. #132099 Add TARGET_HAS_MULHI and TARGET_MASKS_SHIFTS flags to jit lowering
  299. #132445 JIT: Avoid narrowing a retyped AND operand in fgOptimizeEqualityComparisonWithConst
  300. #132676 JIT: fix SIMD12 relocatable-address containment check in ContainCheckIndir
  301. #132625 JIT: turn MEASURE_NOWAY off by default
  302. #132612 JIT: check LIR unused values in the post-phase checks
  303. #132615 JIT: remove duplicate fgDebugCheckLinks calls
  304. #132624 JIT: remove duplicate node uniqueness check
  305. #130023 Add compile-time and run-time methods to obtain Vector<T> byte length
  306. #132569 JIT: speed up fgDebugCheckNodesUniqueness and drop dead morphTrees arg
  307. #132395 JIT: fix switch peeling stealing the wrong operand of its compare
  308. #132536 JIT: Mark GT_LCLHEAP with GTF_CALL and GTF_GLOB_REF
  309. #132511 JIT: Simplify containment and NOP checks
  310. #132485 [release/10.0] Fix array size overflow in objectalloc.cpp
  311. #132520 JIT: Always fill uninitialized JIT memory with 0xcd
  312. #132469 JIT: replace fgWalkTree callbacks with GenTreeVisitor implementations
  313. #132513 JIT: stop validating the whole flow graph after every block compaction
  314. #132506 JIT: Merge fgDebugCheckTypes into fgDebugCheckFlags
  315. #132509 SuperPMI: speed up reportMetadata metric lookup
  316. #132523 JIT: Guard named intrinsic lookup
  317. #132437 Fix codegen of NotMask on x64
  318. #132409 JIT: Enable async inlining by default
  319. #132167 Merge JIT Generics tests into explicit runners
  320. #132168 Merge SIMD test sources into grouped runners
  321. #132259 JIT: fill two gaps in wasm codegen
  322. #132326 JIT: Do not skip reference prolog zero inits in async methods
  323. #123920 Add format attribute to printf-style wrappers and fix format string errors
  324. #132313 JIT: Fix VN map select cache inconsistency on budget exhaustion
  325. #132281 JIT: don't use a jump-threaded block's sharpened predicate for dominator-based inference
  326. #131356 Fix edge likelihoods out of optimize bools
  327. #132269 JIT: Strip GC/reloc flags from the size in arm64 instGen_Set_Reg_To_Imm
  328. #131538 JIT: Support general inlining in runtime async
  329. #132146 Enable hardware intrinsics for ExtractMostSignificantBits test
  330. #132223 JIT: Add an inlining stress mode for general runtime async inlining
  331. #131976 [LoongArch64] Fix some crash CoreRoot tests under --gcstresslevel=4 like Methodical_{r2|d2}.
  332. #132246 JIT: Fix build break in instruction table
  333. #132222 JIT: Do not mark an async call no-return because its callee has no ret
  334. #131835 Rename INS_FLGAS in consistent style for xarch emitter.
  335. #132166 Revert "JIT: cache the wasm image base in a local"
  336. #132170 [Wasm RyuJIT]: Remove more unreachable NYI_WASM entries
  337. #132101 Modernize JIT Directed merged runners
  338. #132100 JIT: fold wasm local address frame offsets into store memargs
  339. #131978 JIT: fold wasm address modes into the load/store memarg
  340. #132027 JIT: Fix async analysis filter dataflow
  341. #131909 Fix for out of range rotate amount during lower
  342. #130512 Accelerate System.Half arithmetic and conversions with FP16
  343. #132014 JIT: rewrite Runtime_131713 without the XML serializer
  344. #131884 JIT: home dead-at-entry wasm GC register params reported to the GC
  345. #131342 Remove boxing for awaited custom awaiters
  346. #131841 JIT: fix three wasm R2R codegen bugs
  347. #131810 [release/10.0] JIT: bail out of CEA cloning if alloc block has multiple successors
  348. #131750 JIT: expand multi-dimensional array intrinsics for large and struct elements
  349. #130481 Recognize bit-manipulation idioms and emit bzhi/bts/btc/btr/bt on xarch
  350. #129688 arm64: Optimize ARM64 compare mask ExtractMostSignificantBits consumers
  351. #131754 JIT: Avoid unnecessary argument spills in indirect call transformer
  352. #131781 [release/10.0] JIT: fix bit-test switch lowering when the bit table is inverted
  353. #131688 JIT: elide redundant wasm Virtual IP updates
  354. #131762 JIT: bail out of CEA cloning if alloc block has multiple successors
  355. #131749 JIT: Handle op1 struct-values in commas in physical promotion
  356. #131736 JIT: fix bit-test switch lowering when the bit table is inverted
  357. #131769 Restore stress guards on process-isolated IL tests
  358. #131731 JIT: Treat the indicator variable as transition-preserved
  359. #130985 Fix codegen assert reloading a spilled signed small cast source
  360. #131661 JIT: use user arg indices when reading call args
  361. #129915 [APX] refactor APX-EVEX emitter
  362. #131442 Normalize TYP_BYREF to TYP_INT in implicit casts even on x86
  363. #131625 Don't share a value number between raw addresses taken in different pinned regions
  364. #131558 [Wasm RyuJit] repair multi-entry try regions
  365. #131559 JIT: evaluate wasm localloc call args to temps
  366. #131488 [arm64] Use the scalar shift when scaling Vector64 shuffle indices
  367. #131560 [RyuJIT Wasm] Remove some unreachable uses of NYI_WASM
  368. #131540 Remove NYI_WASM_SIMD and JitNyiWasmSimdToR2RUnsupported
  369. #128152 JIT: Use a "resumed?" indicator variable in runtime async instead of lvaAsyncContinuationArg != nullptr
  370. #131482 Fix Avx.Compare miscompile with non-constant FloatComparisonMode
  371. #131536 [Tests] Disable inproc crash reporter for intentional crashing tests
  372. #131485 JIT: include element info when comparing array address trees
  373. #131513 JIT: rewrite wasm local.set/local.get as local.tee
  374. #131433 [Wasm RyuJit] assert function regions are contiguous after control flow
  375. #131383 [wasm] Fix JIT helper calls modeled with the wrong return arity
  376. #131471 JIT: don't merge assertions from orphaned pred blocks
  377. #128148 Arm64: Initialization of TYP_SIMD/TYP_MASK locals
  378. #131167 [Wasm Ryujit]: implement runtime-async codegen for R2R Wasm
  379. #127520 Arm64 SVE: Support scalable constant vectors and masks
  380. #131369 JIT: cache the wasm image base in a local
  381. #131362 JIT: recompute exit weight in loop inversion
  382. #130871 JIT: Remove boxing from value type instance calls
  383. #131374 [wasm] Report on-frame GC slots as pinned
  384. #130991 [RyuJIT Wasm] Handle all codegen cases for PackedSimd.Shuffle
  385. #131353 JIT: fold wasm R2R address constants into load memarg offsets
  386. #131328 [wasm][R2R] Fix Vector128 argument alignment in crossgen ArgIterator
  387. #130834 JIT: handle adjacent no-GC regions (#115719)
  388. #131260 SPMI: Allow superpmi.py -details to point to a directory
  389. #130493 JIT: Optimize always suspending helpers
  390. #131314 Correct comment formatting in Runtime_131285.csproj
  391. #131124 JIT: Delete unused functions
  392. #131287 [wasm] Fix R2R codegen trap on async try/return/catch normal-completion branch
  393. #130007 Clear some stale GTF_ORDER_SIDEEFF flags
  394. #130958 [LoongArch64] Fix the m_usComponentSize load in RhpNewArrayFast.
  395. #131171 Fix 130437 testcase for gcstress
  396. #131227 JIT: recognize generic Vector128 intrinsics on wasm
  397. #131164 JIT: don't let tail-merge try and remove the OSR-protected entry block
  398. #131113 [Wasm RyuJit] HW Intrinsic Regalloc: Properly consume temporary from addr operand
  399. #130813 [wasm] Model class-init helper as value-returning in the JIT
  400. #128326 Arm64: Improve support for HW_Flag_ReturnsPerElementMask
  401. #131173 Merge Runtime_131137 into the shared regression runner
  402. #131155 Fix HWIntrinsic codegen for elided scalar/vector reinterprets
  403. #131166 Remove unused args from specifier-less JITDUMP calls
  404. #130832 Fix two latent HWIntrinsic miscompiles in gentree.cpp
  405. #130984 Mark additional commutative xarch and arm64 intrinsics and instructions
  406. #130837 JIT: Cleanup and harden lowering
  407. #130838 Fold floating-point comparisons with a constant NaN in morph
  408. #130907 JIT: fix profile propagation in LowerSwitch
  409. #130990 Reject folding local addresses the emitter cannot encode
  410. #130946 Fix side-effect reordering when folding constant-zero-mask BlendVariable
  411. #130949 Use user-arg accessors when expanding runtime lookups
  412. #131104 JIT: formatting fixes in rangecheck.h
  413. #130843 Fix three xarch lowering correctness/invariant issues
  414. #130822 [RyuJIT Wasm] SIMD Element-Wise Loads and Stores
  415. #130754 [release/10.0] Ignore PhiArgs without matching actual preds
  416. #130324 [release/10.0] JIT: fix widened IV init placed in unreachable block
  417. #130322 [release/10.0] JIT: don't propagate promoted struct LCL_VAR into FIELD_LIST uses
  418. #130867 [release/10.0] Fix tree type in impStoreNullableFields and impLoadNullableFields for simds
  419. #130801 Remove dead/unreachable code in several JIT phases
  420. #130382 Disable large alloc tests on 32b targets
  421. #130836 Wire up orphaned Runtime_129288 regression test
  422. #130803 Remove unused functions from the xarch emitter
  423. #130938 [wasm] Implement WasmBase LeadingZeroCount/TrailingZeroCount JIT intrinsics
  424. #131000 Implement wasm codegen for sub-16 SIMD load/store
  425. #130924 [wasm] Publish __stack_pointer before SuppressGCTransition native calls
  426. #130874 Add missing helix-platforms variables to superpmi-collect
  427. #130913 SuperPMI: fix spurious arm asmdiffs from out-of-range BL relocs
  428. #130974 Update outerloop convert tests for saturating float/double casts
  429. #130936 Add regression test for #124749
  430. #130945 [Wasm RyuJit] fix throw helper layout for side try entries in enclosed regions
  431. #130934 Clarify WASM shuffle/swizzle codegen comments
  432. #130955 Skip profiler leave callback on wasm instead of asserting
  433. #130899 Add RuntimeHelpers.IsRuntimeAsync intrinsic
  434. #130962 Fix build break in PackedSimdTests in outerloop
  435. #128799 Detect and use zicond in ifconversion on riscv64
  436. #130878 [wasm] Correctly materialize integer constants in genCodeForConstant
  437. #130873 Handle cross-targeting Vector<T> size mismatch in the JIT
  438. #130850 Light up remaining SIMD intrinsics on WASM
  439. #130811 JIT: Make GenTreeCall::Equals compare gtCallMoreFlags and more
  440. #128604 JIT, AOT preinit, and CoreLib: Saturate float/double conversions to small integral types
  441. #129717 [WASM] Use relocs for tableBase / imageBase / stackPointer globals
  442. #130417 JIT: preserve null check for unused GT_BLK arg when inlining
  443. #130797 Remove dead switch-table disassembly code from the xarch emitter
  444. #130653 [RyuJIT Wasm] Add JIT\HardwareIntrinsics\Wasm Tests
  445. #130482 JIT: Allow inlining of JIT-inserted AsyncHelpers.TransparentAwait
  446. #130644 Treat Guid and Int128 types as bitwise equatable
  447. #130467 Arm64: keep SP in base operand for large-offset adds
  448. #130202 JIT: add NativeAOT support for non-shared GVM devirtualization
  449. #130081 Recognize Task<->ValueTask adaptions in async versions
  450. #130272 [RyuJIT Wasm] SIMD Codegen for ExtractScalar/ReplaceScalar, Unsigned Compares
  451. #130590 Enable Enum.HasFlag box elision in Tier0
  452. #130444 JIT: Remove transparent scalar/vector conversion HWINTRINSIC nodes during lowering
  453. #130535 JIT: Coalesce stores from struct copies
  454. #130422 JIT: Fold insertps chains that source a scalar extracted from another vector
  455. #130536 Use impInlineRoot()->compIsAsync() in fgGetStaticsCCtorHelper
  456. #128449 JIT: Misc if-conversion
  457. #130546 [LoongArch64] Fix the risk of sign-extension in CodeGen::genIntCastOverflowCheck() with CHECK_SMALL_INT_RANGE signed to signed cast case.
  458. #129852 Remove 128-bit limit on Vector<T> size for ARM64
  459. #130579 [Wasm RyuJit] fix validation error in Microsoft.CodeAnalysis.dll
  460. #130332 Arm64: use FEAT_CSSC cnt/ctz for PopCount and TrailingZeroCount
  461. #130426 Fix range-check monotonicity for multiplication
  462. #130446 JIT: always home register params on wasm
  463. #130401 JIT: Replace promoted fields after morphing indirs in local morph
  464. #130391 [wasm] Bail R2R for all SIMD parameter sizes, not just SIMD16
  465. #130352 JIT: fix GC holes in large stack-target struct block ops
  466. #130326 Create ">=0" assertion for unpromoted Span.Length
  467. #130317 JIT: Make all static base helpers byref typed in async functions
  468. #130267 [Wasm] handle non-catch try entries in SCC switch dispatch
  469. #130070 JIT: Disallow splitting prologs and epilogs across multiple fragments
  470. #130161 Add base lightup support for most of the WASM intrinsics
  471. #130256 Replace SetCallEntrypointForR2R with gtNewUserCallNode
  472. #128375 JIT: don't propagate promoted struct LCL_VAR into FIELD_LIST uses
  473. #130107 JIT: Coalesce adjacent constant-address stores
  474. #130190 JIT: fix widened IV init placed in unreachable block
  475. #130301 Arm64: Only call checkRMWRegisters for RMW intrinsics
  476. #130255 Cleanup the AvxOnlyCompatible checks to ensure that we don't try to generate unsupported nodes for AVX only hardware
  477. #130205 JIT: fold trivial redundant comparisons on loop induction variables
  478. #130252 JIT: Centralize WellKnownArg definitions
  479. #130074 JIT: use type argument as the exact type for Activator.CreateInstance<T>
  480. #130282 Enable JIT\jit64 caninline test on NativeAOT
  481. #130218 Fix variant interface GVM resolution in managed type system
  482. #130260 Block tail merges across handler and filter
  483. #130222 Fix incorrect constant folding of CompareScalar True/False comparisons
  484. #130217 JIT: fix incorrect bounds check removal for Phi indices with a negative minimum
  485. #130181 JIT: Enable tail await optimization in async versions in tier 0
  486. #130153 [Wasm/RyuJit] treat all try side-entries as try entries for throw helpers
  487. #130186 JIT: initialize emitCurIG at declaration
  488. #129325 More cleanups in rangecheck/assertionprop
  489. #127746 JIT: bail on flipping post-layout JTRUE when reversal needs new IR
  490. #130146 Update OutOfMemoryException test
  491. #130049 Update the JIT to share the xplat vector tables across platforms
  492. #130131 Fix build break in codegenwasm
  493. #128721 [RyuJit/WASM] Remove the special prolog handling
  494. #130120 Re-enable Arm64 hwintrinsic tests
  495. #130101 [wasm] Bail R2R for SIMD16 store-indirect and SIMD16 parameters
  496. #130059 Fix x86 codegen non-determinism across host architecture
  497. #130086 Ensure that op_Equality/Inequality can actually import as expected
  498. #129932 Arm64: enable ldp/stp peephole for negative offsets
  499. #129255 JIT: Enable implicit tailcalls for tail-awaits
  500. #129864 [arm64] Avoid sign-extending TYP_INT register moves
  501. #130013 wasm: order funclets in RPO by funclet index
  502. #127164 arm64: Clean up SVE embedded masked codegen
  503. #129948 SPMI: per-context tpdiff examples
  504. #130000 wasm: report the generic context to the GC stack walk
  505. #129901 Handle Synchronized async versions
  506. #130003 JIT: fix asserts in fgRemoveUnreachableTry
  507. #129141 JIT: Tail-merge robust crossJumpVictim selection
  508. #129703 [Wasm RyuJIT] Codegen for Const Vector Create and Basic Packed SIMD operations
  509. #129528 JIT: Reset compGenTreeID on failed Materialize to avoid leaking gen tree IDs
  510. #128702 Support devirtualization for virtual methods that require an instantiating stub
  511. #129898 JIT: Remove InlineCandidateInfo::exactContextNeedsRuntimeLookup
  512. #129856 JIT: ensure magic divisors are always marked DONT_CSE for all targets
  513. #129890 Save async contexts for ValueTask methods
  514. #129914 JIT: fix stale VN in redundant branch opts dominating-branch simplification
  515. #128365 Add AvxVnni.V512 hardware intrinsics
  516. #129917 JIT/wasm: keep EH frame slots at the bottom after frame alignment
  517. #129903 JIT/wasm: pass type-context arg for LDVIRTFTN calls
  518. #129807 JIT/wasm: Pri-1 R2R codegen fixes (BBJ_CALLFINALLY back-edge, null-check loads)
  519. #129868 JIT: only skip loop inversion when the IV test is the bottom test
  520. #129819 [arm64] Emit cmlt/cmle/fcmlt/fcmle for compares against zero
  521. #129793 Remove undefined-behavior UnsafeTests.DangerousAs to stabilize jitstress leg
  522. #129589 Arm64 R2R: fold adrp+add+ldr into adrp+ldr[#:lo12:]
  523. #129810 JIT: Intrinsify Task/ValueTask factory methods
  524. #129870 [release/9.0-staging] Sve: Fix Gather/Scatter tests
  525. #129643 JIT: Fix "Inconsistent profile data" during Head and tail merge (Tier1-OSR)
  526. #129533 JIT: fix opt-repeat divide-by-zero miscompile (#129386)
  527. #129816 JIT/wasm: bail R2R for calls passing 16-byte SIMD by value
  528. #129815 [wasm R2R] Align 8-byte stack locals and keep frames 16-aligned
  529. #129517 arm64: Add float/double cast optimisation tests
  530. #129788 Fix IsKnownNonNull assert for SIMD base addresses
  531. #129716 JIT/wasm: more EH related fixes
  532. #129722 JIT: invert oversize loops with their own bounds checks
  533. #129563 Ensure other types of SIMD related loads, stores, and indirections are marked as used with SIMD intrinsics
  534. #128384 Compile runtime async versions of synchronous task-returning methods
  535. #128668 Have IDIV/DIV REX2 and EVEX compatible
  536. #129447 Improve nullability analysis in ValueNumStore::IsKnownNonNull
  537. #129555 Fix wasm R2R JIT validation issues
  538. #128700 Fix Arm64 conditional select lowering
  539. #128524 JIT: Switch to rangecheck when profitable
  540. #129285 JIT: remove getUnboxedEntry and getInstantiatedEntry
  541. #129646 JIT: fix edge likelihoods for loop cloning
  542. #129655 JIT: add fgRemoveUnreachableTry phase
  543. #129636 Disable scope info for wasm targets
  544. #129644 Fix x64 ToScalar XMM→GPR codegen for contained CreateScalar operand
  545. #129658 Properly annotate CreateGeometricSequence
  546. #129553 Enforce 75-minute timeout floor for windows_arm64 global build jobs
  547. #129058 JIT: Fix for IV opt to preserve updates to a counter that is live into an EH handler
  548. #129494 [Wasm RyuJit] Add Wasm SIMD NYIs, fix PushArgs assert, fix Async/Try/SCC issue.
  549. #128970 Fix ARM32 JIT errors with native int-div
  550. #128962 Fix for 127747
  551. #129354 Use SSA-based TryGetRange to fold conditions
  552. #129519 JIT: don't narrow relocatable handle constants in optNarrowTree
  553. #129127 Clarify JIT-EE method placement in add-new-jit-ee-api skill
  554. #129472 JIT: boost loop inversion cost limit for some loops
  555. #129397 JIT: unpin locals whose value is provably non-movable
  556. #129361 JIT: extend OptimizeConstCompare to remove TYP_BYTE casts
  557. #129059 [Wasm RyuJIT] Spill live ref/byref values to pinned stack slots at calls
  558. #129201 [Wasm RyuJit] Enable Hardware Intrinsics
  559. #129449 Wasm R2R: emit trailing end after retless BBJ_CALLFINALLY
  560. #129392 JIT: recognize CORINFO_HELP_BOX in impThrowIfNull
  561. #129349 JIT: clone array loops with stride > 57
  562. #129390 Improve RangeOps::Merge
  563. #129371 Don't give up on folding obj.GetType if obj can be null
  564. #129391 Wasm R2R: implement support for resume after catch
  565. #129429 JIT: mark R2R direct-call target loads as non-faulting
  566. #129034 Remove NativeAOT debug handle tagging
  567. #129343 [RyuJIT Wasm]: Fix incorrect assert for shift amount type in codegenwasm.cpp
  568. #129134 [Wasm RyuJit] Enable native wasm fast tail calls
  569. #129348 JIT: don't claim rotates set ZF on xarch
  570. #129312 JIT: forward-sub cheap address temps with multiple uses
  571. #129309 JIT: extend loop cloning for span+stride>1 and ±const limits
  572. #123138 arm64: Add support for BFI and BFX instruction
  573. #129253 JIT: Clean up range check overflow caching and bounds check special cases
  574. #129303 JIT: tighten IV range from "i != bound" assertions in rangecheck
  575. #129248 JIT: encapsulate all fields in GenTreeIntCon
  576. #129206 JIT: Allow optNarrowTree CAST case to match ULONG and LONG equivalently
  577. #129306 JIT/wasm: fix quadratic cost in fgWasmEhFlow
  578. #129269 JIT: stop walking parent try regions across funclet boundaries
  579. #129268 JIT: clone "i != arr.Length" loops to eliminate bounds checks
  580. #128906 Update GetRangeFromAssertions to handle some basic TYP_LONG scenarios where it FitsIn<int32_t>
  581. #129092 JIT: Tail-merge don't produce empty blocks
  582. #129205 [Wasm RyuJit]: don't skip top-level catchrets when marking try resumption
  583. #129157 JIT: Clear out return value references from continuations
  584. #129022 JIT: weight loop body cost by block frequency in cloning heuristic
  585. #129144 Clean up in rangecheck
  586. #128811 JIT: Fix debug igDataSize assignment to use code size (so exclude gc vars)
  587. #128479 Represent managed-return-values as native vars
  588. #129188 JIT: Fix arm64 shifted-register containment for long compares
  589. #129187 JIT: fix loop cloning condition for decreasing loops with array-length limit
  590. #129095 [Wasm R2R] Fix two crossgen2 NYIs: FloatTraits::NaN and bitcast of same-size types
  591. #129148 JIT: Narrow long comparisons of sign-extended ints to int in VN
  592. #129102 [Wasm R2R] Retype integer constant struct returns to the native return type
  593. #128909 [release/10.0] JIT: Fix CEA cloning when enumerator local is reassigned from an untracked source
  594. #129061 [Wasm / RyuJit]: implement CKFINITE
  595. #129062 [Wasm RyuJit] Match how other targets consume initblk fill values
  596. #129139 Remove redundant commandline.h include in streamingsuperpmi.cpp
  597. #129106 [Wasm R2R] Fix unsigned add/sub overflow codegen
  598. #126348 JIT: Improve codegen for xarch vector byte multiply
  599. #129018 Change jit fuzzer build to pull Antigen from jitutils
  600. #129084 Remove PHASE_MERGE_THROWS
  601. #129097 JIT: remove unsound NOT(relop) VN simplification
  602. #129101 Misc improvements for rangecheck
  603. #128848 Track bool-returning and scalar-T-returning HW intrinsics via flags
  604. #129100 [Wasm R2R] Report tracked GC vars as untracked in GC info
  605. #129075 [Wasm R2R] Bypass dynamic LDVIRTFTN helper
  606. #129073 [Wasm RyuJit] Zero init gc shadow stack slots in the prolog
  607. #129025 JIT/wasm: place SCC dispatcher inside Wasm try region when needed
  608. #129060 [Wasm / RyuJit]: don't mark CORINFO_HELP_GETREFANY pure on wasm
  609. #129007 Enable must-throw for unsupported intrinsics under optimizations
  610. #128515 JIT: Merge all RETURN/THROW blocks
  611. #129026 SuperPMI: re-apply JitWasmNyiToR2RUnsupported when ignoring stored config
  612. #128960 [release/10.0] JIT: bail from optOptimizeBoolsCondBlock on GTF_UNSIGNED inputs
  613. #129044 JIT test: remove SimdLoadBetweenBCs from ElidedBoundsChecks
  614. #128967 JIT: Superpmi wasm32 neardiff
  615. #127439 JIT: coalesce constant-indexed bounds checks within a block
  616. #128729 JIT: fix loop cloning type assumptions
  617. #128965 Remove dead SIMD contiguous-args code in the JIT
  618. #128533 JIT: Canonicalize '(A & pow2) == pow2' -> '(A & pow2) != 0'
  619. #128963 JIT: fix slow preheader EH bookkeeping in loop cloning
  620. #128951 JIT: fix profile likelihood and overflow UB in optimizebools
  621. #128959 JIT: Fix assertion failed 'genActualType(cmp->gtGetOp1()) == genActualType(cmp->gtGetOp2())'
  622. #128928 JIT: bail from optOptimizeBoolsCondBlock on GTF_UNSIGNED inputs
  623. #128922 Have ComputeRange call into GetRangeFromAssertions for non dependent/symbolic cases
  624. #128924 JIT: remove optCanProp* short-circuit flags
  625. #128926 [Wasm RyuJit] Enable codegen for wasm funclets by default
  626. #128770 Fix incorrect CoercedConstantValue usage in IsVNIntegralConstant
  627. #128904 Wasm spmi use checked jit
  628. #128778 JIT: don't kill FP/SIMD/mask regs across x64 write barriers
  629. #128513 JIT: add post-inline cleanup for no-return calls
  630. #127079 JIT: Fix CEA cloning when enumerator local is reassigned from an untracked source
  631. #128870 JIT: add profitability heuristic for loop cloning
  632. #128678 Ensure Log2 can actually be imported as intrinsic
  633. #128731 Fix for 127141 - Remove unneeded spilling around profiler hooks
  634. #128855 JIT/x64: Route FMA embedded-rounding fallback through genFmaIntrinsic
  635. #128091 JIT: Enable fgOptimizeRelationalComparisonWithCasts for EQ/NE
  636. #128819 Wasm spmi collection2
  637. #128857 Fix SPMI wasm cross-target collection and revise must-expand assert
  638. #128722 Ensure that applicable intrinsics have their ranges tracked by vn/assertionprop/rngchk
  639. #128836 JIT: Remove a TODO comment
  640. #128805 JIT: fix x86 GC hole in stack-to-stack struct copies
  641. #128683 JIT: Unify LowerBlockStore for init across targets
  642. #128684 Arm64: Don't use GT_LEA for masks
  643. #128791 JIT: remove dead BBJ_RETURN-in-loop assertion in loop cloner
  644. #128715 [Wasm RyuJIT] Fix declared local off-by-one in reverse P-Invoke case
  645. #128669 Add wasm coverage to SuperPMI CI pipelines
  646. #128724 [Release/10.0] JIT: Fix bug in GenTree::Equals (#128654)
  647. #128677 Improve the handling of popcnt and tzcnt on arm64
  648. #128701 JIT: avoid redundant nullcheck in LowerBlockStoreAsGcBulkCopyCall
  649. #128712 JIT: Make async tail merging layout equality assert a check
  650. #128541 JIT: skip stack allocation for unused locals
  651. #128662 JIT: Encapsulate VNFuncApp fields and clean up FuncIs/IsVNBinFunc usage
  652. #128654 JIT: Fix bug in GenTree::Equals
  653. #126947 JIT: Unify to instParamLookup and improve GVM test coverage
  654. #128542 Remove CORINFO_HELP_ASSIGN_BYREF
  655. #128620 Ensure other cases of lzcnt/tzcnt/popcount are handled by rangecheck
  656. #128459 JIT: skip loop inversion for loops that are already bottom-tested
  657. #128607 JIT: remove some useless code in escape analysis
  658. #128532 JIT: generalize cloning conditions slightly
  659. #128382 [Wasm RyuJit] emit virtual IP ranges in the unwind info
  660. #128587 JIT: fix overlap check in store coalescing
  661. #128522 Make O1K_LCLVAR local assertprop only
  662. #128559 JIT: Tail merge async suspensions
  663. #128573 JIT: Defer fgSetOptions until after 2nd-pass EH removal
  664. #128521 Help ensure we get the smallest encoding for commutative instructions
  665. #128561 Add jit.instructions.md for JIT folder
  666. #128537 Fix invalid IL test
  667. #127889 JIT: Avoid TLS access when restoring async contexts
  668. #128500 JIT: fold isinst using reaching VNs and compareTypesForCast
  669. #128485 JIT: Propagate class handle when split tree
  670. #128229 [Wasm RyuJIT] Packed Simd Encodings
  671. #128472 [Wasm RyuJIT] Implement contained bitcast operands
  672. #128419 Fix for index out of bounds in sve scatter test
  673. #128447 JIT: Disallow CSE of multireg nodes
  674. #117261 Improve Math.BigMul performance on x64
  675. #127883 JIT: Unify post-order cmp transforms into fgOptimizeRelationalComparison
  676. #128395 Don't pessimize available registers for key GT_INTRINSIC
  677. #128381 [release/8.0-staging] Fix loop hoist memory-dependence tracking
  678. #128274 JIT: Rename fields in indirect call transformer
  679. #128271 Remove -Wno-unused-function -Wno-tautological-compare -Wno-unused-value for clrjit projects
  680. #127915 JIT: Transform 'SELECT(cond, cns, cns)' to 'cns'
  681. #125298 JIT: Add debug consistency checks for GT_RETURN nodes
  682. #128350 Remove unnecessary delayfree from xarch FMA and TERNLOG instructions
  683. #128303 JIT: Canonicalize loop backedges to a single latch block
  684. #126552 [RyuJit] Remove the single-IG prolog restriction
  685. #128348 [Wasm RyuJit] codegen for callfinally
  686. #125729 Pass Vector<T> by reference when InstructionSet_VectorT is available
  687. #127838 Arm64: [PAC-RET] Jit Support
  688. #128307 Break apart gtFoldExprConst into a couple functions so its more manageable
  689. #128243 Delete obsolete CORINFO_TYPE_... entries
  690. #128270 JIT: don't make local copy when devirtualizing boxed interface call
  691. #127536 [AMD64][APX] Introduce the remaining NCI instructions - CTEST, CFCMOV
  692. #128238 Use GTF_IND_TGT_NOT_HEAP/GTF_IND_TGT_HEAP for GT_STORE_BLK
  693. #127003 [release/10.0] Fix assertions generated by optCreateJumpTableImpliedAssertions
  694. #128222 Add support for generating pointers to funclets, and calling funclets by index
  695. #128250 JIT: Move invariant nodes and LCL_VARs in LiftLIREdges
  696. #127932 JIT: Decouple "DoNotEnregister" from liveness
  697. #128242 [Wasm/Ryujit] Fix if stack depth; fix pep call ordering
  698. #128230 Delete unused WBF_NoBarrier_CheckNotHeapInDebug and CORINFO_HELP_ASSIGN_REF_ENSURE_NONHEAP
  699. #128187 Move the folding of double not/neg to gtFoldExpr
  700. #128165 Unify O2K_CHECKED_BOUND_ADD_CNS and O2K_VN
  701. #128066 Fix runtime-async on riscv64 with R2R
  702. #128172 [Wasm RyuJIT] Fix: Use passed in ig when possible for IF_CODE_SIZE in emitDispIns()
  703. #128146 JIT: restore type-based upper bound for RSZ in range analysis
  704. #127416 JIT: Fix morph FIELD_ADDR null check reordering for stores
  705. #128117 Ensure we don't do duplicate checks for various multi-reg paths
  706. #127935 Build wasm JIT and crossgen2 R2R of System.Private.CoreLib for browser target
  707. #128073 JIT: Eagerly invalidate GC state for registers trashed by linker relaxation sequences
  708. #127949 Arm64:[PAC-RET] Instruction encodings
  709. #128037 Arm64: Fix loads and stores for stack allocated vectors and masks
  710. #127967 [ci-scan] Exclude Vector3Interop from GC stress via ActiveIssue (refs #127827)
  711. #128013 Ensure gtInitializeStoreNode consistently handles SIMD/Mask stores for all platforms
  712. #128033 Remove some dead code from impAppendStmt and impStoreStruct
  713. #124415 More precise ranges for casts in GetRangeFromAssertions
  714. #127950 Fold more branches via O1K_VN <relop> O2K_VN comparisons
  715. #127615 JIT: Fix GenTree::IsPow2 functions
  716. #127972 Cleanup VisitOperandUses and GenTreeVisitor to check common kinds rather than be a big switch
  717. #127684 JIT: remove MorphAddrContext
  718. #127959 JIT: fix value-probing schema index collision with handle histogram
  719. #127917 Support spill temps with unknown size on ARM64
  720. #127933 JIT: Fix SELECT to relop changing type and check cond evaluates to 0 or 1
  721. #127910 JIT: Fix untracked uses of lvLiveInOutOfHndlr
  722. #127050 Fix for 126457
  723. #127751 [Wasm RyuJIT] Codegen for reverse pinvoke/unmanagedcallersonly
  724. #127621 JIT: Clean up and unify SELECT opts
  725. #125491 Implement UnknownSizeFrame for locals with unknown size
  726. #127810 JIT: Remove redundant null checks for non-null PHIs
  727. #127450 JIT: Restore EH-region check in if-conversion flow validation
  728. #127659 Intrinsify string.FastAllocateString
  729. #127781 Centralize a couple JIT arithmetic helpers
  730. #127773 [Wasm RyuJit] add length prefix to each jit code contribution
  731. #125554 [RyuJit] Decouple igNum from strict layout order
  732. #126932 [Wasm RyuJIT] Wire up GC info encoding/decoding for Wasm
  733. #127648 Fix INT_MIN overflow in RangeCheck::BetweenBounds
  734. #127555 JIT: Use VisitLocalDefNodes from gtTreeHasLocalStore
  735. #127689 Ensure we preserve vn and produce a canonical tree
  736. #127661 Update fgMorphSmpOp to canonicalize commutative and compare ops at the start of POST-ORDER
  737. #127590 JIT: make funclet order correspond to EH clause order
  738. #127487 JIT: avoid store forward stall for struct params in GS frames
  739. #127624 JIT: Simplify ifdef for R2R needsIndirectionCellArg condition in morph
  740. #127632 Remove an unnecessary switch from GenTree:TryGetUse
  741. #127619 [Wasm RyuJit] Use local's register type when loading it's value
  742. #127528 Split the xarch specific data from register.h into its own file to match other platforms
  743. #124738 JIT: Treat store in JTRUE block as the ElseOperation in if-conversion
  744. #127533 JIT: VN-level (a-x)+x and a-(a-x) folds
  745. #127384 Update OSR docs for recent changes
  746. #126778 [Wasm RyuJit] Call Codegen Fixes for R2R
  747. #127412 [Wasm RyuJIT]: Fix LIR Semantics in Stackifier Output
  748. #125180 JIT: Accelerate floating->long casts on x86
  749. #126360 JIT: Fully share helper-based suspension tails
  750. #127033 [X64][APX] Fix a few bugs in APX implementation and improve the control knob design
  751. #127148 Replace PERFSCORE_MEMORY_* macros with PerfScoreMemoryAccessKind enum class
  752. #127267 Rename JitDasmWithAddress to JitDisasmWithAddress
  753. #127524 JIT: fix Vector{128,256}.ConvertToInt32 use-before-def in gtNewSimdCvtNode
  754. #127213 Remove unwind calls from patchpoint OSR transition
  755. #127477 Handle GT_CCMP in GenTree::OperConsumesFlags on TARGET_AMD64
  756. #126529 JIT: Transform 'CONST - x' to 'xor x, CONST' when possible
  757. #127488 JIT: eliminate redundant bounds checks for span.Slice(span.Length - cns)
  758. #126041 Add size/perf optimized FinishSuspension helpers for runtime async
  759. #127327 Extend System.Runtime.Intrinsics.Arm to support nint and nuint (ArmBase + Crc32)
  760. #126562 JIT: STORE_LCL_FLD coalescing
  761. #127434 JIT: Use normalized block for comparison in RBO dominator walk
  762. #127433 JIT: relax inliner heuristics for callees on [Intrinsic] types
  763. #127413 Add [NoInlining] to Marvin.ComputeHash32OrdinalIgnoreCase
  764. #127402 Allow O2K_CONST_VEC to handle all TYP_SIMD32/64
  765. #127329 JIT: Reduce Vector256/512 Sum to a shared per-lane reduction on x64
  766. #127181 JIT: relop simplification during redundant branch opts
  767. #127357 JIT: Fix profile inconsistency asserts in flow graph optimization
  768. #127363 JIT: Tolerate edge/successor count mismatch at OSR entry in profile reconstruction
  769. #127293 JIT: Fix uninitialized gtCallMethHnd for indirect calls causing false recursive tail call detection
  770. #126062 JIT: Handle edge cases around BlendVariableMask rewrite
  771. #127124 JIT: propagate vector constants through assertion prop
  772. #127184 JIT: Avoid changing def registers that may assume an implicit kill
  773. #127074 JIT: Optimize async OSR resumptions
  774. #127307 ignore unreachable preds in optVisitReachingAssertions
  775. #127117 JIT: eliminate redundant span checks across loop back-edge phis
  776. #127043 [Wasm RyuJit] EH, Unwind, and unwindable frames
  777. #127185 [Wasm RyuJit] Fix: Move implReadyToRunUnsupported() call for funclets before wasm codegen
  778. #127219 JIT: Remove doesMethodHavePartialCompilationPatchpoints
  779. #127158 JIT: Disallow using float callee saves in x64 functions with patchpoints
  780. #126880 JIT: Restore arm64, LA64 and RISCV64 OSR callee saves from tier0 frame
  781. #126907 JIT: refine SSA info in ambiguously threaded blocks
  782. #127167 Ensure that we properly set the morphed flag
  783. #127103 JIT: Renable SSA-aware PHI jump threading
  784. #127052 Remove _DbgBreakCheck calls from the JIT
  785. #127109 Cleanup the xarch perfscore handling
  786. #125333 JIT: Avoid resolving def-use conflicts by changing use registers
  787. #127111 Fix the encoding of vcvtps2ph
  788. #127048 Improve simd and floating-point costing for xarch
  789. #127094 Add support for utilizing F16C instructions on xarch
  790. #126398 arm64: Refactor mov/movprfx for embedded masked operations
  791. #126371 JIT: Fix disasm of reg to reg SIMD narrowing/widening instructions
  792. #126566 [RISC-V] Use fine-grained fence variants for memory barriers
  793. #127042 Revert "Accelerate Half with FP16 ISA"
  794. #126263 superpmi: Refactor parallelsuperpmi.cpp to eliminate scattered #ifdef blocks and remove PAL usage
  795. #126772 [RyuJIT] Make the mis-sized struct return logic only apply to return values that are structs
  796. #127018 JIT: Fix loongarch64 Windows build
  797. #122649 Accelerate Half with FP16 ISA
  798. #126938 [LoongArch64] Initialize emitter::getInsExecutionCharacteristics() to support perfscore.
  799. #126939 [release/10.0] JIT: Handle writes to promoted locals in AliasSet::AddNode
  800. #126885 [release/10.0] JIT: fix bug in loop cloning with down-counting loops
  801. #126913 [release/9.0-staging] JIT: fix bug in loop cloning with down-counting loops
  802. #126914 [release/8.0-staging] JIT: fix bug in loop cloning with down-counting loops
  803. #126917 Improve constant foldings in optAssertionPropGlobal_RelOp
  804. #126920 JIT: Precompute func-app element stride in VN chunks
  805. #125516 [RISC-V] Support displaying GTF_ADDRMODE_NO_CSE flag on GT_CAST
  806. #125347 JIT: Simplify if-conversion
  807. #126645 JIT: Handle writes to promoted locals in AliasSet::AddNode
  808. #126910 Ensure that AVX512_MoveMask and other key mask internal mask intrinsics are hoistable
  809. #126514 [Wasm RyuJIT] Fixes for two codegen asserts
  810. #126865 Ensure IsVectorPerElementMask does a correct check and use it in more places
  811. #126856 JIT: Elide BOUNDS_CHECK(0, length) when length != 0 is proven by assertion
  812. #126770 JIT: fix bug in loop cloning with down-counting loops
  813. #126812 JIT: support SSA-aware PHI jump threading
  814. #124637 Add instruction set detection for SVE_AES, SVE_SHA3, SVE_SM4, SHA3, SM4
  815. #126796 JIT: Remove VNFOA_SharedStatic, ValueNumStore::IsSharedStatic, and shared-static column from valuenumfuncs.h
  816. #125211 [RyuJit] Remove gtCallAddr
  817. #126803 ARM64: Use fmov instead of umov for 64-bit integer ToScalar
  818. #126736 Arm64 jitstress outerloop
  819. #125613 [release/10.0] Fix missing call to write barrier in static constructors of ref structs
  820. #126663 [Wasm RyuJit] fix funclet prolog and epilog codegen
  821. #126587 JIT: dominating redundant branch opt
  822. #126666 Add watchdog instrumentation to OleTxTests.Recovery to capture hang diagnostics
  823. #125174 JIT: Share suspension epilogs in async transformation
  824. #125226 Support InstructionSet_VectorT in HWIntrinsic importer
  825. #126588 JIT: use funclet iterator in genCodeForBBlist
  826. #126445 Wasm funclet RA
  827. #126374 [Wasm RyuJit] transform multi-entry try regions
  828. #126434 Fix for arm64 outerloop jitstress failures in FMA
  829. #125821 JIT: Remove dead Vector2/3 lowering code
  830. #126530 Revert "Inline CORINFO_HELP_ARRADDR_ST helper call, remove WriteBarrier FCall"
  831. #126491 JIT: Funclet iterators
  832. #126392 [Wasm RyuJIT] Fix for assert in obscure block store scenario
  833. #126375 [Wasm RyuJIT] Set wasm's vector width to 128 bits
  834. #126326 [Wasm RyuJIT] Partially implement genStructReturn
  835. #126123 Fix R2R TypeMap fallback for unresolved assembly targets
  836. #126327 [Wasm RyuJit] multi entry try preliminaries
  837. #126028 JIT: Refactor async analysis
  838. #126130 [Wasm RyuJit] Use CORINFO_HELP_INITCLASS for fgGetSharedCCtor
  839. #121985 JIT: Constant fold SequenceEqual with the help of VN
  840. #126000 ARM64: Display method names for indirect calls in JitDisasm output
  841. #126223 JIT: yet another pass of unreachable block removal
  842. #126266 JIT: Remove SSE2 CmpOp/BinOp emulation code
  843. #126296 Fix jit format
  844. #126169 [RyuJit] Coalesce adjacent no-GC regions on x86
  845. #125981 SPMI: Fixes to record-and-replay workflows
  846. #125874 SPMI: Fix invalid format string and allow parsing full uint32 range
  847. #126256 [Wasm RyuJit] more INDEX_ADDR codegen fixes
  848. #126209 [Wasm RyuJIT] Basic implementation of int to float casts without overflow checks
  849. #124794 Rewrite ConvertToDecimal in XslNumber.cs to use safe code with string.Create and for loop
  850. #126238 [Wasm RyuJit] Fix another TARGET_64BIT issue
  851. #126254 [Wasm RyuJIT] Fix build break
  852. #125770 [Wasm RyuJIT] Address block store issues from #125756
  853. #126215 [Wasm RyuJit] Track stackifier temps by actual jit type
  854. #126111 [Wasm RyuJit] bounds checks for GT_INDEX_ADDR
  855. #126064 Fix OperIs(GT_CNS_INT, GT_CNS_LNG) assertion during import
  856. #123717 arm64: Refactor mov/movprfx for unmasked operations
  857. #125175 [release/10.0] Fix bug in LowerCallMemmove
  858. #126059 [Wasm RyuJit] codegen for loading a register-sized struct local
  859. #125747 Fix for 125246
  860. #125836 [Wasm RyuJIT] initblk fix for #125756
  861. #125615 JIT: Add a runtime async optimization to skip saving unmutated locals into reused continuations
  862. #125862 Add support for generalized variable length wasm instructions to the emitter
  863. #125962 JIT: Fix jit-format break
  864. #124332 arm64: Fold negative variable eq/ne to 0
  865. #124642 JIT: Fix fgFoldCondToReturnBlock for multi-statement return blocks
  866. #125446 Improve runtime-async support on linux-riscv64
  867. #125869 JIT: Null out exceptions in reused continuations and reenable continuation reuse
  868. #125852 [Wasm RyuJit] Debug SPC Wasm Validation Fixes
  869. #125812 Fix issue where i32.trunc_sat_f32_s wasn't properly encoded
  870. #125831 JIT: Disable continuation reuse temporarily
  871. #125594 [Wasm RyuJit] eh flow
  872. #125752 [Wasm RyuJit] Add JIT config JitWasmNyiToR2RUnsupported
  873. #125556 Add runtime async support for saving and reusing continuation instances
  874. #125669 [Wasm RyuJIT] Insert r2r ind cell at the end of the args list instead of after 'this'
  875. #125531 [Wasm RyuJIT] Generate element section and populate function pointer table
  876. #125422 Fix crossgen2 JitHost OOM crash: throw on allocation failure
  877. #125497 JIT: Create suspensions/resumptions separately from processing calls in async transformation
  878. #125418 Fix missing call to write barrier in static constructors of ref structs
  879. #125219 JIT: Preference SDSU intervals away from killed registers
  880. #125364 Exclude noisy crypto rules from the repo's CodeQL runs
  881. #125515 JIT: Bump JIT-EE GUID after removal of hardware intrinsics
  882. #125428 arm64: Fix and/bic issue when bailing early
  883. #125425 fix #125301 (NOL bug for partial loads/stores)
  884. #124516 Support for variable sized locals in morph and lclmorph
  885. #125442 Add html.escape to func_name in superpmi diffs report
  886. #123546 arm64: Remove widening casts before truncating
  887. #125382 [LoongArch64] Fix the risk of boundary overflow for Unsigned GT_LE in CodeGen::genCodeForCompare().
  888. #125072 JIT: Fix invalid IR and weights in if-conversion
  889. #124213 [LoongArch64] Place async resumption info in read-write section which reference #123433
  890. #125362 JIT: Stop treating GT_LCLHEAP as inherently exceptional and remove dead StackOverflowException flag
  891. #125398 [Wasm RyuJIT] Move genHomeRegisterParams' output out of the prolog
  892. #125342 [Wasm RyuJIT] Disable multireg ret assert on Wasm because we don't have multireg rets there
  893. #125284 [RyuJIT] Add a JIT flag for Portable Entry Points and set it by default for Browser compilations
  894. #125378 JIT: Fix patchpoint info out-of-bounds read for special OSR locals
  895. #125331 JIT: Disallow op2 for mulx to be in edx
  896. #125141 JIT: Expand simple stfld addresses early and spill data nodes to retain evaluation order
  897. #125279 [Wasm RyuJIT] Handle case where a TYP_STRUCT return has had its type information erased and we are trying to write it to a TYP_STRUCT local
  898. #125153 [Wasm RyuJIT] Call signatures
  899. #125278 [Wasm RyuJIT] Fix classifier embiggening small structs that are being passed-as a wasm primitive
  900. #125292 JIT: Clean up comments and remove redundancy in flowgraph.cpp
  901. #125267 JIT: Disallow partial compilation of runtime async functions
  902. #125214 JIT: Handle some def/use conflicts less conservatively in LSRA
  903. #125239 [RyuJit] Remove dead code from genCallInstruction
  904. #125102 [Wasm RyuJit] More crossgen assert fixes
  905. #125235 Fix bounds checks in #123085 repro
  906. #125222 Change bbAssertionOut if BBJ_COND was folded in morph
  907. #125225 Fix getAsyncOtherVariant JIT-EE implementation in VM
  908. #125146 [RyuJit/WASW] Insert temporaries when needed in the stackifier
  909. #125156 Add visit budget checks to IsMonotonicallyIncreasing and ComputeDoesOverflow
  910. #125172 SVE: truncate Extract results for small types
  911. #125187 [RyuJit/WASM] Disable args sorting
  912. #125188 JIT: fix don't remove issue for loop/try split
  913. #125100 More flowgraph cleanup
  914. #125093 Ignore PhiArgs without matching actual preds
  915. #125056 Improve range analysis for checked bounds in GlobalAP
  916. #125028 JIT: Lower aggressiveRefCnt minimum threshold for CSE promotion
  917. #124977 Fix JIT assert resolving to C runtime assert after jitshared introduction
  918. #124804 Add AVX512 BMM API
  919. #125073 [RyuJit] Add VisitOperandUses
  920. #124114 JIT: Accelerate uint->floating casts on pre-AVX-512 x86
  921. #124943 Add metric usage diff to SPMI reports
  922. #125101 Wasm loop aware rpo compute last block in loop quickly
  923. #125078 [Wasm RyuJIT] Block Stores follow-ups
  924. #124601 arm64: Fixed using ands/bics to compare against static unsigned constant 0
  925. #125055 Use SetValueTruncating instead of SetIconValue
  926. #124880 [Wasm RyuJit] annotate (some) local accesses with var nums
  927. #123146 JIT: Normalize optimizations for ConditionalSelect and BlendVariable
  928. #125045 JIT: Fix late devirtualization debug info for inlining
  929. #122485 [RISC-V] Enable instruction printing for the Generate code phase in JIT dumps
  930. #124846 [Wasm RyuJIT] Block Stores pt. 2
  931. #125027 Array stack iterators
  932. #124915 Cleanup basic block
  933. #125044 JIT: Fix reordering of inlinee return expression with async save/restore
  934. #124953 Remove ActiveIssue from src/tests/JIT/jit64/regress/vsw/373472/test.il
  935. #124798 JIT: Refactor async-to-sync call optimization
  936. #124670 Move diagnostics out of DownlevelLibraryImportGenerator into a separate analyzer
  937. #124975 Introduce DOTNET_JitReportMetrics
  938. #124882 Cleanup flowgraph files
  939. #124955 [release/11.0-preview2] Revert "Increase number of assertions (GlobalAP) + VN cache (#124132)"
  940. #124928 Revert "Increase number of assertions (GlobalAP) + VN cache (#124132)"
  941. #123830 Change Interpreter to share code with the JIT and use its CompAllocator
  942. #124809 JIT: Fix bugs and typos in objectalloc.cpp, extract AnalyzePseudoForCloning
  943. #124250 [Wasm RyuJit] localloc
  944. #124865 Sve2 Scatters need a temp register for indices
  945. #124715 Fix assert in JIT:new() during early initialization
  946. #124504 Arm64: Add SM4 and SHA3 instructions
  947. #124765 Fix Arm64: for 124357
  948. #124731 [Wasm RyuJit] internal registers, integer binop overflow checks
  949. #124837 [Wasm RyuJit] more fixes from crossgen2 of spc
  950. #124820 Delete CORINFO_HELP_ENDCATCH
  951. #124567 JIT: Convert multi-target switches to branchless checks
  952. #124599 Less optimistic checking for const vector to pattern checks
  953. #124763 [Wasm RyuJit] Crossgen fixes
  954. #124759 JIT: minor cleanups in RBO
  955. #123738 [Wasm RyuJIT] Block stores
  956. #124705 Allow range check cloning for BBJ_RETURN blocks
  957. #124711 More AP-related optimizations
  958. #124710 [RyuJit/WASM] Rewrite SP references into PHYSREGs
  959. #124679 [Wasm RyuJit] fix some issues with calls
  960. #124683 Add CORINFO_RUNTIME_LOOKUP::helperEntryPoint
  961. #124575 [Wasm RyuJit] codegen for some intrinsics
  962. #124673 [Wasm RyuJit] fix issue in RewriteLocalStackStore
  963. #124514 JIT: Add a "default value analysis" for async hoisting
  964. #124534 [Wasm RyuJit] Null checks for calls, overflow checks for int to int casts, resolve stack ordering, more
  965. #124582 Write ExpectedExitCode marker in GenerateMarkerFiles
  966. #124605 Fix abandoned jit regression tests
  967. #124572 [RyuJit] Remove the gtCallCookie operand from calls
  968. #124383 SPMI: Fix relocs into new RO data section
  969. #124533 Fix RangeOps::Multiply to handle dependent ranges
  970. #124571 Eliminate redundant bounds checks for arr[^N-1] after arr[^N]
  971. #124503 JIT: Disqualify inlinees using AsyncSuspend intrinsic
  972. #124472 JIT: Handle LCL_ADDR in TreeLifeUpdater
  973. #124502 JIT: Check for inline candidates in gtTreeContainsAsyncCall
  974. #124481 [RyuJit/WASM] Register allocation for multiply used operands
  975. #123992 Add InstructionSet_VectorT for ARM64
  976. #124399 Fix aspnet2 spmi collection script
  977. #124387 Track span.Length's negativity + remove unsound assumption from MergeEdgeAssertions
  978. #124298 [RyuJit/WASM] Implement internal registers
  979. #124432 Fix tree type in impStoreNullableFields and impLoadNullableFields for simds
  980. #124373 JIT: Remove dataSection flexible array
  981. #123515 [Wasm RyuJit] Wasm abi classifier
  982. #124362 [Wasm RyuJit] More codegen and fixes for crossgen replay
  983. #124283 Avoid resolving direct awaits to thunks
  984. #124360 JIT: don't set liveness on unreachable blocks
  985. #124343 SPMI: Run superpmi-diffs on windows-x64
  986. #124205 [Wasm RyuJIT] Start generating relocations for call_indirect
  987. #124286 JIT: some GT_INDEX_ADDRs won't need throw helpers
  988. #124335 JIT: Fix immediate assert when relocs are enabled
  989. #122638 Introduce local variables with unknown size
  990. #124291 Remove nullcheck effect from array len (global AP)
  991. #124290 JIT: clear must keep flag for blocks created in GDV expansions
  992. #124306 Fix failed ElidedBoundsChecks test
  993. #124132 Increase number of assertions (GlobalAP) + VN cache
  994. #124242 Remove bounds checks for 'i + cns < len' pattern
  995. #124249 JIT: refactor GS enabling
  996. #124253 JIT: fix liveness in throw helper blocks
  997. #123907 Fix bug in LowerCallMemmove
  998. #124184 Remove redundant checked contexts around casts + more bounds checks
  999. #124183 JIT: Enable inlining of runtime async methods without awaits
  1000. #124177 Lower GT_RETURN_SUSPEND blocks to BBJ_RETURN
  1001. #124138 [Wasm RyuJit] bitcast
  1002. #124189 [Wasm RyuJit] avoid reverse ops
  1003. #124224 Arm64 Sve: Fix double/float mask test issues
  1004. #124147 Optimize redundant checked context in JIT
  1005. #124001 JIT: Compute async liveness for address exposed locals
  1006. #124137 [Wasm RyuJit] fix local type encoding
  1007. #124115 [Wasm Ryujit] Wasm SP is implicitly referenced
  1008. #124113 [Wasm RyuJit] Ind write barriers, direct helper calls, more
  1009. #124065 Remove O1K_BOUND_OPER_BND and O1K_BOUND_LOOP_BND
  1010. #124028 Wasm call fixes and more
  1011. #123554 [RyuJit] Remove indentation from target*.h files
  1012. #124018 [RyuJit/WASM] Simple RA
  1013. #123500 JIT: Run full liveness in MinOpts for async
  1014. #123666 Fix SVE long narrowing for shifts
  1015. #124030 Introduce OAK_
  1016. #123999 JIT: Further liveness cleanups
  1017. #123955 Make all fields in AssertionDscOp1 and AssertionDscOp2 private
  1018. #123993 SPMI: Fix TOC logic broken after PAL cleanup
  1019. #123948 JIT: Refactor liveness into its own class
  1020. #123926 JIT: Standardize Compiler* field naming to m_compiler
  1021. #123945 Don't ping codegen team for codegen label
  1022. #123044 [Wasm RyuJIT] Initial scaffolding for calls and helper calls
  1023. #123856 Remove a few TP-heavy GlobalAP optimizations
  1024. #122501 [RISC-V] Fix 64-bit offset handling in emitLoadImmediate
  1025. #123433 Place async resumption info in read-write section outside Windows for AOT
  1026. #123183 Support devirtualizing non-shared GVMs in R2R
  1027. #123908 [Wasm Ryujit] Flow cannot fall through into a throw helper
  1028. #123845 Introduce factory methods to create assertions
  1029. #117571 JIT: Move remaining xarch floating->integral cast implementation to lowering
  1030. #123022 Arm64: Fix for 96441
  1031. #123840 SPMI: Update superpmi-diffs queue to OSX.15.Arm64.Open
  1032. #123781 JIT: remove anticipated demand for throw helpers
  1033. #123839 SPMI: Update superpmi-collect queues from osx13 to osx15
  1034. #123794 Fix bug in RangeOps::ShiftRight
  1035. #122970 Remove CHK_FLOW_UPDATE stress mode
  1036. #123584 Fix RangeOps::And incorrect range analysis for bitwise AND
  1037. #123716 Use CORJIT_PREJIT_UNSUPPORTED instead of CORJIT_LIMITATION for BlockNonDeterministicIntrinsics
  1038. #123723 [Draft] Try installing Dotnet SDK on Helix Runners for SPMI Benchmark Collections
  1039. #123624 Fold comparisons into single-value ranges in GetRangeFromAssertions
  1040. #123743 JIT: fix retyping issue in object stack allocation
  1041. #123053 [Wasm RyuJit] throw helper / null check preliminaries
  1042. #123664 Fix JIT assertion for AsyncContinuation in GDV compatibility check
  1043. #123489 JIT: Remove tailcall "has GC safe point" assert
  1044. #122088 [RISC-V] Enhance the utilization of sh1/2/3add(.uw) instructions
  1045. #123707 SPMI: Fix aligned allocation size computation
  1046. #123631 [Wasm RyuJit] proper fix for register homing
  1047. #123378 Add allocMem support for allocating multiple data chunks
  1048. #123585 Add FileCheck test for elided bounds checks
  1049. #123603 JIT: Fix mask AND with zero incorrectly folded
  1050. #123569 [RyuJit Wasm] don't home locals if none are referenced
  1051. #123563 Analyze reaching PHI in GetRangeFromAssertions (more branch foldings)
  1052. #123565 [RyuJit Wasm] fix off by one error for br_table
  1053. #122862 [RyuJIT Wasm] Cast Operations Follow Up
  1054. #123522 [Wasm RyuJit] Fix emission of block and label instructions
  1055. #123233 Add binary VN funcs to TryGetRangeFromAssertions
  1056. #123536 SPMI: Tolerate missing getContinuationType calls
  1057. #123262 [Wasm RyuJit] Wasm stack arg preliminaries
  1058. #123434 JIT: Always use CallInfo method handle for devirtualization
  1059. #123343 JIT: Make async insertion of "rethrow BB" deterministic to fix debug info
  1060. #123362 Fix array size overflow in objectalloc.cpp
  1061. #123270 [RyuJit/WASM] Establish SP & FP and home register arguments
  1062. #123261 [RyuJit/WASM] Some fixes related to local addresses and stores
  1063. #123240 Fix assertions generated by optCreateJumpTableImpliedAssertions
  1064. #122382 [RISC-V] correct comments in emitLoadImmediate
  1065. #123021 [Wasm RyuJIT] Implement 8 and 16-bit loads/stores; implement indirect loads/stores
  1066. #122973 JIT: Make GS cookie phase run on LIR and move it after async phase
  1067. #123154 Fix aspnet2 spmi
  1068. #123152 [RyuJit/WASM] Refactoring in preparation for parameter homing
  1069. #123099 [RyuJit/WASM] Add stackification algorithm skeleton
  1070. #122263 Optimize more bound checks around logical operators
  1071. #123128 JIT: Fix build after two conflicting changes
  1072. #122381 Fix impGetNodeAddr flag handling for volatile, unaligned, and initclass accesses
  1073. #122023 JIT: Devirtualize non-shared generic virtual methods
  1074. #121527 [Bounds checks] Merge inter-block assertions in RangeCheck::MergeAssertion
  1075. #122779 Intrinsify Enum.Equals to avoid boxing
  1076. #123019 [Wasm RyuJIT] Add missing TYP_REF check
  1077. #122967 JIT: Mark non-internal block as imported
  1078. #122924 JIT: Fix GC info for fast tailcalls with contained targets
  1079. #122994 [Wasm RyuJIT] Add some missing types to ins_Load and ins_Store
  1080. #121584 arm64: Refactor GenerateHWIntrinsicTests_Arm.cs
  1081. #122897 [RyuJIT Wasm] Disable genCreateAddrMode on Wasm
  1082. #121739 [RISC-V] Enable constant CSE in RISCV64
  1083. #122167 Inline nullable allocators
  1084. #122946 JIT: fix CEA to handle a few more cases
  1085. #122368 Handle some additional floating-point and SIMD morph optimizations
  1086. #122297 JIT: Expand recognition of struct field offset pattern
  1087. #121672 JIT: Support async contexts for OSR/EnC
  1088. #122873 JIT: Fix alignment for custom layouts with GC refs in them
  1089. #122902 JIT: SpanHelpers.Fill may escape its last argument
  1090. #122866 JIT: Look harder for async calls in ASYNC_CONTINUATION lowering
  1091. #122869 JIT: Disallow tailcall stress in async methods
  1092. #122872 Remove FEATURE_EH_WINDOWS_X86 and all non-funclet EH code from JIT
  1093. #122627 Wasm various fixes
  1094. #122865 JIT: Detect degenerate switches in switch recognition
  1095. #122490 JIT: Zero continuation return after profiler leave callback
  1096. #122727 Handle more async nodes on community platforms
  1097. #121745 Move RBM_SAVED_LOCALLOC_SP reservation from compCompile to setFrameType in LSRA
  1098. #122589 [RISC-V] Comment and formatting cleanup in codegenriscv64
  1099. #122652 Add NoInlining and NoOptimization attributes to ClearWaitSleepJoinState method