Alex Zinenko 8c2ea32072 Emit LLVM IR equivalent of sizeof when lowering alloc operations ... Originally, the lowering of `alloc` operations has been computing the number of bytes to allocate when lowering based on the properties of MLIR type. This does not take into account type legalization that happens when compiling LLVM IR down to target assembly. This legalization can widen the type, potentially leading to out-of-bounds accesses to `alloc`ed data due to mismatches between address computation that takes the widening into account and allocation that does not. Use the LLVM IR's equivalent of `sizeof` to compute the number of bytes to be allocated: %0 = getelementptr %type* null, %indexType 0 %1 = ptrtoint %type* %0 to %indexType adapted from http://nondot.org/sabre/LLVMNotes/SizeOf-OffsetOf-VariableSizedStructs.txt PiperOrigin-RevId: 274159900		2019-10-11 06:33:26 -07:00
..
GPUToCUDA	Use llvm.func to define functions with wrapped LLVM IR function type	2019-10-10 01:34:06 -07:00
GPUToNVVM	Add AllReduceOp to GPU dialect with lowering to NVVM.	2019-09-26 00:17:50 -07:00
GPUToROCDL	[ROCm] Adding pass to lower GPU Dialect to ROCDL Dialect.	2019-10-02 01:50:30 -07:00
GPUToSPIRV	Make SPIR-V lowering infrastructure follow Vulkan SPIR-V validation.	2019-10-09 11:25:58 -07:00
LoopsToGPU	Replace linalg.for by loop.for	2019-07-16 13:44:57 -07:00
StandardToLLVM	Emit LLVM IR equivalent of sizeof when lowering alloc operations	2019-10-11 06:33:26 -07:00
StandardToSPIRV	Add lowering of constant ops to SPIR-V.	2019-10-10 17:19:57 -07:00
VectorToLLVM	LLVM dialect: prefix auxiliary operations with "mlir."	2019-09-03 09:10:56 -07:00
convert-to-cfg.mlir	NFC: rename Conversion/ControlFlowToCFG to Conversion/LoopToStandard	2019-10-03 01:35:03 -07:00