758ddba381
LLVM defines several default datalayouts for integer and floating point types that are not being considered when importing into MLIR. This patch remedies this. Reviewed By: ftynse Differential Revision: https://reviews.llvm.org/D120832
12 lines
445 B
LLVM
12 lines
445 B
LLVM
; RUN: mlir-translate -import-llvm %s | FileCheck %s
|
|
|
|
; CHECK: dlti.dl_spec =
|
|
; CHECK: #dlti.dl_spec<
|
|
; CHECK: #dlti.dl_entry<"dlti.endianness", "little">
|
|
; CHECK: #dlti.dl_entry<i64, dense<64> : vector<2xi32>>
|
|
; CHECK: #dlti.dl_entry<f80, dense<128> : vector<2xi32>>
|
|
; CHECK: #dlti.dl_entry<i8, dense<8> : vector<2xi32>>
|
|
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
|
|
|
|
declare void @foo()
|