mirror of
https://github.com/llvm/llvm-project.git
synced 2025-04-25 07:26:06 +00:00
158d72d728
Set the writable and dead_on_unwind attributes for sret arguments. These indicate that the argument points to writable memory (and it's legal to introduce spurious writes to it on entry to the function) and that the argument memory will not be used if the call unwinds. This enables additional MemCpyOpt/DSE/LICM optimizations.
32 lines
1.4 KiB
C
32 lines
1.4 KiB
C
// REQUIRES: arm-registered-target
|
|
// RUN: %clang_cc1 -triple thumbv7-apple-darwin9 \
|
|
// RUN: -target-abi apcs-gnu \
|
|
// RUN: -target-cpu cortex-a8 \
|
|
// RUN: -mfloat-abi soft \
|
|
// RUN: -target-feature +soft-float-abi \
|
|
// RUN: -ffreestanding \
|
|
// RUN: -emit-llvm -w -o - %s | FileCheck %s
|
|
|
|
#include <arm_neon.h>
|
|
|
|
// CHECK: define{{.*}} void @f0(ptr dead_on_unwind noalias writable sret(%struct.int8x16x2_t) align 16 %agg.result, <16 x i8> noundef %{{.*}}, <16 x i8> noundef %{{.*}})
|
|
int8x16x2_t f0(int8x16_t a0, int8x16_t a1) {
|
|
return vzipq_s8(a0, a1);
|
|
}
|
|
|
|
// Test direct vector passing.
|
|
|
|
typedef float T_float32x2 __attribute__ ((__vector_size__ (8)));
|
|
typedef float T_float32x4 __attribute__ ((__vector_size__ (16)));
|
|
typedef float T_float32x8 __attribute__ ((__vector_size__ (32)));
|
|
typedef float T_float32x16 __attribute__ ((__vector_size__ (64)));
|
|
|
|
// CHECK: define{{.*}} <2 x float> @f1_0(<2 x float> noundef %{{.*}})
|
|
T_float32x2 f1_0(T_float32x2 a0) { return a0; }
|
|
// CHECK: define{{.*}} <4 x float> @f1_1(<4 x float> noundef %{{.*}})
|
|
T_float32x4 f1_1(T_float32x4 a0) { return a0; }
|
|
// CHECK: define{{.*}} void @f1_2(ptr dead_on_unwind noalias writable sret(<8 x float>) align 32 %{{.*}}, <8 x float> noundef %{{.*}})
|
|
T_float32x8 f1_2(T_float32x8 a0) { return a0; }
|
|
// CHECK: define{{.*}} void @f1_3(ptr dead_on_unwind noalias writable sret(<16 x float>) align 64 %{{.*}}, <16 x float> noundef %{{.*}})
|
|
T_float32x16 f1_3(T_float32x16 a0) { return a0; }
|