[llvm] [msan] Add baseline output for neon_vst_float.ll (PR #100210)
Thurston Dang via llvm-commits
llvm-commits at lists.llvm.org
Tue Jul 23 14:12:29 PDT 2024
https://github.com/thurstond created https://github.com/llvm/llvm-project/pull/100210
Followup to https://github.com/llvm/llvm-project/pull/100189/
>From d940b21510c995eed218a14f457992f4f7627ec7 Mon Sep 17 00:00:00 2001
From: Thurston Dang <thurston at google.com>
Date: Tue, 23 Jul 2024 21:11:18 +0000
Subject: [PATCH] [msan] Add baseline output for neon_vst_float.ll
Followup to https://github.com/llvm/llvm-project/pull/100189/
---
.../MemorySanitizer/AArch64/neon_vst_float.ll | 1091 ++++++++++++++++-
1 file changed, 1089 insertions(+), 2 deletions(-)
diff --git a/llvm/test/Instrumentation/MemorySanitizer/AArch64/neon_vst_float.ll b/llvm/test/Instrumentation/MemorySanitizer/AArch64/neon_vst_float.ll
index 54b570584ec21..e9881930b700a 100644
--- a/llvm/test/Instrumentation/MemorySanitizer/AArch64/neon_vst_float.ll
+++ b/llvm/test/Instrumentation/MemorySanitizer/AArch64/neon_vst_float.ll
@@ -1,9 +1,9 @@
-; No assertions yet because the test case crashes MSan
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 4
;
; Test memory sanitizer instrumentation for Arm NEON VST_{2,3,4} and
; VST_1x{2,3,4} instructions, including floating-point parameters.
;
-; RUN: opt < %s -passes=msan -S | FileCheck %s
+; RUN: opt < %s -passes=msan -S -disable-verify | FileCheck %s
;
; UNSUPPORTED: {{.*}}
;
@@ -37,271 +37,1358 @@ target triple = "aarch64--linux-android9001"
; -----------------------------------------------------------------------------------------------------------------------------------------------
define void @st1x2_v1f64(<1 x double> %A, <1 x double> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x2_v1f64(
+; CHECK-SAME: <1 x double> [[A:%.*]], <1 x double> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0:[0-9]+]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0:![0-9]+]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4:[0-9]+]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x2.v1f64.p0(<1 x double> [[A]], <1 x double> [[B]], ptr [[A]])
+;
+; EDITOR'S NOTE: the next call is invalid because the parameters (shadows) are integer, but the called function
+; expects floating-point parameters.
+;
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x2.v1f64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x2.v1f64.p0(<1 x double> %A, <1 x double> %B, ptr %a)
ret void
}
define void @st1x2_v1i64(<1 x i64> %A, <1 x i64> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x2_v1i64(
+; CHECK-SAME: <1 x i64> [[A:%.*]], <1 x i64> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x2.v1i64.p0(<1 x i64> [[A]], <1 x i64> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x2.v1i64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x2.v1i64.p0(<1 x i64> %A, <1 x i64> %B, ptr %a)
ret void
}
define void @st1x2_v2f64(<2 x double> %A, <2 x double> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x2_v2f64(
+; CHECK-SAME: <2 x double> [[A:%.*]], <2 x double> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x2.v2f64.p0(<2 x double> [[A]], <2 x double> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x2.v2f64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x2.v2f64.p0(<2 x double> %A, <2 x double> %B, ptr %a)
ret void
}
define void @st1x2_v2i64(<2 x i64> %A, <2 x i64> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x2_v2i64(
+; CHECK-SAME: <2 x i64> [[A:%.*]], <2 x i64> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x2.v2i64.p0(<2 x i64> [[A]], <2 x i64> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x2.v2i64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x2.v2i64.p0(<2 x i64> %A, <2 x i64> %B, ptr %a)
ret void
}
define void @st1x3_v1f64(<1 x double> %A, <1 x double> %B, <1 x double> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x3_v1f64(
+; CHECK-SAME: <1 x double> [[A:%.*]], <1 x double> [[B:%.*]], <1 x double> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x3.v1f64.p0(<1 x double> [[A]], <1 x double> [[B]], <1 x double> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x3.v1f64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], <1 x i64> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x3.v1f64.p0(<1 x double> %A, <1 x double> %B, <1 x double> %C, ptr %a)
ret void
}
define void @st1x3_v1i64(<1 x i64> %A, <1 x i64> %B, <1 x i64> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x3_v1i64(
+; CHECK-SAME: <1 x i64> [[A:%.*]], <1 x i64> [[B:%.*]], <1 x i64> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x3.v1i64.p0(<1 x i64> [[A]], <1 x i64> [[B]], <1 x i64> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x3.v1i64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], <1 x i64> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x3.v1i64.p0(<1 x i64> %A, <1 x i64> %B, <1 x i64> %C, ptr %a)
ret void
}
define void @st1x3_v2f64(<2 x double> %A, <2 x double> %B, <2 x double> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x3_v2f64(
+; CHECK-SAME: <2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x3.v2f64.p0(<2 x double> [[A]], <2 x double> [[B]], <2 x double> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x3.v2f64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], <2 x i64> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x3.v2f64.p0(<2 x double> %A, <2 x double> %B, <2 x double> %C, ptr %a)
ret void
}
define void @st1x3_v2i64(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x3_v2i64(
+; CHECK-SAME: <2 x i64> [[A:%.*]], <2 x i64> [[B:%.*]], <2 x i64> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x3.v2i64.p0(<2 x i64> [[A]], <2 x i64> [[B]], <2 x i64> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x3.v2i64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], <2 x i64> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x3.v2i64.p0(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C, ptr %a)
ret void
}
define void @st1x4_v1f64(<1 x double> %A, <1 x double> %B, <1 x double> %C, <1 x double> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x4_v1f64(
+; CHECK-SAME: <1 x double> [[A:%.*]], <1 x double> [[B:%.*]], <1 x double> [[C:%.*]], <1 x double> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x4.v1f64.p0(<1 x double> [[A]], <1 x double> [[B]], <1 x double> [[C]], <1 x double> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x4.v1f64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], <1 x i64> [[TMP4]], <1 x i64> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x4.v1f64.p0(<1 x double> %A, <1 x double> %B, <1 x double> %C, <1 x double> %D, ptr %a)
ret void
}
define void @st1x4_v1i64(<1 x i64> %A, <1 x i64> %B, <1 x i64> %C, <1 x i64> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x4_v1i64(
+; CHECK-SAME: <1 x i64> [[A:%.*]], <1 x i64> [[B:%.*]], <1 x i64> [[C:%.*]], <1 x i64> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x4.v1i64.p0(<1 x i64> [[A]], <1 x i64> [[B]], <1 x i64> [[C]], <1 x i64> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x4.v1i64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], <1 x i64> [[TMP4]], <1 x i64> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x4.v1i64.p0(<1 x i64> %A, <1 x i64> %B, <1 x i64> %C, <1 x i64> %D, ptr %a)
ret void
}
define void @st1x4_v2f64(<2 x double> %A, <2 x double> %B, <2 x double> %C, <2 x double> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x4_v2f64(
+; CHECK-SAME: <2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], <2 x double> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x4.v2f64.p0(<2 x double> [[A]], <2 x double> [[B]], <2 x double> [[C]], <2 x double> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x4.v2f64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], <2 x i64> [[TMP4]], <2 x i64> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x4.v2f64.p0(<2 x double> %A, <2 x double> %B, <2 x double> %C, <2 x double> %D, ptr %a)
ret void
}
define void @st1x4_v2i64(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C, <2 x i64> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st1x4_v2i64(
+; CHECK-SAME: <2 x i64> [[A:%.*]], <2 x i64> [[B:%.*]], <2 x i64> [[C:%.*]], <2 x i64> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x4.v2i64.p0(<2 x i64> [[A]], <2 x i64> [[B]], <2 x i64> [[C]], <2 x i64> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st1x4.v2i64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], <2 x i64> [[TMP4]], <2 x i64> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st1x4.v2i64.p0(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C, <2 x i64> %D, ptr %a)
ret void
}
define void @st2_v16i8(<16 x i8> %A, <16 x i8> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v16i8(
+; CHECK-SAME: <16 x i8> [[A:%.*]], <16 x i8> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <16 x i8>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <16 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v16i8.p0(<16 x i8> [[A]], <16 x i8> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v16i8.p0(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v16i8.p0(<16 x i8> %A, <16 x i8> %B, ptr %a)
ret void
}
define void @st2_v1f64(<1 x double> %A, <1 x double> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v1f64(
+; CHECK-SAME: <1 x double> [[A:%.*]], <1 x double> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v1f64.p0(<1 x double> [[A]], <1 x double> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v1f64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v1f64.p0(<1 x double> %A, <1 x double> %B, ptr %a)
ret void
}
define void @st2_v1i64(<1 x i64> %A, <1 x i64> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v1i64(
+; CHECK-SAME: <1 x i64> [[A:%.*]], <1 x i64> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v1i64.p0(<1 x i64> [[A]], <1 x i64> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v1i64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v1i64.p0(<1 x i64> %A, <1 x i64> %B, ptr %a)
ret void
}
define void @st2_v2f32(<2 x float> %A, <2 x float> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v2f32(
+; CHECK-SAME: <2 x float> [[A:%.*]], <2 x float> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v2f32.p0(<2 x float> [[A]], <2 x float> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v2f32.p0(<2 x i32> [[TMP2]], <2 x i32> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v2f32.p0(<2 x float> %A, <2 x float> %B, ptr %a)
ret void
}
define void @st2_v2f64(<2 x double> %A, <2 x double> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v2f64(
+; CHECK-SAME: <2 x double> [[A:%.*]], <2 x double> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v2f64.p0(<2 x double> [[A]], <2 x double> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v2f64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v2f64.p0(<2 x double> %A, <2 x double> %B, ptr %a)
ret void
}
define void @st2_v2i32(<2 x i32> %A, <2 x i32> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v2i32(
+; CHECK-SAME: <2 x i32> [[A:%.*]], <2 x i32> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v2i32.p0(<2 x i32> [[A]], <2 x i32> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v2i32.p0(<2 x i32> [[TMP2]], <2 x i32> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v2i32.p0(<2 x i32> %A, <2 x i32> %B, ptr %a)
ret void
}
define void @st2_v2i64(<2 x i64> %A, <2 x i64> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v2i64(
+; CHECK-SAME: <2 x i64> [[A:%.*]], <2 x i64> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v2i64.p0(<2 x i64> [[A]], <2 x i64> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v2i64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v2i64.p0(<2 x i64> %A, <2 x i64> %B, ptr %a)
ret void
}
define void @st2_v4f16(<4 x half> %A, <4 x half> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v4f16(
+; CHECK-SAME: <4 x half> [[A:%.*]], <4 x half> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v4f16.p0(<4 x half> [[A]], <4 x half> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v4f16.p0(<4 x i16> [[TMP2]], <4 x i16> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v4f16.p0(<4 x half> %A, <4 x half> %B, ptr %a)
ret void
}
define void @st2_v4f32(<4 x float> %A, <4 x float> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v4f32(
+; CHECK-SAME: <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v4f32.p0(<4 x float> [[A]], <4 x float> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v4f32.p0(<4 x i32> [[TMP2]], <4 x i32> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v4f32.p0(<4 x float> %A, <4 x float> %B, ptr %a)
ret void
}
define void @st2_v4i16(<4 x i16> %A, <4 x i16> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v4i16(
+; CHECK-SAME: <4 x i16> [[A:%.*]], <4 x i16> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v4i16.p0(<4 x i16> [[A]], <4 x i16> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v4i16.p0(<4 x i16> [[TMP2]], <4 x i16> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v4i16.p0(<4 x i16> %A, <4 x i16> %B, ptr %a)
ret void
}
define void @st2_v4i32(<4 x i32> %A, <4 x i32> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v4i32(
+; CHECK-SAME: <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v4i32.p0(<4 x i32> [[A]], <4 x i32> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v4i32.p0(<4 x i32> [[TMP2]], <4 x i32> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v4i32.p0(<4 x i32> %A, <4 x i32> %B, ptr %a)
ret void
}
define void @st2_v8f16(<8 x half> %A, <8 x half> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v8f16(
+; CHECK-SAME: <8 x half> [[A:%.*]], <8 x half> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v8f16.p0(<8 x half> [[A]], <8 x half> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v8f16.p0(<8 x i16> [[TMP2]], <8 x i16> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v8f16.p0(<8 x half> %A, <8 x half> %B, ptr %a)
ret void
}
define void @st2_v8i16(<8 x i16> %A, <8 x i16> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v8i16(
+; CHECK-SAME: <8 x i16> [[A:%.*]], <8 x i16> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v8i16.p0(<8 x i16> [[A]], <8 x i16> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v8i16.p0(<8 x i16> [[TMP2]], <8 x i16> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v8i16.p0(<8 x i16> %A, <8 x i16> %B, ptr %a)
ret void
}
define void @st2_v8i8(<8 x i8> %A, <8 x i8> %B, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st2_v8i8(
+; CHECK-SAME: <8 x i8> [[A:%.*]], <8 x i8> [[B:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i8>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP4:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP5:%.*]] = xor i64 [[TMP4]], 193514046488576
+; CHECK-NEXT: [[TMP6:%.*]] = inttoptr i64 [[TMP5]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP7:%.*]], label [[TMP8:%.*]], !prof [[PROF0]]
+; CHECK: 7:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 8:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v8i8.p0(<8 x i8> [[A]], <8 x i8> [[B]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st2.v8i8.p0(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], ptr [[TMP6]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st2.v8i8.p0(<8 x i8> %A, <8 x i8> %B, ptr %a)
ret void
}
define void @st3_v16i8(<16 x i8> %A, <16 x i8> %B, <16 x i8> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v16i8(
+; CHECK-SAME: <16 x i8> [[A:%.*]], <16 x i8> [[B:%.*]], <16 x i8> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <16 x i8>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <16 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <16 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v16i8.p0(<16 x i8> [[A]], <16 x i8> [[B]], <16 x i8> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v16i8.p0(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v16i8.p0(<16 x i8> %A, <16 x i8> %B, <16 x i8> %C, ptr %a)
ret void
}
define void @st3_v1f64(<1 x double> %A, <1 x double> %B, <1 x double> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v1f64(
+; CHECK-SAME: <1 x double> [[A:%.*]], <1 x double> [[B:%.*]], <1 x double> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v1f64.p0(<1 x double> [[A]], <1 x double> [[B]], <1 x double> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v1f64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], <1 x i64> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v1f64.p0(<1 x double> %A, <1 x double> %B, <1 x double> %C, ptr %a)
ret void
}
define void @st3_v1i64(<1 x i64> %A, <1 x i64> %B, <1 x i64> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v1i64(
+; CHECK-SAME: <1 x i64> [[A:%.*]], <1 x i64> [[B:%.*]], <1 x i64> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v1i64.p0(<1 x i64> [[A]], <1 x i64> [[B]], <1 x i64> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v1i64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], <1 x i64> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v1i64.p0(<1 x i64> %A, <1 x i64> %B, <1 x i64> %C, ptr %a)
ret void
}
define void @st3_v2f32(<2 x float> %A, <2 x float> %B, <2 x float> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v2f32(
+; CHECK-SAME: <2 x float> [[A:%.*]], <2 x float> [[B:%.*]], <2 x float> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v2f32.p0(<2 x float> [[A]], <2 x float> [[B]], <2 x float> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v2f32.p0(<2 x i32> [[TMP2]], <2 x i32> [[TMP3]], <2 x i32> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v2f32.p0(<2 x float> %A, <2 x float> %B, <2 x float> %C, ptr %a)
ret void
}
define void @st3_v2f64(<2 x double> %A, <2 x double> %B, <2 x double> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v2f64(
+; CHECK-SAME: <2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v2f64.p0(<2 x double> [[A]], <2 x double> [[B]], <2 x double> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v2f64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], <2 x i64> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v2f64.p0(<2 x double> %A, <2 x double> %B, <2 x double> %C, ptr %a)
ret void
}
define void @st3_v2i32(<2 x i32> %A, <2 x i32> %B, <2 x i32> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v2i32(
+; CHECK-SAME: <2 x i32> [[A:%.*]], <2 x i32> [[B:%.*]], <2 x i32> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v2i32.p0(<2 x i32> [[A]], <2 x i32> [[B]], <2 x i32> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v2i32.p0(<2 x i32> [[TMP2]], <2 x i32> [[TMP3]], <2 x i32> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v2i32.p0(<2 x i32> %A, <2 x i32> %B, <2 x i32> %C, ptr %a)
ret void
}
define void @st3_v2i64(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v2i64(
+; CHECK-SAME: <2 x i64> [[A:%.*]], <2 x i64> [[B:%.*]], <2 x i64> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v2i64.p0(<2 x i64> [[A]], <2 x i64> [[B]], <2 x i64> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v2i64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], <2 x i64> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v2i64.p0(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C, ptr %a)
ret void
}
define void @st3_v4f16(<4 x half> %A, <4 x half> %B, <4 x half> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v4f16(
+; CHECK-SAME: <4 x half> [[A:%.*]], <4 x half> [[B:%.*]], <4 x half> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v4f16.p0(<4 x half> [[A]], <4 x half> [[B]], <4 x half> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v4f16.p0(<4 x i16> [[TMP2]], <4 x i16> [[TMP3]], <4 x i16> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v4f16.p0(<4 x half> %A, <4 x half> %B, <4 x half> %C, ptr %a)
ret void
}
define void @st3_v4f32(<4 x float> %A, <4 x float> %B, <4 x float> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v4f32(
+; CHECK-SAME: <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v4f32.p0(<4 x float> [[A]], <4 x float> [[B]], <4 x float> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v4f32.p0(<4 x i32> [[TMP2]], <4 x i32> [[TMP3]], <4 x i32> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v4f32.p0(<4 x float> %A, <4 x float> %B, <4 x float> %C, ptr %a)
ret void
}
define void @st3_v4i16(<4 x i16> %A, <4 x i16> %B, <4 x i16> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v4i16(
+; CHECK-SAME: <4 x i16> [[A:%.*]], <4 x i16> [[B:%.*]], <4 x i16> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v4i16.p0(<4 x i16> [[A]], <4 x i16> [[B]], <4 x i16> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v4i16.p0(<4 x i16> [[TMP2]], <4 x i16> [[TMP3]], <4 x i16> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v4i16.p0(<4 x i16> %A, <4 x i16> %B, <4 x i16> %C, ptr %a)
ret void
}
define void @st3_v4i32(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v4i32(
+; CHECK-SAME: <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]], <4 x i32> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v4i32.p0(<4 x i32> [[A]], <4 x i32> [[B]], <4 x i32> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v4i32.p0(<4 x i32> [[TMP2]], <4 x i32> [[TMP3]], <4 x i32> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v4i32.p0(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C, ptr %a)
ret void
}
define void @st3_v8f16(<8 x half> %A, <8 x half> %B, <8 x half> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v8f16(
+; CHECK-SAME: <8 x half> [[A:%.*]], <8 x half> [[B:%.*]], <8 x half> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v8f16.p0(<8 x half> [[A]], <8 x half> [[B]], <8 x half> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v8f16.p0(<8 x i16> [[TMP2]], <8 x i16> [[TMP3]], <8 x i16> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v8f16.p0(<8 x half> %A, <8 x half> %B, <8 x half> %C, ptr %a)
ret void
}
define void @st3_v8i16(<8 x i16> %A, <8 x i16> %B, <8 x i16> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v8i16(
+; CHECK-SAME: <8 x i16> [[A:%.*]], <8 x i16> [[B:%.*]], <8 x i16> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v8i16.p0(<8 x i16> [[A]], <8 x i16> [[B]], <8 x i16> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v8i16.p0(<8 x i16> [[TMP2]], <8 x i16> [[TMP3]], <8 x i16> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v8i16.p0(<8 x i16> %A, <8 x i16> %B, <8 x i16> %C, ptr %a)
ret void
}
define void @st3_v8i8(<8 x i8> %A, <8 x i8> %B, <8 x i8> %C, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st3_v8i8(
+; CHECK-SAME: <8 x i8> [[A:%.*]], <8 x i8> [[B:%.*]], <8 x i8> [[C:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i8>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <8 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP5:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP6:%.*]] = xor i64 [[TMP5]], 193514046488576
+; CHECK-NEXT: [[TMP7:%.*]] = inttoptr i64 [[TMP6]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP8:%.*]], label [[TMP9:%.*]], !prof [[PROF0]]
+; CHECK: 8:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 9:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v8i8.p0(<8 x i8> [[A]], <8 x i8> [[B]], <8 x i8> [[C]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st3.v8i8.p0(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], <8 x i8> [[TMP4]], ptr [[TMP7]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st3.v8i8.p0(<8 x i8> %A, <8 x i8> %B, <8 x i8> %C, ptr %a)
ret void
}
define void @st4_v16i8(<16 x i8> %A, <16 x i8> %B, <16 x i8> %C, <16 x i8> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v16i8(
+; CHECK-SAME: <16 x i8> [[A:%.*]], <16 x i8> [[B:%.*]], <16 x i8> [[C:%.*]], <16 x i8> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <16 x i8>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <16 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <16 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <16 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v16i8.p0(<16 x i8> [[A]], <16 x i8> [[B]], <16 x i8> [[C]], <16 x i8> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v16i8.p0(<16 x i8> [[TMP2]], <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i8> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v16i8.p0(<16 x i8> %A, <16 x i8> %B, <16 x i8> %C, <16 x i8> %D, ptr %a)
ret void
}
define void @st4_v1f64(<1 x double> %A, <1 x double> %B, <1 x double> %C, <1 x double> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v1f64(
+; CHECK-SAME: <1 x double> [[A:%.*]], <1 x double> [[B:%.*]], <1 x double> [[C:%.*]], <1 x double> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v1f64.p0(<1 x double> [[A]], <1 x double> [[B]], <1 x double> [[C]], <1 x double> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v1f64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], <1 x i64> [[TMP4]], <1 x i64> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v1f64.p0(<1 x double> %A, <1 x double> %B, <1 x double> %C, <1 x double> %D, ptr %a)
ret void
}
define void @st4_v1i64(<1 x i64> %A, <1 x i64> %B, <1 x i64> %C, <1 x i64> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v1i64(
+; CHECK-SAME: <1 x i64> [[A:%.*]], <1 x i64> [[B:%.*]], <1 x i64> [[C:%.*]], <1 x i64> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <1 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <1 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v1i64.p0(<1 x i64> [[A]], <1 x i64> [[B]], <1 x i64> [[C]], <1 x i64> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v1i64.p0(<1 x i64> [[TMP2]], <1 x i64> [[TMP3]], <1 x i64> [[TMP4]], <1 x i64> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v1i64.p0(<1 x i64> %A, <1 x i64> %B, <1 x i64> %C, <1 x i64> %D, ptr %a)
ret void
}
define void @st4_v2f32(<2 x float> %A, <2 x float> %B, <2 x float> %C, <2 x float> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v2f32(
+; CHECK-SAME: <2 x float> [[A:%.*]], <2 x float> [[B:%.*]], <2 x float> [[C:%.*]], <2 x float> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v2f32.p0(<2 x float> [[A]], <2 x float> [[B]], <2 x float> [[C]], <2 x float> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v2f32.p0(<2 x i32> [[TMP2]], <2 x i32> [[TMP3]], <2 x i32> [[TMP4]], <2 x i32> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v2f32.p0(<2 x float> %A, <2 x float> %B, <2 x float> %C, <2 x float> %D, ptr %a)
ret void
}
define void @st4_v2f64(<2 x double> %A, <2 x double> %B, <2 x double> %C, <2 x double> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v2f64(
+; CHECK-SAME: <2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], <2 x double> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v2f64.p0(<2 x double> [[A]], <2 x double> [[B]], <2 x double> [[C]], <2 x double> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v2f64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], <2 x i64> [[TMP4]], <2 x i64> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v2f64.p0(<2 x double> %A, <2 x double> %B, <2 x double> %C, <2 x double> %D, ptr %a)
ret void
}
define void @st4_v2i32(<2 x i32> %A, <2 x i32> %B, <2 x i32> %C, <2 x i32> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v2i32(
+; CHECK-SAME: <2 x i32> [[A:%.*]], <2 x i32> [[B:%.*]], <2 x i32> [[C:%.*]], <2 x i32> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <2 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v2i32.p0(<2 x i32> [[A]], <2 x i32> [[B]], <2 x i32> [[C]], <2 x i32> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v2i32.p0(<2 x i32> [[TMP2]], <2 x i32> [[TMP3]], <2 x i32> [[TMP4]], <2 x i32> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v2i32.p0(<2 x i32> %A, <2 x i32> %B, <2 x i32> %C, <2 x i32> %D, ptr %a)
ret void
}
define void @st4_v2i64(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C, <2 x i64> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v2i64(
+; CHECK-SAME: <2 x i64> [[A:%.*]], <2 x i64> [[B:%.*]], <2 x i64> [[C:%.*]], <2 x i64> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <2 x i64>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <2 x i64>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v2i64.p0(<2 x i64> [[A]], <2 x i64> [[B]], <2 x i64> [[C]], <2 x i64> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v2i64.p0(<2 x i64> [[TMP2]], <2 x i64> [[TMP3]], <2 x i64> [[TMP4]], <2 x i64> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v2i64.p0(<2 x i64> %A, <2 x i64> %B, <2 x i64> %C, <2 x i64> %D, ptr %a)
ret void
}
define void @st4_v4f16(<4 x half> %A, <4 x half> %B, <4 x half> %C, <4 x half> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v4f16(
+; CHECK-SAME: <4 x half> [[A:%.*]], <4 x half> [[B:%.*]], <4 x half> [[C:%.*]], <4 x half> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v4f16.p0(<4 x half> [[A]], <4 x half> [[B]], <4 x half> [[C]], <4 x half> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v4f16.p0(<4 x i16> [[TMP2]], <4 x i16> [[TMP3]], <4 x i16> [[TMP4]], <4 x i16> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v4f16.p0(<4 x half> %A, <4 x half> %B, <4 x half> %C, <4 x half> %D, ptr %a)
ret void
}
define void @st4_v4f32(<4 x float> %A, <4 x float> %B, <4 x float> %C, <4 x float> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v4f32(
+; CHECK-SAME: <4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> [[C:%.*]], <4 x float> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v4f32.p0(<4 x float> [[A]], <4 x float> [[B]], <4 x float> [[C]], <4 x float> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v4f32.p0(<4 x i32> [[TMP2]], <4 x i32> [[TMP3]], <4 x i32> [[TMP4]], <4 x i32> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v4f32.p0(<4 x float> %A, <4 x float> %B, <4 x float> %C, <4 x float> %D, ptr %a)
ret void
}
define void @st4_v4i16(<4 x i16> %A, <4 x i16> %B, <4 x i16> %C, <4 x i16> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v4i16(
+; CHECK-SAME: <4 x i16> [[A:%.*]], <4 x i16> [[B:%.*]], <4 x i16> [[C:%.*]], <4 x i16> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v4i16.p0(<4 x i16> [[A]], <4 x i16> [[B]], <4 x i16> [[C]], <4 x i16> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v4i16.p0(<4 x i16> [[TMP2]], <4 x i16> [[TMP3]], <4 x i16> [[TMP4]], <4 x i16> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v4i16.p0(<4 x i16> %A, <4 x i16> %B, <4 x i16> %C, <4 x i16> %D, ptr %a)
ret void
}
define void @st4_v4i32(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C, <4 x i32> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v4i32(
+; CHECK-SAME: <4 x i32> [[A:%.*]], <4 x i32> [[B:%.*]], <4 x i32> [[C:%.*]], <4 x i32> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <4 x i32>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <4 x i32>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v4i32.p0(<4 x i32> [[A]], <4 x i32> [[B]], <4 x i32> [[C]], <4 x i32> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v4i32.p0(<4 x i32> [[TMP2]], <4 x i32> [[TMP3]], <4 x i32> [[TMP4]], <4 x i32> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v4i32.p0(<4 x i32> %A, <4 x i32> %B, <4 x i32> %C, <4 x i32> %D, ptr %a)
ret void
}
define void @st4_v8f16(<8 x half> %A, <8 x half> %B, <8 x half> %C, <8 x half> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v8f16(
+; CHECK-SAME: <8 x half> [[A:%.*]], <8 x half> [[B:%.*]], <8 x half> [[C:%.*]], <8 x half> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v8f16.p0(<8 x half> [[A]], <8 x half> [[B]], <8 x half> [[C]], <8 x half> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v8f16.p0(<8 x i16> [[TMP2]], <8 x i16> [[TMP3]], <8 x i16> [[TMP4]], <8 x i16> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v8f16.p0(<8 x half> %A, <8 x half> %B, <8 x half> %C, <8 x half> %D, ptr %a)
ret void
}
define void @st4_v8i16(<8 x i16> %A, <8 x i16> %B, <8 x i16> %C, <8 x i16> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v8i16(
+; CHECK-SAME: <8 x i16> [[A:%.*]], <8 x i16> [[B:%.*]], <8 x i16> [[C:%.*]], <8 x i16> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 64) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i16>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <8 x i16>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 48) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v8i16.p0(<8 x i16> [[A]], <8 x i16> [[B]], <8 x i16> [[C]], <8 x i16> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v8i16.p0(<8 x i16> [[TMP2]], <8 x i16> [[TMP3]], <8 x i16> [[TMP4]], <8 x i16> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v8i16.p0(<8 x i16> %A, <8 x i16> %B, <8 x i16> %C, <8 x i16> %D, ptr %a)
ret void
}
define void @st4_v8i8(<8 x i8> %A, <8 x i8> %B, <8 x i8> %C, <8 x i8> %D, ptr %a) sanitize_memory {
+; CHECK-LABEL: define void @st4_v8i8(
+; CHECK-SAME: <8 x i8> [[A:%.*]], <8 x i8> [[B:%.*]], <8 x i8> [[C:%.*]], <8 x i8> [[D:%.*]], ptr [[A:%.*]]) #[[ATTR0]] {
+; CHECK-NEXT: [[TMP1:%.*]] = load i64, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 32) to ptr), align 8
+; CHECK-NEXT: [[TMP2:%.*]] = load <8 x i8>, ptr @__msan_param_tls, align 8
+; CHECK-NEXT: [[TMP3:%.*]] = load <8 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 8) to ptr), align 8
+; CHECK-NEXT: [[TMP4:%.*]] = load <8 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 16) to ptr), align 8
+; CHECK-NEXT: [[TMP5:%.*]] = load <8 x i8>, ptr inttoptr (i64 add (i64 ptrtoint (ptr @__msan_param_tls to i64), i64 24) to ptr), align 8
+; CHECK-NEXT: call void @llvm.donothing()
+; CHECK-NEXT: [[TMP6:%.*]] = ptrtoint ptr [[A]] to i64
+; CHECK-NEXT: [[TMP7:%.*]] = xor i64 [[TMP6]], 193514046488576
+; CHECK-NEXT: [[TMP8:%.*]] = inttoptr i64 [[TMP7]] to ptr
+; CHECK-NEXT: [[_MSCMP:%.*]] = icmp ne i64 [[TMP1]], 0
+; CHECK-NEXT: br i1 [[_MSCMP]], label [[TMP9:%.*]], label [[TMP10:%.*]], !prof [[PROF0]]
+; CHECK: 9:
+; CHECK-NEXT: call void @__msan_warning_noreturn() #[[ATTR4]]
+; CHECK-NEXT: unreachable
+; CHECK: 10:
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v8i8.p0(<8 x i8> [[A]], <8 x i8> [[B]], <8 x i8> [[C]], <8 x i8> [[D]], ptr [[A]])
+; CHECK-NEXT: call void @llvm.aarch64.neon.st4.v8i8.p0(<8 x i8> [[TMP2]], <8 x i8> [[TMP3]], <8 x i8> [[TMP4]], <8 x i8> [[TMP5]], ptr [[TMP8]])
+; CHECK-NEXT: ret void
+;
call void @llvm.aarch64.neon.st4.v8i8.p0(<8 x i8> %A, <8 x i8> %B, <8 x i8> %C, <8 x i8> %D, ptr %a)
ret void
}
+;.
+; CHECK: [[PROF0]] = !{!"branch_weights", i32 1, i32 1048575}
+;.
More information about the llvm-commits
mailing list